1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

58

59

60

61

62

63

64

65

66

67

68

69

70

71

72

73

74

75

76

77

78

79

80

81

82

83

84

85

86

87

88

89

90

91

92

93

94

95

96

97

98

99

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

145

146

147

148

149

150

151

152

153

154

155

156

157

158

159

160

161

162

163

164

165

166

167

168

169

170

171

172

173

174

175

176

177

178

179

180

181

182

183

184

185

186

187

188

189

190

191

192

193

194

195

196

197

198

199

200

201

202

203

204

205

206

207

208

209

210

211

212

213

214

215

216

217

218

219

220

221

222

223

224

225

226

227

228

229

230

231

232

233

234

235

236

237

238

239

240

241

242

243

244

245

246

247

248

249

250

251

252

253

254

255

256

257

258

259

260

261

262

263

264

265

266

267

268

269

270

271

272

273

274

275

276

277

278

279

280

281

282

283

284

285

286

287

288

289

290

291

292

293

294

295

296

297

298

299

300

301

302

303

304

305

306

307

308

309

310

311

312

313

314

315

316

317

318

319

320

321

322

323

324

325

326

327

328

329

330

331

332

333

334

335

336

337

338

339

340

341

342

343

344

345

346

347

348

349

350

351

352

353

354

355

356

357

358

359

360

361

362

363

364

365

366

367

368

369

370

371

372

373

374

375

376

377

378

379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

400

401

402

403

404

405

406

407

408

409

410

411

412

413

414

415

416

417

418

419

420

421

422

423

424

425

426

427

428

429

430

431

432

433

434

435

436

437

438

439

440

441

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

464

465

466

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

486

487

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

505

506

507

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

552

553

554

555

556

557

558

559

560

561

562

563

564

565

566

567

568

569

570

571

572

573

574

575

576

577

578

579

580

581

582

583

584

585

586

587

588

589

590

591

592

593

594

595

596

597

598

599

600

601

602

603

604

605

606

607

608

609

610

611

612

613

614

615

616

617

618

619

620

621

622

623

624

625

626

627

628

629

630

631

632

633

634

635

636

637

638

639

640

641

642

643

644

645

646

647

648

649

650

651

652

653

654

655

656

657

658

659

660

661

662

663

664

665

666

667

668

669

670

671

672

673

674

675

676

677

678

679

680

681

682

683

684

685

686

687

688

689

690

691

692

693

694

695

696

697

698

699

700

701

702

703

704

705

706

707

708

709

710

711

712

713

714

715

716

717

718

719

720

721

722

723

724

725

726

727

728

729

730

731

732

733

734

735

736

737

738

739

740

741

742

743

744

745

746

747

748

749

750

751

752

753

754

755

756

757

758

759

760

761

762

763

764

765

766

767

768

769

770

771

772

773

774

775

776

777

778

779

780

781

782

783

784

785

786

787

788

789

790

791

792

793

794

795

796

797

798

799

800

801

802

803

804

805

806

807

808

809

810

811

812

813

814

815

816

817

818

819

820

821

822

823

824

825

826

827

828

829

830

831

832

833

834

835

836

837

838

839

840

841

842

843

844

845

846

847

848

849

850

851

852

853

854

855

856

857

858

859

860

861

862

863

864

865

866

867

868

869

870

871

872

873

874

875

876

877

878

879

880

881

882

883

884

885

886

887

888

889

890

891

892

893

894

895

896

897

898

899

900

901

902

903

904

905

906

907

908

909

910

911

912

913

914

915

916

917

918

919

920

921

922

923

924

925

926

927

928

929

930

931

932

933

934

935

936

937

938

939

940

941

942

943

944

945

946

947

948

949

950

951

952

953

954

955

956

957

958

959

960

961

962

963

964

965

966

967

968

969

970

971

972

973

974

975

976

977

978

979

980

981

982

983

984

985

986

987

988

989

990

991

992

993

994

995

996

997

998

999

1000

1001

1002

1003

1004

1005

1006

1007

1008

1009

1010

1011

1012

1013

1014

1015

1016

1017

1018

1019

1020

1021

1022

1023

1024

1025

1026

1027

1028

1029

1030

1031

1032

1033

1034

1035

1036

1037

1038

1039

1040

1041

1042

1043

1044

1045

1046

1047

1048

1049

1050

1051

1052

1053

1054

1055

1056

1057

1058

1059

1060

1061

1062

1063

1064

1065

1066

1067

1068

1069

1070

1071

1072

1073

1074

1075

1076

1077

1078

1079

1080

1081

1082

1083

1084

1085

1086

1087

1088

1089

1090

1091

1092

1093

1094

1095

1096

1097

1098

1099

1100

1101

1102

1103

1104

1105

1106

1107

1108

1109

1110

1111

1112

1113

1114

1115

1116

1117

1118

1119

1120

1121

1122

1123

1124

1125

1126

1127

1128

1129

1130

1131

1132

1133

1134

1135

1136

1137

1138

1139

1140

1141

1142

1143

1144

1145

1146

1147

1148

1149

1150

1151

1152

1153

1154

1155

1156

1157

1158

1159

1160

1161

1162

1163

1164

1165

1166

1167

1168

1169

1170

1171

1172

1173

1174

1175

1176

1177

1178

1179

1180

1181

1182

1183

1184

1185

1186

1187

1188

1189

1190

1191

1192

1193

1194

1195

1196

1197

1198

1199

1200

1201

1202

1203

1204

1205

1206

1207

1208

1209

1210

1211

1212

1213

1214

1215

1216

1217

1218

1219

1220

1221

1222

1223

1224

1225

1226

1227

1228

1229

1230

1231

1232

1233

1234

1235

1236

1237

1238

1239

1240

1241

1242

1243

1244

1245

1246

1247

1248

1249

1250

1251

1252

1253

1254

1255

1256

1257

1258

1259

1260

1261

1262

1263

1264

1265

1266

1267

1268

1269

1270

1271

1272

1273

1274

1275

1276

1277

1278

1279

1280

1281

1282

1283

1284

1285

1286

1287

1288

1289

1290

1291

1292

1293

1294

1295

1296

1297

1298

1299

1300

1301

1302

1303

1304

1305

1306

1307

1308

1309

1310

1311

1312

1313

1314

1315

1316

1317

1318

1319

1320

1321

1322

1323

1324

1325

1326

1327

1328

1329

1330

1331

1332

1333

1334

1335

1336

1337

1338

1339

1340

1341

1342

1343

1344

1345

1346

1347

1348

1349

1350

1351

1352

1353

1354

1355

1356

1357

1358

1359

1360

1361

1362

1363

1364

1365

1366

1367

1368

1369

1370

1371

1372

1373

1374

1375

1376

1377

1378

1379

1380

1381

1382

1383

1384

1385

1386

1387

1388

1389

1390

1391

1392

1393

1394

1395

1396

1397

1398

1399

1400

1401

1402

1403

1404

1405

1406

1407

1408

1409

1410

1411

1412

1413

1414

1415

1416

1417

1418

1419

1420

1421

1422

1423

1424

1425

1426

1427

1428

1429

1430

1431

1432

1433

1434

1435

1436

1437

1438

1439

1440

1441

1442

1443

1444

1445

1446

1447

1448

1449

1450

1451

1452

1453

1454

1455

1456

1457

1458

1459

1460

1461

1462

1463

1464

1465

1466

1467

1468

1469

1470

1471

1472

1473

1474

1475

1476

1477

1478

1479

1480

1481

1482

1483

1484

1485

1486

1487

1488

1489

1490

1491

1492

1493

1494

1495

1496

1497

1498

1499

1500

1501

1502

1503

1504

1505

1506

1507

1508

1509

1510

1511

1512

1513

1514

1515

1516

1517

1518

1519

1520

1521

1522

1523

1524

1525

1526

1527

1528

1529

1530

1531

1532

1533

1534

1535

1536

1537

1538

1539

1540

1541

1542

1543

1544

1545

1546

1547

1548

1549

1550

1551

1552

1553

1554

1555

1556

1557

1558

1559

1560

1561

1562

1563

1564

1565

1566

1567

1568

1569

1570

1571

1572

1573

1574

1575

1576

1577

1578

1579

1580

1581

1582

1583

1584

1585

1586

1587

1588

1589

1590

1591

1592

1593

1594

1595

1596

1597

1598

1599

1600

1601

1602

1603

1604

1605

1606

1607

1608

1609

1610

1611

1612

1613

1614

1615

1616

1617

1618

1619

1620

1621

1622

1623

1624

1625

1626

1627

1628

1629

1630

1631

1632

1633

1634

1635

1636

1637

1638

1639

1640

1641

1642

1643

1644

1645

1646

1647

1648

1649

1650

1651

1652

1653

1654

1655

1656

1657

1658

1659

1660

1661

1662

1663

1664

1665

1666

1667

1668

1669

1670

1671

1672

1673

1674

1675

1676

1677

1678

1679

1680

1681

1682

1683

1684

1685

1686

1687

1688

1689

1690

1691

1692

1693

1694

1695

1696

1697

1698

1699

1700

1701

1702

1703

1704

1705

1706

1707

1708

1709

1710

1711

1712

1713

1714

1715

1716

1717

1718

1719

1720

1721

1722

1723

1724

1725

1726

1727

1728

1729

1730

1731

1732

1733

1734

1735

1736

1737

1738

1739

1740

1741

1742

1743

1744

1745

1746

1747

1748

1749

1750

1751

1752

1753

1754

1755

1756

1757

1758

1759

1760

1761

1762

1763

1764

1765

1766

1767

1768

1769

1770

1771

1772

1773

1774

1775

1776

1777

1778

1779

1780

1781

1782

1783

1784

1785

1786

1787

1788

1789

1790

1791

1792

1793

1794

1795

1796

1797

1798

1799

1800

1801

1802

1803

1804

1805

1806

1807

1808

1809

1810

1811

1812

1813

1814

1815

1816

1817

1818

1819

1820

1821

1822

1823

1824

1825

1826

1827

1828

1829

1830

1831

1832

1833

1834

1835

1836

1837

1838

1839

1840

1841

1842

1843

1844

1845

1846

1847

1848

1849

1850

1851

1852

1853

1854

1855

1856

1857

1858

1859

1860

1861

1862

1863

1864

1865

1866

1867

1868

1869

1870

1871

1872

1873

1874

1875

1876

1877

1878

1879

1880

1881

1882

1883

1884

1885

1886

1887

1888

1889

1890

1891

1892

1893

1894

1895

1896

1897

1898

1899

1900

1901

1902

1903

1904

1905

1906

1907

1908

1909

1910

1911

1912

1913

1914

1915

1916

1917

1918

1919

1920

1921

1922

1923

1924

1925

1926

1927

1928

1929

1930

1931

1932

1933

1934

1935

1936

1937

1938

1939

1940

1941

1942

1943

1944

1945

1946

1947

1948

1949

1950

1951

1952

1953

1954

1955

1956

1957

1958

1959

1960

1961

1962

1963

1964

1965

1966

1967

1968

1969

1970

1971

1972

1973

1974

1975

1976

1977

1978

1979

1980

1981

1982

1983

1984

1985

1986

1987

1988

1989

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

2013

2014

2015

2016

2017

2018

2019

2020

2021

2022

2023

2024

2025

2026

2027

2028

2029

2030

2031

2032

2033

2034

2035

2036

2037

2038

2039

2040

2041

2042

2043

2044

2045

2046

2047

2048

2049

2050

2051

2052

2053

2054

2055

2056

2057

2058

2059

2060

2061

2062

2063

2064

2065

2066

2067

2068

2069

2070

2071

2072

2073

2074

2075

2076

2077

2078

2079

2080

2081

2082

2083

2084

2085

2086

2087

2088

2089

2090

2091

2092

2093

2094

2095

2096

2097

2098

2099

2100

2101

2102

2103

2104

2105

2106

2107

2108

2109

2110

2111

2112

2113

2114

2115

2116

2117

2118

2119

2120

2121

2122

2123

2124

2125

2126

2127

2128

2129

2130

2131

2132

2133

2134

2135

2136

2137

2138

2139

2140

2141

2142

2143

2144

2145

2146

2147

2148

2149

2150

2151

2152

2153

2154

2155

2156

2157

2158

2159

2160

2161

2162

2163

2164

2165

2166

2167

2168

2169

2170

2171

2172

2173

2174

2175

2176

2177

2178

2179

2180

2181

2182

2183

2184

2185

2186

2187

2188

2189

2190

2191

2192

2193

2194

2195

2196

2197

2198

2199

2200

2201

2202

2203

2204

2205

2206

2207

2208

2209

2210

2211

2212

2213

2214

2215

2216

2217

2218

2219

2220

2221

2222

2223

# Copyright (C) 2003-2005 Peter J. Verveer 

# 

# Redistribution and use in source and binary forms, with or without 

# modification, are permitted provided that the following conditions 

# are met: 

# 

# 1. Redistributions of source code must retain the above copyright 

# notice, this list of conditions and the following disclaimer. 

# 

# 2. Redistributions in binary form must reproduce the above 

# copyright notice, this list of conditions and the following 

# disclaimer in the documentation and/or other materials provided 

# with the distribution. 

# 

# 3. The name of the author may not be used to endorse or promote 

# products derived from this software without specific prior 

# written permission. 

# 

# THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS 

# OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED 

# WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 

# ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY 

# DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 

# DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE 

# GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 

# INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, 

# WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING 

# NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS 

# SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. 

 

from __future__ import division, print_function, absolute_import 

import warnings 

 

import numpy 

from . import _ni_support 

from . import _nd_image 

from . import filters 

 

__all__ = ['iterate_structure', 'generate_binary_structure', 'binary_erosion', 

'binary_dilation', 'binary_opening', 'binary_closing', 

'binary_hit_or_miss', 'binary_propagation', 'binary_fill_holes', 

'grey_erosion', 'grey_dilation', 'grey_opening', 'grey_closing', 

'morphological_gradient', 'morphological_laplace', 'white_tophat', 

'black_tophat', 'distance_transform_bf', 'distance_transform_cdt', 

'distance_transform_edt'] 

 

 

def _center_is_true(structure, origin): 

structure = numpy.array(structure) 

coor = tuple([oo + ss // 2 for ss, oo in zip(structure.shape, 

origin)]) 

return bool(structure[coor]) 

 

 

def iterate_structure(structure, iterations, origin=None): 

""" 

Iterate a structure by dilating it with itself. 

 

Parameters 

---------- 

structure : array_like 

Structuring element (an array of bools, for example), to be dilated with 

itself. 

iterations : int 

number of dilations performed on the structure with itself 

origin : optional 

If origin is None, only the iterated structure is returned. If 

not, a tuple of the iterated structure and the modified origin is 

returned. 

 

Returns 

------- 

iterate_structure : ndarray of bools 

A new structuring element obtained by dilating `structure` 

(`iterations` - 1) times with itself. 

 

See also 

-------- 

generate_binary_structure 

 

Examples 

-------- 

>>> from scipy import ndimage 

>>> struct = ndimage.generate_binary_structure(2, 1) 

>>> struct.astype(int) 

array([[0, 1, 0], 

[1, 1, 1], 

[0, 1, 0]]) 

>>> ndimage.iterate_structure(struct, 2).astype(int) 

array([[0, 0, 1, 0, 0], 

[0, 1, 1, 1, 0], 

[1, 1, 1, 1, 1], 

[0, 1, 1, 1, 0], 

[0, 0, 1, 0, 0]]) 

>>> ndimage.iterate_structure(struct, 3).astype(int) 

array([[0, 0, 0, 1, 0, 0, 0], 

[0, 0, 1, 1, 1, 0, 0], 

[0, 1, 1, 1, 1, 1, 0], 

[1, 1, 1, 1, 1, 1, 1], 

[0, 1, 1, 1, 1, 1, 0], 

[0, 0, 1, 1, 1, 0, 0], 

[0, 0, 0, 1, 0, 0, 0]]) 

 

""" 

structure = numpy.asarray(structure) 

if iterations < 2: 

return structure.copy() 

ni = iterations - 1 

shape = [ii + ni * (ii - 1) for ii in structure.shape] 

pos = [ni * (structure.shape[ii] // 2) for ii in range(len(shape))] 

slc = [slice(pos[ii], pos[ii] + structure.shape[ii], None) 

for ii in range(len(shape))] 

out = numpy.zeros(shape, bool) 

out[slc] = structure != 0 

out = binary_dilation(out, structure, iterations=ni) 

if origin is None: 

return out 

else: 

origin = _ni_support._normalize_sequence(origin, structure.ndim) 

origin = [iterations * o for o in origin] 

return out, origin 

 

 

def generate_binary_structure(rank, connectivity): 

""" 

Generate a binary structure for binary morphological operations. 

 

Parameters 

---------- 

rank : int 

Number of dimensions of the array to which the structuring element 

will be applied, as returned by `np.ndim`. 

connectivity : int 

`connectivity` determines which elements of the output array belong 

to the structure, i.e. are considered as neighbors of the central 

element. Elements up to a squared distance of `connectivity` from 

the center are considered neighbors. `connectivity` may range from 1 

(no diagonal elements are neighbors) to `rank` (all elements are 

neighbors). 

 

Returns 

------- 

output : ndarray of bools 

Structuring element which may be used for binary morphological 

operations, with `rank` dimensions and all dimensions equal to 3. 

 

See also 

-------- 

iterate_structure, binary_dilation, binary_erosion 

 

Notes 

----- 

`generate_binary_structure` can only create structuring elements with 

dimensions equal to 3, i.e. minimal dimensions. For larger structuring 

elements, that are useful e.g. for eroding large objects, one may either 

use `iterate_structure`, or create directly custom arrays with 

numpy functions such as `numpy.ones`. 

 

Examples 

-------- 

>>> from scipy import ndimage 

>>> struct = ndimage.generate_binary_structure(2, 1) 

>>> struct 

array([[False, True, False], 

[ True, True, True], 

[False, True, False]], dtype=bool) 

>>> a = np.zeros((5,5)) 

>>> a[2, 2] = 1 

>>> a 

array([[ 0., 0., 0., 0., 0.], 

[ 0., 0., 0., 0., 0.], 

[ 0., 0., 1., 0., 0.], 

[ 0., 0., 0., 0., 0.], 

[ 0., 0., 0., 0., 0.]]) 

>>> b = ndimage.binary_dilation(a, structure=struct).astype(a.dtype) 

>>> b 

array([[ 0., 0., 0., 0., 0.], 

[ 0., 0., 1., 0., 0.], 

[ 0., 1., 1., 1., 0.], 

[ 0., 0., 1., 0., 0.], 

[ 0., 0., 0., 0., 0.]]) 

>>> ndimage.binary_dilation(b, structure=struct).astype(a.dtype) 

array([[ 0., 0., 1., 0., 0.], 

[ 0., 1., 1., 1., 0.], 

[ 1., 1., 1., 1., 1.], 

[ 0., 1., 1., 1., 0.], 

[ 0., 0., 1., 0., 0.]]) 

>>> struct = ndimage.generate_binary_structure(2, 2) 

>>> struct 

array([[ True, True, True], 

[ True, True, True], 

[ True, True, True]], dtype=bool) 

>>> struct = ndimage.generate_binary_structure(3, 1) 

>>> struct # no diagonal elements 

array([[[False, False, False], 

[False, True, False], 

[False, False, False]], 

[[False, True, False], 

[ True, True, True], 

[False, True, False]], 

[[False, False, False], 

[False, True, False], 

[False, False, False]]], dtype=bool) 

 

""" 

if connectivity < 1: 

connectivity = 1 

if rank < 1: 

return numpy.array(True, dtype=bool) 

output = numpy.fabs(numpy.indices([3] * rank) - 1) 

output = numpy.add.reduce(output, 0) 

return output <= connectivity 

 

 

def _binary_erosion(input, structure, iterations, mask, output, 

border_value, origin, invert, brute_force): 

input = numpy.asarray(input) 

if numpy.iscomplexobj(input): 

raise TypeError('Complex type not supported') 

if structure is None: 

structure = generate_binary_structure(input.ndim, 1) 

else: 

structure = numpy.asarray(structure, dtype=bool) 

if structure.ndim != input.ndim: 

raise RuntimeError('structure and input must have same dimensionality') 

if not structure.flags.contiguous: 

structure = structure.copy() 

if numpy.product(structure.shape, axis=0) < 1: 

raise RuntimeError('structure must not be empty') 

if mask is not None: 

mask = numpy.asarray(mask) 

if mask.shape != input.shape: 

raise RuntimeError('mask and input must have equal sizes') 

origin = _ni_support._normalize_sequence(origin, input.ndim) 

cit = _center_is_true(structure, origin) 

if isinstance(output, numpy.ndarray): 

if numpy.iscomplexobj(output): 

raise TypeError('Complex output type not supported') 

else: 

output = bool 

output = _ni_support._get_output(output, input) 

 

if iterations == 1: 

_nd_image.binary_erosion(input, structure, mask, output, 

border_value, origin, invert, cit, 0) 

return output 

elif cit and not brute_force: 

changed, coordinate_list = _nd_image.binary_erosion( 

input, structure, mask, output, 

border_value, origin, invert, cit, 1) 

structure = structure[tuple([slice(None, None, -1)] * 

structure.ndim)] 

for ii in range(len(origin)): 

origin[ii] = -origin[ii] 

if not structure.shape[ii] & 1: 

origin[ii] -= 1 

if mask is not None: 

mask = numpy.asarray(mask, dtype=numpy.int8) 

if not structure.flags.contiguous: 

structure = structure.copy() 

_nd_image.binary_erosion2(output, structure, mask, iterations - 1, 

origin, invert, coordinate_list) 

return output 

else: 

tmp_in = numpy.empty_like(input, dtype=bool) 

tmp_out = output 

if iterations >= 1 and not iterations & 1: 

tmp_in, tmp_out = tmp_out, tmp_in 

changed = _nd_image.binary_erosion( 

input, structure, mask, tmp_out, 

border_value, origin, invert, cit, 0) 

ii = 1 

while ii < iterations or (iterations < 1 and changed): 

tmp_in, tmp_out = tmp_out, tmp_in 

changed = _nd_image.binary_erosion( 

tmp_in, structure, mask, tmp_out, 

border_value, origin, invert, cit, 0) 

ii += 1 

return output 

 

 

def binary_erosion(input, structure=None, iterations=1, mask=None, output=None, 

border_value=0, origin=0, brute_force=False): 

""" 

Multi-dimensional binary erosion with a given structuring element. 

 

Binary erosion is a mathematical morphology operation used for image 

processing. 

 

Parameters 

---------- 

input : array_like 

Binary image to be eroded. Non-zero (True) elements form 

the subset to be eroded. 

structure : array_like, optional 

Structuring element used for the erosion. Non-zero elements are 

considered True. If no structuring element is provided, an element 

is generated with a square connectivity equal to one. 

iterations : {int, float}, optional 

The erosion is repeated `iterations` times (one, by default). 

If iterations is less than 1, the erosion is repeated until the 

result does not change anymore. 

mask : array_like, optional 

If a mask is given, only those elements with a True value at 

the corresponding mask element are modified at each iteration. 

output : ndarray, optional 

Array of the same shape as input, into which the output is placed. 

By default, a new array is created. 

border_value : int (cast to 0 or 1), optional 

Value at the border in the output array. 

origin : int or tuple of ints, optional 

Placement of the filter, by default 0. 

brute_force : boolean, optional 

Memory condition: if False, only the pixels whose value was changed in 

the last iteration are tracked as candidates to be updated (eroded) in 

the current iteration; if True all pixels are considered as candidates 

for erosion, regardless of what happened in the previous iteration. 

False by default. 

 

Returns 

------- 

binary_erosion : ndarray of bools 

Erosion of the input by the structuring element. 

 

See also 

-------- 

grey_erosion, binary_dilation, binary_closing, binary_opening, 

generate_binary_structure 

 

Notes 

----- 

Erosion [1]_ is a mathematical morphology operation [2]_ that uses a 

structuring element for shrinking the shapes in an image. The binary 

erosion of an image by a structuring element is the locus of the points 

where a superimposition of the structuring element centered on the point 

is entirely contained in the set of non-zero elements of the image. 

 

References 

---------- 

.. [1] http://en.wikipedia.org/wiki/Erosion_%28morphology%29 

.. [2] http://en.wikipedia.org/wiki/Mathematical_morphology 

 

Examples 

-------- 

>>> from scipy import ndimage 

>>> a = np.zeros((7,7), dtype=int) 

>>> a[1:6, 2:5] = 1 

>>> a 

array([[0, 0, 0, 0, 0, 0, 0], 

[0, 0, 1, 1, 1, 0, 0], 

[0, 0, 1, 1, 1, 0, 0], 

[0, 0, 1, 1, 1, 0, 0], 

[0, 0, 1, 1, 1, 0, 0], 

[0, 0, 1, 1, 1, 0, 0], 

[0, 0, 0, 0, 0, 0, 0]]) 

>>> ndimage.binary_erosion(a).astype(a.dtype) 

array([[0, 0, 0, 0, 0, 0, 0], 

[0, 0, 0, 0, 0, 0, 0], 

[0, 0, 0, 1, 0, 0, 0], 

[0, 0, 0, 1, 0, 0, 0], 

[0, 0, 0, 1, 0, 0, 0], 

[0, 0, 0, 0, 0, 0, 0], 

[0, 0, 0, 0, 0, 0, 0]]) 

>>> #Erosion removes objects smaller than the structure 

>>> ndimage.binary_erosion(a, structure=np.ones((5,5))).astype(a.dtype) 

array([[0, 0, 0, 0, 0, 0, 0], 

[0, 0, 0, 0, 0, 0, 0], 

[0, 0, 0, 0, 0, 0, 0], 

[0, 0, 0, 0, 0, 0, 0], 

[0, 0, 0, 0, 0, 0, 0], 

[0, 0, 0, 0, 0, 0, 0], 

[0, 0, 0, 0, 0, 0, 0]]) 

 

""" 

return _binary_erosion(input, structure, iterations, mask, 

output, border_value, origin, 0, brute_force) 

 

 

def binary_dilation(input, structure=None, iterations=1, mask=None, 

output=None, border_value=0, origin=0, 

brute_force=False): 

""" 

Multi-dimensional binary dilation with the given structuring element. 

 

Parameters 

---------- 

input : array_like 

Binary array_like to be dilated. Non-zero (True) elements form 

the subset to be dilated. 

structure : array_like, optional 

Structuring element used for the dilation. Non-zero elements are 

considered True. If no structuring element is provided an element 

is generated with a square connectivity equal to one. 

iterations : {int, float}, optional 

The dilation is repeated `iterations` times (one, by default). 

If iterations is less than 1, the dilation is repeated until the 

result does not change anymore. 

mask : array_like, optional 

If a mask is given, only those elements with a True value at 

the corresponding mask element are modified at each iteration. 

output : ndarray, optional 

Array of the same shape as input, into which the output is placed. 

By default, a new array is created. 

border_value : int (cast to 0 or 1), optional 

Value at the border in the output array. 

origin : int or tuple of ints, optional 

Placement of the filter, by default 0. 

brute_force : boolean, optional 

Memory condition: if False, only the pixels whose value was changed in 

the last iteration are tracked as candidates to be updated (dilated) 

in the current iteration; if True all pixels are considered as 

candidates for dilation, regardless of what happened in the previous 

iteration. False by default. 

 

Returns 

------- 

binary_dilation : ndarray of bools 

Dilation of the input by the structuring element. 

 

See also 

-------- 

grey_dilation, binary_erosion, binary_closing, binary_opening, 

generate_binary_structure 

 

Notes 

----- 

Dilation [1]_ is a mathematical morphology operation [2]_ that uses a 

structuring element for expanding the shapes in an image. The binary 

dilation of an image by a structuring element is the locus of the points 

covered by the structuring element, when its center lies within the 

non-zero points of the image. 

 

References 

---------- 

.. [1] http://en.wikipedia.org/wiki/Dilation_%28morphology%29 

.. [2] http://en.wikipedia.org/wiki/Mathematical_morphology 

 

Examples 

-------- 

>>> from scipy import ndimage 

>>> a = np.zeros((5, 5)) 

>>> a[2, 2] = 1 

>>> a 

array([[ 0., 0., 0., 0., 0.], 

[ 0., 0., 0., 0., 0.], 

[ 0., 0., 1., 0., 0.], 

[ 0., 0., 0., 0., 0.], 

[ 0., 0., 0., 0., 0.]]) 

>>> ndimage.binary_dilation(a) 

array([[False, False, False, False, False], 

[False, False, True, False, False], 

[False, True, True, True, False], 

[False, False, True, False, False], 

[False, False, False, False, False]], dtype=bool) 

>>> ndimage.binary_dilation(a).astype(a.dtype) 

array([[ 0., 0., 0., 0., 0.], 

[ 0., 0., 1., 0., 0.], 

[ 0., 1., 1., 1., 0.], 

[ 0., 0., 1., 0., 0.], 

[ 0., 0., 0., 0., 0.]]) 

>>> # 3x3 structuring element with connectivity 1, used by default 

>>> struct1 = ndimage.generate_binary_structure(2, 1) 

>>> struct1 

array([[False, True, False], 

[ True, True, True], 

[False, True, False]], dtype=bool) 

>>> # 3x3 structuring element with connectivity 2 

>>> struct2 = ndimage.generate_binary_structure(2, 2) 

>>> struct2 

array([[ True, True, True], 

[ True, True, True], 

[ True, True, True]], dtype=bool) 

>>> ndimage.binary_dilation(a, structure=struct1).astype(a.dtype) 

array([[ 0., 0., 0., 0., 0.], 

[ 0., 0., 1., 0., 0.], 

[ 0., 1., 1., 1., 0.], 

[ 0., 0., 1., 0., 0.], 

[ 0., 0., 0., 0., 0.]]) 

>>> ndimage.binary_dilation(a, structure=struct2).astype(a.dtype) 

array([[ 0., 0., 0., 0., 0.], 

[ 0., 1., 1., 1., 0.], 

[ 0., 1., 1., 1., 0.], 

[ 0., 1., 1., 1., 0.], 

[ 0., 0., 0., 0., 0.]]) 

>>> ndimage.binary_dilation(a, structure=struct1,\\ 

... iterations=2).astype(a.dtype) 

array([[ 0., 0., 1., 0., 0.], 

[ 0., 1., 1., 1., 0.], 

[ 1., 1., 1., 1., 1.], 

[ 0., 1., 1., 1., 0.], 

[ 0., 0., 1., 0., 0.]]) 

 

""" 

input = numpy.asarray(input) 

if structure is None: 

structure = generate_binary_structure(input.ndim, 1) 

origin = _ni_support._normalize_sequence(origin, input.ndim) 

structure = numpy.asarray(structure) 

structure = structure[tuple([slice(None, None, -1)] * 

structure.ndim)] 

for ii in range(len(origin)): 

origin[ii] = -origin[ii] 

if not structure.shape[ii] & 1: 

origin[ii] -= 1 

 

return _binary_erosion(input, structure, iterations, mask, 

output, border_value, origin, 1, brute_force) 

 

 

def binary_opening(input, structure=None, iterations=1, output=None, 

origin=0, mask=None, border_value=0, brute_force=False): 

""" 

Multi-dimensional binary opening with the given structuring element. 

 

The *opening* of an input image by a structuring element is the 

*dilation* of the *erosion* of the image by the structuring element. 

 

Parameters 

---------- 

input : array_like 

Binary array_like to be opened. Non-zero (True) elements form 

the subset to be opened. 

structure : array_like, optional 

Structuring element used for the opening. Non-zero elements are 

considered True. If no structuring element is provided an element 

is generated with a square connectivity equal to one (i.e., only 

nearest neighbors are connected to the center, diagonally-connected 

elements are not considered neighbors). 

iterations : {int, float}, optional 

The erosion step of the opening, then the dilation step are each 

repeated `iterations` times (one, by default). If `iterations` is 

less than 1, each operation is repeated until the result does 

not change anymore. 

output : ndarray, optional 

Array of the same shape as input, into which the output is placed. 

By default, a new array is created. 

origin : int or tuple of ints, optional 

Placement of the filter, by default 0. 

mask : array_like, optional 

If a mask is given, only those elements with a True value at 

the corresponding mask element are modified at each iteration. 

 

.. versionadded:: 1.1.0 

border_value : int (cast to 0 or 1), optional 

Value at the border in the output array. 

 

.. versionadded:: 1.1.0 

brute_force : boolean, optional 

Memory condition: if False, only the pixels whose value was changed in 

the last iteration are tracked as candidates to be updated in the 

current iteration; if true all pixels are considered as candidates for 

update, regardless of what happened in the previous iteration. 

False by default. 

 

.. versionadded:: 1.1.0 

 

Returns 

------- 

binary_opening : ndarray of bools 

Opening of the input by the structuring element. 

 

See also 

-------- 

grey_opening, binary_closing, binary_erosion, binary_dilation, 

generate_binary_structure 

 

Notes 

----- 

*Opening* [1]_ is a mathematical morphology operation [2]_ that 

consists in the succession of an erosion and a dilation of the 

input with the same structuring element. Opening therefore removes 

objects smaller than the structuring element. 

 

Together with *closing* (`binary_closing`), opening can be used for 

noise removal. 

 

References 

---------- 

.. [1] http://en.wikipedia.org/wiki/Opening_%28morphology%29 

.. [2] http://en.wikipedia.org/wiki/Mathematical_morphology 

 

Examples 

-------- 

>>> from scipy import ndimage 

>>> a = np.zeros((5,5), dtype=int) 

>>> a[1:4, 1:4] = 1; a[4, 4] = 1 

>>> a 

array([[0, 0, 0, 0, 0], 

[0, 1, 1, 1, 0], 

[0, 1, 1, 1, 0], 

[0, 1, 1, 1, 0], 

[0, 0, 0, 0, 1]]) 

>>> # Opening removes small objects 

>>> ndimage.binary_opening(a, structure=np.ones((3,3))).astype(int) 

array([[0, 0, 0, 0, 0], 

[0, 1, 1, 1, 0], 

[0, 1, 1, 1, 0], 

[0, 1, 1, 1, 0], 

[0, 0, 0, 0, 0]]) 

>>> # Opening can also smooth corners 

>>> ndimage.binary_opening(a).astype(int) 

array([[0, 0, 0, 0, 0], 

[0, 0, 1, 0, 0], 

[0, 1, 1, 1, 0], 

[0, 0, 1, 0, 0], 

[0, 0, 0, 0, 0]]) 

>>> # Opening is the dilation of the erosion of the input 

>>> ndimage.binary_erosion(a).astype(int) 

array([[0, 0, 0, 0, 0], 

[0, 0, 0, 0, 0], 

[0, 0, 1, 0, 0], 

[0, 0, 0, 0, 0], 

[0, 0, 0, 0, 0]]) 

>>> ndimage.binary_dilation(ndimage.binary_erosion(a)).astype(int) 

array([[0, 0, 0, 0, 0], 

[0, 0, 1, 0, 0], 

[0, 1, 1, 1, 0], 

[0, 0, 1, 0, 0], 

[0, 0, 0, 0, 0]]) 

 

""" 

input = numpy.asarray(input) 

if structure is None: 

rank = input.ndim 

structure = generate_binary_structure(rank, 1) 

 

tmp = binary_erosion(input, structure, iterations, mask, None, 

border_value, origin, brute_force) 

return binary_dilation(tmp, structure, iterations, mask, output, 

border_value, origin, brute_force) 

 

 

def binary_closing(input, structure=None, iterations=1, output=None, 

origin=0, mask=None, border_value=0, brute_force=False): 

""" 

Multi-dimensional binary closing with the given structuring element. 

 

The *closing* of an input image by a structuring element is the 

*erosion* of the *dilation* of the image by the structuring element. 

 

Parameters 

---------- 

input : array_like 

Binary array_like to be closed. Non-zero (True) elements form 

the subset to be closed. 

structure : array_like, optional 

Structuring element used for the closing. Non-zero elements are 

considered True. If no structuring element is provided an element 

is generated with a square connectivity equal to one (i.e., only 

nearest neighbors are connected to the center, diagonally-connected 

elements are not considered neighbors). 

iterations : {int, float}, optional 

The dilation step of the closing, then the erosion step are each 

repeated `iterations` times (one, by default). If iterations is 

less than 1, each operations is repeated until the result does 

not change anymore. 

output : ndarray, optional 

Array of the same shape as input, into which the output is placed. 

By default, a new array is created. 

origin : int or tuple of ints, optional 

Placement of the filter, by default 0. 

mask : array_like, optional 

If a mask is given, only those elements with a True value at 

the corresponding mask element are modified at each iteration. 

 

.. versionadded:: 1.1.0 

border_value : int (cast to 0 or 1), optional 

Value at the border in the output array. 

 

.. versionadded:: 1.1.0 

brute_force : boolean, optional 

Memory condition: if False, only the pixels whose value was changed in 

the last iteration are tracked as candidates to be updated in the 

current iteration; if true al pixels are considered as candidates for 

update, regardless of what happened in the previous iteration. 

False by default. 

 

.. versionadded:: 1.1.0 

 

Returns 

------- 

binary_closing : ndarray of bools 

Closing of the input by the structuring element. 

 

See also 

-------- 

grey_closing, binary_opening, binary_dilation, binary_erosion, 

generate_binary_structure 

 

Notes 

----- 

*Closing* [1]_ is a mathematical morphology operation [2]_ that 

consists in the succession of a dilation and an erosion of the 

input with the same structuring element. Closing therefore fills 

holes smaller than the structuring element. 

 

Together with *opening* (`binary_opening`), closing can be used for 

noise removal. 

 

References 

---------- 

.. [1] http://en.wikipedia.org/wiki/Closing_%28morphology%29 

.. [2] http://en.wikipedia.org/wiki/Mathematical_morphology 

 

Examples 

-------- 

>>> from scipy import ndimage 

>>> a = np.zeros((5,5), dtype=int) 

>>> a[1:-1, 1:-1] = 1; a[2,2] = 0 

>>> a 

array([[0, 0, 0, 0, 0], 

[0, 1, 1, 1, 0], 

[0, 1, 0, 1, 0], 

[0, 1, 1, 1, 0], 

[0, 0, 0, 0, 0]]) 

>>> # Closing removes small holes 

>>> ndimage.binary_closing(a).astype(int) 

array([[0, 0, 0, 0, 0], 

[0, 1, 1, 1, 0], 

[0, 1, 1, 1, 0], 

[0, 1, 1, 1, 0], 

[0, 0, 0, 0, 0]]) 

>>> # Closing is the erosion of the dilation of the input 

>>> ndimage.binary_dilation(a).astype(int) 

array([[0, 1, 1, 1, 0], 

[1, 1, 1, 1, 1], 

[1, 1, 1, 1, 1], 

[1, 1, 1, 1, 1], 

[0, 1, 1, 1, 0]]) 

>>> ndimage.binary_erosion(ndimage.binary_dilation(a)).astype(int) 

array([[0, 0, 0, 0, 0], 

[0, 1, 1, 1, 0], 

[0, 1, 1, 1, 0], 

[0, 1, 1, 1, 0], 

[0, 0, 0, 0, 0]]) 

 

 

>>> a = np.zeros((7,7), dtype=int) 

>>> a[1:6, 2:5] = 1; a[1:3,3] = 0 

>>> a 

array([[0, 0, 0, 0, 0, 0, 0], 

[0, 0, 1, 0, 1, 0, 0], 

[0, 0, 1, 0, 1, 0, 0], 

[0, 0, 1, 1, 1, 0, 0], 

[0, 0, 1, 1, 1, 0, 0], 

[0, 0, 1, 1, 1, 0, 0], 

[0, 0, 0, 0, 0, 0, 0]]) 

>>> # In addition to removing holes, closing can also 

>>> # coarsen boundaries with fine hollows. 

>>> ndimage.binary_closing(a).astype(int) 

array([[0, 0, 0, 0, 0, 0, 0], 

[0, 0, 1, 0, 1, 0, 0], 

[0, 0, 1, 1, 1, 0, 0], 

[0, 0, 1, 1, 1, 0, 0], 

[0, 0, 1, 1, 1, 0, 0], 

[0, 0, 1, 1, 1, 0, 0], 

[0, 0, 0, 0, 0, 0, 0]]) 

>>> ndimage.binary_closing(a, structure=np.ones((2,2))).astype(int) 

array([[0, 0, 0, 0, 0, 0, 0], 

[0, 0, 1, 1, 1, 0, 0], 

[0, 0, 1, 1, 1, 0, 0], 

[0, 0, 1, 1, 1, 0, 0], 

[0, 0, 1, 1, 1, 0, 0], 

[0, 0, 1, 1, 1, 0, 0], 

[0, 0, 0, 0, 0, 0, 0]]) 

 

""" 

input = numpy.asarray(input) 

if structure is None: 

rank = input.ndim 

structure = generate_binary_structure(rank, 1) 

 

tmp = binary_dilation(input, structure, iterations, mask, None, 

border_value, origin, brute_force) 

return binary_erosion(tmp, structure, iterations, mask, output, 

border_value, origin, brute_force) 

 

 

def binary_hit_or_miss(input, structure1=None, structure2=None, 

output=None, origin1=0, origin2=None): 

""" 

Multi-dimensional binary hit-or-miss transform. 

 

The hit-or-miss transform finds the locations of a given pattern 

inside the input image. 

 

Parameters 

---------- 

input : array_like (cast to booleans) 

Binary image where a pattern is to be detected. 

structure1 : array_like (cast to booleans), optional 

Part of the structuring element to be fitted to the foreground 

(non-zero elements) of `input`. If no value is provided, a 

structure of square connectivity 1 is chosen. 

structure2 : array_like (cast to booleans), optional 

Second part of the structuring element that has to miss completely 

the foreground. If no value is provided, the complementary of 

`structure1` is taken. 

output : ndarray, optional 

Array of the same shape as input, into which the output is placed. 

By default, a new array is created. 

origin1 : int or tuple of ints, optional 

Placement of the first part of the structuring element `structure1`, 

by default 0 for a centered structure. 

origin2 : int or tuple of ints, optional 

Placement of the second part of the structuring element `structure2`, 

by default 0 for a centered structure. If a value is provided for 

`origin1` and not for `origin2`, then `origin2` is set to `origin1`. 

 

Returns 

------- 

binary_hit_or_miss : ndarray 

Hit-or-miss transform of `input` with the given structuring 

element (`structure1`, `structure2`). 

 

See also 

-------- 

ndimage.morphology, binary_erosion 

 

References 

---------- 

.. [1] http://en.wikipedia.org/wiki/Hit-or-miss_transform 

 

Examples 

-------- 

>>> from scipy import ndimage 

>>> a = np.zeros((7,7), dtype=int) 

>>> a[1, 1] = 1; a[2:4, 2:4] = 1; a[4:6, 4:6] = 1 

>>> a 

array([[0, 0, 0, 0, 0, 0, 0], 

[0, 1, 0, 0, 0, 0, 0], 

[0, 0, 1, 1, 0, 0, 0], 

[0, 0, 1, 1, 0, 0, 0], 

[0, 0, 0, 0, 1, 1, 0], 

[0, 0, 0, 0, 1, 1, 0], 

[0, 0, 0, 0, 0, 0, 0]]) 

>>> structure1 = np.array([[1, 0, 0], [0, 1, 1], [0, 1, 1]]) 

>>> structure1 

array([[1, 0, 0], 

[0, 1, 1], 

[0, 1, 1]]) 

>>> # Find the matches of structure1 in the array a 

>>> ndimage.binary_hit_or_miss(a, structure1=structure1).astype(int) 

array([[0, 0, 0, 0, 0, 0, 0], 

[0, 0, 0, 0, 0, 0, 0], 

[0, 0, 1, 0, 0, 0, 0], 

[0, 0, 0, 0, 0, 0, 0], 

[0, 0, 0, 0, 1, 0, 0], 

[0, 0, 0, 0, 0, 0, 0], 

[0, 0, 0, 0, 0, 0, 0]]) 

>>> # Change the origin of the filter 

>>> # origin1=1 is equivalent to origin1=(1,1) here 

>>> ndimage.binary_hit_or_miss(a, structure1=structure1,\\ 

... origin1=1).astype(int) 

array([[0, 0, 0, 0, 0, 0, 0], 

[0, 0, 0, 0, 0, 0, 0], 

[0, 0, 0, 0, 0, 0, 0], 

[0, 0, 0, 1, 0, 0, 0], 

[0, 0, 0, 0, 0, 0, 0], 

[0, 0, 0, 0, 0, 1, 0], 

[0, 0, 0, 0, 0, 0, 0]]) 

 

""" 

input = numpy.asarray(input) 

if structure1 is None: 

structure1 = generate_binary_structure(input.ndim, 1) 

if structure2 is None: 

structure2 = numpy.logical_not(structure1) 

origin1 = _ni_support._normalize_sequence(origin1, input.ndim) 

if origin2 is None: 

origin2 = origin1 

else: 

origin2 = _ni_support._normalize_sequence(origin2, input.ndim) 

 

tmp1 = _binary_erosion(input, structure1, 1, None, None, 0, origin1, 

0, False) 

inplace = isinstance(output, numpy.ndarray) 

result = _binary_erosion(input, structure2, 1, None, output, 0, 

origin2, 1, False) 

if inplace: 

numpy.logical_not(output, output) 

numpy.logical_and(tmp1, output, output) 

else: 

numpy.logical_not(result, result) 

return numpy.logical_and(tmp1, result) 

 

 

def binary_propagation(input, structure=None, mask=None, 

output=None, border_value=0, origin=0): 

""" 

Multi-dimensional binary propagation with the given structuring element. 

 

Parameters 

---------- 

input : array_like 

Binary image to be propagated inside `mask`. 

structure : array_like, optional 

Structuring element used in the successive dilations. The output 

may depend on the structuring element, especially if `mask` has 

several connex components. If no structuring element is 

provided, an element is generated with a squared connectivity equal 

to one. 

mask : array_like, optional 

Binary mask defining the region into which `input` is allowed to 

propagate. 

output : ndarray, optional 

Array of the same shape as input, into which the output is placed. 

By default, a new array is created. 

border_value : int (cast to 0 or 1), optional 

Value at the border in the output array. 

origin : int or tuple of ints, optional 

Placement of the filter, by default 0. 

 

Returns 

------- 

binary_propagation : ndarray 

Binary propagation of `input` inside `mask`. 

 

Notes 

----- 

This function is functionally equivalent to calling binary_dilation 

with the number of iterations less than one: iterative dilation until 

the result does not change anymore. 

 

The succession of an erosion and propagation inside the original image 

can be used instead of an *opening* for deleting small objects while 

keeping the contours of larger objects untouched. 

 

References 

---------- 

.. [1] http://cmm.ensmp.fr/~serra/cours/pdf/en/ch6en.pdf, slide 15. 

.. [2] I.T. Young, J.J. Gerbrands, and L.J. van Vliet, "Fundamentals of 

image processing", 1998 

ftp://qiftp.tudelft.nl/DIPimage/docs/FIP2.3.pdf 

 

Examples 

-------- 

>>> from scipy import ndimage 

>>> input = np.zeros((8, 8), dtype=int) 

>>> input[2, 2] = 1 

>>> mask = np.zeros((8, 8), dtype=int) 

>>> mask[1:4, 1:4] = mask[4, 4] = mask[6:8, 6:8] = 1 

>>> input 

array([[0, 0, 0, 0, 0, 0, 0, 0], 

[0, 0, 0, 0, 0, 0, 0, 0], 

[0, 0, 1, 0, 0, 0, 0, 0], 

[0, 0, 0, 0, 0, 0, 0, 0], 

[0, 0, 0, 0, 0, 0, 0, 0], 

[0, 0, 0, 0, 0, 0, 0, 0], 

[0, 0, 0, 0, 0, 0, 0, 0], 

[0, 0, 0, 0, 0, 0, 0, 0]]) 

>>> mask 

array([[0, 0, 0, 0, 0, 0, 0, 0], 

[0, 1, 1, 1, 0, 0, 0, 0], 

[0, 1, 1, 1, 0, 0, 0, 0], 

[0, 1, 1, 1, 0, 0, 0, 0], 

[0, 0, 0, 0, 1, 0, 0, 0], 

[0, 0, 0, 0, 0, 0, 0, 0], 

[0, 0, 0, 0, 0, 0, 1, 1], 

[0, 0, 0, 0, 0, 0, 1, 1]]) 

>>> ndimage.binary_propagation(input, mask=mask).astype(int) 

array([[0, 0, 0, 0, 0, 0, 0, 0], 

[0, 1, 1, 1, 0, 0, 0, 0], 

[0, 1, 1, 1, 0, 0, 0, 0], 

[0, 1, 1, 1, 0, 0, 0, 0], 

[0, 0, 0, 0, 0, 0, 0, 0], 

[0, 0, 0, 0, 0, 0, 0, 0], 

[0, 0, 0, 0, 0, 0, 0, 0], 

[0, 0, 0, 0, 0, 0, 0, 0]]) 

>>> ndimage.binary_propagation(input, mask=mask,\\ 

... structure=np.ones((3,3))).astype(int) 

array([[0, 0, 0, 0, 0, 0, 0, 0], 

[0, 1, 1, 1, 0, 0, 0, 0], 

[0, 1, 1, 1, 0, 0, 0, 0], 

[0, 1, 1, 1, 0, 0, 0, 0], 

[0, 0, 0, 0, 1, 0, 0, 0], 

[0, 0, 0, 0, 0, 0, 0, 0], 

[0, 0, 0, 0, 0, 0, 0, 0], 

[0, 0, 0, 0, 0, 0, 0, 0]]) 

 

>>> # Comparison between opening and erosion+propagation 

>>> a = np.zeros((6,6), dtype=int) 

>>> a[2:5, 2:5] = 1; a[0, 0] = 1; a[5, 5] = 1 

>>> a 

array([[1, 0, 0, 0, 0, 0], 

[0, 0, 0, 0, 0, 0], 

[0, 0, 1, 1, 1, 0], 

[0, 0, 1, 1, 1, 0], 

[0, 0, 1, 1, 1, 0], 

[0, 0, 0, 0, 0, 1]]) 

>>> ndimage.binary_opening(a).astype(int) 

array([[0, 0, 0, 0, 0, 0], 

[0, 0, 0, 0, 0, 0], 

[0, 0, 0, 1, 0, 0], 

[0, 0, 1, 1, 1, 0], 

[0, 0, 0, 1, 0, 0], 

[0, 0, 0, 0, 0, 0]]) 

>>> b = ndimage.binary_erosion(a) 

>>> b.astype(int) 

array([[0, 0, 0, 0, 0, 0], 

[0, 0, 0, 0, 0, 0], 

[0, 0, 0, 0, 0, 0], 

[0, 0, 0, 1, 0, 0], 

[0, 0, 0, 0, 0, 0], 

[0, 0, 0, 0, 0, 0]]) 

>>> ndimage.binary_propagation(b, mask=a).astype(int) 

array([[0, 0, 0, 0, 0, 0], 

[0, 0, 0, 0, 0, 0], 

[0, 0, 1, 1, 1, 0], 

[0, 0, 1, 1, 1, 0], 

[0, 0, 1, 1, 1, 0], 

[0, 0, 0, 0, 0, 0]]) 

 

""" 

return binary_dilation(input, structure, -1, mask, output, 

border_value, origin) 

 

 

def binary_fill_holes(input, structure=None, output=None, origin=0): 

""" 

Fill the holes in binary objects. 

 

 

Parameters 

---------- 

input : array_like 

n-dimensional binary array with holes to be filled 

structure : array_like, optional 

Structuring element used in the computation; large-size elements 

make computations faster but may miss holes separated from the 

background by thin regions. The default element (with a square 

connectivity equal to one) yields the intuitive result where all 

holes in the input have been filled. 

output : ndarray, optional 

Array of the same shape as input, into which the output is placed. 

By default, a new array is created. 

origin : int, tuple of ints, optional 

Position of the structuring element. 

 

Returns 

------- 

out : ndarray 

Transformation of the initial image `input` where holes have been 

filled. 

 

See also 

-------- 

binary_dilation, binary_propagation, label 

 

Notes 

----- 

The algorithm used in this function consists in invading the complementary 

of the shapes in `input` from the outer boundary of the image, 

using binary dilations. Holes are not connected to the boundary and are 

therefore not invaded. The result is the complementary subset of the 

invaded region. 

 

References 

---------- 

.. [1] http://en.wikipedia.org/wiki/Mathematical_morphology 

 

 

Examples 

-------- 

>>> from scipy import ndimage 

>>> a = np.zeros((5, 5), dtype=int) 

>>> a[1:4, 1:4] = 1 

>>> a[2,2] = 0 

>>> a 

array([[0, 0, 0, 0, 0], 

[0, 1, 1, 1, 0], 

[0, 1, 0, 1, 0], 

[0, 1, 1, 1, 0], 

[0, 0, 0, 0, 0]]) 

>>> ndimage.binary_fill_holes(a).astype(int) 

array([[0, 0, 0, 0, 0], 

[0, 1, 1, 1, 0], 

[0, 1, 1, 1, 0], 

[0, 1, 1, 1, 0], 

[0, 0, 0, 0, 0]]) 

>>> # Too big structuring element 

>>> ndimage.binary_fill_holes(a, structure=np.ones((5,5))).astype(int) 

array([[0, 0, 0, 0, 0], 

[0, 1, 1, 1, 0], 

[0, 1, 0, 1, 0], 

[0, 1, 1, 1, 0], 

[0, 0, 0, 0, 0]]) 

 

""" 

mask = numpy.logical_not(input) 

tmp = numpy.zeros(mask.shape, bool) 

inplace = isinstance(output, numpy.ndarray) 

if inplace: 

binary_dilation(tmp, structure, -1, mask, output, 1, origin) 

numpy.logical_not(output, output) 

else: 

output = binary_dilation(tmp, structure, -1, mask, None, 1, 

origin) 

numpy.logical_not(output, output) 

return output 

 

 

def grey_erosion(input, size=None, footprint=None, structure=None, 

output=None, mode="reflect", cval=0.0, origin=0): 

""" 

Calculate a greyscale erosion, using either a structuring element, 

or a footprint corresponding to a flat structuring element. 

 

Grayscale erosion is a mathematical morphology operation. For the 

simple case of a full and flat structuring element, it can be viewed 

as a minimum filter over a sliding window. 

 

Parameters 

---------- 

input : array_like 

Array over which the grayscale erosion is to be computed. 

size : tuple of ints 

Shape of a flat and full structuring element used for the grayscale 

erosion. Optional if `footprint` or `structure` is provided. 

footprint : array of ints, optional 

Positions of non-infinite elements of a flat structuring element 

used for the grayscale erosion. Non-zero values give the set of 

neighbors of the center over which the minimum is chosen. 

structure : array of ints, optional 

Structuring element used for the grayscale erosion. `structure` 

may be a non-flat structuring element. 

output : array, optional 

An array used for storing the output of the erosion may be provided. 

mode : {'reflect','constant','nearest','mirror', 'wrap'}, optional 

The `mode` parameter determines how the array borders are 

handled, where `cval` is the value when mode is equal to 

'constant'. Default is 'reflect' 

cval : scalar, optional 

Value to fill past edges of input if `mode` is 'constant'. Default 

is 0.0. 

origin : scalar, optional 

The `origin` parameter controls the placement of the filter. 

Default 0 

 

Returns 

------- 

output : ndarray 

Grayscale erosion of `input`. 

 

See also 

-------- 

binary_erosion, grey_dilation, grey_opening, grey_closing 

generate_binary_structure, ndimage.minimum_filter 

 

Notes 

----- 

The grayscale erosion of an image input by a structuring element s defined 

over a domain E is given by: 

 

(input+s)(x) = min {input(y) - s(x-y), for y in E} 

 

In particular, for structuring elements defined as 

s(y) = 0 for y in E, the grayscale erosion computes the minimum of the 

input image inside a sliding window defined by E. 

 

Grayscale erosion [1]_ is a *mathematical morphology* operation [2]_. 

 

References 

---------- 

.. [1] http://en.wikipedia.org/wiki/Erosion_%28morphology%29 

.. [2] http://en.wikipedia.org/wiki/Mathematical_morphology 

 

Examples 

-------- 

>>> from scipy import ndimage 

>>> a = np.zeros((7,7), dtype=int) 

>>> a[1:6, 1:6] = 3 

>>> a[4,4] = 2; a[2,3] = 1 

>>> a 

array([[0, 0, 0, 0, 0, 0, 0], 

[0, 3, 3, 3, 3, 3, 0], 

[0, 3, 3, 1, 3, 3, 0], 

[0, 3, 3, 3, 3, 3, 0], 

[0, 3, 3, 3, 2, 3, 0], 

[0, 3, 3, 3, 3, 3, 0], 

[0, 0, 0, 0, 0, 0, 0]]) 

>>> ndimage.grey_erosion(a, size=(3,3)) 

array([[0, 0, 0, 0, 0, 0, 0], 

[0, 0, 0, 0, 0, 0, 0], 

[0, 0, 1, 1, 1, 0, 0], 

[0, 0, 1, 1, 1, 0, 0], 

[0, 0, 3, 2, 2, 0, 0], 

[0, 0, 0, 0, 0, 0, 0], 

[0, 0, 0, 0, 0, 0, 0]]) 

>>> footprint = ndimage.generate_binary_structure(2, 1) 

>>> footprint 

array([[False, True, False], 

[ True, True, True], 

[False, True, False]], dtype=bool) 

>>> # Diagonally-connected elements are not considered neighbors 

>>> ndimage.grey_erosion(a, size=(3,3), footprint=footprint) 

array([[0, 0, 0, 0, 0, 0, 0], 

[0, 0, 0, 0, 0, 0, 0], 

[0, 0, 1, 1, 1, 0, 0], 

[0, 0, 3, 1, 2, 0, 0], 

[0, 0, 3, 2, 2, 0, 0], 

[0, 0, 0, 0, 0, 0, 0], 

[0, 0, 0, 0, 0, 0, 0]]) 

 

""" 

if size is None and footprint is None and structure is None: 

raise ValueError("size, footprint or structure must be specified") 

 

return filters._min_or_max_filter(input, size, footprint, structure, 

output, mode, cval, origin, 1) 

 

 

def grey_dilation(input, size=None, footprint=None, structure=None, 

output=None, mode="reflect", cval=0.0, origin=0): 

""" 

Calculate a greyscale dilation, using either a structuring element, 

or a footprint corresponding to a flat structuring element. 

 

Grayscale dilation is a mathematical morphology operation. For the 

simple case of a full and flat structuring element, it can be viewed 

as a maximum filter over a sliding window. 

 

Parameters 

---------- 

input : array_like 

Array over which the grayscale dilation is to be computed. 

size : tuple of ints 

Shape of a flat and full structuring element used for the grayscale 

dilation. Optional if `footprint` or `structure` is provided. 

footprint : array of ints, optional 

Positions of non-infinite elements of a flat structuring element 

used for the grayscale dilation. Non-zero values give the set of 

neighbors of the center over which the maximum is chosen. 

structure : array of ints, optional 

Structuring element used for the grayscale dilation. `structure` 

may be a non-flat structuring element. 

output : array, optional 

An array used for storing the output of the dilation may be provided. 

mode : {'reflect','constant','nearest','mirror', 'wrap'}, optional 

The `mode` parameter determines how the array borders are 

handled, where `cval` is the value when mode is equal to 

'constant'. Default is 'reflect' 

cval : scalar, optional 

Value to fill past edges of input if `mode` is 'constant'. Default 

is 0.0. 

origin : scalar, optional 

The `origin` parameter controls the placement of the filter. 

Default 0 

 

Returns 

------- 

grey_dilation : ndarray 

Grayscale dilation of `input`. 

 

See also 

-------- 

binary_dilation, grey_erosion, grey_closing, grey_opening 

generate_binary_structure, ndimage.maximum_filter 

 

Notes 

----- 

The grayscale dilation of an image input by a structuring element s defined 

over a domain E is given by: 

 

(input+s)(x) = max {input(y) + s(x-y), for y in E} 

 

In particular, for structuring elements defined as 

s(y) = 0 for y in E, the grayscale dilation computes the maximum of the 

input image inside a sliding window defined by E. 

 

Grayscale dilation [1]_ is a *mathematical morphology* operation [2]_. 

 

References 

---------- 

.. [1] http://en.wikipedia.org/wiki/Dilation_%28morphology%29 

.. [2] http://en.wikipedia.org/wiki/Mathematical_morphology 

 

Examples 

-------- 

>>> from scipy import ndimage 

>>> a = np.zeros((7,7), dtype=int) 

>>> a[2:5, 2:5] = 1 

>>> a[4,4] = 2; a[2,3] = 3 

>>> a 

array([[0, 0, 0, 0, 0, 0, 0], 

[0, 0, 0, 0, 0, 0, 0], 

[0, 0, 1, 3, 1, 0, 0], 

[0, 0, 1, 1, 1, 0, 0], 

[0, 0, 1, 1, 2, 0, 0], 

[0, 0, 0, 0, 0, 0, 0], 

[0, 0, 0, 0, 0, 0, 0]]) 

>>> ndimage.grey_dilation(a, size=(3,3)) 

array([[0, 0, 0, 0, 0, 0, 0], 

[0, 1, 3, 3, 3, 1, 0], 

[0, 1, 3, 3, 3, 1, 0], 

[0, 1, 3, 3, 3, 2, 0], 

[0, 1, 1, 2, 2, 2, 0], 

[0, 1, 1, 2, 2, 2, 0], 

[0, 0, 0, 0, 0, 0, 0]]) 

>>> ndimage.grey_dilation(a, footprint=np.ones((3,3))) 

array([[0, 0, 0, 0, 0, 0, 0], 

[0, 1, 3, 3, 3, 1, 0], 

[0, 1, 3, 3, 3, 1, 0], 

[0, 1, 3, 3, 3, 2, 0], 

[0, 1, 1, 2, 2, 2, 0], 

[0, 1, 1, 2, 2, 2, 0], 

[0, 0, 0, 0, 0, 0, 0]]) 

>>> s = ndimage.generate_binary_structure(2,1) 

>>> s 

array([[False, True, False], 

[ True, True, True], 

[False, True, False]], dtype=bool) 

>>> ndimage.grey_dilation(a, footprint=s) 

array([[0, 0, 0, 0, 0, 0, 0], 

[0, 0, 1, 3, 1, 0, 0], 

[0, 1, 3, 3, 3, 1, 0], 

[0, 1, 1, 3, 2, 1, 0], 

[0, 1, 1, 2, 2, 2, 0], 

[0, 0, 1, 1, 2, 0, 0], 

[0, 0, 0, 0, 0, 0, 0]]) 

>>> ndimage.grey_dilation(a, size=(3,3), structure=np.ones((3,3))) 

array([[1, 1, 1, 1, 1, 1, 1], 

[1, 2, 4, 4, 4, 2, 1], 

[1, 2, 4, 4, 4, 2, 1], 

[1, 2, 4, 4, 4, 3, 1], 

[1, 2, 2, 3, 3, 3, 1], 

[1, 2, 2, 3, 3, 3, 1], 

[1, 1, 1, 1, 1, 1, 1]]) 

 

""" 

if size is None and footprint is None and structure is None: 

raise ValueError("size, footprint or structure must be specified") 

if structure is not None: 

structure = numpy.asarray(structure) 

structure = structure[tuple([slice(None, None, -1)] * 

structure.ndim)] 

if footprint is not None: 

footprint = numpy.asarray(footprint) 

footprint = footprint[tuple([slice(None, None, -1)] * 

footprint.ndim)] 

 

input = numpy.asarray(input) 

origin = _ni_support._normalize_sequence(origin, input.ndim) 

for ii in range(len(origin)): 

origin[ii] = -origin[ii] 

if footprint is not None: 

sz = footprint.shape[ii] 

elif structure is not None: 

sz = structure.shape[ii] 

elif numpy.isscalar(size): 

sz = size 

else: 

sz = size[ii] 

if not sz & 1: 

origin[ii] -= 1 

 

return filters._min_or_max_filter(input, size, footprint, structure, 

output, mode, cval, origin, 0) 

 

 

def grey_opening(input, size=None, footprint=None, structure=None, 

output=None, mode="reflect", cval=0.0, origin=0): 

""" 

Multi-dimensional greyscale opening. 

 

A greyscale opening consists in the succession of a greyscale erosion, 

and a greyscale dilation. 

 

Parameters 

---------- 

input : array_like 

Array over which the grayscale opening is to be computed. 

size : tuple of ints 

Shape of a flat and full structuring element used for the grayscale 

opening. Optional if `footprint` or `structure` is provided. 

footprint : array of ints, optional 

Positions of non-infinite elements of a flat structuring element 

used for the grayscale opening. 

structure : array of ints, optional 

Structuring element used for the grayscale opening. `structure` 

may be a non-flat structuring element. 

output : array, optional 

An array used for storing the output of the opening may be provided. 

mode : {'reflect', 'constant', 'nearest', 'mirror', 'wrap'}, optional 

The `mode` parameter determines how the array borders are 

handled, where `cval` is the value when mode is equal to 

'constant'. Default is 'reflect' 

cval : scalar, optional 

Value to fill past edges of input if `mode` is 'constant'. Default 

is 0.0. 

origin : scalar, optional 

The `origin` parameter controls the placement of the filter. 

Default 0 

 

Returns 

------- 

grey_opening : ndarray 

Result of the grayscale opening of `input` with `structure`. 

 

See also 

-------- 

binary_opening, grey_dilation, grey_erosion, grey_closing 

generate_binary_structure 

 

Notes 

----- 

The action of a grayscale opening with a flat structuring element amounts 

to smoothen high local maxima, whereas binary opening erases small objects. 

 

References 

---------- 

.. [1] http://en.wikipedia.org/wiki/Mathematical_morphology 

 

Examples 

-------- 

>>> from scipy import ndimage 

>>> a = np.arange(36).reshape((6,6)) 

>>> a[3, 3] = 50 

>>> a 

array([[ 0, 1, 2, 3, 4, 5], 

[ 6, 7, 8, 9, 10, 11], 

[12, 13, 14, 15, 16, 17], 

[18, 19, 20, 50, 22, 23], 

[24, 25, 26, 27, 28, 29], 

[30, 31, 32, 33, 34, 35]]) 

>>> ndimage.grey_opening(a, size=(3,3)) 

array([[ 0, 1, 2, 3, 4, 4], 

[ 6, 7, 8, 9, 10, 10], 

[12, 13, 14, 15, 16, 16], 

[18, 19, 20, 22, 22, 22], 

[24, 25, 26, 27, 28, 28], 

[24, 25, 26, 27, 28, 28]]) 

>>> # Note that the local maximum a[3,3] has disappeared 

 

""" 

if (size is not None) and (footprint is not None): 

warnings.warn("ignoring size because footprint is set", UserWarning, stacklevel=2) 

tmp = grey_erosion(input, size, footprint, structure, None, mode, 

cval, origin) 

return grey_dilation(tmp, size, footprint, structure, output, mode, 

cval, origin) 

 

 

def grey_closing(input, size=None, footprint=None, structure=None, 

output=None, mode="reflect", cval=0.0, origin=0): 

""" 

Multi-dimensional greyscale closing. 

 

A greyscale closing consists in the succession of a greyscale dilation, 

and a greyscale erosion. 

 

Parameters 

---------- 

input : array_like 

Array over which the grayscale closing is to be computed. 

size : tuple of ints 

Shape of a flat and full structuring element used for the grayscale 

closing. Optional if `footprint` or `structure` is provided. 

footprint : array of ints, optional 

Positions of non-infinite elements of a flat structuring element 

used for the grayscale closing. 

structure : array of ints, optional 

Structuring element used for the grayscale closing. `structure` 

may be a non-flat structuring element. 

output : array, optional 

An array used for storing the output of the closing may be provided. 

mode : {'reflect', 'constant', 'nearest', 'mirror', 'wrap'}, optional 

The `mode` parameter determines how the array borders are 

handled, where `cval` is the value when mode is equal to 

'constant'. Default is 'reflect' 

cval : scalar, optional 

Value to fill past edges of input if `mode` is 'constant'. Default 

is 0.0. 

origin : scalar, optional 

The `origin` parameter controls the placement of the filter. 

Default 0 

 

Returns 

------- 

grey_closing : ndarray 

Result of the grayscale closing of `input` with `structure`. 

 

See also 

-------- 

binary_closing, grey_dilation, grey_erosion, grey_opening, 

generate_binary_structure 

 

Notes 

----- 

The action of a grayscale closing with a flat structuring element amounts 

to smoothen deep local minima, whereas binary closing fills small holes. 

 

References 

---------- 

.. [1] http://en.wikipedia.org/wiki/Mathematical_morphology 

 

Examples 

-------- 

>>> from scipy import ndimage 

>>> a = np.arange(36).reshape((6,6)) 

>>> a[3,3] = 0 

>>> a 

array([[ 0, 1, 2, 3, 4, 5], 

[ 6, 7, 8, 9, 10, 11], 

[12, 13, 14, 15, 16, 17], 

[18, 19, 20, 0, 22, 23], 

[24, 25, 26, 27, 28, 29], 

[30, 31, 32, 33, 34, 35]]) 

>>> ndimage.grey_closing(a, size=(3,3)) 

array([[ 7, 7, 8, 9, 10, 11], 

[ 7, 7, 8, 9, 10, 11], 

[13, 13, 14, 15, 16, 17], 

[19, 19, 20, 20, 22, 23], 

[25, 25, 26, 27, 28, 29], 

[31, 31, 32, 33, 34, 35]]) 

>>> # Note that the local minimum a[3,3] has disappeared 

 

""" 

if (size is not None) and (footprint is not None): 

warnings.warn("ignoring size because footprint is set", UserWarning, stacklevel=2) 

tmp = grey_dilation(input, size, footprint, structure, None, mode, 

cval, origin) 

return grey_erosion(tmp, size, footprint, structure, output, mode, 

cval, origin) 

 

 

def morphological_gradient(input, size=None, footprint=None, structure=None, 

output=None, mode="reflect", cval=0.0, origin=0): 

""" 

Multi-dimensional morphological gradient. 

 

The morphological gradient is calculated as the difference between a 

dilation and an erosion of the input with a given structuring element. 

 

Parameters 

---------- 

input : array_like 

Array over which to compute the morphlogical gradient. 

size : tuple of ints 

Shape of a flat and full structuring element used for the mathematical 

morphology operations. Optional if `footprint` or `structure` is 

provided. A larger `size` yields a more blurred gradient. 

footprint : array of ints, optional 

Positions of non-infinite elements of a flat structuring element 

used for the morphology operations. Larger footprints 

give a more blurred morphological gradient. 

structure : array of ints, optional 

Structuring element used for the morphology operations. 

`structure` may be a non-flat structuring element. 

output : array, optional 

An array used for storing the output of the morphological gradient 

may be provided. 

mode : {'reflect', 'constant', 'nearest', 'mirror', 'wrap'}, optional 

The `mode` parameter determines how the array borders are 

handled, where `cval` is the value when mode is equal to 

'constant'. Default is 'reflect' 

cval : scalar, optional 

Value to fill past edges of input if `mode` is 'constant'. Default 

is 0.0. 

origin : scalar, optional 

The `origin` parameter controls the placement of the filter. 

Default 0 

 

Returns 

------- 

morphological_gradient : ndarray 

Morphological gradient of `input`. 

 

See also 

-------- 

grey_dilation, grey_erosion, ndimage.gaussian_gradient_magnitude 

 

Notes 

----- 

For a flat structuring element, the morphological gradient 

computed at a given point corresponds to the maximal difference 

between elements of the input among the elements covered by the 

structuring element centered on the point. 

 

References 

---------- 

.. [1] http://en.wikipedia.org/wiki/Mathematical_morphology 

 

Examples 

-------- 

>>> from scipy import ndimage 

>>> a = np.zeros((7,7), dtype=int) 

>>> a[2:5, 2:5] = 1 

>>> ndimage.morphological_gradient(a, size=(3,3)) 

array([[0, 0, 0, 0, 0, 0, 0], 

[0, 1, 1, 1, 1, 1, 0], 

[0, 1, 1, 1, 1, 1, 0], 

[0, 1, 1, 0, 1, 1, 0], 

[0, 1, 1, 1, 1, 1, 0], 

[0, 1, 1, 1, 1, 1, 0], 

[0, 0, 0, 0, 0, 0, 0]]) 

>>> # The morphological gradient is computed as the difference 

>>> # between a dilation and an erosion 

>>> ndimage.grey_dilation(a, size=(3,3)) -\\ 

... ndimage.grey_erosion(a, size=(3,3)) 

array([[0, 0, 0, 0, 0, 0, 0], 

[0, 1, 1, 1, 1, 1, 0], 

[0, 1, 1, 1, 1, 1, 0], 

[0, 1, 1, 0, 1, 1, 0], 

[0, 1, 1, 1, 1, 1, 0], 

[0, 1, 1, 1, 1, 1, 0], 

[0, 0, 0, 0, 0, 0, 0]]) 

>>> a = np.zeros((7,7), dtype=int) 

>>> a[2:5, 2:5] = 1 

>>> a[4,4] = 2; a[2,3] = 3 

>>> a 

array([[0, 0, 0, 0, 0, 0, 0], 

[0, 0, 0, 0, 0, 0, 0], 

[0, 0, 1, 3, 1, 0, 0], 

[0, 0, 1, 1, 1, 0, 0], 

[0, 0, 1, 1, 2, 0, 0], 

[0, 0, 0, 0, 0, 0, 0], 

[0, 0, 0, 0, 0, 0, 0]]) 

>>> ndimage.morphological_gradient(a, size=(3,3)) 

array([[0, 0, 0, 0, 0, 0, 0], 

[0, 1, 3, 3, 3, 1, 0], 

[0, 1, 3, 3, 3, 1, 0], 

[0, 1, 3, 2, 3, 2, 0], 

[0, 1, 1, 2, 2, 2, 0], 

[0, 1, 1, 2, 2, 2, 0], 

[0, 0, 0, 0, 0, 0, 0]]) 

 

""" 

tmp = grey_dilation(input, size, footprint, structure, None, mode, 

cval, origin) 

if isinstance(output, numpy.ndarray): 

grey_erosion(input, size, footprint, structure, output, mode, 

cval, origin) 

return numpy.subtract(tmp, output, output) 

else: 

return (tmp - grey_erosion(input, size, footprint, structure, 

None, mode, cval, origin)) 

 

 

def morphological_laplace(input, size=None, footprint=None, 

structure=None, output=None, 

mode="reflect", cval=0.0, origin=0): 

""" 

Multi-dimensional morphological laplace. 

 

Parameters 

---------- 

input : array_like 

Input. 

size : int or sequence of ints, optional 

See `structure`. 

footprint : bool or ndarray, optional 

See `structure`. 

structure : structure, optional 

Either `size`, `footprint`, or the `structure` must be provided. 

output : ndarray, optional 

An output array can optionally be provided. 

mode : {'reflect','constant','nearest','mirror', 'wrap'}, optional 

The mode parameter determines how the array borders are handled. 

For 'constant' mode, values beyond borders are set to be `cval`. 

Default is 'reflect'. 

cval : scalar, optional 

Value to fill past edges of input if mode is 'constant'. 

Default is 0.0 

origin : origin, optional 

The origin parameter controls the placement of the filter. 

 

Returns 

------- 

morphological_laplace : ndarray 

Output 

 

""" 

tmp1 = grey_dilation(input, size, footprint, structure, None, mode, 

cval, origin) 

if isinstance(output, numpy.ndarray): 

grey_erosion(input, size, footprint, structure, output, mode, 

cval, origin) 

numpy.add(tmp1, output, output) 

numpy.subtract(output, input, output) 

return numpy.subtract(output, input, output) 

else: 

tmp2 = grey_erosion(input, size, footprint, structure, None, mode, 

cval, origin) 

numpy.add(tmp1, tmp2, tmp2) 

numpy.subtract(tmp2, input, tmp2) 

numpy.subtract(tmp2, input, tmp2) 

return tmp2 

 

 

def white_tophat(input, size=None, footprint=None, structure=None, 

output=None, mode="reflect", cval=0.0, origin=0): 

""" 

Multi-dimensional white tophat filter. 

 

Parameters 

---------- 

input : array_like 

Input. 

size : tuple of ints 

Shape of a flat and full structuring element used for the filter. 

Optional if `footprint` or `structure` is provided. 

footprint : array of ints, optional 

Positions of elements of a flat structuring element 

used for the white tophat filter. 

structure : array of ints, optional 

Structuring element used for the filter. `structure` 

may be a non-flat structuring element. 

output : array, optional 

An array used for storing the output of the filter may be provided. 

mode : {'reflect', 'constant', 'nearest', 'mirror', 'wrap'}, optional 

The `mode` parameter determines how the array borders are 

handled, where `cval` is the value when mode is equal to 

'constant'. Default is 'reflect' 

cval : scalar, optional 

Value to fill past edges of input if `mode` is 'constant'. 

Default is 0.0. 

origin : scalar, optional 

The `origin` parameter controls the placement of the filter. 

Default is 0. 

 

Returns 

------- 

output : ndarray 

Result of the filter of `input` with `structure`. 

 

See also 

-------- 

black_tophat 

 

""" 

if (size is not None) and (footprint is not None): 

warnings.warn("ignoring size because footprint is set", UserWarning, stacklevel=2) 

tmp = grey_erosion(input, size, footprint, structure, None, mode, 

cval, origin) 

tmp = grey_dilation(tmp, size, footprint, structure, output, mode, 

cval, origin) 

if tmp is None: 

tmp = output 

 

if input.dtype == numpy.bool_ and tmp.dtype == numpy.bool_: 

numpy.bitwise_xor(input, tmp, out=tmp) 

else: 

numpy.subtract(input, tmp, out=tmp) 

return tmp 

 

 

def black_tophat(input, size=None, footprint=None, 

structure=None, output=None, mode="reflect", 

cval=0.0, origin=0): 

""" 

Multi-dimensional black tophat filter. 

 

Parameters 

---------- 

input : array_like 

Input. 

size : tuple of ints, optional 

Shape of a flat and full structuring element used for the filter. 

Optional if `footprint` or `structure` is provided. 

footprint : array of ints, optional 

Positions of non-infinite elements of a flat structuring element 

used for the black tophat filter. 

structure : array of ints, optional 

Structuring element used for the filter. `structure` 

may be a non-flat structuring element. 

output : array, optional 

An array used for storing the output of the filter may be provided. 

mode : {'reflect', 'constant', 'nearest', 'mirror', 'wrap'}, optional 

The `mode` parameter determines how the array borders are 

handled, where `cval` is the value when mode is equal to 

'constant'. Default is 'reflect' 

cval : scalar, optional 

Value to fill past edges of input if `mode` is 'constant'. Default 

is 0.0. 

origin : scalar, optional 

The `origin` parameter controls the placement of the filter. 

Default 0 

 

Returns 

------- 

black_tophat : ndarray 

Result of the filter of `input` with `structure`. 

 

See also 

-------- 

white_tophat, grey_opening, grey_closing 

 

""" 

if (size is not None) and (footprint is not None): 

warnings.warn("ignoring size because footprint is set", UserWarning, stacklevel=2) 

tmp = grey_dilation(input, size, footprint, structure, None, mode, 

cval, origin) 

tmp = grey_erosion(tmp, size, footprint, structure, output, mode, 

cval, origin) 

if tmp is None: 

tmp = output 

 

if input.dtype == numpy.bool_ and tmp.dtype == numpy.bool_: 

numpy.bitwise_xor(tmp, input, out=tmp) 

else: 

numpy.subtract(tmp, input, out=tmp) 

return tmp 

 

 

def distance_transform_bf(input, metric="euclidean", sampling=None, 

return_distances=True, return_indices=False, 

distances=None, indices=None): 

""" 

Distance transform function by a brute force algorithm. 

 

This function calculates the distance transform of the `input`, by 

replacing each foreground (non-zero) element, with its 

shortest distance to the background (any zero-valued element). 

 

In addition to the distance transform, the feature transform can 

be calculated. In this case the index of the closest background 

element is returned along the first axis of the result. 

 

Parameters 

---------- 

input : array_like 

Input 

metric : str, optional 

Three types of distance metric are supported: 'euclidean', 'taxicab' 

and 'chessboard'. 

sampling : {int, sequence of ints}, optional 

This parameter is only used in the case of the euclidean `metric` 

distance transform. 

 

The sampling along each axis can be given by the `sampling` parameter 

which should be a sequence of length equal to the input rank, or a 

single number in which the `sampling` is assumed to be equal along all 

axes. 

return_distances : bool, optional 

The `return_distances` flag can be used to indicate if the distance 

transform is returned. 

 

The default is True. 

return_indices : bool, optional 

The `return_indices` flags can be used to indicate if the feature 

transform is returned. 

 

The default is False. 

distances : float64 ndarray, optional 

Optional output array to hold distances (if `return_distances` is 

True). 

indices : int64 ndarray, optional 

Optional output array to hold indices (if `return_indices` is True). 

 

Returns 

------- 

distances : ndarray 

Distance array if `return_distances` is True. 

indices : ndarray 

Indices array if `return_indices` is True. 

 

Notes 

----- 

This function employs a slow brute force algorithm, see also the 

function distance_transform_cdt for more efficient taxicab and 

chessboard algorithms. 

 

""" 

if (not return_distances) and (not return_indices): 

msg = 'at least one of distances/indices must be specified' 

raise RuntimeError(msg) 

 

tmp1 = numpy.asarray(input) != 0 

struct = generate_binary_structure(tmp1.ndim, tmp1.ndim) 

tmp2 = binary_dilation(tmp1, struct) 

tmp2 = numpy.logical_xor(tmp1, tmp2) 

tmp1 = tmp1.astype(numpy.int8) - tmp2.astype(numpy.int8) 

metric = metric.lower() 

if metric == 'euclidean': 

metric = 1 

elif metric in ['taxicab', 'cityblock', 'manhattan']: 

metric = 2 

elif metric == 'chessboard': 

metric = 3 

else: 

raise RuntimeError('distance metric not supported') 

if sampling is not None: 

sampling = _ni_support._normalize_sequence(sampling, tmp1.ndim) 

sampling = numpy.asarray(sampling, dtype=numpy.float64) 

if not sampling.flags.contiguous: 

sampling = sampling.copy() 

if return_indices: 

ft = numpy.zeros(tmp1.shape, dtype=numpy.int32) 

else: 

ft = None 

if return_distances: 

if distances is None: 

if metric == 1: 

dt = numpy.zeros(tmp1.shape, dtype=numpy.float64) 

else: 

dt = numpy.zeros(tmp1.shape, dtype=numpy.uint32) 

else: 

if distances.shape != tmp1.shape: 

raise RuntimeError('distances array has wrong shape') 

if metric == 1: 

if distances.dtype.type != numpy.float64: 

raise RuntimeError('distances array must be float64') 

else: 

if distances.dtype.type != numpy.uint32: 

raise RuntimeError('distances array must be uint32') 

dt = distances 

else: 

dt = None 

 

_nd_image.distance_transform_bf(tmp1, metric, sampling, dt, ft) 

if return_indices: 

if isinstance(indices, numpy.ndarray): 

if indices.dtype.type != numpy.int32: 

raise RuntimeError('indices must of int32 type') 

if indices.shape != (tmp1.ndim,) + tmp1.shape: 

raise RuntimeError('indices has wrong shape') 

tmp2 = indices 

else: 

tmp2 = numpy.indices(tmp1.shape, dtype=numpy.int32) 

ft = numpy.ravel(ft) 

for ii in range(tmp2.shape[0]): 

rtmp = numpy.ravel(tmp2[ii, ...])[ft] 

rtmp.shape = tmp1.shape 

tmp2[ii, ...] = rtmp 

ft = tmp2 

 

# construct and return the result 

result = [] 

if return_distances and not isinstance(distances, numpy.ndarray): 

result.append(dt) 

if return_indices and not isinstance(indices, numpy.ndarray): 

result.append(ft) 

 

if len(result) == 2: 

return tuple(result) 

elif len(result) == 1: 

return result[0] 

else: 

return None 

 

 

def distance_transform_cdt(input, metric='chessboard', return_distances=True, 

return_indices=False, distances=None, indices=None): 

""" 

Distance transform for chamfer type of transforms. 

 

Parameters 

---------- 

input : array_like 

Input 

metric : {'chessboard', 'taxicab'}, optional 

The `metric` determines the type of chamfering that is done. If the 

`metric` is equal to 'taxicab' a structure is generated using 

generate_binary_structure with a squared distance equal to 1. If 

the `metric` is equal to 'chessboard', a `metric` is generated 

using generate_binary_structure with a squared distance equal to 

the dimensionality of the array. These choices correspond to the 

common interpretations of the 'taxicab' and the 'chessboard' 

distance metrics in two dimensions. 

 

The default for `metric` is 'chessboard'. 

return_distances, return_indices : bool, optional 

The `return_distances`, and `return_indices` flags can be used to 

indicate if the distance transform, the feature transform, or both 

must be returned. 

 

If the feature transform is returned (``return_indices=True``), 

the index of the closest background element is returned along 

the first axis of the result. 

 

The `return_distances` default is True, and the 

`return_indices` default is False. 

distances, indices : ndarrays of int32, optional 

The `distances` and `indices` arguments can be used to give optional 

output arrays that must be the same shape as `input`. 

 

""" 

if (not return_distances) and (not return_indices): 

msg = 'at least one of distances/indices must be specified' 

raise RuntimeError(msg) 

 

ft_inplace = isinstance(indices, numpy.ndarray) 

dt_inplace = isinstance(distances, numpy.ndarray) 

input = numpy.asarray(input) 

if metric in ['taxicab', 'cityblock', 'manhattan']: 

rank = input.ndim 

metric = generate_binary_structure(rank, 1) 

elif metric == 'chessboard': 

rank = input.ndim 

metric = generate_binary_structure(rank, rank) 

else: 

try: 

metric = numpy.asarray(metric) 

except: 

raise RuntimeError('invalid metric provided') 

for s in metric.shape: 

if s != 3: 

raise RuntimeError('metric sizes must be equal to 3') 

 

if not metric.flags.contiguous: 

metric = metric.copy() 

if dt_inplace: 

if distances.dtype.type != numpy.int32: 

raise RuntimeError('distances must be of int32 type') 

if distances.shape != input.shape: 

raise RuntimeError('distances has wrong shape') 

dt = distances 

dt[...] = numpy.where(input, -1, 0).astype(numpy.int32) 

else: 

dt = numpy.where(input, -1, 0).astype(numpy.int32) 

 

rank = dt.ndim 

if return_indices: 

sz = numpy.product(dt.shape, axis=0) 

ft = numpy.arange(sz, dtype=numpy.int32) 

ft.shape = dt.shape 

else: 

ft = None 

 

_nd_image.distance_transform_op(metric, dt, ft) 

dt = dt[tuple([slice(None, None, -1)] * rank)] 

if return_indices: 

ft = ft[tuple([slice(None, None, -1)] * rank)] 

_nd_image.distance_transform_op(metric, dt, ft) 

dt = dt[tuple([slice(None, None, -1)] * rank)] 

if return_indices: 

ft = ft[tuple([slice(None, None, -1)] * rank)] 

ft = numpy.ravel(ft) 

if ft_inplace: 

if indices.dtype.type != numpy.int32: 

raise RuntimeError('indices must of int32 type') 

if indices.shape != (dt.ndim,) + dt.shape: 

raise RuntimeError('indices has wrong shape') 

tmp = indices 

else: 

tmp = numpy.indices(dt.shape, dtype=numpy.int32) 

for ii in range(tmp.shape[0]): 

rtmp = numpy.ravel(tmp[ii, ...])[ft] 

rtmp.shape = dt.shape 

tmp[ii, ...] = rtmp 

ft = tmp 

 

# construct and return the result 

result = [] 

if return_distances and not dt_inplace: 

result.append(dt) 

if return_indices and not ft_inplace: 

result.append(ft) 

 

if len(result) == 2: 

return tuple(result) 

elif len(result) == 1: 

return result[0] 

else: 

return None 

 

 

def distance_transform_edt(input, sampling=None, return_distances=True, 

return_indices=False, distances=None, indices=None): 

""" 

Exact euclidean distance transform. 

 

In addition to the distance transform, the feature transform can 

be calculated. In this case the index of the closest background 

element is returned along the first axis of the result. 

 

Parameters 

---------- 

input : array_like 

Input data to transform. Can be any type but will be converted 

into binary: 1 wherever input equates to True, 0 elsewhere. 

sampling : float or int, or sequence of same, optional 

Spacing of elements along each dimension. If a sequence, must be of 

length equal to the input rank; if a single number, this is used for 

all axes. If not specified, a grid spacing of unity is implied. 

return_distances : bool, optional 

Whether to return distance matrix. At least one of 

return_distances/return_indices must be True. Default is True. 

return_indices : bool, optional 

Whether to return indices matrix. Default is False. 

distances : ndarray, optional 

Used for output of distance array, must be of type float64. 

indices : ndarray, optional 

Used for output of indices, must be of type int32. 

 

Returns 

------- 

distance_transform_edt : ndarray or list of ndarrays 

Either distance matrix, index matrix, or a list of the two, 

depending on `return_x` flags and `distance` and `indices` 

input parameters. 

 

Notes 

----- 

The euclidean distance transform gives values of the euclidean 

distance:: 

 

n 

y_i = sqrt(sum (x[i]-b[i])**2) 

i 

 

where b[i] is the background point (value 0) with the smallest 

Euclidean distance to input points x[i], and n is the 

number of dimensions. 

 

Examples 

-------- 

>>> from scipy import ndimage 

>>> a = np.array(([0,1,1,1,1], 

... [0,0,1,1,1], 

... [0,1,1,1,1], 

... [0,1,1,1,0], 

... [0,1,1,0,0])) 

>>> ndimage.distance_transform_edt(a) 

array([[ 0. , 1. , 1.4142, 2.2361, 3. ], 

[ 0. , 0. , 1. , 2. , 2. ], 

[ 0. , 1. , 1.4142, 1.4142, 1. ], 

[ 0. , 1. , 1.4142, 1. , 0. ], 

[ 0. , 1. , 1. , 0. , 0. ]]) 

 

With a sampling of 2 units along x, 1 along y: 

 

>>> ndimage.distance_transform_edt(a, sampling=[2,1]) 

array([[ 0. , 1. , 2. , 2.8284, 3.6056], 

[ 0. , 0. , 1. , 2. , 3. ], 

[ 0. , 1. , 2. , 2.2361, 2. ], 

[ 0. , 1. , 2. , 1. , 0. ], 

[ 0. , 1. , 1. , 0. , 0. ]]) 

 

Asking for indices as well: 

 

>>> edt, inds = ndimage.distance_transform_edt(a, return_indices=True) 

>>> inds 

array([[[0, 0, 1, 1, 3], 

[1, 1, 1, 1, 3], 

[2, 2, 1, 3, 3], 

[3, 3, 4, 4, 3], 

[4, 4, 4, 4, 4]], 

[[0, 0, 1, 1, 4], 

[0, 1, 1, 1, 4], 

[0, 0, 1, 4, 4], 

[0, 0, 3, 3, 4], 

[0, 0, 3, 3, 4]]]) 

 

With arrays provided for inplace outputs: 

 

>>> indices = np.zeros(((np.ndim(a),) + a.shape), dtype=np.int32) 

>>> ndimage.distance_transform_edt(a, return_indices=True, indices=indices) 

array([[ 0. , 1. , 1.4142, 2.2361, 3. ], 

[ 0. , 0. , 1. , 2. , 2. ], 

[ 0. , 1. , 1.4142, 1.4142, 1. ], 

[ 0. , 1. , 1.4142, 1. , 0. ], 

[ 0. , 1. , 1. , 0. , 0. ]]) 

>>> indices 

array([[[0, 0, 1, 1, 3], 

[1, 1, 1, 1, 3], 

[2, 2, 1, 3, 3], 

[3, 3, 4, 4, 3], 

[4, 4, 4, 4, 4]], 

[[0, 0, 1, 1, 4], 

[0, 1, 1, 1, 4], 

[0, 0, 1, 4, 4], 

[0, 0, 3, 3, 4], 

[0, 0, 3, 3, 4]]]) 

 

""" 

if (not return_distances) and (not return_indices): 

msg = 'at least one of distances/indices must be specified' 

raise RuntimeError(msg) 

 

ft_inplace = isinstance(indices, numpy.ndarray) 

dt_inplace = isinstance(distances, numpy.ndarray) 

# calculate the feature transform 

input = numpy.atleast_1d(numpy.where(input, 1, 0).astype(numpy.int8)) 

if sampling is not None: 

sampling = _ni_support._normalize_sequence(sampling, input.ndim) 

sampling = numpy.asarray(sampling, dtype=numpy.float64) 

if not sampling.flags.contiguous: 

sampling = sampling.copy() 

 

if ft_inplace: 

ft = indices 

if ft.shape != (input.ndim,) + input.shape: 

raise RuntimeError('indices has wrong shape') 

if ft.dtype.type != numpy.int32: 

raise RuntimeError('indices must be of int32 type') 

else: 

ft = numpy.zeros((input.ndim,) + input.shape, dtype=numpy.int32) 

 

_nd_image.euclidean_feature_transform(input, sampling, ft) 

# if requested, calculate the distance transform 

if return_distances: 

dt = ft - numpy.indices(input.shape, dtype=ft.dtype) 

dt = dt.astype(numpy.float64) 

if sampling is not None: 

for ii in range(len(sampling)): 

dt[ii, ...] *= sampling[ii] 

numpy.multiply(dt, dt, dt) 

if dt_inplace: 

dt = numpy.add.reduce(dt, axis=0) 

if distances.shape != dt.shape: 

raise RuntimeError('indices has wrong shape') 

if distances.dtype.type != numpy.float64: 

raise RuntimeError('indices must be of float64 type') 

numpy.sqrt(dt, distances) 

else: 

dt = numpy.add.reduce(dt, axis=0) 

dt = numpy.sqrt(dt) 

 

# construct and return the result 

result = [] 

if return_distances and not dt_inplace: 

result.append(dt) 

if return_indices and not ft_inplace: 

result.append(ft) 

 

if len(result) == 2: 

return tuple(result) 

elif len(result) == 1: 

return result[0] 

else: 

return None