|
# http://pyrocko.org - GPLv3 # # The Pyrocko Developers, 21st Century # ---|P------/S----------~Lg----------
''' <html> <head> <style type="text/css"> body { margin-left:10px }; </style> </head> <body> <h1 align="center">STA/LTA</h1> <p> Detect onsets automatically using the Short-Time-Average/Long-Time-Average ratio.<br/> This snuffling uses the method: <a href="http://emolch.github.io/pyrocko/v0.3/trace.html#pyrocko.trace.\ Trace.sta_lta_centered" style="text-decoration:none"> <pre>pyrocko.trace.Trace.sta_lta_centered</pre> </a></p> <p> <b>Parameters:</b><br /> <b>· Highpass [Hz]</b> - Apply high pass filter before analysing.<br /> <b>· Lowpass [Hz]</b> - Apply low pass filter before analysing.<br /> <b>· Short Window [s]</b> - Window length of the short window.<br /> <b>· Ratio</b> - Long window length is the short window length times the <b>Ratio</b>.<br /> <b>· Level</b> - Define a trigger threshold. A marker is added where STA/LTA ratios exceed this threshold. <br /> <b>· Processing Block length</b> - Subdivide dataset in blocks for analysis. <br /> <b>· Show trigger level traces </b> - Add level traces showing the STA/LTA ration for each trace.<br /> <b>· Apply to full dataset</b> - If marked entire loaded dataset will be analyzed. <br /> <b>· Scaling/Normalization method</b> - Select how output of the STA/LTA should be scaled.</ br> </p> <p> A helpfull description of how to tune the STA/LTA's parameters can be found in the the following ebook chapter by Amadej Trnkoczy: <a href="http://ebooks.gfz-potsdam.de/pubman/item/escidoc:4097:3/component/\ escidoc:4098/IS_8.1_rev1.pdf">Understanding and parameter setting of STA/LTA trigger algorithm</a> </p> </body> </html> '''
'Short Window [s]', 'swin', 30., 0.01, 2*h)) 'Ratio', 'ratio', 3., 1.1, 20.)) 'Level', 'level', 0.5, 0., 1.)) 'Processing Block length (rel. to long window)', 'block_factor', 10., 2., 100.,)) 'Show trigger level traces', 'show_level_traces', False)) 'Apply to full dataset', 'apply_to_all', False)) 'Variant', 'variant', 'centered', ['centered', 'right'])) 'Scaling/Normalization method', 'scalingmethod', '[0-1]', scalingmethods)) Switch('Detect on sum trace', 'apply_to_sum', False))
'''Main work routine of the snuffling.'''
self.cleanup()
swin, ratio = self.swin, self.ratio lwin = swin * ratio tpad = lwin
data_pile = self.get_pile() tmin, tmax = data_pile.get_tmin() + tpad, data_pile.get_tmax() - tpad viewer = self.get_viewer()
if not self.apply_to_all: vtmin, vtmax = viewer.get_time_range() tmin = max(vtmin, tmin) tmax = min(vtmax, tmax)
tinc = min(lwin * self.block_factor, tmax-tmin)
show_level_traces = self.show_level_traces
if show_level_traces and tmax-tmin > lwin * 150: self.error( 'Processing time window is longer than 50 x LTA window. ' 'Turning off display of level traces.') show_level_traces = False
markers = [] for traces in data_pile.chopper( tmin=tmin, tmax=tmax, tinc=tinc, tpad=tpad, want_incomplete=False, trace_selector=lambda x: not (x.meta and x.meta.get( 'tabu', False))):
sumtrace = None isum = 0 for tr in traces: if viewer.lowpass is not None: tr.lowpass(4, viewer.lowpass, nyquist_exception=True)
if viewer.highpass is not None: tr.highpass(4, viewer.highpass, nyquist_exception=True)
if self.variant == 'centered': tr.sta_lta_centered( swin, lwin, scalingmethod=scalingmethod_map[self.scalingmethod]) elif self.variant == 'right': tr.sta_lta_right( swin, lwin, scalingmethod=scalingmethod_map[self.scalingmethod])
tr.chop(tr.wmin, min(tr.wmax, tmax))
if not self.apply_to_sum: markers.extend(trace_to_pmarkers(tr, self.level, swin))
tr.set_codes(location='cg') tr.meta = {'tabu': True}
if sumtrace is None: ny = int((tr.tmax - tr.tmin) / data_pile.deltatmin) sumtrace = trace.Trace( deltat=data_pile.deltatmin, tmin=tr.tmin, ydata=num.zeros(ny))
sumtrace.set_codes( network='', station='SUM', location='cg', channel='') sumtrace.meta = {'tabu': True}
sumtrace.add(tr, left=None, right=None) isum += 1
if sumtrace is not None: sumtrace.ydata /= float(isum) if self.apply_to_sum: markers.extend( trace_to_pmarkers(sumtrace, self.level, swin, [('*', '*', '*', '*')]))
if show_level_traces: self.add_trace(sumtrace)
self.add_markers(markers)
if show_level_traces: self.add_traces(traces)
markers = [] tpeaks, apeaks = tr.peaks(level, swin) for t, a in zip(tpeaks, apeaks): ids = nslc_ids or [tr.nslc_id] mark = Marker(ids, t, t, ) print(mark, a) markers.append(mark)
return markers
|