Modified plot output.
This commit is contained in:
parent
f4a76680ad
commit
2ef647251e
@ -9,7 +9,6 @@
|
|||||||
"""
|
"""
|
||||||
|
|
||||||
import warnings
|
import warnings
|
||||||
|
|
||||||
import matplotlib.pyplot as plt
|
import matplotlib.pyplot as plt
|
||||||
import numpy as np
|
import numpy as np
|
||||||
from obspy.core import Stream, UTCDateTime
|
from obspy.core import Stream, UTCDateTime
|
||||||
@ -410,6 +409,8 @@ def getnoisewin(t, t1, tnoise, tgap):
|
|||||||
# get noise window
|
# get noise window
|
||||||
inoise, = np.where((t <= max([t1 - tgap, 0])) \
|
inoise, = np.where((t <= max([t1 - tgap, 0])) \
|
||||||
& (t >= max([t1 - tnoise - tgap, 0])))
|
& (t >= max([t1 - tnoise - tgap, 0])))
|
||||||
|
if np.size(inoise) < 1:
|
||||||
|
inoise, = np.where((t>=t[0]) & (t<=t1))
|
||||||
if np.size(inoise) < 1:
|
if np.size(inoise) < 1:
|
||||||
print ("getnoisewin: Empty array inoise, check noise window!")
|
print ("getnoisewin: Empty array inoise, check noise window!")
|
||||||
|
|
||||||
@ -605,7 +606,7 @@ def wadaticheck(pickdic, dttolerance, iplot):
|
|||||||
wfitflag = 1
|
wfitflag = 1
|
||||||
|
|
||||||
# plot results
|
# plot results
|
||||||
if iplot > 1:
|
if iplot > 0:
|
||||||
plt.figure(iplot)
|
plt.figure(iplot)
|
||||||
f1, = plt.plot(Ppicks, SPtimes, 'ro')
|
f1, = plt.plot(Ppicks, SPtimes, 'ro')
|
||||||
if wfitflag == 0:
|
if wfitflag == 0:
|
||||||
@ -756,11 +757,12 @@ def checkPonsets(pickdic, dttolerance, iplot):
|
|||||||
[xjack, PHI_pseudo, PHI_sub] = jackknife(Ppicks, 'VAR', 1)
|
[xjack, PHI_pseudo, PHI_sub] = jackknife(Ppicks, 'VAR', 1)
|
||||||
# get pseudo variances smaller than average variances
|
# get pseudo variances smaller than average variances
|
||||||
# (times safety factor), these picks passed jackknife test
|
# (times safety factor), these picks passed jackknife test
|
||||||
ij = np.where(PHI_pseudo <= 2 * xjack)
|
ij = np.where(PHI_pseudo <= 5 * xjack)
|
||||||
# these picks did not pass jackknife test
|
# these picks did not pass jackknife test
|
||||||
badjk = np.where(PHI_pseudo > 2 * xjack)
|
badjk = np.where(PHI_pseudo > 5 * xjack)
|
||||||
badjkstations = np.array(stations)[badjk]
|
badjkstations = np.array(stations)[badjk]
|
||||||
print ("checkPonsets: %d pick(s) did not pass jackknife test!" % len(badjkstations))
|
print ("checkPonsets: %d pick(s) did not pass jackknife test!" % len(badjkstations))
|
||||||
|
print(badjkstations)
|
||||||
|
|
||||||
# calculate median from these picks
|
# calculate median from these picks
|
||||||
pmedian = np.median(np.array(Ppicks)[ij])
|
pmedian = np.median(np.array(Ppicks)[ij])
|
||||||
@ -795,19 +797,22 @@ def checkPonsets(pickdic, dttolerance, iplot):
|
|||||||
|
|
||||||
checkedonsets = pickdic
|
checkedonsets = pickdic
|
||||||
|
|
||||||
if iplot > 1:
|
if iplot > 0:
|
||||||
p1, = plt.plot(np.arange(0, len(Ppicks)), Ppicks, 'r+', markersize=14)
|
p1, = plt.plot(np.arange(0, len(Ppicks)), Ppicks, 'ro', markersize=14)
|
||||||
p2, = plt.plot(igood, np.array(Ppicks)[igood], 'g*', markersize=14)
|
if len(badstations) < 1 and len(badjkstations) < 1:
|
||||||
|
p2, = plt.plot(np.arange(0, len(Ppicks)), Ppicks, 'go', markersize=14)
|
||||||
|
else:
|
||||||
|
p2, = plt.plot(igood, np.array(Ppicks)[igood], 'go', markersize=14)
|
||||||
p3, = plt.plot([0, len(Ppicks) - 1], [pmedian, pmedian], 'g',
|
p3, = plt.plot([0, len(Ppicks) - 1], [pmedian, pmedian], 'g',
|
||||||
linewidth=2)
|
linewidth=2)
|
||||||
for i in range(0, len(Ppicks)):
|
for i in range(0, len(Ppicks)):
|
||||||
plt.text(i, Ppicks[i] + 0.2, stations[i])
|
plt.text(i, Ppicks[i] + 0.01, '{0}'.format(stations[i]))
|
||||||
|
|
||||||
plt.xlabel('Number of P Picks')
|
plt.xlabel('Number of P Picks')
|
||||||
plt.ylabel('Onset Time [s] from 1.1.1970')
|
plt.ylabel('Onset Time [s] from 1.1.1970')
|
||||||
plt.legend([p1, p2, p3], ['Skipped P Picks', 'Good P Picks', 'Median'],
|
plt.legend([p1, p2, p3], ['Skipped P Picks', 'Good P Picks', 'Median'],
|
||||||
loc='best')
|
loc='best')
|
||||||
plt.title('Check P Onsets')
|
plt.title('Jackknifing and Median Tests on P Onsets')
|
||||||
plt.show()
|
plt.show()
|
||||||
raw_input()
|
raw_input()
|
||||||
|
|
||||||
@ -943,8 +948,18 @@ def checkZ4S(X, pick, zfac, checkwin, iplot):
|
|||||||
isignal = getsignalwin(tz, pick, checkwin)
|
isignal = getsignalwin(tz, pick, checkwin)
|
||||||
|
|
||||||
# calculate energy levels
|
# calculate energy levels
|
||||||
|
try:
|
||||||
zcodalevel = max(absz[isignal])
|
zcodalevel = max(absz[isignal])
|
||||||
|
except:
|
||||||
|
ii = np.where(isignal <= len(absz))
|
||||||
|
isignal = isignal[ii]
|
||||||
|
zcodalevel = max(absz[isignal - 1])
|
||||||
|
try:
|
||||||
encodalevel = max(absen[isignal])
|
encodalevel = max(absen[isignal])
|
||||||
|
except:
|
||||||
|
ii = np.where(isignal <= len(absen))
|
||||||
|
isignal = isignal[ii]
|
||||||
|
encodalevel = max(absen[isignal - 1])
|
||||||
|
|
||||||
# calculate threshold
|
# calculate threshold
|
||||||
minsiglevel = encodalevel * zfac
|
minsiglevel = encodalevel * zfac
|
||||||
|
Loading…
Reference in New Issue
Block a user