Less than a week ago, we were able to show some of our work to Dr. Monica D. Kohler, a Senior Research Fellow in mechanical and civil Engineering, Caltech. She being a seismic expert in the CSN group, was interested in the effects of seismic waves on tall buildings, and whether it was possible to observe oscillations of tall building in the frequency transforms. She suggested the use of Spectrograms: the Frequency Time Distribution of the seismic data from the SAC files.
Fortunately, the Obspy package itself contains 'obspy.imaging.spectrogram' method, which allows to read and plot spectrograms directly from SAC files. This optimized for speed method takes stream and bunch of other arguments(for plotting) as input and plots spectrogram within a second for 15 min SAC data.
Though some work has to be done if you wish to transfer these plots to Matplotlib.pyplot, or arrange them on Subplots. This because obspy uses its own plotting methods which doesn't plot on a subplot by default.
The workaround is to use 'plt.gca()' to get axis, and then use obspy to plot. Also make sure to give 'show=False' as an argument to spectrogram, to popping up windows at every subplot. A sample code looks for single subplot looks like this:
Fortunately, the Obspy package itself contains 'obspy.imaging.spectrogram' method, which allows to read and plot spectrograms directly from SAC files. This optimized for speed method takes stream and bunch of other arguments(for plotting) as input and plots spectrogram within a second for 15 min SAC data.
 |
| A Spectrogram showing activity during Anza (Mar 11, 2013) tremors for all 3 channels. Notice the red and yellow shades mark the beginning of quake, when according to K Sigma |
Though some work has to be done if you wish to transfer these plots to Matplotlib.pyplot, or arrange them on Subplots. This because obspy uses its own plotting methods which doesn't plot on a subplot by default.
The workaround is to use 'plt.gca()' to get axis, and then use obspy to plot. Also make sure to give 'show=False' as an argument to spectrogram, to popping up windows at every subplot. A sample code looks for single subplot looks like this:
plt.subplot(311)
ax_HNN=plt.gca()
spectrogram(tr_HNN.data,tr_HNN.stats.sampling_rate,show=False,axes=ax_HNN)
ax_HNN=plt.gca()
spectrogram(tr_HNN.data,tr_HNN.stats.sampling_rate,show=False,axes=ax_HNN)
We are planning to plot spectrograms for some CSN sensors in Los Angeles Department of Water & Power (LADWP), to figure some correlation between frequency and respective floor of installation of the sensors.
hello, i am following your post indications as how to plot the spectogram, and works fine, but i dont see anything plotted nor the screen or to a file, how do you save your plot??
ReplyDeletethanks