How to get the errors from the unwrapped interferogram about licsbas HOT 7 CLOSED

For example:

import gdal
import numpy as np

geotiff = 'date1_date2.geo.unw.tif'
data = gdal.Open(geotiff).ReadAsArray()
data[data==0] = np.nan
print('STD: {:.2f} rad'.format(np.nanstd(data)))

from licsbas.

from licsbas.

I see. Assuming that the std can be calculated by (1-coh**2)/(2*coh**2) (e.g., Spaans and Hooper, 2016; https://doi.org/10.1002/2015JB012752):

import gdal
import numpy as np

cc_geotiff = 'data1_data2.geo.cc.tif'
cc = gdal.Open(cc_geotiff).ReadAsArray().astype(np.float32)/255
cc[cc<0.01] = 0.01
cc[cc>0.99] = 0.99
std = np.sqrt((1-cc**2)/(2*cc**2))

from licsbas.

Thank you for your answer.
I'm curious about why the cc values are grouped into < 0.01 and > 0.99 ?
Besides, why the cc value divides 255, and then use the results to calculate the std? I found cc values are limited into a range of -128 to 127. It means, after dividing 255, the cc value should be in a span of -0.50196 to 0.49803. I'm not sure they should be grouped into two groups by abs(cc) < 0.01 and abs(cc) > 0.99, right?

from licsbas.

0.01 and 0.99 are just for avoiding infinity and zero division in std.

The format of cc is uint8 which ranges 0-255 (see below). Perhaps you have read the cc as int8

https://comet.nerc.ac.uk/comet-lics-portal-product-details/
“geo.cc.tif: This is the coherence image of the interferometric pair. The values varies between 0-255 in which 0 refers to the lowest coherence values and 255 indicates the highest values of coherence.”

from licsbas.

Hi, I would like to know. If I prepare the geo.cc.tif for LiCSBAS by myself (Because LiCSAR don't cover my area). I should make the cc between 0-1 or 0-255? I use cc_wave and it give me the result is between 0-1. Do I need to change it to 0-255 for LiCSBAS?
Thank you

from licsbas.

Either is fine because both formats can be handled.
https://github.com/yumorishita/LiCSBAS/blob/master/bin/LiCSBAS02_ml_prep.py#L425
https://github.com/yumorishita/LiCSBAS/blob/master/bin/LiCSBAS11_check_unw.py#L225

from licsbas.