search

yumorishita / LiCSBAS

LiCSBAS: InSAR time series analysis package using LiCSAR products
https://doi.org/10.3390/RS12030424
GNU General Public License v3.0
224 stars 109 forks source link

velocity map sliced in half #397

Open faraksoch1881 opened 2 weeks ago

faraksoch1881 commented 2 weeks ago

The area i'm interested is lies in two frames 121D_06267_131313 and 019D_06217_131313 but 121D_06267_131313 contains considerable amount of area i'm interested in. Hence i processed 121D_06267_131313 frame, but the output results are clipped in east-west direction in vel.mskd while vel contains whole image. What could be the source of error and how to resolved it ? 456421983_1766818537184864_6397117318222094522_n

as1234554321 commented 2 weeks ago

sorry oot @faraksoch1881

  1. how do you display this image on the computer?, on mine there is only 1 image of 1 part, not combined as above
  2. What version of licsbas are you using comet licsbas/licsbas/licsbas2?

thankyou

faraksoch1881 commented 2 weeks ago

sorry oot @faraksoch1881

  1. how do you display this image on the computer?, on mine there is only 1 image of 1 part, not combined as above
  2. What version of licsbas are you using comet licsbas/licsbas/licsbas2?

thankyou

The time series can be produced by using LiCSBAS_plot_ts.py command which displayed half of image. but when i check inside the folder TS_GMC1 folder there is two png images and when i viewed one image, i could see that all images were cut into half except vel. So that means my frame has whole image but somehow using mask or any other issue is cutting it into half which i have no idea how to solve so i posted that png image here

I'm using LiCsbas2 version.

yumorishita commented 1 week ago

Please upload the full log.

as1234554321 commented 1 week ago

hi sir, can i ask something about the result (png) in AOI. I have processed using batch first try, just changed do05op clip to y, everything else was default. png results below: mask_ts_masked

second try, change part 03-05 "y" , get gacos "y", others default do03op_GACOS="y" # y/n do04op_mask="y" # y/n do05op_clip="y" #y/n

p01_start_date="20190105" #default: 20141001 p01_end_date="20230309" # default: today p01_get_gacos="y" # y/n

The result is like the following PNG image: mask_ts_masked

In your opinion, which one is better, or do you have other suggestions for changing parameter values? thank you for your help sir.

yumorishita commented 1 week ago

The coverage looks better in the upper one. There seems to be no other significant difference.

as1234554321 commented 1 week ago

The coverage looks better in the upper one. There seems to be no other significant difference.

thankyou sir, true, no other significant difference because I processed it with parameter values are default. The difference lies in the gacos and masking process. in tutorial (the gacos and masking optional) This step is optional and highly recommended to do it but the result is a lot of white areas. Are there any suggestions for reducing the white/blank area?

thankyou sir

yumorishita commented 1 week ago

https://github.com/yumorishita/LiCSBAS/issues/244#issuecomment-2009502904

as1234554321 commented 1 week ago
  1. in LiCSBAS15_mask_ts.py is it better to high percentage of kept pixels or masked pixels?
  2. Is there a minimum standard percentage value that is good for masked pixels or kept pixels? sorry for basic questions thankyou
faraksoch1881 commented 1 week ago

Please upload the full log.

202409021636batch_LiCSBAS_01_16.log Here is the full log where my map is cut into half in horizontally.

faraksoch1881 commented 1 week ago

hi sir, can i ask something about the result (png) in AOI. I have processed using batch first try, just changed do05op clip to y, everything else was default. png results below: mask_ts_masked

second try, change part 03-05 "y" , get gacos "y", others default do03op_GACOS="y" # y/n do04op_mask="y" # y/n do05op_clip="y" #y/n

p01_start_date="20190105" #default: 20141001 p01_end_date="20230309" # default: today p01_get_gacos="y" # y/n

The result is like the following PNG image: mask_ts_masked

In your opinion, which one is better, or do you have other suggestions for changing parameter values? thank you for your help sir.

I also tried to use GACOS data by using batch processing but it failed saying "No Gacos folder". So did you first download the GACOS data inside GACOS folder and then ran the batch processing or else?

as1234554321 commented 1 week ago

I also tried to use GACOS data by using batch processing but it failed saying "No Gacos folder". So did you first download the GACOS data inside GACOS folder and then ran the batch processing or else?

I run the batch then automatically download the gacos data

faraksoch1881 commented 1 week ago

I also tried to use GACOS data by using batch processing but it failed saying "No Gacos folder". So did you first download the GACOS data inside GACOS folder and then ran the batch processing or else?

I run the batch then automatically download the gacos data

could you please share me your batch file here !!

faraksoch1881 commented 1 week ago

Hi Dr.yumorishita

Here is the full log 202409021636batch_LiCSBAS_01_16.log

yumorishita commented 1 week ago

Was the image you showed at the top mask_ts_masked.png? If so, please show mask_ts.png.

faraksoch1881 commented 1 week ago

Was the image you showed at the top mask_ts_masked.png? If so, please show mask_ts.png.

I'm off my computer right now. but masked png look like the image from first row second column. Like this

Screenshot 2024-09-05 070655
yumorishita commented 1 week ago

No, mask_ts.png.

faraksoch1881 commented 1 week ago

No, mask_ts.png.

I will upload mask_ts.png few hours later as away from my computer now.

faraksoch1881 commented 1 week ago

mask_ts Hi prof. I have attached here mask_ts.png

yumorishita commented 1 week ago

Try increasing p15_n_loop_err_thre. https://github.com/yumorishita/LiCSBAS/blob/5db6516a5a9bad31445d916e61e7ee03c3daa1fd/batch_LiCSBAS.sh#L58

faraksoch1881 commented 1 week ago

Try increasing p15_n_loop_err_thre.

https://github.com/yumorishita/LiCSBAS/blob/5db6516a5a9bad31445d916e61e7ee03c3daa1fd/batch_LiCSBAS.sh#L58

Thank you Dr. yumorishita. I changed it to 8, results was same then to 12 and pixel seems to appear and when i finally change it to 20. I could see the full map. would you please kindly tell me what was wrong and how increasing '20' solved the problem...and since putting 20 seems to solve the problem so if i keep increasing the results will improve or how much i can increase this value ? Is there upper limit? mask_ts

My log files are for 21, 12 and 8 respectively 202409061118batch_LiCSBAS_01_16.log 202409061103batch_LiCSBAS_01_16.log 202409061045batch_LiCSBAS_01_16.log

yumorishita commented 1 week ago

Please read the LiCSBAS paper, especially sections 2.4.2 and 2.4.5.

faraksoch1881 commented 1 week ago

Thank you Dr. Yumorishita for all your help and guidance.

faraksoch1881 commented 4 days ago

Please read the LiCSBAS paper, especially sections 2.4.2 and 2.4.5.

Though i markdown as closed, a final doubt i have is I look at the png files inside folder GEOCml1clip and found two interferograms divided as center by a line. is that a source of this problem? will removing them and running processing can solve it and give better results ? Below are the two interferograms which i suspect be the culprit.

20171003_20171015 unw 20171027_20171202 unw

The value of interferogram based on 11ifg_stats.txt is


# ifg dates         bperp   dt unw_cov  coh_av
20171003_20171015    -9.1   12   0.954   0.274 
20171027_20171202   -37.6   36   0.971   0.254
yumorishita commented 1 day ago

I think the problem is in the LiCSAR products. Please contact COMET.