How do I use pixel offset tracking?

[Xin-Wang-520]

I will use the dinsar method of isce and input topsApp.py insarApp.xml in the terminal to generate some files, and finally find dinsar related content in merge. I would like to ask whether the POT method is also generated in this step. Is it fine_offsets and coarse_offsets?If it is a follow-up process, how to get the final result, such as grd in the form of gmt drawing.

[Xin-Wang-520]

@EJFielding Hello, Dr., could you please explain the use of pot method in ISCE？

[EJFielding]

Hello Xin-Wang,

The topsApp.py application includes the option to run "dense offsets" that is another name for pixel offset tracking. You can add the "do dense offsets" True option to your topsApp.xml input parameter file.

You can look at the https://github.com/isce-framework/isce2/blob/main/applications/topsApp.py code to see all the options that can be added to the topsApp.xml file.

[EJFielding]

You can add this to your topsApp.xml file:

<!--# dense offsets-->
    <property name="do dense offsets">True</property>
    <property name="Ampcor window width">64</property>    <!-- EXAMPLE INPUT-->
    <property name="Ampcor window height">64</property>   <!-- EXAMPLE INPUT-->
    <property name="Filter window size">5</property>   <!-- EXAMPLE INPUT-->

[Xin-Wang-520]

You can add this to your topsApp.xml file:

<!--# dense offsets-->
    <property name="do dense offsets">True</property>
    <property name="Ampcor window width">64</property>    <!-- EXAMPLE INPUT-->
    <property name="Ampcor window height">64</property>   <!-- EXAMPLE INPUT-->
    <property name="Filter window size">5</property>   <!-- EXAMPLE INPUT-->

My topsApp.xml is as follows, if written correctly？

[Xin-Wang-520]

You can add this to your topsApp.xml file:

<!--# dense offsets-->
    <property name="do dense offsets">True</property>
    <property name="Ampcor window width">64</property>    <!-- EXAMPLE INPUT-->
    <property name="Ampcor window height">64</property>   <!-- EXAMPLE INPUT-->
    <property name="Filter window size">5</property>   <!-- EXAMPLE INPUT-->

Is it still like the picture below or is it not correct? How to modify it? Can Doctor write a complete topApp.xml of POT？

[EJFielding]

You can try running that. I think that the Ampcor window width and height are better at 64 for Sentinel-1.

[bqh6741]

Hello, Dr.,, I try according to what you said, the results obtained here, could you please help me to know how to use GMT in which file to extract the azimuth and range displacement, I try to extract the “dense offsets.bil.geo” and“ filt dense _ offsets.bil.geo ”these two files in the two bands respectively, but the feeling value is not quite right,, thank you very much

@EJFielding

[EJFielding]

Yes, the -10000 pixel value is a "no data" value that should be ignored, and there are also other pixels with bad measurements due to poor quality matches. The band1 offsets are in range (across-track) and the band2 offsets are in azimuth (along-track), and both are in units of the pixel size in those directions. For Sentinel-1 Interferometric Wideswath (IW) mode, the range pixel size is 2.3 meters and the azimuth pixel size is 14 meters.

See my paper on the 2019 Mw 7.1 and Mw 6.4 Ridgecrest earthquakes:

Fielding, E. J., Z. Liu, O. L. Stephenson, M. Zhong, C. Liang, A. Moore, S.-H. Yun, and M. Simons (2020). Surface deformation related to the 2019 Mw 7.1 and 6.4 Ridgecrest Earthquakes in California from GPS, SAR interferometry, and SAR pixel offsets, Seismol. Res. Lett. 91, no. 4, 2035-2046, doi:10.1785/0220190302.

[bqh6741]

Thank you very much for your response. I wish you every success in your research.！

[Xin-Wang-520]

Doctor, do you remember wrong? The chart I drew shows that it is Azimuth with a large deformation. Is it azimuth band 1 and range band 2？

[EJFielding]

Don't look at the outliers of the measurements. Those are bad matches. Range is band 1 and azimuth is band 2. You can compare to the ARIA displacement products: https://aria-share.jpl.nasa.gov/20230206_Turkey_EQ/Displacements/Sentinel1/

[Xin-Wang-520]

Don't look at the outliers of the measurements. Those are bad matches. Range is band 1 and azimuth is band 2. You can compare to the ARIA displacement products: https://aria-share.jpl.nasa.gov/20230206_Turkey_EQ/Displacements/Sentinel1/ I did the Turkey earthquake, which you say is Azimuth on the left and rang on the right. But according to the picture drawn by others, it should be rang on the left. What's the explanation? Isn't the band wrong or something？

[bqh6741]

Hello Doctor, I would like to ask for your help again. May I know how you removed the water area in your work? I have tried using @earth_mask_15s.grd in GMT to create a mask file, such as "gmt grdmath @earth_mask_01m_g 0 GT = ocean_land.grd". However, when I multiplied it with my offset_range.grd, it showed that the resolution was not consistent. I don't want to reduce the resolution of my offset result, and I can't find a higher resolution earth_mask file to download. I wonder if you have any methods or suggestions. If you have time after you finish your work, could you please kindly give me some guidance? Thank you very much for your help @EJFielding

[Xin-Wang-520]

Don't look at the outliers of the measurements. Those are bad matches. Range is band 1 and azimuth is band 2. You can compare to the ARIA displacement products: https://aria-share.jpl.nasa.gov/20230206_Turkey_EQ/Displacements/Sentinel1/

thank you. Doctor, how to eliminate those are bad matches. The water in the bottom left corner of the picture, as well as the other blue spots

[EJFielding]

One way to exclude bad matches is to apply a water mask such as with grdlandmask in GMT.

[DaiXiaoF]

不要看测量值的异常值。这些都是糟糕的比赛。范围为波段 1，方位角为波段 2。您可以与 ARIA 位移产品进行比较：https://aria-share.jpl.nasa.gov/20230206_Turkey_EQ/Displacements/Sentinel1/

谢谢。医生，如何消除这些是糟糕的匹配。图片左下角的水，以及其他蓝色斑点 Hello, may I ask how you visualize the deformation result? I use the command "gdal_translate -of GMT -b 2 filt_dense_offsets.bil.geo filt_dense_offsets.bil.geo.grd". Convert the.geo file to.grd file, and then use "ncview filt_dense_offsets.bil.geo.grd" command, the program reported an error, and then I use arcgis to open the.grd file, and the deformation graph is shown like this. Can you share with me your visualization method and your topApp profile? Thank you.

[EJFielding]

One thing to remember is that the ISCE2 output has "-10000" values for pixels where there was no match (on the edges of the scene, so you may need to adjust your color scale. I use the ".nc" suffix for the GMT format as it is a type of netCDF file. The ".grd" suffix might confuse your display software with another file type.

[17797112023]

Hello Doctor, I opened "filt_dense_offsets.bil.geo.vrt" in ArcGIS and the result was displayed like this, as if many blocks were splicing together. In addition, the conversion failed when the command gdal_translate -of GMT -b 2 filt_dense_offsets.bil.geo filt_dense_offsets.bil.geo.grd was executed in the merged folder.

[17797112023]

不要看测量值的异常值。这些都是糟糕的比赛。范围为波段 1，方位角为波段 2。您可以与 ARIA 位移产品进行比较：https://aria-share.jpl.nasa.gov/20230206_Turkey_EQ/Displacements/Sentinel1/

谢谢。医生，如何消除这些是糟糕的匹配。图片左下角的水，以及其他蓝色斑点 Hello, may I ask how you visualize the deformation result? I use the command "gdal_translate -of GMT -b 2 filt_dense_offsets.bil.geo filt_dense_offsets.bil.geo.grd". Convert the.geo file to.grd file, and then use "ncview filt_dense_offsets.bil.geo.grd" command, the program reported an error, and then I use arcgis to open the.grd file, and the deformation graph is shown like this. Can you share with me your visualization method and your topApp profile? Thank you.

I had the same problem

[DaiXiaoF]

是的，像素值是应忽略的“无数据”值，并且由于质量匹配不佳，还有其他像素的测量值不佳。波段 1 偏移在范围内（跨轨道），波段 2 偏移在方位角（沿轨道）中，并且两者都以这些方向上的像素大小为单位。对于Sentinel-1干涉宽幅（IW）模式，范围像素大小为2.3米，方位角像素大小为14米。-10000

请参阅我关于2019年Mw 7.1和Mw 6.4 Ridgecrest地震的论文：

Fielding, E. J., Z. Liu, O. L. Stephenson, M. Zhong, C. Liang, A. Moore, S.-H. Yun, and M. Simons (2020). Surface deformation related to the 2019 Mw 7.1 and 6.4 Ridgecrest Earthquakes in California from GPS, SAR interferometry, and SAR pixel offsets, Seismol. Res. Lett. 91, no. 4, 2035-2046, doi:10.1785/0220190302.

Hello, Professor. In the results obtained by the POT method, what does "dense offsets.bil. geo" and "filt dense _ offsets.bil. geo" represent respectively? Which file do I extract the deformation results from? Will the POT method generate an interferogram? How can I get the LOS directional deformation from the results generated by POT?

[17797112023]

您好，这个我也还在学习中，我也还没处理明白

---- Replied Message ---- | From | @.> | | Date | 09/09/2023 17:25 | | To | isce-framework/isce2 @.> | | Cc | 17797112023 @.>, Comment @.> | | Subject | Re: [isce-framework/isce2] How do I use pixel offset tracking? (Issue #662) |

是的，像素值是应忽略的“无数据”值，并且由于质量匹配不佳，还有其他像素的测量值不佳。波段 1 偏移在范围内（跨轨道），波段 2 偏移在方位角（沿轨道）中，并且两者都以这些方向上的像素大小为单位。对于Sentinel-1干涉宽幅（IW）模式，范围像素大小为2.3米，方位角像素大小为14米。-10000

请参阅我关于2019年Mw 7.1和Mw 6.4 Ridgecrest地震的论文：

Fielding, E. J., Z. Liu, O. L. Stephenson, M. Zhong, C. Liang, A. Moore, S.-H. Yun, and M. Simons (2020). Surface deformation related to the 2019 Mw 7.1 and 6.4 Ridgecrest Earthquakes in California from GPS, SAR interferometry, and SAR pixel offsets, Seismol. Res. Lett. 91, no. 4, 2035-2046, doi:10.1785/0220190302.

Hello, Professor. In the results obtained by the POT method, what does "dense offsets.bil. geo" and "filt dense _ offsets.bil. geo" represent respectively? Which file do I extract the deformation results from? Will the POT method generate an interferogram? How can I get the LOS directional deformation from the results generated by POT?

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: @.***>