Open silverbean-j opened 11 months ago
The order of inputs should be as follows: first images, then cameras, so:
/data/test/ref_img.tif /data/test/tar_img.tif /data/test/ref_img.xml /data/test/tar_img.xml
followed by output prefix.
Somehow, however, in your case the workflow continues. Then it runs into large disparity issues:
Refined search range using a histogram and --outlier-removal-params: (Origin: (-8285, -573) width: 14754 height: 1031)
That is bad.
I will suggested using the mapprojected images workflow: https://stereopipeline.readthedocs.io/en/latest/next_steps.html#mapproj-example
One should ensure dem_geoid gets used as stated there to correct any input DEM if applicable, then one should mapproject left and right image at same resolution as suggested further down. It is strongly suggested to examine the two produced images. They should be very similar if overlaid. And note that the invocation of parallel_stereo will be different, as shown in that example, with a DEM at the end.
Hello,
Thank you for providing a response in the first place. I'm trying to follow the solution you provided, but it seems like either I'm not understanding it properly or I might have conveyed something incorrectly.
Firstly, according to that document, it seems I should have a DEM for the respective imagery. However, I don't have the DEM for that imagery (hence, it needs to be generated). Previously, the command I provided actually resulted in the imagery being successfully produced (I'll attach the DEM that I actually generated additionally). However, currently, it's not working.
I even tried cutting the imagery as per the instructions in the document, but it still isn't working. I'm not sure if the disparity values are still causing an issue.
Could it be that I'm misunderstanding something? Also, just to note, I don't have the .cub file.
Thank you so much for your assistance!
You need .cub files only for outer planets, not for Earth. See if the documentation at https://stereopipeline.readthedocs.io/en/latest/examples.html can help you get started. Choose the right section for your data. Maybe you have RPC data.
Normally with ASP you don't need an initial DEM if you have good left and right images that are quite similar, have good convergence angle and good features ( https://stereopipeline.readthedocs.io/en/latest/examples/stereo_pairs.html),
A DEM is needed only when the disparity range is too large, but that assumes you already have a good dataset.
As to where to get initial DEM, if you need them, that is discussed at https://stereopipeline.readthedocs.io/en/latest/next_steps.html#mapproj-example and further down.
I suggest you read these documentation pages, and practice some more.
Check your left and right input images. They should be similar. You can also try running clips, using stereo_gui ( https://stereopipeline.readthedocs.io/en/latest/tools/stereo_gui.html).
Check for the convergence angle printed on screen.
If still no luck, you can share the precise command you used, and maybe a screenshot from stereo_gui showing left and right images, and the stereo convergence angle as printed on screen.
On Sun, Dec 10, 2023 at 8:10 PM silverbean-j @.***> wrote:
Hello,
Thank you for providing a response in the first place. I'm trying to follow the solution you provided, but it seems like either I'm not understanding it properly or I might have conveyed something incorrectly.
Firstly, according to that document, it seems I should have a DEM for the respective imagery. However, I don't have the DEM for that imagery (hence, it needs to be generated). Previously, the command I provided actually resulted in the imagery being successfully produced (I'll attach the DEM that I actually generated additionally). However, currently, it's not working.
ASP_Crop_log.txt https://github.com/NeoGeographyToolkit/StereoPipeline/files/13630855/ASP_Crop_log.txt
I even tried cutting the imagery as per the instructions in the document, but it still isn't working. I'm not sure if the disparity values are still causing an issue.
Could it be that I'm misunderstanding something? Also, just to note, I don't have the .cub file.
Thank you so much for your assistance!
ASP_MS_Result_with_QTReader.png (view on web) https://github.com/NeoGeographyToolkit/StereoPipeline/assets/97012621/8d71133e-bfe6-4c6b-86f1-3c55b990ce0c ASP_MS_Result_with_QGIS.png (view on web) https://github.com/NeoGeographyToolkit/StereoPipeline/assets/97012621/3ad58f2b-e257-4b34-a215-2c190531ca11
— Reply to this email directly, view it on GitHub https://github.com/NeoGeographyToolkit/StereoPipeline/issues/419#issuecomment-1849294260, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAKDU3EE5A3RX4ECTHCE5BLYI2BSZAVCNFSM6AAAAABAKOG3FOVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTQNBZGI4TIMRWGA . You are receiving this because you commented.Message ID: @.***>
Hello team,
I'm currently working on generating a DEM using ASP. However, I've encountered an issue. The "-PC_sub.tif" output was generated successfully, but the "-PC.tif" output wasn't created, and an error message occurred. Previously, during testing, I used the version
StereoPipeline-3.4.0-alpha-2023-09-17-x86_64-Linux
, but now I'm encountering this problem using theStereoPipeline-3.3.0-Linux version
. I'm running ASP within a Docker container using an Ubuntu image.I formulated the command following Chapter 5.2 "Processing raw" in your tutorial:
The imagery I'm working with is similar to WorldView, and I aim to generate both Panchromatic and Multi-Spectral images. Could you please verify if the command I'm using is incorrect or flawed? I'd appreciate it.
Additionally, if you require the images and RPC files I'm using, could you provide an email address where I can send them?
Unfortunately, even using | tee stereo_output.log, the log file isn't generated, so I'll attach the contents displayed in the command prompt. ASP_log.txt
I would be truly grateful for your assistance. Thank you.