Hello, I figured a way around this, and to be completely honest this issue was coming up from using the DIMAP files (for Pleiades NEO and HR) for the panchromatic and multispectral images and so were loaded with the DIMAP driver.
1. My images for input to pansharpening had no geotransform, as they were generated through several math processing steps using the DIMAP files (for Pleiades NEO and HR) for the panchromatic and multispectral images and so were loaded with the DIMAP driver. The Geotransform was of the form (0.0, 1.0, 0.0, 0.0, 0.0, 1.0), and this makes gdal_pansharpen unhappy. 2. I set the geotransform to that of one of the tif files (formatted with RxCx, using the first one) that contains a geotransform that makes sense for the origin of the image and RPCs, i.e., R1C1 i.e., the upper leftmost image in the delivery. This geotransform matches well enough to make pansharpening happy across the full dataset extent. 3. On the pansharpened result, remove the geotransform with gdal_edit.py -unsetgt <file> and copy in the RPCs from the original DIMAP file. This produces a file that can be used downstream for orthorectification or any other operations utilizing the RPCs. Thanks for the input! Cheers, Joe ________________________________ From: Joe McGlinchy <j...@aidash.com> Sent: Tuesday, February 25, 2025 14:53 To: Rahkonen Jukka <jukka.rahko...@maanmittauslaitos.fi>; Howard Butler <how...@hobu.co>; gdal-dev@lists.osgeo.org <gdal-dev@lists.osgeo.org> Subject: Re: [gdal-dev] pansharpening on images with RPCs post GDAL 3.6 Hi Jukka, That is almost exactly along the lines of what I was thinking: fake the geotransform and CRS data to get past the non-existent CRS data within this level of imagery, and then copy back in the appropriate metadata. Thanks for your response! I'll give it a try. Cheers, Joe ________________________________ From: Rahkonen Jukka <jukka.rahko...@maanmittauslaitos.fi> Sent: Tuesday, February 25, 2025 14:51 To: Joe McGlinchy <j...@aidash.com>; Howard Butler <how...@hobu.co>; gdal-dev@lists.osgeo.org <gdal-dev@lists.osgeo.org> Subject: Re: [gdal-dev] pansharpening on images with RPCs post GDAL 3.6 EXTERNAL SENDER. DO NOT click links, or open attachments, if the sender is unknown, or the message seems suspicious in any way. DO NOT provide your user ID or Password. Hi, I have never done anything with RPCs, but maybe you could remove RPCs with something like “gdal_translate -co profile=baseline input.tif output.tif” (delete the aux.xml or use that magic config option that I do not remember out of my head), define some fake geotransformation by defining same ullr coordinates and crs for both PAN and MS images, then pansharpen, and finally copy the RPCs back like you have done before. I apologize if this does not make any sense. -Jukka Rahkonen- Lähettäjä: gdal-dev <gdal-dev-boun...@lists.osgeo.org> Puolesta Joe McGlinchy via gdal-dev Lähetetty: tiistai 25. helmikuuta 2025 23.38 Vastaanottaja: Howard Butler <how...@hobu.co>; gdal-dev@lists.osgeo.org Aihe: Re: [gdal-dev] pansharpening on images with RPCs post GDAL 3.6 My apologies for the repost, was not aware that is against the rules of the list-serv. noted! The workflow I described, pansharpening on data in image space, is a common practice. This being deprecated was surprising to see as a user, but I understand the reasoning in the migration notes. If I end up determining a workflow that allows similar functionality with gdal > 3.6, I'll be sure to respond accordingly for the benefit of the community. As for the rest of the reply, from one working professional to another, I'll look past it. thanks! -Joe ________________________________ From: Howard Butler <how...@hobu.co<mailto:how...@hobu.co>> Sent: Tuesday, February 25, 2025 14:25 To: Joe McGlinchy <j...@aidash.com<mailto:j...@aidash.com>> Cc: gdal-dev@lists.osgeo.org<mailto:gdal-dev@lists.osgeo.org> <gdal-dev@lists.osgeo.org<mailto:gdal-dev@lists.osgeo.org>> Subject: Re: [gdal-dev] pansharpening on images with RPCs post GDAL 3.6 EXTERNAL SENDER. DO NOT click links, or open attachments, if the sender is unknown, or the message seems suspicious in any way. DO NOT provide your user ID or Password. Joe, Ask the AI from AiDASH to tell you how to do it? Oh, right, the GDAL list is your training corpus, not the inference data. Sorry for being snotty, but please refrain from reposting questions to the list like this. If no one answered your question the first time, the LLM hallucinates and pukes out-of-order GDAL arguments at you, and you cannot dig through the source code and documentation to figure it out yourself, you should hire a consultant. https://gdal.org/en/stable/community/service_providers.html is a list of a few, and inspecting the commit history of the repository should give you some inspiration for others to privately contact. Howard On Feb 25, 2025, at 2:51 PM, Joe McGlinchy via gdal-dev <gdal-dev@lists.osgeo.org<mailto:gdal-dev@lists.osgeo.org>> wrote: Any advice on how to pansharpen image space data with RPCs in gdal >= 3.7?? Repeating below. Hello, I saw in the migration notes from GDAL 3.6 -> 3.7: - Pansharpening now requires that panchromatic and multispectral bands have valid geotransform (in early versions, it was assumed in the case of missing geotransform that they covered the same geospatial extent). The undocumented VRT pansharpened MSShiftX and MSShiftY options (and the corresponding C++ GDALPansharpenOptions::dfMSShiftX and dfMSShiftY members) have been removed, due to using the inverted convention as one would expect, and being sub-par solution compared to using geotransform to correlate pixels of panchromatic and multispectral bands. Previously, I had been taking a PAN and MS set of images with RPCs and running gdal_pansharpen, then copying over the RPCs from the PAN file, then continuing on in my workflow. This no long works past GDAL 3.6, so is there a recommended workflow for doing this? I've been able to "gdalwarp -rpc" the files prior to calling gdal_pansharpen, but i'd like to retain the image in image space with RPCs if possible. thanks, Joe ________________________________ From: Joe McGlinchy <j...@aidash.com<mailto:j...@aidash.com>> Sent: Friday, February 14, 2025 10:37 To: gdal-dev@lists.osgeo.org<mailto:gdal-dev@lists.osgeo.org> <gdal-dev@lists.osgeo.org<mailto:gdal-dev@lists.osgeo.org>> Subject: pansharpening on images with RPCs post GDAL 3.6 Hello, I saw in the migration notes from GDAL 3.6 -> 3.7: - Pansharpening now requires that panchromatic and multispectral bands have valid geotransform (in early versions, it was assumed in the case of missing geotransform that they covered the same geospatial extent). The undocumented VRT pansharpened MSShiftX and MSShiftY options (and the corresponding C++ GDALPansharpenOptions::dfMSShiftX and dfMSShiftY members) have been removed, due to using the inverted convention as one would expect, and being sub-par solution compared to using geotransform to correlate pixels of panchromatic and multispectral bands. Previously, I had been taking a PAN and MS set of images with RPCs and running gdal_pansharpen, then copying over the RPCs from the PAN file, then continuing on in my workflow. This no long works past GDAL 3.6, so is there a recommended workflow for doing this? I've been able to "gdalwarp -rpc" the files prior to calling gdal_pansharpen, but i'd like to retain the image in image space with RPCs if possible. thanks, Joe _______________________________________________ gdal-dev mailing list gdal-dev@lists.osgeo.org<mailto:gdal-dev@lists.osgeo.org> https://lists.osgeo.org/mailman/listinfo/gdal-dev
_______________________________________________ gdal-dev mailing list gdal-dev@lists.osgeo.org https://lists.osgeo.org/mailman/listinfo/gdal-dev
