alicevision
[question] Does Meshroom support building photogrammetry model from images split from 360 video?
Describe the problem I tried to follow the youtube [https://www.youtube.com/watch?v=Acv7lsCixZM&ab_channel=J%C3%A9r%C3%B4meTabeaud)] which it uses a 360 video and stitched a photogrammetry model. In the video it uses another software but i'm wondering if Meshroom supports creating model from spilt 360 images.
Screenshots If applicable, add screenshots to help explain your problem.
Dataset If applicable, add a link or few images to help better understand where the problem may come from.
Log "[2024-11-19 14:23:22.790558] [0x00005620] [trace] Embedded OCIO configuration file: 'C:\Users\braundtlau\Downloads\Meshroom-2023.3.0-win64\Meshroom-2023.3.0\aliceVision/share/aliceVision/config.ocio' found. Program called with the following parameters:
- evCorrection = 0
- input = "d:/Projects/Meshroom/kt_test1/test2/MeshroomCache/StructureFromMotion/864faad48f8eac9a9ea23e930858b27612a6ab58/sfm.abc"
- maskExtension = "png"
- maxCoresAvailable = Unknown Type "unsigned int" (default)
- maxMemoryAvailable = 18446744073709551615 (default)
- output = "d:/Projects/Meshroom/kt_test1/test2/MeshroomCache/PrepareDenseScene/e6a8c0b029e1c729258950b08bd6aa535af9f839"
- outputFileType = "exr"
- rangeSize = 40
- rangeStart = 0
- saveMatricesTxtFiles = 0
- saveMetadata = 1
- verboseLevel = "info"
Hardware : Detected core count : 16 OpenMP will use 16 cores Detected available memory : 42338 Mo
[14:23:22.801274][warning] The number of rigs is incoherent: [14:23:22.801274][warning] 0 rigs declared and 1 rigs used. [14:23:22.801274][error] The input SfMData file 'd:/Projects/Meshroom/kt_test1/test2/MeshroomCache/StructureFromMotion/864faad48f8eac9a9ea23e930858b27612a6ab58/sfm.abc' cannot be read. " )
Desktop (please complete the following and other pertinent information):
- OS: [e.g. win 10, osx, ]
- Python version [e.g. 2.6]
- Qt/PySide version [e.g. 5.15.2]
- Meshroom version: please specify if you are using a release version or your own build
- Binary version (if applicable) [e.g. 2019.1]
- Commit reference (if applicable) [e.g. 08ddbe2]
Additional context Add any other context about the problem here.
I'm not sure what's your error. The main error message is "The number of rigs is incoherent". But yes, it should work if you drag&drap of all the sub-folders of the rig in input of a photogrammetry pipeline. Please provide more details, if you want some help for troubleshooting.
Hello; let me re-open an follow up that ticket, I am following more or less the same procedure than the initial post from @sinki-blau starting from 360 panorama (small test project, less than 30 raw pano generated by a consumer grade camera, more exactly 29 equirectangular images)
Split360images.exe node is set up to create the virtual rig, no pb with that step. I am testing a split into 8 pinhole virtual camera rig. then come FeatureExtraction node and ImageMatching node, here again without any issue. but when I come to FeatureMatching, issue is pending with a limitation of *.feat and *.desc of 40 images... even thought the virtual rig is composed of 29 * 8 cameras = 232 pinhole images I do not really catch why that node is limited to 40 images for that KEY POINT detection... and tried to figure that out digging into the github... I found that https://github.com/alicevision/meshroom/issues/1240 but it's not clear for me and the workaround proposed by @remmel is not working (copy-paste)... because I still do not detect keypoint more than 40 images...
Should you wish to open a now ticket, please advice. The idea is to prove out we can align 360° camera using the virtual rig concept and Split360images.exe node
