Error on Meshing Node [Depth map fusion gives an empty result]

[Kanagawa-Wave]

Problem I am reconstructing on a set of 17 images taken on my phone. And the application errored on node "Meshing", returning "Depth map fusion gives an empty result." Seems like my depth maps weren't properly generated? Any clue about how can I fix it? Btw this is my first time using meshroom, so maybe I did sth wrong?

Log

ERROR:root:Error on node computation: Error on node "Meshing_1":
Log:
Program called with the following parameters:
 * addLandmarksToTheDensePointCloud = 0
 * angleFactor = 15
 * boundingBox =  Unknown Type "struct BoundingBox"
 * colorizeOutput = 0
 * contributeMarginFactor = 2
 * densifyNbBack = 0 (default)
 * densifyNbFront = 0 (default)
 * densifyScale = 1 (default)
 * depthMapsFolder = "C:/Users/Yuqi_Lu/Documents/Dev/3D-Scans/Xbox Controller/Reconstruction/MeshroomCache/DepthMapFilter/cbff7ffae6ecc70a2f48e50129932a7c8f6162c3"
 * estimateSpaceFromSfM = 1
 * estimateSpaceMinObservationAngle = 10
 * estimateSpaceMinObservations = 3
 * exportDebugTetrahedralization = 0
 * fullWeight = 1
 * helperPointsGridSize = 10
 * input = "C:/Users/Yuqi_Lu/Documents/Dev/3D-Scans/Xbox Controller/Reconstruction/MeshroomCache/StructureFromMotion/b6dd616ee696b2fafe71042d06e36b6194d2d9ac/sfm.abc"
 * invertTetrahedronBasedOnNeighborsNbIterations = 10
 * maskBorderSize = 1 (default)
 * maskHelperPointsWeight = 0 (default)
 * maxInputPoints = 50000000
 * maxNbConnectedHelperPoints = 50
 * maxPoints = 5000000
 * maxPointsPerVoxel = 1000000
 * minAngleThreshold = 1
 * minSolidAngleRatio = 0.2
 * minStep = 2
 * minVis = 2 (default)
 * nPixelSizeBehind = 4
 * nbSolidAngleFilteringIterations = 2
 * output = "C:/Users/Yuqi_Lu/Documents/Dev/3D-Scans/Xbox Controller/Reconstruction/MeshroomCache/Meshing/f875c14aba8ad9a2ddae785ba6a75a689e6d53d0/densePointCloud.abc"
 * outputMesh = "C:/Users/Yuqi_Lu/Documents/Dev/3D-Scans/Xbox Controller/Reconstruction/MeshroomCache/Meshing/f875c14aba8ad9a2ddae785ba6a75a689e6d53d0/mesh.obj"
 * partitioning =  Unknown Type "enum EPartitioningMode"
 * pixSizeMarginFinalCoef = 4
 * pixSizeMarginInitCoef = 2
 * refineFuse = 1
 * repartition =  Unknown Type "enum ERepartitionMode"
 * saveRawDensePointCloud = 0
 * seed =  Unknown Type "unsigned int"
 * simFactor = 15
 * simGaussianSize = 10
 * simGaussianSizeInit = 10
 * universePercentile = 0.999 (default)
 * verboseLevel = "info"
 * voteFilteringForWeaklySupportedSurfaces = 1
 * voteMarginFactor = 4

[15:08:31.084074][info] Found 1 image dimension(s):
[15:08:31.084074][info]         - [4032x3024]
[15:08:31.088075][info] Overall maximum dimension: [4032x3024]
[15:08:31.088075][warning] repartitionMode: 1
[15:08:31.088075][warning] partitioningMode: 1
[15:08:31.088075][info] Meshing mode: multi-resolution, partitioning: single block.
[15:08:31.088075][info] bounding Box : length: -10, width: -10, height: -10
[15:08:31.091075][info] Creating dense point cloud.
[15:08:31.091075][info] fuseFromDepthMaps, maxVertices: 5000000
[15:08:31.119082][info] simFactor: 15
[15:08:31.119082][info] nbPixels: 24385536
[15:08:31.119082][info] maxVertices: 5000000
[15:08:31.119082][info] step: 2
[15:08:31.119082][info] realMaxVertices: 6096384
[15:08:31.119082][info] minVis: 2
[15:08:31.119082][info] Load depth maps and add points.
[15:08:31.200100][info] Filter initial 3D points by pixel size to remove duplicates.
[15:08:31.200100][info] Build nanoflann KdTree index.
[15:08:31.801234][info] KdTree created for 6096384 points.
[15:08:31.805235][info] Filtering done.
[15:08:31.832241][info] 6096384 invalid points removed.
[15:08:31.843244][info] 3D points loaded and filtered to 0 points.
[15:08:31.843244][info] Init visibilities to compute angle scores
[15:08:31.843244][info] NANOFLANN: KdTree created.
[15:08:31.843244][info] Create visibilities (0/2)
[15:08:31.845247][info] Create visibilities (1/2)
[15:08:31.917260][info] Visibilities created.
[15:08:31.917260][info] Compute max angle per point
[15:08:31.917260][info] angleFactor: 15
[15:08:31.917260][info] 0 points filtered based on the number of observations (minVis).
[15:08:31.917260][info] 0 invalid points removed.
[15:08:31.917260][info] Filter by angle score and sim score
[15:08:31.917260][info] Build nanoflann KdTree index.
[15:08:31.917260][info] KdTree created for 0 points.
[15:08:31.917260][info] Filtering done.
[15:08:31.917260][info] 0 invalid points removed.
[15:08:31.917260][info] The number of points is below the max number of vertices.
[15:08:31.917260][info] 3D points loaded and filtered to 0 points (maxVertices is 5000000).
[15:08:31.917260][info] Create final visibilities
[15:08:31.917260][info] NANOFLANN: KdTree created.
[15:08:31.918260][info] Create visibilities (0/2)
[15:08:31.918260][info] Create visibilities (1/2)
[15:08:31.993277][info] Visibilities created.
[15:08:31.993277][fatal] Depth map fusion gives an empty result.

Desktop Enviroments:

• OS: win10 21H2
• Meshroom ver: 2021.1.0
• running on rtx2070s, R9 3900x

[asoftbird]

This seems to be related to the pixel size filter removing all points during the filtering step;

[15:08:31.200100][info] Filter initial 3D points by pixel size to remove duplicates. [15:08:31.200100][info] Build nanoflann KdTree index. [15:08:31.801234][info] KdTree created for 6096384 points.
[15:08:31.805235][info] Filtering done.
[15:08:31.832241][info] 6096384 invalid points removed.
[15:08:31.843244][info] 3D points loaded and filtered to 0 points.

I'm having this issue too, I've experimented with settings that I think are related to the pixel filter to various values (pixSizeMargin settings, max number of points), but I wasn't able to get the Meshing step to work. I've logged the process using trace logging mode, but aside from showing IO operations this did not show any extra information.

[ttsesm]

Hi @Ben-YuqiLu, @asoftbird,

Did you manage to find any workaround for this issue. I've stumbled on this as well.

Thanks.

[ttsesm]

Most likely is a memory issue, reducing the amount of images I did not get the error back again.

[stale[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.