Questions about DensifyPointCloud

[Yuanzaizhichi508]

I used DensifyPointCloud to reconstruct the data from colmap, but an error occurred. What is the reason? 17:33:29 [App ] Build date: Mar 14 2023, 09:28:26 17:33:29 [App ] CPU: 12th Gen Intel(R) Core(TM) i9-12900H (20 cores) 17:33:29 [App ] RAM: 15.36GB Physical Memory 2.00GB Virtual Memory 17:33:29 [App ] OS: Linux 5.4.0-146-generic (x86_64) 17:33:29 [App ] Disk: 17:33:29 [App ] SSE & AVX compatible CPU & OS detected 17:33:29 [App ] Command line: DensifyPointCloud scence.mvs --cuda-device -2 17:33:29 [App ] Camera model loaded: platform 0; camera 0; f 0.761x0.761; poses 20 17:33:29 [App ] Scene loaded from interface format (12ms): 20 images (20 calibrated) with a total of 17.58 MPixels (0.88 MPixels/image) 5501 points, 0 vertices, 0 faces 17:33:29 [App ] The camera directions mean is unbalanced; the scene will be considered unbounded (no ROI) 17:33:29 [App ] Point-cloud composed of 5501 points with:

• visibility info (23043 views - 4.19 views/point): 0 points with 1- views (0.00%) 91 points with 2 views (1.65%) 1661 points with 3 views (30.19%) 3749 points with 4+ views (68.15%) 2 min / 4.18887 mean (1.2683 std) / 14 max
• normals visibility info: 0 points with 3- degrees (0.00%) 1 points with 10 degrees (0.00%) 15 points with 25 degrees (0.07%) 121 points with 40 degrees (0.53%) 22355 points with 60 degrees (97.01%) 526 points with 90+ degrees (2.28%) 17:33:29 [App ] Preparing images for dense reconstruction completed: 20 images (90ms) 17:33:29 [App ] Selecting images for dense reconstruction completed: 20 images (6ms) Geometric-consistent estimated depth-maps 20 (100%, 5s352ms)
  free(): invalid pointer 已放弃 (核心已转储)

[navitohighbury]

HI! I also meet the same question. Do you know how to deal with it now? Thanks!