Day 3/100: 25M points in one file. Where to start?
FOSS4G 2019: Connor Manning & Adam Steer demo'd PDAL/Entwine. Hooked instantly.
Sequence matters: classify, filter, mesh, interpolate.
Filter too early → lose detail. Too late → noise corrupts.
570M points. 19h processing. Open-source JSON pipelines. All documented. Visualization: seconds. Processing: 19 hours.
https://blog.maptheclouds.com/learning/lidar-pdal-experiments-yosemite-valley
During my studies, I connected with the company that captured a point cloud of the iconic Agfa-Gevaert factory nearby. It’s fascinating to see these two technologies converge in a single file.
more info: https://jeroenbocken.com/works/3d_laser_scan_pointcloud_agfa-gevaert.html
TIL that there are 360° LIDAR scanners below $450. With open source programs. I desperately need to figure out why I do need one.
I was running out of RAM and out of swap while trying to generate a copc out of a laz with #pdal (the laz weighs 4.7GB). So I created a 15GB swapfile by simply adding
swapDevices = [{
device = "/var/lib/swapfile_15G";
size = 15*1024;
}];
to configuration.nix, rebuild, and voilà. The process is currently running very smoothly, peak consuming the whole RAM as expected (64GB) + 12GB out of the available 24GB.
cc @sguimmara @autra
New open‑source breakthrough: SAM3D can isolate individual objects—like a tall lamp—in a point‑cloud scene, far beyond coarse class segmentation. See how promptable segmentation lets you pick out a human, a chair, or any item in a Meta‑style 3D environment. Dive into the details! #SAM3D #PromptableSegmentation #PointCloud #TallLamp
🔗 https://aidailypost.com/news/sam3d-isolates-specific-itemslike-tall-lampbeyond-broad-class
伊那の名物料理人気店「青い塔」のソースロースカツ丼がやってきたので飯テロ儀式。
#Scaniverse #foss4gj #foss4gshinshu #PointCloud #ソースカツ丼 #FuruhashiLab #古橋研究室 #AoyamaGSC #地球社会共生学部
SandFuchsFroschy 
| 
10W/542Wh
MapTheClouds
bokke
grin
Vincent Picavet
💧🌏 Greg Cocks
Jeremy ⁂
AI Daily Post
mapconcierge