Build

Linux:

mkdir build
cd build
cmake ../
make

Windows:

mkdir build
cd build
cmake ../

Then open the generated sln file in Visual Studio.
Make sure "Release" build is selected.
Build "extract_profile".

Usage

Minimal

./extract_profile <input> -o <output> --coordinates "{x0, y1}, {x1, y}, ..." --width <scalar> 
./extract_area <input> -o <output> --area "{tx,rz,ry,0,-rz,ty,rx,0,-ry,-rx,tz,0,dx,dy,dz,1}"

Extract points in certain LOD level ranges

./extract_profile <input> -o <output> --coordinates "{x0, y1}, {x1, y}, ..." --width <scalar> --min-level <integer> --max-level <integer>

Extract points with different attributes

./extract_profile <input> -o <output> --coordinates "{x0, y1}, {x1, y}, ..." --width <scalar> --output-attributes <attributes> --output-format <format>

input: A point cloud generated with PotreeConverter 2.
output: Can be files ending with *.las, *.laz, *.csv, *.json, *.potree. If omitted, the result will be printed directly to the console.
format: LAS, LAZ, CSV, JSON, POTREE (default). If omitted, the files endings will be used.
attributes: Point attributes like: position, rgb, intensity, classification, ... Default: same as input point cloud
min-level, max-level: Level range including the min and max levels. Can be omitted to process all levels.

With --get-candidates, you'll get the number of candidate points, i.e., the number of points inside all nodes intersecting the profile. The actual number of points might be orders of magnitudes lower, especially if --width is small.

./extract_profile <input> --coordinates "{x0, y1}, {x1, y}, ..." --width <scalar> --min-level <integer> --max-level <integer> --get-candidates

A practical example:

./extract_profile ~/dev/tmp/retz -o ~/dev/tmp/retz.laz --coordinates "{-37.601, -100.733, 4.940},{-22.478, 75.982, 8.287},{66.444, 54.042, 5.388},{71.294, -67.140, -2.481},{165.519, -26.288, 0.253}" --width 2 --min-level 0 --max-level 3

programmerg/CPotree

Build

Usage