The filter script utilizes the Point Cloud Library to remove noise in point clouds of cotton seedling trays. It first uses ConditionalRemoval to filter based on position and color, then RadiusOutlierRemoval, then StatisticalOutlierRemoval.
The following video shows the filtering at each step: Video
The clouds folder holds point clouds of the tray Rep1Tray15_2.
The example subfolder has the example clouds used in the video, with original.ply being zipped due to its large size.
The optimization subfolder has the resultant point clouds of the unselected parameter combinations from the filter parameter optimization process. For instance, ror0.01_300.ply is the result of setting the RadiusOutlierRemoval parameters to 0.01 for radius and 300 for neighbors, sor2_2.4 is StatsticalOutlierRemoval with 2 neighbors and 2.4 distance, and sor_ror.ply is having StatisticalOutlierRemoval come before RadiusOutlierRemoval instead of the other way around.
The program folder includes the files needed to run the filter.
- Ensure you have PCL and CMake installed
- In program/filter.cpp, replace the path with the PLY file that you are trying to filter
- In terminal,
cdintoprogram/build - run
cmake .. - run
make - run
./filter - After the program is finished running, the filtered PLY will be located at
program/build/filtered.ply