Dear GIS Professionals,
Have you ever tried converting a 3D Mesh model of a city or town to a Point Cloud (.las/.laz)?
I discovered that Skyline’s TerraExplorer can convert 3DML/SLPK files to .las/.laz formats.
I recently converted a 3D Mesh model to a 0.1m point cloud (.las) using TerraExplorer and encountered an error while using the Tree Point Classification model to extract tree points. The error message was as follows:
ERROR 050163: The input data does not have the following attributes required by the classification model: numberOfReturns. Failed to execute (ClassifyPointCloudUsingTrainedModel).
All 3D Mesh converted point clouds have the number of returns set to 1.
Any assistance would be greatly appreciated! Alternatively, if there could be a new Tree Point Classification model that doesn’t require the number of returns but utilizes RGB with X, Y, Z for 3D Mesh converted point clouds, that would be fantastic!
I highly recommend using high-accuracy LiDAR with this Tree Point Classification deep learning model to extract tree point clouds (class code 5) and to extract objects from point clouds (geoprocessing) to 2D Concave Hull Polygons with a clustering distance of 0.5m and an alpha radius of 0.5m.
You can download point clouds from LiDAR BC here to try it out.
Maybe 3D Mesh Point Cloud just won't work with this classification model, but it brings a great potential/idea to add one more benefits of having a 3D mesh model if it can help extracting tree coverages, buildings and other municipal assets.
Any help and suggestions are appreciated!