Which scalar fields Contextcapture uses for point cloud processing? I know it uses RGB,Intensity, GPStime but does it benefit if there is any extra fields like scan angle or classification?
Seems there is improvement in Update 13 that scan import now supports scans with unknown unique source position (auto detection of source position)
From help:
The scanner position must be known to allow 3D reconstruction with ContextCapture. When it's not accessible in the input file, the scanner position can be specified manually or detected automatically by ContextCapture by checking the corresponding options. Be careful in this case that the input file corresponds to one scan acquired from one unique position. To be able to automatically detect the scanner position, the scan must contain only scanner position that is clearly defined by an ellipse of high-density points on the ground. If you are importing multiple files at once, the autodetect feature is still available and will detect one position per file.
Still it is not clear does it uses any other fields. Problem is that many scans are already registered and exported in las format with position lost and there is no option to set multiple scan positions. There would be need for more intelligent detection that it detects multiple scan position and points from this scan position from GPStime or angle.