Scanning Design

I used different research designs during my three years in computational forestry. I tend to use higher scan resolution than other people working in this field. Although large scan files are more time demanding for the external preprocessing - i.e. registration and preliminary extraction of the target tree - but they offer you more information if you adjust your research goals at a later time.

My software can scale down point clouds with reduced information loss. If runtime performance is an issue for you, you can do this for better performance. But you still have better quality raw data available if needed later.

The scan resolution I have chosen most of the times resulted in 20.000 x 20.000 points per scan. The scanner used (Z+F Imager 5010) also provided an option to perform scans faster. I never did this, as this is supposed to result in more noise in your point clouds. A single scan takes for me 6.44 min of scanning time. As I use multiple scans per tree, an hour data collection per individual might be possible.

You can see the scan design used in the agro forestry system in the plot below. To my experience today, trees positioned like those labeled with D and B, tend to have poorer quality, as one of their sides might be not covered total by the scanner.

Scanning design, the TLS and tree positions are marked.
 
This template downloaded form free website templates