CloudCompare forum

Hi all,

Can anyone comment on the validity of the following process for scaling an SFM-generated point cloud to a LIDAR-generated point cloud. I am building SFM models from historical imagery of mountains in the alps, and comparing them with modern lidar scans to see volumetric changes in glaciers and rock walls.

1) Open georeferenced LIDAR in CC, allowing global shift
2) Open SFM-generated point cloud in same space
3) Measure a known section of LIDAR point cloud
4) Scale SFM model using "multiply/scale" to get within an order or magnitude
5) Manually move SFM model into the same close space as LIDAR
6) Commence Point pair picking alignment process, being sure to set LIDAR as "reference" and SFM as "aligned", clicking "ADJUST SCALE"
7) After this step I repeat same with Fine registration

Yes, that looks good!

Maybe you could skip step 6 and replace it by manual registration (if it's not too cumbersome). As it may be faster than the 'Align' tool, and sufficient since you use ICP afterwards.