3D Forest v 0.5 Parameters
tool for manage and process TLS lidar data from forest environment
Brought to you by:
janekt,
m-a-r-t-in
Jan,
Do you have suggestions for or examples of the new parameters in 0.5? I ran multiple variations of input distance, multiplication values, number of elements, and distance from terrain on a 1m TLS cloud subsampled in CloudCompare to 0.1 and also 0.01 to make a smaller file size. It either crashed or made one file. Same cloud with default values in v 0.42 generates 650 individual point clouds, and generally did a good job identifying individual trees, shrubs, low veg, and build environment (buildings, poles fences, overhead utility lines).
Cordially,
Jeff
Hi Jeff,
input parameters depends maily on point density and on units. firts aasuption is that cloud are saved in meters. ideal input distance is such distance (voxel size) that collect at least 3 point in one voxel for computing any description. So if you have spacing about 1 cm among points input distance can be about 2 cm. but be carefull since point density can vary close to scanner and in the top of tree. So default value about 4 cm should be ok.
Multiplication value gives information about voxel neighbors,. If cloud is rather dense, you can use smaller value, but stil about 2 -4 should be ok.
distance from terrain and number of elements are more related to the schrub dentity and stem visibility. increase of distance can help find more elements that are connected to each other and number of element reffer to number of voxels connected with given value of description in given distance (inputsize * multiplicator).
So not sure what can cause the problem. maybe send me some sample cloud for estimating such values.
Jan