SURFACING LIDAR

A general overview of the workflow employed.

Learn how to import the lidar scan data and a discussion of what it represents.

A discussion of how matrices encode transformation information and how we can extract this information to find the position of the lidar scanners.

In order to surface a lidar scan you need to create good normals. Set up a for loop to process each of the scans. Inside the for loop use the particle fluid surface as a means of generating normals that you will transfer back to the original lidar scan.

Compute the direction from the points in the scan to the position of the lidar scanner. You will use this to compute which surfaces are occluded from the scanner. Included in this video is a discussion of how you can create a direction vector from two points.

Use the vector computed in the previous video to calculate which surfaces are occluded from the scanner and remove them.

Now that the superfluous geometry has been removed you can transfer the normals back to the original lidar scan. Rather than use attribute transfer use the vex functions xyzdist() and primuv() to get a more accurate transfer.

Surfacing and cleaning up the entire lidar scan.

Surface the lidar scan in sections with PGD. This makes it possible to parallelize the surfacing processes using a render farm.

Taking the output from PDG reassemble it to end up with the finished surface