Aligning Images in PC Painter

Are there any tips or suggestions for aligning imagery from the camera to the lidar model?
I am trying to get vertical surfaces as close as possible, but I find it challenging to align the images to the point cloud. It seems that what is good for one image does not apply to the next, or the “best” fit is still off the mark.

The red polygon is traced from the “intensity” view, then the RGBA is applied. This is as close as I could get the images and the lidar model to match.

Here, I traced the RGBA image and compared my alignment to the intensity view. Dang it… I’m still off!!

Here is an example of how I align my images. I can get this pretty close, almost exact… What I am curious about, did I change too many variables? Do you only mess with one set of variables? The adjustments seems random every time I do this (this is #7). I did level the R2A as best I could, but the vehicle does move off level when in motion (silly physics). However, it seems that when I adjust one image, the rest go out of alignment…

How do you do this to get maximum precision? What is a reasonable expectation here?

I just looked at your I-70 model… wow… so I know you know :slight_smile: :wink:

You may also want to consider looking at the camera distortion settings but it is an iterative process for these adjustments.

The way we do it, not that it is the perfect or best way is to adjust the roll, pitch yaw then x,y,z then roll pitch yaw again, then xyz again, then we adjusted the camera distortion amounts to fine tune the results. In aerial LidAR datasets road lines and other things that have high intensity returns help alot to visaully que the alignment. All of this for the camera bore sighting rely’s on the proper boresight alignment on the LiDAR prior to aligning the camera.

With your mobile LiDAR dataset you should have pretty good dataset to work with.