Aligning a till image to a point cloud

Community Showcase
, ,
1

I use Reality Capture for automobile crash reconstruction. Suppose I have one or more still images of a crash scene that shows tire marks and vehicle resting positions that were not measured by the police. By the time I get retained to reconstruct the case, the tire mark evidence is gone and so are the vehicles. Let’s say I go to the scene with my Trimble 3D scanner and obtain a point cloud of the crash scene. Is there a workflow with Reality Capture that I can import the point cloud and the still image(s), align the point cloud to a still image (get the point cloud in the same perspective as the still image) that contains the evidence that I want to locate and then transfer the evidence as points to the point cloud? I would then need to be able to obtain the Cartesian coordinates of the transferred evidence points.

Thank you all for any help!

Andy

2

Hi ASRich327,
it won’t be possible to align just one image into the model. For that you will need to have minimally two images, but still it could be hard to align, as the capture needs to follow basic photogrammetry rules how to capture the images.
For such merging you will need to have some control points, which will be visible on both datasets or you can use georeferenced data to merge them.
You can check also these tutorials: https://www.youtube.com/@CapturingReality/search?query=laser