I have acces to scans in the .las format and pictures taken at the same place. is it possible for RealityScan to align the images to the scan without ground control points or camera positions ?
Hello @tomthomas99 , it would be easy if you had a terrestrial laser scan in E57 or PTX format.
There is still a chance with .las, but it depends.
Since .las is an unstructured format, to import it into RealityScan, you already need to have images in RS in the same coordiante system that can see the point cloud, so you can convert the point cloud to a structured format that RS can process.
If the .las point cloud is an aerial lidar scan, in that case import it with the option to generate camera poses in a grid above the point cloud.
More about that here: Introducing RealityScan 2.0 | Course
Import the images and align them. There is a slight chance that the alignment will work automatically, but if not, you will need to generate your own GCPs on the imported .las point cloud and mark the GPS on the point cloud generated from the images. After that, re-run alignment to get a single component.
From that point, you can export a trajectory for the images or export XMPs that contain the images’ georeferencing.
Hello Jakub,
Thank you for your answer.
I can have access to .e57 files. My problem is that my images are very badly georeferenced, and I don’t even think they use the same coordinate system. However they do see the point cloud. That is why I would like to ignore the coordinates of the images and just align them on the scan using its georeferencing. Is it possible to do ?
The scan is not a LiDAR aerial scan, it is taken on ground by a Leica RTC360.
In this case, this is possible with ease. Import the structured E57 scans. Have one file per scan station. Import them to RS with the import LiDAR tool. Set the registration to exact and set the correct coordinate system. Then import the images, select all of them, and set their prior pose to unknown.
If your laser scans were scanned in color and the images were taken from similar positions, they should align automatically. If you are trying to add images from a drone, it is very likely that it will create a separate alignment component and not align with the laser scan component. In that case, you will need to place control points in overlapping areas and realign them to merge them. Once that is done, you can export the XMPs or a text file with the new georeferencing of the images.