New price target $2, page-52

You name it .. dream up an analysis and no doubt machine learning/AI will help.

If using photogrammetry and/or if capturing imagery, you can analyse against the original images themselves or against the point cloud or indeed both.

Pointerra already have powerline workflows for vegetation encroachment, but that one is pretty simple - there has been automated classification for vegetation for a good while now - same for power line extract ... clash the two and presto - you have encroaching vegetation.

Little more out of the box - mates of mine at SiteSee are working with photogrammetry-based workflows for cell towers - they use the ML/AI to do rust detection from the imagery for example, and use the point cloud to detect and record the position of the antennas and other infrastructure on the masts.

https://www.sitesee.io/

The big one though as far I can see for now is the automated conversion of scan-to-bim - building information models - so automatically detecting walls, doors, windows, columns, pipes, floors, lights, fire sprinklers, stairs .. then producing a 3D "CAD" model. Similar process for automated extract from mobile laser scanning in verticals like road and rail.

Another one - automated construction validation - upload/keep current your construction model and supply regular point clouds and "clash" the two to compare if stuff is built in the right place. So scenario: the HVAC guy does his ducts, but you find that is about to clash with the fire stuff which is going in next week - old mate HVAC come back immediately and put your stuff in where it is supposed to be. This is a perfect scenario for automate processing. This might mean a collaboration with say Autodesk - they host the models - models are available via API (Autodesk Forge), models get sucked up and compared - 3DP API sends back clash/QA issues again via Forge back in to Autodesk ecosystem for reporting/fix.

Again relatively straight forward - products exist at the desktop level - ClearEdge Verity for example - but to take that to a subscription/usage service is a no-brainer.

Market for construction validation is still very thin - cost of scanning is still probably too high, but the next-gen mobile scanners would likely make it viable - eg NavVis VLX (only $175k .. meh).