we have constructed a 90 camera based 3d scanner and we are experiencing issues with the photogrammetry. here is our setup: – 90 raspberry pi cameras taking uncompressed jpegs (3280 x 2464) approx 15mb for each picture. – custom build illuminated enclosure approx 10ft across – custom firmware based on raspistill we are trying to figure out the proper settings to run the cameras at as well as what settings to run the photogrammetry at using photoscan. we are getting approx 10,000 tie points for the sparse cloud. We have seen other scans where this is much higher. We also are having issues with entire colors being lost, such as one scan of someone we worked with who was wearing a dark blue shirt and while being on a white background there were no points at all where his shirt was.
Any help will be apprecited.
I didn’t find the right solution from the Internet.