Unity SDK 1.8
ML OS 1.3.0-dev2
I'm capturing images for CV and it's been quite difficult to get what I need consistently. I started using a video capture method that is ok, but not ideal. It sometimes results in crashes or "Long frame drop" problems that some of the team is aware of. It also shows a recording icon constantly when that's not accurate and is confusing to customers.
What I really need is to be able to efficiently, on-demand, non-blocking, just grab images. Both with holograms and without. The primary purpose is computer vision. Secondary is for customers to save the results.
When I tried to transition to an image-based version, I couldn't find any clear documentation or examples. I did my best to convert from my video-focused version, but it fails in both MLCameraPreCaptureAEWB and MLCameraCaptureImage. I go through all the camera setup and connections I'm aware of, but it still tells me the CV camera is not available. Permissions are in place.