This repository has been archived by the owner on Aug 22, 2024. It is now read-only.
Transforming depth to color results in loss of depth data on certain devices #840
Labels
External Issue
Known problem with usage from an external dependency
Describe the bug
Transforming a capture from the depth camera perspective to the color camera perspective introduces holes that weren't present in the source depth map when executing on certain devices.
To Reproduce
Run the TransformDepthToColor solution in the Azure-Kinect-Samples repo on a (or at least mine) Surface Laptop 2. The resulting image is full of holes that did not exist in the raw depth map.
Running the same program on any other device we've tested results in a transformed image with minimal holes (just like the source depth map).
Expected behavior
All devices with supported hardware should have the same results when transforming depth to color image.
Logs
Screenshots
Desktop (please complete the following information):
RGB camera firmware: 1.6.98
Depth camera firmware: 1.6.70
Depth config file: 6109.7
Audio firmware: 1.6.14
Build Config: Production
Certificate Type: Microsoft
Additional context
So far, we've only been able to produce this issue on a Surface Laptop 2 (Intel Core i5 / 8GB RAM). All other devices we've tested have worked as expected.
I've compared the original depth maps for the first and second examples, and they are virtually identical, which suggests the holes are being created as part of the transformation. Not sure why this would be hardware dependent though.
The text was updated successfully, but these errors were encountered: