The point cloud created from generated (x,y,z) world coordinates, using the Kinect v1 RGB and depth video frames, gives an impression of a plain 2D image instead of a proper 3D model. It seems depth is not properly filled. This may partially be due to the fact that Kinect encodes depth distances too far or too close as 2047 and 0 respectively. Thus, I treat both values as zero and takes them into consideration while generating (x,y,z) and so it affects point cloud. Can anyone provide some insights why depth is not filled up and how a volume blending can be done to have proper 3D model? Any advices, pointers, or links would be highly appreciated. Thanks
This happens due to the low resolution (matrix depth and RGB) of the Kinect, besides which the computer can influence due to its hardware.
RealSense 3D cameras would be better for this type of use according to manufacturers, however, I have not tested it yet.
Thanks JLSL for a quick insight. Indeed, I am using a cropped frame (128X128) for generating (x,y,z) where I can't see much depth filled really. The Kinect v1 originally produces 640X480 RGB and depth images and I tried generating (x,y,z) from full images but still the results are not very promising. Can you advice on what techniques can be employed to remove such effects? I did not understand how a computer can influence with its HW? Thanks
- Badges
- Science topic
- Similar topics
- Biological Science
- Histology
- Fixatives