I am currently conducting KINECT research at the university using Hero 3D. (kinect v2)
I want to match RGB & Depth. However, due to different resolutions, it is difficult to match them (1920x1080, 512x424)
So we divided the points, colors, into calibrated clouds.
I checked the color values.
However, the X and Y values in the array
The values for X, Y, and Z of the points do not match.
Can you suggest an easier use?
The cloud of points structure consists in two 1D arrays of points and RGB colors. The color array has a length corresponding to the product of the image dimensions you selected. The point arrays has 3X that size (because of the X, Y, Z components for each point). You can reshape the arrays into 2D but you have to take the 3X size into account. Below is an example of how you can do it using the 480x270 resolution:
The result is show below in the RGB image and the corresponding Z-value image.
Notice that the depth sensor has a resolution of 512x424, meaning that using the 1920x1080 image resolution gives a better RGB resolution but position information remains 512x424 (about 10 pixels of the RGB images will have the same X-Y-Z value, in average).