I`m tring to get depth using depth camera of blender instead of cv2.StereoBM_create.compute and depth map looks so much better. but when get 3d points by it with using of a new function (depth_to_3d) the 3d points are not equal the coordinates of objects in my blender scene.these arrays of numbers needs to become equal so can go to the next step. I’m willing to detect some boxes in blender that simulated as an animation and each box’s location changes in x and y axis. finally identify each of them with particular ids. I used stereo vision of open cv for this purpose and for camera calibration implement blender camera calibration of GitHub link: