Depth Map Information from Image Documentation Improvement

[saimouli]

Documentation needs to be improved describing the depth image. May be a code snippet on how to deduce the depth values will be helpful

System information

• OS Platform and Distribution : Linux Ubuntu 18.04):

Steps to reproduce the issue:

• Download the sample from https://apollo.auto/synthetic.html
• Corresponding depth values are very big to be in the camera coordinates

Supporting materials (screenshots, command lines, code/script snippets):

img_depth = cv2.imread("synthetic_laneDataset/Apollo_Sim_3D_Lane_Release/depth/00/0000000.png").astype(np.float32)/255
depth_label_map = (img_depth[:, :, 2] + img_depth[:, :, 1]/255)*655.36
cam_z = depth_label_map[ lane[j], lane[i] ]

Is camz the depth value in camera coordinates? If not how to retrieve this? Detailed documentation will be helpful