Hello.<br>I just reviewed the stereo_image_proc tutorial and image_pipeline/Camera info webpage and there&#39;s something I&#39;m unclear with.<br><br>I understand I get 3D point cloud from stereo_image_proc with respect to left imager optical frame.<br>
Now, I&#39;m interested in getting 2D depth map with pixel values set equal to depth of the point cloud. (No. I do NOT want disparity map.)<br><br>Only way I can think of doing this is multiply K from left/camera_info sensor_msgs with the vector [X/Z, Y/Z, 1] from the point cloud and set the pixel value to Z. <br>
Is there another transform or parameter I need to introduce? <br>Or even better, is there a ROS implementation of this functionality already?<br><br>Thank you.<br><br>