Image ROI based on depth information

asked 2017-01-20 10:44:30 -0600

ipa-hsd gravatar image

I am using a SICK TIM laser scanner to get a 2D scan of flat rectangular object. I know the dimensions of the object, the transformation between the laser scanner and the camera and the camera's calibration matrix. Based the laser scan data, I know the exact position and orientation of the object in the real world.

image description

This image shows the base of the object with respect to the laser scanner.

Based on this information, how can I determine the ROI of the object in the image? I can't do any kind of edge detection or feature detection before the ROI is found out. I am using OpenCV for this purpose.

The problem seems to be reverse of stereo vision where I have to get the depth information from a camera image.

edit retag flag offensive close merge delete