ROS Resources: Documentation | Support | Discussion Forum | Index | Service Status | ros @ Robotics Stack Exchange |
1 | initial version |
With a stereo camera you can either recognize objects using the processed pointcloud or directly on the image. For that, I recommend you to take a look at PCL and OpenCV detection algorithms or example implementations. What kind of objects do you want to recognize?
2 | No.2 Revision |
With a stereo camera you can either recognize objects using the processed pointcloud or directly on the image. For that, I recommend you to take a look at PCL and OpenCV detection algorithms or example implementations. What kind of objects do you want to recognize?
With cans I would start detecting ellipses in the image, or if the cans have a specific colour I would use the color and the depth information to group different cans together.
This question is too localized.