Revision history [back]

Hi, Program flow in ROS is that, it understands the images as a streams of frames (messages) from the camera. You have to process these frames according to the algorithm that you wanted. So these frames coming into the ROS will not be in Mat format for you to use the canny function of OpenCV directly. Hence as mentioned by Ahendrix you have to convert them to the Mat format using cv_bridge and process the frame. Once you have processed it you may want to convert it to ros msg again and publish it to the other nodes.

Hope it helps!! Karthik