In one example, a method may include receiving one or more laser scan of a scene, receiving two or more camera images of the scene, determining one or more decision regions where the camera images are different from one another, detecting edges of the decision regions where the camera images are different from one another, comparing the decision regions where the camera images are different from one another inside of the detected edges with a corresponding region in the laser scan to determine which of the camera images includes a desired region that more closely corresponds to the laser scan, and generating a corrected image including the desired region that more closely corresponds to the laser scan.