Crack detection and tracking realize two functions: crack detection and crack tracking, image enhancement of crack images, highlighting of its detailed features, extraction of crack edges, and comparison of the accuracy of edge extraction. Coordinate and skeletonize the crack feature, and convert it into a data array, which is convenient for storing in image files, which is conducive to the transmission of graphics.
The robot realizes crack tracking, and sends the crack trajectory instruction to the robot to realize the tracking of the robot. The detected crack feature data is converted into the coordinate information required by the robot and sent to the robot.
Image Processing
During the welding process, strong light will be emitted, and the collected image will cause the image to become white and bright, which will affect the subsequent image processing. Aiming at this phenomenon, it is first necessary to preprocess the acquired image to enhance the image features.
For white and bright images, the crack feature is enhanced by combining the prior theory of primary color and dehazing technology, and the edge of the enhanced crack feature is extracted to analyze and compare the sub-pixel edge detection.