Weld pool image processing and feature extraction based on the vision of the CO2 welding

Abstract

In order to capture a clear image of the weld pool, the experimental system based on the reflection of the arc visual and CCD image acquisition was established in this chapter. As to the molten pool image characteristics of CO2 arc welding, this chapter introduces a method for welding pool image processing which may access the pool edge information more accurately. Firstly, it is based on the fuzzy entropy filter to remove noise in the multilevel image, modify the image gray value, and enhance the edge of the welding pool image; secondly, it uses the canny operator edges positioning and adopts the mathematical morphology corrosion and expansion comprehensive processing edge of image; thirdly, the edge detection is used for grayscale morphological gradient; and finally, the molten pool characteristic parameters are calculated. Thus it can lay foundation for the real-time control of the welding.