Underwater welding seam image acquisition and recognition based on binocular vision

Abstract

According to the requirements of robots seam tracking underwater,a binocular stereo vision sensor is developed.An auxiliary light source of halogen Tungsten Lamp and suitable filter is used.The noises such as arc light,spatter and so on are removed and good seam images can be captured by this sensing system.Due to the characteristics of Fuzzy and low contrast about underwater welding seam images,fuzzy enhancement processing technology is adopted.Meanwhile,Pal processing algorithms is improved to enhance the efficiency of image processing and clear edge images are obtained.After edge detection,the welding seam center line is obtained accurately by scanning left and right welding seam feature points.The research laid the foundation for space welding seam tracking underwater.