Research On The Detection And Identify For Vision Information Of Arc Welding Robot Seams

The seam tracking control of arc-welding robot is the key during welding process,it is also the important component with automatically controlled of welding quality.Realize the control of welding seam tracking, welding seam detection and recognitiontechnology are one of the key among them. It is a trend technology that implementingseams detection and recognition using vision sensor.The system adopt vision sensing experimental based on the structure light, CCDreceive the stripe light that above the work piece of diffuse reflection on the differentwelding seams form, through the signal gather and image process, can know themiddle position of the welding seam. The working environment of the arc-weldingrobot is very abominable for vision Sensor, it is strong arc light, splashing and fog thatare interfered. This thesis has studied the result of treatment to welding seam imageby the image process method of the image segmentation, image enhancement, imagefilter and involves the treatment method of morphology.After the image processing, should carry on further characteristic analysis,namely draw the welding seam characteristic point, which provide information onneeding for the auto tracking of the welding seam, seeking position before welding ofarc-welding robot system and the parameter of control. This thesis has studied thealgorithms on the characteristic parameter of several kinds of commonly used weldingseam forms, such as butt joint, overlap joint, V groove, the automate recognition andcharacteristic parameter of the welding seam at the same time.The software system adopt Visual C++ 6.0 to develop a set of detection andrecognition of welding seams based on vision sensing of the structure light, thesystematic man-machine interface is friendly.