Weld Image Identification And Compression Algorithm With Crawling Robot

This paper had studied recognition and compression algorithm of weld image based on the car for welding.First of all, the paper aimed to research welding image process of weld seam based on crawling robot by Binocular Vision. Several filters had been studied, such as mean filter, medium value filter and homomorphic filter. A cascade algorithm composed by medium value filter and homomorphic filter was used to welding image de-noising, which had been proved effective. Besides, an adaptive threshold arithmetic for segmented image was put forward, which had a good robustness, namely cascade OTSU. Several differential operators has been studied, such as LOG, Roberts, Canny, Sobel. Furthermore, the canny differential operators has been used to detect edge and thinning, and the experimental results could meet the requirements. Al-alloy butter plate welding experiments proved that welding torch could track seam tracking well.In the second place, a large number of experiments had been done to verify many classical algorithms such as Huffman Coding, Arithmetic Coding, Run-length Encoding, DPCM, DCT, KLT, DWHT, Wavelet Transform,etc. A specialized software platform had been developed based on VC2008. According to the topic request,a kind of weld compression algorithm had been designed based on Wavelet Transform. Generally, weld compression ratio can reach 64:1, and it could be closed to 92:1 if a weld image had been compressed after Feature Extraction. Experiments show that the algorithm not only for weld image, and for some molten pool image all have certain effect.Eventually, a new weld identification algorithm of robot vision on given region had been presented and it had been proved to be effective on MATLAB.