Study On Application Of Image Processing In Examination Of Resistance Spot Weld Quality

Resistance spot welding is a technology applied extensively in mechanical manufacture industry, and non-destructive testing of resistance spot weld quality has been a hot research project in welding field. In this paper, spot welding tests were carried out and weld quality was tested on a versatile test machine by taking low carbon steel as the specimens. Meanwhile, the welding spot surface images were acquired, pre-processed and changed to gray images. The suitability of homomorphism filter on denoising for welding spot surface images of low-carbon steel plate was analyzed theoretically and the homomorphism filtering was performed to the obtained images. The traditional edge-detection algorithm has been improved in order to suit for the processing of low-carbon steel welding spot image, the comparison was carried out between traditional edge-detection algorithm and improved one, the result shows that the improved algorithm is better for the edge detection. Taking the characteristic zone area of welding spot as eigenvalues, its relationship with tensile strength of the welding spot was studied, and it proves that the welding spot surface image can be used to examine the weld quality.