摘要：
Because of unmaturity technology in the pipeline of the spot welding machine,problems such as solder skips and weld penetration happen,which will lead a bad effect on the product's life and beauty.It is necessary to have a quality detection on the solder joint.Considering the low efficient of traditional detection,a way of machine vision is introduced.An algorithm of image processing is also developed.Firstly,smoothing the image to get rid of the noise.Then using the method of Otsu to segment the threshold and get an inverted triangle distance transformation image which converts the pixel information into gray level information lastly,using watershed algorithm to accurately segment the solder joints.Finally,all the solder joints are classified by their area and other features.Experiment proofs that this algorithm is superior to the traditional methods,which can detect different kinds of lined solder joints efficiently,the industrial efficiency is also greatly improved.
展开
