基于图象分割的图象处理法在焊缝识别中的应用

提出了无须外加光源的基于图象处理的焊缝边缘识别方法和边缘特征提取过程，对采用小波滤波去噪后的焊缝图象采用符合实时性要求的焊缝图象处理方法，包括基于图象分割的焊缝图象分区法以及根据实际焊缝图象的实际性质加以分析得出的基于ＣＣＤ扫描性质的焊缝边缘识别法，这些方法既简单又适用，在实际焊缝跟踪中能得到良好的效果。通过多次实验得出：此种方法算法简单，响应速度快，适合于实际应用，能快速获得清晰的焊颖边缘，可以

Application of image processing based on the image segmentation in weld detection

An edge detection algorithm based on the principle theory of image processing is put into the use of seam tracking when there is no other optical source. Proceeding the noise elimination using the wavelet transform after we get the original image according to the real time demand, also we divide the weld image into three parts based on the image segmentation method, on the base of analysis the real weld image and the character of CCD we get the method of weld recognition. From the experiment we come to the conclusion that the weld line edge can be extracted efficiently and clearly, the ambient device can be simplified greatly.