基于小波变换的大型锻件图像边缘检测

大型锻件热态在线尺寸测量一直是在线生产的难题,近来,提出了用CCD测量尺寸的原理,其中CCD图像的边缘检测是其中的关键环节。文章根据Mallat小波理论中图像边缘对应于小波变换模的局部极大值点的理论,采用B样条小波对大型锻件图像进行了多分辨率小波变换,并根据不同尺度下噪声的小波系数变化较大的特点,采用固定阈值,有效抑制了噪声,成功地检测出图像的单像素边缘。试验结果表明,用小波变换处理大型锻件热态CCD图像的效果较好。

Large-size forging in edge detection based on wavelet transform

On-line measurement of hot forgings has always been a problem. Not long ago the principle of CCD was employed to on-line measurement, in which the key is the edge-detection of images. In this paper, the principle of edge detection using wavelet transformation is used, it can suppress noises effectively and detect single pixel contour of the edge successfully,according to graphical edge corresponding local maximum of wavelet transform modulus in Mallat's wavelet theory,multi-resolving wavelets transforming large-size forging drawing adopting B spline wavelets, based on wavelets coefficients of noise incrementing under different scale, and applying fixed thresholds. Experiments demonstrate that good results can be got by applying wavelet transformation into processing the CCD images in hot condition of large forgings.