基于小波分析的皮革疵点机器视觉检测方法

The Method of Leather Defects Visual Detection Based On Wavelet Analysis

提出了一种利用小波变换检测皮革疵点的方法。为了得到能够将疵点区别于正常皮革纹理的小波基,本文使用遗传算法对小波基优化,通过选择适当的适应度函数,得到最佳小波滤波器,使得疵点皮革图像经小波变换后突出疵点信息。实验表明,该方法能够正确检测出皮革图像中的疵点。

A method of leather defect detection based on wavelets transform has been introduced in this paper. For obtaining the wavelet transform coefficients distinguishing defect from natural texture,wavelet transform coefficients were optimized by genetic algorithm. Selecting appropriate fitness function,optimized wavelet filter was designed,then defects were detected by using wavelet transform. This method was proved to be efficient by the experiments.