基于小波奇异性的纸病检测

讨论了在纸幅随机纹理背景下纸病的检测,提出利用纸病处的奇异性来区分其和背景纹理.首先使用光滑函数与纸病信号进行卷积运算,然后选取能够保留纸病奇异性特征且同时削弱随机纹理所产生起伏的适当尺度下的信号,并对其实施进一步小波变换,去除大部分纹理起伏所对应的极大值线,最后利用极大值线与纵轴相交的截距来判断纸病.

Paper Defects Detection Based on Singularity Characterization

This paper describes the paper defects detection in stochastic textures. Paper defects can be distinguished from the background texture by singularity characterization. The original signals with paper defects are convoluted with the smooth function firstly, then some signals are selected which both preserve the singularity of paper defects and weaken small signal fluctuations . Then a wavelet transform is applied to the selecbecl signals.The most of wavelet fransform modulus maxima lines corresponding to the stochastic textures are removed . Finally the intercept of maxima lines is utilized to estimate paper defects.