摘要
通过帘子布图像频谱及其特征值,判断帘子布是否含有疵点.在TMS320C54x Simulator环境中,采用二维FFT算法,针对帘子布正常纹理和疵点纹理对图像频谱带来的不同影响,找到一组特征值平均灰度P1、经向频率峰值P2、频率成分总和P8.当P1、P2、P8的值依次超出了1556-1700、550-640、2900-3040范围,帘子布存在疵点.实验结果表明采用该方法可以实现帘子布疵点的有效检测.
Whether or not that the cord fabric includes defects can be distinguished through spectrum and features of cord fabric image.Under the TMS320C54x Simulate environment,a group of features are discovered through the analysis of the difference between the faultless cord fabric and the defects in spatial frequency domain with two-dimensional FFT,the average light intensity P1,horizontal frequency peak value P2,all frequency summation P8.When the range of P1,P2,P8 is beyond 1556-1700、550-640、2900-3040 each other...
出处
《电子器件》
CAS
2007年第5期1782-1785,共4页
Chinese Journal of Electron Devices
