基于两次二维Daubechies小波变换的纸病在线辨识方法研究

针对造纸生产线上纸病在线检测系统光源不均匀性和纸病难以识别的问题,通过建立纸病图像背景模型实现动态补偿光源的不均匀成分,并根据不同纸病图像的灰度分布规律,两次使用二维小波变换增强其灰度轮廓信息和提取灰度特征,以实现纸病的在线辨识。实验结果表明,该方法可以显著提高纸病辨识精度。

Study on the On-line Detection of Paper Defects Using Twice 2D Daubechies Wavelet Transformation

Considering the problems of paper defects on-line detection system, including the light source was uneven and wrinkle defect was difficult to be identified, paper defect image background-model was built to dynamically compensate the uneven light, and based on the gray scale distribution of the images of various defects two-dimensional wavelet transformation was used twice to enhance the gray scale contour information of the image and its gray scale characteristics were extracted to realize the paper defects detection on-line. The experimental results showed that the method could improve paper defects identification accuracy.