基于小波变换的棉花异纤检测算法研究

研究棉花中异性纤维的检测问题,提高检测准确度。针对棉花异纤检测,当棉花中混有头发丝等与棉花纹理相似的线状异纤时,传统的单一尺度的边缘检测算法不能将线状异纤准确检测出来,造成检测准确度不高的问题。为解决上述问题,提出一种小波变换的棉花异纤检测方法,通过将棉花图像分割为多个连通区域,并对各区域进行多尺度小波变换,然后计算并设定每个区域的检测阈值,最后利用检测阈值完成异纤检测,可避免传统单一尺度的边缘检测方法不能准确识别线状异纤的问题。实验证明,改进方法能够准确地将线状异纤准确检测出来,具有很高的检测准确度,取得了满意的结果。

Research on Foreign Fiber Detection Algorithm Based on Wavelet Transform

Research foreign fiber detection problem to improve the detecting accuracy.When foreign fibers are mixed with cotton,the traditional edge detection algorithm can not accurately detect the threadiness foreign fibers,causing low detection accuracy.This paper presentrf a cotton foreign fiber detect method based on wavelet transform,The cotton image was segmented into several connected areas,and multi-scale wavelet transform was executed with these connected areas.Then the detection thresholds were calculated for each areas.Finally,the detection threshold was use to complete the detecting,which prevents the problem that traditional single-scale edge detection method can not accurately identify the threadiness foreign fibers.Experiments show that the method can accurately detect threadiness foreign,fibers,has higher detection accuracy and satisfactory results.