应用聚类统计分析的棉花异纤图形检测算法

在棉纺企业原棉异性纤维剔除工艺过程中,异性纤维种类及特征多样,难以构造统一的识别模型,为此,提出了一种基于聚类统计分析的棉花异性纤维图形检测算法。通过获取原棉纤维及异性纤维在RGB颜色模型空间的各分量值,进行数值聚类统计分析,采用RGB彩色图像阈值聚类统计分类的方法将获取的图片信息分为3类,进而判断有无异性纤维,再经过形态学等预处理修缮图像,对棉花中异性纤维的特征进行提取,较准确地得到异性纤维的面积、质心坐标和周长等参数,为异性纤维的清除提供条件。实验结果表明,该算法能较准确地识别异性纤维。

Algorithm for pattern detection of cotton foreign fibers based on cluster statistic analysis

Due to the foreign fibers types and characteristics of diversity, it was very difficult to construct a unified recognition model for eliminating raw cotton foreign fibers in cotton spinning enterprises processes. This paper proposes an image processing algorithm based on the cluster statistic analysis of cotton foreign fibers. The numerical statistical analysis was performed acquiring each component walues raw cotton fibers and the image information was divided into three categories by using RGB color image threshold statistical classification method. Then the foreign fibers was determined. Finally, by adopting image preprocessing for taking a better image and extracting the characteristics of foreign fibers in cotton, the area size, circumference and perimeter of foreign fibers were obtained. This should be the precondition for further eliminating raw cotton foreign fibers. The experimental results show that the algorithm can accurately identify foreign fibers.