一种基于Canny的原棉杂质图像分割方法研究

摘要

针对原棉杂质人工检验方式效率和效果不能保证的问题,提出一种基于Canny的原棉图像分割方法.该方法首先通过计算灰度的二阶微分增强图像,然后采用梯度按阈值取舍的方法检测图像边缘,在此基础上,结合"非极大值抑制"和形态学的连接操作,确保图像分割的效果.原棉杂质图像分割实验结果表明,该方法能够有效地将杂质与原棉背景分离,获取的杂质边缘清楚、流畅,有利于后期原棉杂质的分类与识别,同时,为棉纤维检验领域中机器视觉检测系统的研发提供了技术支撑和数据参考.%Considering the low efficiency and low effectiveness of manual inspection of raw cotton impurities,this paper proposes a Canny-based method of raw cotton image segmentation.According to the method,the second-order differential enhanced image of gray is to be calculated first,then the image border is to be tested with the method of gradient threshold choice method,and non-maximum suppression and join morphological operation are to be combined on this basis,to ensure the effect of image segmentation.Simulation experiment results of raw cotton impurities image segmentation show that the method works to remove impurities from raw cotton background,and the impurities edges obtained with the method are clear and fluent,which is good for classifying and recognizing raw cotton impurities subsequently,and provides technical support and data reference for the development and research of machine visual inspection system in the field of cotton fiber inspection.