摘要
为解决人为采摘、室外曝晒、分散储运等出现的问题,提出利用机器视觉来实现棉花异性纤维自动检测技术,包括图像预处理、图像增强、图像分割、图像数据的特征提取,目标模式识别方法。通过实验,SVM分类器训练出来的实验结果表明异性纤维识别的准确率有明显提高;最后再次实验,对检出的结果进行性能分析和评估,有效地提高了异性纤维检测方法的正确率,并对结果计算出相应的平均识别准确率,性能提升尤为明显。本文提出的棉花异性检测方法是在已有的检测方法基础上进行优化,有很好的应用价值。
In order to solve the problems of artificial picking,outdoor exposure,distributed storage and transportation,it is proposed to realize the automatic elimination technology of cotton heterosexual fiber by using machine vision principle,including image preprocessing,image enhancement,image segmentation,feature extraction of image data and target pattern recognition method.Through experiments,the experimental results trained by SVM classifier show that the accuracy of heterogeneous fiber recognition is significantly improved.Finally,the experiment is repeated,and the performance of the detected results is analyzed and evaluated,which effectively improves the accuracy of the method for detecting foreign fibers.And the corresponding average recognition accuracy is calculated for the results,and the performance improvement is particularly obvious.The cotton heterosexual detection method proposed in this paper proposes an optimization method based on the existing detection methods,which has good application value.
作者
张云
许江淳
王志伟
史鹏坤
Zhang Yun;Xu Jiangchun;Wang Zhiwei;Shi Pengkun(Faculty of Information Engineering and Automation,Kunming University of Science and Technology,Kunming,650500,China)
出处
《中国农机化学报》
北大核心
2018年第9期61-65,共5页
Journal of Chinese Agricultural Mechanization
关键词
机器视觉
图像增强
图像分割
特征提取
模式识别
SVM分类器
machine vision
image enhancement
image segmentation
feature extraction
pattern recognition
SVM classifier