当前位置: 首页 > 文章 > 核桃外部裂纹检测方法研究 长江大学学报(自科版)农学卷 2018 (14) 64-68+77+5
Position: Home > Articles > Method for External Cracks Detection of Walnuts Journal of Yangtze University(Natural Science Edition)Agricultural Science Volumn 2018 (14) 64-68+77+5

核桃外部裂纹检测方法研究

作  者:
张新伟;易克传;刘向东;赵学观;高连兴
单  位:
新疆大学科学技术学院;北京农业智能装备技术研究中心;安徽科技学院机械工程学院;沈阳农业大学工程学院
关键词:
核桃(Juglans regia);缺陷;检测;数学形态学
摘  要:
针对核桃(Juglans regia)在采收或加工过程中表面易出现裂纹或碎壳的问题,采用6尺度数学形态学方法对核桃的表面缺陷进行了检测。选择5个3×3模板构成不同方向的结构元素并依次进行6个尺度的膨胀操作,通过实体加权融合与信息熵结合的融合方法进行图像融合,得到最终的核桃表面裂纹图像,最后再对图像进行开启操作,除去图像中存在的细小的干扰物体。结果发现:多尺度数学形态学方法总体检测准确率为94%,远高于局部阈值法(65%)、分水岭(87%)和Roberts(86%)等传统方法,为核桃的在线品质检测提供了科学依据。
译  名:
Method for External Cracks Detection of Walnuts
作  者:
ZHANG Xin-wei;YI Ke-chuan;LIU Xiang-dong;ZHAO Xue-guan;GAO Lian-xing;School of Mechanical Engineering,Anhui Science and Technology University;School of Science and Technology,Xinjiang University;Beijing Research Center of Intelligent Equipment for Agriculture;College of Engineering,Shenyang Agriculture University;
单  位:
ZHANG Xin-wei%YI Ke-chuan%LIU Xiang-dong%ZHAO Xue-guan%GAO Lian-xing%School of Mechanical Engineering,Anhui Science and Technology University%School of Science and Technology,Xinjiang University%Beijing Research Center of Intelligent Equipment for Agriculture%College of Engineering,Shenyang Agriculture University
关键词:
walnut(Juglans regia);;flaw;;detection;;mathematical morphology
摘  要:
In allusion to the problems of walnut surface(Juglans regia)being prone to a cracking or broken shell in the process of harvesting or processing,the surface defects of walnuts were detected by 6-scale mathematical morphology,and five 3×3 templates were chosen to form the structural elements at different directions and a six scale expansion operation was carried out in turn.Image fusion was carried out by combining entity weighted fusion with information entropy,and the crack image of walnut surface was obtained.Finally,the image was opened to remove the small interference objects in the image.The test results showed that the overall detection accuracy of the multi-scale mathematical morphology was 94%,which was more higher than that of traditional defect detection methods,such as local threshold method of(65%),watershed of(87%)and Roberts of(86%).It provided a scientific basis for the online quality testing of walnuts.

相似文章

计量
文章访问数: 13
HTML全文浏览量: 0
PDF下载量: 0

所属期刊

推荐期刊