推荐星级:
  • 1
  • 2
  • 3
  • 4
  • 5

基于双阈值自适应分割的轴承滚子表面缺陷提取技术研究

更新时间:2019-09-19 07:08:49 大小:2M 上传用户:杨义查看TA发布的资源 标签:表面缺陷提取技术 下载积分:0分 评价赚积分 (如何评价?) 打赏 收藏 评论(0) 举报

资料介绍

文档为基于双阈值自适应分割的轴承滚子表面缺陷提取技术研究详解文档,是一份不错的参考资料,感兴趣的可以下载看看,,,,,,,,,,,,,

部分文件列表

文件名 大小
基于双阈值自适应分割的轴承滚子表面缺陷提取技术研究.pdf 2M

部分页面预览

(完整内容请下载后查看)
Computer Science and Application 
A Novel Bearing Roller Surface Defect  
Extraction Method Based on Double  
Threshold Adaptive Segmentation  
Liyan Yi1,3, Chang Liu2,3,4, Yucheng Zhang2,3, Xiang Yu2,3, Chao Liu1,3  
1School of Communications and Information Engineering, Chongqing University of Posts and  
Telecommunications, Chongqing  
2Beijing Key Laboratory of Mobile Computing and Pervasive Device, Beijing  
3Wireless Center of Institute of Computing Technology, Chinese Academy of Sciences, Beijing  
4University of Chinese Academy of Sciences, Beijing  
Received: Jan. 25th, 2019; accepted: Feb. 5th, 2019; published: Feb. 13th, 2019  
Abstract  
Aiming at the problems of incomplete defects and slow processing speed in the detection of indus-  
trial defects, this paper proposes a defect extraction algorithm based on double threshold adap-  
tive segmentation. According to the histogram of the actual workpiece, the standard histogram is  
obtained by fitting, and then the histogram of the actual workpiece and the standard histogram  
adaptive analysis are used to quickly locate the bilateral error threshold, thus avoiding high gray  
scale defects or low gray missing detection of degree defects. The experimental results show that  
the proposed method can effectively improve the integrity and accuracy of defect extraction and  
has a good processing speed.  
Keywords  
Defect Extraction, Threshold Segmentation, Histogram, Gaussian Fitting, Bearing Roller  
基于双阈值自适应分割的轴承滚子表面缺陷提  
取技术研究  
易礼燕1,3,刘 畅2,3,4,张玉成2,3,余 翔1,刘 超1,3  
1重庆邮电大学通信与信息工程学院,重庆  
2移动计算与新型终端北京市重点实验室,北京  
3中国科学院计算技术研究所无线通信技术研究中心,北京  
文章引用: 易礼燕, 刘畅, 张玉成, 应分割的轴承滚子表面缺陷提取技术研究[J]. 计算  
机科学与应用, 2019, 9(2): 314-322. DOI

全部评论(0)

暂无评论

上传资源 上传优质资源有赏金

  • 打赏
  • 30日榜单

推荐下载