大幅皮革表面缺陷检测数学算法优化与应用研究  

Optimization and Application of Mathematical Algorithm for Surface Defect Detection of Large Scale Leather

作　　者：任艳 李向超 REN Yan;LI Xiangchao(Shaanxi Post and Telecommunication College,Xianyang 712000,China;Xi'an No26 Middle School,Xi'an 710001,China)

机构地区：[1]陕西邮电职业技术学院,陕西咸阳712000 [2]西安市第二十六中学,陕西西安710001

出　　处：《中国皮革》2024年第4期29-32,37,共5页China Leather

基　　金：陕西省“十三五”规划2018年度课题(SGH18B009);信息化平台下的高职院校高等数学课程改革与实践项目(2022SZX461)。

摘　　要：针对传统CCD皮革表面缺陷检测系统应用算法成像视场小、无法进行大幅皮革检测等问题,以CCD系统为基础融入改进Faster R-CNN算法对原有的数学算法进行优化,并以整张牛革为检测对象对优化以后的缺陷检测系统检测效果进行分析。结果表明,在原有检测系统基础上应用改进后的Faster R-CNN算法不仅实现了大幅皮革的缺陷检测,还能够在一定程度上提升系统检测精度和效率。In view of the problems such as the small imaging field of view of the traditional CCD leather surface de-fect detection system and the inability to carry out large scale leather detection,this paper optimized the original math-ematical algorithm based on CCD system by using the improved Faster R-CNN algorithm,and analyzed the detection effect of the optimized defect detection system by taking the whole cow leather as the detection object.The results show that the application of the improved Faster R-CNN algorithm on the basis of the original detection system can not only realize the defect detection of leather,but also improve the detection accuracy and efficiency of the system to a certain extent.

关 键 词：大幅皮革 表面缺陷检测 改进Faster R-CNN算法 系统应用 

分 类 号：TS57[轻工技术与工程—皮革化学与工程]