摘要
FastRCNN或FasterRCNN算法常用于对皮革材料表面缺陷进行检测,但其存在检测精度不高、准确率一般等问题。本文以深度学习改进传统算法得到改进Faster RCNN算法,并将改进后的算法用于皮革材料缺陷机器视觉检测工艺中,通过原图位置坐标改进、池化层算法设计及卷积神经网络特征提取等步骤实现了皮革材料缺陷机器视觉检测的优化。
Fast RCNN or Faster RCNN algorithm is often used to detect surface defects of leather materials,but there are some problems such as low detection accuracy and average accuracy.In this paper,deep learning was used to improve the traditional algorithm Faster RCNN algorithm,and the improved algorithm was applied to the machine vision detection process of leather material defects.The optimization of machine vision detection of leather material defects was realized through the improvement of original map position coordinates,pooling layer algorithm design and convolutional neural network feature extraction.
作者
朱春燕
ZHU Chunyan(College of Intelligent Science and Information Engineering,Xi'an Peihua University,Xi'an 710125,China)
出处
《中国皮革》
CAS
2023年第12期26-29,共4页
China Leather
基金
西安培华学院校级科研项目(phk2203)。
