电子测量与仪器学报

伊欣同,单亚峰.基于改进Faster R CNN的光伏电池内部缺陷检测[J].电子测量与仪器学报,2021,35(1):40-47

基于改进Faster R CNN的光伏电池内部缺陷检测

Photovoltaic cell internal defect detection based on improved Faster R CNN

DOI：

英文关键词:deep learning residual channel attention module photovoltaic cells defect detection near infrared image

基金项目:国家自然科学基金（51974151）资助项目

作者	单位
伊欣同	辽宁工程技术大学电气与控制工程学院葫芦岛125000
单亚峰	辽宁工程技术大学电气与控制工程学院葫芦岛125000

Author	Institution
Yi Xintong	Faculty of Electrical and Control Engineering,Liaoning University of Technology, Huludao 125000,China
Shan Yafeng	Faculty of Electrical and Control Engineering,Liaoning University of Technology, Huludao 125000,China

摘要点击次数: 1315

全文下载次数: 4

中文摘要:

光伏电池近红外图像中复杂异构背景使内部缺陷检测成为一项极具挑战性的问题，为此，提出了一种基于深度学习的目标检测框架残差通道注意力Faster R CNN(residual channel attention faster R CNN, RCA Faster R CNN)，该网络通过卷积层池化层提取图像特征，再送入新颖的残差通道注意力RCA模块进行复杂背景特征抑制和缺陷特征突出，进而区域推荐网络推荐出更加精确的包含缺陷的候选框，最后利用分类与定位网络实现高精度的缺陷分类和位置估计。实验结果表明，RCA Faster R CNN的缺陷检测精度提升到了8329%，证明了所提方法的有效性。

英文摘要:

The complex heterogeneous background in the near infrared images of photovoltaic solar cells makes the detection of internal defects become a very challenging problem.Thus, an object detection framework based on deep learning residual channel attention Faster R CNN (RCA Faster R CNN) is proposed, which employs convolution layer and pooling layer to extract the image features, and sends them to the novel residual channel attention (RCA) module for complex background feature suppression and defect feature highlighting, then the region proposal network recommends a more accurate proposal containing defects, finally the classification and position network is applied to achieve high precision defect classification and position estimation.The experimental results show that the defect detection accuracy of RCA Faster R CNN has improved to 8329%, which proves the effectiveness of the proposed method.

查看全文查看/发表评论下载PDF阅读器