摘要
针对光伏电池片易受材料、生产工艺等因素影响,产生诸多缺陷且检测难度高等问题,设计了一套以嵌入式AI设备Jetson Xavier NX为控制单元的智能检测分拣系统。该系统利用电致发光成像技术采集图像,选用以ResNet50为主干的Faster RCNN网络结构作为目标检测算法,并融入特征金字塔网络(FPN)提高网络对多尺度缺陷的特征表达能力。并采用六轴机械臂将各类残片分拣至指定区域。经实验测试,检测的平均精度为92.4%,能够满足实际生产需求。
An intelligent detection and sorting system based on embedded AI device Jetson Xavier NX was designed to solve the problem that photovoltaic cells are susceptible to many defects and difficult to detect due to factors such as materials and production process.In this system,electroluminescence imaging technology is used for image acquisition,ResNet50 based Faster RCNN network structure is selected as the target detection algorithm,and the Feature Pyramid Network(FPN) is used to improve the feature expression ability of multi-scale defects.Use six-axis manipulator to sort all kinds of debris to the designated area.The test results show that the average accuracy is 92.4%,which can meet the actual production needs.
作者
孙晨
邓宽
SUN Chen;DENG Kuan(School of Mechanical,Yancheng Institute of Technology,Yancheng 224051,China;School of Electronic Information Engineering,Jinling Institute of Technology,Nanjing 211169,China)
出处
《电子设计工程》
2024年第4期129-134,共6页
Electronic Design Engineering
