摘要
阐述基于YOLO的电路板瑕疵小目标检测方法。电路板图像通过深层网络,多尺度下修改锚定框,能更好地检测小目标。实验表明,基于YOLO的电路板瑕疵小目标检测方法的平均精确率达到99.45%。
This paper expounds the YOLO based method for detecting small targets of circuit board defects.The circuit board image is modified at multiple scales through deep networks to better detect small targets.The experiment shows that the average accuracy of YOLO based circuit board defect small target detection method reaches 99.45%.
作者
吴慧君
WU Huijun(School of Electronic Information Engineering,Zhangzhou Vocational and Technical College,Fujian 363000,China)
出处
《集成电路应用》
2024年第5期216-217,共2页
Application of IC
基金
福建省教育厅2022年福建省中青年教师教育科研项目(JAT220685)。
关键词
计算机技术
目标识别
YOLO
瑕疵检测
computer technology
object recognition
YOLO
defect detection