基于机器视觉的生咖啡豆缺陷检测模型研究

打开文本图片集

中图分类号：S126 文献标识号：A 文章编号：1001-4942（2025）07-0152-07

AbstractIn order to improve the defect detection accuracy green cfee bean during sorting process， a cfe bean defect detection model based on improved YOLOv5s was proposed. Firstly，the collected images were iterated enhanced to create a data setcontaining diffrent cfeebean defects.Secondly，asmalltarget detection layer was added to the original network structure YOLOv5s to improve the performance the algorithm indense scenes；the coordinate attention（CA） mechanism was introduced into the network， the feature expression ability the model was enhanced；the activation function was replaced by Mish function to promote better transmisson information to the neural network，thus improved the accuracy generalization performance the model. The experimental results showed that compared with the original model，the improved model increased the accuracy average accuracy by 6.8 1.1 percentage points，，respectively.The average confidence could also be improved under complex lighting conditions reached to more than 0.90. The improved model was suitable for detecting unsorted green cee beans could provide a theoretical basis for future mechanized sorting cee beans.

KeyWordsGreen cfee beans； Defect detection; Machine vision; Deep learning; CA mechanism; Mi-sh function

农作物视觉检测识别是当前研究的热门领域，涵盖计算机视觉、图像处理、机器学习和人工智能等多个学科。（剩余9053字）