农业装备与车辆工程2025，Vol.63Issue(3)：6-9,4.DOI:10.3969/j.issn.1673-3142.2025.03.002
基于改进YOLOv8的被遮挡柑橘果实检测算法研究
Research on the detection algorithm of occluded citrus fruit based on improved YOLOv8
摘要
Abstract
In order to obtain the accurate feature information of citrus fruits,Achieve accurate identification and localisation of occluded citrus,and achieve the purpose of accurate picking,an improved model for the identification of occluded citrus fruits was proposed.In this model,the BAM attention mechanism is introduced into the YOLOv8 object detection model to solve the problem of poor recognition effect of occluded citrus,and the SVM instance segmentation model was used to extract the accurate contour of occluded citrus fruit to achieve accurate recognition of occluded citrus fruits.The simulation results showed that compared with the YOLOv8 object detection model,the average centroid positioning error and fruit diameter error rate were reduced by 16.29 mm and 8.03%,respectively.关键词
柑橘/YOLOv8/BAM注意力机制/支持向量机Key words
citrus/YOLOv8/BAM attention mechanism/support vector machine分类
信息技术与安全科学引用本文复制引用
陈乾辉,吴德刚..基于改进YOLOv8的被遮挡柑橘果实检测算法研究[J].农业装备与车辆工程,2025,63(3):6-9,4.基金项目
2023年度河南省本科高校研究性教学改革研究与实践项目(164) （164）
商丘工学院2022年高等教育教学改革研究与实践项目"基于'一流课程'建设的'线上线下'相融合的研究性教学模式研究与实践"(2022JGXM01) （2022JGXM01）
