基于改进YOLOv8算法的枸杞果实目标检测  

Target Detection of Lycium Barbarum Fruit Based on Improved YOLOv8 Algorithm

作　　者：秦玮健 康峰[1] 王亚雄[1] 陈冲冲 仝思源 QIN Wei-jian;KANG Feng;WANG Ya-xiong;CHEN Chong-chong;TONG Si-yuan(School of Technology,Beijing Forestry University,Beijing 100083,China)

机构地区：[1]北京林业大学工学院,北京100083

出　　处：《机械工程与自动化》2024年第6期24-26,30,共4页Mechanical Engineering & Automation

基　　金：宁夏回族自治区“揭榜挂帅”项目(2022BBF01002-02)。

摘　　要：针对枸杞果实较小、遮挡严重等问题,提出一种改进YOLOv8算法对枸杞红果进行识别。引入Inner-IoU提升网络收敛速度;嵌入DAT模块,对提取到的冗余特征进行筛选;引入DBB模块来扩大感受野;增加一个更小尺寸的检测头提高检测精度。实验表明,改进后的YOLOv8算法的mAP达到90.2%。In order to solve the problems of small fruit size and serious occlusion of lycium barbarum fruit,an improved YOLOv8 algorithm was proposed to identify wolfberry red fruit.Inner-IoU is introduced to improve the network convergence speed,DAT module is embedded to filter the redundant features extracted,DBB module was introduced to expand the receptive field,and a smaller detection head is added to improve the detection accuracy.Experiment shaws that the mAP of the improved YOLOv8 algorithm reaches 90.2%.

关 键 词：YOLOv8算法 目标检测 枸杞 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]