基于YOLOv8的河流漂浮垃圾检测算法研究  

Detection of Floating Garbage in Rivers Based on Improved YOLOv8

作　　者：席凯凯 狄巨星[1] 杨阳[1] XI Kaikai;DI Juxing;YANG Yang(Hebei University of Architecture and Engineering,Zhangjiakou 075000,China)

机构地区：[1]河北建筑工程学院,河北张家口075000

出　　处：《长江信息通信》2024年第11期28-31,共4页Changjiang Information & Communications

摘　　要：为了解决无人机航拍的河流漂浮垃圾检测的难题,文章提出了一种改进YOLOv8的目标检测算法,用于检测河流漂浮的小目标垃圾。在原有YOLOv8算法模型的基础上,添加了SENet注意力机制,提高了小目标检测的表征能力,并进行了超参数的微调。改进后的模型相比原有的YOLOv8s算法,平均精度(mAP)提升了4.4%,达到80%。因此,改进的YOLOv8s更适用于小目标检测。In order to solve the problem of river floating garbage detection based on UAV aerial photography,this paper proposes an improved target detection algorithm of YOLOv8for the detection of small target garbage floating in rivers.On the basis of the original YOLOv8algorithm model,the SENet attention mechanism is added to improve the characterization ability of small target detection,and the hyperparameters are fine-tuned.Compared with the original YOLOv8salgorithm,the average accuracy(mAP)of the improved model is increased by 4.4%to 80%.Therefore,the improved YOLOv8sis more suitable for small target detection.

关 键 词：YOLOv8 SENet 目标检测 超参数 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]