自然复杂环境下油茶果识别的重参数化算法

打开文本图片集

中图分类号：S24 文献标识码：A 文章编号：2095-5553（2026）02-0078-08

DOI：10.13733/j.jcam.issn.2095-5553.2026.02.012

Abstract：Toaddress thechalenges inthemachinepickingrecognitiontaskofCamelliaoleiferafruits innatural environments，suchas dense fruit adhesion，leafand branchocclusion，fruitcolor diference，anduneven lighting，and in light of the current research issues of insufcient detection acuracy and robustness incomplex scenarios，an improved YOLOv8n model，namely YOLOv8—COD，is proposed.In this model，hyper parameters in C2f module areadjusted andlightweightconvolutional moduleisintegrated.Heavyparameterizationmoduleisused toreplaceconvolutional module in backbone network，so as to maintain computational eficiency whileimproving model detectionaccuracy. Adding Global Attention Mechanism（GAM） into the feature fusion module andreplacing CIoU with GIoU—Focal can helpthemodel focus oncamelia fruitand improve therecognitionrateofthe modelundertheconditionsof fruit oclusion and adhesion． Compared with the traditional YOLOv8n，its precision rate，recall rate，and mAP are increased by 0.2% ， （204号 3.3% ，and 2.1% respectively. In complex natural environment，the missed detection probability of YOLOv8—COD decreased significantlycompared with YOLOv8n，and the detection accuracy was improved，which can efectivelyrealize the detection and identification of Camellia oleifera fruits.

Keywords： Camelia oleifera fruits；YOLOv8n; detection and recognition;YOLOv8—COD；RepVGG

0 引言

在我国南方广袤的土地上，尤其是湖南、广西、河南、江西等地区，油茶果作为当地的主要经济作物之一被广泛种植[1]。（剩余12817字）