基于YOLOv8n绿橙检测算法研究

打开文本图片集

中图分类号：TP391.4；S666.4 文献标识码：A 文章编号：2095-5553（2025）07-0145-08

Abstract：Thecolorofthegreeorangefruitsis similar tothatof the branchesand leaves，making identification challnging.Tosolve theproblemsofhigh-precision detectionand modellightweight improvementof green orange， this studyselects the“green orange”from the orange basein Meishan City，Sichuan Province as the researchobject and proposesan improved target detection algorithm named ESN-YOLO based on YOLOv8n. First，GSConv and VoVGSCSP modules are introduced into the Neck section of the model to significantly reduce its parameters.Then，the SimAMattentionmechanism is integrated into theNeck network layer toenhance thedetection performanceof the model. Finally，thelossfunctionis modified from CIoU toEIoUtoimprove the model's precisionandacuracy intargetdetection. The results show that the modified ESN—YOLO model achieves an accuracy of 96.9% ： mAP@0.5 of 99.7% ，and mAP@0.5：0.95 of 83.6% . Compared with the original YOLOv8n model，these values represent improvements of （204号 3.2% ， 0.5% ，and 3.6% ，respectively. The model size of the improved algorithmis 5730KB ，and the model parameters are 2.6 MB，which are reduced by 11.4% and 13.9% ，respectively. ESN—YOLO achieved an effective balancebetweenmodel lightweightand target detectionacuracy.Theimproved model is deployedontheembedded device Jetson Nano，where the detection speed for a single orange photo is 732ms ，with an accuracy rate exceeding 90% ：

Keywords：green orange；target detection；model lightweight；deep learning

0 引言

用传统的采摘方式，该方式效率较低且准确率有限，难以实现实时监测的需求。（剩余13401字）