摘要
为实现柑橘果实的精准快速识别,提出了一种改进YOLOv8网络模型。首先将YOLOv8网络模型中的部分传统卷积替换为ODConv全维动态卷积,以增强YOLOv8网络模型在复杂的自然环境下的适应能力,然后将YOLOv8的CIoU损失函数替换为MPDIoU损失函数,解决了CIoU损失函数在特殊情况下退化的问题,接着通过消融试验,分别验证了ODConv全维动态卷积与MPDIoU损失函数的有效性,改进后YOLOv8n、YOLOv8s、YOLOv8m、YOLOv8l、YOLOv8x的平均识别精度mAP分别从86.40%、88.92%、88.97%、88.99%、89.11%提高至88.25%、89.32%、89.57%、89.90%、90.12%。试验结果表明,ODConv全维动态卷积与MPDIoU损失函数能有效提高YOLOv8网络模型在自然环境下的柑橘果实识别能力。
In order to achieve precise and fast identification of citrus fruits,an improved YOLOv8 was proposed.Firstly,certain tradi⁃tional convolutions in the YOLOv8 were replaced with ODConv full-dimensional dynamic convolutions to enhance the model’s adapt⁃ability in complex natural environments.Subsequently,the CIoU loss function of YOLOv8 was substituted with the MPDIoU loss func⁃tion to address the degradation issue of the CIoU loss function in specific scenarios.Furthermore,the effectiveness of ODConv full-di⁃mensional dynamic convolutions and MPDIoU loss function was verified through a series of ablation experiments.The average recogni⁃tion accuracy(mAP)of the improved models,YOLOv8n,YOLOv8s,YOLOv8m,YOLOv8l,and YOLOv8x,was increased from 86.40%,88.92%,88.97%,88.99%,89.11%to 88.25%,89.32%,89.57%,89.90%,90.12%,respectively.Experimental results demonstrated that ODConv full-dimensional dynamic convolutions and MPDIoU loss function significantly enhanced the citrus fruit identification capability of the YOLOv8 in natural environments.
作者
余圣新
韦莹莹
方辉
李敏
柴秀娟
曾志康
覃泽林
YU Sheng-xin;WEI Ying-ying;FANG Hui;LI Min;CHAI Xiu-juan;ZENG Zhi-kang;QIN Ze-lin(Agricultural Science and Technology Information Research Institute of GAAS,Nanning 530000,China;Agricultural Information Institute of CAAS,Beijing 100000,China)
出处
《湖北农业科学》
2024年第8期23-27,共5页
Hubei Agricultural Sciences
基金
广西创新驱动发展专项资金项目(桂科AA22036002
桂科AA20108003)
广西壮族自治区农业科学院科技发展基金项目(桂农科2023JZ09)
广西壮族自治区农业科学院稳定资助科研团队基金项目(桂农科2021YT077)。
作者简介
余圣新(1994-),男,广西蒙山人,工程师,硕士,主要从事农业信息化、图像识别的研究工作,(电话)18378816668(电子信箱)49134017@qq.com。
