基于改进 YOLOv8 的有毒蘑菇识别方法
Journal: Agricultural Science DOI: 10.32629/as.v9i2.3753
Abstract
蘑菇种类繁多且是否有毒难以凭外观辨认,发展蘑菇识别技术尤为重要。然而,蘑菇常生长于密集遮挡的环境中,其形态多样、尺度变化显著,给识别带来了巨大挑战。为此,本文通过引入DAttention模块和SlimNeck模块对YOLOv8模型进行改进。结果显示,改进后的模型在精确率、召回率和平均精度均值上均优于基本模型,分别达到了92.5%、87.7%、92.1%。同时,与YOLOv5、YOLOv7、及YOLOv10模型相比,改进后的模型mAP@50提高2.1~14个百分点。本文为有毒蘑菇识别提供一种准确的解决方案。
Keywords
YOLOv8；有毒蘑菇识别；注意力机制；SlimNeck模块
Full Text
PDF - Viewed/Downloaded: 0 TimesReferences
[1] 董雪媛,王治宝.常见蘑菇的营养成分及有效成分的研究现状[J].神经药理学报,2020,10(03):37-41.
[2] 刘倩,王全庆.毒蕈中毒:躲避致命陷阱之科学应对指南[J].家庭医学(下半月),2025,(10):41-42.
[3] 樊帅昌.基于深度残差网络与迁移学习的毒蕈图像识别[D].浙江农林大学,2019.
[4] Kiss N,Czúni L.Mushroom image classification with CNNs: A case-study of different learning strategies[C]//2021 12th International Symposium on Image and Signal Processing and Analysis(ISPA).IEEE,2021:165-170.
[5] 赵成棋.基于深度学习的蘑菇识别模型研究及系统构建[D].北京林业大学,2022.
[6] 黄治国.基于深度学习模型的蘑菇图像检测[D].云南大学,2024.
[7] 程前.基于改进YOLOv8的密集场景下的行人检测方法研究[D].北京交通大学,2024.
[8] Xia Z,Pan X,Song S,et al.Vision transformer with deform able attention[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition.IEEE,2022:4794-4803.
[9] 杨茜,熊炜,孟圣哲,等.基于改进YOLOv8的绝缘子缺陷检测方法[J].电子测量技术,2025,48(07):86-97.
[10] 赵子瀚.基于YOLOv8的玉米虫害检测方法研究[D].佳木斯大学,2025.
[11] Li,Hulin,et al.Slim-neck by GSConv: A lightweightdesign for real-time detect or architectures[J].Journal of Real-Time Image Processing,2024,21(3):62.
[12] 赵斌.基于多源信息融合的羊只体尺测量研究[D].内蒙古师范大学,2024.
[13] 林海鹏.基于深度学习的光学镜片表面缺陷检测方法[D].广东工业大学,2025.
[14] FanY, Zhang L,Zheng C, et al. Real-time and accura-te meal detection for meal-assisting robots[J].Journal of Food Engineering,2024,371:111996.
[15] 蔡宇华.基于深度学习的西瓜叶片病害的识别研究[D].西安石油大学,2025.
[16] 夏清.基于YOLOv8的无人机目标检测跟踪研究与实现[D].扬州大学,2025.
[17] Misra D.Mish: A self regularized non-monotonic acti -vation function[J].ArXiv,2019,arXiv:1908.08681.
[18] Wang C Y, Bochkovskiy A, Liao H Y M. YOLOv7: Trainable bag-of-freebies sets new state-of the-art for r-eal time object detectors[C]//2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition(CVPR).2023:7464-7475.
[19] Liu H,Liu W, Wang W,et al.A method of grape clust-er target detection and picking point location based on impro ved YOLOv8 [C].International Conference on Gui-dance,Navigat ion and Control,2025.
[20] Bellout A,Zarboubi M,Dliou A,et al.YOLOv10—enabled IoT robot car for accurate disease detection in straw-berry cultivation[C].ITM Web of Conferences.EDP Sc-iences,2024,69: 01011.
[2] 刘倩,王全庆.毒蕈中毒:躲避致命陷阱之科学应对指南[J].家庭医学(下半月),2025,(10):41-42.
[3] 樊帅昌.基于深度残差网络与迁移学习的毒蕈图像识别[D].浙江农林大学,2019.
[4] Kiss N,Czúni L.Mushroom image classification with CNNs: A case-study of different learning strategies[C]//2021 12th International Symposium on Image and Signal Processing and Analysis(ISPA).IEEE,2021:165-170.
[5] 赵成棋.基于深度学习的蘑菇识别模型研究及系统构建[D].北京林业大学,2022.
[6] 黄治国.基于深度学习模型的蘑菇图像检测[D].云南大学,2024.
[7] 程前.基于改进YOLOv8的密集场景下的行人检测方法研究[D].北京交通大学,2024.
[8] Xia Z,Pan X,Song S,et al.Vision transformer with deform able attention[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition.IEEE,2022:4794-4803.
[9] 杨茜,熊炜,孟圣哲,等.基于改进YOLOv8的绝缘子缺陷检测方法[J].电子测量技术,2025,48(07):86-97.
[10] 赵子瀚.基于YOLOv8的玉米虫害检测方法研究[D].佳木斯大学,2025.
[11] Li,Hulin,et al.Slim-neck by GSConv: A lightweightdesign for real-time detect or architectures[J].Journal of Real-Time Image Processing,2024,21(3):62.
[12] 赵斌.基于多源信息融合的羊只体尺测量研究[D].内蒙古师范大学,2024.
[13] 林海鹏.基于深度学习的光学镜片表面缺陷检测方法[D].广东工业大学,2025.
[14] FanY, Zhang L,Zheng C, et al. Real-time and accura-te meal detection for meal-assisting robots[J].Journal of Food Engineering,2024,371:111996.
[15] 蔡宇华.基于深度学习的西瓜叶片病害的识别研究[D].西安石油大学,2025.
[16] 夏清.基于YOLOv8的无人机目标检测跟踪研究与实现[D].扬州大学,2025.
[17] Misra D.Mish: A self regularized non-monotonic acti -vation function[J].ArXiv,2019,arXiv:1908.08681.
[18] Wang C Y, Bochkovskiy A, Liao H Y M. YOLOv7: Trainable bag-of-freebies sets new state-of the-art for r-eal time object detectors[C]//2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition(CVPR).2023:7464-7475.
[19] Liu H,Liu W, Wang W,et al.A method of grape clust-er target detection and picking point location based on impro ved YOLOv8 [C].International Conference on Gui-dance,Navigat ion and Control,2025.
[20] Bellout A,Zarboubi M,Dliou A,et al.YOLOv10—enabled IoT robot car for accurate disease detection in straw-berry cultivation[C].ITM Web of Conferences.EDP Sc-iences,2024,69: 01011.
Copyright © 2026 邢文静
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License