数字农科院2.0

A new training strategy: Coordinating distillation techniques for training lightweight weed detection model

文献类型： 外文期刊

作者： Peng Zhou;Yangxin Zhu;Chengqian Jin;Yixiang Gu;Yinuo Kong;Yazhou Ou;Xiang Yin;Shanshan Hao

作者机构：

关键词： Channel pruning;Knowledge distillation;Site-specific weed management;YOLOv5

期刊名称： Crop Protection

ISSN： 0261-2194

年卷期： 2025 年 190 卷

页码：

收录情况： SCIE（2025版）

摘要： The excessive use of herbicides and the continuous growth of weeds pose a severe threat to agricultural production while also presenting potential environmental and public health risks. Site-Specific Weed Management effectively addresses this issue but relies heavily on real-time and accurate weed detection algorithms. This study innovatively proposed a new training strategy (TASA) to develop a lightweight weed detection algorithm based on YOLOv5. The heart of TASA's innovation was to alleviate information conflicts among multiple distillation techniques and between distillation techniques and student models and to stop distillation at the appropriate time. We used channel pruning technology to compress the model volume and introduced Knowledge Distillation (KD) during fine-tuning to recover the model performance maximally. Concurrently, TASA was used to coordinate multiple distillation techniques to assist in recovering the model training. The experimental results indicated that the volume of the Optimized YOLOv5s was reduced by 79.2% compared to the YOLOv5s, while the mean Average Precision (mAP) and F1-score (F1) reached 97.4% and 95.1%, respectively, with only decreased by 1.2% and 1.5%. Additionally, the detection speed on the CPU increased by 86.64%, reaching 38.423 frames per second (FPS). Meanwhile, we had also developed an online detection system based on PyQt5 and deployed it on Raspberry Pi. The system detected weeds in real-time, achieving a mAP of 96.3% and an FPS of 25.521 when the input image resolution was 416 × 416.

分类号：

  • 相关文献

[1]基于区域亮度自适应校正的茶叶嫩芽检测模型. 吕军,方梦瑞,姚青,武传宇,贺盈磊,边磊,钟小玉. 2021

[2]基于YOLOv5深度学习的茶叶嫩芽估产方法. 徐海东,马伟,谭彧,刘星星,郑永军,田志伟. 2022

[3]Semi-supervised learning and attention mechanism for weed detection in wheat. Teng Liu,Xiaojun Jin,Luyao Zhang,Jie Wang,Yong Chen,Chengsong Hu,Jialin Yu. 2023

[4]An Industrial-Grade Solution for Crop Disease Image Detection Tasks. Guowei Dai,Jingchao Fan. 2022

[5]Joint optimization of AI large and small models for surface temperature and emissivity retrieval using knowledge distillation. Wang Dai,Kebiao Mao,Zhonghua Guo,Zhihao Qin,Jiancheng Shi,Sayed M. Bateni,Liurui Xiao. 2025

[6]Compressing recognition network of cotton disease with spot-adaptive knowledge distillation. Xinwen Zhang,Quan Feng,Dongqin Zhu,Xue Liang,Jianhua Zhang. 2024

[7]Identification of Pine Wilt Disease Infected Wood Using UAV RGB Imagery and Improved YOLOv5 Models Integrated with Attention Mechanisms. Peng Zhang,Zhichao Wang,Yuan Rao,Jun Zheng,Ning Zhang,Degao Wang,Jianqiao Zhu,Yifan Fang,Xiang Gao. 2023

[8]Research on machine vision and deep learning based recognition of cotton seedling aphid infestation level. Xin Xu,Jing Shi,Yongqin Chen,Qiang He,Liangliang Liu,Tong Sun,Ruifeng Ding,Yanhui Lu,Chaoqun Xue,Hongbo Qiao. 2023

[9]A Detection Model for Cucumber Root-Knot Nematodes Based on Modified YOLOv5-CMS. Wang C.,Sun S.,Zhao C.,Mao Z.,Wu H.,Teng G.. 2022

[10]Detection and Counting of Maize Leaves Based on Two-Stage Deep Learning with UAV-Based RGB Image. Xu, Xingmei,Wang, Lu,Shu, Meiyan,Liang, Xuewen,Ghafoor, Abu Zar,Liu, Yunling,Ma, Yuntao,Zhu, Jinyu. 2022

[11]Deep-Learning-Based Rice Phenological Stage Recognition. Jiale Qin,Tianci Hu,Jianghao Yuan,Qingzhi Liu,Wensheng Wang,Jie Liu,Leifeng Guo,Guozhu Song. 2023

[12]Tea Bud Detection and 3D Pose Estimation in the Field with a Depth Camera Based on Improved YOLOv5 and the Optimal Pose-Vertices Search Method. Chen, Zhiwei,Chen, Jianneng,Li, Yang,Gui, Zhiyong,Yu, Taojie. 2023

[13]Lightweight SM-YOLOv5 Tomato Fruit Detection Algorithm for Plant Factory. Xinfa Wang,Zhenwei Wu,Meng Jia,Tao Xu,Canlin Pan,Xuebin Qi,Mingfu Zhao. 2023

[14]YOUNG APPLE FRUITS DETECTION METHOD BASED ON IMPROVED YOLOv5. Yonghui Du,Ang Gao,Yuepeng Song,Jing Guo,Wei Ma,Longlong Ren. 2024

[15]An Unstructured Orchard Grape Detection Method Utilizing YOLOv5s. Wang, Wenhao,Shi, Yun,Liu, Wanfu,Che, Zijin. 2024

[16]A Novel Deep Learning Method for Detecting Strawberry Fruit. Shuo Shen,Famin Duan,Zhiwei Tian,Chunxiao Han. 2024

[17]Small target tea bud detection based on improved YOLOv5 in complex background. Mengjie Wang,Yang Li,Hewei Meng,Zhiwei Chen,Zhiyong Gui,Yaping Li,Chunwang Dong. 2024

[18]Lightweight Structure and Attention Fusion for In-Field Crop Pest and Disease Detection. Zijing Luo,Yunsen Liang,Naimin Kong,Lirui Liang,Wenjun Peng,Yujie Yao,Chi Qin,Xiaohan Lu,Mingman Xu,Yining Zhang,Chenyang Lin,Chengyao Jiang,Mengyao Li,Yangxia Zheng,Yameng Jiang,Wei Lu. 2025

作者其他论文 更多>>