[1]亢洁,代鑫,徐婷,等.基于一致性学习的半监督玉米田间杂草检测模型[J].江苏农业科学,2025,53(23):244-251.
　Kang Jie,et al.A semisupervised corn field weed detection model based on coherent learning[J].Jiangsu Agricultural Sciences,2025,53(23):244-251.
点击复制

基于一致性学习的半监督玉米田间杂草检测模型()

《江苏农业科学》[ISSN:1002-1302/CN:32-1214/S]

卷:
第53卷
期数:
2025年第23期
页码:
244-251
栏目:
农业工程与信息技术
出版日期:
2025-12-05

文章信息/Info

Title:
A semisupervised corn field weed detection model based on coherent learning
作者:
亢洁; 代鑫; 徐婷; 赫轩; 夏宇; 刘文波; 王佳乐
陕西科技大学电气与控制工程学院，陕西西安 710021
Author(s):
Kang Jie; et al
关键词:
玉米田; 杂草;  半监督目标检测;  教师-学生模型;  一致性学习; 平均精确率均值
Keywords:
-
分类号:
S126；TP391.41
DOI:
-
文献标志码:
A
摘要:
针对玉米田间杂草检测数据标注过程中耗时且复杂的问题，提出了一种基于一致性学习的半监督玉米田间杂草检测模型。该模型以改进后的YOLO v8n为基础，结合了教师-学生模型和一致性学习的思想。在该模型中，教师模型通过对无标签数据进行推理生成伪标签，学生模型通过一致性学习确保在不同数据增强下，学生模型的输出与教师模型的预测结果保持一致。同时，教师模型的权重通过学生模型的指数滑动平均进行更新，从而提高教师模型的稳定性和可靠性。结果表明，本研究模型在仅使用15%的有标签图像比例下，相较于Efficient Teacher、OneTeacher、Unbiased Teacher这3种半监督目标检测模型，平均精确率均值分别提高14.6、3.3、3.4百分点。同时，杂草类别的平均精确率分别提高18.4、5.0、3.4百分点。因此，本研究模型能够在有限的有标签图像比例下，充分利用无标签数据进行训练，保持较高的检测性能，并有效减少数据标注的人工成本，展现其在玉米田间杂草检测任务中的应用潜力。
Abstract:
-

参考文献/References:

［1］樊湘鹏，周建平，许燕，等. 基于改进卷积神经网络的复杂背景下玉米病害识别［J］. 农业机械学报，2021，52(3)：210-217.
［2］亢洁，刘佳，王佳乐，等. 基于多元感受野与EResPANet的草莓病害检测算法研究［J］. 陕西科技大学学报，2024，42(6)：190-198.
［3］Liu W，Anguelov D，Erhan D，et al. SSD：single shot MultiBox detector［M］//Computer Vision~ECCV 2016.Cham：Springer International Publishing，2016：21-37.
［4］Ren S Q，He K M，Girshick R，et al. Faster R-CNN：towards real-time object detection with region proposal networks［J］. IEEE Transactions on Pattern Analysis and Machine Intelligence，2017，39(6)：1137-1149.
［5］Carion N，Massa F，Synnaeve G，et al. End-to-end object detection with transformers［M］//Computer Vision~ECCV 2020.Cham：Springer International Publishing，2020：213-229.
［6］Huang X，Wang X，Lyu W，et al. PP-YOLO v2：a practical object detector［EB/OL］. ［2024-11-20］. https://arxiv.org/pdf/2104.10419.
［7］冀汶莉，刘洲，邢海花. 基于YOLO v5的农田杂草识别轻量化方法研究［J］. 农业机械学报，2024，55(1)：212-222，293.
［8］王秀红，王庆杰，李洪文，等. 基于YOLO v5-Jetson TX2的秸秆覆盖农田杂草检测方法［J］. 农业机械学报，2023，54(11)：39-48.
［9］Li C，Li L，Jiang H，et al. YOLO v6：a single-stage object detection framework for industrial applications［EB/OL］. ［2024-11-10］. https://arxiv.org/pdf/2209.02976.
［10］Tang J L，Wang D，Zhang Z G，et al. Weed identification based on K-means feature learning combined with convolutional neural network［J］. Computers and Electronics in Agriculture，2017，135：63-70.
［11］Wang Y C，Zhang S Q，Dai B S，et al. Fine-grained weed recognition using Swin Transformer and two-stage transfer learning［J］. Frontiers in Plant Science，2023，14：1134932.
［12］Milan M，Petar C，Mihajlo C，et al. Maize germination parameters and early seedlings growth under different levels of salt stress［J］. Ratarstvo i Povrtarstvo，2013，50(1)：49-53.
［13］尚文卿，齐红波. 基于改进Faster R-CNN与迁移学习的农田杂草识别算法［J］. 中国农机化学报，2022，43(10)：176-182.
［14］Lee D H. Pseudo-label：the simple and efficient semi-supervised learning method for deep neural networks［C］//Workshop on Challenges in Representation Learning，ICML.2013，3(2)：896.
［15］Xie Q，Dai Z，Hovy E，et al. Unsupervised data augmentation for consistency training［J］. Advances in Neural Information Processing Systems，2020，33：6256-6268.
［16］Liu Y C，Ma C Y，He Z，et al. Unbiased teacher for semi-supervised object detection［EB/OL］. ［2024-11-15］. https://doi.org/10.48550/arXiv.2102.09480.
［17］Wang Z Y，Li Y L，Guo Y，et al. Data-uncertainty guided multi-phase learning for semi-supervised object detection［C］//2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Nashville：IEEE，2021：4568-4577.
［18］Xu B，Chen M，Guan W，et al. Efficient teacher：semi-supervised object detection for YOLO v5［EB/OL］. ［2024-11-05］. https://doi.org/10.48550/arXiv.2302.07577.
［19］Tarvainen A，Valpola H. Mean teachers are better role models：weight-averaged consistency targets improve semi-supervised deep learning results［J］. Advances in Neural Information Processing Systems，2017，30.
［20］亢洁，代鑫，刘文波，等. 基于改进YOLO v8n的玉米田间杂草检测网络［J］. 江苏农业科学，2024，52(20)：165-172.
［21］Luo G，hou Y，Jin L，et al. Towards end-to-end semi-supervised learning for one-stage object detection［EB/OL］. ［2024-11-11］. 21 https://doi.org/10.48550/arXiv.2302.11299.

相似文献/References:

[1]杨君珑,李小伟.模拟氮沉降对干旱半干旱区几种杂草生长及光合特征的影响[J].江苏农业科学,2015,43(12):157.
　Yang Junlong,et al.Effects of simulated nitrogen deposition on growth and photosynthetic characteristic of weeds in arid and semiarid area[J].Jiangsu Agricultural Sciences,2015,43(23):157.
[2]高婷,王红春,石旭旭,等.水稻机械化插秧栽培及其草害防除[J].江苏农业科学,2013,41(09):60.
　Gao Ting,et al.Cultivation and weeds control of machine transplanting rice[J].Jiangsu Agricultural Sciences,2013,41(23):60.
[3]龚勋,王珩义.不同年生紫茎泽兰不同部位水浸液对甜荞种子萌发的影响[J].江苏农业科学,2016,44(03):161.
　Gong Xun,et al.Effects of aqueous extracts of different organs of Eupatorium adenophorum born in different years on seed germination ofFagopyrum esculentum Moench[J].Jiangsu Agricultural Sciences,2016,44(23):161.
[4]张俊,刘娟,臧秀旺,等.花生田常见杂草防治措施及展望[J].江苏农业科学,2016,44(01):141.
　Zhang Jun,et al.Control measures and prospect of common weeds in peanut field[J].Jiangsu Agricultural Sciences,2016,44(23):141.
[5]焦子伟,张相锋,尚天翠,等.国内外有机农业杂草控制技术研究进展[J].江苏农业科学,2016,44(01):1.
　Jiao Ziwei,et al.Research progress of weed control technology of organic farming at home and abroad[J].Jiangsu Agricultural Sciences,2016,44(23):1.
[6]武旭霞,张东彦,宋健,等.杂草识别专家及信息系统研究进展[J].江苏农业科学,2014,42(05):132.
　Wu Xuxia,et al.Research progress of weed identification expert and information system[J].Jiangsu Agricultural Sciences,2014,42(23):132.
[7]温广月,钱振官,李涛,等.稻田鸭舌草田间发生消长规律及生态学特性[J].江苏农业科学,2015,43(11):201.
　Wen Guangyue,et al.Preliminary study on occurrence and germination behavior of Monochoria vaginalis in paddy field[J].Jiangsu Agricultural Sciences,2015,43(23):201.
[8]刘兴林,桑松,孙涛,等.五氟磺草胺药肥混用对水稻移栽田杂草的防除效果[J].江苏农业科学,2015,43(08):119.
　Liu Xinglin,et al.Control effect of penoxsulam 0.025% GR mixed with fertilizer on annual weeds in transplanted rice fields[J].Jiangsu Agricultural Sciences,2015,43(23):119.
[9]王怀宇,李景丽.基于纹理特征的玉米苗期田间杂草识别[J].江苏农业科学,2014,42(07):143.
　Wang Huaiyu,et al.Weed identification in field during seedling stage of maize based on textural features[J].Jiangsu Agricultural Sciences,2014,42(23):143.
[10]李晓兰,相吉山,张艾明,等.地膜覆盖对玉米田土壤理化性质和线虫群落组成的影响[J].江苏农业科学,2018,46(15):257.
　Li Xiaolan,et al.Effects of plastic film mulching on soil physicochemical properties and nematode community composition in corn fields[J].Jiangsu Agricultural Sciences,2018,46(23):257.
[11]熊战之,袁树忠,钱兰娟,等.硝磺草酮、苯唑草酮对夏玉米田杂草的防除效果[J].江苏农业科学,2013,41(12):134.
　Xiong Zhanzhi,et al.Control effects of mesotrione and topramezone on weeds in summer maize fields[J].Jiangsu Agricultural Sciences,2013,41(23):134.
[12]张军高,漆永红,岳德成,等.玉米田封闭除草剂撒施效果比较[J].江苏农业科学,2015,43(09):154.
　Zhang Jungao,et al.Comparative study on effects of closed herbicides in corn field[J].Jiangsu Agricultural Sciences,2015,43(23):154.
[13]亢洁,代鑫,刘文波,等.基于改进YOLO v8的玉米田间杂草检测网络[J].江苏农业科学,2024,52(20):165.
　Kang Jie,et al.Weed detection network in maize field based on improved YOLO v8[J].Jiangsu Agricultural Sciences,2024,52(23):165.

备注/Memo

备注/Memo:
收稿日期：2024-12-23
基金项目：国家自然科学基金(编号：62376147)；陕西省自然科学基础研究计划(编号：2022JQ-181)；西安市科技计划(编号：23NYGG0070)。
作者简介：亢洁(1973—)，女，陕西渭南人，博士，副教授，硕士生导师，主要从事机器视觉、智慧农业方面的研究。E-mail：kangjie@sust.edu.cn。
更新日期/Last Update: 2025-12-05