基于改进YOLOv8-Track 的芝麻单株蒴果检测计数研究

doi:10.15933/j.cnki.1004-3268.2025.04.016

河南农业科学 ›› 2025, Vol. 54 ›› Issue (4): 155-166.DOI: 10.15933/j.cnki.1004-3268.2025.04.016

• 农业信息与工程·农产品加工 • 上一篇下一篇

基于改进YOLOv8-Track 的芝麻单株蒴果检测计数研究

李琛昊^1,2，王川^1,3，李国强²，赵巧丽²，杨萍^1,2，王凯⁴，常升龙³，郑国清²

（1.河南师范大学计算机与信息工程学院，河南新乡 453007；2.河南省农业科学院农业信息技术研究所/农业农村部黄淮海智慧农业技术重点试验室，河南郑州 450002；3.河南师范大学软件学院，河南新乡 453007；4.河南省土壤肥料站，河南郑州 450002）

收稿日期:2024-09-06 出版日期:2025-04-15 发布日期:2025-05-20
通讯作者: 王川（1976-），男，河南新乡人，副教授，硕士，主要从事人工智能理论及应用研究。E-mail：wangch@htu.edu.cn 王凯（1987-），男，河南商丘人，高级农艺师，主要从事智慧农业技术研究。E-mail：kaiwang19777@qq.com 常升龙（1990-），男，河南濮阳人，副教授，博士，主要从事遥感图像处理理论及其应用研究。E-mail：2024123@htu.edu.cn
作者简介:李琛昊（1999-），男，河南鹤壁人，在读硕士研究生，研究方向：计算机视觉及农业信息化。E-mail：208283128@stu.htu.edu.cn
基金资助:
河南省科技攻关计划项目（252102111168，242102110351，232102110295）；国家自然科学基金项目（62072159）；河南省农业科学院优秀创新团队项目（2024TD07）

Detection and Counting of Sesame Capsules per Plant Based on Improved YOLOv8‐Track Model

LI Chenhao^1,2，WANG Chuan^1,3，LI Guoqiang²，ZHAO Qiaoli²，YANG Ping²，WANG Kai⁴，CHANG Shenglong³，ZHENG Guoqing²

（1.College of Computer and Information Engineering，Henan Normal University，Xinxiang 453007，China；2.Institute of Agricultural Information Technology，Henan Academy of Agricultural Sciences/Key Laboratory of Huang‐Huai‐Hai Smart Agricultural Technology，Ministry of Agriculture and Rural Affairs，Zhengzhou 450002，China；3.College of Software，Henan Normal University，Xinxiang 453007，China；4.Henan Soil and Fertilizer Station，Zhengzhou 450002，China）

Received:2024-09-06 Published:2025-04-15 Online:2025-05-20

摘要/Abstract

摘要： 单株蒴果数是芝麻产量构成的重要因素。为实现单株芝麻蒴果的准确检测计数，使用目标检测、多目标追踪等技术，动态追踪单株蒴果，有助于提高芝麻育种和栽培管理效率。针对芝麻蒴果小目标、生长密集、遮挡重叠等现象，以YOLOv8-Track为基准模型，在特征融合网络中引入小目标检测头和Shuffle attention注意力机制，在模型后处理阶段引入MPDIOU损失函数，构建了SD-YOLOv8-Track模型。然后利用模型ByteTrack多目标追踪算法的ID计数方法，以芝麻单株旋转视频作为模型输入，追踪统计芝麻蒴果数。结果表明，以单幅图片为输入，SD-YOLOv8-Track模型检测蒴果的准确率、召回率、平均精度分别为92.25%、92.4%、92.58%，比原模型YOLOv8-Track分别提高5.94、6.6、6.31百分点。以单株旋转视频为输入，SD-YOLOv8-Track模型的多目标追踪准确率、多目标追踪精确率分别为89.42%、88.23%，比原模型分别提高4.23、4.60百分点。SD-YOLOv8-Track模型检测蒴果的平均计数准确率、漏检率、误检率分别为93.27%、3.85%、2.88%，平均计数准确率比原模型提高5.61百分点，漏检率和误检率比原模型分别降低3.84、1.77百分点。改进后的SD-YOLOv8-Track模型具有较好的芝麻单株蒴果检测性能，适用于芝麻单株蒴果的动态完整计数

关键词: 芝麻蒴果, 检测计数, 多目标追踪, 动态计数, Shuffle attention, MPDIOU, YOLOv8-Track

Abstract: Sesame capsules are an essential factor in the composition of sesame yield. In order to realize the accurate detection and counting of sesame capsules per plant，using object detection，multiple targets tracking and other technologies for dynamic tracking of capsules per plant is helpful to improve the efficiency of sesame breeding and cultivation management. Aiming at the phenomena of sesame capsules，such as small target，dense growth and overlapping occlusion，this study taked YOLOv8‐Track as the benchmark model，introduced small target detection head and Shuffle attention mechanism into the feature fusion network，and introduced MPDIOU loss function in the post‐processing stage of the model to construct SD‐YOLOv8‐Track model.In addition，this study utilized the ID counting method of model ByteTrack multi‐target tracking algorithm to track and count sesame capsules using a single rotating video of sesame as the model input. The results showed that when taking a single picture as input，the accuracy，recall，and mean average precision of the SD‐YOLOv8‐Track model for detecting capsules were 92.25%，92.4%，and 92.58%，respectively，indicating 5.94，6.6，and 6.31 percentage points higher than those of the original model YOLOv8‐Track.For the rotating video as input，the multiple object tracking accuracy and multiple object tracking precision of SD‐YOLOv8‐Track model were 89. 42% and 88.23%，respectively，which were 4.23 and 4.60 percentage points higher than the original model.The accuracy，missed detection rate，and error detection rate of the SD‐YOLOv8‐Track model were 93.27%，3.85%，and 2.88%，respectively.The accuracy rate was 5.61 percentage points higher than that of the original model，and the missed detection rate and false detection rate were 3.84 and 1.77 percentage points lower than that of the original model.The improved SD‐YOLOv8‐Track model performs better in detecting sesame capsules and is suitable for dynamic complete counting of sesame capsules in a plant.

Key words: Sesame capsules, Detection and counting, Multiple targets tracking, Dynamic counting, Shuffle attention, MPDIOU, YOLOv8‐Track

中图分类号:

李琛昊, 王川, 李国强, 赵巧丽, 杨萍, 王凯, 常升龙, 郑国清. 基于改进YOLOv8-Track 的芝麻单株蒴果检测计数研究[J]. 河南农业科学, 2025, 54(4): 155-166.

LI Chenhao, WANG Chuan, LI Guoqiang, ZHAO Qiaoli, YANG Ping, WANG Kai, CHANG Shenglong, ZHENG Guoqing. Detection and Counting of Sesame Capsules per Plant Based on Improved YOLOv8‐Track Model[J]. Journal of Henan Agricultural Sciences, 2025, 54(4): 155-166.

图/表 15

参考文献

［1］王瑞元.中国为全球芝麻产业的发展作出了重要贡献［J］.中国油脂，2019，44（12）：1‐2.
WANG R Y.China has made important contributions to the development of global sesame industry［J］.China Oils 164and Fats，2019，44（12）：1‐2.

［2］胡峰，张建涛，张杰，等.河南省芝麻种植气候适宜度时空变化研究［J］.河南农业科学，2023，52（8）：56‐68.
HU F，ZHANG J T，ZHANG J，et al.Study on spatio‐temporal variation of climatic suitability of sesame planting in Henan Province［J］.Journal of Henan Agricultural Sciences，2023，52（8）：56‐68.
［3］孙华，余意雯，黄萌，等.我国油料作物生产概况和空间集聚特征分析［J］.江苏农业科学，2022，50（23）：67‐74.
SUN H，YU Y W，HUANG M，et al.Study on production situation and spatial agglomeration characteristics of oil crops in China［J］.Jiangsu Agricultural Sciences，2022，50（23）：67‐74.
［4］刘海礁，邱雅洁，周萌，等.基于文献计量分析芝麻的研究热点与趋势［J］.中国油脂，2024，49（8）：103‐110.
LIU H J，QIU Y J，ZHOU M，et al. Status and trends of sesame（Sesamum indicum L.）research based on bibliometric analysis［J］.China Oils and Fats，2024，49（8）：103‐110.
［5］JHANSI A N，SRIJAYA T，MADHAVI A，et al.Influence of phosphorus and sulphur application on sesame yield in high P soils of telangana，India［J］.International Journal of Environment and Climate Change，2023，13（10）：2071‐2079.
［6］ZHANG N，PAN Z Y，LI T H，et al.Improving graph representation for point cloud segmentation via attentive
filtering［C］//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Vancouver，BC，Canada：IEEE，2023：1244‐1254.
［7］XUE M Q，ZHANG H F，SONG J，et al. Meta‐attention for ViT‐backed continual learning［C］//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. New Orleans，LA，USA：IEEE，2022：150‐159.
［8］RIZWAN IQBAL H M，HAKIM A. Classification and grading of harvested mangoes using convolutional neural network［J］.International Journal of Fruit Science，2022，22（1）：95‐109.
［9］闫彬，樊攀，王美茸，等.基于改进YOLOv5m的采摘机器人苹果采摘方式实时识别［J］. 农业机械学报，2022，53（9）：28‐38.
YAN B，FAN P，WANG M R，et al. Real‐time apple picking pattern recognition for picking robot based on improved YOLOv5m［J］.Transactions of the Chinese Society for Agricultural Machinery，2022，53（9）：28‐38.
［10］沈志豪，刘金江，张建洋. 基于改进YOLOX-s的田间麦穗检测及计数［J］.江苏农业科学，2023，51（12）：164‐171.
SHEN Z H，LIU J J，ZHANG J Y. Detection and counting of wheat ears in field based on improved YOLOX‐s［J］.Jiangsu Agricultural Sciences，2023，51（12）：164‐171.
［11］郭明月，刘雅晨，李伟夫，等.基于视频跟踪算法的果园猕猴桃产量实时预估［J］.农业机械学报，2023，54 （6）：178‐185.
GUO M Y，LIU Y C，LI W F，et al. Real‐time production prediction of kiwifruit in orchard based on video tracking algorithm［J］.Transactions of the Chinese Society for Agricultural Machinery，2023，54（6）：178‐185.
［12］郭瑞，于翀宇，贺红，等.采用改进YOLOv4算法的大豆单株豆荚数检测方法［J］.农业工程学报，2021，37（18）：179‐187.
GUO R，YU C Y，HE H，et al.Detection method of soybean pod number per plant using improved YOLOv4 algorithm［J］.Transactions of the Chinese Society of Agricultural Engineering，2021，37（18）：179‐187.
［13］王川，赵恒滨，李国强，等.基于改进YOLOX模型的芝麻蒴果检测方法研究［J］.河南农业科学，2022，51（11）：155‐162.
WANG C，ZHAO H B，LI G Q，et al.Detection method of sesame capsules based on improved YOLOX model［J］.Journal of Henan Agricultural Sciences，2022，51（11）：155‐162.
［14］LI X F，ZHAO S L，CHEN C L，et al.YOLO‐FD：An accurate fish disease detection method based on multi‐task learning ［J］. Expert Systems with Applications，2024，258：125085.
［15］刘拥民，张炜，麻海志，等.基于注意力机制的轻量化YOLOv5s蓝莓检测算法［J］.河南农业科学，2024，53（3）：151‐157.
LIU Y M，ZHANG W，MA H Z，et al.Lightweight YOLOv5s blueberry detection algorithm based on attention mechanism［J］.Journal of Henan Agricultural Sciences，2024，53（3）：151‐157.
［16］BEYENE D A，HUANGRUI，TOLA K D，et al.DCNAM：Automatic detection of pixel level fine crack using a densely connected network with attention mechanism［J］.Structures，2024，68：107073.
［17］WANG F Y，ZHANG D，ZHANG H W，et al.Semantic scene completion with cleaner self［C］//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Vancouver，BC，Canada：IEEE，2023：867‐877.
［18] TIAN Y N，YANG G D，WANG Z，et al.Apple detection during different growth stages in orchards using the improved YOLO‐V3 model［J］.Computers and Electronics in Agriculture，2019，157：417‐426.
［19］SUNIL G C，UPADHYAY A，ZHANG Y，et al.Field‐based multispecies weed and crop detection using ground robots and advanced YOLO models：A data and model‐centric approach［J］.Smart Agricultural Technology，2024，9：100538.
［20］DENG L，LUO S J，HE C H，et al.Underwater small and occlusion object detection with feature fusion and global context decoupling head‐based YOLO［J］.Multimedia Systems，2024，30（4）：208.
［21］BESER B，REIS T，BERBER M N，et al.YOLO‐V5 based deep learning approach for tooth detection and segmentation on pediatric panoramic radiographs in mixed dentition［J］.BMC Medical Imaging，2024，24（1）：172.

［22］REN Y M，LI Y Z，GAO X Y.An MRS‐YOLO model for high‐precision aaste detection and classification［J］.Sensors，2024，24（13）：4339.
［23］ZHU M H，GHAFFARI M，CLARK W A，et al. E2PN：Efficient SE（3）‐equivariant point network［C］//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Vancouver，BC，Canada：IEEE，2023：1223‐1232.
［24］赵永杰，蒲六如，宋磊，等.基于Shuffle-ZoeDepth单目深度估计的苗期玉米株高测量方法［J］.农业机械学报，2024，55（5）：235‐243.
ZHAO Y J，PU L R，SONG L，et al.Measurement method of seedling stage maize height based on Shuffle‐ZoeDepth monocular depth estimation［J］.Transactions of the Chinese Society for Agricultural Machinery，2024，55（5）：235‐243.
［25］LIU F F，WEI H R，ZHAO W Z，et al. WB‐DETR：Transformer‐based detector without backbone［C］//Proceedings of the IEEE Conference on Computer Vision. QC，Canada：IEEE，2021：2959‐2967.
［26］MA Q H，JIN S H，BIAN G，et al.Multi‐scale marine object detection in side‐scan sonar images based on BES‐YOLO［J］. Sensors，2024，24（14）：4428.
［27］JUNG D Y，OH Y J，KIM N H. A study on Gan‐based car body part defect detection process and comparative analysis of YOLO v7 and YOLO v8 object detection performance［J］. Electronics，2024，13（13）：2598.
［28］ZHAO H X，CUI H W，QU K M，et al.A fish appetite assessment method based on improved ByteTrack and spatiotemporal graph convolutional network［J］.Biosystems Engineering，2024，240：46‐55.
［29］XIAO C C，CAO Q，ZHONG Y J，et al.MotionTrack：Learning motion predictor for multiple object tracking［J］.Neural Networks，2024，179：106539.
［30］JANG E，LEE S J，JO H.A new multimodal map building method using multiple object tracking and gaussian process regression［J］.Remote Sensing，2024，16（14）：2622.
［31］于雁南，莫泳彬，严继池，等. 基于改进ShuffleNet V2的柑橘病害识别研究［J］.河南农业科学，2024，53（1）：142‐151.
YU Y N，MO Y B，YAN J C，et al.Identification of citrus diseases based on improved ShuffleNet V2［J］.Journal of Henan Agricultural Sciences，2024，53（1）：142‐151.
［32］钟正扬，云利军，杨璇玺，等. 基于改进YOLOX的自然环境下核桃识别算法研究［J］. 河南农业科学，2024，53（1）：152‐161.
ZHONG Z Y，YUN L J，YANG X X，et al.Research on walnut recognition algorithm in natural environment based on improved YOLOX［J］.Journal of Henan Agricultural Sciences，2024，53（1）：152‐161.

基于改进YOLOv8-Track 的芝麻单株蒴果检测计数研究

Detection and Counting of Sesame Capsules per Plant Based on Improved YOLOv8‐Track Model

PDF

可视化

摘要/Abstract

引用本文

使用本文

图/表 15

参考文献

相关文章 1

编辑推荐

Metrics