YOLOv11-CoordAttention轻量化烟叶目标检测模型

打开文本图片集

关键词：YOLOv11-CoordAttention；轻量化；烟叶；目标检测模型

中图分类号：TP183；TP391.41;TS442 文献标识码：A

文章编号：0439-8114（2026）01-0152-07

DOI：10.14088/j.cnki.issn0439-8114.2026.01.025 开放科学（资源服务）标识码（OSID）：

A lightweight YOLOv11-CoordAttention model for tobacco leaf object detection

ZHANG Qian-zi， ZHU Yun-cong，DUQi-xia，ZHAO Wen-jun，LILi-hua，LI Xue-ming，DENGShao-wen，WANGJian-song，GAOYun-cai，CAOJing（HongtaTobacco（Group）Co.，Ltd.，Yuxi 6531oO，Yunnan，China）

Abstract：ToenhancetheperformanceoftheYOLOv11modelinintellgentgradingtasksfortobaccoleafobjectdetectionandtoaddress theissuesofaccuracyandtimelinessintobacoleafobjectdetectionwithinresource-constrainedenvironments，alightweight YOLOv11-CoodAtentiontobaccleafobjectdetecionmodelasproposed.Thfectivenessofvariouscomponentswasevaluatedby comparingtheimpactofdiferentbackbonenetworks，convolutionalmodules，andatention mechanismsomodelaccuracyandspeed. Ablationxperimentsresetupontisbasistoinvestigatethepracticalefectsofptimzedcombinatios，therebycompresively revealingthemodel’sperformanceinpracticalapplications.TheresultsindicatedthattheYOLOv11-CoordAtentionmodeldemonstrated superior comprehensive performance in the tobacco leaf object detection task，achieving a precision of 100% ，recall of 99.4% ， F1-score of 99.7% ， mAP50 of 99.5% ，with amodel size of 5.2 MB，2.3×10parameters，6.3x10°FLOPs，and a frame rate of 198.2 f/s. ComparedtotheYOLOv11model，theYOLOv11-CoordAtentionmodelimproved precisionby1.2percentagepointsandmeanaverageprecisionby0.percentageponts.ThetainingprocssoftheYOLOv-CorAtentionmodelwasstablefetivendehbitedoutstandingperformance.Thelossesforboththetrainingandvalidationsetssteadilydecreasedandconvergedasthetraining epochsincreasedindicatingasufiientleaingprocsswitoutoverfitingItersofperformancemetrics，temodelitad high precision and recall，achieving high accuracy and low missed detection rates. Its mAP50 and mAP50–95 metricswere both excellent，indicatingpowerfuldetectioncapabilityndighrobustessTheOLOv-CordAtentionmodelcombinedtheadvantagesof beinglightweightfintndate.Itoudustablyosoueostrainddevicdasompeentforobaccfde tection tasks in complex scenarios.

Key words：YOLOv11-CoordAttention;lightweight;tobacco leaf;object detection model

收稿日期：2025-04-26

基金项目：红塔集团科技项目（2022YL02）

作者简介：张千子（1994-），女，云南玉溪人，农艺师，硕士，主要从事烟叶外观质量检验研究，（电子信箱）06001130@hongta.com;通信作者，邓邵文（1991-），男，云南玉溪人，农艺师，主要从事烟叶外观质量检验研究，（电子信箱）06000589@hongta.com。（剩余11183字）