基于Fert-YOLO的高梁育性检测模型研究

打开文本图片集

中图分类号：S514 文献标识码：A 文章编号：1671-8151（2025）04-0046-11

Abstract：[Objective]Asanimportantcropforbothfoodand energyproduction，sorghum'sfertilitydetectioniscrucial forarietybreedingandyieldimprovement.However，traditionaldetectionmethodssuferfrom loweficiencyduetocomplexfield backgrounds，necesitating highly eficientandaccuratedetectiontechnologies.[Method]ThisstudyproposedFert-YOLO，a lightweightdetectionmodelforsorghumfertilityasedonYOLOv8n.First，multipleoflinedataaugmentationmethodswere usedtoenhancedatadiversityand improve themodel’sgeneralizationability.Second，toreducenetworkcomplexitywhileeffectively improvingdetection performance，StarNet wasused toreplaceYOLOv8n's backbone featureextraction network.In thefeature fusionstage，theC2Fmodule was redesigned by incorporating mixed localchannelatention（MLCA）mechanism， strengthening the network'sabilitytocapturecriticalfeatures.Finally，alightweight sharedconvolution detection（LSCD） head was introduced，which sharedconvolutionallayerparameters tosignificantlyreduce modelsizeandcomplexity.Results] TheFert-YOLO modeldemonstrated outstanding performance in sorghum fertilitydetection.Compared to theoriginal YOLOv8n model， it achieved a 1.5% improvement in mean average precision（ （Map0，5） ，further enhancing detection accuracy. Additionally，，the model’s，floating-point operations per second （FLOPs）and parameters were reduced by 40.0% and 47.8% ， respectively，significantly improving inference speedanddeployment eficiency.Whencompared toothercommonsingle-stage lightweightdetectionmodels，Fert-YOLOshowedclearadvantagesinbothdetectionaccuracyandmodelefficiency.Conclu sion]Thisresearchprovidedareliabletechnicalsupportforeficientsorghumfertlitydetectioninfieldconditions，contributing significantly to smart sorghum breeding and precision agriculture.

Keywords：Sorghum，Fertilitydetection，YOLOv8n，Model optimization

高粱［Sorghumbicolor（L.）Moench]是中国主要的粮食作物之一，其可用于食用、酿造、生物燃料、饲草等多方面，在国民经济中占有重要地位[1-2]。（剩余14111字）