基于深度学习的农作物病虫害识别系统设计

打开文本图片集

中图分类号：TP183 文献标志码：A 文章编号：2096-9902（2026）03-0023-0

Abstract：Inordertosolvetheproblemsoflowfficiencyunstableaccuracyandinsuficientpertinenceofpreventionand controlsgestionsintraditionalropestientificationthisrsearchimstopromotethedevelopmentofinteligentgiculture andcombinesdeeplearningtechnologytodevelopaninteligentidentificationsystemforcroppestandinsectpestsbasedonthe VGG16odel.Thestudyfirstconstructedalarge-scalepestanddiseasedatasetcovering13majorcashcropsinGuangxiand containing26Oohigh-precisionannotatedimagesandoptimizedthetrainingprocessoftheconvolutionalneuralnetwork（CNN） modelthroughtransferleamingtechnology;thesystemcanrealizereal-timeuploadandphotorecognitionofpestanddisease images.，ompletefeatureeractiontypeclasificationanddegreeofdamageasessment，andoutputulti-dimesionalsuggestions includingphysical，chemicalandbiologicalcontrol.Testresultsshowthatthemodelrecognitionaccuracyrateofthesystem reaches 95 % ，andthetimeconsumingforsingleimagerecognitioniscontrolledwithinO.5seconds.Insummary，thesystemcan efectively meetthepestanddiseasecontrolneedsof diferentusersand providetechnicalsupportforsmartagriculture.

Keywords：convolutionalneuralntwork（NN）；smartagiculture;croppestidentifcation;inteligentsytemdesin;WeChat Mini Program

随着农业现代化进程的持续推进，农作物病虫害防治工作在提升农产品产量和质量方面的核心地位越发凸显。（剩余9500字）