基于CA_MobileViT模型的水稻钾素营养诊断研究

打开文本图片集

doi：10.13304/j.nykjdb.2024.0146

中图分类号：S126 文献标志码：A 文章编号：1008-0864（2025）08-0080-09

Diagnosisof Potassium Nutrition inRiceBased on CA_MobileViTModel

WU Zheng1，YANG Hongyun²*，SUN Aizhen1，KONG Jie²，HUANG Shumei² （1.SchoolofComputerandInformationEnginering，JiangxiAgricultural University，Nanchang330o45，China;2.hlf Software，Jiangxi Agricultural University，Nanchang 33OO45，China）

Abstract： In order to rapidly and accuratelely diagnose and recognize potassium stress degree of rice，The Huanghuazhan varietyoflatericewasusedas material，and4potassiumfertilization levels weresetrespectivelyfor base fertilizer and topdressng at jointing stage.Taking scanned images of3spread leavesof main stem at tiller stage and jointingstageasdataset，and MobileViTastheskeleton，the CA MobileViTmodelwasconstructedafter introducingCoordinate Attentionto the BNLayer of each 3×3 convolutioninthelayerof MobileViTand introduce transferlearning.Theresults were verified bycomparing EfficientNet-V2，ConvNeXt，MobileViTand CA - MobileViT models.The results showed that the accuracy rates of the 4 models were 98.4% ， 98.5% ， 94.2% and 95.3% ， respectively. The accuracy of CA_MobileViT model was 95.3% ，1.1 percent point higher than that of MobileViT model，but3.1percentpoint lowerthanthatofEficientNet-V2modeland3.2percentpointlowerthanthatof ConvNeXt model.However，the numberof parameters of CA_MobileViTmodels was about 1/4of the EffcientNet-V2 and ConvNeXt large models，and the training time was reduced by about 1/3 .Theimproved CA_MobileViT model hasahigh accuracy forthe diagnosis of potassium stress degree of rice，and could effectively guide the scientific potassium topdressing management of rice，and also provideauniversaland feasible method for the rapid and accurate diagnosis of nutrition of other crops.

eyWords：rice；potassium nutrition diagnosis；lightweight hybrid model;transferlearning；Coordinate Attentiol

钾素作为农作物生长发育所必需的矿物质元素之一，在植物生命活动中极其重要，它可以促进碳水化合物的合成和运输，促进蛋白质的合成1土壤自身富含多种矿物质，是植物根系获取钾素的主要来源。（剩余14065字）