摘要
针对玉米精准施肥模型与土壤多个参数有复杂的关系,且动态性数据量庞大、维度高、具有非线性强的特点,本文基于云计算技术的数据存储、分析、共享的功能,运用神经网络较强的处理非线性问题的能力,对玉米土壤养分数据进行处理,构建玉米土壤实时监控云平台,进行基于优化RBF神经网络的玉米土壤养分施肥模型研究,明确模拟施肥量与玉米产量、土壤养分含量的关系,将该模型在农安县陈家店村应用,为玉米的精准施肥提供决策依据。结果表明,基于云计算与RBF神经网络集成的玉米精准施肥模型与传统网络进行比较减少了误差,节约了时间,可为玉米精准施肥提供咨询指导,促进对玉米精准农田管理的实施。
The precise fertilization model for maize has a complex relationship with multiple parameters of soil,also,it possesses dynamic data and high-dimension characteristics,indicating strong non-linearity.This article utilizes the functions of data storage,analysis,and sharing provided by cloud computing technology,uses neural networks,which have a robust ability to process soil nutrient data in a nonlinear fashion.Firstly,a cloud platform for real-time soil monitoring was built,followed by the researches on soil nutrient fertilization models optimized by the Radial Basis Function(RBF) neural network.Finally,the model simulates the relationship between the fertilizer rate,maize yield,and the soil nutrient content.The model was applied to Chenjiadian village,Nong'an county,which is the national spark program demonstration base,providing a decision-making basis for precise fertilization of maize.The experimental application results showed that compared with the traditional networks,the integration of the precision fertilization model for maize based on cloud computing and RBF neural networks decreased errors and saved time.Thus,it can provide consultation and guidance for precise fertilization of maize and promote the implementation of precise farmland management of maize.
作者
贾海峰
李玥峤
黄帅
陈桂芬
赵姗
李楠
JIA Hai-feng;LI Yue-qiao;HUANG Shuai;CHEN gui-fen;ZHAO Shan;LI Nan(Jilin Institute of Education,Changchun 130022;Jilin Academy of Agricultural Sciences,Changchun 130033;School of Information Technology,Jilin Agricultural University,Changchun 130118,China)
出处
《玉米科学》
CAS
CSCD
北大核心
2023年第6期128-134,共7页
Journal of Maize Sciences
基金
国家“863”项目(2006AA10A309)
国家星火计划(2015GA660004)
吉林省重点科技研发项目(20180201073SF)。
关键词
玉米
云计算
精准施肥模型
RBF神经网络
土壤空间变异
Maize
Cloud computing
Precision fertilization model
RBF neural network
Soil spatial variability
