基于介电特性和BPNN建模的小麦含水率在线检测  

Online Detection of Wheat Moisture Content Based on Dielectric Properties and BPNN Modeling

作　　者：姬虹[1,2] 李康 宋东方 王万章[2] 李保谦 冯伟[2] Ji Hong;Li Kang;Song Dongfang;Wang Wanzhang;Li Baoqian;Feng Wei(Henan Polytechhnic,Zhengzhou 450046,China;College of Mechanical Engineering,Henan Agricultural University,Zhengzhou 450002,China)

机构地区：[1]河南职业技术学院,郑州450046 [2]河南农业大学机电工程学院,郑州450002

出　　处：《农机化研究》2025年第8期119-129,共11页Journal of Agricultural Mechanization Research

基　　金：中央引导地方科技发展资金项目(Z20221341068);郑州市科技惠民计划项目(2022KJHM0045);河南省科技厅科技攻关项目(242102110359);河南省科技厅重点软科学项目(242400411017)。

摘　　要：为满足小麦籽粒含水率在线检测需求,设计了一种搭载在联合收获机上基于介电特性的同侧圆弧电容式小麦籽粒含水率在线检测传感器。对6个不同品种小麦进行了温度、频率、电容、容重4个因素对含水率检测影响的实验研究,采用BP神经网络法建立了含水率与温度、频率、电容、容重4因素关系的预测模型,其训练集和测试集的决定系数R 2为0.896和0.893,均方根误差RMSE为1.317和1.342,预测模型稳定性和预测能力较强。研究表明:将温度、频率、容重这3因素引入的电容法联合收获机在线小麦含水率检测系统,能有效提高整体系统的检测精度和重复性。通过对不同小麦品种含水率检测影响因素相关性分析和数学模型的建立与优化,提高了电容法小麦含水率检测精度,为联合收获机小麦含水率检测系统中电容式传感器软硬件设计提供了理论依据。In order to meet the needs of online detection of moisture content of wheat grain,an online detection sensor of wheat grain moisture content based on the dielectric characteristics of the same side circular arc capacitance was designed on the combined harvester.The effects of four factors(temperature,frequency,capacitance and bulk density)on moisture content detection of six different varieties of wheat were studied.The BP neural network method was used to establish a wheat moisture content prediction model based on the relationship between moisture content and four factors(temperature,frequency,capacitance and bulk density).The determination coefficients(R 2)of its training and testing sets were 0.896 and 0.893.The Root Mean Square Error(RMSE)were 1.317 and 1.342,indicating that the prediction model had strong stability and prediction ability.The results showed that the temperature,frequency and bulk density were introduced into the online wheat moisture content detection system of the capacitive method combined harvester,which can effectively improve the detection accuracy and repeatability of the overall system.The correlation analysis of the influencing factors of moisture content detection of different wheat varieties and the establishment and optimization of mathematical models were carried out to improve the detection accuracy of wheat moisture content by capacitive method,and to provide theoretical basis for the software and hardware design of capacitive sensor in wheat moisture content detection system of combined harvester.

关 键 词：小麦含水率 在线检测 介电特性 BP神经网络 

分 类 号：S223.2[农业科学—农业机械化工程] S220.3[农业科学—农业工程]