基于BP神经网络番茄果实横纵径生长的建模研究

目的]准确快速地预测果实的大小,提高温室番茄种植的经济效益。方法]以2002和2003年果实横纵径部分试验资料为样本,根据BP神经网络的建模原理建立果实横纵径生长的模型,通过仿真试验检验了网络的仿真性能。结果]以MATLAB6.5作为系统仿真平台,对果实横纵径进行了建模仿真。利用神经网络工具箱命令:m,b,r]=postreg(A,T)对系统的仿真输出与目标输出作了线性回归分析,分析结果为:m=0.9987,b=0.0051,r=1.0000;m=1.0000,b=-1.7724e-004,r=1.0000。结论]基于BP神经网络的果实横纵径模型具有较好的仿真效果,在果实横纵径的预测中具有很好的应用前景。

Research on the Modeling of Transverse and Longitudinal Diameter Growth of Tomato Fruit Based on BP Neural Network

Objective] The purpose of the study was to predict fruit size accurately and rapidly and enhance the economic benefit of planting tomato in greenhouse.Method] With a part of experimental data about transverse and longitudinal diameters of fruit in 2002 and 2003 as samples,the model of transverse and longitudinal diameter growth of fruit was set up according to the modeling principle of BP neural network(BPNN) and the simulation performance of the network was checked up through simulation experiment.Result] With MATLAB6.5 as simulation platform for the system,the modeling and simulation were conducted on the transverse and longitudinal diameter of fruit.The linear regression analysis on the simulation and target output was done by the command of m,b,r] = postreg(A,T) from NN toolbox and the analysis results were as follows:m=0.998 7,b=0.005 1,r=1.000 0;m=1.000 0,b=-1.772 4e-004,r=1.000 0.Conclusion] The transverse and longitudinal diameter model of fruit based on BPNN had relatively better simulation effect and very good application prospect in predicting the transverse and longitudinal diameter of fruit.