基于PSO-GA-BP神经网络的大米水分含量预测  

Prediction of rice moisture content based on PSO-GA-BP neural network

作　　者：黄艳雁[1] 姚健 HUANG Yan-yan;YAO Jian(Guangxi Vocational&Technical Institute of Industry,Nanning 530001,China)

机构地区：[1]广西工业职业技术学院,广西南宁530001

出　　处：《粮食与饲料工业》2023年第3期56-60,共5页Cereal & Feed Industry

摘　　要：为保障大米质量安全,提出一种基于BP神经网络的大米品质检测方法。方法以大米水分含量作为大米品质指标,通过结合粒子群优化算法(PSO)和遗传算法(GA)优化BP神经网络的阈值和权值,实现了BP神经网络的改进,提高BP神经网络的收敛速度,解决BP神经网络容易陷入局部最优的问题。然后,将改进的BP神经网络应用于大米水分含量预测中,实现了大米水分含量的准确预测。仿真结果表明,所提的改进BP神经网络模型相较于标准BP神经网络模型和LSTM-BP神经网络模型,在预测集和验证集上的均方根误差更小,分别为0.007和0.005,其大米水分含量预测值与真实值接近,可准确检测大米水分含量,为大米品质检测奠定了理论基础。In order to ensure the quality and safety of rice,a rice quality detection method based on BP neural network was proposed.Taking the moisture content of rice as the rice quality index,the threshold and weight of BP neural network were optimized by combining Particle Swarm Optimization(PSO)and Genetic Algorithm(GA)to realize the improvement of BP neural network,improve the convergence speed of BP neural network,and solve the problem that BP neural network was easy to fall into local optimization.Then,the improved BP neural network was applied to the prediction of rice moisture content,and the accurate prediction of rice moisture content was realized.The simulation results showed that compared with the standard BP neural network model and LSTM-BP neural network model,the improved BP neural network model had smaller root mean square error in the prediction set and the verification set,which were 0.007 and 0.005,respectively.The predicted value of rice moisture content was close to the true value,which could accurately detect the moisture content of rice,laying a theoretical foundation for rice quality detection.

关 键 词：大米品质检测 BP神经网络 高光谱图像 粒子群优化算法 遗传算法 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]