基于深度学习BP人工神经网络的农产品质量检测  

Quality Determining of Agricultural Products Based on Deep Learning of BP Neural Network

作　　者：方孝荣[1] 

机构地区：[1]金华职业技术学院,浙江金华321007

出　　处：《金华职业技术学院学报》2016年第3期73-76,共4页Journal of Jinhua Polytechnic

基　　金：国家高技术研究发展计划(863计划)"多尺度农田信息获取与融合技术"(2013AA102301)子课题"农田信息快速感知技术与装备开发与示范应用"(2013AA10230108)

摘　　要：农产品质量检测涉及到农产品加工、品质控制以及成分分析等非稳态系统,可以采用不同分析技术获取农产品质量相关的参数,通过数据分析实现对农产品质量的检测。现代农产品质量检测要求快速、准确,但根据检测获得的参数构建农产品质量信息的分析模型是研究的难题和关键。BP神经网络是人工神经网络中常用的一种模型,它以误差的反向传播不断地训练和调整网络,使结果最优;为确保模型有效性,利用深度学习的人工神经网络建立农产品生产、加工过程中农产品质量检测分析的模型,进而提高农产品的质量检测。Detection of agricultural products quality consists of agricultural products processing, quality control and composition analysis, which is an on-steady state system. Agricultural product quality parameters can be obtained by different analytical instruments, and the quality could be determined through analyzing the acquired parameters. Modern agricultural product quality requires fast and accurate detection, and how to use the measured parameters to build prediction models for quality detection is essential and important. BP （back propagation） neural network is a kind of error inverse propagation training algorithm for the muhilayer feed forward network, and is currently the most widely used model of neural network. Through deep learning, artificial neural network can be used to establish a model for detection analysis in the process of manufacturing and processing agricultural products, which will promote the further development of the agricultural industry. This paper mainly introduces the structure of neural network based on deep learning and its application in agricultural industry.

关 键 词：BP人工神经网络 深度学习 农产品品质检测 

分 类 号：TP274[自动化与计算机技术—检测技术与自动化装置] S126[自动化与计算机技术—控制科学与工程]