基于BP人工神经网络的某型雷达降水估计误差修订建模  

Model of Error Correction for Radar Precipitation Estimation Based on BP Neural Network

作　　者：夏光滨[1] 方勇[2] 

机构地区：[1]91550部队15分队 [2]91550部队91分队

出　　处：《舰船电子工程》2016年第11期65-67,102,共4页Ship Electronic Engineering

摘　　要：文章分析了雷达降水估计中误差的来源,详细讨论了神经网络的结构、学习算法、权值调整和程序流程的具体设计,构建了对雷达降水估计与地面降水实测数据之间的误差修订模型,神经网络的初步训练取得了较好的应用效果,是减少非系统性误差、提高雷达使用效能和增强保障能力的有效方法。This paper analyses the origins of the error for radar precipitation estimation, specifies the fabric of the neu-ral network and the adjusting of weights with learning-arithmetic, constructs a model of error correction for radar precipitati-on estimation The primary training of the network has been proved availability to application and it is an effective way to reduce the non systematic error, improve radar performance and enhance security capablity.

关 键 词：神经网络 降水估计 误差修订 

分 类 号：TN95[电子电信—信号与信息处理]