无损检测中的基于模糊神经网络数据融合技术

Multisensor Data Fusion Based on the Fuzzy Neural Network in NDT

数据融合是一门正在飞速发展的信息处理技术 ,它能从多源信息中融合信息 ,减少信号的不确定度并再现出一个全面的信源。传统的数据融合模型大都建立在统计理论模型之上。存在的缺点主要表现在要求先验知识 ,应用领域受到限制。基于模糊神经网络的数据融合是一个新的领域 ,本文给出了一种适用于多传感器数据融合的模糊神经网络结构。文末给出了它在无损检测中的应用。文章对该模型的实现及其特点进行了详细讨论。它在无损检测中的应用表明该模型解决了传统模型中存在的问题 ,同时表明该模型也可用在其它许多领域。

Data fusion is a fast-growing signal processing technique.It combines information from multi sources,reduces signal uncertainty and improves the overall performance of the sources.Traditional data fusion models mostly base on the statistic theory and have some shortcomings.For example,they need prior knowledge and the application areas are limited.Data fusion based on the fuzzy neural network is a new area.This paper presents a new fuzzy neural network model adapted to multisensor data fusion and gives its application in NDT.The realization of this model and its characteristics are discussed in detail.Its application in NDT illustrates that this model solves many problems of the traditional model and can be used in many other areas.