管类内表面检测图象分类器的研制

本文论述了多层前馈神经网络在工业图象视觉检测中的应用，并介绍了一个对高精度管道类内表面进行瑕点、条纹、裂痕等疵病进行实时检测的高性能图象识别系统。该系统采用多层感知器作为规则检测器对被测图象进行特征提取和分类识别，充分利用了人工种经网络所具有的信息分布式存储、大规模自适应并行处理以及高度的容错性等性能。

The Study of An Image Classifyer for Detecting Inner Walls of Pipes

This paper is concerned with an application of a multilayer feedforward neural network for the vision detection of industrial pictures, and introducesa high characteristic image processing and recognition system which can be used for real-time inspecting blemishes,streaks and cracks, etc. to the inner walls of high-accuracy pipes. Take fullyadvantage of the function of artificialneural network, suchas the information distribured memory, large scale self-adapting parallelprocessing, high fault-tolerant ability and so forth, this system uses amultilayer perceptron as a regular detector to extract features of the images inspected and classify the images.