STRIP SURFACE DEFECT RECOGNITION USING MARKING WINNER IN SELF-ORGANIZING NEURAL NETWORK BASED ON IMAGE INFORMATION

摘要

High-quality strip as automotive, household electrical appliances and other industries important raw materials, the surface quality is unable to meet the demand of users. The ultimate goal is Zero defect. Currently, as to surface defect of strip steel, it is difficult to accurately and effectively recognize which kind of defect and how serious defect on the strip steel. It is also the first step to import the surface quality of strip steel, in this paper, we propose a practical scheme self-organizing neural network (SOFM), which could prevent the situation of linear Indivisible or multi irregular space clustering error. The SOFM is an unsupervised learning algorithm of Artificial Neural Network (ANN). It allows the data organized onto a feature graph while conserving most of the topological characters of the original data space. In order to get multi subspace clustering, the SOFM is not unsupervised in this paper. The SOFM with winner neuron marking method is given. An example is presented in paper to illustrate the advantage of this method.