基于灰度共生矩阵和BP神经网络集成的纹理图像分*类

Texture Image Classification Based on Gray Level Co-occurrence Matrix and BP Neural Network Ensemble

通过对灰度共生矩阵的分析,提取图像的纹理特征参数,并用BP神经网络集成的方法对Brodatz纹理库图像进行分类,仿真结果显示,其分类效果优于单一的BP神经网络,可有效提高分类识别率。

This paper investigates the collection of textural parameters according to the analysis of grey level co - occurrence matrix and classification of Brodatz texture database by means of BP neural network ensemble. The emulation result shows that the classification effect is superor than that from simple BP neural network, the classification discrimination is increased effectively.