为了实现图像处理与分析的自动化、智能化,把数据挖掘引入到图像处理与分析领域,针对图像纹理数据的特点,在深入研究数据挖掘软件的基础上,提出了一种图像纹理特征数据挖掘的理想的系统结构;介绍了基于图像(遥感图像)数据挖掘的软件原型系统RSImageMiner中图像纹理特征数据挖掘模块的各子功能模块的设计与实现;并以图像纹理分割为例,给出实验结果及分析。实验表明,该软件原型能得到较满意的图像纹理特征,并能有效地指导图像处理与分析。
For the intelligence and automatization of image processing and analyzing, data mining is introduced into image processing and analyzing. Aimed at the characteristic of image texture data, based on the indepth studies for data mining software, a novel ideal model of image texture feature data mining system was designed; each subsystem of image texture data mining module in RSImageMiner was designed and developed; then with examples of image segmentation, the result and some analysis of experiment are listed. And it suggests that preferable image texture features could be produced, and these features could guide image processing and analyzing effectively.