摘要
作者在本文中讨论了利用电磁波散射的数值计算、高光谱、人工智能进行遥感图像理解的现状 ,说明了地理信息系统 (GIS)数据融入遥感图像理解的重要性 ,概述了 GIS数据融入所面临的问题 ,指出了 GIS融入遥感图像理解对处理系统的一般要求 ,分析了基于数据挖掘的专家系统、神经网络、进化计算的特点和性能。结合神经网络和进化计算能融合多源数据、高度并行、自适应、自组织能力和知识处理的能力 ,以及进化计算通过重组、变异和复制算法具有优化选择的功能 ,构建了基于进化计算的神经网络的 GIS数据 ,并融入模型。模型中 GIS属性数据惯穿遥感图像理解的全过程 ,是一种高层次的数据融合 ,且 GIS数据特征的提取、处理和遥感图像理解是高度并行的。它是一种实现 GIS数据融入遥感图像理解的有效途径。基于数据挖掘的专家系统也具有上述功能 ,但存在数据挖掘的困难。
This paper discusses the condition of image understanding of remote sensing (RS) using numerical analysis of electromagnetic wave, high spectrum and artificial intelligence. It summarizes the problems of the data from geographic information system (GIS) fusing into RS and analyses the characteristics and capability of the expert system based on the data excavating, neural networks and evolution computation. Neural networks and evolution computation have the characteristics of self-adaptive, parallel, the capabilities of fusing the various data and processing the knowledge. Evolution computation has the capability of optimizing selection by regrouping, variation and coping. The model of neural networks based on evolution computation was built. The data of GIS is fused through out the modeling when the image understanding is running. The fusion is high-level one of GIS with RS. The model can realize the data fusion of GIS into image understanding of RS with a better performance than that by the expert system based on the data excavating.
出处
《物探化探计算技术》
CAS
CSCD
2002年第1期62-67,共6页
Computing Techniques For Geophysical and Geochemical Exploration
