摘要
首先探讨了基于像素的多源遥感影像高频调制融合法 ,根据成像系统特性和Heisenberg测不准原理 ,设计的高斯滤波器对高分辨率影像滤波的方法是合理有效的。在研究BP神经网络的基础上 ,采用动量法和学习率自适应调整的策略 ,提高了BP神经网络学习算法收敛速度 ,并增强了算法的可靠性。提出并实现了多源遥感影像像素级融合分类与决策级分类融合两种分类方法 ,并进行了比较。采用LandsatTM 3 ,4,5和航空SAR影像进行试验 ,结果表明两种分类方法是行之有效的 ,均适用于多源遥感影像分类。
With the availability of multi_sensor,multi_temporal,multi_resolution and multi_spectral image data from operational Earth observation satellites ,the image fusion has become a valuable tool in remote sensing image evaluation .It is a relatively new and rapidly developing research field in remote sensing .In this paper,a pixel_level fusion algorithm of multi_source images in remote sensing based on high frequency modulation is studied.According to the characters of imaging system and principle of Heisenberg,a Gaussian filter is designed and used in the algorithm,which is proved to be effective.A back_propagation feed forward artificial neural network using momentum and adjusting learning rate by self_adaptation is studied.The speed and reliability of BP neural network are improved.A pixel_level fusion procedure and a decision_level fusion procedure for classification of multi_source remotely sensed images are proposed.A multi_source image set including Landsat TM3,4,5 and SAR has been used in classification.Compared with their classification accuracy obtained by the two procedures,the results show that the two procedures applied in classification of multi_source remotely sensed images are effective.
出处
《武汉大学学报（信息科学版）》
EI
CSCD
北大核心
2001年第5期430-434,共5页
Geomatics and Information Science of Wuhan University
基金
国家测绘局测绘科技发展基金资助项目 ( 980 15 )
关键词
高通滤波
影像融合
BP神经网络
遥感影像
high frequency modulation
image fusion
BP neural network
classification
