TM与SAR图像融合多方法研究及其效果定量评价

Multi-method study of TM and SAR image fusion and effectiveness quantification evaluation

本文针对TM多光谱与SAR高分辨率的特征,研究了彩色技术、数学运算和图像变换三类共11种融合算法,并基于ERDAS-SpatialModeler、ENVI-IDL和MATHLAB等技术分别实现这些算法。从信息量、清晰度和逼真度共四项指标出发,定量评价了各个方法融合方法的侧重点。利用极差变换法处理指标,建立综合评价模型,得出单方法RGB(0.75)法相对最优、Brovey(0.10)法最劣以及三类方法彩色技术(2.59)最优;图像变换(2.25)次之;算术运算(0.82)最不可取的结论。

This paper studies the three types of fusion algorithms——color technology;mathmatics computation and image sharping totally eleven methods in view of the hyperspectral of TM and the high-resolution of SAR,and realizing these theories based on softwares of ERDAS-Spatial Modeler,ENVI-IDL and MATHLAB.Given effectiveness quantification evaluation for each method by four standards of information amounts;clarity and fidelity.It is proved that RGB（0.75） is the relative best method,while Brovey（0.10） worst;and Color technology（2.59） best;image sharping（2.25） secondary and mathmatics computation（0.82） the most not desirable in three types of fusion algorithms after processing standards and building model.Study shows scientific significance to the fusion methods selection of TM and SAR.