摘要
近年来融合技术不断创新,方法多种多样,但是对融合结果缺乏统一的定量评价方法。在分析与总结当前常用的遥感图像融合结果定量评价方法的基础上,给出了亮度信息、空间细节信息和光谱信息等定量评价参数,并且通过程序实现了定量评价。以IHS和PCA两种融合方法对QuickBird图像多光谱波段和全色波段融合作为实验,进行定量评价分析,结果表明所提出的定量评价参数能够较准确地反映图像融合情况,比定性评价有效、全面,可为选择恰当的融合方法提供依据。
With the rapid development of technique on remote sensing image fusion, the methods of fusion have been ameliorated from pixel level to decision level recently. But how to quantificationally evaluate the resultant image impersonality and comprehensively is also a question. This paper brings forward brightness information, spa-texture information and spectrum distortion information by analysing and summarizing current methods of quantitative evaluation on remote sensing image fusion. Those parameters have been actualized via a program. At last, the multi-spectral bands and panchromatic band image of QuickBird sensors are fused with IHS （Intensity-Hue-Saturation）fusion and PCA （Primary Component Analysis）fusion. By evaluation the resultant with the parameters,it shows that quantitative evaluation is more available and comprehensive than qualitative evaluation. So the quantitative evaluation is suitable to evaluate the fused image. And it can support how to choose the appropriate method of image fusion.
出处
《遥感技术与应用》
CSCD
2007年第3期460-465,共6页
Remote Sensing Technology and Application
基金
国家自然科学基金项目"高分辨率遥感数据决策级融合方法研究"(40501047)资助
