基于HSI颜色空间和小波变换的多光谱图像和偏振图像融合实验研究  

Polarimetric and Multi-spectral Image Fusion Based on HSI Color System and Wavelet Transform

作　　者：陈振跃[1] 王霞[1] 邹晓风[1] 

机构地区：[1]北京理工大学光电学院,北京100081

出　　处：《光子学报》2010年第S1期43-47,共5页Acta Photonica Sinica

基　　金：武器装备预研基金(No.9140A01060110BQ0108);北京理工大学基础研究基金(No.20090442017)资助

摘　　要：偏振成像技术是当前光电成像领域研究的热点之一.为了更好地实现目标探测识别,本文研究了基于HSI颜色空间的偏振图像和多光谱图像的融合算法,对每一个波段内的偏振图像进行融合并映射到HSI颜色空间;然后对实验所选取的三个波段的H、S、I图层进行小波系数融合,并将融合图像在RGB颜色空间输出显示.给出了融合图像的客观评价参数,通过与前人的实验对比,表明该算法更加适合于人眼的视觉特性.融合图像用颜色凸显目标的偏振信息,能够将目标和背景明显区分开来,很大程度地提高了目标探测识别的质量.Polarization imaging technique is a hot spot in the field of optoelectronic imaging.Due to its great significance in the quality of object detection and recognition,polarization image and multi-spectral image fusion algorithms based on HSI color space were discussed.In each band,four line polarized images were fused to get three images which stood for intensity(I) image,degree of polarization(DoLP) image and angle of polarization(AoP) image,and they were mapped into HSI color space.Then a wavelet transform based method was introduced to reconstruct the wavelet coefficients of H,S,I layer in the three bands respectively.The final fusion result was shown in RGB color space.Experiments indicate that the proposed algorithm integrates spectral information and polarization information effectively and can improve the quality of target detection and recognition.

关 键 词：偏振图像 图像融合 HSI颜色空间 小波变换 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]