摘要
提出一种针对当前不同卫星传感器数据融合的新方法。该方法基于Brovey融合方法的思想,充分考虑了不同卫星传感器全色影像与多光谱影像的光谱范围差异以及光谱响应差异,通过公式推导建立了基于权重系数β和比例系数w两个因子的全色影像与多光谱影像的关系式,并根据这两个因子重新构建了Brovey融合过程中的乘积系数。改进后的方法有效地改善了传统Brovey融合方法的光谱畸变问题。将上述方法应用于北京1号、SPOT4/5、Landsat5(TM)以及环境一号卫星数据之间的4例融合实验中,并与Brovey融合、Modified IHS融合方法进行定性和定量评价,结果表明其综合性能优于其他方法,在细节融入度高的基础上,仍能保持良好的光谱信息,而且保留了Brovey融合快速的优点,易于推广和应用。
We proposes a novel fusion method for multi-sensor image fusion, whose basic thought comes from Brovey transfor- mation. This method exploits a new calculation formula of the product coefficient in Brovey transformation through considering spectral range differences and spectral response differences of different sensors in multi-sensor image fusion.Based on the spec- tral range differences and spectral response differences of various sensors, the inferential reasoning formula was proposed. The relationship between panchromatic and multi-spectral images was built using weighting coefficient fl and proportion coefficient w. Finally, the product coefficient in Brovey transformation was rebuilt by the two factors.The novel multi-sensor image fusion method has greatly reduced spectral distortion in Brovey transformation. To evaluate its performance and efficiency, four experi- ments using proposed fusion method, Brovey transformation and Modified IHS were carried out, including fusion of Beijing-l small satellite （BJ-l） multi-spectral （MS） image and SPOT 5 panchromatic （PAN） image, the SPOT 4 MS image and BJ-l PAN image, the Landsat 5 （TM） MS image and BJ-1 PAN image, and the HJ-lA MS image and SPOT 5 PAN image. Qualitative and quantitative evaluations of the tests were also conducted. The results showed that the proposed method is better than the other two methods generally. It shows good ability in spectral preservation and high spatial details immersion. The simpleness and timesaving make it high potential in future applications.
出处
《遥感学报》
EI
CSCD
北大核心
2012年第2期343-360,共18页
NATIONAL REMOTE SENSING BULLETIN
基金
北京市科技计划(编号:Z111101061710001)~~
关键词
多卫星传感器
光谱范围差异
光谱响应差异
乘积系数
multi-sensor, spectral range differences, spectral response differences, product coefficient
