高分辨率遥感影像融合存在的问题及改进措施

Shortages of the Data Fusion Approaches to High-resolution Satellite Images and Expected Improvements

影像融合技术作为富集和优化多源遥感影像信息的一种有效途径一直受到遥感界和图像处理领域的关注,并且已开发出多种融合方法。但是随着高分辨率遥感卫星的发展,当前一些融合方法对新型高分辨率遥感影像如IKONOS、Quick-Bird的融合效果欠佳,颜色差异是存在的主要问题。本文从高分辨率遥感影像光谱特征的分析入手找到颜色偏差的原因,同时给出了多种改进措施,然后分别利用传统的和改进的方法进行了融合试验,并对试验结果进行对比分析,以期找到对新型高分辨率遥感影像融合更加适用的方法。

With the development of remote sensing technologies, the image fusion becomes particularly important to acquire large-scale and detailed geospatial information for a variety of applications. But current available image fusion techniques are not efficient for the new high-resolution satellite images, like IKONOS, QuickBird, as well. Significant color distortion is one of the major shortages. In this paper, the reasons causing color distortion have been discussed, and some improved methods are also introduced. After that some experiments about QB Pan and MS image fusion have been presented with both the traditional techniques and improved methods. The results have shown that the new methods can reduce the color distortion significantly and the quality of fused images has improved.