气象卫星红外云图与水汽图的融合技术研究

Study of Fusion of Meteorological Satellite Infrared Image and Water Vapor Image

将多分辨率分析融合方法和多尺度几何分析融合方法应用于气象卫星水汽图和红外云图的融合中,并用主观视觉、平均互信息和Xydeas-Petrovic指标对各种融合算法的性能分别进行了定性和定量评价。结果表明,与源图像相比,融合图像取得了更好的视觉效果,图像中包含了更多的信息量,云体清晰度和云的层次感得到了提高,纹理变得细致了。从平均互信息和Xydeas-Petrovic指标看,多尺度几何分析融合方法的效果较多分辨率分析融合方法更好。

The multi-resolution analysis fusion method and multi-scale geometric analysis fusion method are used for fusing the infrared cloud images and water vapor cloud images from FY-2E meteorological satellite.Moreover,the mean mutual information and Xydeas-Petrovic index are used to evaluate the performance of different fusion algorithms qualitatively and quantitatively.The results show that the visual effect of the fused image is actually improved relatively to the raw images.The fused image contains more abundant information about the cloud and the texture and hierarchy of cloud are improved as well. In light of two objective evaluation indexes,the multi-scale geometric analysis fusion method is better than the multi-resolution analysis fusion method in fusion effectiveness.