The Fusion Of Satellite Infrared Image And Visible Image

Severe convective weather disasters, such as storms, seriously threat to people's lives and property, even often cause serious damage to people's property. Therefore, strong convective weather monitoring and early warning systems research and deve-lopment, is the current hot spot of the academic study.Meteorological satellite has features with a large range of space monitoring,and provides meteorological cloud data which has continuity in its spatial and temporal. This article mains to make the various channels of Meteosat visible and infrared data fusion,from the original meteorological satellite cloud data, in order to improve data accuracy and precision, and lay a good foundation for the monitoring and early warning for the strong con-vective clouds.In the strong convective atmosphere activity, meteorological data which collects infrared and visible light data by meteorological satellite is the key data, which is used to monitor and warn a severe convective storms and other weathers.Besides,in the application,we mainly use brightness temperature of the infrared data to monitor the convective clouds. However, in the current meteorological satellite technology,the infrared images have low-resolution,while the meteorological satellites, visible light images have high-resolution. In this paper, we fuse the useful visible light image information to the infrared images, by the way of wavelet transform and multiresolu-tion analysis, to improve the resolution of infrared images in the visual effect.This article mains to discuss the formation principles, characteristics and differ-ences of the visible and infrared satellite cloud images,in order to make the foundation for the image fusion later.Besides, this article also studies the range of the infrared and visible light fusion, since meteorological satellite sensor visible channel has a disadvantage that it only works during the day, the resulting visible image has the problem of oblique. By the way of making the contrast of a series of visible and infrared images in the summer, we get the valid time interval of the visible data in the Beijing-Tianjin region.This article introduces the single-wavelet to the research for the meteorological satellite cloud images fusion, by using wavelet decomposition for the respective pairs of visible images and infrared images, we can get their low-frequency components and high-frequency components. Based on the actual application, in order to retain basic information infrared images in the fusion image,we make that the low-frequency parts of the fusion images adopt to all the low-frequency infrared, and high-frequency parts of the fusion image are obtained in this way, by which we use the method of local variance to select the high-frequency part of the infrared and visible images, to obtain the fusion of high-frequency components,finally it can be restructed to a higher resolution image by the inverse wavelet transform.Unfortunately, the disadvantage of the single-wavelet method is that, with the increase of decomposition level of wavelet transform,the loss of information of the fusion image will increase after reconstruc-tion, and even possibly produces the distortion phenomenon,in this basis,to improve single-wavelet transform, we use the multi-wavelet transform on the fusion of visible and infrared cloud images.This article uses a single wavelet image decomposition and reconstruction for reference, we apply the multi-wavelet transform to the fusion of meteorological satellite cloud images,to improve the accuracy of the fusion image. As the multi-wavelet transform has the characteristics of the multi-dimensional data flow, we need to do pre-filtering for the visible and infrared satellite images, so we do many experi-ments for several different pre-filtering methods,in order to prepare for the multi-wavelet transform fusion image. Finally, we study the image fusion quality evaluation method in depth,and introduce a series of qualitative and quantitative evaluation criteria and methods for the fused images, in the last,we compare the images which can be obtained by pairs of interpolation to be legitimate, the single-wavelet method based on infrared information and multi-wavelet method respectively, use the method of the subjective and objective evaluation to evaluate the fusion images, and compare the advantages and disadvantages of the three fusion methods, it is very important to access a number of important parameters, which are better for the warning and tracking for strong convective weather.