Research On The Removal Of Fog And Cloud In Satellite And Aviatic Images

Cloud and fog are the main factors in the influence on the clarity of the satellite and aviation images. This paper studies the removal of fog and cloud in the satellite and aviation photography images.Fog and cloud mainly have an influence on images by the scattering, absorption and radiating. In the band of the visible light, the influence of absorption and radiating are ignored, which is relative to scattering of the impact. The scattering exerts an influence to the visible light through three respects: Fog and cloud scattering the light forming no image causes their own formation of image; Fog and cloud causes the attenuation of the index law of the light forming image; The light forming image is continued the formation of image after fog and cloud scattering. Three mentioned respects cause the decline of contrast and resolution and the change of the luminance.In the course of obtaining of the aviation photography image, as a result of the granule scattering to the image, the images' quality decline obviously, which show in follow ways the decrease of Brightness, contrast gradient and resolution and images' blurring. This paper first removes the influences of the fog by image enhancement. The main feature of the degradation images is that the degradation influence of the fog is related with the distance between the objects and the camera. According to this correlation and the features of the degradation images by fog the method of multi-scale process was used in this paper. The self-adapted median filter is used then to reduce the noise of the result. This paper analyzed and evaluated the results of the methods in direct and objective way. The result of the tests shows that the methods used in this paper all has some effects on removing fog. The multi-scale process is better on restoring the resolution, brightness and contrast, removing fog and enhancing the object image.In the satellite images thin cloud and mist concentrates relatively in the low...