Research On The Enhancement And Color Normalization Algorithm Of Remote Sensing Image

With the development of computer technology and the upgrading of information level,remote sensing has been developed rapidly as one of the most rapid means of acquiring spatial information,and has been applied in surveying,military reconnaissance,land and resource management and other fields.As the carrier of remote sensing information,the image quality of remote sensing image directly influences the later processing and use of remote sensing image.In this paper,the algorithm of automatic fog enhancement and color consistency of remote sensing images in aerial and aerospace Photogrammetry is studied.Low altitude photogrammetry is susceptible to the effects of fog and haze on image quality,the accompanying atmospheric scattering and absorption in fog and haze result in the decrease of contrast,unclear target features,blurred detail information and color distortion,which caused serious disturbance to the subsequent processing of the image.In this paper,a haze removal algorithm is designed for the aerial image of the measurement area based on the theory of the prior principle of dark primary colors.By using the Otsu method,the image Dark Channel segmentation is automatically removed from the image which does not conform to the dark primary color,and the global optimal estimation of the transmission rate is realized by the whole image of the surveying area,and the transmission rate is optimized for the part of the single image which does not conform to the dark color.Finally,the atmospheric physical model is used to enhance the processing of the original image.After the experiment,the image visual quality was improved obviously,and the extraction quantity of the feature points was improved remarkably.Large-scale satellite imagery is affected by the seasons,light and weather conditions,such as the impact of color distortion,the same area of the image tone often there are significant differences,resulting in the mosaic of the positive image of the visual continuity is poor,the performance of a "large cotton cloth",seriously affecting the interpretation and use of satellite images.In this paper,a color consistency processing algorithm of a large range of satellite images using datum tones is proposed,which can be used as the base tone image of the existing or network-exposed and color-consistent low-resolution satellite,and the original image is corrected by replacing the low-frequency information of the original image and constructing the independent linear model of image.In order to deal with the color consistency of the whole image,the method does not have the problem of accumulative error and is suitable for the color consistency processing of large area satellite image.The aerial image is affected by the stability of flight platform,the flight angle and the photosensitive characteristic of the imaging system.The difference of brightness and hue is often existed in the image of the same area,in this paper,we propose a method to estimate the color correction coefficient of the region image by increasing the random observation.By using the overlapped region of aerial image,the local illumination compensation of the single image is realized,the corresponding datum-tone base map is generated,and the color correction of the original image is realized by using the color mapping technique,which effectively solves the problem of "bottom pot effect" in aerial image.Experiments show that the algorithm has the characteristics of fast processing speed,stable effect and strong robustness.In this paper,we study the radiation-related problems such as enhancement of optical remote sensing image and color consistency in photogrammetry,and design a targeted algorithm,which verifies the validity and reliability of this algorithm.