Research On Radiometric Correction Of The Key Techniques For Degraded Remote Sensing Image

It is more and more important to carry out the analysis of the earth's surface by remote sensing at present.Satellite,space shuttle,unmanned aerial vehicles and so on are able to get many remote sensing images.However,due to the complexity of the atmospheric environment,the result of optical remote sensing images can not truly reflect the surface information.Especially when there are clouds,fog and haze in the atmosphere,it is a great interference to the optical remote sensing imaging.It has a great influence on the ground object's quantitative analysis,change detection and mapping.The use of remote sensing images can be effectively reduced.In order to reduce or eliminate the influence of atmospheric environment on remote sensing image,we research on atmospheric correction and thin cloud removal of remote sensing images.Recovering ground information from degraded remote sensing images to improve the quality of remote sensing image and reduce the waste of resources.The main research contents and innovations are as follows:1.The study of common atmospheric correction algorithm and thin cloud removal algorithm is carried out.Experimenting on these correction algorithms,we find there are some defects on these correction algorithms when eliminating thin cloud.Especially when the fog or cloud is thick,the result is not very satisfactory.2.Aiming at the problem of atmospheric correction based on look-up table,slow construction and low precision.An atmospheric correction algorithm for remote sensing image based on atmospheric parameter regression equation is proposed.In the 6S atmospheric correction model,the aerosol thickness is used as the independent variable to fit the three parameters of the atmospheric parameters atmospheric backscattering ratio,the path radiation term equivalent reflectivity and atmospheric transmission.The algorithm is used to study the atmospheric correction of GF-1 multi-spectral images,The results show that the algorithm can effectively correct the influence of atmospheric on GF-1 multi-spectral image.It have a good consistency with FLAASH correction results.After the atmospheric correction of the algorithm,the ability to extract image vegetation information is more prominent.3.Research on algorithm of thin cloud removal for remote sensing image based on Haze Optimized Transformation(HOT)algorithm and Background Suppressed Haze Thickness Index(BSHTI)algorithm.In the study of BSHTI image optimization algorithm,it is found that although the morphological method can optimize the BSHTI image,but morphological knowledge also makes the algorithm more complex.In this paper,Red and Blue Spectral Difference(RBSD)and Normalized Difference Vegetation Index(NDVI)were introduced into the optimization of BSHTI images obtained by GF-1 images.The experiments show that the BSHTI image can be optimized by selecting the appropriate threshold.We use virtual cloud point(VCP)method to remove cloud,visual and quantitative evaluation parameters show that the ability of the extract ground information from the image after the cloud removal is improved.In this paper,the image of the image before and after the defog is matched respectively.The results show that the matching quality of the high score image is significantly improved after the process of remove cloud and defog.4.The algorithms of we research are integrated,and the degraded remote sensing image restoration system is designed and implemented.The system mainly includes the function module has the radiation processing module and the cloud processing module,The system interface is friendly and highly interactive.