Study Of Thin Cloud Removal For High Spatial Resolution Satellite Images

Since GF-1 satellite was in service,it has been widely used in the fields of disaster detection,resource exploration and environmental protection.As for high-resolution satellite data,it makes it possible to observe the changes of surface details on the small spatial scale and to detect the environmental impact of human activities,has important significance.However,since the quality of the image data which the high spatial resolution device captured will be influenced by cloud and lead to different degrees of decline.In the area of thick cloud cover,the underlying surface information is completely lost.In the area of thin cloud cover,Although the quality of the image is degenerated,the image still preserve available information of underlying surface quality.In order to improve the level of quantitative interpretation and utilization ratio of image,remove the influence of thin cloud on satellite image effectively,this paper studies the removal of thin cloud.The main contents and conclusions of the paper are as follows:(1)Firstly,we introduce the type and characteristic of the image,and then summarize the degradation model of the conventional remote sensing image and the image model affected by the thin cloud.From the aspects of spatial and frequency characteristics,the characteristics of remote sensing image cloud area are analyzed.(2)Then,a method of remove thin cloud based on Mallat wavelet transform is proposed.First,it is possible to decompose the image into high frequency detail and low frequency approximation components.Based on the facts that cloud noise occupies lower frequency part in the distribution characteristics while scenery information makes up the relative high part,this kind of algorithm processes cloudy zones by using linear methods according to cloud thickness in the largest scale of low frequency sub-band image;as for high frequency sub-band,different scales of non-linear enhancing operators are applied in order to improve images' sharpness and reduce the impact of residual cloud.Finally,median filtering is added to process the image to reduce the influence of high frequency mutation cloud.Taking the GF-1 as an example,this kind of method can achieve better results.(3)Designing software to remove thin clouds,it will realize TIFF images read?display?save?Interception of sub images?wavelet decomposition and reconstruction?two methods of thin cloud removal and exiting procedures implemented in function,and giving the results of the corresponding graph.(4)Taking the GF-1 as an example,this kind of algorithm is utilized to process GF-1 images.Experiment shows that kind of algorithm can remove thin cloud while it can also preserve image details the same time,which indicates it is better than traditional wavelet transform method.