Research On Thin Cloud Removal In The Remote Sensing Image

Remoting sensing technology has been widely used in many fields of national economy and the society development.But the cost of the remote sensing data is too high,so that it is very important to take full advantage of the data.Cloud cover is one of the reason of the lack of the remote sensing data,it is very important to find a way to remove the effects of cloud cover in current condition.The work in this thesis focuses on the removing of cloud cover in single image.This paper introduce a way to remove the effects of cloud by frequency transform.Here,information of cloud cover is removed by high-pass filting and information of background is enhanced by homomorphic filtering and pixels replacing.Furthermore, a new way is particularly discussed based on wavelet transform.Final,propose a method which is combine homomorphic filtering with wavelet transform. Then choose three RGB single band image,and experiment to remove thin cloud with following methods:Improved homomorphic filtering,Wavelet transform,Edge detection combine with wavelet transform,Hmomorpgic filtering combine with wavelet transform,Homomorphic filtering combine with wavelet transform which add edge detection.The conclusions of the dissertation is listed as following:(1)Replace pixels can improve the result of classical homomorphic filter,but the level is not ideal.(2)The method of wavelet transform get a better effect.The new method which combine homomorphic filter with wavelet analysis remove cloud effectly as well as reserve the details as much as possible.This eassy has such special points as:(1)The new method can deal with simultaneously on the high and low frequency.It both remove cloud validly and reserve the details as much as possible. (2)Chose some better methods which have a good performance in single-band grayscale imange's cloud removing to combine with HIS in order to restorate cloud and reserve color information.Experiments shows that it gets a good result.(3)Based on the idea of improving the details information, the eassy integrate wavelet modulus maximum algorithm in image edge detection into various methods.