Removal of Cloud Cover And Noise From Remote Sensing Images
The advancement in modern science has seen a considerable advancement in technology. In today’s era, it is essential to be able to capture good quality images as the application of the same lies in various fields. There has been considerable work done to improve the quality of the satellite images taken but limitations persist. The main reason for these limitations is the ambivalence in algorithms and the inaccuracy in the methods to measure the system. The cameras and the other digital equipment are also responsible for adding the deformities in the image. The weather conditions along with any kind of external factor is responsible for deformities like- blur, fog, noise, haze, etc. The noise in the image are very difficult to completely remove from the image. In this paper we will talk about the images that are collected via remote sensors. We will basically focus on satellite images. We will be making use of Dark Channel Prior algorithm. The above chosen algorithm is a result of the survey done through which it was understood that DCP was successful in removal of fog or cloud cover and noise both at the same time. Moreover, the comparison of the output from the DCP method, done by qualitative and quantitative comparison with other algorithms resulted in an enhanced version of the images. Post the application of DCP to the image we apply Guided filters to the image. This is done to enhance the image through edge preservation. It is a wavelet-based method to obtain a sharp and clear image. We will make use of MATLAB to help us in the approach. The main objective of this is that the satellite images suffer distortions and we need to retrieve the original image so that we can perform operations on it like, segmentation. This is an essential pre-processing step which every image must go through to obtain clarity and consistency in it. Our estimations predict that this method will be at a competitive level with the other existing approaches.