Satellite imaging is one of the most attractive sources of information for the governmental agencies and the commercial companies since the lunch of high resolution commercial satellites. It is very important especially for the military applications. Satellite images may have unwanted signals (noise) in addition to useful information due to several reasons such as bad sensor function (detectors and electronics), and imaging environment. Several noise removal methods can be used to eliminate or reduce the effect of noise over the image before information extraction.
In this paper a comparative study among four types of noise removal filters is carried out. The investigated filters are Median Filter, Wiener Filter, Average Filter and Bilateral Filter. These filters are applied on a test set of four high resolution remote sensing images acquired by different satellites (GeoEye.1, Ikonos, Spot.5 and World View2). The test images are contaminated by four types of noise: Salt and Pepper noise (SPN), Shot Noise (Poisson noise), Speckle Noise and Gaussian Noise. The results of applying the four filters are compared, evaluated and analyzed. The evaluation is conducted with the help of Mean Square Errors (MSE), Peak-Signal to Noise Ratio (PSNR), structural similarity index measure (SSIM), discrepancy(D) and Universal Image Quality Index (UIQI).
Satellite Image noise, Average filter, Median filter, Wiener filter, Bilateral filter, PSNR, MSE, SSIM, Discrepancy-Universal Image Quality Index (UIQI)
