Author : Mohammad Reza Mobasheri, Erfan Amraei, Seyyed Amir Zendehbad [ Special \\\"9thSASTech-2015, Iran\\\"-2015] [Page No : 7-16] [2015]

Nowadays satellite images are widely used to monitor local phenomenon and conduct environmental studies. Unfortunately the existence of some problems in electronic system of the sensors reduces the data quality and appearance of noise in the images. The most common type of noise in these images is the striping noise created as a result of some problems such as the detectors mismatch, improper calibration of the sensor and its fatigue during the lifetime of the satellite. One of the methods to striping noise removal is image filtering. In this paper, we compare the performance of the wavelet based filters, the frequency domain and spatial domain filters in striping noise removal. For this purpose, the images acquired by Landsat 4 Multi Spectral Scanner as well as the simulated images for striping noise are used. Finally, to evaluate the results some statistical parameters such as mean, standard deviation, relative error, RMSE, MSE and PSNR are used that indicate the priority of wavelet-based methods over other methods.