Noise identification and removal by using higher order statistics and ICA Transformation for satellite images
T. Venkatakrishnamoorthy, Galiveeti Umamaheswara Reddy · 2019
Satellite images play a crucial role in meteorological operations for analyzing the atmospheric related information, urban & vegetation studies and sea surface temperatures. This information is adversely affected by noise pixels that are additionally added into an image, the spatial and spectral characteristics and pixel information are also changed by these components. The identification of noises is very important, if the noise is identified, suitable filters are applied to remove the noise from the actual image. The identification of impulse and speckle noises are very difficult if it exceeds above 40%.Using this proposed technique, the speckle and impulse noises are identified by the Higher order statistics which is based on smoothing factor and later it removes noises using Independent component analysis transform method. The proposed method gives the best PSNR when compared with other standard methods.