A multiresolution method for internal wave detection and color display from satellite images

Abstract

In the present work an algorithm for wave detection is developed. The followed methodology is based on the multiresolution processing, the wavelet transform and the RGB model. This algorithm has been tested in three SAR images provided by the COSMOSkymed satellite from the ASI. First, Gaussian masks are generated in order to detect edges in the horizontal and vertical directions. This work is repeated in four resolution levels. Each level depends on the size and the variance of the gaussian mask. Once the horizontal and vertical edges are found, it is followed a non-maximal supresion processing to discard not relevant edges. This images are enhanced by morphological processing. To integrate the levels into a single image in order to interpretate the obtained results it is used the RGB model. Each one of the first three levels is introduced in a RGB image channel. Finally the pixels that appear in white are the corresponding to the most important edges, that is, the waves.