Coupling the stochastic matched filter and the à Trous algorithm for SAS image de-noising
Fabien Chaillan, Philippe Courmontagne · OCEANS 2006 - Asia Pacific · 2006
The wide research domain concerning SAS image de-noising shows the complexity of the problem. SAS devices designed to explore underwater world generate noise-corrupted data, strongly disturbed by the speckle noise, which affect both radiometric and spatial resolutions. Although many de-noising filtering techniques exist in the signal processing society and have shown their efficiency, they suffer of an important restrictive problem: the spatial resolution degradation. The matter is to design a processing having a strong de-noising power while preserving the spatial resolution. To perform this task, we present in this study a new way of SAS image de-noising consisting in coupling an efficient filtering technique, the stochastic matched filtering method, with a multi-resolution analysis technique, the a Trous algorithm. Comparison with some classical approaches on real SAS data reveal the efficiency of such an idea.