Satellite-based Synthetic Aperture Radar (SAR) provides a powerful vessel surveillance capability in front of time consuming traditional reconnaissance methods. Nevertheless, due to the presence of speckle and to the reduced dimensions of the targets compared to the sensor resolution, the automatic interpretation of SAR images is often complicated even though vessels undetected are sometimes visible by eye. Therefore, the main difficulties of automatic ship detection will be reviewed and the main drawbacks will be identified. As a result of this preliminary analysis, the application of wavelet tools is found to be a useful mathematical object for this purpose. So, according to the capabilities of the wavelet analysis, the objective is to provide other tools for a better understanding and further exploitation of satellite high resolution images. More specifically, a novel method for ship detection based on multiscale tools is proposed, justified and tested over simulated and real images.
展开▼
