A Connected Component Based Offshore Platforms Extraction Method of GF2 Multi-channel Optical Images

In order to solve the problem of extraction of offshore platforms, this paper proposes a new automatic extraction algorithm for offshore platforms based on high resolution optical remote sensing images. The algorithm includes these key procedures: the multi-channel processing, the OSF(Order Statistical Filtering) algorithm and the morphological operation. First, the multi-channel difference value is used as the input of the OSF to enhance image quality, then the wave is removed by the statistical characteristics of the connected component. Finally, the ship targets are removed based on the relative invariance of the position of the platform target. The effectiveness of the automatic extraction method is verified by GF2 image experiments.