An Infrared Ship Detection Algorithm Based on VIMI Infrared Multispectral Images

The Visual and Infrared Multispectral Imager (VIMI) on the GF-5 Satellite and the Hyperspectral Observation Satellite had a wide spectral range (0.45 μm~12.5 μm). Images with 12 bands from visible spectrum to thermal infrared spectrum could be obtained simultaneously based VIMI. It can monitor the maritime traffic in ports and related sea areas. A modified Infrared ship detection algorithm based on multiple infrared bands images was proposed in this paper. This algorithm was divided into two steps: Sea-Land segmentation; ship detection. The basic principle of infrared multispectral target recognition is the quantitative comparison and analysis of the spectral vectors and the standard spectral database. A modified algorithm was proposed in this paper to detect the ships. In the modified algorithm, a combination spectral operator was constructed with the spectral angle mapper and the spectral information entropy. This operator took use of their information complementary and creates their advantages to overcome the disadvantages of traditional SAM information divergence. Finally, the effectiveness of this algorithm was proved by the experiments with VIMI infrared images.