| abstract |
The invention discloses a seawater pollution area identification method and equipment based on high-resolution remote sensing images, and belongs to the field of digital image processing. The method firstly uses the supervised learning algorithm to automatically classify the remote sensing images of sea and land, and through the iterative clustering process, the classification results can reach a higher level of accuracy, and at the same time, compared with the existing sea and land boundary analytical classification methods, the amount of computation is relatively low. Then, using the difference in chlorophyll concentration and the brightness of pollutant shadows between the marine pollution area and the surrounding seawater, the normalized vegetation index related to chlorophyll, the normalized water shadow index related to brightness, and the The segmentation of the image interpretation idea is combined with the saliency mechanism based on human vision. The extraction of marine pollution areas is realized through threshold segmentation, and the areas with good water quality and severely polluted areas are extracted respectively, and the pollution transition areas are further obtained. The method of the invention provides a convenient and accurate reference for the prevention and treatment of marine pollution. |