- Method for automatic inland water boundary extraction using remote sensed hyperspectral images

摘要

The present invention relates to a method for automatically extracting a water-land boundary of a freshwater area using a hyperspectral remote sensing image, (a) receiving superspectral image data from an ultraspectral camera by an input device, and preprocessing the image by a preprocessing module ( Radiation correction, atmospheric correction, and geometric correction); (b) the decision tree pre-classification module includes pre-classifying the water region and the land region using the hyperspectral image that has been subjected to the pre-processing, but includes misclassification; (c) the decision tree reclassification module further subdividing the pre-classified and misclassified water region into at least one of water, artificial structures, and land shadows; (d) the decision tree reclassification module is subdivided into the pre-classified and misclassified terrestrial area into at least one of green algae, soil, and vegetation, and (e) the decision tree reclassification module comprises Finally, by combining water and green algae, determining the water area and integrating the rest into the terrestrial area, the decision tree that classifies water and terrestrial from the hyperspectral image is expanded and subdivided. It has the effect of supplementing the problem of land misclassification.