Remote sensing classification method based on relative entropy

A remote sensing classification method based on relative entropy, comprising the following steps:

• step 1, determining sample points corresponding to n types of ground objects in a study area, and determining series remote sensing parameters, wherein the series remote sensing parameters are configured to identify characteristics of the n types of ground objects, n is a positive integer greater than 1, the series remote sensing parameters comprise series data formed by hyperspectral data reflecting spectral curve differences among the n types of ground objects, and the n types of ground objects comprise an evergreen forest, a deciduous forest, a closed shrubland, an open shrubland, a woody savanna, a savanna, a grassland, an agricultural land, a water, and a bare area;
• step 2, extracting, based on the sample points, remote sensing parameter values corresponding to each type of the n types of ground objects from the series remote sensing parameters, obtaining a standard time series plot of each type of the n types of ground objects according to the extracted remote sensing parameter values, and taking the standard time series plots of the n types of ground objects as a first distribution;
• step 3, taking remote sensing parameter values corresponding to to-be-classified pixels of the study area as a second distribution, determining, based on the second distribution and the first distribution, KL values of the to-be-classified pixels corresponding to each type of the n types of ground objects by using a KL-divergence formula, and obtaining n KL layers of the study area according to the KL values of the to-be-classified pixels corresponding to the n types of ground objects;
• and
• step 4, comparing n KL values of each of the to-be-classified pixels, and classifying the to-be-classified pixel to be a type of ground objects with a minimum KL value of the n KL values, and thereby obtaining classification results of the study area.