Fine Extraction Method of Remote Sensing Image Information Based on Surface Elements

Abstract

The invention discloses a method for finely extracting remote sensing image information based on surface elements. The method includes: selecting a high-resolution remote sensing image for remote sensing image segmentation, and obtaining several segmentation units after segmentation, and each segmentation unit is called a remote sensing image object; count the spectral histogram inside each remote sensing image object, and the spectral histogram is used as the first classification feature of the remote sensing image object; based on the internal spectral histogram of the remote sensing image object, perform the supervised classification of the curve matching algorithm on the remote sensing image object; according to For the preliminary classification results of remote sensing images, record the category numbers of the preliminary classification results for several adjacent remote sensing image objects in four directions for each remote sensing image object, and extract the spatial correlation between remote sensing image objects, and the spatial correlation is taken as The second classification feature of remote sensing image objects; combine the spectral histogram and spatial correlation, and use the curve matching algorithm to supervise the classification of remote sensing image objects again.

Description

technical field [0001] The invention relates to the technical field of satellite remote sensing image extraction, in particular to a method for finely extracting remote sensing image information based on surface elements. Background technique [0002] The popularization of high-resolution remote sensing images provides a basis for fine-grained monitoring of surface elements. How to effectively extract surface element information from high-resolution remote sensing images is the current research focus in the field of remote sensing. Object-oriented classification of remote sensing images (object -based image classification) is a method for information extraction of high-resolution remote sensing images. This method first divides remote sensing images into remote sensing image objects according to clustering rules. Each remote sensing image object has high internal homogeneity, and adjacent remote sensing images The heterogeneity among image objects is high, and the smallest u...

AI Technical Summary

Problems solved by technology

[0003] Aiming at the above-mentioned technical problems in related technologies, the present invention proposes a method for finely extracting remote sensing image information based on surface elements, which can solve the technical problem that the extraction accuracy of surface elements by traditional object-oriented classification features is not high

Images