Remote sensing image picture element decomposing method based on simulation of real scenes

Abstract

The invention relates to a remote sensing image picture element decomposing method based on the simulation of real scenes. The method comprises the following steps: 1. investigating a sample plot on the ground to obtain crown type and parameters of the sample plot; 2. simplifying the geometry of a target vegetation; 3. simulating the terrain of the sample plot in a three-dimensional modeling software; 4. simulating a Moso bamboo stand scene in the three-dimensional modeling software; 5. obtaining an analog abundance value; 6. calculating the spectral reflectance of end member; and 7. decomposing least square picture element with all constraints. The invention relates to the three-dimensional modeling of real scenes and the application thereof in remote sensing science. The method applies the a priori knowledge of investigation data of sample plot in end member extraction and introduces the three-dimensional analog model in two-dimensional linear spectral decomposition, thus solving the problem that real end member is hard to obtain through the existing picture element decomposition method. The method can be adopted to obtain end member with better quality and achieve better mixed picture element decomposition effect.

Description

【Technical field】 [0001] The invention relates to a three-dimensional modeling remote sensing image pixel decomposition method of a real scene. 【Background technique】 [0002] Generally, remote sensing images are an information carrier that describes the attributes of ground objects and targets in units of pixels. However, in the pixel scale, there are usually multiple surface object types or different states of the same surface object type that affect the information acquired by remote sensing images. Generally, the phenomenon that multiple ground object signals are mixed in a pixel is called "mixed pixel problem". Mixed pixels not only affect the accuracy of object recognition, but also the main obstacle to the quantitative development of remote sensing science. In order to solve a series of problems caused by mixed pixels, people have developed corresponding pixel decomposition methods for different problems. Zhao Yingshi (2003) summarized five types of mixed pixel dec...

AI Technical Summary

Problems solved by technology

[0007] The present invention proposes a remote sensing image pixel decomposition method based on simulating real scenes in order to solve the problem that the existing mixed pixel decomposition method is difficult to accurately obtain endmembers and is not suitable for the required pure pixel in the image.

Images