Modis Satellite Remote Sensing Image Labeling Method Based on Local Spectrum Decomposition and Scoring

Abstract

The invention discloses a MODIS satellite remote sensing image labeling method based on local spectrum decomposition and scoring, including the steps of manually labeling training samples, sample feature extraction and classifier training steps, MODIS satellite remote sensing image probability prediction and area labeling steps; MODIS satellite remote sensing images are manually labeled, and then the collected images are subjected to feature processing based on spectral decomposition and a classifier is trained. Finally, in actual use, the trained classifier is used to classify and label the MODIS satellite remote sensing images to be predicted. Compared with the prior art, the method of the present invention has a novel feature extraction method and a high degree of automatic classification test.

Description

technical field [0001] The invention relates to an area labeling method based on spectrum decomposition and scoring in pattern recognition, and is particularly suitable for the ice sea cloud probability prediction and area labeling problems of MODIS satellite remote sensing images. Background technique [0002] Currently, 51 countries have participated in polar scientific expeditions focusing on Antarctic scientific expeditions, including most developed countries and major developing countries in the world. It is related to global changes and the future of mankind. The display and competition on the international stage has far-reaching significance and great influence in politics, science, economy, diplomacy, military and other aspects. With the development of science and technology, in recent years, countries have attached great importance to the use of the most advanced technology to provide guarantee for polar scientific expeditions and use technology to escort scientific...

AI Technical Summary

Problems solved by technology

[0003] In order to achieve the purpose of automatic probability prediction and regional labeling of MODIS satellite remote sensing images, it is necessary to identify the texture information that can effectively distinguish ice, sea and cloud regions in the image. The block spectrum energy of the regional image is concentrated and the optical density is low; the image block spectrum energy of the thick ice and thick cloud area is relatively concentrated and the optical density is high; the spectral energy of the thin ice and thin cloud area is scattered and the optical density distribution is wide; and there is ice (cloud) The approximate rank of the local area of ​​the channel is anisotropic, and the approximate rank calculated in different directions has a large variance

Images