Infrared remote sensing image change detection method based on multi-source target characteristic support

Abstract

The invention provides an infrared remote sensing image change detection method based on a multi-source target characteristic support. The method comprises the following steps of: S1, carrying out change detection on visible light remote sensing images so as to obtain an initial detection result; S2, converting change targets and invariable targets obtained by visible light change detection to infrared remote sensing images according to visible light and infrared remote sensing image registration; S3, extracting multi-source target characteristics; S4, respectively selecting a part of the change targets and invariable targets of the visible light as positive samples, selecting a part of the invariable targets of the visible light as negative samples and utilizing the selected samples as training samples; S5, training SVM to obtain a change detection model; and S6, carrying out change detection to obtain the infrared remote sensing image change detection result. The method provided by the invention can be used for detecting image changes caused by changes of ground features and detecting image changes caused by changes of the working states of the ground features, and is high in stability and wide in application range.

Description

technical field [0001] The present invention relates to a method for detecting changes in infrared remote sensing images, specifically a method for detecting changes in infrared remote sensing images based on multi-source target features and machine learning. The related technologies involve remote sensing image information processing, target feature extraction, and pattern recognition , machine learning and other technical fields. Background technique [0002] The purpose of remote sensing image change detection is to automatically detect changed ground objects by analyzing and calculating remote sensing images taken at different times in the same scene. important application value. At present, most remote sensing image change detection systems are processed for visible light remote sensing images, which can detect image changes caused by changes in ground objects. In recent years, with the continuous development of aerospace technology and sensor technology, people have ...

AI Technical Summary

Problems solved by technology

[0006] The beneficial effect of the present invention is that the present invention solves the automatic detection of changes in infrared remote sensing images through automatic detection of infrared remote sensing image changes based on multi-source (visible light, infrared) target features, combined with visible light remote sensing image change detection, feature extraction and machine learning technology The problem is that it can not only detect image changes caused by changes in ground features, but also detect image changes caused by changes in the working status of ground features, avoiding the inability to effectively detect changes caused by ground features using a single infrared remote sensing image. The image changes caused by it and the use of a single visible light remote sensing image cannot effectively detect the image changes caused by changes in the working status of ground objects

Images