Remote sensing image variation detection method based on region similarity

Abstract

The invention discloses a remote sensing image variation detection method based on region similarity. The detection method comprises the steps of: selecting tow-sence remote sensing images with different time phases in a monitored region, wherein the image data of the remote sensing images include multispectral data and panchromatic wave band data; carrying out the image preprocessing on the selected remote sensing images; combining all wave bands of the two remote sensing images in two periods after preprocessing into a one-scene image, dividing the one-scene image to establish the one-to-one mapping relationship of targets of the remote sensing images in two periods; establishing the feature set description of each target according to division results; calculating the region similarity between the targets according to the spectrum features of the targets, and carrying out algebraic calculus according to the textual features and NDVI features of the targets; building up a variable region detection criterion by setting a region similarity threshold value and combining the results of the algebraic calculus, and extracting the variable regions of the remote sensing images in two periods. The detection method can effectively eliminate the phenomenon of salt and pepper and improve the extract precision of the remote sensing image variable regions.

Description

technical field [0001] The invention relates to the technical field of remote sensing image processing, in particular to a remote sensing image change detection method based on regional similarity. Background technique [0002] Remote sensing information change detection is an important research topic in remote sensing applications. Using satellite remote sensing images in the same area and different time phases for change detection is an important tool for dynamic monitoring of land and other resources, resource change investigations, environmental monitoring, and basic geographic database updates. The key technology in the application of earth observation has a wide range of application fields. Judging from the existing change detection methods, for medium and low resolution remote sensing images, the method mainly stays in the pixel-level data-guided method, and lacks the knowledge-guided feature-level change detection method. [0003] In recent years, the development of...

AI Technical Summary

Problems solved by technology

However, any classification errors in the single-temporal imagery will show up in the final change detection map, which is called the "salt and pepper" phenomenon, so the main disadvantage of this method is that the accuracy of change detection depends on the accuracy of two independent classification results. Accuracy, there is a phenomenon of accumulation of classification errors

Images