Multispectral Image Change Detection Method Based on Probabilistic Segmentation and Gaussian Mixture Clustering

Abstract

The invention belongs to the technical field of identification methods or devices using electronic equipment, and discloses a multi-spectral image change detection method based on probability segmentation and Gaussian mixture clustering; first, two original multi-spectral images in the same region but at different times are input, Using CVA and SAM to construct a hybrid difference image HDS; secondly, using the statistical region merging algorithm to perform multi-scale segmentation on the difference image to map the difference image to the superpixel space; finally, using the K-means algorithm to initialize the Gaussian mixture model to overcome its easy convergence in The shortcomings of local optimal solutions, fitting the probability and statistics distribution of the superpixel feature space, and using the Bayesian discriminant rule based on the minimum error rate to obtain the change detection results. The present invention makes better use of the amplitude change information and angle change information of the spectral vector, and can obtain the local structural features of the image by using the superpixel segmentation, thereby effectively improving the detection accuracy of the change area in the SAR image.

Description

technical field [0001] The invention belongs to the technical field of identification methods or devices using electronic equipment, and in particular relates to a multispectral image change detection method based on probability segmentation and Gaussian mixture clustering. Background technique [0002] At present, the existing technologies commonly used in the industry are as follows: With the rapid development of remote sensing technology in recent years, the amount of remote sensing data is increasing day by day, and is widely used in fields such as environmental monitoring, atmospheric analysis and urban planning. Among them, the change detection of multispectral remote sensing images is applied to military and civilian fields, mainly involving the positioning of natural disaster areas such as floods, fires and earthquakes, the analysis of urban expansion, and the evaluation of strike effects in military applications. Therefore, research on multispectral remote sensing im...

AI Technical Summary

Problems solved by technology

However, this method is to segment the original dual-temporal remote sensing image separately, and then calculate the distance measure value between the superpixels. Due to the difference in the shape and structure of the target between the dual-temporal images, it is necessary to calculate the similar distance before the boundary configuration. accurate, which will destroy the geometric information of the target

Images