SAR image change detection method based on image cutting

Abstract

The invention discloses a SAR image change detection method based on image cutting. By using the prior art, detection precision is low. Through using the method of the invention, the above problem can be solved. The method comprises the following steps: using an image cutting method to carry out block pretreatment on the SAR image and the changed SAR image; then, using the SAR image and the changed SAR image after the blocking to construct an adjacent map before change and the changed adjacent map respectively; connecting corresponding peaks of the two adjacent maps, and fusing the two adjacent maps so as to obtain a two-dimensional image of the SAR image and the changed SAR image; finally, using a NJW spectral clustering method to carry out cutting on the two-dimensional image so as to obtain labels of changed image blocks; extracting a kind of data blocks which indicate a change area in the changed image so as to obtain a final change detection result image. In the prior art, constructing a difference image and carrying out dimension reduction on the image can cause information loss and cumulative errors. By using the method of the invention, the above problem can be avoided and detection precision of the change detection can be increased.

Description

technical field [0001] The invention relates to the technical field of image processing, in particular to the processing of SAR images, which can be mainly applied to the monitoring of ecology and environment and the evaluation and prevention of natural disasters. Background technique [0002] Because synthetic aperture radar SAR is not affected by atmospheric conditions and cloud cover and other related conditions, multi-temporal SAR image technology is playing an increasingly important role in people's production and life. The SAR image change detection technology is playing a vital role in related applications, such as ecological and environmental monitoring and natural disaster evaluation and prevention, are playing an increasingly important role. Image change detection is to study the changes between two or more images of the same scene at different time periods. The acquisition of SAR images of the same scene at different time periods makes it possible to provide real-...

AI Technical Summary

Problems solved by technology

[0004] The defects of the current change detection method mainly include: for the difference graph classification method, the separability of the difference graph directly affects the change detection results, and the construction of a high-separability difference graph is still a difficult problem. The loss of image information is inevitable in the process of classification; for the post-classification comparison method, the effect of each phase classification plays a crucial role in the subsequent classification results, and image classification itself is a relatively difficult task. The problem of classification accuracy is difficult to guarantee, and there is a problem of error accumulation in the process of comparison

Images