Remote sensing image change detection method and device based on DT-CWT (dual-tree complex wavelet transform) and MRF (Markov random field)

Abstract

The invention discloses a remote sensing image change detection method and device based on DT-CWT (dual-tree complex wavelet transform) and MRF (Markov random field). According to the method and device of the invention, the multi-directional expression and multi-scale analysis of image information can be facilitated; correlations between pixels are fully utilized; the retention of high-frequency information and the removal of noises can be effectively balanced; edge detection is smoother; change detection results have good regional consistency; a false detection rate can be greatly reduced, and the influence of registration errors can be removed; and the accuracy of remote sensing image change detection can be greatly improved.

Description

technical field [0001] The invention relates to image processing technology, in particular to a remote sensing image change detection method and device combined with DT-CWT and MRF. Background technique [0002] With the continuous improvement of remote sensing data acquisition methods and the continuous shortening of update cycles, change detection technology, as an important application in remote sensing image processing and analysis, has been widely used in land use and land cover change monitoring, forest or vegetation change detection, and disaster relief. Governance, town and road changes, geographic database updates, and many other areas. Since the unsupervised change detection method can obtain the final change detection result without any additional prior knowledge except for the use of the two-temporal original image, the method is more efficient, so it has become a research hotspot. Usually unsupervised change detection is change detection based on pixel differen...

AI Technical Summary

Problems solved by technology

The threshold method is to distinguish the change and non-change pixels in the difference image by selecting an appropriate threshold, so as to obtain the final change area. Incomplete segmentation, while the detection results are heavily dependent on the chosen threshold

The transformation method is mainly a multi-scale transformation method, such as wavelet transformation. This type of method can perform multi-scale analysis and can better overcome the influence of factors such as sensor noise and registration error, but it does not consider the correlation between pixels. Rough edges, many false detection pixels

The image modeling method is to simulate the distribution information related to the spatial context by establishing a model. For example, the Markov Random Field (MRF) fully considers the interaction between pixels in the neighborhood and overcomes the problem of isolated pixels. , but the expression of mutation and detail information is weak, and the detection of false changes is more

In order to solve the problems existing in the above methods, remote sensing image change detection methods combining wavelet transform and MRF have been proposed. False changes caused by factors such as single pixel isolation, noise, and registration errors, but because high-frequency information is directly discarded in the fusion process, and only low-frequency information is used for experiments, it will inevitably cause information loss. In addition, there is a direction selection in wavelet transform. Sexuality and translation sensitivity, so that change detection details are still underexpressed

Images