SAR (Synthetic Aperture Radar) image segmentation method based on multi-scale feature fusion

Abstract

The invention discloses an SAR (Synthetic Aperture Radar) image segmentation method based on multi-scale feature fusion, comprising the steps of: firstly, extracting the textural features of an image by utilizing fast discrete curvelet transform (FDCT), extracting the statistic features of the image by utilizing stationary wavelet transform (SWT), then, fusing the two type of multi-scale features into high-dimensional feature vectors, and finally, segmenting by adopting a fuzzy C-mean clustering method. At the same time of eliminating broken pieces in a homogenizing area, the boundary is more precise and smoother. The method is favorable for effectively preventing the extraction of textural information from the interference of speckle noises. By fusing SWT coefficient statistical features, the textural information and the grey statistical information of the SAR image are effectively utilized and a favorable segmentation effect can be achieved.

Description

technical field [0001] The invention relates to a SAR image segmentation method, which is a SAR image segmentation method based on multi-scale feature fusion. Background technique [0002] Synthetic Aperture Radar (SAR) is an imaging system that works in the microwave band and has the ability to acquire all-weather, all-time, multi-view, and multi-resolution data. Therefore, it has immeasurable application value in the reconnaissance and surveillance of national defense, military affairs, environment, disasters, etc. SAR images are also widely concerned by researchers because of their high spatial resolution and rich detail information. Segmentation of SAR images is an important aspect of SAR applications. But because SAR is a coherent imaging system, SAR images are seriously affected by its inherent coherent speckle noise, which reduces the quality of SAR images and makes segmentation more difficult. Therefore, the research on SAR image segmentation has important theoreti...

AI Technical Summary

Problems solved by technology

[0003] There are more or less shortcomings in the existing SAR image segmentation algorithms: the threshold-based segmentation method is relatively simple, but it is most susceptible to the influence of coherent speckle noise. Since this method does not consider the context information in the SAR image, there is no The texture in the SAR image is analyzed, so the segmentation effect is poor; the segmentation method based on edge detection is also susceptible to the interference of coherent speckle noise, and there is a contradiction between missed detection and false detection, it is difficult to achieve satisfactory results ; SAR image segmentation based on texture feature clustering, because it involves the extraction of SAR texture features, can improve the effect of gray threshold segmentation to a certain extent, but there are also problems such as feature selection, feature combination and clustering algorithm selection ; SAR image segmentation based on the Markov random field model involves more knowledge of probability and statistics and the model is more complex

[0004] To sum up, the existing SAR image segmentation methods do not make full use of the rich texture information and gray information contained in the SAR image itself, the feature extraction of the SAR image is not comprehensive, and the segmentation effect is not good, so that it cannot give an efficient The features used for SAR image segmentation

Images