Segmentation method for synthetic aperture radar images in consideration of multi-scale Markov field

Abstract

The invention discloses a segmentation method for SAR images in consideration of multi-feature multi-scale MRF field, which is used for the treatment of remote sensing images, the method comprises the steps of: (1) extracting texture information from original images by utilizing a gray level co-occurrence matrix, and performing multi-scale segmentation on the resultant texture characteristic images; (2) detecting edge information of the original images by means of wavelet transformation in order to obtain edge characteristic images; (3) incorporating texture multi-scale segmentation result with edge characteristics by a linear weighting method, namely, corresponding pixel points of the two images are directly subjected to weighted stack to obtain final segmentation result. The invention reduces erroneous segmentation rate of SAR images and outstandingly remains detailed information of the images like edge information while eliminating speckle noise.

Description

technical field [0001] The invention relates to a synthetic aperture radar image segmentation method, in particular to a SAR image segmentation method considering multi-features and multi-scale Markov fields, and belongs to the field of remote sensing image processing. Background technique [0002] Synthetic Aperture Radar (SAR) has the characteristics of all-weather, large area, high resolution, and ability to penetrate the surface clouds and fog, so that it can obtain a large amount of information in a short time, which also makes the application of SAR images become a research field in the field of remote sensing. hotspots. With the development of spaceborne SAR systems at home and abroad, a large amount of spaceborne SAR image data has been obtained. Compared with the rapid growth trend of SAR data sources, the research on SAR image processing and analysis technology, which is the basis of SAR data application, is relatively lagging behind. SAR image segmentation is an ...

AI Technical Summary

Problems solved by technology

The method based on the MRF field has high segmentation accuracy for images with few categories, and makes full use of statistical and texture characteristics, but the MAP estimation is computationally intensive and time-consuming.

Based on the multi-scale MRF field method, a new multi-scale random field model is used to replace the MRF model, and the MAP estimate is replaced by a continuous MAP (SMAP) estimate derived from a new estimation standard, which solves the calculation problem of the MRF model , but the speckle suppression is not sufficient, and the edge detail information of the segmentation result is easily lost

Images