Ultra-pixel fast generation method for polarized SAR image

Abstract

The invention provides an ultra-pixel fast generation method for a polarized SAR image. The method includes the following steps of 1. initialization: setting the elements in an unstable point set as all the pixel points in the polarized SAR image, naming the elements in the unstable point set as unstable points, and setting the unstable point labels as the label values of the unstable point corresponding pixels; 2. updating the unstable point labels; 3. updating the image area center, the unstable point set, and the number of iterations; and 4. performing the postprocessing based on the non-similarity measurement. An image area obtained by the above-mentioned four steps is an ultra-pixel. Therefore, the ultra-pixel having the small size and being different from the image area of the surrounding neighborhood can be reserved.

Description

Technical field [0001] The invention relates to the technical field of image segmentation, and more specifically, to a method for rapidly generating superpixels of a polarized SAR (Synthetic Aperture Radar, synthetic aperture radar) image. Background technique [0002] Polarized SAR has four channels: HH (horizontal transmission and horizontal reception) polarization mode, HV (horizontal transmission and vertical reception) polarization mode, VH (vertical transmission and horizontal reception) polarization mode, VV (vertical transmission and vertical reception) polarization Way, so it can provide more information. Polarization SAR image superpixel generation is a method of clustering SAR image pixels to obtain a series of similar image regions with specific meanings. Polarization SAR image superpixel generation is essentially an image segmentation method. With the rise of polarized SAR images, polarization SAR image processing based on superpixels has also received more and more...

AI Technical Summary

Problems solved by technology

Since the polarimetric SAR image is affected by coherent speckle noise, when the SLIC method is directly used, the edge fit of the generated superpixels is low

Therefore, scholars have proposed some improved SLIC algorithms for polarimetric SAR image superpixel generation (see the literature Superpixel segmentation for polarimetric SAR imagery using local iterative clustering.IEEE Geosci.Remote Sens.Lett.2015,12,13-17), The modified Wishart distance is used to replace the Euclidean distance of the CIELAB color space in the SLIC method. This method generates superpixels with high edge fit, but the calculation efficiency is too low.

Images