SAR image super-resolution reconstruction method based on Euclidean distance regularization

Abstract

The invention discloses an SAR image super-resolution reconstruction method based on Euclidean distance regularization. First of all, multiple frames of an SAR image, with closer time phases are input and registered, and the multiple frames of the SAR image are placed in a grid and are processed to an SAR observation image with uniformly distributed pixels; then on the basis that an SAR image degradation model is established, a high-resolution SAR image reconstruction formula based on Euclidean distance double-side fully differential regularization is constructed; next, the observation image is initialized, and afterwards, according to the reconstruction formula, iteration estimation is performed on the high-resolution SAR image by use of a calculable optimization mode; and finally, outputting an estimated image as an SAR image super-resolution reconstruction result. According to the invention, a double-side fully differential mode of a Manhattan distance in SAR image reconstruction is changed to a Euclidean distance form, and prior knowledge is more reasonably described; and a substitution formula of the reconstruction formula is constructed for the purpose of solving the problem of differential solving in an optimization iteration process, and thus on an optimization solving calculable basis, the accuracy of the SAR image super-resolution reconstruction result is improved.

Description

technical field [0001] The invention relates to the technical field of remote sensing image processing, in particular to a SAR image super-resolution reconstruction method based on Euclidean distance regularization. Background technique [0002] Synthetic Aperture Radar (SAR) is a microwave imaging radar that uses Doppler effect and pulse compression technology to achieve high resolution. It has all-day and all-weather remote sensing imaging capabilities. With the continuous improvement of scientific research and application requirements, the spatial resolution of SAR images needs to be further improved. The following methods can be used: 1. Directly improve the antenna length, synthetic aperture length and extended signal bandwidth of the radar system, but this method is limited by hardware Technical bottleneck and development cycle, and the cost is high; 2. Using the original echo data (non-image data) of the SAR imaging system to obtain high-resolution SAR images through...

AI Technical Summary

Problems solved by technology

[0004] However, the problem lies in: 1) the grayscale measure of the Manhattan distance cannot accurately represent the gradient modulus of the image; 2) the geometric measure of the Manhattan distance will make the correlation between pixels no longer depend on the length of the straight line segment between pixels

Obviously, the above problems will have a certain impact on the description of prior knowledge, thereby reducing the accuracy of SAR image reconstruction

Images