SAR image super-resolution method based on neural network

Abstract

The invention provides an SAR image super-resolution method based on a neural network, and belongs to the field of image super-resolution. Aiming at a set neural network structure, a loss function isconstructed by utilizing a mean square error between a predicted value and a real value, a model mapping problem is converted into an optimization problem of the loss function, weights and bias valuesof all layers of the neural network are determined, and finally low resolution-high-resolution mapping relation is obtained; and the SAR image to be processed is input into a network to obtain a super-resolution result. Compared with the prior art, the method has the advantages that feature propagation can be effectively enhanced, the training speed can be increased, and a high-quality SAR imagesuper-resolution result can be obtained.

Description

technical field [0001] The invention belongs to the field of image super-resolution, in particular to a neural network-based SAR image super-resolution method. Background technique [0002] Synthetic aperture radar (SAR) is an imaging radar with high mapping resolution. Compared with traditional optical remote sensing, SAR has the characteristics of all-day and all-weather work, and can penetrate some obstacles, so it is widely used in military and civilian fields. As far as SAR images are concerned, the image resolution that can be obtained determines the breadth and depth of its application, so the research on SAR image super-resolution is of great significance. [0003] There are many existing image super-resolution methods, which can be roughly divided into three categories: interpolation-based methods, reconstruction-based methods, and learning-based methods. Among them, the learning-based method uses the data set corresponding to the low-resolution-high-resolution (L...

AI Technical Summary

Problems solved by technology

[0004] In the prior art, there is a method for image super-resolution reconstruction using a neural network, but the network of this method is too shallow to fully learn the nonlinear mapping relationship between LR images and HR images; and the one-way propagation of the network makes The initial features of the image are very weak in the later layers, which affects the quality of the training results; there is also a method of simultaneous supervision and weighted reconstruction of multiple intermediate results, which improves the training effect and performance, but due to inter-layer propagation Only the information of the previous layer is used, and there are still problems such as insufficient feature propagation

Images