SAR (Synthetic Aperture Radar) image compression method based on super prior architecture

Abstract

The invention discloses an SAR image compression method based on a super prior architecture, and the method specifically comprises the steps: constructing a self-codec convolutional network which comprises an encoder, an arithmetic codec and a decoder; entropy modeling of the hyper-priori network: the entropy modeling of the hyper-priori network is divided into four parts, namely a hyper-parameter encoder, a hyper-parameter arithmetic codec, a hyper-parameter decoder and a probability model; meanwhile, the model predicts a mean value and a variance by using a single Gaussian mixture model to simulate signal distribution so as to realize entropy modeling; distortion optimization is carried out, the model is subjected to back propagation according to a loss function, training and optimization are carried out continuously, and a better compression effect is obtained. According to the SAR image compression method, entropy modeling is carried out on variables in a potential space, adjustment is carried out according to different compression objects to further improve the compression rate, the network structure is designed and improved, the feature extraction and generalization ability of the network is improved, and the compression rate and the compression quality of the SAR image are considered.

Description

technical field [0001] The invention belongs to the technical field of image compression, and in particular relates to a SAR image compression method based on a super prior framework. Background technique [0002] Synthetic aperture radar (SAR) images are increasingly important in various remote sensing applications. However, as the ability to transmit image data has rapidly increased, the storage speed of ground stations has not achieved a commensurate growth rate. Most image compression algorithms depend on accurate signal representation, and SAR image compression algorithms based on discrete wavelet transform (DWT) have been extensively studied. However, because SAR images contain rich textures, DWT and the traditional compression algorithm JPEG, JPEG2000 also cannot obtain effective representation. Therefore, it is of great significance to develop a compression algorithm that can obtain a higher compression ratio while maintaining the SAR image quality. [0003] In re...

AI Technical Summary

Problems solved by technology

However, since SAR images contain rich textures, DWT and the traditional compression algorithm JPEG, JPEG2000 also cannot obtain an effective representation

Images