Distributed hyperspectral image coding and decoding method based on three-dimensional wavelet transform

Abstract

The invention discloses a distributed hyperspectral image coding and decoding method based on three-dimensional wavelet transform. After spectrum wave filtering is carried out for a hyperspectral image at the coding end, information transformed by wavelet is divided into two parts of a high frequency part and a low frequency part. The low frequency part and the high frequency part are respectively as a keyframe and a W-Z frame to carry out independent set partitioning in hierarchical trees (SPIHT) coding. A Wyner-Ziv coder improves code-error-resistance property by adopting a Turbo coding mechanism. At the decoding end, low frequency information is used for predicting the high frequency part and side information is formed, and the combined decoding is carried out with the high frequency part. The high and low frequency parts which are carried out spectrum wave filtering are respectively encoded, and lower computation complexity of a coding end is guaranteed. Due to the fact that combined decoding of the low frequency part and the high frequency part is carried out at the decoding end, spectrum relevancy of the hyperspectral image is fully played. Due to the fact that turbo channel coding is adopted, robustness of coding code stream to channel noise is strengthened, and gradability and flexibility are provided.

Description

technical field [0001] The invention belongs to a hyperspectral remote sensing data compression method, in particular to a distributed hyperspectral image encoding and decoding method based on three-dimensional wavelet transform that can improve the compression efficiency of an existing encoding algorithm and has a wide application range. Background technique [0002] The rapid development of imaging spectroscopy has brought remote sensing technology into the stage of hyperspectral remote sensing. Space communication, especially deep space communication, has the characteristics of heterogeneous, time-varying, limited bandwidth, high bit error rate, limited resources and limited processing capacity of spaceborne equipment, etc., which brings challenges to the compression and transmission of hyperspectral remote sensing images. Huge difficulties: On the one hand, in practical applications, the compression process needs to be completed on the star, and the energy supply on the ...

AI Technical Summary

Problems solved by technology

However, this encoding method has a large amount of calculation and high complexity at the encoding end, and it puts forward higher requirements on the processing capacity, storage capacity and energy supply of the on-board encoder.

Therefore, the traditional hyperspectral image coding method still has the following shortcomings: first, the simple two-dimensional compression method cannot use inter-spectral information, and the compression efficiency is low, which puts a lot of pressure on the downlink bandwidth; second, the three-dimensional coding The coding complexity of the method, large amount of calculation, long processing time, and high requirements for the on-board power supply and processor

At present, distributed source coding, as a coding method suitable for cross-correlation sources, is attracting more and more attention in the industry, but so far there is no distributed high-efficiency coding for hyperspectral images based on 3D wavelet transform method

Images