Quick difference value vector quantitative compression coding method of ultra-spectrum signal

Abstract

The invention provides a quick difference value vector quantitative compression scheme of an ultra-spectrum signal, which comprises the steps of: in a Hadamard domain, constructing an initial code book of initial vector quantization, carrying out quick clustering, generating a final code book and an index of initial vector quantization; obtaining difference value image data from the restored image of the initial vector quantization and an original image to be used as an input vector of the difference value vector quantization, carrying out Hadamard transformation on the input vector; and constructing an initial code book of the difference value vector quantization, and carrying out quick clustering to generate a final code book and an index of the difference value vector quantization. The quick difference value vector quantitative compression scheme provided by the invention is combined with the Hadamard transformation property and the characteristic of reducing the quantization error by the difference value vector quantization; and compared with the traditional LBG algorithm, the quick difference value vector quantitative compression scheme has the advantages that under the condition of same compression ratio, the computing quantity can be greatly reduced, the image restoring quality can be improved, and the purpose of quick compression and coding is achieved.

Description

technical field [0001] The invention belongs to the field of hyperspectral remote sensing image processing, and in particular relates to a hyperspectral image data compression method based on difference vector quantization technology. Background technique [0002] With the wide application of remote sensing technology in the fields of earth resources management, environmental detection, and military detection, the data characteristics of multispectral image data in the past have only improved in the spatial resolution of the image, and its data characteristics have not satisfied people in production and scientific research. For example, the images of SPOT and Landset TM only contain 4 to 7 discrete bands, and people are interested in objects whose absorption characteristic width is 20nm to 40nm, so when the objects in the band When the target spectral characteristics are similar, the application of multispectral images is limited. At this time, more spectral bands (usually i...

AI Technical Summary

Problems solved by technology

[0007] Aiming at the existing hyperspectral image compression technology, the recovery quality of the image is not high and the amount of calculation is large when a high compression ratio is obtained, and a hyperspectral image fast compression coding method based on difference vector quantization is proposed. In the case of constant compression ratio, improve the restoration quality of hyperspectral images,

Images