Hyperspectral Image Compression Coding Method Based on Multivariate Vector Quantization

Abstract

The invention relates to a hyperspectral image compression encoding method, in particular to a multivariate vector quantization hyperspectral image compression encoding method. Multivariate vector quantization hyperspectral image compression coding method, including reading hyperspectral image data; constructing a compression dictionary: according to two multivariate vector quantization models and dictionary atom selection strategies, multiple regression is used to calculate the dictionary used for each pixel compression Atoms and their coefficients are compressed and encoded. The present invention rebuilds the vector quantization model and proposes two multivariate vector quantization models. The omitted error information is less than that in the traditional vector quantization method, and in the newly proposed multivariate vector quantization model, the selected dictionary atoms The coefficient of is also not limited, and its value is calculated based on the spectrum itself and the dictionary composition, thereby ensuring the quality of the reconstructed image and reducing image distortion caused by compression coding.

Description

Technical field [0001] The invention relates to a hyperspectral image compression coding method, specifically a multivariate vector quantization hyperspectral image compression coding method. Background technique [0002] Hyperspectral image compression coding technology, as an important research direction of hyperspectral image processing, has long attracted wide attention from experts, scholars and engineers from various countries, and has been widely used in agriculture, mineral exploration, military defense and other fields. Hyperspectral images can provide a large amount of detailed information about features by using imaging and spectroscopy techniques. It can reflect the characteristics of ground objects in hundreds of thousands of electromagnetic spectrum bands. However, with the continuous development of hyperspectral technology, people's requirements for hyperspectral images have gradually increased, which has resulted in the continuous expansion of space, spectral res...

AI Technical Summary

Problems solved by technology

In the traditional vector quantization method, because the error is directly omitted, the compression will cause the degradation of the image quality that cannot be recovered, which will have a great impact on the subsequent image processing.

Images