Contourlet Transform Based Listless Block Cube Tree Coding for Hyperspectral Images

Abstract

The performance of compression algorithms at low bit rates is a critical benchmark, particularly for hyperspectral imaging where the fidelity of reconstruction is paramount. Although wavelet-based approaches are prevalent in the literature, they frequently present a trilemma of undesirable trade-offs, suffering from insufficient coding efficiency, exorbitant memory requirements, or high computational overhead. The proposed compression algorithm employees advance wavelet transform to leverage both spectral and spatial redundancies found in HS data cubes. Present study explores the utilization of block tree coding algorithm and contourlet transform to compress HS images. The primary goal is to enhance the coding efficiency while minimizing storage and transmission requirements. The proposed compression algorithm is evaluated on four benchmarks against eight state of art other compression algorithms on three performance metrics named coding efficiency, coding memory and coding complexity. In addition, it has low encoding/decoding time than other compression algorithm. From the simulation result, it has been clear that proposed compression algorithm has 2% to 4% increase in coding efficiency compared to the other state of art compression algorithms.