Lossy hyperspectral image compression based on intra-band prediction and inter-band fractal encoding. (August 2016)

• Record Type: Journal Article
• Authors: Zhao, Dongyu; Zhu, Shiping; Wang, Fengchao
• Is Part Of: Computers & electrical engineering. Volume 54(2016)
• Page Start: 494
• Page End: 505
• Publication Date: 2016-08
• Subjects: Hyperspectral image -- Lossy compression -- Fractal encoding -- Prediction; Computer engineering -- Periodicals; Electrical engineering -- Periodicals; Electrical engineering -- Data processing -- Periodicals; Ordinateurs -- Conception et construction -- Périodiques; Électrotechnique -- Périodiques; Électrotechnique -- Informatique -- Périodiques; Computer engineering; Electrical engineering; Electrical engineering -- Data processing; Periodicals; Electronic journals; 621.302854
• DOI: 10.1016/j.compeleceng.2016.03.012
• Languages: English
• ISSNs: 0045-7906

Abstract

Hightlights: A lossy hyperspectral image compression scheme based on intra-band prediction and inter-band fractal is proposed. Intra-band prediction exploiting spatial correlation is applied to I-bands. Inter-band fractal encoding with a local search algorithm is applied to P-bands. SVM classification accuracy for the reconstructed image is slightly higher than that for the original one. Abstract: Recently, hyperspectral image compression has become an urgent issue for remote sensing applications. A lossy hyperspectral image compression scheme based on intra-band prediction and inter-band fractal encoding is put forward in this paper. The hyperspectral image is firstly partitioned into several groups of bands (GOBs). Intra-band prediction is applied to the first band in each GOB, exploiting spatial correlation, while inter-band fractal encoding with a local search algorithm is applied to the other bands in each GOB, making use of the local similarity between two adjacent bands. The fractal parameters are signed Exp-Golomb entropy encoded. To improve the decoded quality, the prediction error and fractal residual are further transformed, quantized, and entropy encoded. Experimental results illustrate that the proposed scheme can obtain a better compression performance with low complexity compared with other well-known methods. In addition, the effect of compression on SVM (Support Vector Machine) classification is presented. Graphical abstract: