Hyper-spectral image compression based on band selection and slant Haar type orthogonal transform

ABSTRACT

There is information redundancy in both spatial and spectral aspects of hyperspectral images. Considering a fixed proportion in sequential forward method may not find the optimal bands, we modify a band selection (BS) method by introducing a parameter to adjust the proportion of standard deviation to correlation, which may select key bands quickly and accurately. In addition, slant Haar type orthogonal transforms (SHTOT) have slant base vectors suitable to express the image brightness with gradual change. However, SHTOT have attracted little attention of scholars’. This paper introduces SHTOT with fast algorithm and varied parameters to the further compression of band images from a space point of view. Comparative experiments were performed with other BS strategies and state-of-the-art orthogonal transforms, such as DCT, DWT, Walsh, slant transform, Haar type orthogonal transforms. Final experimental results achieved on the commonly used data sets validate that the proposed approach has a faster speed, high compression ratio and good image quality. Additionally, it is more appropriate to choose different SHTOT for the specific application. The sparse SHTOT generated by certain parameter values are more suitable for the high requirement in image quality, but the compression ratio isn’t very high, while dense SHTOT produced by another parameter values are fitter for the opposite cases.

KEYWORDS:

• Hyper-spectral image
• band selection
• image compression
• slant haar type orthogonal transform

Acknowledgements

Thank the anonymous reviewers for the modification of this paper. This investigation is supported by the National Natural Science Foundation of China under Grant No 61202051.

Disclosure statement

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data Availability statement

Data will be made available on request.

CRediT authorship contribution statement

Xiuqiao Xiang: Methodology, Writing – original draft, Conceptualization, Software, Review & editing. Yuhong Jiang: Software, Formal analysis, Supervision. Baochang Shi: Formal analysis, Conceptualization.

Additional information

Funding

The work was supported by the National Natural Science Foundation of China [61202051].