Mathematical Transform Based Set Partition Hyperspectral Image Compression Algorithms : A Comparative Review

Abstract

Hyperspectral imaging has been extensively investigated as an emerging, promising technique for multiple domains. Originally derived from remote sensing, this technology now incorporates both machine vision and point spectroscopy within its scope of use. The end result is the transfer of hyperspectral images in large quantities from sensors to analysis centers and then, finally, to data centers using this method. Compression algorithms are used to ensure that these large-sized hyperspectral images are stored in a manner that is both efficient and secure. Compression algorithms using mathematical transforms works for lossy and lossless environment, which make it proper choice for the image sensors. Set partitioned based compression algorithms is a sub category of mathematical transform based algorithms which utilized the property of wavelet transform set structure to achieve the compression of image. This survey focuses on different hyperspectral image compression algorithms which utilized mathematical transform to achieve the compression. Additionally, we evaluate the most effective algorithms in terms of their coding efficiency and provide a detailed analysis of the primary factors that influence compression performance. Also, coding memory and embeddedness is also covered in the comparative analysis.