A Study On Hyperspectral Image Compression Algorithm Based On Fractal Theory

Hyperspectral image is the sequence of image acquired by spectrometer in different wavelength, which increase the information of spectrum based on two dimensional remote-sensing image. The data of high resolution and continuously spectrum images is very huge, and bring a lot of trouble to it's storage and transmission. So it is important to search an effective compression of hyperspectral image on the existing condition.Compared to the traditional remote-sensing data source, hyperspectral data is redundancy in space and spectrum. So in order to get a good effect of images compress in hyperspectral image compression, we must consider to removal the dependence of both space and spectrum..This paper based on fractal theory, use the high compressive ratio character of fractal compression, by the character of hyperspectral image, we applied fractal theory to hyperspectral image and designed a three dimensional fractal compression coding algorithm, as well as made an experiment on hyperspectral image data through IDL software, which removed the space correlation of image and removed the correlation of wave-range image as well. Algorithm showed that the use of three-dimensional fractal coding, be very high compression rate, and maintain a high spectral curve for object recognition capability.