High Spectral Image Compression Method

As an emerging discipline combines electromagnetism, optics and signal processing, hyperspectral remote sensing imagery has been widely used in military, geology, medicine, agriculture, marine biology, atmospheric and environmental monitoring, etc. However, hyperspectral remote sensing imaging technology severely restricted the application and development due to its massive data which will bring great difficulty in date storage and transmission. Therefore an efficient method for hyperspectral image compression is an urgent need to resolve the issue.In this paper the compression algorithm of hyperspectral image is studied. Firstly, the data storage format of hyperspectral image is studied, and spatial and spectral correlation of hyperspectral image is also analysed. Secondly, the compression algorithm of hyperspectral image based on linear prediction is carried out a detailed analysis, and some encoding methods of hyperspectral data are introduced.Some preprocessing methods are used to improve the performance of hyperspectral image compression before hyperspectral image compression, a novel hyperspectral compression algorithm based on hybrid encoding which combines with the methods of the band optimized grouping and the wavelet transform is proposed,and the experimental results show that the proposed approach achieves a good performance in reconstructed image quality. Then, the compression algorithm of hyperspectral image based on compressive sensing theory is studied. The compressed sensing theory and the key elements of the theory are introduced, and the experimental results of different reconstruction algorithms based on compressed sensing theory are analysed. Finally, the compression algorithm of hyperspectral image based on coded aperture snapshot spectral imaging is researched.The principle of coded aperture snapshot spectral imaging is introduced, and the theoretical model of coded aperture snapshot spectral imaging technology is built.The experimental result of compression algorithm based on coded aperture snapshot spectral imaging technology is analysed.