Research On Lossless Compression Of Hyperspectral Image Of Land Remote Sensing Satellite

With the rapid development of remote sensing technology, the ability of exploring the land surface information has been extended into space. Hyperspectral images obtained by the Land Remote Sensing Satellite has been widely used in the surveying underground mineral resources, monitoring and assisting agricultural development, studying and forecasting a variety of natural disaster and environmental pollution, drawing a variety of geological features Figure. Extensive application of hyperspectral images promotes the rapid advancement of hyperspectral imager technology, requirements for quality of Land Remote Sensing Satellite hyperspectral image becomes higher and higher, its spatial resolution and spectral resolution of the image is growing, at the same time, which leads that hyperspectral data increase exponentially. Contradictions between the vast amounts of Land Remote Sensing Satellite hyperspectral image data and the limited onboard memory capacity and satellite transmission channel band width, and the rich feature of Land Remote Sensing Satellite hyperspectral image and the extensive application value of information, which makes efficient hyperspectral image lossless compression technology become a serious problem to be solved.To solve the transmission and storage problems resulting from massive hyperspectral remote sensing data, using the significant spectral correlation within the hyperspectral images, a lossless compression scheme based on hybrid prediction is proposed. Intra-band prediction is used only for the first image along the spectral line using a median predictor.And inter-band hybrid prediction which is the combination of a linear prediction and a context prediction is applied to other band images. Second order linear predictor is used to obtain a prediction reference value. The final prediction is obtained through context prediction model according to the prediction reference value. The residual image of hybrid prediction is coded by the Huffman coding. Experimental results on AVIRIS hyperspectral images show that the proposed algorithm achieves better lossless compression, the average compression ratio is 3.17, which is 0.05~0.48 higher than those from other lossless compression algorithms such as 3D-CALIC, LUT, C-DPCM, JPEG-LS. It is an effective lossless compression method for hyperspectral images.