| Aiming at the problems of storage massive hyperspectral remote sensing image and lower data compression,a hyperspectral remote sensing image compression algorithm based on adaptive predictor is proposed from the perspective of three-dimensional data.Firstly,the correlation between each band is calculated,and the reference band index table is established according to the optimal band correlation.The inter spectral prediction is carried out according to the optimal reference band index table,and the residual image after inter spectral prediction is obtained.Secondly,according to the characteristics of the image itself,three-dimensional space predictor or spectral space joint predictor is selected to predict the residual image data after inter spectral prediction of the optimal reference band.The selection of three-dimensional space predictor and spectral space joint predictor is by the statistics of the reference optimal reference band index table.When the correlation between each band is larger,indicated the data there is a larger correlation in the three-dimensional space,so choose the three-dimensional space predictor,otherwise choose the spectrum space joint predictor.These two optimized predictors take into account different types of data compression.Finally,the information entropy of the generated residual image has been greatly reduced compared with the original image,the arithmetic coding is used for coding the residual image and obtain the compressed data.The experimental results show that the compression effect of the two predictors in the optimized adaptive predictor algorithm is a little different,the compression effect obtained by optimizing the adaptive predictor compression algorithm is significantly higher than that of the common similar compression methods.It can be concluded that the algorithm in this paper is an effective lossless compression algorithm for hyperspectral remote sensing image.There are 23 pictures,11 tables and 67 references in this paper. |
PDF Full Text Request