干涉光谱图像序列压缩的实验研究  

Experimental Study on Interference Spectral Image Series Compression

作　　者：杜述松[1] 袁艳[2] 相里斌[3] 陶然[1] 张文喜[1] 

机构地区：[1]中国科学院西安光学精密机械研究所 [2]北京航空航天大学,北京100083 [3]中国科学院高新技术产业局,北京100864

出　　处：《光子学报》2007年第9期1672-1676,共5页Acta Photonica Sinica

基　　金：国家自然科学基金(60532070)资助

摘　　要：基于均匀值量化压缩算法和AR模型算法,对干涉型超光谱成像仪直接产生的实际干涉图像序列进行了近无损压缩仿真实验.AR模型算法对残差采用Lloyd-Max量化器进行量化,对量化输出结果用Huff man编码器进行编码.分析了AR模型阶数对压缩性能的影响.结果指出,均匀值量化和AR模型算法均能满足干涉光谱图像序列压缩的要求.Based on the uniform quantization and the AR model compression algorithm, simulated experiment is carried out for compressing the interferometric hyperspectral image series which were directly produced by the inlerferometric hyper spectro-imager. In AR model the Lloyd-Max quantizer is use to quantize the predicted residuals. The outpul of the quantizer is encoded by the Huffman encoder. The compression quality is analyzed. The uniform quantization and the AR model compression algorilhm can satisfied the requirement of the compression for the interferometric hyperspectral image series.

关 键 词：干涉超光谱图像 数据压缩 均匀值量化 AR模型 HUFFMAN编码 

分 类 号：TH919.8[机械工程]