多波段和查找表预测值校正的高光谱图像压缩  

Compression of hyperspectral images based on multi-band images and lookup table prediction correction

作　　者：张海涛[1] 曹博甫 刘万军[1] 

机构地区：[1]辽宁工程技术大学软件学院,辽宁葫芦岛125105

出　　处：《激光与红外》2017年第9期1181-1185,共5页Laser & Infrared

基　　金：国家自然科学基金项目(No.61172144)资助

摘　　要：为了提高高光谱图像无损压缩的压缩比,提出了一种利用多波段图像和查找表预测值校正的高光谱图像无损压缩方法。首先,利用当前被压缩波段图像的前三个波段图像计算LAIS;然后对获取到的查找表预测值进行两次校正;最后求出预测残差,采用自适应算术编码方法进行编码。对该方法进行了理论分析和实验验证。结果表明,针对美国国家航空航天局提供的AVIRIS图像中的Moffett Field图像的不同区域和不同波段进行实验,所提出的方法比LAIS-LUT压缩比提高了0.09~0.34。In order to improve the compression ratio of lossless compression of hyperspectral images, a lossless com- pression method of hyperspectral images was proposed based on multi - band images and lookup table prediction cor- rection. First, LAIS was calculated by using images from three previous bands of current bands ; and then the lookup ta- ble prediction value was corrected two times ; finally the prediction residual error was calculated, and the data were en- coded with adaptive arithmetic coding method. Theoretical analysis and experimental verification show that, compared with LAIS- LUT,the compression ratio of the proposed method is increased by 0. 09 -0. 34 in different areas and bands of Moffett Field image from NASA＇ s AVIRIS images.

关 键 词：高光谱图像 无损压缩 预测方法 查找表 空间相关性 谱间相关性 

分 类 号：TP751.1[自动化与计算机技术—检测技术与自动化装置]