基于预测编码算法的地震勘探数据无失真压缩理论与方法

海量数据给地震勘探数据的传输、存储和处理提出了严峻挑战.地震数据压缩是解决海量地震数据传输和存储问题的关键.本文定义了SEG Y地震勘探数据文件中的头段数据所占比例(简称头段比例),给出了头段比例计算式,并导出了SEG Y文件压缩倍数与头段比例、头段压缩倍数和样点压缩倍数之间的关系式,从而发现SEG Y文件压缩倍数随样点压缩倍数变化的理论极限是头段压缩倍数与头段比例之比值,并据此从理论上阐明了对头段数据进行高效无失真压缩的必要性.更重要的是,本文对SEG Y数据文件中的头段数据进行了研究,发现了卷头数据和道头数据各自的统计规律,为对头段数据实现高倍压缩提供了重要的理论依据.在此基础上,本文提出了一种适合于对SEG Y头段数据进行高效压缩的方法,实验结果表明,在保证无失真的情况下,本文方法可对SEG Y头段数据实现30~1000倍的压缩,这远高于用Winzip和WinRAR压缩SEG Y头段数据所达到的压缩倍数.

Lossless compression of SEG Y header identification

In this paper, the header ratio is defined as the ratio between the SEG Y header data volume and the SEG Y file size, a formula describing the relationship between the header ratio, the compression ratio of SEG Y header data and the compression ratio of the SEG Y file is derived. It is discovered from the formula that the theoretical limit of the SEG Y file compression ratio is the quotient of the SEG Y header data compression ratio divided by the header ratio, therefore, it is necessary to compress the header data efficiently in order to get a high compression of the SEG Y file. Furthermore, the statistical properties of the SEG Y reel header data and trace header data are analyzed. An efficient lossless compression method for SEG Y header data, known as Header Identification Data Lossless Prediction Coding (HIDLPC) method, is proposed based on the statistical properties of the SEG Y header data. Experimental results show that the lossless compression ratio by using HIDLPC to compress SEG Y header data is between 30 and 1000, much higher than the corresponding compression ratios by using Winzip and WinRAR.