摘要
将模式识别方法应用于地球物理资料处理,近年来已引起了国内外学术界的注意。模式识别可以使我们充分利用在野外获得的资料,从中“榨取”出更多的地质信息。它不仅可能自动提取我们所需要的地球物理资料,而且还可能对所提取出来的资料自动进行“解释”。本文主要说明在用计算机进行地球物理资料的自动处理和解释时,可以把其中一些问题归结为模式识别问题,从而可以运用模式识别的许多原理和方法。文中简要介绍了统计模式识别和结构模式识别两大类方法的一般原理。较详细地介绍了用句法方法识别波形和标记的松驰算法,以及它们在地球物理资料自动处理中可能的应用。这是一个值得注意的、有价值的研究领域。
Abstract
The application of pattern recognition to geophysical data processing has arrested the academic circles' attention, both at home and abroad, in recent years. With the aid of the pattern recognition, we can make full use of the field data and much more geology information could be "squeezed out". By ultilizing the pattern recognition, not only the required geophysical data could be automatically extracted, but also the interpretation could be performed in the same way.In the auto processing and interpretating of geophysical data by computer, some of the problems might be attributed to the pattern recognition so that basic principles as well as methods related to the pattern recognition could thus be used to advantage.In this paper, the general principles on the two major categories, statistical pattern recognition and structural pattern recognition,are briefly introduced. The loose algorithm for discerning waveform and mark-sense by syntax-directed method and its possible application in auto processing of geophysical data are discussed in some detail. It is indeed a field of science which merits attention and is open to further studies.
论文详情