摘要
本文研究最大熵图像重建用于地球物理学中井间观测问题,着重研究了剑桥算法.对算法中拉格朗日乘子的确定给出了新方法;解决了算法的收敛条件和收敛范围;提出了减少算法计算量和内存量的方法,并编制了实用程序MECT.通过数值模型的研究,最后重建出了某矿区的剖图,与BPT和ART方法的结果进行了比较.证明最大熵图像重建法的优越性在于能有效地抑制噪声,分辨率高,边界影响程度小.若改善观测条件,将提高最大熵成像法的分辨率.
Maximum entropy image reconstruction (MEIR) is a newly developed imaging method in the world. It adopts entropy of imaging (functional), reaching extremum, as criterion to produce image of good quality and wide dynamic range with stronger exploration in the theory. This paper deals with problems of application of MEIR to well-to-well computerized tomography in geophysics. It places emphasis on the Cambridge algorithm which is the most representative one in all of MEIR algorithms. Convergent condition and range are discussed. The method is given to reduce computer cost and memory capacity. A practical program of MECT is designed. This paper is connected with the electromagnetic profile project of the State Seismological Bureau. The authors adopted program MECT to obtain a cross section of a mining area, and then the result is compared with the results of Back Projection Technique (BPT) and Algebraic Reconstruction Technique (ART), so MEIR's superiority in suppressing noise, having high resolution and less boundary effect is proved. It is discovered that if measurement condition be improved such as to increase the data collection in the measurement system, to set measurement points on the surface, the resolution of MEIR will be farther improved.
出处
《地球物理学报》
SCIE
EI
CAS
CSCD
北大核心
1992年第2期234-241,T002,共9页
Chinese Journal of Geophysics
基金
地震科学联合基金
关键词
井间
层析
成像
最大熵
电磁波
Maximum Entropy image reconstruction, Crosshole tomography, Elec-tromagnetic profile, Nurrerical modelling.
