Efficient ERT measuring and inversion strategies for 3D imaging of buried antiquities
Efficient ERT measuring and inversion strategies for 3D imaging of buried antiquities
复制标题
用于埋藏古物 3D 成像的高效 ERT 测量和反演策略
DOI：
--
复制
发表时间：
2007
期刊：
影响因子：
--
通讯作者：
A. Sarris
中科院分区：
文献类型：
--
作者：
N. Papadopoulos;P. Tsourlos;G. Tsokas;A. Sarris
本文研究了从考古遗址收集的平行和/或正交二维表面电阻率层析成像数据的最佳处理技术（2D与3D反演）。为了在相对较短的时间内收集沿X、Y或XY方向的平行二维剖面，对标准电阻计地球物理仪器进行了简单的改进，采用了极-极阵列。灵敏度分析表明，为了获得可靠的地下三维电阻率图像，平行二维线之间的距离必须更小，或者最多等于基本电极间距。通过模拟和反演合成数据和真实数据证实了这一点。将二维反演剖面的后验组合得到的准三维图像与全三维反演电阻率模型进行直接比较，表明三维反演算法在重建埋藏考古结构方面具有优势，即使在复杂的考古遗址中也是如此。由于许多考古目标固有的三维性，准三维图像会受到人工制品的影响。单一测量方向与完整的三维处理和解释方案相结合，足以详细成像三维地下电阻率变化。此外，准牛顿雅可比矩阵更新技术的实现将处理时间缩短了一半，而精度和分辨率没有明显损失。
ABSTRACT The optimum processing technique (2D vs. 3D inversion) to interpret and visualize parallel and/or orthogonal two‐dimensional surface Electrical Resistivity Tomography data collected from archaeological sites is examined in this work. A simple modification of a standard resistance‐meter geophysical instrument was implemented in order to collect parallel two‐dimensional sections along the X‐, Y‐ or XY‐direction in a relatively short time, employing a pole–pole array. The sensitivity analysis showed that the distance between the parallel 2D lines must be smaller or, at the most, equal to the basic inter‐electrode spacing in order to produce reliable 3D resistivity images of the subsurface. This was confirmed by modelling and inversion of both synthetic and real data. Direct comparisons of the quasi‐3D images, resulting from a posteriori combination of the inverted 2D sections, with the full 3D inverted resistivity models indicated the superiority of the 3D inversion algorithm in the reconstruction of buried archaeological structures, even in complex archaeological sites. Due to the inherent three‐dimensionality of many archaeological targets, quasi‐3D images suffer from artefacts. The combination of a single survey‐direction with a full 3D processing and interpretation scheme is adequate to image the 3D subsurface resistivity variation in detail. Furthermore, the implementation of a quasi‐Newton Jacobian matrix update technique reduced the processing time by one‐half without any significant loss of accuracy and resolution.