关键词: 无人机, 倾斜摄影, 数字露头, 三维露头建模, 多点地质统计学, 训练图像, 三角洲储层构型, 坪头乡

Abstract: Aiming at shortcomings in the traditional research methods of outcrops, using the agility, overall perspective and high-degree quantification of UAV oblique photography, the authors have completed the collection, processing and geological interpretation of delta outcrops, and established the three-dimensional digitized outcrop model and quantitative geological knowledge database of reservoirs. Through the fine analysis and comparative study of multiple outcrop sections, and the use of such methods as lithofacies combination, hierarchical interface, interaction of plane and section, sequential indicator stochastic simulation and multiple-point geostatistical simulation, this study establishes the three-dimensional multiple-point statistical model of the outcrop area. The study shows that:(1) The three-dimensional coordinate system of the oblique photographic data collected by UAV after processing is fully coupled with the information of high-precision images, realizing the three-dimensional digitalization, visualization and quantification of the geological information of outcrops. Under the setting of well points, in combination with the lithofacies constraints of well point, this paper carries out a comparison of outcrops with cross-well base-level cycles and the tracking of substrata, and establishes the spatial stratigraphic framework and quantitative reservoir geological knowledge database of the outcrop area. (2) Combining the analysis of traditional outcrop sections and the quantitative reservoir geological knowledge database of modern UAV oblique photography is a complementary means of establishing the spatial distribution of geological attributes and geological models in the study area, which improves the rationality of research results. Three types of genetic sandbodies with great differences in architecture is developed from bottom to top in the outcrop area, i.e., crossed-superimposed distributary channels, distributary channels superimposed with sheet sand bars, and skeletal sandbodies such as thin sheet sandbody or isolated distributary channels. (3) UAV oblique photography model expands the study of two-dimensional irregular outcrop sections into three-dimensional space, and establishes a three-dimensional sequential indicator stochastic simulation model, which not only enhances the correlation between the sections, but also provides strong support for studying the plane sedimentary microfacies of outcrops, and finely describing the distribution of sandbody architecture by the interaction of plane and section. (4)Based on the interaction of plane and section, as well as data from the quantitative reservoir geological knowledge database, this paper determines the structure model of three-dimensional training images, finally conducts three-dimensional multiple-point statistical simulation, and explores the facies-controlling porosity parameters and "sweet spot" distribution rules. The results show that there are small differences in the realization of the multiple-point model, the distribution of sand bodies in each layer, the quantitative reservoir geological knowledge database, and also the statistical data values of various models; the model achieved by multiple-point 3D simulation is highly reliable.

Key words: UAV, oblique photography, digital outcrop, three-dimensional outcrop modeling, multiple-point geostatistics, training image, delta reservoir architecture, Pingtou township