CN-121982260-A - Geological three-dimensional exploration image reconstruction method

CN121982260ACN 121982260 ACN121982260 ACN 121982260ACN-121982260-A

Abstract

The invention belongs to the technical field of image processing, and particularly relates to a geological three-dimensional exploration image reconstruction method which comprises the steps of calculating credibility weights, self-adaptive adjustment parameters and geological complexity of data points, determining contribution weights of all the data points to grid nodes, screening adjacent points of the grid nodes, constructing an initial three-dimensional grid based on geological attribute values, calculating gradient vectors of all the nodes of the initial three-dimensional grid, constructing a structure tensor matrix, decomposing to obtain characteristic values and characteristic vectors, dividing a structure area according to the characteristic values, calculating relative weights and normalized adjustment factors of all the directions, combining preset area type factors to obtain smooth adjustment coefficients of all the grid nodes in all the directions, constructing a diffusion matrix by the smooth adjustment coefficients of all the directions and the corresponding characteristic vectors, and iteratively optimizing the initial three-dimensional grid by an anisotropic diffusion technology to obtain a reconstructed geological three-dimensional exploration image. The invention improves the accuracy of reconstructing the geological three-dimensional exploration image.

Inventors

WANG WENFEI
WANG YONGJUN
PENG ZHENZHOU
LI HUANHUAN
FENG XIAOQIANG
GUO YUQI
GUO YUHUI

Assignees

陕西省一三九煤田地质水文地质有限公司
陕西省煤田地质集团有限公司

Dates

Publication Date: 20260505
Application Date: 20260408

Claims (10)

1. A method for reconstructing a geological three-dimensional survey image, comprising: the method comprises the steps of carrying out weighted summation on normalized signal-to-noise ratio and stability coefficient of data points to obtain reliability weight of the data points, calculating to obtain average data point density, taking reciprocal of the average data point density as self-adaptive adjustment parameter, setting geological complexity according to geological survey report, and calculating contribution weight of each data point to be interpolated grid node; Obtaining gradient vectors of grid nodes in an initial three-dimensional grid through three-dimensional Sobel operator calculation, constructing a structure tensor matrix, decomposing characteristic values to obtain characteristic values and characteristic vectors of the grid nodes, and dividing a layered structure area, a fracture structure area and a uniform structure area according to the relative sizes of the characteristic values; calculating normalized adjustment factors of the feature vectors corresponding to all directions based on the feature values, and multiplying the normalized adjustment factors by preset region type factors to obtain smooth adjustment coefficients of grid nodes in all directions; according to the smooth adjustment coefficient and the characteristic vector of each grid node in each direction, a diffusion matrix of each grid node is constructed, an anisotropic diffusion technology is adopted to carry out iterative optimization on an initial three-dimensional grid, a final three-dimensional grid is obtained, and a reconstructed geological three-dimensional exploration image is obtained through three-dimensional visualization mapping.
2. The method of reconstructing a geological three-dimensional survey of claim 1, wherein obtaining a confidence weight for a data point comprises: ; In the formula, Is a data point Reliability weights of (2); Is a data point The normalized signal-to-noise ratio of (2) and the weight coefficient is 0.6; Is a data point Is a stability coefficient of 0.4; Is an index of data points in the original survey data set.
3. The method for reconstructing a geological three-dimensional exploration image according to claim 1, wherein said method for obtaining stability coefficients comprises: obtaining geological attribute values of all data points in the original exploration data set according to physical detection results Is arranged as a center Is to data points in the three-dimensional neighborhood of Normalized values of ratios of the geologic attribute values of (c) to the average of the geologic attribute values of all data points within its stereo vicinity as data points Is a stability coefficient of (c).
4. A method of reconstructing a geological three-dimensional survey image according to claim 1, wherein said calculating the contribution weights of each data point to the grid node to be interpolated comprises: ; In the formula, Is a data point Grid node to be interpolated Is a contribution weight of (2); Is a data point With the grid node to be interpolated Euclidean distance between them; is a distance decay parameter; is a self-adaptive adjustment parameter; is the degree of geological complexity; Is a data point Reliability weights of (2); is a natural exponential function; Is an index of data points in the original survey data set.
5. The method for reconstructing a geological three-dimensional exploration image according to claim 1, wherein said screening a set of neighboring points based on euclidean distance comprises: computing grid nodes to be interpolated Euclidean distance between each data point in original exploration data set and grid node to be interpolated Centering, taking all data points with Euclidean distances smaller than a preset searching radius as grid nodes to be interpolated Obtaining a grid node to be interpolated Is a set of neighboring points of (c).
6. The method for reconstructing a geological three-dimensional survey according to claim 1, wherein the dividing of the layered structure region, the fractured structure region and the uniform structure region according to the relative sizes of the feature values comprises grid nodes The characteristic values of (a) are respectively , wherein, Presetting a preset proportion threshold value, a first preset difference value threshold value and a second preset difference value threshold value based on the actual demand of geological exploration And (3) with Is greater than a preset ratio threshold, and And (3) with When the difference value of the grid nodes is smaller than a first preset difference threshold value, determining that the grid nodes are positioned in the layered structure area, when And (3) with The difference value of (2) is smaller than a first preset difference threshold value, and And (3) with When the ratio of the grid nodes is larger than a preset ratio threshold value, judging that the grid nodes are positioned in the fracture structure area, and when And (3) with The difference value of (2) is smaller than a second preset difference threshold value, And (3) with And when the difference value of the grid nodes is smaller than a second preset difference value threshold value, judging that the grid nodes are positioned in the uniform structure area.
7. The method for reconstructing a geological three-dimensional exploration image according to claim 1, wherein said calculating normalized adjustment factors of feature vectors corresponding to respective directions based on feature values comprises: Will be the first Directional grid node Dividing the eigenvalues of (a) by the grid nodes in three directions The sum of the characteristic values of (a) is the first calculation result The relative weight of the directions is such that, Is the index of the direction to which the feature vector corresponds, The values of (1), (2) and (3) are grid nodes in three directions The characteristic values of (a) are respectively Subtracting the first from the value 1 The maximum normalization operation is carried out on the result obtained by the relative weight of the direction, and the finally obtained normalization value is the first Normalized adjustment factor of direction.
8. A method of reconstructing a geological three-dimensional survey according to claim 1, wherein said predetermined region type factor comprises: when grid node In the case of a layered structure region, the region type factor When grid node In the case of a fracture structure region, the region type factor When grid node When belonging to a uniform structural region, the region type factor 。
9. The method for reconstructing a geological three-dimensional exploration image according to claim 1, wherein said constructing a diffusion matrix of each grid node according to the smoothed adjustment coefficient and the eigenvector of each grid node in each direction comprises: ; Wherein, the Is a grid node Is a diffusion matrix of (a); Is the first Directional grid node Is a smooth adjustment coefficient of (2); Is the first Directional grid node Is a feature vector of (1); Representing the transpose.
10. A method of reconstructing a geological three-dimensional survey according to claim 1, wherein said obtaining a reconstructed geological three-dimensional survey comprises: The method comprises the steps of carrying out iterative optimization on an initial three-dimensional grid through an anisotropic diffusion technology based on a diffusion matrix of all grid nodes, setting iteration stop conditions, stopping iteration when the number of iterations reaches 10, finally obtaining a final three-dimensional grid after optimization convergence, and finally inputting geological attribute values of each grid node in the final three-dimensional grid to a three-dimensional graph rendering engine, and mapping to generate a reconstructed geological three-dimensional exploration image.

Description

Geological three-dimensional exploration image reconstruction method Technical Field The invention relates to the technical field of image processing. More particularly, the present invention relates to a geological three-dimensional survey image reconstruction method. Background The three-dimensional geological exploration image reconstruction is a core technology of geological resource exploration, is applied to the fields of underground structure analysis and the like, and is difficult to comprehensively reflect the spatial distribution characteristics of a geological body along with the complexity of a geological structure, and the three-dimensional reconstruction technology can intuitively present the three-dimensional form, the fault distribution and the like of the geological body and can provide more comprehensive and accurate data support for various geological decisions. At present, a conventional radial basis function interpolation algorithm is generally adopted for reconstructing a geological three-dimensional exploration image, and the specific operation comprises the steps of collecting discrete exploration data and taking the discrete exploration data as control points, selecting a preset radial basis function, such as a Gaussian function, as a kernel function, wherein the numerical value of the radial basis function only depends on the distance between the control points, determining a weight coefficient corresponding to each control point by solving a linear equation set constructed on the basis of all the control points, and calculating the data value of any unknown position in the whole three-dimensional space by utilizing the weight coefficients and the kernel function so as to obtain the reconstructed geological three-dimensional exploration image. The conventional radial basis function interpolation method has inherent limitations that firstly, in the step of kernel function selection and parameter setting, parameters are preset by relying on manual experience, the lack of self-adaptive matching capability on geologic structure characteristics easily causes excessive smoothness of interpolation results in complex structural areas, namely fault boundary blurring and lithology layering detail loss, and secondly, in the integral process of constructing and solving an interpolation model, the mechanism of the interpolation method only depends on pure mathematical mapping of discrete data points, and the prior knowledge of geologic structures cannot be effectively integrated, so that the reconstructed geologic three-dimensional exploration image and the real geologic structure have obvious deviation. Disclosure of Invention In order to solve the technical problems that the conventional method relies on manual experience to set parameters and does not integrate with geological priori knowledge, so that a reconstructed image is excessively smooth in a complex area, faults are blurred and the deviation from a real structure is large, the invention provides the scheme in the following aspects. In a first aspect, the invention provides a geological three-dimensional exploration image reconstruction method, which comprises the steps of preprocessing exploration data of an exploration area to obtain an original exploration data set, obtaining credibility weights, self-adaptive adjustment parameters and geological complexity of data points, calculating contribution weights of grid nodes to be interpolated of the data points, screening adjacent point sets according to Euclidean distances between the grid nodes to be interpolated and the data points in the original exploration data set, carrying out weighted summation on geological attribute values of all the data points in the adjacent point sets based on the contribution weights to obtain geological attribute values of the grid nodes to be interpolated, constructing an initial three-dimensional grid, obtaining gradient vectors of grid nodes in the initial three-dimensional grid through three-dimensional Sobel operator calculation, constructing a structure tensor matrix, carrying out eigenvalue decomposition, obtaining eigenvalue and eigenvector of each grid node, dividing a layered structure area, a fracture structure area and a uniform structure area according to the relative sizes of the eigenvalue, calculating relative weights and normalization adjustment factors of all directions based on the eigenvalue, combining the area type factors of all nodes in all directions, obtaining adjustment coefficients of the initial three-dimensional grid nodes according to the contribution weights, carrying out weighted summation on the geological attribute values of all the data points in all directions, obtaining a final three-dimensional expansion grid, carrying out three-dimensional iterative optimization on the final three-dimensional grid construction by adopting a three-dimensional diffusion algorithm to obtain a final three-dimensional network, and fina