High-precision well-seismic data matching method

Abstract

The invention discloses a high-precision well-seismic data matching method, which is characterized in that the method can accurately match well data content corresponding to a minimum-resolution seismic grid and makes a reservoir mark. The method realizes high-precision well-seismic data matching, can confirm the number and location of well data corresponding to each seismic grid, and realizes high-precision seismic grid data labels; the reservoir and non-reservoir are not expressed by 0 and 1; and each seismic grid reservoir label is floating-point data within the range from 0 to 1, and accurate sample point support can be provided for machine learning.

Description

technical field [0001] The invention belongs to the fields of geophysical exploration and machine learning, and in particular relates to a high-precision well seismic data matching method. Background technique [0002] Due to the difference in principles and methods between seismic exploration and well logging, seismic data can reflect horizontal and vertical formation information by taking advantage of dense plane acquisition, but its vertical resolution is low, while well data has high vertical resolution but horizontal resolution. Low. Conventional well-seismic multi-scale matching processing methods use time-depth conversion to give the matching situation of certain position depth and time, or reduce the resolution of well data to achieve multi-scale matching, which is not as accurate as seismic grid and high-precision well data. The matching of data also makes it difficult to determine whether the sample point is the data label of the reservoir during machine learning....

AI Technical Summary

Problems solved by technology

[0002] Due to the difference in principles and methods between seismic exploration and well logging, seismic data can reflect horizontal and vertical formation information by taking advantage of dense plane acquisition, but its vertical resolution is low, while well data has high vertical resolution but horizontal resolution. Low

Conventional well-seismic multi-scale matching processing methods use time-depth conversion to give the matching situation of certain position depth and time, or reduce the resolution of well data to achieve multi-scale matching, which is not as accurate as seismic grid and high-precision well data. The matching of data also makes it difficult to determine whether the sample point is the data label of the reservoir during machine learning

Images