Sequence stratigraphic framework construction method and system based on Unet network
A sequence stratigraphic framework and construction method technology, applied in neural architecture, neural learning methods, biological neural network models, etc., can solve problems such as inaccurate stratigraphic framework, difficulty in interpreting horizons, and low signal-to-noise ratio of seismic data
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0032] This embodiment provides a method for constructing a sequence stratigraphic framework based on the Unet network, comprising the following steps:
[0033] S1 forms seismic data and sequence stratigraphic framework label pairs based on the seismic data of the work area with a sequence stratigraphic framework.
[0034] Among them, the process of forming seismic data and sequence stratigraphic framework label pairs includes two steps: preprocessing of seismic data with sequence stratigraphic framework and production of work area label data.
[0035] The preprocessing process of seismic data in the work area with sequence stratigraphic framework includes denoising processing, data standardization processing and data range control. Denoising processing is used to remove random noise data, large and small outliers of data; data normalization processing is through L2 norm normalization processing, Z-score normalization or minimum-maximum normalization processing to compress the...
Embodiment 2
[0047] Based on the same inventive concept, this embodiment discloses a sequence stratigraphic framework construction system based on the Unet network, including:
[0048] a label pair forming module, configured to form seismic data and sequence stratigraphic framework label pairs according to the seismic data of the work area with the sequence stratigraphic framework;
[0049] The model building module is used for substituting the seismic data and the sequence stratigraphic framework label into the initial Unet network model, and the Unet network model is trained to obtain the final Unet network model;
[0050] The data acquisition module to be constructed is used to obtain the seismic data of the work area to be constructed and perform data preprocessing;
[0051] The stratigraphic framework output module substitutes the seismic data to be constructed into the final Unet network model to obtain the sequence stratigraphic framework.
Embodiment 3
[0053] This embodiment provides an application of a sequence stratigraphic framework construction method based on deep learning technology in the case of a simple underground structure, which specifically includes the following steps:
[0054] S1: if Figure 4 As shown, denoising processing, data standardization processing and data range control are performed on the seismic data. Denoising processing is used to remove random noise data, large and small outliers of data; data normalization processing is through L2 norm standardization processing, Z-score standardization or minimum-maximum normalization processing to compress the data to a certain value range space middle. Data range control uses the two horizons that have been interpreted in the target area to control the range of seismic data. The seismic data above and below the two horizons are assigned a value of 0 to form the preprocessed seismic data.
[0055] S2: if Figure 5 As shown, the saved Unet network model par...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More