Seismic facies interpretation provides a reference for analyzing geological conditions and predicting oil and gas reservoirs. The application of deep learning in seismic facies interpretation reduces a lot of manual work and interpreters’ subjective effects existing in conventional methods. Convolutional neural network (CNN) is widely used interpretation technique in deep learning. However, conventional CNN is not the best model for interpreting massive dataset due to its low efficiency and low classification accuracy. Given this issue, we propose an effective scheme, which is based on an enhanced encoder-decoder structure named DeepLabv3+ developed by Google. This encoder-decoder structure performs higher accuracy and efficiency than CNN models and simple semantic segmentation models such as deconvolutional neural network for extracting multi-scale contextual information and recovering more pixel-level details in prediction results. By training diversity samples after data augmentation and tuning parameters, the application of our scheme on F3 dataset demonstrates its good performance for high accuracy and efficiency.

Presentation Date: Wednesday, September 18, 2019

Session Start Time: 1:50 PM

Presentation Time: 4:45 PM

Location: 221D

Presentation Type: Oral

This content is only available via PDF.
You can access this article if you purchase or spend a download.