ملخص
We consider the problem of 3D seismic inversion from prestack data using a very small number of seismic sources. The proposed solution is based on a combination of compressed-sensing and machine learning frameworks, known as compressed-learning (Calderbank et al., 2009; Calderbank and Jafarpour, 2012; Zisselman et al., 2018; Mdrafi and Gurbuz, 2021; Douarre et al., 2021; Mou and Zhang, 2023). The solution jointly optimizes a dimensionality reduction operator and a 3D inversion encoder-decoder implemented by a deep convolutional neural network (DCNN). Dimensionality reduction is achieved by learning a sparse binary sensing layer that selects a small subset of the available sources, then the selected data is fed to a DCNN to complete the regression task. The end-to-end learning process provides a reduction by an order-of-magnitude in the number of seismic records used during training, while preserving the 3D reconstruction quality comparable to that obtained by using the entire dataset.
| اللغة الأصلية | الإنجليزيّة |
|---|---|
| الصفحات (من إلى) | 1054-1058 |
| عدد الصفحات | 5 |
| دورية | SEG Technical Program Expanded Abstracts |
| مستوى الصوت | 2023-August |
| المعرِّفات الرقمية للأشياء | |
| حالة النشر | نُشِر - 14 ديسمبر 2023 |
| منشور خارجيًا | نعم |
| الحدث | 3rd International Meeting for Applied Geoscience and Energy, IMAGE 2023 - Houston, الولايات المتّحدة المدة: 28 أغسطس 2023 → 1 سبتمبر 2023 |
بصمة
أدرس بدقة موضوعات البحث “Deep Compressed Learning for 3D Seismic Inversion'. فهما يشكلان معًا بصمة فريدة.قم بذكر هذا
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver