Inversion of Oceanic Parameters Represented by CTD Utilizing Seismic Multi-Attributes Based on Convolutional Neural Netw
- PDF / 802,092 Bytes
- 9 Pages / 595.22 x 842 pts (A4) Page_size
- 64 Downloads / 174 Views
Inversion of Oceanic Parameters Represented by CTD Utilizing Seismic Multi-Attributes Based on Convolutional Neural Network AN Zhenfang1), 2), ZHANG Jin1), 2), *, and XING Lei1), 2) 1) Key Laboratory of Submarine Geosciences and Prospecting Techniques, Ministry of Education and College of Marine Geosciences, Ocean University of China, Qingdao 266100, China 2) Evaluation and Detection Technology Laboratory of Marine Mineral Resources, Pilot Qingdao National Laboratory for Marine Science and Technology (Qingdao), Qingdao 266071, China (Received January 30, 2019; revised May 10, 2019; accepted May 6, 2020) © Ocean University of China, Science Press and Springer-Verlag GmbH Germany 2020 Abstract In Recent years, seismic data have been widely used in seismic oceanography for the inversion of oceanic parameters represented by conductivity temperature depth (CTD). Using this technique, researchers can identify the water structure with high horizontal resolution, which compensates for the deficiencies of CTD data. However, conventional inversion methods are modeldriven, such as constrained sparse spike inversion (CSSI) and full waveform inversion (FWI), and typically require prior deterministic mapping operators. In this paper, we propose a novel inversion method based on a convolutional neural network (CNN), which is purely data-driven. To solve the problem of multiple solutions, we use stepwise regression to select the optimal attributes and their combination and take two-dimensional images of the selected attributes as input data. To prevent vanishing gradients, we use the rectified linear unit (ReLU) function as the activation function of the hidden layer. Moreover, the Adam and mini-batch algorithms are combined to improve stability and efficiency. The inversion results of field data indicate that the proposed method is a robust tool for accurately predicting oceanic parameters. Key words oceanic parameter inversion; seismic multi-attributes; convolutional neural network
1 Introduction Oceanic parameters, including temperature, salinity, density, and velocity can be obtained directly by conductivity temperature depth (CTD) instruments or indirectly by inversion utilizing seismic data. Although the vertical resolution of CTD data is higher than that of seismic data, its horizontal resolution is far lower. Joint CTD-seismic inversion combines the advantages of both to obtain oceanic parameters with high resolution. However, traditional inversion methods, such as constrained sparse spike inversion (CSSI) and full waveform inversion (FWI), typically assume the existence of prior deterministic mapping operators between the geophysical responses and the geophysical parameters, such as a convolution operator and a wave equation operator. For some oceanic parameters, however, such as temperature and salinity, it is difficult to establish their mapping relationships and seismic responses by mathematical modeling. A recent trend in many scientific fields has been to solve inverse problems using data-driven methods and a * Corr
Data Loading...
