Abstract
The deep convolutional neural networks (CNNs) have been shown excellent performances for image denoising. However, the denoising CNN model trained with a specific noise level cannot deal with the images which have spatiotemporally variant random noise and low signal-to-noise ratio (SNR), such as seismic images. To this end, we propose a patch-based denoising CNN method, namely PDCNN. Specifically, we cluster the overlapping patches of noisy image into K classes where the image patches have close noise levels in each class, and then choose a suitable model for denoising the corresponding class from a series of well-trained CNN models. By embodying the structural statistics, we propose a CNN model selection criterion with a structural-dependent parameter. In contrast to the manual model selection process, the more accurate CNN model is chosen automatically and effectively. The capability of the PDCNN is demonstrated on synthetic and field seismic images. Experimental results show that the proposed method largely benefits from using multiple CNN models to jointly denoise, and leads to the satisfactory denoising performance in spatiotemporally variant seismic random noise reduction and structural signal preservation.
Highlights
Seismic exploration is the significant way for probing subsurface structures to explore oil and gas resources
The single convolutional neural networks (CNNs) model named denoising CNN models (DnCNNs)-B, which is trained by a range of noise levels for blind Gaussian denoising in [21], can deal with spatiotemporally variant random noise, but it still has some trouble in well generalization for complicated real noisy images
THE PROPOSED PATCH-BASED DENOISING CNN METHOD Based on the superior denoising ability of the single CNN model for its matching noise level, we propose a patch-based denoising CNN method (PDCNN) to remove the spatiotemporally variant random noise in seismic images, by combining the patch clustering and multiple denoising CNN models
Summary
Seismic exploration is the significant way for probing subsurface structures to explore oil and gas resources. Y. Zhang et al.: Patch Based Denoising Method Using Deep CNN for Seismic Image performance in preserving the details of seismic structural signals [11], [12]. The single CNN model named DnCNN-B, which is trained by a range of noise levels for blind Gaussian denoising in [21], can deal with spatiotemporally variant random noise, but it still has some trouble in well generalization for complicated real noisy images. Motivated by patch-based denoising, we propose a patchbased denoising CNN method, namely PDCNN, which combines the patch clustering and multiple CNN models to remove the spatiotemporally variant random noise in seismic images. Comparing with traditional seismic denoising methods and state-of-art denoising model, our method performs better in terms of signal preservation and spatiotemporally variant random noise reduction
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
