Erratic seismic noise, following a (known or unknown) non-Gaussian distribution, poses a formidable challenge to conventional methods of random noise attenuation. Many erratic noise cancellation methods, for instance, robust reduced-rank and sparsity-promoting filtering, have been proven to achieve promising results in overcoming this challenge. Among them, deep learning (DL) methods require no assumptions about the underlying clear seismic image and are also more robust against erratic and random noise. However, the success of existing DL-based denoising methods strongly depends on supervised learning from a large number of ground-truth seismic images affected by erratic noise and their clean counterparts, which are typically unavailable in a real-world setting. As an alternative, this article presents an unsupervised DL method for erratic-plus-Gaussian noise removal based on a robust deep convolutional autoencoder (RDCAE). In the RDCAE, the mean squared error (mse) loss in a classic DCAE is replaced by the smooth Welsch function to exploit the concept of robust image denoising. In this way, the erratic noise is downweighted by means of a curbed weight defined in terms of the Welsch function. In contrast, the random noise is diluted by combining the mean square in the Welsch function and the total variation (TV). Subsequently, the training procedures required for solving the RDCAE are derived on the basis of the backpropagation (BP) algorithm for a neural network. Experiments conducted on both synthetic and real field datasets are reported to illustrate the efficacy of the proposed method.

The industry treats the distributed acoustic sensing (DAS) system, which uses an optical fiber cable in vertical seismic profiling (VSP) data acquisition, as a new and fast-growing technology. The high-quality data set acquired from the DAS acquisition system can produce high-precision VSP images and obtain more detailed checkshots. However, in field data, the acquired VSP data set suffers from strong coherent DAS coupling noise. Many factors may cause coherent DAS coupling noise, such as the cable slapping and ringing due to the physical placement, the regular swing of the wireline in the well, and the uncoupling between the fiber cable and the casing. This DAS coupling noise reduces the signal-to-noise ratio and affects the subsequent processing and interpretation. Removing the DAS coupling noise can help to improve the quality of the VSP data set acquired with the DAS system. We have developed a sparse-optimization-based DAS coupling noise removing method. In the DAS-based VSP data set, the effective signal and the coupling noise have distinct morphological characteristics. The effective VSP signal has a wide bandwidth, whereas the DAS coupling noise appears in some narrow frequency bands in the frequency domain. The continuous wavelet transform and the discrete cosine transform can sparsely represent the effective VSP signal and DAS coupling noise, respectively. Therefore, we choose these two transforms as two sparse dictionaries and combine them to form an overcomplete dictionary. The morphological component analysis (MCA) can use the morphological difference between different components and the overcomplete dictionary to sparsely represent all components in the complicated signal. Based on the MCA theory, we use the block coordinate relaxation algorithm to separate the effective VSP signal and DAS coupling noise. Applications on a synthetic data set and two field data sets have validated the effectiveness of our method.

The commonly used stable factor methods for the inverse Q-filter achieve good performance in seismic data processing; however, the constant gain-limit assumption in these methods is not associated with the effective frequency band of seismic data and cannot obtain desirable results with high resolution and high signal-to-noise ratio (S/N). Our extended stable factor method for the inverse Q-filter extends these methods by introducing two parameters and constant or self-adaptive gain limit to achieve the desirable high-resolution and high-S/N result. The extended stable factor method for the inverse Q-filter can be implemented in the frequency or time-frequency domain; the latter implementation achieves a higher S/N. Analysis of synthetic signals and field seismic data application illustrate that our method can produce a desirable high-resolution and high-S/N result.

Automatic extraction of high-precision roads from remote sensing images is crucial for path planning and road monitoring. However, there is room to improve the accuracy and generalization of existing methods in segmentation due to the challenges posed by ground object occlusion and complex backgrounds. Most existing methods rely on convolutional neural networks (CNNs), but the limitations of convolution prevent direct semantic interaction at a distance. In contrast, Mix-Transformer obtains long-term modeling capability through the self-attention mechanism, and inspired by it, we propose a multiscale self-adaptive network (UMiT-Net) based on the U-shaped structure. First, UMiT-Net extracts global features with the efficient Mix-Transformer backbone. Second, the dilated attention module (DAM) is used in the bottleneck of the network to fuse semantic features further to ensure the connectivity of the road. Third, in the decoder, to improve the accuracy of road segmentation, we construct the multiscale self-adaptive module (MSAM), which summarizes rich scene understanding from dense contexts with strip windows conforming to road morphology, and embed an edge enhancement module (EEM) to correct road edges. Finally, we design patch expanding (PE), which solves the problem of heavy computation of upsampling due to high resolution. The experimental results show that our UMiT-Net is substantially ahead of other state-of-the-art methods and has a significant improvement in generalization ability.