Search

A horizontal towed linear coherent hydrophone array is often employed to estimate the spatial intensity distribution of incident plane waves scattered from the geological and biological features in an ocean acoustic waveguide using conventional beamforming. However, due to the physical limitations of the array aperture, the spatial resolution after conventional beamforming is often limited by the fat main lobe and the high sidelobes. Here, we propose a method originated from computer vision deblurring based on deep learning to enhance the spatial resolution of beamformed images. The effect of image blurring after conventional beamforming can be considered a convolution of beam pattern, which acts as a point spread function (PSF), and the original spatial intensity distributions of incident plane waves. A modified U-Net-like network is trained on a simulated dataset. The instantaneous acoustic complex amplitude is assumed following circular complex Gaussian random (CCGR) statistics. Both synthetic data and experimental data collected from the South China Sea Experiment in 2021 are used to illustrate the effectiveness of this approach, showing a maximum 700% reduction in a 3 dB width over conventional beamforming. A lower normalized mean square error (NMSE) is provided compared with other deconvolution-based algorithms, such as the Richardson–Lucy algorithm and the approximate likelihood model-based deconvolution algorithm. The method is applicable in various acoustic imaging applications that employ linear coherent hydrophone arrays with one-dimensional conventional beamforming, such as ocean acoustic waveguide remote sensing (OAWRS). [ABSTRACT FROM AUTHOR]

Electron microscopy is indispensable for examining the morphology and composition of solid materials at the sub-micron scale. To study the powder samples that are widely used in materials development, scanning electron microscopes (SEMs) are increasingly used at the laboratory scale to generate large datasets with hundreds of images. Parsing these images to identify distinct particles and determine their morphology requires careful analysis, and automating this process remains challenging. In this work, we enhance the Mask R-CNN architecture to develop a method for automated segmentation of particles in SEM images. We address several challenges inherent to measurements, such as image blur and particle agglomeration. Moreover, our method accounts for prediction uncertainty when such issues prevent accurate segmentation of a particle. Recognizing that disparate length scales are often present in large datasets, we use this framework to create two models that are separately trained to handle images obtained at low or high magnification. By testing these models on a variety of inorganic samples, our approach to particle segmentation surpasses an established automated segmentation method and yields comparable results to the predictions of three domain experts, revealing comparable accuracy while requiring a fraction of the time. These findings highlight the potential of deep learning in advancing autonomous workflows for materials characterization. [ABSTRACT FROM AUTHOR]

Deep learning-based video coding methods have demonstrated superior performance compared to classical video coding standards in recent years. The vast majority of the existing deep video coding (DVC) networks are based on convolutional neural networks (CNNs), and their main drawback is that since CNNs are affected by the size of the receptive field, they cannot effectively handle long-range dependencies and local detail recovery. Therefore, how to better capture and process the overall structure as well as local texture information in the video coding task is the core issue. Notably, the transformer employs a self-attention mechanism that captures dependencies between any two positions in the input sequence without being constrained by distance limitations. This is an effective solution to the problem described above. In this paper, we propose end-to-end transformer-based adaptive video coding (TAVC). First, we compress the motion vector and residuals through a compression network built on the vision transformer (ViT) and design the motion compensation network based on ViT. Second, based on the requirement of video coding to adapt to different resolution inputs, we introduce a position encoding generator (PEG) as adaptive position encoding (APE) to maintain its translation invariance across different resolution video coding tasks. The experiment shows that for multiscale structural similarity index measurement (MS-SSIM) metrics, this method exhibits significant performance gaps compared to conventional engineering codecs, such as × 2 6 4 , × 2 6 5 , and VTM-15.2. We also achieved a good performance improvement compared to the CNN-based DVC methods. In the case of peak signal-to-noise ratio (PSNR) evaluation metrics, TAVC also achieves good performance. [ABSTRACT FROM AUTHOR]

Traffic sign detection is a crucial task in computer vision, finding wide-ranging applications in intelligent transportation systems, autonomous driving, and traffic safety. However, due to the complexity and variability of traffic environments and the small size of traffic signs, detecting small traffic signs in real-world scenes remains a challenging problem. In order to improve the recognition of road traffic signs, this paper proposes a small object detection algorithm for traffic signs based on the improved YOLOv7. First, the small target detection layer in the neck region was added to augment the detection capability for small traffic sign targets. Simultaneously, the integration of self-attention and convolutional mix modules (ACmix) was applied to the newly added small target detection layer, enabling the capture of additional feature information through the convolutional and self-attention channels within ACmix. Furthermore, the feature extraction capability of the convolution modules was enhanced by replacing the regular convolution modules in the neck layer with omni-dimensional dynamic convolution (ODConv). To further enhance the accuracy of small target detection, the normalized Gaussian Wasserstein distance (NWD) metric was introduced to mitigate the sensitivity to minor positional deviations of small objects. The experimental results on the challenging public dataset TT100K demonstrate that the SANO-YOLOv7 algorithm achieved an 88.7% mAP@0.5, outperforming the baseline model YOLOv7 by 5.3%. [ABSTRACT FROM AUTHOR]

Haze-free images have become a prerequisite for many computer vision tasks; therefore, single image dehazing is particularly important in the field. However, existing deep learning dehazing methods face the following problems: (1) the substantial potential of prior dehazing methods in recovering image visibility is disregarded; (2) most deep learning dehazing methods primarily use spatial information while disregarding the information in different frequency domains of the images; (3) these methods have limited dehazing capabilities because of a lack of paired hazy and clear images. Therefore, we propose a novel unsupervised dehazing network, DFP-Net, to address the above three issues. Specifically, we embedded the dark channel prior algorithm into our network to combine the advantages of prior and deep learning methods. We carefully designed a dual-branch frequency-domain processing network and a spatial perception fusion network to jointly explore information in different frequency and spatial domains of the images. Furthermore, we combined contrastive learning and adversarial training to alleviate the problem of lacking paired training samples using unpaired training samples. The extensive experimental results demonstrate that DFP-Net outperforms other state-of-the-art methods on synthetic and real datasets, achieving an improvement of approximately 2.8 dB in PSNR and producing visually pleasing results, thereby contributing to enhanced visibility under hazy weather conditions. Code is available at: https://github.com/wsl666/DFP-Net.git. • A novel unsupervised dehazing network combining prior and learning-based approaches. • A dual-branch network processes high- and low-frequency image components separately. • DFP-Net trained with unpaired samples mitigates data scarcity in supervised methods. [ABSTRACT FROM AUTHOR]

Carbon fiber reinforced plastic (CFRP) has become one of the main structural materials for aerospace vehicles. However, some internal defects are prone to occur and have potential to cause significant losses of life and property. Currently, the detection of internal defects for CFRP mainly relies on ultrasonic, and other technologies, while they have disadvantages of low efficiency, and poor adaptability. Therefore, this paper explores a novel method to locate internal defects of CFRP laminates by analyzing vibration signals. Firstly, a signal acquisition scheme is designed. Then, a global interactive attention-based lightweight denoising network (GIALDN) is designed to analyze vibration signals and locate internal defects of CFRP laminates. In GIALDN, the threshold denoising method is used to eliminate noise-related features and improve feature discrimination; a global interactive attention module is designed, which makes the network pay more attention to the valid features while realizing the global interactive connection and obtains the rich contextual features; combining with the convolution layer of de-pooling strategy and multi-layer convolution using the residual connection, the backbone of the network is formed. Finally, an experimental platform is established to test the performance of GIALDN. Results show that the location accuracy of GIALDN can reach 98.68%, which is more than 15% higher than those of VGGnet11 and FaultNet, and is also superior to those of LSTM, RNN, Rsenet18, SEresnet18 and Densenet121. Lastly, the location accuracies of GIALDN on CFRP laminates with the same thickness and different stacking sequences are investigated and a good model applicability can be observed. • GIALDN is designed for locating the internal defects in CFRP laminates. • LDM is proposed to effectively eliminate noise and improve feature discrimination. • GIAM is proposed to obtain the inter-data connection and rich contextual features. • De-pooling and 1DCNN strategies are developed to improve the model performance. • The accuracy of GIALDN is 98.68%, which is higher than the existing models. [ABSTRACT FROM AUTHOR]

The heat generation is significant in the machining process, leading to thermal errors, and finally the geometric precision of machined parts is reduced. So the precision machine tool is a key factor in determining the geometric precision of complex parts. In recent years, the error control method is applied. But the method fails in reducing thermal errors because it cannot effectively process large-volume data, resulting from its low executing efficiency. To solve above issues, a new intelligent digital-twin prediction and reverse control system is designed for thermally induced errors based on the user-edge-cloud architecture to expedite the executing efficiency. The data-driven error modeling method is augmented by an error mechanism-based modeling to express the thermal error as a function with the temperature, armature current, rotational speed, and ambient temperature as independent variables, and then the long-term memorizing behavior of thermal errors is demonstrated. The error model is established based on an improved wavelet threshold denoising (IWTD) and a (long short-term memory) LSTM network to describe the memorizing behavior, and IWTD-LSTM network error prediction model is embedded into the digital-twin system. The digital-twin system and IWTD-LSTM network model were verified on a precision machine tool. With the implementation of the digital-twin system, the thermal error and the volume of the transferred data are reduced by 88.72% and 56.36%, respectively. [ABSTRACT FROM AUTHOR]

The geometric precision of machined gears is reduced by thermal errors. So the prediction and control of thermal errors are essential. But the prediction and control are a process involving the processing of a large-volume thermal data, and then the processing efficiency is low, which severely hinders the geometric precision improvement. To solve this problem, a new mist-edge-fog-cloud system (MEFCS) architecture is proposed for the error prediction and control. A finite element model is established to prove the applicability of bidirectional long short-term memory (Bi-LSTM) network. A cosine and sine gray wolf optimization (CSGWO) algorithm is proposed to optimize the batch size. Then the CSGWO-Bi-LSTM network error model is proposed. The predictive accuracy is 90.80%, 94.57%, 95.77%, 96.79%, 97.51%, 98.45%, and 98.92% for the multiple linear regression model, recurrent neural network, LSTM network, Bi-LSTM network, CSGWO1-Bi-LSTM network, CSGWO2-Bi-LSTM network, and CSGWO3-Bi-LSTM network, respectively. The volume of the transferred data is reduced by 11/16 with the data-based model, and the volume of the transferred thermal data is reduced to 1/10 with the designed system. A precision threshold is set, and the predictive accuracy is improved by 8.31% by the system with the precision threshold compared with the system without the precision threshold. With the proposed MEFCS, the accuracy level of the tooth profile deviation f H α is increased from ISO level 5 to ISO level 3. The total execution time of the mist-cloud structure, mist-edge-cloud structure, mist-fog-cloud structure, and mist-edge-fog-cloud structure is 206 s, 200 s, 186 s, and 167 s, respectively. [Display omitted] • Physical- and data-based models are combined to prove error mechanism. • CSGOW with different control parameters is proposed to optimize batch size. • CSGOW-Bi-LSTM network is used to build error models. • New MEFCS architecture is proposed to enhance system executing efficiency. • MEFCS is developed to realize error prediction and control. [ABSTRACT FROM AUTHOR]