Search

Synthetic Aperture Radar (SAR) is renowned for its all-weather and all-time imaging capabilities, making it invaluable for ship target recognition. Despite the advancements in deep learning models, the efficiency of Convolutional Neural Networks (CNNs) in the frequency domain is often constrained by memory limitations and the stringent real-time requirements of embedded systems. To surmount these obstacles, we introduce the Split_ Composite method, an innovative convolution acceleration technique grounded in Fast Fourier Transform (FFT). This method employs input block decomposition and a composite zero-padding approach to streamline memory bandwidth and computational complexity via optimized frequency-domain convolution and image reconstruction. By capitalizing on FFT's inherent periodicity to augment frequency resolution, Split_ Composite facilitates weight sharing, curtailing both memory access and computational demands. Our experiments, conducted using the OpenSARShip-4 dataset, confirm that the Split_ Composite method upholds high recognition precision while markedly enhancing inference velocity, especially in the realm of large-scale data processing, thereby exhibiting exceptional scalability and efficiency. When juxtaposed with state-of-the-art convolution optimization technologies such as Winograd and TensorRT, Split_ Composite has demonstrated a significant lead in inference speed without compromising the precision of recognition. [ABSTRACT FROM AUTHOR]

The Earth rotation parameters (ERPs), including polar motion (PMX and PMY) and universal time (UT1-UTC), play a central role in functions such as monitoring the Earth's rotation and high-precision navigation and positioning. Variations in ERPs reflect not only the overall state of movement of the Earth, but also the interactions among the atmosphere, ocean, and land on the spatial and temporal scales. In this paper, we estimated ERP series based on very long baseline interferometry (VLBI) observations between 2011–2020. The results show that the average root mean square errors (RMSEs) are 0.187 mas for PMX, 0.205 mas for PMY, and 0.022 ms for UT1-UTC. Furthermore, to explore the high-frequency variations in more detail, we analyzed the polar motion time series spectrum based on fast Fourier transform (FFT), and our findings show that the Chandler motion was approximately 426 days and that the annual motion was about 360 days. In addition, the results also validate the presence of a weaker retrograde oscillation with an amplitude of about 3.5 mas. This paper proposes a hybrid prediction model that combines convolutional neural network (CNN) and long short-term memory (LSTM) neural network: the CNN–LSTM model. The advantages can be attributed to the CNN's ability to extract and optimize features related to polar motion series, and the LSTM's ability to make medium- to long-term predictions based on historical time series. Compared with Bulletin A, the prediction accuracies of PMX and PMY are improved by 42% and 13%, respectively. Notably, the hybrid CNN–LSTM model can effectively improve the accuracy of medium- and long-term polar motion prediction. [ABSTRACT FROM AUTHOR]

Layover detection is crucial in 3D array SAR topography reconstruction. However, existing algorithms are not automated and accurate enough in practice. To solve this problem, this paper proposes a novel layover detection method that combines the complex-valued (cv) neural network and expert knowledge to extract features in the amplitude and phase of multi-channel SAR. First, inspired by expert knowledge, a fast Fourier transform (FFT) residual convolutional neural network was developed to eliminate the training divergence of the cv network, deepen networks without extra parameters, and facilitate network learning. Then, another innovative component, phase convolution, was designed to extract phase features of the layover. Subsequently, various cv neural network components were integrated with FFT residual learning blocks and phase convolution on the skeleton of U-Net. Due to the difficulty of obtaining SAR images marked with layover truths, a simulation was performed to gather the required dataset for training. The experimental results indicated that our approach can efficiently determine the layover area with higher precision and fewer noises. The proposed method achieves an accuracy of 97% on the testing dataset, which surpasses previous methods. [ABSTRACT FROM AUTHOR]

As a critical component in industrial systems, timely and accurate fault diagnosis of rolling bearings is closely related to reliability and safety. Since the equipment usually operates in normal conditions with few fault samples, unbalanced data distribution problems lead to poor fault diagnosis ability. To address the above problems, a two-channel convolutional neural network (TC-CNN) model is proposed. Firstly, the frequency spectrum of the vibration signal is extracted using the Fast Fourier Transform (FFT), and the frequency spectrum is used as the input to the one-dimensional convolutional neural network (1D-CNN). Secondly, the time-frequency image of the vibration signal is extracted using generalized S-transform (GST), and the time-frequency image is used as the input to the two-dimensional convolutional neural network (2D-CNN). Then, feature extraction in the convolution and pooling layers is performed in the above two CNN channels, respectively. The feature vectors obtained from the two CNN models are stitched together in the fusion layer, and the fault classes are identified using an SVM classifier. Finally, using the rolling bearing experimental dataset of Case Western Reserve University (CWRU), the fault diagnosis effect of the proposed TC-CNN model under various data imbalance conditions is verified. In comparison with other related works, the experimental results demonstrate the better fault diagnosis results and robustness of the method. [ABSTRACT FROM AUTHOR]

The massive environmental noise interference and insufficient effective sample degradation data of the intelligent fault diagnosis performance methods pose an extremely concerning issue. Realising the challenge of developing a facile and straightforward model that resolves these problems, this study proposed the One-Dimensional Convolutional Neural Network (1D-CNN) based on frequency-domain signal processing. The Fast Fourier Transform (FFT) analysis is initially utilised to transform the signals from the time domain to the frequency domain; the data was represented using a phasor notation, which separates magnitude and phase and then fed to the 1D-CNN. Subsequently, the model is trained with White Gaussian Noise (WGN) to improve its robustness and resilience to noise. Based on the findings, the proposed model successfully achieved 100% classification accuracy from clean signals and simultaneously achieved considerable robustness to noise and exceptional domain adaptation ability. The diagnosis accuracy retained up to 97.37%, which was higher than the accuracy of the CNN without training under noisy conditions at only 43.75%. Furthermore, the model achieved an accuracy of up to 98.1% under different working conditions, which was superior to other reported models. In addition, the proposed model outperformed the state-of-art methods as the Signal-to-Noise Ratio (SNR) was lowered to −10 dB achieving 97.37% accuracy. In short, the proposed 1D-CNN model is a promising effective rolling bearing fault diagnosis. [ABSTRACT FROM AUTHOR]

The use of surface roughness (Ra) to indicate product quality in the milling process in an intelligent monitoring system applied in-process has been developing. From the considerations of convenient installation and cost-effectiveness, accelerator vibration signals combined with deep learning predictive models for predicting surface roughness is a potential tool. In this paper, three models, namely, Fast Fourier Transform-Deep Neural Networks (FFT-DNN), Fast Fourier Transform Long Short Term Memory Network (FFT-LSTM), and one-dimensional convolutional neural network (1-D CNN), are used to explore the training and prediction performances. Feature extraction plays an important role in the training and predicting results. FFT and the one-dimensional convolution filter, known as 1-D CNN, are employed to extract vibration signals' raw data. The results show the following: (1) the LSTM model presents the temporal modeling ability to achieve a good performance at higher Ra value and (2) 1-D CNN, which is better at extracting features, exhibits highly accurate prediction performance at lower Ra ranges. Based on the results, vibration signals combined with a deep learning predictive model could be applied to predict the surface roughness in the milling process. Based on this experimental study, the use of prediction of the surface roughness via vibration signals using FFT-LSTM or 1-D CNN is recommended to develop an intelligent system. [ABSTRACT FROM AUTHOR]

Building a pattern detection model using a deep learning algorithm for data collected from manufacturing sites is an effective way for to perform decision-making and assess business feasibility for enterprises, by providing the results and implications of the patterns analysis of big data occurring at manufacturing sites. To identify the threshold of the abnormal pattern requires collaboration between data analysts and manufacturing process experts, but it is practically difficult and time-consuming. This paper suggests how to derive the threshold setting of the abnormal pattern without manual labelling by process experts, and offers a prediction algorithm to predict the potentials of future failures in advance by using the hybrid Convolutional Neural Networks (CNN)–Long Short-Term Memory (LSTM) algorithm, and the Fast Fourier Transform (FFT) technique. We found that it is easier to detect abnormal patterns that cannot be found in the existing time domain after preprocessing the data set through FFT. Our study shows that both train loss and test loss were well developed, with near zero convergence with the lowest loss rate compared to existing models such as LSTM. Our proposition for the model and our method of preprocessing the data greatly helps in understanding the abnormal pattern of unlabeled big data produced at the manufacturing site, and can be a strong foundation for detecting the threshold of the abnormal pattern of big data occurring at manufacturing sites. [ABSTRACT FROM AUTHOR]

Visibility is a complex phenomenon inspired by emissions and air pollutants or by factors, including sunlight, humidity, temperature, and time, which decrease the clarity of what is visible through the atmosphere. This paper provides a detailed overview of the state-of-the-art contributions in relation to visibility estimation under various foggy weather conditions. We propose VisNet, which is a new approach based on deep integrated convolutional neural networks for the estimation of visibility distances from camera imagery. The implemented network uses three streams of deep integrated convolutional neural networks, which are connected in parallel. In addition, we have collected the largest dataset with three million outdoor images and exact visibility values for this study. To evaluate the model's performance fairly and objectively, the model is trained on three image datasets with different visibility ranges, each with a different number of classes. Moreover, our proposed model, VisNet, evaluated under dissimilar fog density scenarios, uses a diverse set of images. Prior to feeding the network, each input image is filtered in the frequency domain to remove low-level features, and a spectral filter is applied to each input for the extraction of low-contrast regions. Compared to the previous methods, our approach achieves the highest performance in terms of classification based on three different datasets. Furthermore, our VisNet considerably outperforms not only the classical methods, but also state-of-the-art models of visibility estimation. [ABSTRACT FROM AUTHOR]