Search

It is difficult to visually detect internal defects in a large transformer with a metal closure. For convenient internal inspection, a micro-robot was adopted, and an inspection method based on an image-enhancement algorithm and an improved deep-learning network was proposed in this paper. Considering the dim environment inside the transformer and the problems of irregular imaging distance and fluctuating supplementary light conditions during image acquisition with the internal-inspection robot, an improved MSRCR algorithm for image enhancement was proposed. It could analyze the local contrast of the image and enhance the details on multiple scales. At the same time, a white-balance algorithm was introduced to enhance the contrast and brightness and solve the problems of overexposure and color distortion. To improve the target recognition performance of complex carbon-trace defects, the SimAM mechanism was incorporated into the Backbone network of the YOLOv8 model to enhance the extraction of carbon-trace features. Meanwhile, the DyHead dynamic detection Head framework was constructed at the output of the YOLOv8 model to improve the perception of local carbon traces with different sizes. To improve the defect target recognition speed of the transformer-inspection robot, a pruning operation was carried out on the YOLOv8 model to remove redundant parameters, realize model lightness, and improve detection efficiency. To verify the effectiveness of the improved algorithm, the detection model was trained and validated with the carbon-trace dataset. The results showed that the MSH-YOLOv8 algorithm achieved an accuracy of 91.80%, which was 3.4 percentage points higher compared to the original YOLOv8 algorithm, and had a significant advantage over other mainstream target-detection algorithms. Meanwhile, the FPS of the proposed algorithm was up to 99.2, indicating that the model computation and model complexity were successfully reduced, which meets the requirements for engineering applications of the transformer internal-inspection robot. [ABSTRACT FROM AUTHOR]

Point cloud registration is widely used in autonomous driving, SLAM, and 3D reconstruction, and it aims to align point clouds from different viewpoints or poses under the same coordinate system. However, point cloud registration is challenging in complex situations, such as a large initial pose difference, high noise, or incomplete overlap, which will cause point cloud registration failure or mismatching. To address the shortcomings of the existing registration algorithms, this paper designed a new coarse-to-fine registration two-stage point cloud registration network, CCRNet, which utilizes an end-to-end form to perform the registration task for point clouds. The multi-scale feature extraction module, coarse registration prediction module, and fine registration prediction module designed in this paper can robustly and accurately register two point clouds without iterations. CCRNet can link the feature information between two point clouds and solve the problems of high noise and incomplete overlap by using a soft correspondence matrix. In the standard dataset ModelNet40, in cases of large initial pose difference, high noise, and incomplete overlap, the accuracy of our method, compared with the second-best popular registration algorithm, was improved by 7.0%, 7.8%, and 22.7% on the MAE, respectively. Experiments showed that our CCRNet method has advantages in registration results in a variety of complex conditions. [ABSTRACT FROM AUTHOR]

Collaborative problem-solving is a crucial skill to be developed in education. With the development of information technology, collaborative problem-solving in online environments has received increasing attention in recent years. There has been some foundation of research on computer-supported collaborative learning (CSCL) and collaborative problem-solving (CPS). They both have a focus, but there is a trend to combine them. Further understanding of the current state of research on online collaborative problem-solving is much needed. This study reviews 528 relevant papers from the last decade. The researchers summarized the current state of research and hotspots in online collaborative problem-solving based on the visual analysis of CiteSpace. It was found that the number of papers on online collaborative problem-solving is on the rise, with CPS and CSCL as the main research content and 'feature extraction', 'task analysis', and 'computation modeling' as the hot research topics in the past three years. The current research power on CPS and CSCL is mainly distributed in China and the United States, but the inter-institutional collaboration is little. Based on the research results, this paper proposes four future directions to improve the research in online collaborative problem-solving: identity awareness in online collaborative problem-solving, interdisciplinary online collaborative problem-solving, online collaborative problem-solving for teachers, and online collaborative problem-solving and ICT skills. [ABSTRACT FROM AUTHOR]

Scene-matching navigation is one of the essential technologies for achieving precise navigation in satellite-denied environments. Selecting suitable-matching areas is crucial for planning trajectory and reducing yaw. Most traditional selection methods of suitable-matching areas use hierarchical screening based on multiple feature indicators. However, these methods rarely consider the interrelationship between different feature indicators and use the same set of screening thresholds for different categories of images, which has poor versatility and can easily cause mis-selection and omission. To solve this problem, a suitable-matching areas' selection method based on multi-level saliency is proposed. The matching performance score is obtained by fusing several segmentation levels' salient feature extraction results and performing weighted calculations with the sub-image edge density. Compared with the hierarchical screening methods, the matching performance of the candidate areas selected by our algorithm is at least 22.2% higher, and it also has a better matching ability in different scene categories. In addition, the number of missed and wrong selections is significantly reduced. The average matching accuracy of the top three areas selected by our method reached 0.8549, 0.7993, and 0.7803, respectively, under the verification of multiple matching algorithms. Experimental results show this paper's suitable-matching areas' selection method is more robust. [ABSTRACT FROM AUTHOR]

Depth images obtained from lightweight, real-time depth estimation models and consumer-oriented sensors typically have low-resolution issues. Traditional interpolation methods for depth image up-sampling result in a significant information loss, especially in edges with discontinuous depth variations (depth discontinuities). To address this issue, this paper proposes a semi-coupled deformable convolution network (SCD-Net) based on the idea of guided depth map super-resolution (GDSR). The method employs a semi-coupled feature extraction scheme to learn unique and similar features between RGB images and depth images. We utilize a Coordinate Attention (CA) to suppress redundant information in RGB features. Finally, a deformable convolutional module is employed to restore the original resolution of the depth image. The model is tested on NYUv2, Middlebury, Lu, and a Real-Sense real-world dataset created using an Intel Real-sense D455 structured-light camera. The super-resolution accuracy of SCD-Net at multiple scales is much higher than that of traditional methods and superior to recent state-of-the-art (SOTA) models, which demonstrates the effectiveness and flexibility of our model on GDSR tasks. In particular, our method further solves the problem of an RGB texture being over-transferred in GDSR tasks. [ABSTRACT FROM AUTHOR]

Zero-shot sketch-based image retrieval (ZS-SBIR) is an important computer vision problem. The image category in the test phase is a new category that was not visible in the training stage. Because sketches are extremely abstract, the commonly used backbone networks (such as VGG-16 and ResNet-50) cannot handle both sketches and photos. Semantic similarities between the same features in photos and sketches are difficult to reflect in deep models without textual assistance. To solve this problem, we propose a novel and effective feature embedding model called Attention Map Feature Fusion (AMFF). The AMFF model combines the excellent feature extraction capability of the ResNet-50 network with the excellent representation ability of the attention network. By processing the residuals of the ResNet-50 network, the attention map is finally obtained without introducing external semantic knowledge. Most previous approaches treat the ZS-SBIR problem as a classification problem, which ignores the huge domain gap between sketches and photos. This paper proposes an effective method to optimize the entire network, called domain-aware triplets (DAT). Domain feature discrimination and semantic feature embedding can be learned through DAT. In this paper, we also use the classification loss function to stabilize the training process to avoid getting trapped in a local optimum. Compared with the state-of-the-art methods, our method shows a superior performance. For example, on the Tu-berlin dataset, we achieved 61.2 + 1.2% Prec200. On the Sketchy_c100 dataset, we achieved 62.3 + 3.3% mAPall and 75.5 + 1.5% Prec100. [ABSTRACT FROM AUTHOR]

Diesel engines have a wide range of functions in the industrial and military fields. An urgent problem to be solved is how to diagnose and identify their faults effectively and timely. In this paper, a diesel engine acoustic fault diagnosis method based on variational modal decomposition mapping Mel frequency cepstral coefficients (MFCC) and long-short-term memory network is proposed. Variational mode decomposition (VMD) is used to remove noise from the original signal and differentiate the signal into multiple modes. The sound pressure signals of different modes are mapped to the Mel filter bank in the frequency domain, and then the Mel frequency cepstral coefficients of the respective mode signals are calculated in the mapping range of frequency domain, and the optimized Mel frequency cepstral coefficients are used as the input of long and short time memory network (LSTM) which is trained and verified, and the fault diagnosis model of the diesel engine is obtained. The experimental part compares the fault diagnosis effects of different feature extraction methods, different modal decomposition methods and different classifiers, finally verifying the feasibility and effectiveness of the method proposed in this paper, and providing solutions to the problem of how to realise fault diagnosis using acoustic signals. [ABSTRACT FROM AUTHOR]

In order to solve the problem wherein too many base station antennas are deployed in a massive multiple-input–multiple-output system, resulting in high overhead for downlink channel state information feedback, this paper proposes an uplink-assisted channel feedback method based on deep learning. The method applies the reciprocity of the uplink and downlink, uses uplink channel state information in the base station to help users give feedback on unknown downlink information, and compresses and restores the channel state information. First, an encoder–decoder structure is established. The encoder reduces the network depth and uses multi-resolution convolution to increase the accuracy of channel state information extraction while reducing the number of computations relating to user equipment. Afterward, the channel state information is compressed to reduce feedback overhead in the channel. At the decoder, with the help of the reciprocity of the uplink and downlink, the feature extraction of the uplink's magnitudes is carried out, and the downlink channel state information is integrated into a channel state information feature matrix, which is restored to its original size. The simulation results show that compared with CSINet, CRNet, CLNet, and DCRNet, indoor reconstruction precision was improved by an average of 16.4%, and outside reconstruction accuracy was improved by an average of 21.2% under all compressions. [ABSTRACT FROM AUTHOR]

In MOOC learning, learners' emotions have an important impact on the learning effect. In order to solve the problem that learners' emotions are not obvious in the learning process, we propose a method to identify learner emotion by combining eye movement features and scene features. This method uses an adaptive window to partition samples and enhances sample features through fine-grained feature extraction. Using an adaptive window to partition samples can make the eye movement information in the sample more abundant, and fine-grained feature extraction from an adaptive window can increase discrimination between samples. After adopting the method proposed in this paper, the four-category emotion recognition accuracy of the single modality of eye movement reached 65.1% in MOOC learning scenarios. Both the adaptive window partition method and the fine-grained feature extraction method based on eye movement signals proposed in this paper can be applied to other modalities. [ABSTRACT FROM AUTHOR]

Chip pad inspection is of great practical importance for chip alignment inspection and correction. It is one of the key technologies for automated chip inspection in semiconductor manufacturing. When applying deep learning methods for chip pad inspection, the main problem to be solved is how to ensure the accuracy of small target pad detection and, at the same time, achieve a lightweight inspection model. The attention mechanism is widely used to improve the accuracy of small target detection by finding the attention region of the network. However, conventional attention mechanisms capture feature information locally, which makes it difficult to effectively improve the detection efficiency of small targets from complex backgrounds in target detection tasks. In this paper, an OCAM (Object Convolution Attention Module) attention module is proposed to build long-range dependencies between channel features and position features by constructing feature contextual relationships to enhance the correlation between features. By adding the OCAM attention module to the feature extraction layer of the YOLOv5 network, the detection performance of chip pads is effectively improved. In addition, a design guideline for the attention layer is proposed in the paper. The attention layer is adjusted by network scaling to avoid network characterization bottlenecks, balance network parameters, and network detection performance, and reduce the hardware device requirements for the improved YOLOv5 network in practical scenarios. Extensive experiments on chip pad datasets, VOC datasets, and COCO datasets show that the approach in this paper is more general and superior to several state-of-the-art methods. [ABSTRACT FROM AUTHOR]

The automatic parking system based on vision is greatly affected by uneven lighting, which is difficult to make an accurate judgment on parking spaces in the case of complex image information. To solve this problem, this paper proposes a parking space visual detection and image processing method based on deep learning. Firstly, a 360-degree panoramic system was designed to photograph the vehicle environment. The image has been processed to obtain a panoramic aerial view, which was input as the original image of the parking space detection system. Secondly, the Faster R-CNN (Region-Convolutional Neural Network) parking detection model was established based on deep learning. It was aimed to detect and extract the parking space from the input image. Thirdly, the problems of uneven illumination and complex background were solved effectively by removing the background light from the image. Finally, a parking space extraction method based on the connected region has been designed, which further simplified the parking space extraction and image processing. The experiment results show that the mAP (mean Average Precision) value of the Faster R-CNN model using 101-Floor ResNet as the feature extraction network is 89.30%, which is 2.28% higher than that of the Faster R-CNN model using 50-Floor ResNet as the feature extraction network. The model built in this paper can detect most parking spaces well. The position of the output target box is accurate. In some test scenarios, the confidence of parking space recognition can even reach 100%. In summary, the proposed method can realize the effective identification and accurate positioning of parking spaces. [ABSTRACT FROM AUTHOR]

Channel state information (CSI) provides a fine-grained description of the signal propagation process, which has attracted extensive attention in the field of indoor positioning. The CSI signals collected by different fingerprint points have a high degree of discrimination due to the influence of multi-path effects. This multi-path effect is reflected in the correlation between subcarriers and antennas. However, in mining such correlations, previous methods are difficult to aggregate non-adjacent features, resulting in insufficient multi-path information extraction. In addition, the existence of the multi-path effect makes the relationship between the original CSI signal and the distance not obvious, and it is easy to cause mismatching of long-distance points. Therefore, this paper proposes an indoor localization algorithm that combines the multi-head self-attention mechanism and effective CSI (MHSA-EC). This algorithm is used to solve the problem where it is difficult for traditional algorithms to effectively aggregate long-distance CSI features and mismatches of long-distance points. This paper verifies the stability and accuracy of MHSA-EC positioning through a large number of experiments. The average positioning error of MHSA-EC is 0.71 m in the comprehensive office and 0.64 m in the laboratory. [ABSTRACT FROM AUTHOR]

Semantic segmentation is an important task for the interpretation of remote sensing images. Remote sensing images are large in size, contain substantial spatial semantic information, and generally exhibit strong symmetry, resulting in images exhibiting large intraclass variance and small interclass variance, thus leading to class imbalance and poor small-object segmentation. In this paper, we propose a new remote sensing image semantic segmentation network, called CAS-Net, which includes coordinate attention (CA) and SPD-Conv. In the model, we replace stepwise convolution with SPD-Conv convolution in the feature extraction network and add a pooling layer into the network to avoid the loss of detailed information, effectively improving the segmentation of small objects. The CA is introduced into the atrous spatial pyramid pooling (ASPP) module, thus improving the recognizability of classified objects and target localization accuracy in remote sensing images. Finally, the Dice coefficient was introduced into the cross-entropy loss function to maximize the gradient optimization of the model and solve the classification imbalance problem in the image. The proposed model is compared with several state-of-the-art models on the ISPRS Vaihingen dataset. The experimental results demonstrate that the proposed model significantly optimizes the segmentation effect of small objects in remote sensing images, effectively solves the problem of class imbalance in the dataset, and improves segmentation accuracy. [ABSTRACT FROM AUTHOR]

Remote sensing change detection involves detecting pixels that have changed from a bi-temporal image of the same location. Current mainstream change detection models use encoder-decoder structures as well as Siamese networks. However, there are still some challenges with this: (1) Existing change feature fusion approaches do not take into account the symmetry of change features, which leads to information loss; (2) The encoder is independent of the change detection task, and feature extraction is performed separately for dual-time images, which leads to underutilization of the encoder parameters; (3) There are problems of unbalanced positive and negative samples and bad edge region detection. To solve the above problems, a mutual feature-aware network (MFNet) is proposed in this paper. Three modules are proposed for the purpose: (1) A symmetric change feature fusion module (SCFM), which uses double-branch feature selection without losing feature information and focuses explicitly on focal spatial regions based on cosine similarity to introduce strong a priori information; (2) A mutual feature-aware module (MFAM), which introduces change features in advance at the encoder stage and uses a cross-type attention mechanism for long-range dependence modeling; (3) A loss function for edge regions. After detailed experiments, the F1 scores of MFNet on SYSU-CD and LEVIR-CD were 83.11% and 91.52%, respectively, outperforming several advanced algorithms, demonstrating the effectiveness of the proposed method. [ABSTRACT FROM AUTHOR]

The grade of wheat quality depends on the proportion of unsound kernels. Therefore, the rapid detection of unsound wheat kernels is important for wheat rating and evaluation. However, in practice, unsound kernels are hand-picked, which makes the process time-consuming and inefficient. Meanwhile, methods based on traditional image processing cannot divide adherent particles well. To solve the above problems, this paper proposed an unsound wheat kernel recognition algorithm based on an improved mask RCNN. First, we changed the feature pyramid network (FPN) to a bottom-up pyramid network to strengthen the low-level information. Then, an attention mechanism (AM) module was added between the feature extraction network and the pyramid network to improve the detection accuracy for small targets. Finally, the regional proposal network (RPN) was optimized to improve the prediction performance. Experiments showed that the improved mask RCNN algorithm could identify the unsound kernels more quickly and accurately while handling adhesion problems well. The precision and recall were 86% and 91%, respectively, and the inference time on the test set with about 200 targets for each image was 7.83 s. Additionally, we compared the improved model with other existing segmentation models, and experiments showed that our model achieved higher accuracy and performance than the other models, laying the foundation for wheat grading. [ABSTRACT FROM AUTHOR]

Besides the many advances made in the facial detection and recognition fields, face recognition applied to visual images (VIS-FR) has received increasing interest in recent years, especially in the field of communication, identity authentication, public safety and to address the risk of terrorism and crime. These systems however encounter important problems in the presence of variations in pose, expression, age, occlusion, disguise, and lighting as these factors significantly reduce the recognition accuracy. To prevent problems in the visible spectrum, several researchers have recommended the use of infrared images. This paper provides an updated overview of deep infrared (IR) approaches in face recognition (FR) and analysis. First, we present the most widely used databases, both public and private, and the various metrics and loss functions that have been proposed and used in deep infrared techniques. We then review deep face analysis and recognition/identification methods proposed in recent years. In this review, we show that infrared techniques have given interesting results for face recognition, solving some of the problems encountered with visible spectrum techniques. We finally identify some weaknesses of current infrared FR approaches as well as many future research directions to address the IR FR limitations. [ABSTRACT FROM AUTHOR]

In order to solve the problem of an accurate recognition of tea picking through tea picking robots, an edge device detection method is proposed in this paper based on ShuffleNetv2-YOLOv5-Lite-E for tea with one bud and two leaves. This replaces the original feature extraction network by removing the Focus layer and using the ShuffleNetv2 algorithm, followed by a channel pruning of YOLOv5 at the neck layer head, thus achieving the purpose of reducing the model size. The results show that the size of the improved generated weight file is 27% of that of the original YOLOv5 model, and the mAP value of ShuffleNetv2-YOLOv5-Lite-E is 97.43% and 94.52% on the pc and edge device respectively, which are 1.32% and 1.75% lower compared to that of the original YOLOv5 model. The detection speeds of ShuffleNetv2-YOLOv5-Lite-E, YOLOv5, YOLOv4, and YOLOv3 were 8.6 fps, 2.7 fps, 3.2 fps, and 3.4 fps respectively after importing the models into an edge device, and the improved YOLOv5 detection speed was 3.2 times faster than that of the original YOLOv5 model. Through the detection method, the size of the original YOLOv5 model is effectively reduced while essentially ensuring recognition accuracy. The detection speed is also significantly improved, which is conducive to the realization of intelligent and accurate picking for future tea gardens, laying a solid foundation for the realization of tea picking robots. [ABSTRACT FROM AUTHOR]

To solve the problem of feature distribution discrepancy in cross-corpus speech emotion recognition tasks, this paper proposed an emotion recognition model based on multi-task learning and subdomain adaptation, which alleviates the impact on emotion recognition. Existing methods have shortcomings in speech feature representation and cross-corpus feature distribution alignment. The proposed model uses a deep denoising auto-encoder as a shared feature extraction network for multi-task learning, and the fully connected layer and softmax layer are added before each recognition task as task-specific layers. Subsequently, the subdomain adaptation algorithm of emotion and gender features is added to the shared network to obtain the shared emotion features and gender features of the source domain and target domain, respectively. Multi-task learning effectively enhances the representation ability of features, a subdomain adaptive algorithm promotes the migrating ability of features and effectively alleviates the impact of feature distribution differences in emotional features. The average results of six cross-corpus speech emotion recognition experiments show that, compared with other models, the weighted average recall rate is increased by 1.89~10.07%, the experimental results verify the validity of the proposed model. [ABSTRACT FROM AUTHOR]

To solve the problem of inadequate feature extraction by the model due to factors such as occlusion and illumination in person re-identification tasks, this paper proposed a model with a joint cross-consistency learning and multi-feature fusion person re-identification. The attention mechanism and the mixed pooling module were first embedded in the residual network so that the model adaptively focuses on the more valid information in the person images. Secondly, the dataset was randomly divided into two categories according to the camera perspective, and a feature classifier was trained for the two types of datasets respectively. Then, two classifiers with specific knowledge were used to guide the model to extract features unrelated to the camera perspective for the two types of datasets so that the obtained image features were endowed with domain invariance by the model, and the differences in the perspective, attitude, background, and other related information of different images were alleviated. Then, the multi-level features were fused through the feature pyramid to concern the more critical information of the image. Finally, a combination of Cosine Softmax loss, triplet loss, and cluster center loss was proposed to train the model to address the differences of multiple losses in the optimization space. The first accuracy of the proposed model reached 95.9% and 89.7% on the datasets Market-1501 and DukeMTMC-reID, respectively. The results indicated that the proposed model has good feature extraction capability. [ABSTRACT FROM AUTHOR]

In order to categorize feature classes by capturing subtle differences, hyperspectral images (HSIs) have been extensively used due to the rich spectral-spatial information. The 3D convolution-based neural networks (3DCNNs) have been widely used in HSI classification because of their powerful feature extraction capability. However, the 3DCNN-based HSI classification approach could only extract local features, and the feature maps it produces include a lot of spatial information redundancy, which lowers the classification accuracy. To solve the above problems, we proposed a spatial attention network (SATNet) by combining 3D OctConv and ViT. Firstly, 3D OctConv divided the feature maps into high-frequency maps and low-frequency maps to reduce spatial information redundancy. Secondly, the ViT model was used to obtain global features and effectively combine local-global features for classification. To verify the effectiveness of the method in the paper, a comparison with various mainstream methods on three publicly available datasets was performed, and the results showed the superiority of the proposed method in terms of classification evaluation performance. [ABSTRACT FROM AUTHOR]

In the early named entity recognition models, most text processing focused only on the representation of individual words and character vectors, and paid little attention to the semantic relationships between the preceding and following text in an utterance, which led to the inability to handle the problem of multiple meanings of a word during recognition. To address this problem, most models introduce the attention mechanism of Transformer model to solve the problem of multiple meanings of a word in text. However, the traditional Transformer model leads to a high computational overhead due to its fully connected structure. Therefore, this paper proposes a new model, the BERT-Star-Transformer-CNN-BiLSTM-CRF model, to solve the problem of the computational efficiency of the traditional Transformer. First, the input text is dynamically generated into a character vector using the BERT model pre-trained in large-scale preconditioning to solve the problem of multiple meanings of words, and then the lightweight Star-Transformer model is used as the feature extraction module to perform local feature extraction on the word vector sequence, while the CNN-BiLSTM joint model is used to perform global feature extraction on the context in the text. The obtained feature sequences are fused. Finally, the fused feature vector sequences are input to CRF for prediction of the final results. After the experiments, it is shown that the model has a significant improvement in precision, recall and F1 value compared with the traditional model, and the computational efficiency is improved by nearly 40%. [ABSTRACT FROM AUTHOR]

Objective: The shallow underwater environment is complex, with problems of color shift, uneven illumination, blurring, and distortion in the imaging process. These scenes are very unfavorable for the reasoning of the detection network. Additionally, typical object identification algorithms struggle to maintain high resilience in underwater environments due to picture domain offset, making underwater object detection problematic. Methods: This paper proposes a single-stage detection method with the double enhancement of anchor boxes and features. The feature context relevance is improved by proposing a composite-connected backbone network. The receptive field enhancement module is introduced to enhance the multi-scale detection capability. Finally, a prediction refinement strategy is proposed, which refines the anchor frame and features through two regressions, solves the problem of feature anchor frame misalignment, and improves the detection performance of the single-stage underwater algorithm. Results: We achieved an effect of 80.2 mAP on the Labeled Fish in the Wild dataset, which saves some computational resources and time while still improving accuracy. On the original basis, UWNet can achieve 2.1 AP accuracy improvement due to the powerful feature extraction function and the critical role of multi-scale functional modules. At an input resolution of 300 × 300, UWNet can provide an accuracy of 32.4 AP. When choosing the number of prediction layers, the accuracy of the four and six prediction layer structures is compared. The experiments show that on the Labeled Fish in the Wild dataset, the six prediction layers are better than the four. Conclusion: The single-stage underwater detection model UWNet proposed in this research has a double anchor frame and feature optimization. By adding three functional modules, the underwater detection of the single-stage detector is enhanced to address the issue that it is simple to miss detection while detecting small underwater targets. [ABSTRACT FROM AUTHOR]

Hyperspectral image classification methods based on deep learning have led to remarkable achievements in recent years. However, these methods with outstanding performance are also accompanied by problems such as excessive dependence on the number of samples, poor model generalization, and time-consuming training. Additionally, the previous patch-level feature extraction methods have some limitations, for instance, non-local information is difficult to model, etc. To solve these problems, this paper proposes an image-level feature extraction method with transfer learning. Firstly, we look at a hyperspectral image with hundreds of contiguous spectral bands from a sequential image perspective. We attempt to extract the global spectral variation information between adjacent spectral bands by using the optical flow estimation method. Secondly, we propose an innovative data adaptation strategy to bridge the gap between hyperspectral and video data, and transfer the optical flow estimation network pre-trained with video data to the hyperspectral feature extraction task for the first time. Thirdly, we utilize the traditional classifier to achieve classification. Simultaneously, a vote strategy combined with features at different scales is proposed to improve the classification accuracy further. Extensive, well-designed experiments on four scenes of public hyperspectral images demonstrate that the proposed method (Spe-TL) can obtain results that are competitive with advanced deep learning methods under various sample conditions, with better time effectiveness to adapt to new target tasks. Moreover, it can produce more detailed classification maps that subtly reflect the authentic distribution of ground objects in the original image. [ABSTRACT FROM AUTHOR]

The depth estimation algorithm based on the convolutional neural network has many limitations and defects by constructing matching cost volume to calculate the disparity: using a limited disparity range, the authentic disparity beyond the predetermined range can not be acquired; Besides, the matching process lacks constraints on occlusion and matching uniqueness; Also, as a local feature extractor, a convolutional neural network lacks the ability of global context information perception. Aiming at the problems in the matching method of constructing matching cost volume, we propose a disparity prediction algorithm based on Transformer, which specifically comprises the Swin-SPP module for feature extraction based on Swin Transformer, Transformer disparity matching network based on self-attention and cross-attention mechanism, and occlusion prediction sub-network. In addition, we propose a double skip connection fully connected layer to solve the problems of gradient vanishing and explosion during the training process for the Transformer model, thus further enhancing inference accuracy. The proposed model in this paper achieved an EPE (Absolute error) of 0.57 and 0.61, and a 3PE (Percentage error greater than 3 px) of 1.74% and 1.56% on KITTI 2012 and KITTI 2015 datasets, respectively, with an inference time of 0.46 s and parameters as low as only 2.6 M, showing great advantages compared with other algorithms in various evaluation metrics. [ABSTRACT FROM AUTHOR]

In order to solve the problem in which structurally incoherent low-rank non-negative matrix decomposition (SILR-NMF) algorithms only consider the non-negativity of the data and do not consider the manifold distribution of high-dimensional space data, a new structurally incoherent low rank two-dimensional local discriminant graph embedding (SILR-2DLDGE) is proposed in this paper. The algorithm consists of the following three parts. Firstly, it is vital to keep the intrinsic relationship between data points. By the token, we introduced the graph embedding (GE) framework to preserve locality information. Secondly, the algorithm alleviates the impact of noise and corruption uses the L1 norm as a constraint by low-rank learning. Finally, the algorithm improves the discriminant ability by encrypting the structurally incoherent parts of the data. In the meantime, we capture the theoretical basis of the algorithm and analyze the computational cost and convergence. The experimental results and discussions on several image databases show that the proposed algorithm is more effective than the SILR-NMF algorithm. [ABSTRACT FROM AUTHOR]

At present, the UAV (Unmanned Aerial Vehicle) has been widely used both in civilian and military fields. Most of the current object detection algorithms used to detect UAVs require more parameters, and it is difficult to achieve real-time performance. In order to solve this problem while ensuring a high accuracy rate, we further lighten the model and reduce the number of parameters of the model. This paper proposes an accurate and lightweight UAV detection model based on YOLOv4. To verify the effectiveness of this model, we made a UAV dataset, which contains four types of UAVs and 20,365 images. Through comparative experiments and optimization of existing deep learning and object detection algorithms, we found a lightweight model to achieve an efficient and accurate rapid detection of UAVs. First, from the comparison of the one-stage method and the two-stage method, it is concluded that the one-stage method has better real-time performance and considerable accuracy in detecting UAVs. Then, we further compared the one-stage methods. In particular, for YOLOv4, we replaced MobileNet with its backbone network, modified the feature extraction network, and replaced standard convolution with depth-wise separable convolution, which greatly reduced the parameters and realized 82 FPS and 93.52% mAP while ensuring high accuracy and taking into account the real-time performance. [ABSTRACT FROM AUTHOR]

In recent years, jellyfish outbreaks have frequently occurred in offshore areas worldwide, posing a significant threat to the marine fishery, tourism, coastal industry, and personal safety. Effective monitoring of jellyfish is a vital method to solve the above problems. However, the optical detection method for jellyfish is still in the primary stage. Therefore, this paper studies a jellyfish detection method based on convolution neural network theory and digital image processing technology. This paper studies the underwater image preprocessing algorithm because the quality of underwater images directly affects the detection results. The results show that the image quality is better after applying the three algorithms namely prior defogging, adaptive histogram equalization, and multi-scale retinal enhancement, which is more conducive to detection. We establish a data set containing seven species of jellyfishes and fish. A total of 2141 images are included in the data set. The YOLOv3 algorithm is used to detect jellyfish, and its feature extraction network Darknet53 is optimized to ensure it is conducted in real-time. In addition, we introduce label smoothing and cosine annealing learning rate methods during the training process. The experimental results show that the improved algorithms improve the detection accuracy of jellyfish on the premise of ensuring the detection speed. This paper lays a foundation for the construction of an underwater jellyfish optical imaging real-time monitoring system. [ABSTRACT FROM AUTHOR]

In the last years, methods for detecting text in real scenes have made significant progress with an increase in neural networks. However, due to the limitation of the receptive field of the central nervous system and the simple representation of text by using rectangular bounding boxes, the previous methods may be insufficient for working with more challenging instances of text. To solve this problem, this paper proposes a scene text detection network based on cross-scale feature fusion (CSFF-Net). The framework is based on the lightweight backbone network Resnet, and the feature learning is enhanced by embedding the depth weighted convolution module (DWCM) while retaining the original feature information extracted by CNN. At the same time, the 3D-Attention module is also introduced to merge the context information of adjacent areas, so as to refine the features in each spatial size. In addition, because the Feature Pyramid Network (FPN) cannot completely solve the interdependence problem by simple element-wise addition to process cross-layer information flow, this paper introduces a Cross-Level Feature Fusion Module (CLFFM) based on FPN, which is called Cross-Level Feature Pyramid Network (Cross-Level FPN). The proposed CLFFM can better handle cross-layer information flow and output detailed feature information, thus improving the accuracy of text region detection. Compared to the original network framework, the framework provides a more advanced performance in detecting text images of complex scenes, and extensive experiments on three challenging datasets validate the realizability of our approach. [ABSTRACT FROM AUTHOR]

Bearing is one of the most important parts of rotating machinery with high failure rate, and its working state directly affects the performance of the entire equipment. Hence, it is of great significance to diagnose bearing faults, which can contribute to guaranteeing running stability and maintenance, thus promoting production efficiency and economic benefits. Usually, the bearing fault features are difficult to extract effectively, which results in low diagnosis performance. To solve the problem, this paper proposes a bearing fault feature extraction method and it establishes a bearing fault diagnosis method that is based on feature fusion. The basic idea of the method is as follows: firstly, the time-frequency feature of the bearing signal is extracted through Wavelet Packet Transform (WPT) to form the time-frequency characteristic matrix of the signal; secondly, the Multi-Weight Singular Value Decomposition (MWSVD) is constructed by singular value contribution rate and entropy weight. The features of the time-frequency feature matrix obtained by WPT are further extracted, and the features that are sensitive to fault in the time-frequency feature matrix are retained while the insensitive features are removed; finally, the extracted feature matrix is used as the input of the Support Vector Machine (SVM) classifier for bearing fault diagnosis. The proposed method is validated by data sets from the time-varying bearing data from the University of Ottawa and Case Western Reserve University Bearing Data Center. The results show that the algorithm can effectively diagnose the bearing under the steady-state and unsteady state. This paper proposes that the algorithm has better fault diagnosis capabilities and feature extraction capabilities when compared with methods that aree based on traditional feature technology. [ABSTRACT FROM AUTHOR]

Landslides are the most common and destructive secondary geological hazards caused by earthquakes. It is difficult to extract landslides automatically based on remote sensing data, which is import for the scenario of disaster emergency rescue. The literature review showed that the current landslides extraction methods mostly depend on expert interpretation which was low automation and thus was unable to provide sufficient information for earthquake rescue in time. To solve the above problem, an end-to-end improved Mask R-CNN model was proposed. The main innovations of this paper were (1) replacing the feature extraction layer with an effective ResNeXt module to extract the landslides. (2) Increasing the bottom-up channel in the feature pyramid network to make full use of low-level positioning and high-level semantic information. (3) Adding edge losses to the loss function to improve the accuracy of the landslide boundary detection accuracy. At the end of this paper, Jiuzhaigou County, Sichuan Province, was used as the study area to evaluate the new model. Results showed that the new method had a precision of 95.8%, a recall of 93.1%, and an overall accuracy (OA) of 94.7%. Compared with the traditional Mask R-CNN model, they have been significantly improved by 13.9%, 13.4%, and 9.9%, respectively. It was proved that the new method was effective in the landslides automatic extraction. [ABSTRACT FROM AUTHOR]

Water area segmentation is an important branch of remote sensing image segmentation, but in reality, most water area images have complex and diverse backgrounds. Traditional detection methods cannot accurately identify small tributaries due to incomplete mining and insufficient utilization of semantic information, and the edge information of segmentation is rough. To solve the above problems, we propose a multi-scale feature aggregation network. In order to improve the ability of the network to process boundary information, we design a deep feature extraction module using a multi-scale pyramid to extract features, combined with the designed attention mechanism and strip convolution, extraction of multi-scale deep semantic information and enhancement of spatial and location information. Then, the multi-branch aggregation module is used to interact with different scale features to enhance the positioning information of the pixels. Finally, the two high-performance branches designed in the Feature Fusion Upsample module are used to deeply extract the semantic information of the image, and the deep information is fused with the shallow information generated by the multi-branch module to improve the ability of the network. Global and local features are used to determine the location distribution of each image category. The experimental results show that the accuracy of the segmentation method in this paper is better than that in the previous detection methods, and has important practical significance for the actual water area segmentation. [ABSTRACT FROM AUTHOR]

Point cloud registration is a key step in the reconstruction of 3D data models. The traditional ICP registration algorithm depends on the initial position of the point cloud. Otherwise, it may get trapped into local optima. In addition, the registration method based on the feature learning of PointNet cannot directly or effectively extract local features. To solve these two problems, this paper proposes SAP-Net, inspired by CorsNet and PointNet++, as an optimized CorsNet. To be more specific, SAP-Net firstly uses the set abstraction layer in PointNet++ as the feature extraction layer and then combines the global features with the initial template point cloud. Finally, PointNet is used as the transform prediction layer to obtain the six parameters required for point cloud registration directly, namely the rotation matrix and the translation vector. Experiments on the ModelNet40 dataset and real data show that SAP-Net not only outperforms ICP and CorsNet on both seen and unseen categories of the point cloud but also has stronger robustness. [ABSTRACT FROM AUTHOR]

Performing predictive maintenance (PdM) is challenging for many reasons. Dealing with large datasets which may not contain run-to-failure data (R2F) complicates PdM even more. When no R2F data are available, identifying condition indicators (CIs), estimating the health index (HI), and thereafter, calculating a degradation model for predicting the remaining useful lifetime (RUL) are merely impossible using supervised learning. In this paper, a 3 DoF delta robot used for pick and place task is studied. In the proposed method, autoencoders (AEs) are used to predict when maintenance is required based on the signal sequence distribution and anomaly detection, which is vital when no R2F data are available. Due to the sequential nature of the data, nonlinearity of the system, and correlations between parameter time-series, convolutional layers are used for feature extraction. Thereafter, a sigmoid function is used to predict the probability of having an anomaly given CIs acquired from AEs. This function can be manually tuned given the sensitivity of the system or optimized by solving a minimax problem. Moreover, the proposed architecture can be used for fault localization for the specified system. Additionally, the proposed method can calculate RUL using Gaussian process (GP), as a degradation model, given HI values as its input. [ABSTRACT FROM AUTHOR]

To improve the recognition rate of lower limb actions based on surface electromyography (sEMG), an effective weighted feature method is proposed, and an improved genetic algorithm support vector machine (IGA-SVM) is designed in this paper. First, for the problem of high feature redundancy and low discrimination in the surface electromyography feature extraction process, the weighted feature method is proposed based on the correlation between muscles and actions. Second, to solve the problem of the genetic algorithm selection operator easily falling into a local optimum solution, the improved genetic algorithm-support vector machine is designed by championship with sorting method. Finally, the proposed method is used to recognize six types of lower limb actions designed, and the average recognition rate reaches 94.75%. Experimental results indicate that the proposed method has definite potentiality in lower limb action recognition. [ABSTRACT FROM AUTHOR]

Data-driven chatter detection techniques avoid complex physical modeling and provide the basis for industrial applications of cutting process monitoring. Among them, feature extraction is the key step of chatter detection, which can compensate for the accuracy disadvantage of machine learning algorithms to some extent if the extracted features are highly correlated with the milling condition. However, the classification accuracy of the current feature extraction methods is not satisfactory, and a combination of multiple features is required to identify the chatter. This limits the development of unsupervised machine learning algorithms for chattering detection, which further affects the application in practical processing. In this paper, the fractal feature of the signal is extracted by structure function method (SFM) for the first time, which solves the problem that the features are easily affected by process parameters. Milling chatter is identified based on k-means algorithm, which avoids the complex process of training model, and the judgment method of milling chatter is also discussed. The proposed method can achieve 94.4% identification accuracy by using only one single signal feature, which is better than other feature extraction methods, and even better than some supervised machine learning algorithms. Moreover, experiments show that chatter will affect the distribution of cutting bending moment, and it is not reliable to monitor tool wear through the polar plot of the bending moment. This provides a theoretical basis for the application of unsupervised machine learning algorithms in chatter detection. [ABSTRACT FROM AUTHOR]

A sensor registry system (SRS) registers sensor metadata and provides them for a seamless semantic process. Recently, network coverage information-based SRS (NC-SRS) was developed to provide sensor information filtering by combining path prediction and network coverage checks. However, the NC-SRS has problems caused by issues such as termination of OpenSignal service and pre-building road segments. Therefore, this paper proposes a sensor registry system-based predictive information service (SRS-PIS) using a grid. SRS-PIS predicts a path based on the grid, checks the network coverage, and filters the sensor. This paper presents a grid-based real-time path prediction algorithm and an algorithm for grouping network service-disabled areas. To obtain network coverage information, we constructed and implemented a grid-based coverage map through experiment to measure the signal strength. As an evaluation, we compared features among SRS-based systems and SRS-PIS, and compared advantages and disadvantages between segment-based and grid-based methods. [ABSTRACT FROM AUTHOR]

Desert segmentation of remote sensing images is the basis of analysis of desert area. Desert images are usually characterized by large image size, large-scale change, and irregular location distribution of surface objects. The multi-scale fusion method is widely used in the existing deep learning segmentation models to solve the above problems. Based on the idea of multi-scale feature extraction, this paper took the segmentation results of each scale as an independent optimization task and proposed a multi-resolution supervision network (MrsSeg) to further improve the desert segmentation result. Due to the different optimization difficulty of each branch task, we also proposed an auxiliary adaptive weighted loss function (AWL) to automatically optimize the training process. MrsSeg first used a lightweight backbone to extract different-resolution features, then adopted a multi-resolution fusion module to fuse the local information and global information, and finally, a multi-level fusion decoder was used to aggregate and merge the features at different levels to get the desert segmentation result. In this method, each branch loss was treated as an independent task, AWL was proposed to calculate and adjust the weight of each branch. By giving priority to the easy tasks, the improved loss function could effectively improve the convergence speed of the model and the desert segmentation result. The experimental results showed that MrsSeg-AWL effectively improved the learning ability of the model and has faster convergence speed, lower parameter complexity, and more accurate segmentation results. [ABSTRACT FROM AUTHOR]

In recent years, with the rapid development of computer vision, increasing attention has been paid to remote sensing image scene classification. To improve the classification performance, many studies have increased the depth of convolutional neural networks (CNNs) and expanded the width of the network to extract more deep features, thereby increasing the complexity of the model. To solve this problem, in this paper, we propose a lightweight convolutional neural network based on attention-oriented multi-branch feature fusion (AMB-CNN) for remote sensing image scene classification. Firstly, we propose two convolution combination modules for feature extraction, through which the deep features of images can be fully extracted with multi convolution cooperation. Then, the weights of the feature are calculated, and the extracted deep features are sent to the attention mechanism for further feature extraction. Next, all of the extracted features are fused by multiple branches. Finally, depth separable convolution and asymmetric convolution are implemented to greatly reduce the number of parameters. The experimental results show that, compared with some state-of-the-art methods, the proposed method still has a great advantage in classification accuracy with very few parameters. [ABSTRACT FROM AUTHOR]

To solve the problems of low accuracy, low real-time performance, poor robustness and others caused by the complex environment, this paper proposes a face mask recognition and standard wear detection algorithm based on the improved YOLO-v4. Firstly, an improved CSPDarkNet53 is introduced into the trunk feature extraction network, which reduces the computing cost of the network and improves the learning ability of the model. Secondly, the adaptive image scaling algorithm can reduce computation and redundancy effectively. Thirdly, the improved PANet structure is introduced so that the network has more semantic information in the feature layer. At last, a face mask detection data set is made according to the standard wearing of masks. Based on the object detection algorithm of deep learning, a variety of evaluation indexes are compared to evaluate the effectiveness of the model. The results of the comparations show that the mAP of face mask recognition can reach 98.3% and the frame rate is high at 54.57 FPS, which are more accurate compared with the exiting algorithm. [ABSTRACT FROM AUTHOR]

Deep learning models have strong abilities in learning features and they have been successfully applied in hyperspectral images (HSIs). However, the training of most deep learning models requires labeled samples and the collection of labeled samples are labor-consuming in HSI. In addition, single-level features from a single layer are usually considered, which may result in the loss of some important information. Using multiple networks to obtain multi-level features is a solution, but at the cost of longer training time and computational complexity. To solve these problems, a novel unsupervised multi-level feature extraction framework that is based on a three dimensional convolutional autoencoder (3D-CAE) is proposed in this paper. The designed 3D-CAE is stacked by fully 3D convolutional layers and 3D deconvolutional layers, which allows for the spectral-spatial information of targets to be mined simultaneously. Besides, the 3D-CAE can be trained in an unsupervised way without involving labeled samples. Moreover, the multi-level features are directly obtained from the encoded layers with different scales and resolutions, which is more efficient than using multiple networks to get them. The effectiveness of the proposed multi-level features is verified on two hyperspectral data sets. The results demonstrate that the proposed method has great promise in unsupervised feature learning and can help us to further improve the hyperspectral classification when compared with single-level features. [ABSTRACT FROM AUTHOR]

Owing to the increased use of urban rail transit, the flow of passengers on metro platforms tends to increase sharply during peak periods. Monitoring passenger flow in such areas is important for security-related reasons. In this paper, in order to solve the problem of metro platform passenger flow detection, we propose a CNN (convolutional neural network)-based network called the MP (metro platform)-CNN to accurately count people on metro platforms. The proposed method is composed of three major components: a group of convolutional neural networks is used on the front end to extract image features, a multiscale feature extraction module is used to enhance multiscale features, and transposed convolution is used for upsampling to generate a high-quality density map. Currently, existing crowd-counting datasets do not adequately cover all of the challenging situations considered in this study. Therefore, we collected images from surveillance videos of a metro platform to form a dataset containing 627 images, with 9243 annotated heads. The results of the extensive experiments showed that our method performed well on the self-built dataset and the estimation error was minimum. Moreover, the proposed method could compete with other methods on four standard crowd-counting datasets. [ABSTRACT FROM AUTHOR]

When early failures in rolling bearings occur, we need to be able to extract weak fault characteristic frequencies under the influence of strong noise and then perform fault diagnosis. Therefore, a new method is proposed: complete ensemble intrinsic time-scale decomposition with adaptive Lévy noise (CEITDALN). This method solves the problem of the traditional complete ensemble intrinsic time-scale decomposition with adaptive noise (CEITDAN) method not being able to filter nonwhite noise in measured vibration signal noise. Therefore, in the method proposed in this paper, a noise model in the form of parameter-adjusted noise is used to replace traditional white noise. We used an optimization algorithm to adaptively adjust the model parameters, reducing the impact of nonwhite noise on the feature frequency extraction. The experimental results for the simulation and vibration signals of rolling bearings showed that the CEITDALN method could extract weak fault features more effectively than traditional methods. [ABSTRACT FROM AUTHOR]

Robot grasping is an important direction in intelligent robots. However, how to help robots grasp specific objects in multi-object scenes is still a challenging problem. In recent years, due to the powerful feature extraction capabilities of convolutional neural networks (CNN), various algorithms based on convolutional neural networks have been proposed to solve the problem of grasp detection. Different from anchor-based grasp detection algorithms, in this paper, we propose a keypoint-based scheme to solve this problem. We model an object or a grasp as a single point—the center point of its bounding box. The detector uses keypoint estimation to find the center point and regress to all other object attributes such as size, direction, etc. Experimental results demonstrate that the accuracy of this method is 74.3% in the multi-object grasp dataset VMRD, and the performance on the single-object scene Cornell dataset is competitive with the current state-of-the-art grasp detection algorithm. Robot experiments demonstrate that this method can help robots grasp the target in single-object and multi-object scenes with overall success rates of 94% and 87%, respectively. [ABSTRACT FROM AUTHOR]

With the widespread success of deep learning in the two-dimensional field, how to apply deep learning methods from two-dimensional to three-dimensional field has become a current research hotspot. Among them, the polygon mesh structure in the three-dimensional representation as a complex data structure provides an effective shape approximate representation for the three-dimensional object. Although the traditional method can extract the characteristics of the three-dimensional object through the graphical method, it cannot be applied to more complex objects. However, due to the complexity and irregularity of the mesh data, it is difficult to directly apply convolutional neural networks to 3D mesh data processing. Considering this problem, we propose a deep learning method based on a capsule network to effectively classify mesh data. We first design a polynomial convolution template. Through a sliding operation similar to a two-dimensional image convolution window, we directly sample on the grid surface, and use the window sampling surface as the minimum unit of calculation. Because a high-order polynomial can effectively represent a surface, we fit the approximate shape of the surface through the polynomial, use the polynomial parameter as the shape feature of the surface, and add the center point coordinates and normal vector of the surface as the pose feature of the surface. The feature is used as the feature vector of the surface. At the same time, to solve the problem of the introduction of a large number of pooling layers in traditional convolutional neural networks, the capsule network is introduced. For the problem of nonuniform size of the input grid model, the capsule network attitude parameter learning method is improved by sharing the weight of the attitude matrix. The amount of model parameters is reduced, and the training efficiency of the 3D mesh model is further improved. The experiment is compared with the traditional method and the latest two methods on the SHREC15 data set. Compared with the MeshNet and MeshCNN, the average recognition accuracy in the original test set is improved by 3.4% and 2.1%, and the average after fusion of features the accuracy reaches 93.8%. At the same time, under the premise of short training time, this method can also achieve considerable recognition results through experimental verification. The three-dimensional mesh classification method proposed in this paper combines the advantages of graphics and deep learning methods, and effectively improves the classification effect of 3D mesh model. [ABSTRACT FROM AUTHOR]

Detailed information regarding land utilization/cover is a valuable resource in various fields. In recent years, remote sensing images, especially aerial images, have become higher in resolution and larger span in time and space, and the phenomenon that the objects in an identical category may yield a different spectrum would lead to the fact that relying on spectral features only is often insufficient to accurately segment the target objects. In convolutional neural networks, down-sampling operations are usually used to extract abstract semantic features, which leads to loss of details and fuzzy edges. To solve these problems, the paper proposes a Multi-level Feature Aggregation Network (MFANet), which is improved in two aspects: deep feature extraction and up-sampling feature fusion. Firstly, the proposed Channel Feature Compression module extracts the deep features and filters the redundant channel information from the backbone to optimize the learned context. Secondly, the proposed Multi-level Feature Aggregation Upsample module nestedly uses the idea that high-level features provide guidance information for low-level features, which is of great significance for positioning the restoration of high-resolution remote sensing images. Finally, the proposed Channel Ladder Refinement module is used to refine the restored high-resolution feature maps. Experimental results show that the proposed method achieves state-of-the-art performance 86.45% mean IOU on LandCover dataset. [ABSTRACT FROM AUTHOR]

Due to the large number of grid connection of distributed power supply, the existing scheduling methods can not meet the demand gradually. The proposed virtual power plant provides a new idea to solve this problem. The photovoltaic power prediction provides the data basis for the scheduling of the virtual power plant. Prediction intervals of photovoltaic power is a powerful statistical tool used for quantifying the uncertainty of photovoltaic power generation in power systems. To improve the interval prediction accuracy during the non-stationary periods of photovoltaic power, this paper proposes a probabilistic ensemble prediction model, which combines the modules of data preprocessing, non-stationary period discrimination, feature extraction, deterministic prediction, uncertainty prediction, and optimization integration into a general framework. More specifically, in the non-stationary period discrimination module, the method of discriminating the difference of the power ratio difference is introduced and applied for identifying the non-stationary period of the data of photovoltaic output; in the deterministic point prediction module, a stacking- long-short-term memory neural network model is used for point forecasts; in the uncertainty interval prediction module, a BAYES neural network is introduced for probabilistic forecasts; in the optimization integration module, an optimization algorithm named Non-dominated Sorting Genetic Algorithm-II is applied for integrating and optimizing the results of the point forecast and probabilistic forecast. The proposed model is tested using two photovoltaic outputs and weather data measured from a grid-connected photovoltaic system. The results show that the proposed model outperforms conventional forecast methods to predict short-term photovoltaic power outputs and associated uncertainties. The interval width is reduced by 10–20%, and the prediction accuracy is improved by at least 10%; this can be a useful tool for photovoltaic power forecasting. [ABSTRACT FROM AUTHOR]

The turn domain resampling (TDR) method is proposed in the paper on the basis of the existing angle domain resampling for solving the problem of non-fixed fault frequency under variable working conditions. TDR can select the appropriate sampling order according to the influence of frequency conversion, which avoided the error caused by the spline interpolation method. It can provide accurate parameters for the subsequent calculation of the equivalent frequency order. Variable multi-scale morphological filtering (VMSMF) method is proposed for the purpose of further reducing the interference of noise in resampling signal to feature extraction. VMSMF adaptively selects structural elements according to the parameter change of impact signal to make its scale more targeted. It only needs to calculate once using the optimal structural unit for a particular impact, and the filtering accuracy and operating efficiency have been greatly improved. The main steps of this article are as follows. First, the TDR is used to resample the original signal as to get the resampling signal which is still submerged by the strong noise. In the second step, VMSMF is used to filter the resampling signal to obtain the signal with less noise interference. Finally, the fault characteristics of the filtering signal was extracted and compared with the possible fault frequency calculated by the sampling parameters provided by resampling, so as to determine the fault type of the planetary gearbox. By analyzing the simulation signal and the experimental signal respectively, this method can find out the corresponding fault characteristics effectively. [ABSTRACT FROM AUTHOR]

Recently, Hyperspectral Image (HSI) classification has gradually been getting attention from more and more researchers. HSI has abundant spectral and spatial information; thus, how to fuse these two types of information is still a problem worth studying. In this paper, to extract spectral and spatial feature, we propose a Double-Branch Multi-Attention mechanism network (DBMA) for HSI classification. This network has two branches to extract spectral and spatial feature respectively which can reduce the interference between the two types of feature. Furthermore, with respect to the different characteristics of these two branches, two types of attention mechanism are applied in the two branches respectively, which ensures to extract more discriminative spectral and spatial feature. The extracted features are then fused for classification. A lot of experiment results on three hyperspectral datasets shows that the proposed method performs better than the state-of-the-art method. [ABSTRACT FROM AUTHOR]

The bearing is a component of the support shaft that guides the rotational movement of the shaft, widely used in the mechanical industry and also called a mechanical joint. In bearing fault diagnosis, the accuracy much depends on the feature extraction, which always needs a lot of training samples and classification in the commonly used methods. Neural networks are good at latent feature extraction and fault classification, however, they have problems with instability and over-fitting, and more labeled samples must be trained. Switchable normalization and semi-supervised learning are introduced to solve the above obstacles in this paper, which proposes a novel bearing fault diagnosis method based on switchable normalization semi-supervised generative adversarial networks (SN-SSGAN) with 1-dimensional representation of vibration signals as input. Experimental results showed that the proposed method has a desirable 99.93% classification accuracy in the case of less labeled data from the public data set of West Reserve University, which is better than the state-of-the-art methods. [ABSTRACT FROM AUTHOR]

With the popularity of using deep learning-based models in various categorization problems and their proven robustness compared to conventional methods, a growing number of researchers have exploited such methods in environment sound classification tasks in recent years. However, the performances of existing models use auditory features like log-mel spectrogram (LM) and mel frequency cepstral coefficient (MFCC), or raw waveform to train deep neural networks for environment sound classification (ESC) are unsatisfactory. In this paper, we first propose two combined features to give a more comprehensive representation of environment sounds Then, a fourfour-layer convolutional neural network (CNN) is presented to improve the performance of ESC with the proposed aggregated features. Finally, the CNN trained with different features are fused using the Dempster–Shafer evidence theory to compose TSCNN-DS model. The experiment results indicate that our combined features with the four-layer CNN are appropriate for environment sound taxonomic problems and dramatically outperform other conventional methods. The proposed TSCNN-DS model achieves a classification accuracy of 97.2%, which is the highest taxonomic accuracy on UrbanSound8K datasets compared to existing models. [ABSTRACT FROM AUTHOR]