An Autonomous Feature Detection Method for Slow-Moving Small Target on Sea Surface Based on Kernelized Contextual Bandit

Adaptive ORB feature detection with a variable extraction radius in RoI for complex illumination scenes

In recent years, with the progress of society and the development of the economy, the number of cars in China has been increasing. In wireless communication networks, the choice of wireless nodes has a greater impact on the improvement of system performance (such as channel capacity, coverage area, etc.). The intelligent vehicle target detection system can perceive and recognize the surrounding objects such as pedestrians and vehicles through sensors, which is the basis for realizing the unmanned driving of intelligent vehicles. In a wireless environment where multiple wireless nodes coexist, current research focuses on how long it takes to re-plan the selection of wireless nodes (for mobile environments) and how to allocate and manage wireless nodes, such as where under the conditions of wireless, who will wireless with whom, and in what way (such as auto focus, digital fusion, or other) wireless, etc.; the dedicated centralized controller (such as the base station of the infrastructure-based wireless access network) determines the wireless partner (i.e., centralized type selection), or the wireless node decides by itself (i.e., distributed selection); select the appropriate number of wireless nodes to take into account the system performance gain and implementation complexity. The popularity of automobiles has brought great convenience to people's lives, but it has also brought greater traffic pressure. At present, the amount of domestic automobile traffic has increased exponentially, and urban traffic congestion has become more serious, and even caused serious traffic accidents, directly affecting people's quality of life. The emergence of intelligent transportation systems has effectively alleviated road traffic pressure and reduced the incidence of traffic accidents. As an important part of the intelligent transportation system, the research of intelligent vehicles has received extensive attention. Intelligent vehicles use a variety of sensors installed on the body to sense the environment, and realize intelligent driving functions such as lane line detection, obstacle detection, dynamic cruise control, and unmanned driving, which is conducive to reducing the incidence of traffic accidents and improving the safety of vehicle driving. At present, in the research of target recognition and tracking of intelligent vehicles, the traditional target detection method is mainly based on artificial feature extraction, which is difficult to describe more complex or higher-order image features, and the tracking effect is not good, which limits the target detection. An intelligent vehicle is an intelligent system with functions such as environment perception, planning decision-making, and operation control, and is an important part of the intelligent transportation system. Identify the effect. In response to these problems, we present a study of the target recognition and tracking of intelligent vehicles based on grid map and lidar sensor technology. To verify the proposed method in the actual scene, we identify and track the vehicle ahead on the road through a car equipped with a lidar sensor. The meaning of target detection: target detection, also called target extraction, is an image segmentation based on target geometric and statistical features. The results show that the proposed research method can accurately identify the vehicle in front and detect the moving targets such as pedestrians. The tracking trajectory is also consistent with the expected route. The results show that the laser radar-based vehicle target recognition and tracking in this study is effective. In the 10 frames of images, the average time taken for the recognition of one frame of image is 9.8 ms, while the traditional method is 12.9 ms. Technology can improve reliable road information for intelligent vehicle driving and improve the transportation efficiency and safety performance of intelligent vehicles. There are two main target detection methods for lidar: detection methods based on feature extraction and detection methods based on grid maps.

Benefiting from the development of big data, edge computing, and deep learning, splendid breakthroughs have been made in automatic speech recognition (ASR) in recent years. Since then, more and more smart products have chosen speech as the interface for human-computer interaction, which causes popularity of edge intelligence (EI) enhanced automatic speech recognition. While people are enjoying the social changes brought by speech recognition technology, a factor of instability quietly emerged called audio adversarial example which is a type of audio deliberately generated by attackers via adding subtle perturbations to the original audio signal. The added perturbations which sound like certain noise that cannot be precepted by human but will cause ASR system make wrong transcription. Three detection algorithms for audio adversarial examples are proposed in this thesis, namely, the robust detection algorithm based on WER (word error rate), the feature detection algorithm based on ADR (adversarial ratio), and the collaborative detection algorithm based on neural network. The experiment results show that three detection algorithms proposed in this thesis have a great discrimination on audio adversarial examples and achieve high AUC scores. Among them, the cooperative detection is the best and the feature detection is the worst. In addition, we found that robust detection algorithm tends to have a higher accuracy score but a lower recall score, while feature detection algorithm tends to have the converse performance. Moreover, since the proposed collaborative detection method combines the advantages of the robust detection and feature detection methods, it presents a better performance with respect to accuracy, recall, and F1 score.

This paper explores the concept of a Forward Scattering Radar (FSR) system with a rotational transmitter for target detection and localization. Most of the research and development in FSR used a fixed dedicated transmitter; therefore, the detection of stationary and slow moving target is very difficult. By rotating the transmitter, the received signals at the receiver contain extra information carried by the Doppler due to the relative movement of the transmitter-target-receiver. Hence, rotating the transmitter enhances the detection capability especially for a stationary and slow-moving target. In addition, it increases the flexibility of the transmitter to control the signal direction, which broadens the coverage of FSR networks. In this paper, a novel signal processing for the new mode of FSR system based on the signal’s joint time-frequency is proposed and discussed. Additionally, the concept of the FSR system with the rotational transmitter is analyzed experimentally for the detection and localization of a stationary target, at very low speed and a low profile target crossing the FSR baseline. The system acts as a virtual fencing of a remote sensor for area monitoring. The experimental results show that the proposed mode with the new signal processing scheme can detect a human intruder. The potential applications for this system could be used for security and border surveillance, debris detection on an airport runway, ground aerial monitoring, intruder detection, etc.

Based on the discreteness of existing form and the continuity of energy distribution of digital image, this paper presents a new local feature detection method. The proposed method realizes the local feature detection of optical image, radar image and infrared image based on gray gradient sequence and 1-D Hilbert-huang transform method. Experimental results show that this method can describe and detect common local feature of single-sensor images and multi-sensor images efficiently for image registration application.

Synthetic aperture radar ground moving target indication (SAR-GMTI) is an attractive mode for radar system to obtain high-resolution regional images and detection of moving targets simultaneously. Moreover, the 360° observation capability of circular SAR (CSAR) makes it possible for moving target trajectory reconstruction. In order to address the problem that slow-moving targets, which are buried in strong stationary ground clutter, are difficult to be detected, a framework of multichannel SAR-GMTI is presented in this article. For ground moving target detection in the presence of stationary clutter, the proposed method first adopts clutter suppression interferometry and relaxation-based cyclic algorithm to suppress clutterand retrieve parameters (e.g., radial velocity and Doppler information). Then, to further reduce the false alarm rate, multiple moving target tracking algorithm is performed in range–Doppler domain. Finally, the moving target trajectory is reconstructed by using the proposed two-stage parameter estimation method based on Doppler characteristic and CSAR geometry, and the mathematical analysis of this method is performed. The proposed method is robust and insensitive to the signal-to-noise ratio. It does not rely on priori road information, so it has wide applicability, especially in military application. In addition, the framework is performed in echo domain and independent of imaging processing. The experimental results on simulated data are presented to evaluate the estimation accuracy of the proposed method, and the real data processing results are provided to demonstrate the validity and feasibility of the proposed method.

Non-cooperative target position and attitude estimation are essential technologies for the effective implementation of space target operations. However non-cooperative target lacks prior information and information interaction and require certain real-time performance with hardware limitations in space target operation. This presents significant challenges in accurately estimation the pose of non-cooperative targets so the model-based method and deep learning have certain limitations. To address the challenges in real scenarios, we utilize YOLOv7 for preliminary object detection and propose the traditional feature-based methods to tackle the important task of identifying essential characteristics for determining the position and attitude of non-cooperative spacecraft. A feature extraction method specifically tailored for spacecraft solar panels commonly observed is proposed based on line detection and clustering and the combination of length and angle measure functions. The quadrilateral structure closest to the solar panel can be extracted by the measure function. The physical simulation platform is built and based on the experimental results, YOLOv7 demonstrates good accuracy in the detection of solar panels with low real-time, the proposed solar panel position method effectively extracts and locates the features of the solar panel with minimal pixel error. It possesses a certain degree of robustness and real-time performance that can facilitate the subsequent estimation of position and attitude.

In this paper, we propose a novel framework of localization algorithm based on Compressive Sensing (CS) by K-Nearest Neighbor (KNN) classification in Wireless Sensor Networks (WSNs), which can effectively alleviate the problem of off-grid. Many papers about CS-based localization assume that each target is in the center of the grid. However, the assumption is almost impossible. In this paper, the targets can be located anywhere in the area. The target location is regarded as an unknown sparse signal in the discrete spatial domain. Practically, the recovered discrete signal is not exactly K-sparse. In order to make high level of localization accuracy, we localize the targets in three different situations and choose the cluster region according to the distribution of the power of the targets. Then for each classification, we obtain the final location by KNN. This approach can make great use of the location information and compensate the error induced by the problem of off-grid. Finally, simulation and performance analysis show the effectiveness and robustness of our proposed localization algorithm.

Recently, passive bistatic radar (PBR) has become an emerging technology because it allows target detection and localization with advantages of low cost, covert detection, low vulnerability to electronic jamming and high valuable counter to stealth technology. However, these researchers mainly focus on the illuminators of opportunity such as FM radio, television broadcast, mobile network basestation, wireless fidelity, long term evolution, global navigation satellite, etc. Different from the aforementioned illuminators, this paper introduces a new PBR using a non-cooperative frequency agile phased array radar illuminator. The types of illuminator of opportunity are enriched, whereas the difficulties in PBR signal processing also come along. On one hand, since PBR works in a non-cooperative mode, there is a problem that the instantaneous parameters of transmitting signal are unknown at the receiver. On the other hand, the frequency agility technology destroys the coherency between pulses and the ability of rapidly changing beam scanning makes it difficult for PBR to realize space synchronization. Under such scenarios, conventional radar signal processing methods cannot be conducted effectively. To address these problems, this paper then proposes a feasible target detection and localization method, which is tested by field experiments. Preliminary results validate the proposed method can be applied to PBR system successfully.

China’s sea surface area is vast, the need to monitor the area is too large, and the traditional human monitoring method consumes a lot of manpower. Additionally, the monitoring period is too long; the monitoring efficiency is too low; and long-term human monitoring can easily cause visual fatigue, as well as missed detection and error detection. At present, the detection of sea surface targets generally includes infrared, visible light and other different means, which can obtain the image information of sea surface targets in different ways. The infrared target detection of the sea surface can be processed in the spatial domain and frequency domain, respectively, but the image resolution is not high in general, and the detection effect is not good because it is easily affected by weather. In this paper, we propose a maritime target detection method based on embedded vision. Based on visible video images, this paper realizes the rapid detection and recognition of sea surface targets. Clouds and waves in ocean images are filtered by adding an image preprocessing module. Compared with the traditional two-frame difference method, this algorithm has better detection capability for sea surface targets. Experiments were carried out in different weather conditions to detect moving ships at sea. By comparing the number of detection boxes and the detection accuracy, the accuracy of this method reaches 90.2 percent. By designing a single camera location algorithm for the marine environment, the world coordinate location of the marine target is realized. On this basis, the communication function is added to realize the intelligent monitoring of the sea surface without human intervention.

This work is related to the method of facial feature detection based on distance vector fields (DVFs), recently proposed by Asteriadis et al. We briefly present the concept and describe improvements that we introduced to the original solution. The main advantages of our approach are the reduced computational complexity of the DVF extraction algorithm as well as the enhanced precision of the resultant vector field.

In this paper we present a new method for accurate real-time facial feature detection. Our method is based on local feature detection and enhancement. Previous work in this area, such as that of Viola and Jones, require looking at the face as a whole. Consequently, such approaches have increased chances of reporting negative hits. Furthermore, such algorithms require greater processing power and hence they are especially not attractive for real-time applications. Through our recent work, we have devised a method to identify the face from real-time images and divide it into regions of interest (ROI). Firstly, based on a face detection algorithm, we identify the face and divide it into four main regions. Then, we undertake a local search within those ROI, looking for specific facial features. This enables us to locate the desired facial features more efficiently and accurately. We have tested our approach using the Cohn-Kanade’s Extended Facial Expression (CK+) database. The results show that applying the ROI has a relatively low false positive rate as well as provides a marked gain in the overall computational efficiency. In particular, we show that our method has a 4-fold increase in accuracy when compared to existing algorithms for facial feature detection.

Pixel versus object — A comparison of strategies for the semi-automated mapping of archaeological features using airborne laser scanning data

With the increase in urban rail transit construction, instances of tunnel disease are on the rise, and cracks have become the focus of tunnel maintenance and management. Therefore, it is essential to carry out crack detection in a timely and efficient manner to not only prolong the service life of the tunnel but also reduce the incidence of accidents. In this paper, the design and structure of a tunnel crack detection system are analyzed. On this basis, this paper proposes a new method for crack identification and feature detection using image processing technology. This method fully considers the characteristics of tunnel images and the combination of these characteristics with deep learning, while a deep convolutional network (Single-Shot MultiBox Detector (SSD)) is proposed based on deep learning for object detection in complex images. The experimental results show that the test set accuracy and training set accuracy of the support vector machine (SVM) in the classification comparison test are up to 88% and 87.8%, respectively; while the test accuracy of Alexnet’s deep convolutional neural network-based classification and identification is up to 96.7%, and the training set accuracy is up to 97.5%. It can be seen that this deep convolutional network recognition algorithm based on deep learning and image processing is better and more suitable for the detection of cracks in subway tunnels.

Sentiment analysis is widely used today to make data-driven decisions in different industries, starting from marketing and including brand management, reputation monitoring, and customer satisfaction analysis. Its growing importance is closely linked with so-called ‘word-of-mouth’ communication, from reading online reviews to writing comments on social networks. Effective separation of sentiments ensures that companies' responses are timely and critical patterns are seen in big data sets. Statistical measures, information gain, correlation-based approaches, etc, have been employed for the feature selection. Still, the problem associated with text data mining is that they don’t convey the text's relative difficulty and additional features. To fill this gap, our research proposes a new feature selection technique through Ant Colony Optimization (ACO) and K Nearest Neighbour (KNN) performed on 28,000 customer reviews in different product categories. The results, therefore, showed an overall accuracy of 80.1%, with the Support Vector Machine (SVM) set at 80.5% on each selected feature, which was slightly higher than the Convolutional Neural Network (CNN), which scored a 78.41% accuracy. SVM remains on the mark of 83%, and for CNN, the rate achieved on the same was 80.8% when both were applied to the entire dataset. These facts rejected the infallibility of the simple and complex algorithms used singly in the sentiment classification, indicating that more sophisticated algorithms like ACO and KNN can provide business solutions to improve their service delivery based on customers’ feedback.