Secure task offloading scheme for traffic video surveillance via reinforcement learning

Enormous advance has proven throughout the years in the area of traffic surveillance by the growth of intelligent traffic video surveillance system. In the current work, through the traffic videos, the traffic video surveillance automatically keyed out the vehicles like ambulance and trucks, which in turn assisted us in directing the vehicles at the time of emergency. Nevertheless, it doesn’t provide us a vital solution for the regulating the traffic. Moreover, this idea just identifies the vehicles, but it couldn’t notice the accidents expeditiously. Therefore in the proposed work, expeditious traffic video surveillance and monitoring system are presented along with dynamic traffic signal control and accident detection mechanism. Hybrid median filter has been utilized at the beginning for pre-processing of traffic videos, and to remove the noise. Hybrid support vector machine (SVM with extended Kalman filter) has been utilized to chase the vehicles. Next, the histogram of flow gradient features are drew-out to categories the vehicles. According to the traffic density and through video files, vehicles are computed, and then for emergency vehicles, the traffic signal gets switched dynamically. To realize the arrival of ambulances, the cameras have been set to catch traffic videos minimum at 500 m of the signal and deep learning neural networks has been employed. Hence dynamic signal control has been incorporated expeditiously. Likewise, multinomial logistic regression has been utilized in real-time live streaming videos, to identify the accidents correctly. The observational solution shows that the proposed intelligent traffic video surveillance system render expeditious dynamic control of traffic signals and it raises the identification of accidents correctly.

We consider the problem of secure geographic routing (GR) in mobile wireless ad-hoc networks (WANETs). This paper presents a dependable and pragmatic trust and reputation aware geographic routing (TARA-GR) method for mobile WANETs to achieve self-adaptability, scalability, lightweight communication, and efficiency. It supports the GR decisions by the dynamic computation of consolidated trust value (CTV) of each node in the network by integrating direct and indirect trust values. With the calculated trust value, each source or forwarding node selects the trusted nodes to route the packets from the source to the destination. TARA-GR is integrated with a well known greedy perimetre stateless routing (GPSR) protocol. This integration is termed as TARA-GPSR. The simulation results have shown that TARA-GPSR significantly improved the network performance metrics, such as packet delivery fraction, throughput, and energy consumption.

Artificial intelligence and machine learning have great ability to solve real-time problems related to various fields and have many applications. One such subfield is deep learning, deep learning has many architectures like DNN, CNN and RNN. One of the major applications of deep learning is object detection. YOLO a SOTA object detector considered to be a smart convolutional neural network having capability of detecting objects in the image, classifying them accordingly and localizing the object perfectly with annotations. YOLOv4 is employed in this work for multiple object detection in image and video for traffic surveillance applications trained using a custom dataset created with Indian road traffic images. 9 different classes namely Car, Bus, Van, Truck, Two-wheeler, Auto, Person, Bicycle and Mini truck are considered in creating custom dataset. Proposed custom model trained for 500 epochs with a custom image dataset has achieved a training mAP of 92% and a training loss of 0.001. Achieved an accuracy of 99.28%, sensitivity of 94.44% and MCC of 96.78% on unseen or test dataset. Proposed model predicted all the classes with an average precision of 98.32%. This proves that our model is robust and proficient in detecting road traffic classes and has scope in developing traffic surveillance systems.

The Internet of Things (IoT) networks are the most prone to internal as well as external attacks in addition to energy consumption problems. Conventional security solutions are not able to address these issues effectively due to the limited resources of sensor nodes participating in IoT communications. In this work, an Efficient and Secure Fog Based Routing Mechanism (ESFRM) is proposed to protect the network from faulty internal as well as external attacks. Every node participating in IoT communications calculates the comprehensive trust value of the next intermediate node which is the addition of direct trust, indirect trust and energy trust values before forwarding the data. This comprehensive trust value is then compared with the comprehensive threshold trust value to decide whether the particular node is a rogue node or a valid normal node. Further, an enhanced RSA (Rivest, Shamir, Adleman) algorithm is implemented to provide three levels of data security from Cluster Head (CH) to fog node, from fog node to cloud server and directly from CH to cloud server. For this purpose, an efficient CH selection technique is also proposed in this work. The proposed methodology is compared with the Secure Energy-efficient Fog-based Routing (SEFR) protocol and Trust-aware Secure Routing Protocol (TSRP). The evaluation results show that the proposed ESFRM outperforms the conventional schemes with respect to energy consumption, malicious node detection and transmission rate.

The Trust-Aware Aggregation Authentication Protocol for SIoT Networks is a security process intended for SIoT platforms. It concentrates on ensuring the reliability of communication and data aggregation between interrelated IoT devices. This protocol deploys authentication systems for verifying the identities of devices and integrates trust-aware mechanisms to estimate the trustworthiness of data exchanged from the social environment of SIoT. By establishing a trustworthy and secure communication structure, this protocol improves the entire integrity and security of SIoT networks, addressing potential vulnerabilities connected with social communications between IoT devices. Therefore, this study develops an enhanced Trust Management using Locally Weighted Salp Swarm Algorithm with Deep learning (ETM-LWSSADL) technique for SIoT Networks. The ETM-LWSSADL technique computes direct and indirect trust values and is assessed depending upon different weighing factors for maximizing the application performance and creating a secure data transmission process. During authentication process, when the SIoT device with total trust value (TTV)is not greater than the threshold value (THV) or authentication token is invalid, the gateways then disregard the node. Besides, bidirectional gated recurrent unit (BiGRU) model is applied to generate a THV on collected traffic data. Moreover, the ETM-LWSSADL technique exploits the LWSSA technique for optimum hyper parameter selection of the BiGRU algorithm. To highlight the enhanced performance of the ETM-LWSSADL methodology, an extensive range of simulations can be involved.

Road segmentation plays an important role in many computer vision applications, either for in-vehicle perception or traffic surveillance. In camera-equipped vehicles, road detection methods are being developed for advanced driver assistance, lane departure, and aerial incident detection, just to cite a few. In traffic surveillance, segmenting road information brings special benefits: to automatically wrap regions of traffic analysis (consequently, speeding up flow analysis in videos), to help with the detection of driving violations (to improve contextual information in videos of traffic), and so forth. Methods and techniques can be used interchangeably for both types of application. Particularly, we are interested in segmenting road regions from the remaining of an image, aiming to support traffic flow analysis tasks. In our proposed method, road segmentation relies on a superpixel detection based on a novel edge density estimation method; in each superpixel, priors are extracted from features of gray-amount, texture homogeneity, traffic motion and horizon line. A feature vector with all those priors feeds a support vector machine classifier, which ultimately takes the superpixel-wise decision of being a road or not. A dataset of challenging scenes was gathered from traffic video surveillance cameras, in our city, to demonstrate the effectiveness of the method.

An intelligent transportation system has a major topic called traffic surveillance. In a complex urban Traffic Surveillance System (TSS), booming of vehicle detection and tracking is a dilemma. To overcome the problem, a two stage approach for moving vehicle detection system is proposed in this paper. The proposed system mainly consists of two stages namely, Hypothesis Generation (HG) and Hypothesis Verification (HV). At the first step, hypotheses are generated by shadows beneath the vehicles is darker than the road region concept. In the second step a generated hypothesis is verified as correct or not using Optimal Artificial Neural Network (OANN). The weights corresponding ANN is optimally selected using Grasshopper Optimisation Algorithm (GOA). Through experimental results, it is shown that the proposed moving vehicle detection system has better accuracy compared to other methods.

In recent years, traffic video surveillance has increased significantly. However, most of the footage is reviewed by humans or not at all. Therefore, tools capable of analysing traffic video sequences and autonomously extracting information are required. In this paper, we present an agent-based approach to analysing driver behaviour. Our work differs from normal road monitoring systems in that we are interested in inferences about driver behaviour and in learning normal' driving modes, rather than specific instances of driver actions. Our system provides a behavioural description of traffic scenes. First, we present a kinematic traffic simulator designed to test driving agents. The simulator supports multiple lanes, obstructions and different environmental conditions. Second, we specify the agent's perception and reasoning models. Contrary to current autonomous driving systems, our behavioural models primarily influence agent perception. This approach is supported by recent psychological studies carried out on human drivers. Furthermore it simplifies the system implementation, increasing the ease of testing alternative models. By embedding the agents in the simulator, we observe classical traffic behaviour. Finally, we suggest ways to use the system's results directly or within higher level tools.

Vehicle global 6-DoF pose estimation under traffic surveillance camera

A major challenge that all law-enforcement and intelligence-gathering organizations are facing is to accurately and efficiently analyze the growing volumes of crime data. In many fields, video surveillance can be used for that purpose. Video surveillance systems are now commonly used at various places like banks, hotels, schools, supermarkets. These systems are applicable for real-time monitoring or post checking. Current video surveillance systems have lower intelligence and it need people to monitor them. Today, video surveillance systems are used for traffic cameras, too. Traffic cameras are an innovative and extremely functional application of video surveillance technology. Whether they are recording traffic patterns for the future study and observation or monitoring traffic and issuing tickets for moving violations, traffic cameras are an explosively popular form of video surveillance. One way to find traffic patterns or to prevent traffic accidents and provide better road security is to use data mining techniques. With data mining, stream from many traffic cameras can be processed in real time. Property like this could provide some intelligence upgrade within systems, so they could make decision on their own without the need of people to monitor them. In this paper the authors described the data mining techniques for stream processing, pattern and abnormality detection in traffic surveillance systems. Each video recording and the use of recorded video on the court must be covered by legal regulations. The authors investigated and extracted the appropriate articles from concrete Laws of Republic Serbia, covering this kind of problems.

The model of indirect trust value is done further research in statistics principle and a new model of trust value is proposed. On this basis, the electoral method of cluster node is studied, and then a routing algorithm is given based on considering the trust value, the density, the distance between nodes and the forward capability. Experiment and simulation results show that the established trust value model is considerably reliable, and with the increase of cycles, the trust value gradually tends to 1. In a certain range about the number of nodes, in comparison with references 10 and 11, this algorithm improves convergence speed largely and is stable during the information transfer process. When the network topology changes, this algorithm can do fast rerouting, which shows this optimization algorithm has widely application.

International audience

Concerns the development of a traffic surveillance and control system (TCSS) for the Tsing Yi interchange, which is at the end of the new Tsing Ma bridge which leads to the airport. Hazards include not only traffic incidents but also severe winds including typhoons. The complete TCSS comprises two types of subsystems, surveillance systems to identify the real-time traffic situation and control systems to implement appropriate traffic control measures when and where necessary.

Automated vehicle detection plays an essential role in the traffic video surveillance system. Video communication of these traffic cameras over real-world limited bandwidth networks can frequently suffer network congestion or unstable bandwidth, especially in regard to wireless systems. This often hinders the detection of moving vehicles in variable bit-rate video streams. This paper presents a novel approach for vehicle detection based on probabilistic neural networks through artificial neural networks which can accurately detect moving vehicles not only in high bit-rate video streams but also in low bit-rate video streams. The overall results of detection accuracy analyses demonstrate that the proposed approach has a substantially higher degree of both qualitative and quantitative efficacy than other state-of-the-art methods. For instance, the proposed method achieved Similarity and F1 accuracy rates that were up to 77.69% and 81.85% higher than the other compared methods, respectively.

This paper proposes human tracking and recognition method in a camera network. Human matching in a multi-camera surveillance system is a fundamental issue for increasing the accuracy of recognition in multiple views of cameras. In camera network, videos have different characteristics such as pose, scale and illumination. Therefore it is necessary to use a hybrid scheme of scale invariant feature transform to detection and recognition human's behaviors. The main focus of this paper is to analyze activities for tracking and recognition humans to extract trajectories. Extracting the trajectories help to detect abnormal behavior which may be occluded in single- camera surveillance. KEYWORDS: Camera network, Multi-camera surveillance, Human's behavior, Trajectories extraction. I. INTRODUCTION Tracking and behavior recognition are two fundamental tasks in video surveillance systems which are widely employed in commercial applications for purposes of statistics gathering and processing. The number of cameras and complexity of surveillance systems have been continuously increasing to have better coverage and accuracy. Multi-camera systems become increasingly attractive in machine vision. Applications include multi view object tracking, event detection, occlusion handling and etc. In this paper, we develop method for tracking and recognition by a traffic video surveillance system of two cameras with a partially overlapping field of view. This paper is organized as follows: an overview of the past works in section2. Our proposed architecture and algorithm is presented in section3. Results of subjective evaluations and objective performance measurements with respect to Ground-truth are presented in section4. Section5 contains the conclusion.