Traffic incident detection based on the width characteristic of the moving object marker

Early traffic event detection is essential for transportation networks’ quick response and accurate performance recovery. Social media data (e.g., Twitter data) can be a valuable source for detecting and describing traffic events, such as accidents, congestion, and road closures. Integrating spatial and temporal features inherent in Twitter can enhance models by revealing non-trivial information, patterns, and knowledge, resulting in improved performance. Recent research in Deep Learning (DL) has revealed the strength of learning features directly from data to extract potential hidden features that efficiently infer human activities and interactions and detect the relationships to generate fine-grained information. This research explores the efficiency of integrating Twitter data’s spatial, temporal, and semantic features for traffic event detection using DL. A novel framework employing a transformer-based multi-feature fusion approach is proposed, designed to detect traffic incidents comprehensively via Twitter data. The framework classifies tweets based on multiple dimensions: (1) semantic content is numerically represented and categorized traffic incidents, non-traffic, or traffic conditions and information; (2) spatial characteristics are analyzed through hot-spot analysis techniques, classifying locations into hot or cold spots; and (3) temporal attributes (date and time) are visualized and analyzed through heat maps reflecting incident densities. The performance of the models was then evaluated based on various fusion scenarios combining spatial, temporal, and semantic data using performance metrics such as F-score and accuracy. The results showed that the scenario of transformer-based multi-feature fusion of spatial, temporal, and semantic data for traffic event detection yielded better results. The model achieved a 7.93% accuracy improvement when distinguishing between the two classes, “Traffic Incident” and “Non-Traffic”, and a 6.09% increase in accuracy when classifying across three categories: “Traffic Incident,” “Non-Traffic,” and “Traffic Conditions and information.” These findings highlight the effectiveness of using a multifeatured Twitter dataset for improved detection accuracy.

Highway traffic incident detection based on BPNN

Vehicular ad hoc networks (VANETs) have emerged in the past decades as a significant type of networks, which consists of vehicles with sensors to communicate. For its ad-hoc nature, VANETs have great potential in a large number of applications, within which rear-end collision warning and traffic automatic incidents detection are two major applications. Because of the large number of injury and consequent economic loss, rear-end traffic collision has become an important issue and attracted a large number of attentions. In the past decades, there have been lots of efforts paid on this field. Existing work usually employed mathematical approaches or machine-learning approaches. In this study, we develop a collaborative rear-end collision warning algorithm (CORECWA), which is able to estimate and assess traffic risk in a collaborative and real-time way, and further notify drivers the warning message timely. Experiments results have shown that our algorithm outperforms the predominant method, HONDA algorithm. On the other hand, traffic incidents detection has been a critical problem in the past decades, due to the considerable economical cost and inestimable disgruntlement from numerous drivers. We present a support vector machines (SVM)-based approach for automatic incident detection (AID), in which the traffic data are collected by VANETs techniques. We process collected data and utilize traffic variables in the SVM model to confirm whether an incident occurs. Several experiments have been conducted to evaluate our approach’s performance, and the results show that our approach could outperform the other two approaches in most cases.

Detecting spatiotemporal traffic events using geosocial media data

The first problem of traffic incident management is the detection and confirmation of traffic accidents. The detection method based on coil and video data is limited to practical application due to its high cost and insignificant detection effect. This paper presents a traffic incident detection algorithm based on outlier mining. The algorithm extracts characteristics of traffic event and builds a set of eigenvectors by using the new real-time traffic information released by NavInfo. The algorithm is simple, efficient and easy to deploy. Experimental results show that compared with traffic incident detection based on pattern recognition, the proposed algorithm has higher accuracy and can effectively distinguish between conventional congestion and traffic incidents.

Past studies on automatic traffic incident detection have mainly focused on the incidents on freeways, which are controlled-access roads. There are not many works on urban traffic incident detection. In addition, the traffic data used in detecting an incident are still mostly collected from fixed sensors such as loop detectors. With the advances in mobile sensing and vehicular technology, it is foreseeable that mobile sensors will be used increasingly in the near future. In fact, traffic data will be collected directly by vehicles. Detecting traffic incidents in an urban road network with the traffic data collected by mobile sensors poses several challenges. First, the urban roads are uncontrolled-access roads, which are typically full of flow-disruptive entities such as traffic signals, intersections, crossings, bus stops, etc. These entities can disrupt the traffic flow in a similar way that an incident does, making it more difficult to detect an incident. Second, it is still not clear which traffic variables, collected by mobile sensors, can be used in detecting an incident in an urban environment. In this paper, we investigate and identify the traffic variables that are effective in detecting an incident in an urban road network. Particularly, speed, acceleration, lane-change ratio and travel time are studied. The results show that these four traffic variables are generally effective for traffic incident detection. However, among the four variables, the least effective one is the travel time.

In the current Intelligent Transportation Systems (ITS) traffic monitoring and incident detection are usually supported with mostly traditional and relatively slow reactivity technologies. In this paper we propose a new service, namely THOR (Traffic monitoring Hybrid ORiented service), able to combine two different wireless technologies and to provide real time information about vehicular traffic monitoring and incident detection. THOR relies on LTE (Long Term Evolution) and Dedicated Short Range Communication based VANETs (Vehicular ad-hoc NETworks) in a hybrid approach, which is compliant with ITS standards. This hybrid networking approach can be deployed today and can be ready for tomorrow VANET technology. We test THOR by simulations in a scenario with vehicle flows synthesized from real measured vehicular traffic traces. We provide an LTE load analysis and an assessment of incident detection capabilities. Our results are promising in terms of reactivity, precision and network traffic load sustainability.

With more people owning cars, traffic accidents happen more frequently and result in many more fatalities each year. It is challenging for current roadside traffic incident detection equipment to have many blind spots, especially after an accident, which results in secondary injuries from traffic events not being discovered in time. As a result, a technique for detecting traffic incidents is presented for use on the vehicle side and is based on video captured by the car camera. The backbone of the proposed enhanced YOLOv5l model for the detection model is the EfficientNet structure with the SE layer eliminated, which drastically decreases the number of model parameters and computing work. To increase model detection accuracy, the fused attention technique is implemented at the prediction side. When compared to the YOLOv5s model, the upgraded model's mAP improves by 13% to 87.7%; when compared to the YOLOv5l model, our model's detection speed reduces by 7ms to 22.7ms, allowing it to be placed at the vehicle side for real-time detection of road traffic incidents.

SVM and KNN ensemble learning for traffic incident detection

Methods of treatment (case reports)

SNSJam: Road traffic analysis and prediction by fusing data from multiple social networks

Traffic incident detection is an essentialpart of the intelligenttransportation systemand attracts lots of attention from researchers in different areas. To reduce the computation and to avoid overfitting of the traditional feed-forward neural network, an extreme machine learning is implemented to detect traffic incidents. Using the real-world data, the proposed model is compared with feed-forward neural network and other two commonly used models by several evaluation criteria. The results indicate the proposed model has the highest accuracy, detection rate and lowest false alarm rate. Moreover, it consumes less time. In conclusion, the performance of the proposed model is better than other benchmark models and suit to detect traffic incidents.

A new approach to traffic incident detection is proposed in this paper. The method consists of two stages. In the first stage, a real-time adaptive on-line procedure is used to extract the significant components of traffic states, namely, average velocity and density of moving vehicles. In order to effectively and efficiently account for the time-varying and random nature of traffic incidents, it is necessary to have a real-time on-line adaptive algorithm. In the second stage, we apply a new neural network called fuzzy CMAC to identify traffic incidents. Simulation results show that the performance is very good.

Video surveillance traffic incident detection is an important part of Intelligent Transportation Systems, as it allows for the quick response to emergency incidents and enhances the safety of road users. The use of fixed cameras to monitor traffic creates several problems for traffic incident detection, including vehicle occlusions; changes in lighting, the complexity of how vehicles interact with each other, and the need to represent both sudden collisions and gradual change in traffic flow over time. While deep learning techniques have shown success in this area, most methods have focused on short-term event modelling and do not have the capacity to capture long-term event dependencies seen in real-world roadways. To address these challenges, an intelligent traffic incident detection framework using a deep fusion architecture is introduced. The spatial feature extraction is performed at the frame level using a ResNet50 backbone. Temporal dependencies on successive frames are captured by using transformer-based encoder. Sequential information of motion over time is refined using a Long Short-Term Memory (LSTM) network; thus, providing a temporal-coherence component of sequencing. The intelligent traffic accident detection framework was tested against the Traffic Accident Detection Dataset, built from highway surveillance camera video footage. The evaluation of the framework produced a classification accuracy of 97.93% (with equal precision and recall for accident and normal traffic classes) in realistic surveillance scenarios.

Focusing on retrieving abnormal events from traffic surveillance video databases, a novel video description scheme (DS) is proposed in this work, which divides the object description into moving object description and still region description. Additionally, to address the challenge posed by redundant description data, we developed one group of second order spatial relationship (SR) semantics to represent the changes of SR over time directly; Incorporated with the proposed DS, a semantic model based on description data is developed to implement the low-complexity traffic event detection. Experimental results have shown that the proposed detection model is very efficient.