Traffic Video Surveillance Research Articles

Smoky Vehicle Detection Based on Range Filtering on Three Orthogonal Planes and Motion Orientation Histogram

Smoky vehicle detection is an important task in reducing motor vehicle pollution. This paper presents a method to automatically detect smoky vehicles from the traffic surveillance videos. More specifically, the visual background extractor background subtraction algorithm and some rules are adopted to detect moving vehicle object and locate the key region at the back of the vehicle. Based on sufficient observations of the smoke characteristics in the real scene, three groups of features, including color moments (CMs) features, improved motion orientation histogram features, and the new model range filtering on three orthogonal planes (RF-TOP)-based features, are designed and proposed to distinguish smoky vehicles and non-smoke vehicles. The color information CM features are used as a preliminary sieve to filter out the samples that are obviously non-smoke regions. The other two groups of features are combined to one feature vector to obtain motion information and spatiotemporal information of the key region. Two strategies, including histogram and projection, are designed to extract discriminative dynamic features from the proposed model RF-TOP to characterize the key region. The pruning radial basis function neural network classifier is adopted to classify the extracted features. For the traffic surveillance videos in the daylight with sunny weather, the experimental results show that the proposed methods have better performances and work effectively with lower false alarm rates than existing methods, and the proposed method with histogram strategy achieves the best performance.

Road Traffic Anomaly Detection Based on Fuzzy Theory

The road traffic scenes are usually complex and the traffic video is vulnerable to external factors such as light, weather, and obstructions. It is difficult to extract the traffic parameters and detect the traffic anomaly exactly with the existing image processing and analysis technologies under such uncertainty factors. Considering the advantages of the fuzzy theory in dealing with uncertain information, we use the fuzzy theory to handle the complex issues in traffic video surveillance and put forward a traffic anomaly detection algorithm. First, the fuzzy traffic flow is designed based on the virtual detection lines and the fuzzy theory. Second, the fuzzy traffic density is designed based on the pixel statistics and the fuzzy theory. Third, the target’s fuzzy motion state is designed based on the vehicle trajectory and the fuzzy theory. Besides, the relevant membership functions for these traffic parameters are designed to perform state evaluation. Finally, the traffic anomaly detection algorithm is designed based on above-mentioned fuzzy traffic parameters and fuzzy control rules. The experimental results show that the algorithm proposed performs well in road anomaly detection.

Research on Vehicle Detection Method Based on Background Modeling

Abstract This paper mainly studies the background difference method in the field of intelligent traffic, proposes a background modeling method base on frame difference, and compares it with the statistical average background model and Gaussian distribution background modeling method. vehicle contour obtained by the morphological method. Finally, experiments were carried out on 4 normal road traffic surveillance videos, the effective detection rate used in this paper reaches 93.75%, which has a certain degree of application. The algorithm model need to be further tested in more complex weather and different road conditions.

Regional Detection of Traffic Congestion Using in a Large-Scale Surveillance System via Deep Residual TrafficNet

Despite the huge amount of traffic surveillance videos and images have been accumulated in the daily monitoring, deep learning approaches have been underutilized in the application of traffic intelligent management and control. In this paper, traffic images, including various illumination, weather conditions, and vast scenarios, are extracted from the current surveillance system using in Shaanxi Province and preprocessed to set up a proper training dataset. In order to detect traffic congestion, a network structure is proposed based on residual learning to be pre-trained and fine-tuned. The network is then transferred to the traffic application and re-trained with self-established training dataset to generate the TrafficNet. The accuracy of TrafficNet to classify congested and uncongested road states reaches 99% for the validation dataset and 95% for the testing dataset. The proposed TrafficNet are verified by a regional detection of traffic congestion on a large-scale surveillance system currently using in China. The effectiveness and efficiencies are magnificently demonstrated with quick detection in the high accuracy in the case study. The experimental trial could extend its successful application to traffic surveillance system and has potential enhancement for intelligent transport system in future.

Efficient Moving Vehicle Detection for Intelligent Traffic Surveillance System Using Optimal Probabilistic Neural Network

The vehicle detection system 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. The objective of this paper is to develop an effective method for moving vehicle detection problems that can find high quality solutions (with respect to detection accuracy) at a high convergence speed. To achieve this objective, we propose a method that hybridises the cuckoo search (CS) with Opposition-based learning (OBL), where OBL is improve the performance of the CS algorithm while optimising the weights of the standard PNN model. The proposed system mainly consists of two modules such as: 1) design novel OCS-PNN model; 2) moving vehicle detection using OCS-PNN model. The algorithm is tested on three standard video dataset. For instance, the proposed method achieved the maximum precision of 94%, F-measure of 94% and similarity of 94%.

A learning-based approach for leaf detection in traffic surveillance video

Traffic surveillance video is recorded in uncontrolled outdoor scenarios. If the camera view gets obstructed by the leaves, the video will fail to be used in vehicle tracking and recognition. It is required that the traffic video surveillance systems run self-checking in order to evaluate if the camera view is blocked by leaves or not. In view of this, a two-step learning framework is proposed in this paper to automatically determine whether the video is leaf degraded or leaf free. First, the proposed framework exploits the convolutional neural network to learn the discriminative features of leaf particles. Then the trained model is used to detect candidate leaf patches in the image. Second, a probabilistic approach is used to pool decisions of each candidate leaf patch to generate final leaf detection result in the video. Experimental results are provided to demonstrate that the proposed approach can effectively detect leaves in real-world traffic surveillance video.

Fast scene analysis for surveillance & video databases

In professional/consumer domains, video databases are broadly applied, facilitating quick searching by fast region analysis, to provide an indication of the video contents. For realtime and cost-efficient implementations, it is important to develop algorithms with high accuracy and low computational complexity. In this paper, we analyze the accuracy and computational complexity of newly developed approaches for semantic region labeling and salient region detection, which aim at extracting spatial contextual information from a video. Both algorithms are analyzed by their native DSP computations and memory usage to prove their practical feasibility. In the analyzed semantic region labeling approach, color and texture features are combined with their related vertical image position to label the key regions. In the salient region detection approach, a discrete cosine transform (DCT) is employed, since it provides a compact representation of the signal energy and the computation can be implemented at low cost. The techniques are applied to two complex surveillance use cases, moving ships in a harbor region and moving cars in traffic surveillance videos, to improve scene understanding in surveillance videos. Results show that our spatial contextual information methods quantitatively and qualitatively outperform other approaches with up to 22% gain in accuracy, while operating at several times lower complexity.

Intelligent traffic video surveillance and accident detection system with dynamic traffic signal control

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.

Vehicle classification for large-scale traffic surveillance videos using Convolutional Neural Networks

Vehicle classification plays an important role in intelligent transport system. However, because the conventional vehicle classification methods are not robust to variations such as illumination, weather, noise, and the classification accuracy cannot meet the requirements of practical applications. Therefore, a new vehicle classification method using Convolutional Neural Networks is proposed in this paper, which consists of two steps: pre-training and fine-tuning. In pre-training, GoogLeNet is pre-trained on ILSVRC-2012 dataset to obtain the initial model with the corresponding connection weights. In fine-tuning, the initial model is further fine-tuned on VehicleDataset which is constructed with 13,700 images in this paper to obtain the final classification model. All images in the VehicleDataset are extracted from real highway surveillance videos, including variations of illumination, noise, resolution, angle of video cameras and weather. The vehicles are divided into six categories, i.e., bus, car, motorcycle, minibus, truck and van. The performance evaluation is carried out on the VehicleDataset. The experimental results show that the proposed method can avoid the complicated process of manually extracting features and the average classification accuracy is up to 98.26%, which is 3.42% higher than the conventional methods using “Feature + Classifier”.

Dust particle detection in traffic surveillance video using motion singularity analysis

Dust particle detection in video aims to automatically determine whether the video is degraded by dust particle or not. Dust particles are usually stuck on the camera lends and typically temporally static in the images of a video sequence captured from a dynamic scene. The moving objects in the scene can be occluded by the dusts; consequently, the motion information of moving objects tends to yield singularity. Motivated by this, a dust detection approach is proposed in this paper by exploiting motion singularity analysis in the video. First, the optical model of dust particle is theoretically studied in by simulating optical density of artifacts produced by dust particles. Then, the optical flow is exploited to perform motion singularity analysis for blind dust detection in the video without the need for ground truth dust-free video. More specifically, a singularity model of optical flow is proposed in this paper using the direction of the motion flow field, instead of the amplitude of the motion flow field. The proposed motion singularity model is further incorporated into a temporal voting mechanism to develop an automatic dust particle detection in the video. Experiments are conducted using both artificially-simulated dust-degraded video and real-world dust-degraded video to demonstrate that the proposed approach outperforms conventional approaches to achieve more accurate dust detection.

Mobile content distribution with vehicular cloud in urban VANETs

Plenty of multimedia contents such as traffic images, surveillance video, music and movie will flood into vehicular ad hoc networks. However, content distribution over VANETs is not a easy task, due to the high mobility of vehicles and intermittent connectivity. Infrastructure-based scheme can relieve the problem, but with a large amount of investment. In this paper, we propose a mobile content distribution scheme based on roadside parking cloud (RPC), which is formed by the parked car on the roadside, and mobile cloud (MC), which is formed by moving cars on the road. According to a trip history model, a mobile car can estimate its following trajectory. When it wants to download the content, gateway node of the RPC will work out a downloading schedule, which tells it how much chunks it can download from which RPCs. Moreover, the helper of the mobile car in mobile cloud would deliver specified chunks to it when there is lack of RPC in the following trip. Simulation results show that cloud-based scheme performs better than inter-vehicle communication approach and cluster-based scheme.

Multiple Vehicle Detection and Tracking in Highway Traffic Surveillance Video Based on SIFT Feature Matching

This paper presents a complete method for vehicle detection and tracking in a fixed setting based on computer vision. Vehicle detection is performed based on Scale Invariant Feature Transform (SIFT) feature matching. With SIFT feature detection and matching, the geometrical relations between the two images is estimated. Then, the previous image is aligned with the current image so that moving vehicles can be detected by analyzing the difference image of the two aligned images. Vehicle tracking is also performed based on SIFT feature matching. For the decreasing of time consumption and maintaining higher tracking accuracy, the detected candidate vehicle in the current image is matched with the vehicle sample in the tracking sample set, which contains all of the detected vehicles in previous images. Most remarkably, the management of vehicle entries and exits is realized based on SIFT feature matching with an efficient update mechanism of the tracking sample set. This entire method is proposed for highway traffic environment where there are no non-automotive vehicles or pedestrians, as these would interfere with the results.

Video Surveillance for Vehicle Classification using Hybrid SVM and Enhanced Ensemble Classifier

Video surveillance has obtained significant growth in the application oriented research fields especially in monitoring and security systems. This paper focuses on utilizing video surveillance in roadside traffic areas for identifying the vehicles such as ambulances in order to provide better road safety especially during emergency situations. The analyses of these situations are done using the video cameras fixed at roadsides for the surveillance purposes with a distributed computing process. These video cameras provide larger video files which are to be processed using efficient machine learning approaches for enhanced vehicle tracking process. In this paper, a road traffic video surveillance system is proposed for automatically identifying the road accidents from the live video files. The proposed system utilizes Hadoop-Map Reduce framework for processing the video files distributive and also in parallel. The distributed video files are enhanced using Gaussian filtering. The background subtraction is carried out Markov Random field (MRF) with Bayesian estimation process along with shadow removal and mist removal for environmental disturbances. The scene classification is carried out using Linear Discriminant Analysis (LDA) while a Hybrid Support Vector Machine (HSVM) combining Kalman filter with the SVM is employed for tracking the vehicles. Thus the proposed method efficiently track and classify the foreground objects with use of object classification and detection in traffic scenes especially in identifying the ambulances and heavy vehicles. The results obtained from the experiments on the proposed research shows the efficiency of traffic monitoring using traffic scenes.

Parallel Key Frame Extraction for Surveillance Video Service in a Smart City.

Surveillance video service (SVS) is one of the most important services provided in a smart city. It is very important for the utilization of SVS to provide design efficient surveillance video analysis techniques. Key frame extraction is a simple yet effective technique to achieve this goal. In surveillance video applications, key frames are typically used to summarize important video content. It is very important and essential to extract key frames accurately and efficiently. A novel approach is proposed to extract key frames from traffic surveillance videos based on GPU (graphics processing units) to ensure high efficiency and accuracy. For the determination of key frames, motion is a more salient feature in presenting actions or events, especially in surveillance videos. The motion feature is extracted in GPU to reduce running time. It is also smoothed to reduce noise, and the frames with local maxima of motion information are selected as the final key frames. The experimental results show that this approach can extract key frames more accurately and efficiently compared with several other methods.

APPOS: An adaptive partial occlusion segmentation method for multiple vehicles tracking

In traffic surveillance videos, it is common that the vehicles are occluded partially by each other. Such kind of occlusion situation is a challengeable task in multiple vehicles tracking. Various solutions in dealing with the occlusion for vehicles tracking have been proposed in many literatures. However, most of them are specialized on one tracking method and cannot flexibly adapt to the others. In this paper, we propose an adaptive partial occlusion segmentation method (APPOS) for multiple vehicles tracking. In this method, the occlusion detection process is firstly conducted to discover the occlusion. After that, the candidate regions of the respective occluded vehicles are roughly evaluated by the contour’s optical flow. Finally, the line scanning which uses color contrast among regions is adopted to accurately locate the vehicles. We evaluate the effectiveness and accuracy of APPOS by the experiments on practical and simulating videos.

Anomalous Event Detection in Traffic Video Based on Sequential Temporal Patterns of Spatial Interval Events

Detection of anomalous events from video streams is a challenging problem in many video surveillance applications. One such application that has received significant attention from the computer vision community is traffic video surveillance. In this paper, a Lossy Count based Sequential Temporal Pattern mining approach (LC-STP) is proposed for detecting spatio-temporal abnormal events (such as a traffic violation at junction) from sequences of video streams. The proposed approach relies mainly on spatial abstractions of each object, mining frequent temporal patterns in a sequence of video frames to form a regular temporal pattern. In order to detect each object in every frame, the input video is first pre-processed by applying Gaussian Mixture Models. After the detection of foreground objects, the tracking is carried out using block motion estimation by the three-step search method. The primitive events of the object are represented by assigning spatial and temporal symbols corresponding to their location and time information. These primitive events are analyzed to form a temporal pattern in a sequence of video frames, representing temporal relation between various object’s primitive events. This is repeated for each window of sequences, and the support for temporal sequence is obtained based on LC-STP to discover regular patterns of normal events. Events deviating from these patterns are identified as anomalies. Unlike the traditional frequent item set mining methods, the proposed method generates maximal frequent patterns without candidate generation. Furthermore, experimental results show that the proposed method performs well and can detect video anomalies in real traffic video data.

Probabilistic neural networks based moving vehicles extraction algorithm for intelligent traffic surveillance systems

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 61.75% and 69.38% higher than the other compared methods, respectively.

Robust vehicle tracking algorithm for nighttime videos captured by fixed cameras in highly reflective environments

In this study, the problem of vehicle detection, tracking and speed estimation in the nighttime traffic surveillance videos captured in highly reflective environments is considered. In this case, a robust algorithm is proposed which uses vehicle headlights as their prominent features. The proposed algorithm consists of three main stages. In the first stage, bright objects are segmented by thresholding the grey‐scale image. An effective algorithm is then applied to distinguish between vehicles lights and lights reflected on the road and on the vehicles bodies. In the second stage, the segmented bright objects are tracked using their spatial characteristics and their shapes and then, their speeds are estimated. To correct the camera perspective effect and reduce computational complexity, a projective transformation is used. In the third stage, the lights of each vehicle are grouped and paired using their positions and speeds. Motorbikes are also identified among the unpaired lights in this stage. Finally, the proposed real‐time system is implemented in C and applied to videos captured by traffic surveillance cameras in some highways in Iran. Experimental results reveal that accuracy of the algorithm proposed for vehicle detection is more than 98%.

Vehicles overtaking detection using RGB-D data

Outstanding breakthrough in traffic video surveillance has been made with the development of computer vision techniques. However, there are still some problems to be solved due to the limitations of two-dimensional (2D) video data. With the popularity of three-dimensional (3D) cameras, high-quality color and depth data can be obtained simultaneously in real time through the video stream. In this paper, we propose a vehicles overtaking detection method using RGB-D data captured by the Kinect device in simulated traffic scenes. Vehicles are detected and tracked with a robust traffic scene understanding on RGB-D data. The depth data is utilized to recognize vehicles overtaking by analyzing the posture change of vehicles in different scenes. The principle of vehicles overtaking detection is to fit the line of vehicle side in the coordinate system and calculate the angle between the vehicle side and the road orientation. As a consequence, the posture change of the vehicle can be recognized. The proposed method is evaluated using simulation experiments which show that our work has good performance in vehicles overtaking detection and tracking.

Video traffic analysis using scene and vehicle models

This paper presents a complete system for analyzing a vehicle׳s behavior in the context of real-time traffic video surveillance applications. To obtain the best possible results, it is fundamental to exploit the scene characteristics and the predefined traffic rules. For that purpose, an initial training step is performed that involves estimating the geometrical structure of the road, i.e., the depth relative to the camera, the vanishing points, the road areas itself, the road decomposition (into normal and forbidden traffic lanes or areas), the traffic rules, the typical vehicle trajectories and speeds, and the lane-changing rules. This process leads to a scene model, which is used together with a simple vehicle geometrical model during the vehicle detection, tracking and trajectory estimation phases to improve the robustness against the perspective and occlusion effects. Shadow effects are also accounted for during the moving object detection phase. Finally, this spatio-temporal analysis is used to obtain information that concerns the vehicles׳ behaviors. Experiments show that the information obtained is reliable and can be computed in real time.