Traffic Surveillance Applications Research Articles

Unmanned Aerial Vehicle–Based Traffic Analysis: Methodological Framework for Automated Multivehicle Trajectory Extraction

Unmanned aerial vehicles (UAVs), commonly referred to as drones, are one of the most dynamic and multidimensional emerging technologies of the modern era. This technology has recently found multiple potential applications within the transportation field, ranging from traffic surveillance applications to traffic network analysis. To conduct a UAV-based traffic study, extremely diligent planning and execution are required followed by an optimal data analysis and interpretation procedure. In this study, however, the main focus was on the processing and analysis of UAV-acquired traffic footage. A detailed methodological framework for automated UAV video processing is proposed to extract the trajectories of multiple vehicles at a particular road segment. Such trajectories can be used either to extract various traffic parameters or to analyze traffic safety situations. The proposed framework, which provides comprehensive guidelines for an efficient processing and analysis of a UAV-based traffic study, comprises five components: preprocessing, stabilization, georegistration, vehicle detection and tracking, and trajectory management. Until recently, most traffic-focused UAV studies have employed either manual or semiautomatic processing techniques. In contrast, this paper presents an in-depth description of the proposed automated framework followed by a description of a field experiment conducted in the city of Sint-Truiden, Belgium. Future research will mainly focus on the extension of the applications of the proposed framework in the context of UAV-based traffic monitoring and analysis.

Toward Accurate Device-Free Wireless Localization With a Saddle Surface Model

Device-free wireless localization (DFL) is a technique that can locate a target by analyzing its shadowing effect on wireless links, which causes the variation of link measurements, while removing the requirement of equipping the target with a device. It can provide fundamental data for pervasive computing, smart environment, and traffic surveillance applications. The observation model, which represents the relationship between wireless link measurement and target location, is very important for DFL, since it characterizes the shadowing effect of the target on wireless links and, therefore, determines the performance of the DFL system. In this paper, inspired by measurement results, we propose a saddle surface (SaS) model to describe the shadowing effect. The SaS model characterizes the elaborate information within the spatial impact area and provides more useful observation information for the location estimation algorithm. We incorporate the SaS model into the particle filter framework for location estimation. Extensive experiments in indoor and outdoor scenarios are carried out to evaluate the performance of the proposed schemes. The tracking errors of 0.78 and 0.21 m in the given two scenarios demonstrate the better performance of the proposed SaS model compared with existing models.

A Multi-view Learning Approach to Foreground Detection for Traffic Surveillance Applications

This paper proposes an effective multi-view learning approach to foreground detection for traffic surveillance applications. This approach involves three main steps. First, a reference background image is generated via temporal median filtering, and multiple heterogeneous features (including brightness variation, chromaticity variation, and texture variation, each of which represents a unique view) are extracted from the video sequence. Then, a multi-view learning strategy is devised to online estimate the conditional probability densities for both the foreground and the background. The probability densities of three features are approximately conditionally independent and are estimated with kernel density estimation. Pixel soft labeling is conducted by using Bayes rule, and the pixelwise foreground posteriors are computed. Finally, a Markov random field is constructed to incorporate the spatiotemporal context into the foreground/background decision model. The belief propagation algorithm is used to label each pixel of the current frame. Experimental results verify that the proposed approach is effective to detect foreground objects from challenging traffic environments and outperforms some state-of-the-art methods.

Modeling Fast Boat Motion Impact on Satellite SAR MTI Systems

This letter analyzes the problems of imaging small and fast boats with synthetic aperture radar (SAR), which may experience severe defocusing and/or smearing. This situation is encountered, inter alia , in ocean traffic monitoring and surveillance applications where the detection of small and fast vessels is of great interest. The target modeling and impact of moving vessels in SAR images are presented and discussed using simulated results. A matched filter bank is used before applying moving target indication techniques over multichannel SAR images, refocusing the target signal in the azimuth domain and enhancing the detectability of these types of vessels.

A Genetic Algorithm-Based Moving Object Detection for Real-time Traffic Surveillance

Recent developments in vision systems such as distributed smart cameras have encouraged researchers to develop advanced computer vision applications suitable to embedded platforms. In the embedded surveillance system, where memory and computing resources are limited, simple and efficient computer vision algorithms are required. In this letter, we present a moving object detection method for real-time traffic surveillance applications. The proposed method is a combination of a genetic dynamic saliency map (GDSM), which is an improved version of dynamic saliency map (DSM) and background subtraction. The experimental results show the effectiveness of the proposed method in detecting moving objects.

Fully Automatic Roadside Camera Calibration for Traffic Surveillance

This paper deals with automatic calibration of roadside surveillance cameras. We focus on parameters necessary for measurements in traffic-surveillance applications. Contrary to the existing solutions, our approach requires no a priori knowledge, and it works with a very wide variety of road settings (number of lanes, occlusion, quality of ground marking), as well as with practically unlimited viewing angles. The main contribution is that our solution works fully automatically-without any percamera or per-video manual settings or input whatsoever-and it is computationally inexpensive. Our approach uses tracking of local feature points and analyzes the trajectories in a manner based on cascaded Hough transform and parallel coordinates. An important assumption for the vehicle movement is that at least a part of the vehicle motion is approximately straight-we discuss the impact of this assumption on the applicability of our approach and show experimentally that this assumption does not limit the usability of our approach severely. We efficiently and robustly detect vanishing points, which define the ground plane and vehicle movement, except for the scene scale. Our algorithm also computes parameters for radial distortion compensation. Experiments show that the obtained camera parameters allow for measurements of relative lengths (and potentially speed) with ~2% mean accuracy. The processing is performed easily in real time, and typically, a 2-min-long video is sufficient for stable calibration.

Ghost Vehicle and Shadow Removal Approach for Traffic Surveillance and Monitoring at Various Intersections Using Computer Vision

As traffic surveillance technology continues to grow worldwide vehicle detection, counting, tracking and classification are gaining importance. This paper proposes computer vision based real time vehicle detection, tracking and classification at urban intersections. Firstly, foreground extraction using double subtraction method is proposed, which increases the accuracy of blob detection. Classification based on the geometrical attributes of the vehicle and also quadrant division of the junction is put forward. Setting up dynamic ROIs is discussed, which increased the scope of traffic surveillance for different types of junctions. The proposed system is implemented using Intel Open CV library for image processing and video processing applications. The Practical implementation of the algorithm is made with C++ and computer vision. Several junction surveillance videos are used to evaluate the performance of the traffic surveillance system. In this paper, detection, tracking and classification of objects in with removal of shadow and ghost vehicles at different junction in video surveillance .Proposed work elaborated computer vision approach for Traffic monitoring in traffic surveillance application. Test results in the performance of the proposed algorithm in detection, classification and counting and proved the effectiveness of the traffic surveillance system. Obtained results showed a better performance in terms of accuracy.

Integrating 3D structure into traffic scene understanding with RGB-D data

RGB Video now is one of the major data sources of traffic surveillance applications. In order to detect the possible traffic events in the video, traffic-related objects, such as vehicles and pedestrians, should be first detected and recognized. However, due to the 2D nature of the RGB videos, there are technical difficulties in efficiently detecting and recognizing traffic-related objects from them. For instance, the traffic-related objects cannot be efficiently detected in separation while parts of them overlap, and complex background will influence the accuracy of the object detection. In this paper, we propose a robust RGB-D data based traffic scene understanding algorithm. By integrating depth information, we can calculate more discriminative object features and spatial information can be used to separate the objects in the scene efficiently. Experimental results show that integrating depth data can improve the accuracy of object detection and recognition. We also show that the analyzed object information plus depth data facilitate two important traffic event detection applications: overtaking warning and collision avoidance.

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.

Using Traffic Light Signal to Enhance Intersection Foreground Detection Based on Video Sensor Networks

Foreground detection plays an important role in the traffic surveillance applications, especially in urban intersections. Background subtraction is an efficient approach to segment the background and foreground with static cameras from video sensor networks. But when modelling the background, most statistical techniques adjust the learning rate only based on the changes from video sequences, which is a crucial parameter controlling the updating speed. This causes a slow adaptation to sudden environmental changes. For example, a stopped car fuses into background before moving again, and it lowers the segmentation performance. This paper proposes an efficient way to address the problem by accounting for the physical world signal in traffic junctions. It assigns an adaptive learning rate to each pixel by integrating traffic light signal obtained from sensor networks. Combined with abundant physical world signals, background subtraction method is able to adapt itself to the outside world changes instantly. We test our approach in real urban traffic intersection; experimental results show that the new method increases the accuracy of detection and has a promising future.

Using Polynomial Wigner–Ville Distribution for Velocity Estimation in Remote Toll Applications

This letter presents the use of polynomial Wigner-Ville distribution (PWVD) for accurate velocity estimation as used in remote road traffic management applications. In such applications based on inverse synthetic aperture radar, velocity estimation is central to obtain a suitable level of performance of the signal processing. Moreover, the precision of this velocity estimation is crucial in order to achieve the best detection and estimation of the gauge of the vehicles through the use of radar images. Hence, the PWVD is applied as an instantaneous frequency estimator used in this traffic surveillance application.

A Survey of Visual Traffic Surveillance Using Spatio-Temporal Analysis and Mining

The focus of this survey is on spatio-temporal data mining and database retrieval for visual traffic surveillance systems. In many traffic surveillance applications, such as incident detection, abnormal events detection, vehicle speed estimation, and traffic volume estimation, the data used for reasoning is really in the form of spatio-temporal data (e.g. vehicle trajectories). How to effectively analyze these spatio-temporal data to automatically find its inherent characteristics for different visual traffic surveillance applications has been of great interest. Examples of spatio-temporal patterns extracted from traffic surveillance videos include, but are not limited to, sudden stops, harsh turns, speeding, and collisions. To meet the different needs of various traffic surveillance applications, several application- or event- specific models have been proposed in the literature. This paper provides a survey of different models and data mining algorithms to cover state of the art in spatio-temporal modelling, spatio-temporal data mining, and spatio-temporal retrieval for traffic surveillance video databases. In addition, the database model issues and challenges for traffic surveillance videos are also discussed in this survey.

Practical Camera Calibration From Moving Objects for Traffic Scene Surveillance

We address the problem of camera calibration for traffic scene surveillance, which supplies a connection between 2-D image features and 3-D measurement. It is helpful to deal with appearance distortion related to view angles, establish multiview correspondences, and make use of 3-D object models as prior information to enhance surveillance performance. A convenient and practical camera calibration method is proposed in this paper. With the camera height H measured as the only user input, we can recover both intrinsic and extrinsic parameters of the camera based on redundant information supplied by moving objects in monocular videos. All cases of traffic scene layouts are considered and corresponding solutions are given to make our method applicable to almost all kinds of traffic scenes in reality. Numerous experiments are conducted in different scenes, and experimental results demonstrate the accuracy and practicability of our approach. It is shown that our approach can be effectively adopted in all kinds of traffic scene surveillance applications.

Modified Unsupervised Image Segmentation based on Gaussian Mixture Model for Traffic Surveillance Applications

paper deals with an efficient image segmentation algorithm for video images which is quite useful for video based traffic surveillance applications. It includes video segmentations, morphological operations and labeling. In the field of surveillance system, effective video object segmentation is a conveyance for video analysis and processing. It presents a new algorithm for video object segmentation i.e. unsupervised image segmentation based on Gaussian mixture model with modified EM procedure. It uses the spatial unsupervised GMM clustering technique in which the objective function is modified or the prior term is added in the Bayesian. Firstly we use EM algorithm to estimate the distribution of input image data with which the number of mixture components is automatically determined. Secondly the segmentation is arrived at by clustering each pixel into the appropriate component according to the Minimum Message Length (MML) criterion with the help of appropriate priors like Dirichlet-Normal-Wishart (DNW) prior. The proposed technique automatically decides the best number of clusters for images. The best number of clusters is obtained by using the cluster validity criterion with the help of Gaussian distribution.

Image tracking of motorcycles and vehicles on urban roads and its application to traffic monitoring and enforcement

Image tracking has increasingly gained attention for use in vision‐based traffic monitoring and surveillance applications. For many cities in Asia countries, it is desirable to detect multiple motorcycles as well as cars for urban traffic monitoring and enforcement. In this paper, a novel contour initialization and tracking algorithm is presented to track multiple motorcycles and vehicles at any position on the roadway. This method has the capability to detect moving vehicles of various sizes and to generate their initial contours for image tracking. The proposed method is not constrained by lane boundaries or vehicle size. To track vehicles on roadways, dynamic models are designed to predict the horizontal and vertical positions of vehicle contours. A Kalman filter is designed to update the prediction based on real‐time image measurement. Practical experimental studies using video clips are presented to evaluate the performance of the proposed method. Traffic parameters such as traffic flow, vehicle speeds and traffic density are obtained with satisfactory accuracy.

Estimating Speed and Direction of Small Dynamic Targets through Optical Satellite Imaging

Moving Target Indicators (MTI) are systems used to distinguish movement from stationary scenes and sometimes to derive the spatial attributes of these objects. These systems are currently used in many sectors such as traffic studies, border surveillance, and military applications. The proposed MTI reveals vehicles and their velocities using commercial imagery from a passive optical satellite-mounted sensor. With simple process of image differencing, the MTI can automatically recognize conveyances in motion (speed and direction) represented by polygons formed by a group of pixels from successive images. Micro-change detection with an existing commercial satellite requires special considerations of differences in spatial and spectral resolution between images. Complications involving the movement detection system such as vehicle overlap, vehicle clusters, and zones of low confidence are refined by adding error-reducing modules. This process is tested on a variety of vehicles, their concentrations, and environments, confirming the feasibility of utilizing an MTI with commercial optical satellite imagery for movement recognition and velocity estimation.

Moving target tracking and measurement with a binocular vision system

Visual tracking and measurement is one of the most important tasks in computer vision and finds applications in traffic surveillance, vision-guided mobile robots etc. In this paper, a binocular vision system is used to achieve the visual servoing, mimicking the movements of human eyes. First, a robust scheme, which combines adaptive background subtraction and Camshift algorithms, is proposed to detect and track a moving object. Then, we discuss the problems of camera parameter calibration and position measurement of moving objects. A new approach is introduced to realize the extrinsic parameter calibration of a pan-tilt camera, providing successful tracking and accurate measurement of a moving target, even as it is outside the view of camera. Numerous experiments have been done on the binocular vision system of a humanoid mobile platform, and the results show that the proposed approaches work very well.

Multiple-Target Tracking and Identity Management withApplication to Aircraft Tracking

The problem of tracking and managing the identities of multiple targets is discussed and applied to the passive radar tracking of aircraft. The targets are assumed to be commercial aircraft switching modes during flight, and are thus well modeled by hybrid systems. We propose a computationally efficient algorithm based on joint probabilistic data association for target-measurement correlation. We use the results of this algorithm to simultaneously implement an identity management algorithm based on identity-mass flow, and a multiple-target tracking algorithm based on the residual-mean interacting multiple model algorithm. Together, they constitute the multiple-target tracking and identity management algorithm. The multiple-target tracking and identity management algorithm incorporates suitable local information about target identity, when available, in a manner that decreases the uncertainty in the system as measured by its statistical entropy. For situations in which local information is not explicitly available, a technique based on multiple hypothesis testing is proposed to infer such information. This algorithm allows us to track multiple targets, each capable of multiple modes of operation, in the presence of continuous process noise and of spurious measurements. The multiple-target tracking and identity management algorithm is demonstrated through various scenarios that are motivated by air traffic surveillance applications.

Use of Path Objects for Air Traffic Control

A path object is a set of instructions plus the values of associated parameters that would be used by an aircraft’s flight management system (FMS) or area navigation (RNAV) computer to construct a flight trajectory on the basis of the values of the parameters provided by the pilot or air traffic control (ATC) system. The concept of path objects and its applicability for development of FMS- or RNAV-based flight paths for aircraft routes are described. The use of path objects requires only a small number of parameters for specification of an RNAV route, and the computer then calculates the resulting flight path. Because most RNAV routes are just variations of the same basic shapes, many RNAV routes can be generated from a single path object by changing just one or two parameters. This offers the capability to reduce the size of the navigation database by a significant amount. This also allows the dynamic alteration of three-dimensional FMS-RNAV routes instead of radar vectoring, which represents a significant improvement in ATC procedures. Because of the compact expressions for the path objects, these procedures could be used in a voice communications environment as well as a data link environment. A path object can be thought of as a high-level language with which aircraft and ATC systems communicate flight path intentions. The concept of FMS- or RNAV-stored path objects is a change in thinking about the role of avionics technology. It offers a means of including the precision of the FMS-RNAV technology in the ATC system during the transition from today’s system to the future’s fully automated control system. This is an enabling technology that improves the ability to use and maintain FMS-RNAV in its role as a vital component of the ATC system. It has the potential to enhance future air traffic management-communication navigation surveillance concepts and applications.

Lane detection by orientation and length discrimination

This paper describes a novel lane detection algorithm for visual traffic surveillance applications under the auspice of intelligent transportation systems. Traditional lane detection methods for vehicle navigation typically use spatial masks to isolate instantaneous lane information from on-vehicle camera images. When surveillance is concerned, complete lane and multiple lane information is essential for tracking vehicles and monitoring lane change frequency from overhead cameras, where traditional methods become inadequate. The algorithm presented in this paper extracts complete multiple lane information by utilizing prominent orientation and length features of lane markings and curb structures to discriminate against other minor features. Essentially, edges are first extracted from the background of a traffic sequence, then thinned and approximated by straight lines. From the resulting set of straight lines, orientation and length discriminations are carried out three-dimensionally with the aid of two-dimensional (2-D) to three-dimensional (3-D) coordinate transformation and K-means clustering. By doing so, edges with strong orientation and length affinity are retained and clustered, while short and isolated edges are eliminated. Overall, the merits of this algorithm are as follows. First, it works well under practical visual surveillance conditions. Second, using K-means for clustering offers a robust approach. Third, the algorithm is efficient as it only requires one image frame to determine the road center lines. Fourth, it computes multiple lane information simultaneously. Fifth, the center lines determined are accurate enough for the intended application.