Vehicle Tracking Algorithm Research Articles

Bird’s eye view localization of surrounding vehicles: Longitudinal and lateral distance estimation with partial appearance

On-road vehicle detection is essential for perceiving driving settings, and localizing the detected vehicle helps drivers predict possible risks and avoid collisions. However, there are limited works on vehicle detection with partial appearance, and the method for partially visible vehicle localization has not been explored. In this paper, a novel framework for vehicle detection and localization with partial appearance is proposed using stereo vision and geometry. First, the original images from the stereo camera are processed to form a v-disparity map. After object detection using v-disparity, vehicle candidates are generated with prior knowledge of possible vehicle locations on the image. Deep learning-based verification completes vehicle detection. For each detected vehicle, partially visible vehicle tracking algorithm is newly introduced. To track partially visible vehicles, this algorithm detects the vehicle edge on the ground, defined as the grounded edge, and then selects a reference point for Kalman filter tracking. Finally, a rectangular box is drawn on the bird’s eye view to represent vehicle’s longitudinal and lateral location. The proposed system successfully performs partially visible vehicle detection and tracking. For testing the localization performance, the datasets in a highway and an urban setting are used and provide less than 1.5 m longitudinal error and 0.4 m lateral error in standard deviation.

Particle filter‐based vehicle tracking via HOG features after image stabilisation in intelligent drive system

Tracking vehicles automatically are of great importance in the intelligent transportation systems to guarantee the traffic safety, especially for the self-driving technique. To solve the problem of the tiny jitter of the camera due to the shaking platform and the rough road, and to track the vehicle ahead, a salient feature-based video stabilisation method is introduced to remove the effect of camera motion, and an improved particle filter-based vehicle tracking algorithm via histogram of oriented gradient (HOG) features is proposed in this study. For video stabilisation, first, the features from accelerated segment test algorithm are applied to extract salient points from the two adjacent frames. Then, the fast retina keypoint algorithm is used to match the correspondences and the M-estimator sample consensus algorithm is applied to remove the outliers. Finally, the image with tiny jitter is stabilised by the affine transformation matrix calculated from the retained inliers. After the image stabilisation, an improved particle filter-based tracking method is proposed for vehicle tracking via the HOG feature extracted from an adaptive searching area determined by the expected vehicle position of the previous frame. Experimental results demonstrate that camera motion can be effectively compensated by the feature-based stabilisation method, and the forward vehicle can be tracked with stability and robustness in real time.

Real-time Vehicle Detection and Tracking Algorithm for Forward Vehicle Collision Warning

Real-time Vehicle Detection and Tracking Algorithm for Forward Vehicle Collision Warning

An Efficient Monitoring of Real Time Traffic Clearance for an Emergency Service Vehicle Using IOT

In this paper, a real time emergency vehicle tracking through architecture of instinctive emergency recognition system with high-intensity digital camera it is located in a national high way traffic signal. In a recent survey, there are thousands of people losing their lives due to the delay in the emergency services. Resent survey say that more than 4000 heart attack victims can be saved each year if the delay could be minimized and in the present scenario the number of deaths is in lakhs and this number can be effectively reduced by providing timely and accurate emergency service all the way through avoiding the unnecessary time delay near traffic jams during an emergency situation (Tagne et al. in IEEE Trans Intell Transp Syst 17(3):796–809, 2015). This method clarifies the modeling and working of different units of the emergency vehicle identification system such that optimized emergency vehicle tracking algorithm, with the traffic supervision unit. In this article confer the basic components and their function such that internet of things and their dissimilar layers of protocol, raspberry pi and its architecture, with interfacing sensors such as Siren sound detect (REES-52), the Wireless sensor (NRF905se). It is direct and the most efficient route that is asphalt construct to the central server checks for the location of the vehicle and change the traffic signal. When the emergency vehicle is approaching the traffic lights. The system generate in sequence regarding the traffic emergency situations such as ambulance and siren sound (Emergency Services Review: A comparative review of international ambulance service best practice. http://www.aace.org.uk. Accessed 28 July 2017, 2017). The processed information can be used to divert the live traffic as desirable to avoid the problems related to real time road traffic (Sundar et al. in Sens J IEEE 15(2):1109–1113, 2015).

Research on Video Image Vehicle Tracking Algorithm Based on Random Forest Method

On the basis of apparent tracking theory of discriminant model, we put forward a vehicle video tracking improvement algorithm based on the random forest classifier to seek a better classifier confidence figure, so that we can improve the tracking accuracy. After getting the degree of confidence of the rectangular image block predicted by the classifier, the pixels and the distance information between rectangular image blocks have different impacts on the degree of confidence. According to this feature, using the pixel confidence estimation method based on weighted distance information to improve the confidence estimation accuracy of pixel, and then improve the tracking accuracy.

Infrared measurement and composite tracking algorithm for air-breathing hypersonic vehicles

Air-breathing hypersonic vehicles have capabilities of hypersonic speed and strong maneuvering, and thus pose a significant challenge to conventional tracking methodologies. To achieve desirable tracking performance for hypersonic targets, this paper investigates the problems related to measurement model design and tracking model mismatching. First, owing to the severe aerothermal effect of hypersonic motion, an infrared measurement model in near space is designed and analyzed based on target infrared radiation and an atmospheric model. Second, using information from infrared sensors, a composite tracking algorithm is proposed via a combination of the interactive multiple models (IMM) algorithm, fitting dynamics model, and strong tracking filter. During the procedure, the IMMs algorithm generates tracking data to establish a fitting dynamics model of the target. Then, the strong tracking unscented Kalman filter is employed to estimate the target states for suppressing the impact of target maneuvers. Simulations are performed to verify the feasibility of the presented composite tracking algorithm. The results demonstrate that the designed infrared measurement model effectively and continuously observes hypersonic vehicles, and the proposed composite tracking algorithm accurately and stably tracks these targets.

Vehicle Positioning and Tracking Algorithm in the Visual Positioning System in Ship Cabin

A large cargo ship is widely used for vehicle transporting, and it is necessary to position vehicles in the cabin. For the characteristic of large space, low height of ship cabin and large vehicle, building a multi-camera positioning system for vehicles is the optimal one. In this paper, a high accuracy vehicle positioning algorithm is proposed to locate multiple moving vehicles in the positioning system. Firstly, the conversion relationship between the pixel coordinate and the world coordinate is calculated. Zhang’s calibration method is adopted to correct the lens distortion, the images are stitched to get the complete cabin image and the height of vehicle body is calibrated. Secondly, a method for vehicle positioning and tracking is proposed. Background subtraction method is adopted to detect the preliminary vehicle position, and Spatio-Temporal Context (STC) algorithm is adopted to track the mark on the vehicle body. The experimental results show that in 2.56m*0.62m space of the simulated ship cabin, for three vehicle models with size of 0.45m*0.14m*0.15m, the horizontal positioning accuracy can reach 2cm each, less than 1%. With the proposed algorithm, multiple moving vehicles positioning with high accuracy and high precision in ship cabin can be reached.

A system for the generation of synthetic Wide Area Aerial surveillance imagery

The development, benchmarking and validation of aerial Persistent Surveillance (PS) algorithms requires access to specialist Wide Area Aerial Surveillance (WAAS) datasets. Such datasets are difficult to obtain and are often extremely large both in spatial resolution and temporal duration. This paper outlines an approach to the simulation of complex urban environments and demonstrates the viability of using this approach for the generation of simulated sensor data, corresponding to the use of wide area imaging systems for surveillance and reconnaissance applications. This provides a cost-effective method to generate datasets for vehicle tracking algorithms and anomaly detection methods. The system fuses the Simulation of Urban Mobility (SUMO) traffic simulator with a MATLAB controller and an image generator to create scenes containing uninterrupted door-to-door journeys across large areas of the urban environment. This ‘pattern-of-life’ approach provides three-dimensional visual information with natural movement and traffic flows. This can then be used to provide simulated sensor measurements (e.g. visual band and infrared video imagery) and automatic access to ground-truth data for the evaluation of multi-target tracking systems.

Research on Intelligent Recognition System for Traffic Violation

This paper focuses on the intelligent recognition system of traffic violation on the basis of analyzing urban complex monitoring environment. For the realization that the system integrates the forensics of violation with license plate recognition in all day time, this paper proposes a set of efficient and robust vehicle detection and tracking algorithms. Vehicle detection was realized by the visual features of the license plate and color characteristic of the taillight to detect edge differential motion between image frames. In vehicle tracking, the extraction of large-scale fast stable points by Lucas-Kanade Pyramid method and predictions for tracking taillight based on Kalman method are proposed to realize long-distance stable tracking vehicles in whole day time. The practical applications show the system has excellent performance.

Real Time Vehicle Tracking Algorithm Based on Statistical Characteristics in HSI Color Model

Real Time Vehicle Tracking Algorithm Based on Statistical Characteristics in HSI Color Model

Visual Vehicle Tracking Based on Deep Representation and Semisupervised Learning

Discriminative tracking methods use binary classification to discriminate between the foreground and background and have achieved some useful results. However, the use of labeled training samples is insufficient for them to achieve accurate tracking. Hence, discriminative classifiers must use their own classification results to update themselves, which may lead to feedback-induced tracking drift. To overcome these problems, we propose a semisupervised tracking algorithm that uses deep representation and transfer learning. Firstly, a 2D multilayer deep belief network is trained with a large amount of unlabeled samples. The nonlinear mapping point at the top of this network is subtracted as the feature dictionary. Then, this feature dictionary is utilized to transfer train and update a deep tracker. The positive samples for training are the tracked vehicles, and the negative samples are the background images. Finally, a particle filter is used to estimate vehicle position. We demonstrate experimentally that our proposed vehicle tracking algorithm can effectively restrain drift while also maintaining the adaption of vehicle appearance. Compared with similar algorithms, our method achieves a better tracking success rate and fewer average central-pixel errors.

Likelihood-Field-Model-Based Dynamic Vehicle Detection and Tracking for Self-Driving

Dynamic vehicle detection and tracking is crucial for self-driving in urban environments. The main problem of the previous beam-model-based algorithms is that they cannot detect and track dynamic vehicles that are occluded by other objects. In this paper, we develop a novel dynamic vehicle detection and tracking algorithm to solve this problem for our autonomous land vehicle (ALV), which is equipped with a Velodyne LIDAR and a GPS-aid inertial navigation system. For detection, our improved two-dimensional virtual scan is presented to detect the potential dynamic vehicles with a scan differencing operation. Then, for each potential dynamic vehicle, a novel likelihood-field-based vehicle measurement model is proposed to weight its possible poses. Finally, our newly modified scaling series algorithm and the importance sampling technique are adopted to estimate the initial pose and the corresponding velocity for each vehicle, respectively. The scaling series algorithm coupled with a Bayesian filter (SSBF) was previously used to handle the tactile localization problem in static background scenes. For tracking dynamic vehicles, we improve the SSBF by adding the ego-motion compensation so that the improved algorithm is able to update the pose and velocity for each vehicle in dynamic background scenes. Both the quantitative and qualitative experimental results validate the performance of our dynamic vehicle detection and tracking algorithm on the KITTI datasets and the Velodyne data collected by our ALV in dynamic urban environments.

New preceding vehicle tracking algorithm based on optimal unbiased finite memory filter

In recent years, visual object tracking technologies have been used to track preceding vehicles in advanced driver assistance systems (ADASs). The accurate positioning of preceding vehicles in camera images allows drivers to avoid collisions with the preceding vehicle. Tracking systems typically take advantage of state estimators, such as the Kalman filter (KF) and the particle filter (PF), in order to suppress noises in measurements. In particular, the KF is popular in visual object tracking, because of its computational efficiency. However, the visual tracker based on the KF, referred to as the Kalman tracker (KT), has the drawback that its performance can decrease due to modeling and computational errors. To overcome this drawback, we propose a novel visual tracker based on the optimal unbiased finite memory filter (OUFMF) in the formulation of a linear matrix inequality (LMI) and a linear matrix equality (LME). We call the proposed visual tracker the finite memory tracker (FMT), and it is applied to the preceding vehicle tracking. Through extensive experiments, we demonstrate the FMT’s performance that is superior to that of the KT and other filter-based tracker.

MCMC 기반 파티클 필터를 이용한 지능형 자동차의 다수 전방 차량 추적 시스템

Abstract Intelligent vehicle plans motion and navigate itself based on t he surrounding environment perception. Hence, the precise environment recognition is an essential part of self-dr iving vehicle. There exist many vulnerable road users (e.g. vehicle, pedestrians) on vehicular driving environm ent, the vehicle must percept all the dynamic ob-stacles accurately for safety. In this paper, we propose an multiple vehicle tracking algorithm using microwave radar. Our proposed system includes various special features. F irst, exceptional radar measurement model for ve-hicle, concentrated on the corner, is described by mixture dens ity network (MDN), and applied to particle filte r weighting. Also, to conquer the curse of dimensionality of parti cle filter and estimate the time-varying number o fmulti-target states, reversible jump markov chain monte carlo ( RJMCMC) is used to sampling step of the proposed algorithm. The robustness of the proposed algorithm is demonstr ated through several computer simulations.Key Words : Intelligent Vehicle, Multiple Target Tracking (MTT), Markov C hain Monte Carlo (MCMC), Particle Filter, Mixture Density Network (MDN).Received: Jan. 30, 2014Revised : Apr. 14, 2015Accepted: Apr. 14, 2015

The Role of Path Continuity in Lateral Vehicle Control

Motion planning and tracking control are generally developed separately for autonomous vehicle's navigation. Paths are often designed to satisfy the kinematics of the vehicle. Geometric singularities in the reference trajectory deteriorate the performance of the path tracking controller. An integrated approach is needed in autonomous robotics research to account for the influence of the path properties on the controller behavior. The current research considers the controller performance and resulting stability and disturbances of the vehicle. The compound effect of planning and tracking control is investigated in this article. Systematic changes to the vehicle velocity, desired maneuver and tracking algorithm are made to evaluate the role of path planning continuity on the vehicle's dynamic responses and controller's performance. This research validates previous literature that showed enhancement in controller performance by using continuous reference trajectories. Our findings suggest that the use of cubic parametrically continuous spline paths improves passenger comfort and enhance vehicle's stability. The experimental results report a 90% reduction in both lateral acceleration and side slip.

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%.

차량의 부분 특징을 이용한 터널 내에서의 차량 검출 및 추적 알고리즘

In this paper, an efficient vehicle detection and tracking algorithm for detection incident in tunnel is proposed. The proposed algorithm consists of three steps. The first one is a step for background estimates, low computational complexity and memory consumption Running Gaussian Average (RGA) is used. The second step is vehicle detection step, Adaboost algorithm is applied to this step. In order to reduce false detection from a relatively remote location of the vehicles, local features according to height of vehicles are used to detect vehicles. If the local features of an object are more than the threshold value, the object is classified as a vehicle. The last step is a vehicle tracking step, the Kalman filter is applied to track moving objects. Through computer simulations, the proposed algorithm was found that useful to detect and track vehicles in the tunnel.

단일 레이저 스캐너와 Integrated Probabilistic Data Association Filter를 이용한 도심환경에서의 다중 차량추적

This paper describes a multiple vehicle tracking algorithm using an integrated probabilistic data association filter (IPDAF) in urban environments. The algorithm consists of two parts; a pre-processing stage and an IPDA tracker. In the pre-processing stage, measurements are generated by a feature extraction method that manipulates raw data into predefined geometric features of vehicles as lines and boxes. After that, the measurements are divided into two different objects, dynamic and static objects, by using information of ego-vehicle motion. The IPDA tracker estimates not only states of tracks but also existence probability recursively. The existence probability greatly assists reliable initiation and termination of track in cluttered environment. The algorithm was validated by using experimental data which is collected in urban environment by using single laser scanner.

Moving Vehicle Detection and Tracking Algorithm in Traffic Video

Aiming at the defects and shortages of traditional moving vehicles detection algorithms, by the analysis and comparison of the existing detection algorithms, we propose an algorithm that combined with frames with symmetric difference and background difference to detect moving vehicle in this paper. First, two different difference images by using frames with symmetric difference and background difference are gained respectively and two binary images can be gained by the appropriate threshold, then the contour of moving vehicles can be extracted by applying OR operation in the two binary images. Finally, the precise moving vehicles will be gained by mathematic morphological methods. In this paper we use Harris operator, Feature Points such as edges and corners are extracted, followed by block-matching to track the Feature Points in successive viedo frames. Many vehicles can be tracked at the same time automatically since the information is obtained from video sequences. DOI: http://dx.doi.org/10.11591/telkomnika.v11i6.2613

A Multi-Vehicle Tracking Algorithm Based on QPSO Clustering Algorithm and Particle Position Matching Template

In this paper,we introduce a new tracking algorithm to detect and track multiple Vehicles in highway or special area. For the traditional vehicle tracking algorithm that can not be real-time tracking of vehicles or appear to tracking chaotic situation when most of the scene is in motion and component parts of vehicles are occluded. To overcome this problem, we use a background image separation method to separate the moving vehicle and the background and introduce a QPSO clustering algorithm for gathering together color particles of vehicles moving target,finally we through the particle position matching template to achieve target matching tracking.In an experimental evaluation, we show that algorithm through processing the monitoring image of moving vehicles to show tracking information of moving vehicle in the real-time.