Particle Filter For Target Tracking Research Articles

Application of Particle Filter Algorithm Based on Gaussian Clustering in Dynamic Target Tracking

In order to solve the problem of particle depletion and complexity in traditional particle filtering, a particle filter target tracking method based on Gaussian clustering is proposed in this paper. Combined with adaptive double-sampling and gradmethod, the real time and robustness of dynamic tracking are improved significantly.

Occlusion-Aware Correlation Particle Filter Target Tracking Based on RGBD Data

In recent decades, there have been considerable improvements in target-tracking algorithms. However, aspects such as target occlusion, scale variation, and illumination changes still present significant challenges to existing algorithms. In this paper, we describe an occlusion-aware correlation particle filter target-tracking method based on RGBD data. First, we derive a target occlusion judgment mechanism based on a depth image and the histogram of oriented gradients (HOG) feature. We then formulate the tracking mechanism for the target prediction–tracking–optimization–redetection process using a correlation maximum likelihood estimation particle filter algorithm. We propose an adaptive update strategy whereby the system saves a well-tracked model when no occlusion occurs, and then uses this saved model to replace poorly tracked models in the event of occlusion. Furthermore, we consider the scale variation and adjust the target size according to the depth image, but we leave the HOG feature vector dimension of the target area unchanged. Thus, the problems such as model offset, scale variation, and loss of features are corrected over time. The experimental results demonstrate that the proposed target-tracking algorithm can detect target occlusion and track targets well, requires fewer calculations to perform target prediction–tracking–optimization–redetection, reduces the impact of illumination changes, and achieves better real-time performance and accuracy than many existing algorithms.

Particle Filter Target Tracking Algorithm Based on Dynamic Niche Genetic Algorithm

Particle Filter Target Tracking Algorithm Based on Dynamic Niche Genetic Algorithm

Distributed Particle Filter for Target Tracking: With Reduced Sensor Communications

For efficient and accurate estimation of the location of objects, a network of sensors can be used to detect and track targets in a distributed manner. In nonlinear and/or non-Gaussian dynamic models, distributed particle filtering methods are commonly applied to develop target tracking algorithms. An important consideration in developing a distributed particle filtering algorithm in wireless sensor networks is reducing the size of data exchanged among the sensors because of power and bandwidth constraints. In this paper, we propose a distributed particle filtering algorithm with the objective of reducing the overhead data that is communicated among the sensors. In our algorithm, the sensors exchange information to collaboratively compute the global likelihood function that encompasses the contribution of the measurements towards building the global posterior density of the unknown location parameters. Each sensor, using its own measurement, computes its local likelihood function and approximates it using a Gaussian function. The sensors then propagate only the mean and the covariance of their approximated likelihood functions to other sensors, reducing the communication overhead. The global likelihood function is computed collaboratively from the parameters of the local likelihood functions using an average consensus filter or a forward-backward propagation information exchange strategy.

Integrated particle filter for target tracking in clutter

In a typical surveillance situation the number and the trajectories of targets are a priori unknown. Each measurement has an unknown source; either clutter, a target being tracked or a new target. The tracks are initialised and updated using measurements, thus both true and false tracks exist at any given time. The authors present a particle filter approach which recursively calculates the probability of target existence, which may be used as a track quality measure for the false track discrimination. Single target, joint multitarget and linear multitarget versions are presented. The algorithms assume manoeuvring (multiple trajectory propagation models) targets and state dependent probability of target detection.

Particle Filter Target Tracking Algorithm Based on MCMC Iteration Cubature

In view of the problem of particle degradation and tracking accuracy in the standard particle filter tracking target algorithm, a new improved particle filter algorithm called Iterated Cubature Kalman Particle Filter (ICKPF) is proposed in this paper. The new ICKPF algorithm is based on the Markov Chain Monte Carlo (MCMC), and the cubature rule based on numerical integration method is used to calculate the mean and covariance, which generates the proposal distribution for the particle filter. The current measurements are integrated into the proposal distribution. Therefore, degree of approximation to the system posterior density is improved. Simulation results show that the estimation error of the ICKPF-MCMC algorithm is much better than other algorithms.

Distributed Box Particle Filtering for Target Tracking in Sensor Networks

Distributed target tracking is a significant technique and is widely used in many applications. Combined with the interval analysis, box particle filtering (BPF) has been proposed to solve the problem of Bayesian filtering when the uncertainties in the measurements are intervals; that is, the measurements are interval-based vectors. This paper is targeted for extending the existing BPF based on a single sensor to a distributed sensor network. We propose a distributed BPF (d-BPF) that each sensor communicates with its direct neighbors to collaboratively estimate the states of the target. The feasibility of the proposed distributed BPF is justified, and some numerical simulations are presented to show its effectiveness in target tracking.

A New Particle Filter Target Tracking Algorithm Based on Genetic Algorithm

Aimed to problem that particles exist degradation in particle filter (PF) algorithm for target tracking which is used in Wireless Sensor Networks, a new particle filter target tracking algorithm (GPF) based on genetic algorithm was proposed in this paper. First, based on the description of the PF algorithm, the factors that affect the degradation were ana- lyzed, which were importance function and re-sampling technique. Then, a new importance function which can make the current measured value take effect was constructed and the re-sampling technique was mended by using genetic algo- rithms in the GPF. Finally, the simulation tracking results showed that the GPF algorithm has better tracking accuracy than that of PF.

Particle Filter Target Tracking Algorithm Based on the SIFT and Color Features Fusion

Aiming at the poor robustness problem of using single feature in the target tracking process, a novel tracking algorithm based on color and SIFT features fusion in particle filter framework is presented in complex environments. Color and SIFT features are selected to establish the target model according to their stability, The scale and rotation invariance of SIFT feature and resistance occlusion property of color feature has been fused in the particle filter framework adaptively. According to the dynamic change of the tracking scene, the fusion weights is updated adaptively. Experimental results show the proposed method can track target robustly under complex scene in real-time performance.

Color and motion-based particle filter target tracking in a network of overlapping cameras with multi-threading and GPGPU

This paper describes an efficient implementation of multiple-target multiple-view tracking in video-surveillance sequences. It takes advantage of the capabilities of multiple core Central Processing Units (CPUs) and of graphical processing units under the Compute Unifie Device Arquitecture (CUDA) framework. The principle of our algorithm is 1) in each video sequence, to perform tracking on all persons to track by independent particle filters and 2) to fuse the tracking results of all sequences. Particle filters belong to the category of recursive Bayesian filters. They update a Monte-Carlo representation of the posterior distribution over the target position and velocity. For this purpose, they combine a probabilistic motion model, i.e. prior knowledge about how targets move (e.g. constant velocity) and a likelihood model associated to the observations on targets. At this first level of single video sequences, the multi-threading library Threading Buildings Blocks (TBB) has been used to parallelize the processing of the per-target independent particle filters. Afterwards at the higher level, we rely on General Purpose Programming on Graphical Processing Units (generally termed as GPGPU) through CUDA in order to fuse target-tracking data collected on multiple video sequences, by solving the data association problem. Tracking results are presented on various challenging tracking datasets.

Niche PSO Particle Filter with Particles Fusion for Target Tracking

An improved particle filter algorithm is proposed to track a randomly moving target in video. In particle filter framework, a particle swarm optimization improved by niche technique which implemented by restricted competition selection is integrated. It can move particles into high likelihood area of target and form multi-population distribution, so that the searching capability of particles is enhanced and then the adaptation to the change of dynamic target state is improved. The particles of niching particle swarm optimization and the particles of particle filter are integrated for new particle weight calculation and finally realize a new particle filter for target tracking in video sequence.

Fuzzy Particle Filter for Target Tracking

For the lack of the traditional particle filter and particle filter based on the colour histogram in the background illumination and the deformation and transformation of target, fuzzy particle filter combined with the classical Monte Carlo algorithm employs fuzzy technique to extract the target features. Colour image sequence converts from RGB to L*a*b Colour space of the International Illumination Commission. Fuzzy particle filter and re-sampling weight superimposed is more effective in reducing the number of particles and enhances traditional particle filter capacity to the recognition positioning and real-time tracking. Experimental results displayed that the method in the target shape, attitude or background illumination in intraday variation tracked target in real time. The algorithm enhances the traditional particle filter robustness and real-time in the light of dramatic changes, background constant variation and the target deformation self-adaptive. And the partially obscured target is not lost.

Compressed Iterative Particle Filter for Target Tracking

Particle filtering has been widely used in the non-linear n-Gaussian target tracking problems. The main problem of particle filtering is the lacking and exhausting of particles, and choosing effective proposed distribution is the key point to overcome it. In this paper, a new mixed particle filtering algorithm was proposed. Firstly, the unscented kalman filtering is used to generate the proposed distribution, and in the resample step, a new certain resample method is used to choose the particles with ordered larger weights. GA algorithm is introduced into the certain resample method to keep the variety of the particles. Simuational results have shown that the proposed algorithm has better performances than other three typical filtering algorithms.

Research on Stratified Re-Sampling Particle Filter Target Tracking Algorithm Based on Multiple Clues

To further improve the performance of video object tracking and overcome the instability brought by using a single measurement source, a new particle filter algorithm was proposed which combined the color with contour information. The methord used color likelihood model to predict object, geometric active contour model to update the particle, and Stratified Re-sampling method to overcome the particle degradation. Experimental results show that the proposed algorithm is more robust than the traditional algorithm based on a single clue in occlusion, even though the color of the object is very similar to the background, it still tracks the target accurately.

An Improved Particle Filter for Target Tracking in Sensor Systems

Sensor systems are not always equipped with the ability to track targets. Sudden maneuvers of a target can have a great impact on the sensor system, which will increase the miss rate and rate of false target detection. The use of the generic particle filter (PF) algorithm is well known for target tracking, but it can not overcome the degeneracy of particles and cumulation of estimation errors. In this paper, we propose an improved PF algorithm called PF-RBF. This algorithm uses the radial-basis function network (RBFN) in the sampling step for dynamically constructing the process model from observations and updating the value of each particle. With the RBFN sampling step, PF-RBF can give an accurate proposal distribution and maintain the convergence of a sensor system. Simulation results verify that PF-RBF performs better than the Unscented Kalman Filter (UKF), PF and Unscented Particle Filter (UPF) in both robustness and accuracy whether the observation model used for the sensor system is linear or nonlinear. Moreover, the intrinsic property of PF-RBF determines that, when the particle number exceeds a certain amount, the execution time of PF-RBF is less than UPF. This makes PF-RBF a better candidate for the sensor systems which need many particles for target tracking.

Improved Particle Filter for Target Tracking

A new improved particle filter algorithm with the simplified UT (unscented transformation) and the modified unscented Kalman filter (UKF) proposal distribution is presented. The scaling factor is added to adaptively estimate on line and to improve the filtering performance. An adaptive algorithm is developed. In the bearings-only tracking experiments, the results confirm the improved particle filter algorithm outperforms others.