الفهرس 
ABSTRACTOnly 14 pages are availabe for public view
Abstract
In the field of video tracking and detection, many researchers were seeking to get to the concept
of computer vision. Computer vision is a field that includes methods for acquiring, processing,
analyzing, and understanding images and, in general, high-dimensional data from the real world
in order to produce numerical or symbolic information, e.g., in the forms of decisions. The
research aims to build a working algorithm to track the target and determined it in a sequence of
frames corresponding to a video. The key requirements imposed on the tracker are being able to
track arbitrary targets, track accurately through challenging conditions and performing the
tracking in real-time. With these requirements in mind, we have researched what has been done
in the field and proposed an algorithm for video tracking. Video tracking in general is the
process of estimating, over time, the location of one or multiple objects in a video captured by
an analogue camera. Video tracking is often divided into target representation, target detection
and target tracking. First you create a model of the target. Then you implement techniques for
detecting the object using the target model. Lastly, you use different techniques to predict the
new target location given the current location. After comparing some trackers we chose to
implement the particle filter (PF). The PF is a robust tracker which can handle partial occlusion
and non-Gaussian distributions. Choosing the importance proposal distribution is a key issue for
solving nonlinear filtering problems.
In this work, the first contribution is a novel extension described to the PF tracking algorithm,
which investigates a hybrid approach derived from unscented Kalman filter (UKF) and iterated
extended Kalman filter (IEKF) called hybrid iterated Kalman particle filter (HIKPF) to generate
the proposal distribution. The proposed algorithm leads to an efficient use of the latest
observations and generates more close approximation of the posterior probability density.
The second contribution is a tracking algorithm based on particle filter and optimized
Likelihood. PF approximates a posterior probability density of the state by using samples which
are called particles. Here, the state is treated as the position of the object and the weight is
considered as the likelihood of each particle. For this likelihood, the similarity between the
color histogram of the tracked object and the region around the position of each particle was
calculated by using Bhattacharya distance. To enhance the results, a new parameter is multiplied
by the previous likelihood to increase the particles weight. The system proves to be robust
against problems of partial occlusion and illumination changes. The tracker was tested on a set
of videos with challenging scenarios. It is also of use to handle cases where the target may be
momentarily occluded due to other obstacles in the area. Although the tracker could handle
some scenarios if the parameters were tuned correctly, it did not perform satisfactorily when
dealing with difficult motion.
The third contribution is a new tracking algorithm called particle filter joint color-texture
histogram ”PFJCTH”. This algorithm is presented by using the joint color texture histogram to
represent a target and then applying it to particle filter algorithm. The texture features of the
object are extracted by using the local binary pattern (LBP) technique to represent the object.
The proposed algorithm extracts effectively the edge and corner features in the target region,
which characterizes better and represents more robustly the target. This tracker is capable of
tracking scale change and velocity change of the object
The main focus of this contribution is to find a technique that works well in these scenarios for a
single tracked target. For tracking performance verification, the implementations of these
algorithms are written in Matlab.
The performances of the proposed algorithms were measured by using the root mean
square error (RMSE), in case of HIKPF, PF with optimized likelihood and PFJCTH.