Robust Tracking Method Research Articles

Robust Face Tracking Method Based on Kalman Particle Filter and CamShift

As the tracking mode based on color probability distribution in Continuously Adaptive Mean-SHIFT can not describe object’s nonrigid deformation precisely, an efficient robust face tracking method is proposed, which integrates object’s moving state estimation into Kalman particle filter, and combines Kalman particle filter with CamShift tracking algorithm. Experimental results show that the method presented has good robustness to size and angle variation, rapid movement, partial and fully occlusion of face.

Hybrid tracking system for robust fiducials registration in augmented reality

An effective augmented reality system requires an accurate registration of virtual graphics on real images. In this work, we developed a multi-modal tracking architecture for object identification and occlusion handling. Our approach combines several sensors and techniques to overcome the environment changes. This architecture is composed of a first coded targets registration module based on a hybrid algorithm of pose estimation. To manage partial target occlusions, a second module based on a robust method for feature points tracking is developed. The latest component of the system is the hybrid tracking module. This multi-sensors part handles total target occlusions issue. Experiments with the multi-modal system proved the effectiveness of the proposed tracking approach and occlusion handling in augmented reality applications.

A robust hybrid method for maximum power point tracking in photovoltaic systems

In this paper a new hybrid method for maximum power point tracking in PV systems has been proposed. This method combines offline and online methods in order to estimate duty cycle of converter in maximum power point. In the offline phase, temperature and radiation intensity are the inputs of the system to estimate the approximate maximum power based on analytical equations of solar cell. These equations which give the relation of maximum power with temperature and irradiation can be derived from characteristics of cell provided by manufacturer or experiments. Afterwards the duty cycle of converter would be estimated using circuit equations of measured Thevenin model of the load and battery. Measuring Thevenin equation results in robustness of method respecting variations of load and battery. In the online phase, the classic perturbation and observation (P&O) method will be utilized for fine tuning and tracking of maximum power point. The proposed method has been simulated in MATLAB/SIMULINK workspace and compared with some other MPPT methods. The results reveal that this hybrid method outperforms other methods in term of performance and speed of tracking.

Object Tracking Approach based on Mean Shift Algorithm

Object tracking has always been a hotspot in the field of computer vision, which has a range of applications in real world. The object tracking is a critical task in many vision applications. The main steps in video analysis are: detection of interesting moving objects and tracking of such objects from frame to frame. Most of tracking algorithms use pre-defined methods to process. In this paper, we introduce the Mean shift tracking algorithm, which is a kind of important no parameters estimation method, then we evaluate the tracking performance of Mean shift algorithm on different video sequences. Experimental results show that the Mean shift tracker is effective and robust tracking method.

Robust Tracking Control of a Quadrotor Helicopter

In this paper, a robust tracking control method for automatic take-off, trajectory tracking, and landing of a quadrotor helicopter is presented. The designed controller includes two parts: a position controller and an attitude controller. The position controller is designed by the static feedback control method to track the desired trajectory of the altitude and produce the desired angles for pitch and roll angles. By combining the proportional-derivative (PD) control method and the robust compensating technique, the attitude controller is designed to track the desired pitch and roll angles and stabilize the yaw angle. It is proven that the attitude tracking error of each channel can converge to the given neighborhood of the origin ultimately. Experimental results demonstrate the effectiveness of the designed control method.

A Robust Line Tracking Method based on a Multiple Model Kalman Filter Model for Mobile Projector Systems

Abstract Line tracking is essential for many real world applications of computer vision. In thispaper, we propose a computationally efficient line tracker that can robustly andaccurately track lines in an image. We use amultiple-model-Kalman filter (MMKF) scheme which can handle line tracking accurately and robustly. The basic idea is to run N multiple sub-Kalman filters in parallel. Each filter is configuredto use a different state transition model. All the filters are updated by the measurement atthe same time following the conventional Kalman filter update process. The finalprediction is a combination of outputs from all the Kalman filter modules. Theexperimental result shows that the proposed system can track a line robustly in real time. The result is useful in shape detection and should be suitable for building mobile projector applications.

Fusing Visual Tracking and Navigation for Autonomous Control of An Intelligent Wheelchair

Abstract This paper presents a hybrid navigation and control method for intelligent wheelchair. A robust and semi-supervised vision tracking method using Microsoft Kinect and obstacle avoidance navigation approach are integrated into a two-level control system. Then, position tracking control considering nonholonomic constraints are adopted for the motion control of wheelchairs. An indoor environment is set up for evaluating the application of the proposed approaches. The experimental results show that the proposed method is effective and can be used as a hands-free controller for the intelligent wheelchair.

Tracking of Ocean-Floor Organisms Using a Particle Filter

This paper describes the development of a visual tracking method of an underwater vision system installed into a UVMS (underwater vehicle-manipulator system) to manipulate ocean-floor organisms such as acanthaster. First, an execution procedure of UVMS for approaching the target is shown. Next, a novel and robust target tracking method by using particle filter framework in order to achieve the procedure is proposed. We applied the proposed method to real video images, and examined its effectiveness by comparing it with other methods. The results of some experiments that evaluate the performance of the developed method are given.

713 虹彩欠損補正楕円推定トラッキング法による広範囲対応視線入力システム(OS7 オーガナイズドセッション《精密/微細加工と評価》)

Eye Tracking Technique becomes more feasible with late fast visual image processing technique and various application are planned with linked with geographic, surrounding state, object recognition and handling. Eye tracking performance is based on iris edge extraction , and this is banefully affected by the reflected light and occlusion by eyelid and eyelashes in conventional outline estimation method. This paper proposed the robust eye tracking method (ACEF) with ellipse fitting considered deformation in gazing direction. In addition, 2 type of mouse control methods are compared with relative drive method and direct drive methods.

Pointing Angle Estimation for Human-Robot Interface

In this paper, we propose an approach to track and estimate user pointing direction in 3D Space. In the area of human-robot interaction, user communicates with service robot in their daily life activities to give commands and execute the given task accordingly. Therefore, the ability of user to gives command to service robot naturally can provide an interactive user interface system for real 3D space environment. For this purpose, we aim to perform pointing gesture tracking and after that estimate the user's pointing direction. Our method of pointing direction estimation is based on 3D orientation of hand and shoulder center of user. We make comparison with our previous method to find the best hypothesis. Experimental results show the angular error for the estimation of pointing direction is successfully improved from our previous method. As a result, our natural user interface system can manipulate 3D objects in living room environment thus providing intuitive robotic service for human robot interaction. I. INTRODUCTION Tracking of human gesture has been widely researched either in human-computer interaction or human-robot interaction. Besides voice, gesture is an example of platform to interfacing between human and robot intuitively and naturally. Pointing is one part of gesture which provides information and intention from human to service robot. Pointing-based user interface can easily indicate object and location for the robot delivering robot task in home environment. However, in the real environment, some ambiguity can occur and affect the performance of robotics service. The ambiguity in pointing gesture comes from the angular error in estimation of user's pointing direction particularly when certain objects is closely located. Due to the ambiguous pointing command, correct object and location cannot be identified thus service robot surely unable to deliver task successfully. To overcome this problem, we conduct several experiments on tracking of pointing arm and make comparison study to find the best hypothesis. In this paper, we present a robust tracking method to be used in pointing direction estimation. We measure the pointing angle from the position of hand with respect to the shoulder center. To obtain the 3D coordinates of pointing arm

PSO Algorithm Particle Filters for Improving the Performance of Lane Detection and Tracking Systems in Difficult Roads

In this paper we propose a robust lane detection and tracking method by combining particle filters with the particle swarm optimization method. This method mainly uses the particle filters to detect and track the local optimum of the lane model in the input image and then seeks the global optimal solution of the lane model by a particle swarm optimization method. The particle filter can effectively complete lane detection and tracking in complicated or variable lane environments. However, the result obtained is usually a local optimal system status rather than the global optimal system status. Thus, the particle swarm optimization method is used to further refine the global optimal system status in all system statuses. Since the particle swarm optimization method is a global optimization algorithm based on iterative computing, it can find the global optimal lane model by simulating the food finding way of fish school or insects under the mutual cooperation of all particles. In verification testing, the test environments included highways and ordinary roads as well as straight and curved lanes, uphill and downhill lanes, lane changes, etc. Our proposed method can complete the lane detection and tracking more accurately and effectively then existing options.

Catching a Rat by Its Edglets

Computer vision is a noninvasive method for monitoring laboratory animals. In this article, we propose a robust tracking method that is capable of extracting a rodent from a frame under uncontrolled normal laboratory conditions. The method consists of two steps. First, a sliding window combines three features to coarsely track the animal. Then, it uses the edglets of the rodent to adjust the tracked region to the animal's boundary. The method achieves an average tracking error that is smaller than a representative state-of-the-art method.

TH‐C‐BRA‐09: Robust 3D Lung Tumor Tracking Using Maximum Likelihood Estimation

Purpose: To develop a robust 3D lung tumor tracking method based on a single x‐ray projection using maximum likelihood estimation. Methods: 4DCT images at different phases are registered to yield motion vector fields and a PCA based lung motion model is built. An acquired CBCT projection is compared with the simulated projections under different metrics to yield the correct PCA coefficients and hence the corresponding motion vector fields, which can be used to derive the 3D tumor location at this projection. The metrics we considered include normalized cross correlation and normalized mutual information between the gradient images of the acquired CBCT and that of the simulated 4DCT projection, and the position of diaphragm. A maximum likelihood estimation method is developed to combine the tracking results from all the metrics and integrate the constraints regarding tumor motion smoothness and wave form, leading to robust tumor locations. The method is tested on real patient data. Results: It is found that normalized cross correlation, normalized mutual information, and diaphragm location can be used to determine the PCA coefficients and the tumor position, although the efficacy of any single metric depends on the gantry angles and breathing phases. However, with the combination of different metrics, the tumor tracking accuracy is greatly improved. The resulting mean 3D error is found to be 1.3mm. Conclusions: A robust 3D tumor tracking method is developed based on maximum likelihood estimation to derive tumor location corresponding to an x‐ray projection for lung cancer patient. This work is supported by Varian Medical Systems through a Master Research Agreement.

Interface Exchange as an Indicator for Eddy Heat Transport

AbstractThe ocean contains many large‐scale, long‐lived vortices, called mesoscale eddies, that are believed to have a role in the transport and redistribution of salt, heat, and nutrients throughout the ocean. Determining this role, however, has proven to be a challenge, since the mechanics of eddies are only partly understood; a standard definition for these ocean eddies does not exist and, therefore, scientifically meaningful, robust methods for eddy extraction, characterization, tracking and visualization remain a challenge. To shed light on the nature and potential roles of eddies, we extend our previous work on eddy identification and tracking to construct a new metric to characterize the transfer of water into and out of eddies across their boundary, and produce several visualizations of this new metric to provide clues about the role eddies play in the global ocean.

Robust decentralized multi-model adaptive template tracking

In this paper, a robust and efficient visual tracking method through the fusion of several distributed adaptive templates is proposed. It is assumed that the target object is initially localized either manually or by an object detector at the first frame. The object region is then partitioned into several non-overlapping subregions. The new location of each subregion is found by an EM11Expectation Maximization, see [45] for more information.-like gradient-based optimization algorithm. The proposed localization algorithm is capable of simultaneously optimizing several possible solutions in a probabilistic framework. Each possible solution is an initializing point for the optimization algorithm which improves the accuracy and reliability of the proposed gradient-based localization method to the local extrema. Moreover, each subregion is defined by two adaptive templates named immediate and delayed templates to solve the “drift” problem.22The problem of gradually updating the object appearance model with irrelevant information such as background pixel values [43]. The immediate template is updated by short-term appearance changes whereas the delayed template models the long-term appearance variations. Therefore, the combination of short-term and long-term appearance modeling can solve the template tracking drift problem. At each tracking step, the new location of an object is estimated by fusing the tracking result of each subregion. This fusion method is based on the local and global properties of the object motion to increase the robustness of the proposed tracking method against outliers, shape variations, and scale changes. The accuracy and robustness of the proposed tracking method is verified by several experimental results. The results also show the superior efficiency of the proposed method by comparing it to several state-of-the-art trackers as well as the manually labeled “ground truth” data.

1A1-G11 一つのLRFを用いた人追跡に関する一考察(車輪型/クローラ型移動ロボット(1))

The objective of this study is to develop a robust human tracking method for mobile robot using only one laser range finder (LRF). When a mobile robot runs along a human, it is always necessary to observe positions of the human. Therefore, the LRF must be mounted on a position easy to recognize human. Furthermore, the LRF may lose the position of the target human, and another human might enter between the robot and the target human. Therefore, this paper discusses the mounting position of the LRF, as well as a robust tracking method using the LRF.

Robust Tracking Using Particle Filter with a Hybrid Feature

This paper presents a novel method for robust object tracking in video sequences using a hybrid feature-based observation model in a particle filtering framework. An ideal observation model should have both high ability to accurately distinguish objects from the background and high reliability to identify the detected objects. Traditional features are better at solving the former problem but weak in solving the latter one. To overcome that, we adopt a robust and dynamic feature called Grayscale Arranging Pairs (GAP), which has high discriminative ability even under conditions of severe illumination variation and dynamic background elements. Together with the GAP feature, we also adopt the color histogram feature in order to take advantage of traditional features in resolving the first problem. At the same time, an efficient and simple integration method is used to combine the GAP feature with color information. Comparative experiments demonstrate that object tracking with our integrated features performs well even when objects go across complex backgrounds.

Robust Tracking Method by Mean-Shift Using Spatiograms and Particle-Filter

In this paper, we proposed the real time robust tracking method which combined mean-shift using the spatiograms with the particle-filter, and the proposed method can continue to track non-rigid objects with a temporal occlusion. The proposed method is using a tracker representing the center of the object and samples. At first the tracker tracks the object by using the Mean-Shift algorithm based on the spatiograms. Then, to avoid to be trapped by occlusion, the true target position is searched by using the particle-filter. The proposed method needs only small number of samples so that the computational cost can be reduced largely. We conducted experiments in order to compare the proposed method with the conventional Mean-Shift. As a result, the proposed method was more effective than the conventional method, when the object moved rotation or moved rapidly and a temporal occlusion occurred.

Robust Tracking Method to Moving Object with Random Walk

Robust Tracking Method to Moving Object with Random Walk

Robust consensus tracking of a class of second-order multi-agent dynamic systems

In this paper, we study the problem of robust consensus tracking for a class of second-order multi-agent dynamic systems with disturbances and unmodeled agent dynamics. Contrary to previous approaches, we design continuous distributed consensus protocols to enable global asymptotic consensus tracking. Our focus is on consensus protocol design and stability analysis which also leads to the derivation of sufficient conditions for consensus tracking. We first consider the case of undirected information exchange with a symmetric and positive definite information-exchange matrix. We develop an identifier for each agent to estimate the unknown disturbances and unmodeled agent dynamics. Based on the identifier, we develop a consensus tracking protocol to enable global asymptotic consensus tracking using local information obtained from neighboring agents. The closed-loop stability is proven using Lyapunov analysis theory and an invariance-like theorem. We then extend the approach to the case of directed information exchange, whose information-exchange matrix is only of full rank so that the approach for undirected graphs cannot be directly applied. We show that global asymptotic consensus tracking can still be enabled under the new derived sufficient conditions by designing a new identifier, which utilizes the estimated information exchanged from neighboring agents, and constructing a new Lyapunov function. Examples and numerical simulations are provided to validate the effectiveness of the proposed robust consensus tracking method.