Stable Target Tracking Research Articles

Real-Time Object Detection and Tracking for Unmanned Aerial Vehicles Based on Convolutional Neural Networks

This paper presents a system applied to unmanned aerial vehicles based on Robot Operating Systems (ROSs). The study addresses the challenges of efficient object detection and real-time target tracking for unmanned aerial vehicles. The system utilizes a pruned YOLOv4 architecture for fast object detection and the SiamMask model for continuous target tracking. A Proportional Integral Derivative (PID) module adjusts the flight attitude, enabling stable target tracking automatically in indoor and outdoor environments. The contributions of this work include exploring the feasibility of pruning existing models systematically to construct a real-time detection and tracking system for drone control with very limited computational resources. Experiments validate the system’s feasibility, demonstrating efficient object detection, accurate target tracking, and effective attitude control. This ROS-based system contributes to advancing UAV technology in real-world environments.

Predictive optimization algorithm for beam combination systems based on adaptive fiber optics collimators

Beam combining of fiber lasers based on adaptive fiber optics collimator (AFOC) plays an important role in the laser propagation applications, whereas the system performance is severely affected by the target occlusion during the beam’s long-range atmosphere transmission. In this paper, the sources, characteristics, and potential impacts of the target occlusion were analyzed for the first time, and the predictive stochastic parallel gradient descent (P-SPGD) algorithm was proposed to solve this problem. The basic principle of P-SPGD was to monitor the target occlusion in real-time, and then dynamically choose whether to use the observed gradient or the predicted gradient to update the control voltages of AFOCs. Beam combining experiments based on three AFOCs were conducted in the free space with a distance of 40 m. The results showed that when the moving target was obscured for 3 s, the AFOCs based on the traditional optimization algorithm had lost track of the target, whereas the P-SPGD algorithm can detect this emergency with a probability of 98.8%, and accurately model the motion state of the target. As a result, the AFOCs can still achieve stable target tracking during the period of target occlusion, and keep the combined beams from divergence.

Multi-UAV cooperative target tracking with bounded noise for connectivity preservation

We investigate cooperative target tracking of multiple unmanned aerial vehicles (UAVs) with a limited communication range. This is an integration of UAV motion control, target state estimation, and network topology control. We first present the communication topology and basic notations for network connectivity, and introduce the distributed Kalman consensus filter. Then, convergence and boundedness of the estimation errors using the filter are analyzed, and potential functions are proposed for communication link maintenance and collision avoidance. By taking stable target tracking into account, a distributed potential function based UAV motion controller is discussed. Since only the estimation of the target state rather than the state itself is available for UAV motion control and UAV motion can also affect the accuracy of state estimation, it is clear that the UAV motion control and target state estimation are coupled. Finally, the stability and convergence properties of the coupled system under bounded noise are analyzed in detail and demonstrated by simulations.

Research on motion planning for an indoor spray arm based on an improved potential field method.

The target spraying effect of spray robots mainly depends on the control performance of the spraying arm during the processes of aiming and tracking. To further improve the robustness of the endpoint control and positioning accuracy of the spray arm, an improved potential field algorithm for the motion planning and control of the spray arm is proposed based on prophase research. The algorithm introduces a velocity potential field, visual field constraints and joint position limit constraints into the traditional artificial potential field method. The velocity potential field is used to ensure that the target state of the spraying arm is at the same velocity as the target crop (relative velocity) to achieve stable target tracking. The visual field constraints and joint position limit constraints are utilized to ensure the efficiency of the visual servo control and the movement of the spray arm. The algorithm can plan a feasible trajectory for the spraying arm in Cartesian space and image space, and use the speed controller to control the spraying arm movement along the trajectory for aiming and tracking. Simulation analysis shows that the algorithm can plan better motion trajectories than the servo controller based on image moments in previous studies. In addition, the experimental results show that the algorithm can effectively improve the robustness of targeting and tracking control for the spray robot.

Correlation Filter With Motion Detection for Robust Tracking of Shape-Deformed Targets

Target tracking is an important area of research in computer vision where stable target's tracking has been well solved. But in real world, it is difficult to ensure that the camera or lens could be fixed and the target could maintain its shape in whole video sequence. And as a result, in these unstable cases, robust tracking algorithms have to deal with the problem of target shape-deforming. Once the scenes video sequence contains shape-deformed target, tracking become a real challenging problem. Most previous tracking algorithms based on craft features only used HOG or/and CN features. This paper proposed an algorithm named as Correlation Filtering with Motion Detection (CFMD). This algorithm takes into account the camera shake and target motion information of the video sequence. After removing the effects of lens shake and camera movement, this algorithm can predict the motion information of the target, thereby effectively improving the tracking accuracy and robustness. In CFMD, the target position is determined by the weighted outputs of motion detection and correlation filter tracker. We evaluated our CMFD algorithm on the OTB-100 and VOT-2018 dataset compared with other target tracking algorithms, including Kernel Correlation Filter (KCF), Scale Adaptive with Multiple Features tracker (SAMF), Discriminative Scale Space Tracker (DSST), and Sum of Template and Pixel-wise LEarners (Staple), Learning Spatial-Temporal Regularized Correlation Filters for Visual Tracking(STRCF), Multi-Cue Correlation Filters for Robust Visual Tracking(MCCT). The experimental results showed that our algorithm owns the property of robust tracking of shape-deformed targets in video sequences containing lens shaking or camera moving and it achieves the state-of-the-art precision and tracking effects.

Intelligent visual enhancement system

Vision-based augmented reality is a new kind of visual application technology, which transfers synthetic sensory information into a user’s perception of a real environment. It is perceived by human senses. However, the existing visual enhancement equipment has single function, limited processing capacity and poor interaction. To overcome these shortcomings, we designed a visual enhancement system that integrates cloud computing, AR technology and deep learning. First, small, remote and wireless cameras are used to obtain image data, which need to be uploaded to a cloud. Then, the method of deep learning and feature matching is adopted to carry out facial consistency analysis, which improves the robustness of target detection. And stable target tracking is achieved by time sequence state filtering. Finally, the information of image analysis and processing is transmitted back to the AR device, so the prompts of target text and voice are given for intelligent auxiliary decision-making in time.

Analysis of gyroscope system sliding mode control in a target seeking and tracking process

AbstractThe task of a homing head of an anti‐aircraft short‐range missile is to determine the position of the missile relative to a target. The most modern missiles use scanning heads. The propelling element of such a head in the paper is a gyroscope system. Disturbance acting on a gyroscope system cause its deviations from the set direction. In order to ensure precise and stable target tracking by a head, it is important to optimally select the gyroscope system control parameters, which enables mitigating the impact of such disturbance on the positioning accuracy of its axis. The paper suggests and implements a sliding mode controller in a gyroscope system, used for target seeking and tracking. Tests comparing a control system with a PD controller with optimal parameters and with a sliding mode controller were performed in a Matlab/Simulink environment. Selected test results were shown in graphic form.

Automatic tracking and positioning algorithm of motion object in complex environment

The problem of large error in automatic tracking and positioning of moving targets in complex environment is discussed. A method of target tracking and positioning based on probability analysis for moving images in complex environment is proposed. The likelihood and prior probability characteristics of moving target under complex background is calculated, and the optimal label of foreground jitter image contour is calculated using probability theory to identify the target region of moving image of the current frame in the complex state, and effectively complete the moving image tracking and marking in complex state. Experimental results show that the proposed method can accurately predict foreground objects in complex images, and improve the stability of target tracking.

Generalized Labeled Multi-Bernoulli Extended Target Tracking Based on Gaussian Process Regression

For the problems that Gamma Gaussian Inverse Wishart Cardinalized Probability Hypothesis Density (GGIW-CPHD) filter cannot accurately estimate the extended target shape and has a bad tracking performance under the condition of low SNR, a new generalized labeled multi-Bernoulli algorithm based on Gaussian process regression is proposed. The algorithm adopts the star convex to model the extended target, and realizes the online learning of the Gaussian process by constructing the state space model to complete the estimation of the extended target shape. At the same time, in the low SNR environment, the target motion state is tracked by the good tracking performance of the generalized label Bernoulli filter. Simulation results show that for any target with unknown shape, the proposed algorithm can well offer its extended shape and in the low SNR environment it can greatly improve the accuracy and stability of target tracking.

Research on semi-supervising learning algorithm for target model updating in target tracking

Target detection and tracking technique is one of the hot subjects in image processing and computer vision fields, which has significant research value not only in military areas such as imaging guidance and military target tracking, but also for civil use such as security and monitoring and the intelligent man-machine interaction. In this paper, for target deformation, scale changing, rotation, and other issues in the long-term stable target tracking, a bootstrapping feedback learning algorithm is proposed, which may improve the target model and the classifier discriminating capacity as well as the fault tolerance ability; and it also makes fewer errors during the updating, and then the proof of convergence of the algorithm is given. Experimental results show that among the same tracking algorithms, utilization of the learning method to update the target model and classifier is more stable and more adaptable than unusing it in the processes of target scale changing, deformation, rotation, perspective changing and fuzzy. And compared with the existing conventional method, this method has a better robustness, and a high value in practical application and research.

WE-G-BRF-09: Force- and Image-Adaptive Strategies for Robotised Placement of 4D Ultrasound Probes

Purpose: To allow continuous acquisition of high quality 4D ultrasound images for non-invasive live tracking of tumours for IGRT, image- and force-adaptive strategies for robotised placement of 4D ultrasound probes are developed and evaluated. Methods: The developed robotised ultrasound system is based on a 6-axes industrial robot (adept Viper s850) carrying a 4D ultrasound transducer with a mounted force-torque sensor. The force-adaptive placement strategies include probe position control using artificial potential fields and contact pressure regulation by a PD controller strategy. The basis for live target tracking is a continuous minimum contact pressure to ensure good image quality and high patient comfort. This contact pressure can be significantly disturbed by respiratory movements and has to be compensated. All measurements were performed on human subjects under realistic conditions. When performing cardiac ultrasound, rib- and lung shadows are a common source of interference and can disrupt the tracking. To ensure continuous tracking, these artefacts had to be detected to automatically realign the probe. The detection is realised by multiple algorithms based on entropy calculations as well as a determination of the image quality. Results: Through active contact pressure regulation it was possible to reduce the variance of the contact pressure by 89.79% despite respiratory motion of the chest. The results regarding the image processing clearly demonstrate the feasibility to detect image artefacts like rib shadows in real-time. Conclusion: In all cases, it was possible to stabilise the image quality by active contact pressure control and automatically detected image artefacts. This fact enables the possibility to compensate for such interferences by realigning the probe and thus continuously optimising the ultrasound images. This is a huge step towards fully automated transducer positioning and opens the possibility for stable target tracking in ultrasoundguided radiation therapy requiring contact pressure of 5–10 N. This work was supported by the Graduate School for Computing in Medicine and Life Sciences funded by Germany's Excellence Initiative [DFG GSC 235/1].

An Anti- Occlusion Tracking Algorithm Based on MCD Correlation Matching

In order to improve the stability of target tracking under occlusion conditions，on the basis of researching some target tracking algorithms, this paper presents an algorithm based on MCD correlation matching, which combines multi sub-templates matching and target movement prediction. Besides, for occlusion characteristics, corresponding template matching criterions and updating methods are put forward. Experimental results show that, comparing with the single template method which updating frame by frame, the proposed algorithm has a certain anti-occlusion ability with better stability and continuity of target tracking under occlusion conditions.

A Fuzzy Shape Control Method for Stainless Steel Strip on Sendzimir Rolling Mill

Abstract The strip shape in the stainless steel process has made an issue of the strip quality. The objective of the shape control is to minimize the shape deviation and to maintain symmetrical shape patterns in the lateral direction. The method of the shape recognition employs the least square method. The shape deviation is controlled by the fuzzy shape controller (FSC). The experiments have been performed online for various stainless materials, thicknesses, and strip widths. The test results show very efficient performances in respect of stable target tracking and symmetrical and minimal fluctuation of the strip shape.

Predictive proportional nonlinear control for stable-target tracking of a mobile robot: an experimental study

Two new types of control method have been developed based on model predictive control for stable-target tracking of a nonholonomic mobile robot. One method (Method 1) is a new nonlinear control method. This was developed based on model predictive control (predictive nonlinear control) to predict the next position of a mobile robot using the current velocities of the right and left wheels. This technique uses a tuning guideline in predictive nonlinear control. The other method (Method 2) is a combination of Method 1 and proportional control (predictive proportional nonlinear control). Method 2 involves a tuning guideline not only in a predictive nonlinear controller, but also in a proportional controller. In this technique, the selection of a tuning guideline in the proportional controller is enhanced, and thereby increases the control action in closed-loop responses. In Method 1, the nonlinear controller is derived from Liapunov stability theory, and is used to control the linear and angular velocities for locomotion control. Tuning parameters in the nonlinear controller (in Method 1) are selected to satisfy various design criteria, such as stability, performance, and robustness. Method 1 has certain limitations that result in a decrease of the performance criteria specified. Strong nonlinearities in the mobile robot system result in accumulated errors. To enhance performance further, we developed Method 2 as the solution for decreasing cumulative errors. Hence, the proportional controller is added to Method 1 in the closed-loop form in order to eliminate errors. The advantage of Method 2 is that it can cope with strong nonlinearities in the mobile vehicle system. The results of the performances of Method 1 and Method 2 are shown to demonstrate the effectiveness of both methods, and also the better performance of Method 2. The two new methods are effective in stable-target tracking, yielding an increase in performance and stability.