基于混合高斯模型的目标跟踪系统设计

Design of Target Tracking System Based on Gaussian Mixture Model

随着机器人在各行各业的广泛应用,使其与周围环境的交互变得越来越重要,而视觉可以使其具备非接触式的环境感知能力。针对这一难题提出了一种机器人的视觉系统设计方法,主要工作有:首先建立背景的混合高斯模型,并在追踪过程中实时更新以适应外界环境的变化。然后运用共面P4P重定位的方法,提高了目标定位的精度。其次引入卡尔曼滤波的结构相似度目标追踪算法,增强了系统的鲁棒性和实时性。最后通过实验的方法验证了该视觉系统的有效性。

With the extensive use of Delta robots, the interaction between the robot and the surrounding environment becomes more and more important. Taking it into account that the vision can make it have non-contact environment-aware ability, therefore, this paper presents the design of target tracking system. The main work is as follows Firstly, a Gaussian mixture model is set up and updated in real time to adapt to the changes of the external environment. And then use the method of coplanar P4P relocation to improve the accuracy of target location. Furthermore, Kalman filter is introduced to improve the robustness and real-time performance of the system. Finally, the validity of the designed vision system is verified by the experimental method.