机载目标视频跟踪与控制技术图像处理技术研究

菜单

摘要近几年，随着旋翼飞行器硬件水平的提高和目标视频跟踪技术的不断发展，机载目标视频跟踪系统逐渐受到了广泛关注。本文重点研究了复杂动态背景下的机载目标视频处理技术。首先介绍了基于静态背景下的运动目标跟踪算法，在基于MFC框架的Visual C++ 6.0环境中开发了静态背景下的目标视频跟踪软件，并对飞行中的旋翼无人机进行了跟踪实验。其次介绍了均值漂移(Mean-Shift)算法的理论及其在视频目标跟踪中的应用，然后基于机载目标视频跟踪的实际工程应用，提出一种基于目标模板更新和背景权重的改进Mean-Shift算法，接着根据所提的改进Mean-Shift算法在基于 OpenCV 的 Visual C++ 6.0 集成开发环境下编写了视频跟踪程序，并对机载摄像头拍摄的具有复杂动态背景的目标视频进行了跟踪实验。最后给出了机载目标视频跟踪系统的硬件组成及旋翼无人机与地面站通信的流程。两次跟踪实验结果及巴氏系数、迭代次数的对比结果表明，本文所提的改进Mean-Shift算法在跟踪精度和实时性上较传统Mean-Shift算法均有所提高，同时实验结果也验证所提的跟踪算法对机载目标视频跟踪的有效性。64505

毕业论文关键词无人机目标视频跟踪复杂动态背景机载摄像头跟踪实验

毕业设计说明书（论文）外文摘要

Title The tracking and controlling of airborne target video-tracking system——research of algorithm for target video-tracking

Abstract

In the past few years, more and more attention has been drawn to the airborne target video-tracking system with the better performance of hardware of the unmanned aerial vehicle(UAV) and the technological development of video tracking. The video processing technology of the airborne target tracking system based on the complicated and dynamic background was mainly discussed in this paper. Firstly, an algorithm for the video tracking system based on the static background was presented, and a video-tracking experiment using the tracking software under the MFC frame of visual C++ 6.0 was designed. Secondly, the Mean-Shift theory and its application in video tracking were introduced, and a Mean-Shift algorithm with model updating and background weight added was proposed for engineering application of the airborne target video-tracking system. Then based on the proposed modified algorithm, a tracking program under the OpenCV frame of visual C++ 6.0 was designed, and the tracking experiment with the airborne camera was conducted. Lastly, the pictures of the hardware of the airborne target video-tracking system and the proceeding of the data communication between UAV and the ground station were presented. The experimental results show that the tracking system based on the proposed Mean-Shift algorithm has higher tracking accuracy than that based on the traditional algorithm, and the proposed tracking algorithm has less iterations than the traditional one. Moreover, the experimental results show that the proposed Mean-Shift algorithm can be feasible for the airborne target video-tracking.

Keywords unmanned aerial vehicle target video-tracking complicated and dynamic background airborne camera tracking experiment

1 绪论 2

1.1 研究背景和意义 2

1.2 国内外研究现状 3

1.3 论文的工作安排 5

2 基于静态背景下的运动目标跟踪方法 7

2.1 彩色图像灰度化 7