【作者】 李丽娟；

【导师】 刘晓鸿；

【作者基本信息】 北京邮电大学 ， 计算机科学与技术， 2015， 硕士

【摘要】 随着微型飞行器的研究与应用越来越普遍,飞行器监控系统作为其与地面通信的终端系统也得到了越来越多研究人员的关注。移动通信和终端处理能力飞速发展的今天,移动终端监控系统很好地结合了通信领域和计算机工业。对于传统监控系统需要在指定地点及网线设备条件下才能对目标现场进行监控的情形,移动终端充分发挥了它的平台优势,监控连接只需其在无线网络覆盖的范围,扩展了应用范围和灵活性。论文设计实现了基于Android的微型飞行器的智能监控系统,主要包括以下几个方面：(1)分析监控系统的各项需求,设计了基于Android平台的微型飞行器的智能监控系统总体方案,并且,系统接收飞行器数据和无线摄像头视频数据,并完成摄像头数据在Android终端的实时显示。(2)实现OpenCV在Android平台的应用,并根据系统和模块的需求,设计实现应用程序的交互设计、数据存储查询方案、算法实现、功能性能评估等。(3)建立了基于多特征组合的霍夫森林分类模型,并应用到系统的目标跟踪模块,结合Camshift跟踪算法,根据定位结果和置信度定位目标中心,实现了Android平台上的目标跟踪。同时,设计了多类别目标检测模型,通过多个分类器检测相应类别目标,并且完成检测结果的统计与存储。(4)设计并实现了用户友好性的Android终端界面,通过系统各项指标的评估,完成系统的测试和功能、性能验证。更多还原

【Abstract】 With the increasingly research and application of micro air vehicles, the monitoring system is also becoming more and more important.The monitoring system is a part of communication devise between air vehicles and ground terminal systems.Nowadays,with the rapid development of mobile communications and terminal handling capacity,mobile operator system combines wireless communications and computer industries effectively. Traditional monitoring system need the designated location and the network cable to monitor our machine.Compared to the traditional monitoring system,monitor system in Android has many advantages,such as flexibility and operated easily etc.This paper can conclude as follows:Firstly,we analysis the requirements of the monitoring system, design the overall plan of intelligent monitoring system based on the Android platform.And designed to receive aircraft data and wireless camera video data,make the data displayed in real-time in the Android platform.Secondly,we apply OpenCV in Android platform,and according to the demands of the system and modules,design and implement the application,including interactive design,data storage solutions,algorithms code,functional and performance evaluation. Thirdly,we establish a multi-feature Hough forest classification model,which combined with the Camshift tracking algorithm,and then apply it to the object tracking module.According to parameters and confidences to locate the object center,the model achieve the module of tracking in the Android platform.At the same time,design of the multi-class model for object detection.Finally,design and implement a user-friendly interface of Android devices,and also complete functional and performance testing and evaluating the system of indicators.更多还原