基于Android无线传感器网络目标跟踪识别系统的设计

发布时间：2018-04-23 01:00

本文选题：无线传感器网络 + Android　；参考：《安徽理工大学》2014年硕士论文

【摘要】：无线传感器网络是目前最活跃的科技领域之一,具有由体积小,数据采集、数据传输、数据处理功能的的传感器节点组成,在日常生活,工业生产,学校科研方面都有广泛的应用前景。随着计算机视觉和Android系统的应用发展,加上芯片设计、制造工艺的进步,将Android系统移植到嵌入式设备里已成为趋势。Android系统具有强大的人机交互能力,优良的显示效果和炫丽的图形界面,但自身图像处理、目标跟踪方面的能力相对有限。计算机OpenCV视觉库含有很多通用的算法能高效地完成这一任务,而且也得到Android系统的支持,故在Android系统上使用OpenCV成为可能。运动目标检测是目标跟踪的前提条件,目前,在图像处理及计算机视觉研究领域已有许多目标跟踪算法,如经典的Meanshift,质心算法,基于模板匹配的跟踪算法等,这些算法对目标识别有可靠性但是目标跟踪方面的效果差强人意。TLD目标跟踪算法有三部分组成,即跟踪器、学习过程和检测器,采用跟踪和检测相结合的策略,是一种自适应的、可靠的跟踪技术。本文主要设计了一种作为无线传感器网络节点的Android嵌入式设备,实现目标跟踪识别的功能。主控芯片是三星公司的S3C6410,在其上搭建了Android OpenCV开发包,使用TLD目标跟踪算法,使节点的数据采集的有效性,可交互性都得到了提高。节点之间的数据传输是采用Wi-Fi互联技术,节省了布线费用,同时短距离传输可靠性也能保证。最后试验了设备对目标识别跟踪的效果,证明了本系统识别的有效性。
[Abstract]:Wireless sensor network is one of the most active fields of science and technology at present. It is composed of sensor nodes with small size, data acquisition, data transmission and data processing functions. The school scientific research has the extensive application prospect. With the development of computer vision and Android system, as well as the development of chip design and manufacturing technology, transplanting Android system to embedded devices has become a trend. Excellent display effect and dazzling graphical interface, but its own image processing, target tracking ability is relatively limited. Computer OpenCV vision library contains many common algorithms to accomplish this task efficiently, and it is also supported by Android system, so it is possible to use OpenCV on Android system. Moving target detection is the precondition of target tracking. At present, there are many target tracking algorithms in the field of image processing and computer vision research, such as classical Meanshift, centroid algorithm, template matching based tracking algorithm and so on. These algorithms have reliability for target recognition but the effect of target tracking is unsatisfactory. TLD target tracking algorithm consists of three parts: tracker, learning process and detector, and adopts the strategy of combining tracking and detection. Is an adaptive, reliable tracking technology. In this paper, a Android embedded device is designed as the node of wireless sensor network to realize the function of target tracking and recognition. The main control chip is Samsung S3C6410. the Android OpenCV development kit is built on the chip, and the TLD target tracking algorithm is used to improve the validity and interactivity of the node data acquisition. The data transmission between nodes is based on Wi-Fi technology, which saves the wiring cost and ensures the reliability of short distance transmission. Finally, the effect of the equipment on target recognition and tracking is tested, and the effectiveness of the system is proved.
【学位授予单位】：安徽理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP212.9;TN929.5

【参考文献】