基于DSP的目标检测与跟踪系统设计

发布时间：2018-04-20 12:40

本文选题：智能监控 + DSP　；参考：《湖南大学》2012年硕士论文

【摘要】：视频监控作为安全防范的一种重要的手段越来越受到人们的关注，特别是在机场、银行、地铁等敏感区域的实时监控。随着人们对视频监控系统实时性、准确性、自动化程度等要求的不断提高，基于计算机视觉和视频分析的智能视频监控系统已经成为视频监控领域的重要研究内容，并代表着视频监控未来的发展方向。智能视频监控系统利用计算机视觉和视频分析等技术，在不需要人为参与的情况下，对摄像机采集到的视频图像进行自动分析、处理，得到视频图像中目标的特征信息，并根据目标的运动信息做出反应。为满足智能视频监控系统对实时性、跟踪准确性、可靠性的性能要求，本文设计并实现了一套基于通用DSP芯片的智能监控系统，实现了监控场景中目标的自动检测跟踪，并在显示界面进行标注显示。论文的主要工作包括：首先，设计实现了基于DSP的智能监控系统的硬件平台，并根据智能监控系统的处理流程对硬件平台的各个功能模块进行设计实现。与现有视频监控系统硬件平台相比，基于通用DSP芯片的硬件实现方案具有通用性强、具有良好可扩展性等优点；然后，在视频监控平台的基础上完成了智能监控系统的软件设计，包括视频图像采集、图像的处理及上位机的目标跟踪显示等模块。通过分析背景减除法、帧间差分法及光流法三种目标检测跟踪算法的性能，选择背景减除法作为本系统中运动目标检测跟踪的算法，并通过实验验证了，在实时性及目标检测的准确度方面具有明显的优越性。最后，结合对当前系统硬件平台进行了调试，并验证了系统的的检测跟踪效果。根据本文系统的工作和不足做了系统总结，，并针对未来的研究方向做了展望。
[Abstract]:As an important means of security, video surveillance has attracted more and more attention, especially in sensitive areas such as airport, bank, subway and so on. With the improvement of real-time, accuracy and automation of video surveillance system, intelligent video surveillance system based on computer vision and video analysis has become an important research content in the field of video surveillance. And represents the future direction of video surveillance. The intelligent video surveillance system makes use of computer vision and video analysis techniques to automatically analyze and process the video images collected by the camera without the participation of human beings, so as to obtain the characteristic information of the target in the video images. And according to the moving information of the target to make a response. In order to meet the requirements of real-time, tracking accuracy and reliability of intelligent video surveillance system, this paper designs and implements an intelligent monitoring system based on universal DSP chip, which realizes the automatic detection and tracking of targets in monitoring scene. And in the display interface label display. The main work of the thesis includes: Firstly, the hardware platform of the intelligent monitoring system based on DSP is designed and implemented, and the function modules of the hardware platform are designed and implemented according to the processing flow of the intelligent monitoring system. Compared with the existing hardware platform of video surveillance system, the hardware implementation scheme based on general DSP chip has the advantages of good generality and good expansibility. Then, the software design of the intelligent monitoring system is completed on the basis of the video monitoring platform, including video image acquisition, image processing and target tracking and display module of the upper computer. By analyzing the performance of three target detection and tracking algorithms, background subtraction, inter-frame differential and optical flow, background subtraction is selected as the moving target detection and tracking algorithm in this system. It has obvious superiority in real time and accuracy of target detection. Finally, the hardware platform of the current system is debugged, and the detection and tracking effect of the system is verified. According to the work and deficiency of this paper, the future research direction is prospected.
【学位授予单位】：湖南大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TP391.41;TP368.1

【参考文献】