基于DM6467平台的智能监控视频信号处理技术研究

发布时间：2018-03-17 16:28

本文选题：智能视频监控　切入点：达芬奇技术　出处：《南京理工大学》2017年硕士论文　论文类型：学位论文

【摘要】：智能视频监控作为安全防范系统的重要技术,在无需人为干预的情况下,能对监控摄像头采集到的视频自动分析并发出指令,因而在军事与民用领域具有广泛的应用。结合TMS320DM6467芯片在视频图像处理领域的性能优势,本文基于DM6467平台对智能监控视频信号处理技术进行研究。本文内容主要分为五个部分:第一部分介绍了智能视频监控的研究背景和意义及国内外研究现状;第二部分介绍了几种常见的嵌入式智能视频监控系统解决方案和系统硬件平台的构建,包括视频采集系统以及视频信号处理系统;第三部分首先介绍了帧差法、混合高斯模型算法、核密度估计算法、模糊积分特征算法等常用的运动目标检测算法的基本原理,其次针对经典W4算法不能更新背景模型的不足,提出了本文基于W4算法和帧差法的算法,有效地解决了经典W4算法因视频初始帧中存在运动目标而导致误检的问题。接着简要介绍了视频压缩和网络传输技术;第四部分将本文算法和其他常用的运动目标检测算法对不同的视频源进行matlab仿真实验,对检测结果进行定性分析和量性分析,并进行异常事件检测实验;第五部分介绍了智能监控视频信号处理技术的研究与设计,包括目标检测、异常事件检测、H.264视频压缩和网络传输。重点介绍了 DSP端图像处理算法的移植以及ARM端应用程序的设计。最后开展多组实验,实验结果表明,本文提出的基于W4算法和帧差法的目标检测算法在硬件平台上检测效果良好,其准确率高,完整度好,误检率低,且能很好地服务于异常事件检测;同时,视频编码的压缩比尚可,网络传输基本满足实时性。
[Abstract]:As an important technology of security and prevention system, intelligent video surveillance can automatically analyze and issue instructions to the video captured by the surveillance camera without human intervention. Therefore, it has been widely used in military and civilian fields. Combined with the performance advantage of TMS320DM6467 chip in video image processing, This paper is based on the DM6467 platform to study the intelligent video signal processing technology. The content of this paper is divided into five parts: the first part introduces the research background and significance of intelligent video surveillance and the research status at home and abroad; The second part introduces several common embedded intelligent video surveillance system solutions and the construction of the system hardware platform, including video capture system and video signal processing system. The basic principles of moving target detection algorithms such as mixed Gao Si model algorithm, kernel density estimation algorithm, fuzzy integral feature algorithm and so on. Secondly, the classical W4 algorithm can not update the background model. An algorithm based on W4 algorithm and frame difference method is proposed in this paper, which effectively solves the problem of false detection caused by the moving object in the original frame of the video. Then, the video compression and network transmission technology are briefly introduced. In the 4th part, the matlab simulation experiments of different video sources are carried out with this algorithm and other commonly used moving target detection algorithms, and the detection results are analyzed qualitatively and quantitatively, and the abnormal event detection experiments are carried out. Part 5th introduces the research and design of intelligent video signal processing technology, including target detection. H.264 video compression and network transmission. This paper mainly introduces the transplantation of image processing algorithm on DSP and the design of ARM application program. Finally, many experiments are carried out, and the experimental results show that, The target detection algorithm based on W4 algorithm and frame difference method proposed in this paper has good detection effect on hardware platform, its accuracy is high, integrity is good, false detection rate is low, and it can well serve for abnormal event detection. The compression ratio of video coding is good, and the network transmission basically meets the real-time requirement.
【学位授予单位】：南京理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.41;TN948.6

【参考文献】