基于无线移动视频图像传输的远程监控与智能识别系统

VIP免费
3.0 牛悦 2024-11-19 4 4 2.95MB 69 页 15积分
侵权投诉
摘 要
移动视频智能监控是基于计算机视觉和模式识别领域，并结合现有的 3G 无
线通信技术和嵌入式技术的应用，是视频监控进一步发展的方向。通过对视频图
像高清采集、传输，进而监控和识别不同种类的物体，能及时发现和处理远程监
控现场中的异常情况。本文的研究内容主要分为以下两个部分：无线网络中视频
图像的传输策略和对现场目标物体的检测、识别和跟踪。进而能对异常现象进行
定义和报警。智能视频监控具有广阔的市场前景，可应用于交通管理、安全防护、
实时监控等领域。
无线移动视频智能监控系统中涉及无线移动视频传输和计算机视觉中数字
图像处理技术。本文将就两部分中影响系统鲁棒性的关键算法进行研究，包括对
无线移动视频的鲁棒传输和鲁棒的智能视频监控中复杂天气状况下（夜晚、雨雾
天）对目标进行检测、识别和跟踪，并能在智能监控中实现对目标物体的阴影和
遮挡进行有效处理。然后在搭建 Ti 达芬奇嵌入式硬件平台对算法进行实现和仿真，
并对算法的有效性进行评估。
算法实现的平台选用是 Ti 达芬奇，它是一组专为数字视频技术开发设计的
嵌入式硬件平台，其中 SoC 处理器 DM6446 由开放式的双核构造，同时具有 ARM 和
DSP 模块的数字多媒体处理器。首先将各类算法封装进 DSP 端，同时生成底层的算
法库。并对相应算法进行优化来更好的适应实时的视频处理要求。其次对 ARM 端
的进行 Linux 应用程序的软件开发。最后在这二者的基础上搭建集合视频图像采
集处理、无线移动传输、视频编解码功能的无线移动视频智能监控系统。对本文
提出算法进行了验证和评估。实验结果表明，基于 Ti 达芬奇嵌入式硬件平台实现
的算法具有较好的智能识别性能和在视频处理方面具有较高的实时性和稳定性。
关键词: 智能视频监控 H.264 算法 Ti 达芬奇 嵌入式 Linux
无线视频传输 背景建模 鲁棒性
ABSTRACT
Smart mobile video surveillance is based on computer vision and pattern recognition,
combined with the existing 3G wireless communication technology and embedded technology, is
the direction of further development of video surveillance. Through high-definition video image
capture, transmission, and then monitor and identify different types of objects, and can find and
deal with remote monitoring site in the anomalies. This paper will be divided into two parts: the
wireless network video transmission strategy and target objects on-site detection, identification
and tracking. Anomalies can then be defined and alarm. Intelligent video surveillance has a broad
market prospect, can be applied to traffic management, security, real-time monitoring and other
fields.
Wireless Mobile Video Intelligent Surveillance System involving wireless mobile video
transmission and digital image processing computer vision technology. This will influence the two
key algorithm robustness study, including the robustness of the wireless mobile video transmission
and robust intelligent video surveillance in complex weather conditions on target detection,
identification and tracking, and intelligent control in achieving the target object's shadow and
occlusion of effective treatment. And then set up Ti DaVinci Embedded Hardware implementation
of the algorithm and simulation, and evaluate the effectiveness of the algorithm.
Algorithm selection is Ti Da Vinci platform, which is a group designed for digital video
technology development and design of embedded hardware platforms, including SoC DM6446
processor's dual-core by the open structure of both ARM and DSP module of the digital multimedia
device. First of all, the various algorithms into the DSP side package, while generating the
underlying algorithms library. And the corresponding optimization algorithm to better adapt to
the real-time video processing requirements. Second, the conduct of Linux on the ARM side
application software development. Finally, the two structures based on collection of the video
image acquisition and processing, wireless transmission, video codec function of intelligent wireless
mobile video monitoring system. Of the proposed algorithm is verified and evaluated. Experimental
results show that Ti-based embedded hardware platform DaVinci algorithm has better recognition
performance and intelligent video processing in real time and high stability.
Keywords: Intelligent Video Surveillance ; H.264 Algorithm;
TI DaVinci;Wireless video transmission;
Embedded Linux System;
Backgroulld Modeling ; Robustness;
目 录
摘 要 .................................................................................................................................5
ABSTRACT ......................................................................................................................6
第一章 绪论.....................................................................................................................1
§1.1 引言 ..................................................................................................................................... 1
§1.2 无线移动视频智能监控系统介绍 .....................................................................................2
§1.2.1 嵌入式系统 ................................................................................................................ 2
§1.2.2 无线移动视频传输 .................................................................................................... 3
§1.2.3 智能视频监控概述 .................................................................................................... 4
§1.3 论文组成 ............................................................................................................................. 5
第二章 系统硬件平台选取.............................................................................................7
§2.1 DM6446 芯片介绍 ............................................................................................................ 7
§2.1.1 ARM 模块介绍 ......................................................................................................... 8
§2.1.2 DSP 模块介绍 ........................................................................................................... 9
§2.1.3 视频处理子系统 ...................................................................................................... 10
§2.2 DVEVM 硬件开发板介绍 .............................................................................................. 11
§2.3 本章小结 ........................................................................................................................... 12
第三章 无线视频智能监控系统关键算法研究...........................................................13
§3.1 有损信道传输中视频编码算法研究 ...............................................................................13
§3.1.1 视频编码算法选取 .................................................................................................. 13
§3.1.2 H.264 算法原理简介 .............................................................................................. 15
§3.1.3 FMO 图像片组划分 .................................................................................................16
§3.1.4 不对等保护及 FEC ..................................................................................................17
§3.1.5 基于重要性的 FMO 信道不对等保护方案 ........................................................... 20
§3.2 鲁棒的智能视频监控算法研究 .......................................................................................21
§3.2.1 基于 UKF 的背景建模方法 .................................................................................... 21
§3.2.2 基于物理模型的去雨雾算法 .................................................................................. 22
§3.2.3 基于 Retinex 理论和小波变换的夜间图像增强算法 ........................................... 23
§3.2.4 基于图像边缘和角点特征的去阴影算法 ..............................................................24
§3.2.5 基于目标轮廓的预测跟踪算法 .............................................................................. 26
§3.3 本章小结 ........................................................................................................................... 27
第四章 基于 Ti 达芬奇平台的系统总体设计 ............................................................29
§4.1 系统结构设计 ................................................................................................................... 29
§4.1.1 硬件结构设计 .......................................................................................................... 29
§4.2.1 软件结构设计 .......................................................................................................... 30
§4.2 前端处理线程实现 ...........................................................................................................33
§4.2.1 视频采集 .................................................................................................................. 33
§4.2.2 视频预处理 .............................................................................................................. 34
§4.3 视频处理线程实现 ...........................................................................................................36
§4.3.1 API 调用流程 .......................................................................................................... 36
§4.3.2 算法实现及封装 ...................................................................................................... 37
§4.3.3 应用层实现 .............................................................................................................. 37
§4.4 TS 打包发送线程实现 .................................................................................................... 38
§4.4.1 TS 流介绍 ................................................................................................................ 38
§4.4.2 TS 打包发送实现 .................................................................................................... 39
§4.5 本章小结 ........................................................................................................................... 39
第五章 基于 Ti 达芬奇平台的算法实现与仿真 ........................................................41
§5.1 系统框架设计 ................................................................................................................... 41
§5.2 xDM 算法封装及 Codec Package 生成 ............................................................................42
§5.2.1 xDM 算法库封装 ....................................................................................................42
§5.2.2 配置生成 Codec Package .........................................................................................44
§5.3 DM6446 上H. 264 视频编码手工汇编优化 .................................................................. 44
§5.3.1 DM6446 的手工汇编优化方法 ..............................................................................44
§5.3.2 手工汇编优化的步骤 .............................................................................................. 46
§5.3.3 H. 264 块变换优化后性能比较 .............................................................................. 46
§5.3.4 使用内联函数 .......................................................................................................... 47
§5.4 DSP Server 的构建 .......................................................................................................... 48
§5.4.1 DSP Server 的脚本配置 ..........................................................................................49
§5.4.1 DSP/BIOS 的脚本配置 ...........................................................................................49
§5.5 ARM Linux 应用程序开发 ............................................................................................. 51
§5.5.1 Engine 的集成配置 .................................................................................................51
§5.5.1 Linux 多线程设计 ...................................................................................................51
§5.5.1 Video Thread 程序设计 .......................................................................................... 52
§5.6 演示系统的搭建 ...............................................................................................................53
§5.6.1 开发板的硬件配置 .................................................................................................. 53
§5.6.2 客户端播放器的设计 .............................................................................................. 53
§5.7 算法实验仿真结果及分析 ...............................................................................................55
第六章 总结与展望.....................................................................................................60
参考文献.........................................................................................................................62
作者在读期间科研成果简介.........................................................................................64
致 谢...........................................................................................................................65
摘要：

摘要移动视频智能监控是基于计算机视觉和模式识别领域，并结合现有的3G无线通信技术和嵌入式技术的应用，是视频监控进一步发展的方向。通过对视频图像高清采集、传输，进而监控和识别不同种类的物体，能及时发现和处理远程监控现场中的异常情况。本文的研究内容主要分为以下两个部分：无线网络中视频图像的传输策略和对现场目标物体的检测、识别和跟踪。进而能对异常现象进行定义和报警。智能视频监控具有广阔的市场前景，可应用于交通管理、安全防护、实时监控等领域。无线移动视频智能监控系统中涉及无线移动视频传输和计算机视觉中数字图像处理技术。本文将就两部分中影响系统鲁棒性的关键算法进行研究，包括对无线移动视频的鲁棒传输和鲁棒...

展开>> 收起<<
基于无线移动视频图像传输的远程监控与智能识别系统.pdf

共69页,预览7页

还剩页未读， 继续阅读

相关推荐

更多
立即下载
作者：牛悦 分类：高等教育资料 价格：15积分 属性：69 页 大小：2.95MB 格式：PDF 时间：2024-11-19

开通VIP享超值会员特权

  • 多端同步记录
  • 高速下载文档
  • 免费文档工具
  • 分享文档赚钱
  • 每日登录抽奖
  • 优质衍生服务

作者详情

相关内容

更多

推荐作者

热门标签

/ 69
评分 收藏
分享
立即下载 VIP免费下载
客服
关注