【摘要】：隨著社會公共安全需求的提高,視頻監(jiān)控正不斷延伸到社會的各個角落,傳統(tǒng)的人工值守的視頻監(jiān)控的方式已經不能滿足日益增長的監(jiān)控需求和監(jiān)控規(guī)模的需要。以全天候、無人值守、自動告警、實時等智能化為特點的智能監(jiān)控的需求變得越來越迫切,因此,對智能視頻監(jiān)控技術的研究成為了視頻監(jiān)控領域中一個研究熱點。 本文在研究智能視頻監(jiān)控的各種關鍵技術,對現(xiàn)有的各項研究成果和應用實例進行分析研究的基礎上,構建了基于DSP的智能視頻分析前端系統(tǒng),與網絡視頻服務器和監(jiān)控終端一起,建立了一個完整的智能視頻監(jiān)控系統(tǒng)。 智能視頻監(jiān)控關鍵技術的研究和實現(xiàn)基于ADSP-BF561的智能視頻分析前端是本文的研究重點。本文首先對智能視頻監(jiān)控關鍵技術進行了研究,包括運動目標檢測技術、運動目標跟蹤技術以及各種計算機圖像圖形處理算法。對智能視頻監(jiān)控關鍵技術的研究為構建智能視頻監(jiān)控系統(tǒng)提供了良好的理論基礎,并且為各項功能的實現(xiàn)提出了技術解決路線。構建基于ADSP-BF561的智能視頻分析前端是本文的另一個研究重點,課題針對BF561的雙核特性采用了雙核應用模式構建工程,結合開源軟件工具庫,實現(xiàn)了視頻H.264編碼和視頻分析的并行工作。工程包括視頻數(shù)據采集、智能視頻分析、視頻H.264編碼以及壓縮碼流的傳輸?shù)裙δ苣K,實現(xiàn)了入侵檢測,遺留物檢測,運動目標跟蹤,信號丟失檢測等眾多功能。最后針對DSP對工程進行了優(yōu)化,以提高處理效率,達到實時性要求。
[Abstract]:With the improvement of social public security demand, video surveillance is extending to every corner of the society. The traditional manual video surveillance can not meet the increasing demand for surveillance and the need of monitoring scale. The demand of intelligent surveillance, which is characterized by all-weather, unattended, automatic alarm, real-time and so on, has become more and more urgent. Therefore, the research of intelligent video surveillance technology has become a research hotspot in the field of video surveillance. Based on the research of various key technologies of intelligent video surveillance and the analysis of existing research results and application examples, the intelligent video analysis front-end system based on DSP is constructed in this paper. Together with the network video server and the monitoring terminal, a complete intelligent video monitoring system is established. Research and implementation of key Technologies of Intelligent Video Surveillance the research focus of this paper is intelligent video analysis front end based on ADSP-BF561. In this paper, the key technologies of intelligent video surveillance are studied, including moving target detection, moving target tracking and various image processing algorithms. The research on the key technology of intelligent video surveillance provides a good theoretical basis for the construction of intelligent video surveillance system, and provides a technical solution for the realization of various functions. The construction of intelligent video analysis front end based on ADSP-BF561 is another research focus of this paper. In view of the dual-core characteristic of BF561, the paper adopts dual-core application mode to build the project, and combines with open source software tool library. The parallel work of video H.264 coding and video analysis is realized. The project includes video data acquisition, intelligent video analysis, video H.264 coding and the transmission of compressed code stream. It realizes many functions such as intrusion detection, legacy detection, moving target tracking, signal loss detection and so on. Finally, the project is optimized for DSP to improve the processing efficiency and meet the real-time requirements.
【學位授予單位】：華北電力大學
【學位級別】：碩士
【學位授予年份】：2014
【分類號】：TN948.6

【參考文獻】

相關期刊論文 前4條

1 艾海舟,呂鳳軍,劉偉,郭艷芳;面向視覺監(jiān)視的變化檢測與分割[J];計算機工程與應用;2001年05期

2 宋愷;蔡軼珩;沈蘭蓀;;利用光流-Snakes算法對動態(tài)舌體跟蹤的初步研究[J];計算機工程與應用;2007年13期

3 王亮,胡衛(wèi)明,譚鐵牛;人運動的視覺分析綜述[J];計算機學報;2002年03期

4 黃銘;陸思良;孔凡讓;;魔方還原機器人的視覺子系統(tǒng)設計與實現(xiàn)[J];機械與電子;2013年05期

相關博士學位論文 前1條

1 焦波;面向智能視頻監(jiān)控的運動目標檢測與跟蹤方法研究[D];國防科學技術大學;2009年

，

本文編號：2226450