本文選題：嵌入式　切入點：多線程編程　出處：《中山大學》2014年碩士論文

【摘要】：在人工智能技術和嵌入式技術不斷發(fā)展的今天，，模式識別技術有了長足的發(fā)展，也使得人們對視頻監(jiān)控有了更高的要求。而基于智能識別技術的視頻監(jiān)控在嵌入式產(chǎn)品中的應用也越來越廣泛。具體到監(jiān)獄安防領域，其對視頻監(jiān)控產(chǎn)品更是提出了實時監(jiān)控和預警的需求。在這種需求的激勵系啊，也為了能夠更好地完善監(jiān)獄安防工作，及時發(fā)現(xiàn)被監(jiān)控對象的異常行為并實現(xiàn)實時預警，文中重點對基于視頻處理技術的監(jiān)獄監(jiān)控系統(tǒng)進行設計與實現(xiàn)。在文中系統(tǒng)的設計和實現(xiàn)過程中，所采用的核心芯片為S3C2410，同時以國內某公司所研發(fā)的ARM9開發(fā)板基礎，對一種多層次實時監(jiān)控和預警的智能監(jiān)控網(wǎng)絡系統(tǒng)進行開發(fā)。該系統(tǒng)以嵌入式的Web視頻實時監(jiān)控設備為實現(xiàn)核心，在結合網(wǎng)絡通信技術和圖像識別技術的基礎上，實現(xiàn)以本地視頻智能分析預警為中心、以及區(qū)、市和省級監(jiān)控中心為骨干的多層次網(wǎng)絡。從而能夠更好的滿足監(jiān)獄安防這一特殊行業(yè)對安防產(chǎn)品的要求。 在本文系統(tǒng)中，微處理器采用S3C2410，而嵌入式的操作系統(tǒng)則主要采用Linux系統(tǒng)，從而有效節(jié)約了系統(tǒng)的研發(fā)和使用成本。其中，系統(tǒng)中的軟件部分還包括有對攝像頭的驅動和采集。而采集過程又可以分為單幀/多幀圖像的采集和視頻流采集。 在軟硬件環(huán)境構建的基礎上，系統(tǒng)還對基于C/S模式的網(wǎng)絡視頻實時監(jiān)控、網(wǎng)絡實時圖像智能分析和多級自動報警等功能進行了詳細說明。此外，通過對文中內容的深入研究，文中作者對各種嵌入式編程技術、多線程編程、圖像分割技術等進行了深入分析。同時，還對嵌入式系統(tǒng)的軟件開發(fā)流程以及嵌入式Linux的遠程調試等進行了分析和挖掘，為系統(tǒng)的開發(fā)和調試奠定基礎，有效縮短了系統(tǒng)的研發(fā)周期。
[Abstract]:With the continuous development of artificial intelligence and embedded technology, pattern recognition technology has made great progress. It also makes people have higher requirements for video surveillance. And video surveillance based on intelligent identification technology is more and more widely used in embedded products. It has put forward the demand for real-time monitoring and early warning for video surveillance products. In the incentive system of this demand, ah, in order to better improve prison security work, to detect the abnormal behavior of the monitored objects in time and to realize real-time early warning, This paper focuses on the design and implementation of the prison monitoring system based on video processing technology. In the design and implementation of the system, the core chip is S3C2410, and the base of the ARM9 development board developed by a domestic company. An intelligent monitoring network system with multi-level real-time monitoring and early warning is developed. The system uses embedded Web video real-time monitoring equipment as the core, and combines network communication technology and image recognition technology. To realize the multi-level network with local video intelligent analysis and warning as the center and the district, city and provincial monitoring center as the backbone, which can better meet the requirements of the security products of the special industry of prison security. In this system, the microprocessor uses S3C2410, while the embedded operating system mainly adopts Linux system, which saves the cost of research and development of the system. The software of the system also includes the drive and acquisition of the camera, and the acquisition process can be divided into single-frame / multi-frame image acquisition and video stream acquisition. Based on the construction of software and hardware environment, the functions of real-time monitoring of network video based on C / S mode, intelligent analysis of network real-time images and multi-level automatic alarm are described in detail. In this paper, various embedded programming techniques, multi-thread programming and image segmentation techniques are deeply analyzed. At the same time, the software development flow of embedded system and remote debugging of embedded Linux are analyzed and mined. It lays a foundation for the development and debugging of the system and effectively shortens the research and development cycle of the system.
【學位授予單位】：中山大學
【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7條

1 張伯虎;楊金柱;陳建莉;;智能視頻監(jiān)控系統(tǒng)分析與應用[J];安防科技;2006年11期

2 王滔,于潔;Linux系統(tǒng)下USB攝像頭驅動開發(fā)[J];電子技術應用;2004年11期

3 李雪;王普明;;基于高斯-拉普拉斯算子的圖像邊緣檢測方法[J];河南機電高等專科學校學報;2009年06期

4 郭會茹;;智能視頻監(jiān)控系統(tǒng)在看守所中的應用研究[J];淮陰工學院學報;2011年03期

5 彭立芹,馮文釗,曾志遠;一種改進的LOG遙感數(shù)字圖像邊緣提取算法[J];南京師范大學學報(工程技術版);2004年04期

6 余臘生;劉勇;;基于網(wǎng)絡的智能視頻監(jiān)控系統(tǒng)的設計與實現(xiàn)[J];計算機工程與設計;2009年16期

7 董軍浪;王慶飛;;基于改進遺傳算法的模糊C均值聚類算法[J];西安工程大學學報;2008年05期

相關博士學位論文 前1條

1 李志華;智能視頻監(jiān)控系統(tǒng)目標跟蹤與分類算法研究[D];浙江大學;2008年

本文編號：1681764