基于图像识别的摄像头故障检测系统设计

基于图像识别的摄像头故障检测系统设计(任务书,开题报告,论文12000字)
摘 要
摄像头视频监控系统是一种防范能力较强的综合系统,近年来随着人们的安全意识不断提高,摄像头视频监控系统已逐步渗透到国民生活的诸多领域,银行监控、社区安保监控、交通违章监控等等,面对诸多的摄像头应用,要依靠人工去检测每个摄像头的监控画面是否正常,这是十分艰难的工作。
因此,本文设计和实现了一套基于图像识别的摄像头故障检测系统,能够自动检测摄像头所拍摄的图像是否正常,判断画面是否出现异常以及人为干扰,并实时发出告警提示用户进行相应的维护与检修。
监控摄像头作为视频监控系统的最前端,比较容易受到外界的干扰,特别是一些可疑人员为了躲避监控,通常会对摄像头做出各种干扰动作,常见的摄像头干扰包括,摄像头镜头旋转或位置移动,摄像头镜头遮挡,摄像头镜头被喷漆等等。相关研究工作发现,干扰前后图像的特征会发生明显变化,基于这些研究工作,本文利用SIFT图像特征转换函数获取图像的特征,通过图像特征的变化来判断摄像头是否发生干扰,并结合PCA-SIFT来对图像特征向量进行降维,减少计算量,提高效率。 [来源:http://www.doc163.com]
除了人为的干扰外,监控摄像头本身还会受到网络传输,硬件故障,周围环境等方面的影响,而产生图像异常,这些异常包括静帧、黑屏、马赛克、色偏等等。影响监控质量,导致有效信息丢失。通过对视频质量检测的常用方法进行研究,本文设计和实现了一套监控摄像头画面异常检测的方案,该方案在应用层视屏终端进行监控图像质量分析,实时检测图像异常的出现,并即时做出提示。
关键词:干扰检测、视频质量、SIFT、图像异常检测
Abstract
Video surveillance system is an integrated system of prevention ability, in recent years, with the increasing awareness of people's security, video surveillance systems have penetrated into all areas of national life, banking supervision, residential security monitoring, traffic violation monitoring, etc. With so many camera applications, to detect whether each camera monitor screen is normal, will be very difficult.
Therefore, we designed and implemented a Surveillance camera interference detection system that can automatically detect the captured images to determine whether the picture is abnormal and human disturbance.
Surveillance cameras are in the front of video monitoring system. They are easy to be disturbed. Especially some suspicious persons in order to avoid monitoring, usually make various interference action on the camera, including camera rotation or lens shade and so on. Related studies found that a significant change in image characteristics may occur after the interference, based on these studies we can using SIFT image feature functions to obtain characteristics of the image transformation the image feature to determine whether the camera interference occurs, and combining PCA-SIFT to the image feature vector dimension reduction, reducing the amount of calculation and improve efficiency.
In addition to human disturbance, the surveillance camera itself is also affected the surrounding environment, network transmission, hardware failure. It produce an image anomalies which include static frame, black, mosaic, color cast, and so on. Resulting in loss of effective information. Through the research on the quality of video detection, we designed and implemented a surveillance camera picture anomaly detection scheme which can detect static frames, black, mosaic, color cast, and can immediately issue an alarm. [资料来源:Doc163.com]
Key Words: Interference detection, video quality, ,SIFT,abnormal image detection


目 录
第1章 绪论 1
1.1 课题研究背景及意义 1
1.2 国内外研究现状 1
1.3 本文主要工作 2
1.4 本文结构安排 2
第2章 相关背景技术和工作 4
2.1 相关技术介绍 4
2.1.1 图像特征提取 5
2.2.2 特征转换算法 7
2.2.3 特征转换算法 7
2.2 小结 7
第3章 核心技术 8
3.1 基于SIFT的摄像头干扰检测 8
3.1.1 对SIFT算法的总的概述 9
3.2 视频图像异常检测 10
3.2.1 黑屏 10
3.2.2 静帧 11
3.5.3 马赛克 12
3.5.4 色偏检测 15
第4章 系统设计与实现 16
4.1 系统需求 16
4.2 系统整体设计 16
4.2.1系统流程 18
4.3 图像采集模块 21
4.3.1 图像采集核心代码 21
4.4 摄像头干扰检测模块 22
4.4.1 特征变换核心算子 23
4.4.2 实验结果 23
4.5 图像故障检测模块 28
4.5.1 故障检测模块核心代码 28
4.6 小结 33
第5章 总结与展望 34
5.1 全文总结 34
5.2未来展望 34
参考文献 36
致 谢 40 [来源:http://Doc163.com]