Matlab视频中的运动目标检测方法研究
关键词 视频运动目标检测；背景减法；积累差异；Matlab
毕业设计说明书（论文）外文摘要
Title Video Moving Object Detection Methods
Abstract
Video moving object detection is the key technology of the visual surveillance system in the bottom of the whole system ,is the basic of a species of follow-up high-level processing or application such as video analysis,video coding,video retrieval and video surveillance.The content of this thesis is using background subtraction detecting moving object in video.First we based on differences accumulation informations to make background modeling,then Otsu method was used to threshold the background subtracting image.After that twice region growing operation was used for making connective regions.Finally, moving objects were related minimal Eculidean distance and motion trajectories were painted. Test platform is Matlab，is used to design and develop a basis for moving object detection and tracking system. After the user input a video file, the system will automatically extract the moving objects in the video automatically, calculate their bounding boxes and number .
Keywords Video moving object detection ; background subtraction ; accumulative differen ; Matlab
目 次
1 绪论 1
1.1 研究背景及意义 1
1.2 视频运动目标检测的研究现状 2
1.3 论文研究内容及章节安排 3
1.3.1 论文研究的内容 3
1.3.2 论文的章节安排 3
2 基于积累差异的视频运动目标检测 4
2．1 背景减法 4
2.1.1 背景减法原理 4
2.2.1 背景模型及更新 5
2．2 基于积累差异的背景建模 5
2.2.1 积累差异 5
2.2.2 积累差异背景建模 6
2.3 Otsu自适应阈值化及目标轮廓提取 7
2.3.1 Otsu阈值化算法 7
2.3.2 目标轮廊提取 8
2.4 运动目标分割 8
2.5 目标质心关联 9
2.5.1 质心标记 9
2.5.2质心关联 9
2.6 运动目标跟踪 10
2.7 小结 11
3 Matlab平台实现视频中运动目标的检测 12
3.1 算法测试平台构架 12
3.2 平台实现及使用的Matlab功能函数 13
3.2.1 平台实现 13
3.2.2 Matlab功能函数 13
3.3 算法框架 14
3.4 实验结果及分析 14
3.5 小结 16
结 论 17