基于openCV运动物体检测的方法研究
关键词：灰度化，预处理，模糊处理，帧差法，高斯背景建模，光流
Study on the method of object detection based on openCV
Abstract: With the development of science and technology，as the human alternative, the robots have been widely used in more and more fields. And the robot technology is also increasingly developed.Robot vision is a very important part in a robot system. It mainly includes image acquisition, pre-processing and image detection part. This paper mainly studies the robot vision system for image preprocess and detection of moving objects. This paper studied the color model and the video image sequence, the image adjustments and changes, make the image more easily identify and deal with, avoid the interference of various adverse effects on the image parameter, in order to facilitate the information selection or inhibition. The use of Visual Studio software, through image processing, target detection, and position。Then we implanted OpenCV image processing library, in the analysis of the background difference, frame difference,optical flow method based features and other detection methods, the moving target detection are using thesedifferent methods, and the experiments proved the feasibility and reliability.
key words:Gray, Pretreatment,Fuzzy processing,Frame difference method,The background difference method,Optical flow method
目录
1. 绪论 5
1.1 研究背景与意义 5
1.2 研究现状分析 5
1.3研究目的及论文组织 6
2.视频图像的处理基础 8
2.1色彩基础 8
2．2颜色空间及其模型 9
2.2.1 RGB颜色模型 10
2．2.2 HSI颜色模型 11
2．2.3RGB与HIS颜色模型的转化 11
2.3图像灰度化 13
2.4视频图像序列 14
2.5本章总结 14
3. 基于OPenCV的图像预处理 16
3.1 平滑处理 16
3．1.1 简单模糊 17
3.1.2中值模糊 17
3.1.3高斯模糊 18
3.1.4双边滤波 19
3.1.5 4种模糊处理的选择 19
3.2图像二值化处理 20
3.3图像形态学 20
3.4本章总结 22
4.运动目标的提取与检测 23
4.1帧差法 23
4.2背景差分 25
4.2.1 高斯背景模型的建立 26
4.2.2 opencv中高斯模型的应用 27
4.3光流场法 28
4.3.1 Lucas-Kanade方法 29