基于混合高斯建模方法的运动目标检测方法研究与实现毕业论文导师张卫国.docVIP

摘 要 运动目标检测是计算机视觉研究领域的基础，它是从图像序列中去除静止的背景区域，将运动区域的前景检测并提取出来，如何有效地把感兴趣的目标如人、物等前景目标从复杂场景中分割出来，并对目标行为做出相应的检测，是计算机视觉研究的热点和难点。 本文对视频图像读取图像灰度化等，使用高斯滤波方法搭建在Visual C++6. 0上OpenCV软件进行设计OpenCV；混合高斯背景模型 ABSTRACT Moving target detection is the foundation of computer vision research field, which is still to remove the background from an image sequence region, the prospect of a motion area detected and extracted, how effectively the target of interest, such as human, material and other foreground objects from a complex scene carved out, and make the appropriate test target behavior, is hot and difficult computer vision research. In this paper, the video image reading, image processing such as Gray did a brief explanation, and respectively, mean filter, median filter and Gaussian filter three different methods to compare the advantages and disadvantages of each method, At the same time proposed to use a Gaussian filter. Finally, build in Visual C ++ 6. OpenCV software 0 of design, using a mixed Gaussian background modeling methods to achieve the detection of moving targets. In the background of the various modeling methods, Gaussian mixture background model with its simple, flexible and efficient characteristics of a classical method in the field, can be a good influence to eliminate background small perturbations generated by the moving target detection, more accurate moving objects detected in complex environments, robust and real-time. Key Words: Moving Object Detection; OpenCV; The background model 目录1 绪论 1 1.1 研究背景和意义 1 1.2 国内外研究现状与发展趋势 1 1.2.1 国内外研究现状 1 1.2.2 研究点 3 1.3 研究内容 4 1.4 本章小结 4 2 视频图像检测预处理 5 2.1 视频读取 5 2.2 图像灰度化 6 2.4图像去噪 7 2.4.1 中值去噪 8 2.4.2 均值去噪 9 2.4.3 高斯去噪 10 2.5本章小结 11 3 运动目标检测算法研究 12 3.1 几种常见的的介绍 12 3.1.1 光流法 12 3.1.2 背景减除法 13 3.1.3 帧差法 13 3.2 基于自适应高斯背景模型的目标检测 14 3.2.1 单高斯背景模型 15 3.2.2 混合高斯背景模型 15 3.3本章小结 15 4 基于混合高斯背景建模方法的运动目标检测 4.1 OpenCV介绍 17 4.2图像二值化 18 4.3 形态学运算 18 4.3.1 图像腐蚀 19 4.3.2 图像膨胀 19 4.3.3 开运算和闭运算 19 4.3.4 形态学运算的部分代码 20 4.3.5 形态学