受全球都市化的影响，大量的人群聚集易造成人群拥挤、发生踩踏伤亡等不幸事件，同时人群密度的增大也将会使城市的公共交通迎来短期的人流高峰，人群的高度拥挤若得不到及时有效的疏散，对城市的治安将会造成较大的威胁。因此，自动的对人群信息进行有效地分析从而估计出人群密度，已经成为智能视频监控的研究热点。 本文研究的重点是如何对视频监控中的人群密度进行估计。首先需要对监控场景的视频图像提取人群前景图像，然后对其提取人群密度特征，最后将获得的密度特征值送入分类器中，从而得到人群密度。 在人群前景提取方面，本文首先采用加权平均法对人群图像进行灰度化，然后采用中值滤波法进行噪声的消除，最后采用自适应的视频帧差法来构造背景图像，通过背景减操作获得人群前景图像。 在人群密度的特征提取方面，本文首先介绍了人群密度监控技术的研究现状及基本理论，通过分析可知，基于纹理的分析方法对于高密度且有遮挡的人群场景非常有效。因此，本文采用基于灰度共生矩阵的纹理分析方法来提取人群密度特征，通过实验研究确定了灰度共生矩阵的最佳参数，并选择了能量、熵、对比度及逆差矩4个重要的特征作为人群图像的纹理特征。 在模式识别的分类问题上，本文采用支持向量机作为估计人群密度的分类器，按照分类规则，使用训练样本建立了支持向量机分类器模型，通过实验采用 “粗调加细调”方法测试研究了径向基核函数的核参数和惩罚参数C的最佳组合的选取。 最后，为了验证本文算法的有效性，分别对两个不同的人群视频进行了实验，测试样本的准确率均达到了95%以上。实验结果表明本文方法简单有效，便于应用在实际场景中，为有关部门能更好地保障公共安全提供了有力的帮助。

For the impact of global urbanization，dense crowds moving and rendezvousing would easily cause serious trampling accidents. These unfortunate accidents happen frequently in the world. At the same time, the increase of crowd density will make the city's public transport face short-term peak. If the congested crowd can not be evacuated in time, it will pose a greater threat to urban public security. Thus, automatic analysis of the crowd information and estimating of the crowd density have become a research focus in the intelligent video surveillance. The focus of this thesis is how to estimate the crowd density of video surveillance. Firstly, extract the crowd foreground image of video image in monitoring scene, and then extract the crowd density characteristics of the foreground image. Finally, the crowd density characteristics are sent to classifier, thereby obtaining the crowd density. In the aspect of crowd foreground extraction, the weighted average method is adopted to gray the crowd image, then the median filtering method is used to eliminate noise, and finally, adaptive change detection mask method is applied to construct the background image, thereby getting crowd foreground image by background subtraction operation. As to the crowd density characteristics extraction, the development and basic principles of crowd density monitoring system both in China and abroad is introduced. The texture-based analysis method is very effective for the high-density scene. Therefore, the texture analysis method based on Gray Level Co-occurrence Matrix is used to extract the crowd density characteristics. The best parameters of GLCM are determined by experiments, and the four important characteristics of the energy, entropy, contrast and homogeneity are selected as the texture features of crowd image. Regarding pattern recognition classification, the support vector machine (SVM) is adopted as classifier to estimate the crowd density. According to the classification rules, the support vector machine classifier model is established by use of training samples. By experiments, “The coarse and fine tuning” method is used to test and study the selection of the best combination of the radial basis function’s kernel parameter and penalty parameter C. Finally, in order to verify the effectiveness of the algorithm, experiments are carried out on two different crowd videos. The accuracy rate of test samples reaches more than 95%. The experimental results show that the method is simple and effective with convenience for application in the actual scene.