一种基于多层背景模型前景检测算法.pdfVIP

一种基于多层背景模型的前景检测算法 杨 涛，李 静，潘泉，程咏梅 （西北工业大学自动化学院，西安，710072 ） 摘 要：动态场景中的前景检测是后继处理的基础和制约整个智能视频监控系统稳定性、可靠性的关键。 为了在保证运动目标检测的基础上，进一步检测出前景中的静止目标并消除“鬼影（Ghost）”，提出一种 基于多层背景模型的前景检测算法。将背景分为参考背景和动态背景两层，分别采用单高斯和混合高斯模 型进行背景建模。在线检测时采用动态背景提取变化前景，用动态背景与参考背景之间高斯分布的差异提 取静止前景，同时，通过逐层分析、比较输入像素与两层背景模型分布的相互关系，快速消除 Ghost，降 低虚警。实验表明，多层背景模型具有良好的检测性能和实时性，为后继跟踪、分类等处理提供了坚实的 基础。目前，以该算法为核心构建了一个实时目标检测、跟踪系统 GreatWall system，对图像大小为 320x240 的视频序列的平均处理速度达到 15帧/秒。 关键词：多层背景模型；背景建模；混合高斯；静止前景检测 中图分类号：TP391 文献标识码：A A Novel Multiple Layer Background Model For Foreground Detection YANG Tao, LI Jing, PAN Quan, CHENG Yongmei College of Automation, Northwestern Polytechnical Unviersity, Xi’an, 710072, China yangtaonwpu@163.com, jinglinwpu@163.com Abstract: Foreground detection is an important research problem in visual surveillance. In this paper, we present a novel multiple layer background model to detect and classify foreground into three classes, moving object, static object and ghost. The background is divided into two layers, reference background and dynamic background. Single Gaussian model and Gaussian mixture model are used respectively. Compared with many existing background models, a unique characteristic of the proposed algorithm is that through analyzing the Gaussian distributions of the two layers, static object and ghost are correctly labeled. A real-time object detection and tracking system named as GreatWall system is developed and tested under indoor and outdoor scenes with various scenarios. Extensive experimental results demonstrate that the proposed algorithm is effective and efficient and the processing speed of the system reaches 15fps for the image size o