摘要
介绍了在视觉监控领域经常用到的几种基于自适应背景模型的经典方法,如非参数模型、W4模型、单高斯模型和混合高斯模型等。通过实验,着重从系统的敏感性、实时性和运动对象分割的完整性等方面分析比较了上述方法在检测过程中的优势及不足之处,以期在工程实践中有一定的参考作用。
The paper introduces several classical methods for vision- surveillance using adaptive background models,such as the non-parametric model,W4 models, single C-aussian model and mixture Gaussian model and so on. Compares the proposed methods＇predominance and deficiency on the performance of detection sensitivity, real -time of the system and the full of the segmentation of moving target by experiments, which probably has some benefits for the project.
出处
《现代电子技术》
2007年第5期123-127,共5页
Modern Electronics Technique
关键词
视觉监控
背景减除
混合高斯模型
运动物体检测
vision - surveillance
background subtraction
mixture Gaussian model
moving target detection
