基于改进的混合高斯模型背景建模在运动目标检测中的实现

混合高斯模型方法广泛应用于现代智能视频监控系统中的运动目标检测，但受制于不可改变的客观环境影响---尤其是光照、背景对象的轻微扰动等影响，往往检测效果不佳，为了解决该问题，使运动目标检测工作可以更好地适用于多种环境，笔者提出了基于改进的混合高斯模型背景建模方法，利用二维中值滤波和形态学原理对强点干扰噪声进行有效过滤，同时保护好信号变化的边界。实验表明：该方法可有效提高抵制噪声干扰效果，具有准确的目标检测效果。

GAUSSIAN MIXTURE MODEL BASED ON IMPROVED BACKGROUND MODELING IN THE REALIZATION OF MOVING OBJECT DETECTION

Gaussian mixture model approach is widely used for moving tanget detection in modern intelligent video surveillance system.Due to the immutable objective environmental impact,especially,light,background object minor disturbances and other effects,the detection is ineffective.In order to make the moving target detection work be better suited for a variety of environments,an improved Gaussian mixture model based background modeling method is proposed,using 2D median filtering and morphological principle of the strengths effectively to filter interference noise and protect the signal change of the border.Experimental results show that this method can effectively improve the resist noise with accurate target detection.