一种基于改进混合高斯背景模型的运动目标检测方法研究

提出了一种固定摄像头下自适应的运动目标检测方法。该方法基于改进的混合高斯背景模型,通过在线更新模型学习率,实现背景模型更新。最后用背景差分法检测出运动目标。实验结果表明,与传统混合高斯模型的运动目标检测方法相比,该方法有较好的自适应性,能快速适应场景的变化。

Detection of moving objects based on improved Gaussian mixture model

This paper proposes adaptive sports target detection methods in the fixed cameras. The method is based on the improved the Gaussian mixture background model, through the online updating model adopted, realizing the background model updating. Then, the target motion is detected with the background finite difference. The experimental results show that compared with the detection of moving objects of the traditional Gaussian mixture model method, the method has better adaptability. It can quickly adapt to the change of scene.