摘要
论文提出了一种将单高斯建模法与混合高斯建模法相结合的运动目标检测方法。针对实际场景中的部分动态背景会给运动目标检测带来很大干扰,混合高斯建模法提取目标不完整、实时性差的问题,结合单高斯建模法处理的优点,将图像自动分割为静态背景区域和动态背景区域,分别检测,并将得到的运动目标检测图像进行拼接及后续处理。实验结果表明,用该方法提取目标完整度高,实时性好,对动态背景干扰不敏感,具有较好的鲁棒性。
This paper presents a moving target detection method based on combination of single Gaussian model and Gaussian mixture model. In practical situations, part of the dynamic background will bring a lot of interference to moving target detection, and traditional Gaussian mixture modeling method can＇t extract the complete target with poor real-time. Considering the advantages of treatment with single Gaussian model, the image is cut into static and dynamic background area which is processed with single Gaussian modeling method and Gaussian mixture modeling method respectively. After that, the moving target image stitching is gotten and some follow-up treatment treatments are done. Experimental results show that using this method can extract full target and have high real-time, it is insensitive to dynamic background interference, and has good robustness.
出处
《计算机与数字工程》
2014年第5期791-795,共5页
Computer & Digital Engineering
基金
国家自然科学基金项目(编号:61370180)资助
关键词
运动目标检测
单高斯模型
混合高斯模型
动态背景
图像分割
moving target detection, single gaussian model, gaussian mixture model, dynamic background, image segmentation
