摘要
背景减除法常采用混合高斯模型作为背景模型来进行目标检测,它可以自适应学习并表示分布复杂的背景.混合高斯模型在光线变化缓慢的情况下表现很好,但是在光线快速变化的情况下,由于高斯背景无快速更新机制,无法应对光线迅速变化的情况.通过对混合高斯模型进行优化,提出了一种改进的混合高斯模型检测算法,并通过实验证明了新算法明显提高了运动目标检测的准确度.
Gaussian mixture model is frequently used in background subtraction as the background model to realize objects detection. It can be adaptive to learning and expresses complex - distributional background. Gaussian mixture model performs well on the condition of light change slowly, but it is unable to cope with rapid -change light. The Gaussian background without rapid updating mechanism. This paper tries to optimize Gaussian mixture model (GMM) and puts forward an improved Gaussian mixture model detection algorithm. Experiments prove that the new algorithm improves the accuracy of moving objects detection significantly.
出处
《河南工程学院学报(自然科学版)》
2013年第3期65-68,共4页
Journal of Henan University of Engineering:Natural Science Edition
基金
安徽省自然科学研究项目(KJ2012Z267)
关键词
混合高斯模型
背景减除法
运动目标检测
gaussian mixture model
background subtraction
moving objects detection