自适应背景抽取算法

Adaptive Background Subtraction Algorithm

针对室内环境的特点 ,一种新的自适应背景抽取算法提出了 .该算法只需利用像素点采样值的亮度分量 ,采用统计的方法给场景每个像素点的亮度值建模 ,通过时间滤波器保持其序列均值和偏差 ;根据光照的变化对像素点亮度值的影响的分布情况 ,实现光照突变检测和像素点的自适应更新 ;采用将视频图像分块的方法 ,实现了对运动对象的快速抽取和跟踪 .与已有算法相比 ,该算法具有较低的时间、空间复杂度和可调的误检率 .实验表明 。

A new adaptive background subtraction algorithm aiming at characteristics of an indoor circumstance is presented in this paper. Our algorithm uses only luminance components of sampled image pixels and models every pixel with a statistical model, whose run mean and standard deviation are maintained by a time filter. When any sudden light change of circumstance occurs, the algorithm can detect and update pixels statistical parameters adaptively, according to the distribution of pixels luminance changes arising from the light change. With the method of partitioning a video frame into many small regular pixel blocks, our algorithm can extract motion objects from the background fast and track them effectively. It is shown that our algorithm can be realized with lower time and space complexity and adjustable object detection error rate with comparison to other background subtraction algorithms. Test results see the good performances of our method used in a motion tracking system.