基于背景遗传模型的运动目标检测算法

Moving Targets Detection Algorithm Based on Background Genetic Model

获取每个像素或者一组像素的时间序列模型是在背景中检测复杂运动目标的较好方法。但是,这些方法计算复杂度高而且耗时、耗资源。为了在简化计算基础上获得更精准检测效果,本文采用了基于背景遗传模型的运动目标检测算法。此算法通过当前帧图像宏块与背景模型相应宏块的最大相关度情况,决定是否进行遗传操作;对不满足适应度评估条件的宏块进行遗传操作。并采用Codebook算法对此宏块进行前景检测。通过仿真结果表明,本文算法与传统的Codebook运动目标检测算法相比,运算量更少,背景更新更有效,具有更高的精确度和鲁棒性。

A time series model for each pixel or a group of pixels is a good method to detect the complex moving targets in the background. However, these methods are cost computational complexity. In order to obtain a more accurate detection result, a moving target detection algorithm based on background genetic model is adopted in this paper. In this algorithm, the maximum correlation degree of the macro block is determined by the current frame and the background model, and the genetic operation is determined. Codebook algorithm is used to detect the foreground of macro block. The simulation results show that the calculation of proposed algorithm is less than the traditional Codebook moving target detection algorithm, and the algorithm is more efficient and has higher accuracy and robustness.