一种改进W4算法的运动目标检测研究

针对经典W4背景建模算法只能克服光照强度的微小变化以及背景的轻微运动等问题,提出了一种新的运动目标检测算法。首先,利用均值法进行背景初始化选出静止像素集合,消除背景中运动目标的干扰;其次,给定背景初始帧,用经典W4算法计算出每个像素点的最小灰度值、最大灰度值以及最大帧间差分值;然后,对每个像素点提取的最小灰度值和最大灰度值进行线性加权,并且与均值法得到的初始背景相结合建立稳定的背景模型,克服了移动、阴影、光照突变等影响;最后,比较当前帧与背景模型从而检测出准确的运动目标。实验证明,与其它均值法、经典W4算法以及混合高斯背景建模方法相比较,改进方法不仅耗时短而且取得了较为理想的检测效果。

A MovingTarget Detection Research Based on Improved W4 Algorithm

A new improved algorithm based on the classical W4 is proposed to detect the moving targets because of the deficiency in light intensity and the slight movement of background. Firstly, using the multiple frames averaging method to initialize the background model, which can overcome the effect that caused by moving objects in the back_ground;Secondly, calculated the minimum, the maximum and the maximum frame difference gray value;Then, com_bined with the initial background which is established by average method to build another stable background;Finally, the moving targets can be accurately detected by comparing the current frame and the background. Experimental results show that compared with the average method, W4 algorithm and the gaussian mixture model, the proposed method not only saves time but also has obtained the good effect.