适应光照突变的运动目标检测算法

Motion Object Detection Algorithm for Sudden Illumination Changes

针对光照突然变化条件下的运动目标检测存在的问题,提出一种基于局部二值模式(LBP)的背景建模算法.首先分别计算背景帧和当前帧的LBP纹理特征图,然后将其对应像素进行异或运算,得到的当前像素作为前景的概率,并根据该概率自适应地更新背景,可以使其很快接近真实背景,再用背景减除法得到目标.实验结果表明,文中算法能有效地处理光照的突然变化,背景更新速度快,检测出的目标接近真实目标.

As the result of motion object detection can be greatly affected by sudden illumination changes, a background model based on local texture feature described by local binary pattern（LBP） which is gray-insensitive is proposed in this paper. Every coming frame and its background is converted into LBP first. The probability of a pixel belonging to foreground is generated by the operation of exclusive OR of the current image and the background＇s LBP. Then the background can be updated adaptively and quickly based on the probability. Finally, we can get the foreground objects easily by a background subtraction. Experimental results demonstrated the effectiveness of the proposed approach in dealing with sudden illumination changes. It can relative precisely obtain moving objects and update background model quickly.