动态阈值的Vibe运动目标检测

Vibe算法是一种高效的像素级背景建模算法，但是它同混合高斯模型一样，不能适应光线突变的问题，在对Vibe算法的基础上提出了一种在静态场景下对光照变化鲁棒的运动目标检测方法。该方法首先利用Vibe模型建立背景样本集，并利用Vibe模型对判别为背景的像素对背景帧进行更新。其次视频当前图像帧与背景帧差分，并采用Otsu算法计算图像的分割阈值来检测运动目标。实验结果表明，改进的方法能够很好地消除由于环境光照变化引起的“曝光”现象，提高了运动目标检测的精确度，并且改进的算法对室内场景下的阴影也有较好的抑制作用。

Vibe moving object detection method based on dynamic threshold

Vibe algorithm is regarded as an effective pixel-level background modeling algorithm. But it can not be adapted to the light mutations as same as the mixed Gauss model. On the basis of Vibe algorithm, it proposes that a robust moving target detection method for light charges in static scenes. Firstly, the Vibe mode is used to establish the background frame, and it updates background frame using the pixel which judged as background by Vibe model. Secondly, differences of the current image frame and the background video frame are put. And Otsu algorithm is used to calculate the image segmentation threshold to detect moving targets. Experimental result shows that the improved method can eliminate the “exposure” phenomenon caused by the change of the environment, improves the accuracy of moving target detection, and has a good effect on the shadow of the indoor scene.