视频捕捉的对象中，背景通常保持不变。一般分析中关注移动的前景物体，威力提取出前景物体，需要建立背景的模型，将模型和当前帧进行比对检测前景物体。前景提取应用非常广泛，特别是在智能监控领域中。



如果有不含前景物体的背景图片，提取前景的工作相对容易，只需要比对当前帧和背景图片的不同，调用函数absdiff实现。但是大多数情况，获得背景图片是不可能的，比如在复杂的场景下，或者有光线条件的变化。因此，就需要动态的变换背景。一种简单的办法是对所观察到的图片取平均，但这样做也有很多弊端，首先，这种办法在计算背景图片的前需要输入大量的图片，其次我们进行取平均的过程中不能有前景物体进入。所以一种相对好的办法是动态建立背景图片并实时更新。-Video capture of the object and the background usually remain unchanged. General analysis focus on the prospect of moving objects, power extract prospect object, it is necessary to set up the background model, the model and comparing the current frame detection prospect object. Foreground extraction are widely used, especially in the field of intelligent monitoring.



If there is no prospect of object containing the background picture, extraction prospect work relatively easy, only need than current frame and background picture is different, absdiff called function realization. But in most cases, it is impossible for the background picture, such as in a complex scene, or a light conditions change. Therefore, it needs to the dynamic transformation background. A simple solution is to take pictures of the observed average, but this also has a lot of disadvantages, first of all, this kind of method in the calculation of the background picture first need to enter a lot of pictures, secondl