基于综合匹配的鸟类视频中运动物体的检测

The Object Detection in Birds Video Based on Synthesized Matching Method

鄱阳湖是中国最大的湿地。每年秋末冬初,会有成千上万只候鸟来此越冬,该地区的自然生态系统和这些鸟类的活动有着密切的联系,因此,研究该地区的鸟类是了解该地生态系统的一个重要方式。本文采用帧间差分法对该地所拍摄鸟类视频数据进行运动物体检测。由于视频在获取过程中存在摄像机的移动,所以在处理过程中首先需要估计背景的运动并加以补偿。背景运动的估计是在常用的块匹配方法的基础上作了一定的改进,通过多个匹配块的匹配结果来计算运动矢量。为提高正确率,块的匹配在差分图像上进行。最后,利用在鄱阳湖录取的视频对本文所述方法进行了实验,实验结果表明本文所述方法对背景的去除和运动物体的检测的有效性。

Poyang lake is the biggest wetland of China.There are thousands of migrant birds come here for hibernation in the autumn and winter.The natural ecosystem of the region has great relationship with these birds activities,so it is important to study the birds for the ecosystem research.We attempt to detect the moving birds from the video by frame difference method.Because of the movement of the camera lens during shooting,the background motion estimation and compensation have to be applied first.The background motion estimation is a modified block-matching algorithm,calculating the motion vector based on many matching results of the multiple matching blocks.The block matching algorithm is calculated on the difference image for improving accuracy.At last,an experiment of the method presented in the paper is conducted,using the video shooting at Poyang lake.The experiment shows that this method is efficient on background removal and moving object detection.