基于相关配准的帧间差分算法

针对帧间差分法在摄像头运动时受动态背景严重干扰的问题,提出了一种基于图像配准的运动目标检测算法。首先将中值滤波后的连续两帧图像配准,配准时先在前一帧图像中选取背景,即背景图像,用区域相关法将后一帧图像与背景图像配准;接着将配准后的2帧图像差分得到帧间差分图像,即帧差图像,再用数学形态学的开运算去掉帧差图像中的一些细小噪声;最后将连续两帧去噪后的帧差图像逻辑与运算,得到运动目标检测结果。实验结果表明,在摄像头运动时的动态背景下,该算法有效地抑制了动态背景的干扰,准确地检测出了运动目标的边界,提高了运动目标检测在动态背景下的应用价值。

The Inter-frame Difference Algorithm Based on Correlative Registration

To solve the problem that the inter-frame difference algorithm is seriously interfered by the dynamic background when the camera is moving, this paper proposes a moving object detection algorithm based on image registration. Firstly, take image registration of two consecutive frames after the median filtering. The regional correlation method is put forward in this paper to register the latter frame with the background image, which is selected from the previous frame. Then inter-frame difference algorithm is adopted to get frame difference image, and mathematical morphological opening operation is used to remove some small noise in the frame difference image. Lastly, the result of moving target detection is gotten from a logical AND operation of two consecutive de-noised frame difference image. The experiment results show that the proposed algorithm effectively suppresses the interference of the dynamic background and accurately detects the boundary of the moving target in the dynamic background of a moving camera. It increases the application value of the moving target detection in dynamic background.