动态背景的序列图像中运动小目标检测算法

Detection method for moving small target in infrared image sequences with dynamic background

提出了一种动态背景下红外运动小目标的检测算法。该算法在连续四帧图像配准的基础上,对配准之后的图像序列进行沿时间轴的一维小波变换,实现目标和背景的分离。然后对主要包含运动目标信息的图像按像素相乘的方法得到目标增强图像,对其分割后提取运动小目标。实验结果表明,该方法能够有效的对红外慢速和快速运动的小目标进行检测。

A method to detect moving small target from infrared image sequences with dynamic background is proposed in this paper.The image sequences are processed with the 1-D wavelet transform to separate the target from background,which is based on four frame images are registered at the first step.And then the images that contain the moving target information are pixel multiplied,in order to obtain the target enhancement image and extract the moving small target after it is segmented.The result of this experiment indicates that the proposed method can detect and locate a slow or fast moving small target effectively.