摘要
针对动态背景下运动目标检测的问题,最大限度地降低背景对运动目标检测的影响,提出了一种基于相位相关法和傅里叶梅林变换的动态背景下运动目标检测算法。动态背景下运动目标检测的主要部分是背景运动补偿,首先利用相位相关法和傅里叶梅林变换估计全局运动参量,然后根据全局运动参量利用双线性内插法进行背景匹配,最后对配准后的图像利用帧间差分法提取运动目标。实验表明,该算法具有一定的鲁棒性,能有效地检测动态背景下的运动目标。
According to the problem of moving object detection under dynamic background, decreasing the effect of background in moving object detection, a method for moving object detection under dynamic background based on phase correlation and Fourier-Mellin trans- form is introduced. The main part of moving object detection under dynamic background is background compensation. Firstly, use the al- gorithm of phase correlation and Fourier-Mellin transform to estimate the global motion vector. Secondly, use bilinear interpolation method and the global motion vector to match the two adjacent frame backgrounds. Finally ,use registered images and frame difference algorithm to get the moving object. The result of simulation shows that this algorithm is robust and can detect moving object under dynamic background effectively.
出处
《计算机技术与发展》
2013年第7期104-107,共4页
Computer Technology and Development
基金
装甲兵工程学院创新基金(2012CJ42)
关键词
动态背景
运动目标检测
相位相关
傅里叶梅林
dynamic background
moving object detection
phase correlation
Fodrier-Mellin
