【作者】 李峥； 彭绍武； 孙彬； 桑农；

【Author】 Li Zheng, Peng Shao-wu, Sun Bin, Sang Nong(Key Laboratory of Ministry of Education for Image Processing and Intelligent Control, Wuhan 430074,lz9721 Institute for Pattern Recognition and Artificial Intelligence, Huazhong University of Science and Technology, Wuhan 430074)

【机构】 图像信息处理与智能控制教育部重点实验室； 华中科技大学图像识别与人工智能研究所；

【摘要】 在利用速度滤波器估计运动时,为了获得较高的估计精度,需要使用大量的、具有不同速度响应的速度滤波器,这将带来计算量巨大的困难。本文提出利用遗传算法来寻求局部最优的速度滤波器,并将其用于一维遮挡运动的速度估计问题。把运动图象对速度滤波器的匹配误差作为遗传算法中的适应度函数。用遗传算法在可计算的范围内找出和运动图象匹配误差最小的速度滤波器,该速度滤波器对应的速度最接近运动图象的真实速度。该方法不需要知道运动信号的表达式,适应性比较强。文中讨论了将此方法推广至二维遮挡运动时可能遇到的问题。更多还原

【Abstract】 It will bring forth great difficulties that a large number of velocity filters which have different velocity response have been used to obtain the higher estimation precision. This paper brings out a new method to pursuit the local optimal velocity filters by using genetic algorithm, and this method has been used to solve the problem of one dimension motion velocity estimation. This method takes the matching error between the motion image and the velocity filter as the adaptability function. The genetic algorithm can find the velocity filter which has the minimum matching error to the motion image. So the velocity which corresponds to the velocity filter is the velocity which is the nearest to the actual velocity. This method does not need the representation of the motion signals, so the adaptation of this method is good. And this paper also discussed the problems which can be met when this method is applied to two-dimension occlusion motion.更多还原