基于兴趣点检测的目标识别算法

基于兴趣点检测的目标识别算法 A Target Recognition Algorithm Based on Interesting Points 修春波 XIU Chunbo 1 first-author 由霖 YOU Lin 1 赵鑫 ZHAO Xin 1 天津工业大学电气工程与自动化学院 天津工业大学电气工程与自动化学院 School of Electrical Engineering and Automation,Tianjin Polytechnic University School of Electrical Engineering and Automation,Tianjin Polytechnic University 为了有效地确定目标位置,提出了一种基于兴趣点检测的目标识别方法.根据像素点的灰度和边缘强度信息,确定兴趣点的分布位置.以兴趣点为中心,利用位势函数构建目标函数,目标函数取最大值的位置确定为目标位置.提出了一种改进的概率混沌优化算法求解目标函数的最大值.概率混沌优化算法将搜索空间分为原始搜索空间和精搜索空间,分别以不同的概率同时对两个空间进行搜索,并逐渐增大精搜索空间的搜索概率,从而可在确保算法具有全局寻优能力的前提下加快寻优速度.研究结果表明,该方法可有效地确定目标位置,且结果不受目标平移、旋转、缩放、甚至变形等影响,仿真结果验证了该方法的有效性. A new target detection algorithm based on interesting points is proposed to determine the position of the target effectively.The interesting points are determined according to its gray and edge information.An object function is constructed by potential functions whose centers are the interesting points.The target position is determined where object function is max.A probability chaos optimization algorithm is proposed to optimize the object function.The fine searching space and the origin space are searched simultaneously according to different searching probability,and the searching probability of the fine searching space is raised continuously,which can ensure the algorithm has the global searching ability and high searching efficiency.The position of the target can be determined by the method.And translation,rotation,scaling,and deformation of the target cannot affect the recognition result.Simulation results show that the proposed algorithm is valid and feasible. 目标识别 兴趣点 边缘检测 混沌优化 target recognition interesting points edge detection chaos optimization 10.14188/j.1671-8836.2011.06.012 TP391.41 国家自然科学基金资助项目（60808020） 2011-06-01 2021-04-01 6 57