一种用于弹道终端飞机红外图像机轴检测的算法  

Algorithm of airplane axis detection for infrared images of airplane at missile terminal

作　　者：张瑞[1] 邱锋[1,2] 杨星[1] 刘琳琳 

机构地区：[1]西北工业大学航天学院,陕西西安710072 [2]西安志峰路桥有限公司,陕西西安710072

出　　处：《现代电子技术》2014年第8期61-64,共4页Modern Electronics Technique

基　　金：航空基金(2012ZA53004)

摘　　要：以制导引信一体化(GIF)为背景,利用红外成像导引头图像信息,对飞机目标姿态识别技术展开研究;在提取出飞机目标图像骨架的基础上,提出一种改进的Hough变换与最小二乘拟合相结合的算法检测出了飞机轴线。利用Hough变换鲁棒且不需启发信息的特点进行初步检测,确定直线存在的大致区域;利用最小二乘拟合法确定直线区域内特征点回归直线的精确参数;对该算法进行仿真。仿真结果表明,提出的该算法提高了飞机轴线检测的检测率和检测精度,降低了对Hough变换的分辨率要求,可以减小算法整体的空间开销。Based on the infrared imaging guidance system,the attitude recognition technology of target airplane was studied by using image information of infrared imaging seeker on the back ground of guidance integrated fuzing（GIF）technology. The skeleton of a target airplane image is extracted by a thinning algorithm,and then an algorithm which combines the improved Hough transform with the least square fitting is proposed to detect the airplane-axis. The preliminary detection is carried out ac-cording to the robust and heuristics-free characteristics of Hough transformation to determine image regions where a line may exist. The least square fitting is used to confirm the accurate parameters of feature point regression straight-line in a straight-line re-gion. The simulation results show that the algorithm has improved the detection rate and detection precision of airplane＆#39;s axis, decreased the requirement of resolution to Hough transformation,and reduced the space demand of the algorithm.

关 键 词：姿态识别 跟踪 HOUGH变换 最小二乘拟合 

分 类 号：TN919-34[电子电信—通信与信息系统]