目标识别与定位的快速融合算法  

A Fast Fusion Algorithm for Target Identification and Localization

作　　者：史战果[1,2] 张筱[1] 吴迪[1] 王胜永[1] 朱磊[1] 张秀彬[1] 

机构地区：[1]上海交通大学电子信息与电气工程学院,上海200240 [2]南京军区兵种训练基地,南京212400

出　　处：《弹箭与制导学报》2012年第3期201-204,共4页Journal of Projectiles,Rockets,Missiles and Guidance

摘　　要：针对目前的车载火控系统不能自动识别目标类型、准确定位造成反应速度慢和命中率低的问题,文中提出了一种基于图像信息的目标识别与定位的快速融合算法。该算法通过火控系统配置的图像采集装置实施对前方目标的实时监视,采用偏序关系认准、模板匹配和视差测距算法融合实现对目标类型的识别与距离检测,为火控系统的自动化、智能化水平的提升提供了一种有效的技术手段。该算法运算周期小于50ms,识别准确率高达98%,在2000m视距内距离检测误差小于1m。To solve the problem of slow response and low hitting probability caused by current vehicular fire control system can not achieve automatic target type identification and accurate target location, a fast fusion algorithm for target identification and location based on target identification of image information was proposed. The algorithm uses image collecting devices of the fire control system to realize the real- time monitor of the targets ahead. It also uses integration of partial relation, template matching and parallax ranging to identify the type of target and measure the distance. The algorithm provides a very effective technique to improve the automation and intelligence level of the vehicular fire control system. The operation period of algorithm is less than 50ms with the identification accuracy rate more than 98% and the ranging tolerance less than one meter on the condition that the range of visibility is within 2000 meters.

关 键 词：偏序关系 视差测距 模板匹配 算法融合 

分 类 号：TJ810.371[兵器科学与技术—武器系统与运用工程]