UCAV环境感知中路标选取及定位方法  

Method to select and position landmark in UCAV environment perception

作　　者：吴德伟[1] 周阳[1] 杜佳[1] 李伟龙[1] 

机构地区：[1]空军工程大学信息与导航学院,陕西西安710077

出　　处：《系统工程与电子技术》2014年第10期2048-2052,共5页Systems Engineering and Electronics

基　　金：国家自然科学基金(61273048)资助课题

摘　　要：为实现无人作战飞机(unmanned combat aerial vehicles,UCAV)认知导航环境感知中路标的选取和位置求解,提出一种基于增量多层判别回归(incremental hierarchical discriminant regression,IHDR)树的路标选取和定位方法。利用快速鲁棒性特征(speeded up robust feature,SURF)算法提取的特征点描述矢量及其在地理空间中的位置组合为路标信息,构造IHDR树,实现已感知信息的存储记忆。UCAV进行环境感知时,若路标信息已知,以已知路标描述矢量对IHDR树检索,通过已知路标位置与检索输出位置之间误差的大小进行路标选取。若路标信息未知,以SURF算法提取环境中路标的描述矢量,并对IHDR树检索,通过比值法对输出路标进行提纯,以提纯后路标的位置作为对应路标的定位结果输出。仿真结果表明,该方法可有效选取设定位置误差内的路标,低噪声干扰下路标的定位具有高正确率。In order to realize landmarks＇ selecting and location solving during the environment perception in unmanned combat aerial vehicles （UCAV） cognitive navigation, a method based on the incremental hierarchical discriminant regression （IHDR） tree is proposed to select and position landmarks. The IHDR tree is built by the landmarks＇ information which is combined by the description vector of feature points got by the speeded up robust feature （SURF） algorithm and its position in geographic space, so the perception information can be saved and remembered. If the landmarks＇ information is known in the UCAV environment perception, the in formation is used to retrieve the IHDR tree, then the landmarks can be selected by analyzing the error＇s size between the known location and the output location. If the landmarks＇ information is unknown, the IHDR tree is retrieved by the description vector of landmarks got by the SURF algorithm, then the output landmarks are purified by the ratio method, and the location of the purified landmarks is used as the positioning results. The si- mulation results show that the proposed method can select landmarks within the scope of the set location error effectively, and the positioning of landmarks can obtain high accuracy in low noise.

关 键 词：环境感知 增量多层判别回归树 快速鲁棒性特征 路标选取 路标定位 

分 类 号：V279[航空宇航科学与技术—飞行器设计]