多特征的UAV快速目标识别算法仿真  

Fast UAV Target Recognition Algorithm Based on Multiple Features

作　　者：李大健[1] 刘慧霞[1] 张清[2] 席庆彪[1,2] 郭海帆[3] 

机构地区：[1]西北工业大学第365研究所,西安710065 [2]西北工业大学自动化学院,西安710072 [3]西南电子设备研究所,成都610036

出　　处：《火力与指挥控制》2012年第3期35-38,共4页Fire Control & Command Control

基　　金：国家自然科学基金资助项目(61074155)

摘　　要：针对UAV(Unmanned Aerial Vehicle)侦察图像快速目标识别问题,重点展开基于多特征的UAV快速目标识别算法的仿真研究。算法结合图像的不变矩特征和SIFT特征,首先用不变矩特征构造适应度函数并利用遗传算法的全局搜索能力,在侦察图像中进行搜索,快速提取出可能包含目标的感兴趣区域(ROI,Region of Interest);然后采用尺度不变特征变换算法(SIFT,Scale Invariant Feature Transform)在ROI区域中进行匹配识别,从而确定目标的精确位置。仿真结果表明:算法具有较强的鲁棒性,能有效地识别飞机目标并显著减少识别时间,为UAV系统提供了一种近实时的目标识别方法。To settle the problem of identifing the target in the UAV（Unmanned Aerial Vehicle） reconnaissance image in real time,a UAV fast target identify algorithm is proposed,which is based on invariant moments and the SIFT（Scale Invariant Feature Transform） features of the image.Firstly,similarity function is designed with invariant moments,and a fitness function of GA（genetic algorithm） is taken to search the image globally,the region of interest（ROI）,which may contain the target,can be extracted quickly.Secondly the location of the object can be determined accurately by SIFT algorithm in the ROI region.Simulation results show that the method is of strong robustness and can effectively identify the plane targets in airport with the ability to reduce the recognition time significantly.

关 键 词：无人机 目标识别 不变矩 遗传算法 SIFT 

分 类 号：TP753[自动化与计算机技术—检测技术与自动化装置]