摘要
针对前视红外成像制导中地面待机目标的轮廓提取问题,提出一种新的多目标轮廓提取方法。首先利用相对定位分割技术,将待机目标区域从复杂地面背景中分离出来,然后采用灰度阈值法初步检测目标数量和方位,再通过边缘点聚类分割算法将多目标边缘图像进行分割,最后基于距离差分最小提取目标轮廓。实验结果表明,针对复杂地面场景,该方法可以实现对不同数量、不同类型、不同尺寸地面待机目标的外轮廓提取。基本满足前视红外成像制导对轮廓提取算法实时性好、轮廓完整度高、抗干扰能力强等要求。
According to the difficulties of the contour extraction for the ground standby target in forward looking infra-red(FLIR)imaging guidance,a new contour extraction approach for multi targets is proposed.Firstly,the target areas are separated from the complex ground background with the relative positioning and segmentation technique.Then,the gray threshold method is used to detect the number and orientation of targets.Next,the multi targets edge image is segmented by the edge points clustering algorithm.Finally,the target contour is extracted based on the minimum distance difference.In view of the complex ground scene,the experiments demonstrate that the method can realize the contour extraction of different quantity,different types,and different sizes of ground standby target.It can satisfy FLIR imaging guidance for contour extraction algorithm requirements of higher contour integrity and good real-time,as well as strong anti-jamming and stabilization.
作者
陈世伟
张胜修
杨小冈
齐乃新
CHEN Shiwei ZHANG Shengxiu YANG Xiaogang QI Naixin(DeparLmenL of Control Engineering , RockeC Force University of Engineering , Xi’an 710025, China)
出处
《系统工程与电子技术》
EI
CSCD
北大核心
2017年第7期1647-1652,共6页
Systems Engineering and Electronics
基金
国家自然科学基金(61401470)资助课题
关键词
相对定位分割
边缘聚类
距离差分
轮廓提取
relative positioning a n d segmentation
edge clustering
distance difference
contour extraction