基于多特征直方图的红外图像目标跟踪  

Target Tracking in Infrared Image Based on Multi-Feature Histogram Algorithm

作　　者：王杰[1] 孙艳丽[1] 周伟[2] 

机构地区：[1]海军航空工程学院信息基础实验部,山东烟台264001 [2]海军航空工程学院融合研究所,山东烟台264001

出　　处：《兵器装备工程学报》2017年第8期103-106,共4页Journal of Ordnance Equipment Engineering

摘　　要：根据红外图像的特点,采用单一颜色难以描述复杂的跟踪环境,尤其是目标周围有相似亮度的背景干扰时,会造成跟踪失败。针对该问题,提出在MS跟踪理论框架下将颜色特征和梯度特征联合起来描述目标。实验表明:该方法能准确地跟踪海上红外运动目标。According to the characteristics of the infrared image,using single color description is difficult to adapt to the complex tracking environment. Especially that the brightness around the target is similar to background,it will cause the tracking failure. A method combined color features and gradient features to describe the target under the MS track theoretical framework is proposed. The experimental results show the algorithm can track infrared moving target at sea.

关 键 词：红外图像 多特征直方图 目标跟踪 MEANSHIFT算法 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]