面向光电图像目标检测跟踪的距离信息处理  

Distance information processing method oriented target detection and tracking in photoelectric image

作　　者：冉欢欢[1,2,3] 游安清[1,2] 霍义华[3] 黄自力[3] 

机构地区：[1]中国工程物理研究院应用电子学研究所,四川绵阳621999 [2]中国工程物理研究院高功率微波技术国防科技重点实验室,四川绵阳621999 [3]西南技术物理研究所控制与信息技术部,四川成都610041

出　　处：《太赫兹科学与电子信息学报》2015年第5期769-774,共6页Journal of Terahertz Science and Electronic Information Technology

基　　金：双动态目标跟踪技术研究资助项目(2014HPM-09)

摘　　要：针对复杂背景下,光电图像目标检测跟踪的难题,分析了基于光电图像目标检测跟踪的基本模型。依据距离信息的空间维度和包含的信息量,提出将距离信息划分为标量距离信息、矢量距离信息以及阵列距离信息的分类方法;系统地研究了面向光电图像目标检测跟踪的距离信息处理方法。通过分析研究得出光电图像目标尺寸与标量距离信息的关系;提出将图像高层特征匹配模型从平面扩展到三维立体空间的图像匹配方法;总结了阵列距离信息在目标检测识别中的应用价值。通过面向光电目标检测跟踪的距离信息处理方法的分析研究,为融合距离信息的光电图像目标尺度分析、图像匹配以及目标种类辨识提供了理论依据。A basic model of target detection and tracking of photoelectric image has been analyzed to solve the problems existed in target detection and tracking of photoelectric image in complex background. The distance information is divided into three categories including scalar distance information, vector distance information and array distance information according to spatial dimensions and the amount of contained information. The distance information processing method oriented target detection and tracking in optical information has been researched systematically. The relationship between the target scale in photoelectric image and the scalar distance information is obtained. The image matching method using the high-level feature matching model in photoelectric image is expanded from 2D to 3D space. The applications of array distance information in target detection and recognition are summarized. This work can provide a theoretical basis for target scale analysis, image matching and identification of target species in photoelectric images.

关 键 词：光电图像 距离 信息融合 目标跟踪 目标检测 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]