基于分级轮廓分形特征的舰船目标识别方法  

An Approach to Naval Vessel Target Recognition Based on Its Hierarchical Contour Fractal Features

作　　者：汪国有[1] 李捷[1] 张天序[1] 

机构地区：[1]华中理工大学图像识别与人工智能研究所图像信息处理与智能控制教育部开放实验室

出　　处：《华中理工大学学报》1999年第12期1-3,共3页Journal of Huazhong University of Science and Technology

摘　　要：采用一种共线性检查算法去除舰船轮廓底边，利用轮廓曲率的分级分形特征建立前视成像末制导过程中舰船目标的分级模型数据库，提出了一种基于分级分形特征的大型舰船目标识别方法．对舰船目标的实验结果表明了该识别方法的有效性和正确性．An algorithm for collinearity detection which can be used to eliminate the hemline of naval vessel contour is proposed. Then, by utilizing the curvature based hierarchical fractal features of a target contour not containing its hemline, a hierarchical model database for the vessel target during forward looking imaging terminal guidance is developed. A hierarchical fractal based method for large naval vessel targets recognition is also proposed. The results of experiments with two types of naval vessel verify the validity and correctness of the method proposed.

关 键 词：分形 轮廓特征 目标识别 舰船 目标检测 

分 类 号：U675.7[交通运输工程—船舶及航道工程] TP391.41[交通运输工程—船舶与海洋工程]