飞机图像识别中离散点曲率计算方法研究  

Research of Discrete-points Curvature Calculating Methods for Image Recognition of Airplane

作　　者：朱延娟[1] 倪周松 

机构地区：[1]同济大学航空航天与力学学院,上海200092

出　　处：《机电一体化》2016年第6期13-18,共6页Mechatronics

摘　　要：针对飞机图像识别中离散点曲率计算,阐述了基于弦长、卷积积分和自适应模糊直线段3种离散点曲率计算方法。在分析和研究每一种方法的算法原理和计算过程的基础上,对飞机图像进行离散点曲率计算并标记特征点,分析各离散点曲率计算方法的特点与适用范围,为物体识别与检测等提供可参考的依据。Focusing on discrete-points curvature calculating for image recognition of airplane, this article states three calculating methods of discrete-points curvature which are approaches based on chord length, convolution inte- gral and self-adaptive fuzzy segment. It describes the algorithm principles and calculation processes of every method, computes discrete-points curvature of an airplane image, and locates feature points. Finally, it analyses their charac- teristics and application scopes, which provides reference basis for object recognition and detection.

关 键 词：曲率计算 离散点 自适应 图像识别 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]