摘要
本文阐述了一种简单的线条特征法,该算法是通过边缘检测,细化、符号化三过程实现的,而符号化过程是借助于边缘检测的方向码通过扫描、连接、补缺三步骤完成的,根据所选择的控制参数,可以形成基于最短线条长度的各种层次的符号化描述,灵活性强,速度快,效率高,特别适用于双目立体视觉的研究中,用来获得线条基元。
A simple algorithm for line-segment acquisition is illustrated in this paper.It is composed of threeprocedures of edge detection,thinning and symbolization,in which the time-segments of an image aresimply produced by means of direction code of edge points through three steps of scanning,linking andmending.Different levels of image representation,which,is especially suitable for line-segment-basedbinocular stereo vision system as its primitives,can be achieved accordihg to the paramters chosen.It is afast effective algorithm being more flexible and applicable.
出处
《机器人》
EI
CSCD
北大核心
1989年第4期32-35,39,共5页
Robot
关键词
机器视觉
边缘检测
特征抽取
细化
robot vision
edge detection
thinning
feature acquisition