摘要
为了从复杂环境中准确识别出火炬,选取火炬的颜色特征与几何形状特征作为识别火炬的依据,将两者组合成有意义的图像特征,采用HIS颜色空间,中值滤波,Sobel边缘提取,Hough变换等方法,同时提出了一种有效的空穴填补法,实现了识别火炬的目的,它是进一步进行机器人视觉伺服的基础.
In order to recognize the torch from complex environment, this article chooses the color features and geometry shape features of the torch as dependence for the torch＇s recognition. These two features are then composed into significative image features. Adopting the HIS color space, median filter, Sobel edge distill, Hough transform and using an effective method to fill the cavity, the recognition of the torch can be realized. It is the base of robot＇s vision servo.
出处
《东南大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2005年第A02期151-154,共4页
Journal of Southeast University：Natural Science Edition
基金
国家高技术研究发展计划(863计划)资助项目(2003AA420010-04)
关键词
火炬
颜色特征
形状特征
识别
torch
color features
shape features
recognition
