摘要
图像的识别分类总是以被识别物体的某些特征为基础。因此,零件图像识别特征的选取与提取是零件图像识别分类的关键技术,本文对零件图像的几何特征和形状特征进行了研究。设计了零件图像处理程序,提取了区域面积、边界周长、轮廓个数,圆形度参数、区域重心坐标、区域的外接矩形参数6个直观易测的零件图像特征参数。实验结果表明,本文提出的方法是有效的,为下一步零件图像的识别分类奠定了基础。
Image recognition is based on some features of the recognized objects.Therefore, selecting and extracting the recognition features of the part images is the key technology for image recognition.The geometric features and shape features are studied in this paper.Part image processing software is designed.Six visual and easily measured characteristics of the part images,such as region area,boundary perimeter,contour number,circularity,region center of gravity coordinates,region boundary-rectangle coordinates are extracted.The results show that the method mentioned in this paper is valid,which makes preparations for the next part image recognition of assembly line.
出处
《河南科技大学学报（自然科学版）》
CAS
2008年第6期7-9,共3页
Journal of Henan University of Science And Technology:Natural Science
基金
河南省重大科技攻关计划项目(0622020100)
关键词
图像识别
零件图像
几何特征
形状特征
Image recognition
Part images
Geometric features
Shape features
