摘要
提出并初步实现了一个面向图纸管理、采用在线图形模糊识别方法和相关反馈技术的手绘图形输入工具。给出了基于笔画速度及曲率变化的特征点检测方法及基本图元的几何识别特征。由于草图具有的模糊特性会造成对用户检索意图理解的偏差,引入模糊集合,定义了基本图元的模糊识别隶属函数。根据模糊集合的隶属度自动地完成草图识别,利用模糊集的距离预测用户的输入意图。在图档管理系统中的应用表明:手绘草图的输入及图元的模糊识别具有很好的实用价值。
A prototype of sketch-based graphic input tool is realized for drawings management, which is based upon online graphic fuzzy recognition and relevante feedback. A method to locate breakpoints of freehand drawings by using curvature and speed information in a stroke is described. After definition of geometrical recognition features of primitive objects and considering the ambiguity of sketch that causes a problem about understanding user's retrieval intention precisely in sketch retrieval, the fuzzy membership function is designed and the membership degree is calculated for recognizing priml"tive objects. The distance between two fuzzy sets can be used to reveal user's latent retrieval requirement. The algorithm is partly applied in the system of engineering drawings management system. The application results show that the method of sketch-based graphic input and fuzzy recognition described here has good value in engineering application.
出处
《南京理工大学学报》
EI
CAS
CSCD
北大核心
2008年第4期458-463,共6页
Journal of Nanjing University of Science and Technology
基金
国防基础科研重大项目
关键词
模糊
图元识别
相关反馈
fuzziness
primitives recognition
relevant feedback