摘要
本文提出的定性描述图表示方法能在无模型指导的情况下有效地实现遮挡物体的分离,并将识别过程转化为图或子图同构的问题,还提出了用“提纯”与“搜索”相结合的递归过程来判定图或子图同构。实验证明了该方法的有效性和稳定性。
In this paper, a new 3D object qualitative representation called Qualitative Description Graph (QDG) is presented. This symbolic rep esentation can separate effectively the objects with occlusion and simplify greatly the identi ication of objects in complex scenes. To improve the recognition efficieney, s recursive scheme integrating 'purifying' with 'searching' is proposed. Several experiments show that this ap(?)roach is effective, robust and efficient.
出处
《自动化学报》
EI
CSCD
北大核心
1993年第1期63-70,共8页
Acta Automatica Sinica
基金
高等学校博士学科点专项科研基金
关键词
定性描述图
定性识别
模式识别
the representation of 3D objects
Qualitative Description Graph
qualitative identification
graph or subgraph isomorpiism.
