摘要
本文对多面体的空间定位及其识别作了较为深入的研究,提出了一种基于物体几何模型的立体图象柔性匹配及识别算法。依据物体的局部特征和关系图,采用二维半的方法由底向上地恢复物体的深度信息,并建立对物体几何模型的描述,据此,达到了对物体准确识别。
In this paper, a stereo image matching and scene recognition algorithm for
polyhedrons is preposed. According to the local features and their relative graphs of
three dimensional object, the depth of the object is recovered by 2.5 dimensional
method and the description of geometric model of the object is built. Experim-
ental results show that the three dimensional object can be exactly recogized.
出处
《西安交通大学学报》
EI
CAS
CSCD
北大核心
1991年第1期55-62,共8页
Journal of Xi'an Jiaotong University
基金
国家自然科学基金
关键词
计算机视觉
几何模型
定位
物体
model
stereo match
spatial location
recognition
