摘要
三维信息处理中,距离是重要的特征信息之一。本文提出了一种基于遗传算法的三维表面距离测量方法 ,根据特定条件在模型表面上逼近两点间最近距离。首先给出和模型相交的平面束数学描述;其次选择特定平面作为种子,以三维表面距离构建适应度函数,利用遗传算法在平面束中找出最优平面;最后结合局部性原理,降低算法复杂度。通过实际三维模型的实验,取得了99.71%的正确率,验证了方法的有效性。
In three-dimensional information processing, the surface distance is an important characteristic parameter. This paper brings a measurement for surface distance along 3D model based on genetic algorithm by special qualifications. Fristly, to get the plane’s formula which intersected with 3D modal. Secondly, to select the special planes as seeds for genetic algorithm, use surface distance to build adapt function, use genetic algorithm to optimize the plane parameters. Finally, to reduce the complexity of the method by using locality principle. The experimental results on practical 3D model show a 99.71% veracity, proving the validity of this method.
出处
《中国体视学与图像分析》
2010年第2期121-126,共6页
Chinese Journal of Stereology and Image Analysis
关键词
遗传算法
表面距离
三维模型
局部性原理
genetic algorithm
surface distance
3D modal
localized theory
作者简介
林增刚（1977-），男，汉族，陕西西安人。研究方向：三维信息处理、计算机视觉与模式识别等。
