作　　者： ; ; ;

机构地区： 中国计量学院机电工程学院

出　　处： 《机械设计与研究》 2011年第1期10-12,共3页

摘　　要： 基于SOFM神经网络构建的矩形网格模型可以实现密集散乱点数据自组织压缩,生成期望疏密程度和精度的双有序点列,但该模型存在边缘误差。为减小矩形网格的边缘误差,改进了矩形网格模型的训练模式,提出了3步训练模式。第1步采用整个测量点集,对矩形网格模型中的所有神经元进行整体训练;第2步采用测量点集中的边界点集,对矩形网格模型中的网格边界神经元进行训练;第3步采用边界点集中的角点点集,对矩形网格模型中的网格边界角点神经元进行训练。算例表明,应用该训练模式,可以有效减小矩形网格的边缘误差,矩形网格逼近散乱数据点集的逼近精度得到提高并覆盖数据点集整体分布范围。 An approach based on the self-organizing feature map （SOFM） neural network has been developed to reconstruct rectangular mesh for the dense 3D scattered data. However the approach suffers from boundary problems. A three-steps training method is proposed in order to reduce the boundary error. First all the neurons of the mesh model are trained directly over the 3D scattered digitized points. Next only the boundary neurons of the mesh model undergo training by the bounda^3＇ points of the digitized points. Last only the corner neurons of the mesh model are trained by the corner points of the boundary, points. As a result of applying the proposed training method, the boundary error is greatly reduced and the mesh is drawn toward the sampled object with higher precision comparing with the original SOFM training algorithm. The feasibility of the developed training method was demonstrated by two examples.

关 键 词： 逆向工程 矩形网格 神经网络 边缘误差 散乱数据

领　　域： [自动化与计算机技术] [自动化与计算机技术]