摘要
针对三维激光测高拟合问题,本文采用BP神经网络进行拟合,该方法可以有效地消除或减弱数据中的系统误差影响,提高计算精度。实际算例证明,在数据无粗差情况下,BP神经网络的内符合与外符合精度优于传统的二次曲面拟合的精度。
BP neural network is used in height fitting to control systematic errors in the 3D laser scanner, and it can improve the precision of fitting. The precisions of inner and outer coincidence of BP neural network are better than those of quadratical form without outliers.
出处
《测绘科学与工程》
2007年第4期22-24,共3页
Geomatics Science and Engineering
关键词
三维激光扫描
BP神经网络
二次曲面拟合
高程拟合
3D laser scanning
BP neural network
the fitting of quadratical form
height fitting