三维建模道路地形三维可视化分析

3D Visualized Analysis of Road Terrain by 3D Modeling

摘要:

研究一种可以适用于道路地形三维建模的DEM可视化模型精度优化方案。通过对道路地形三维DEM模型的现状和问题分析，发现在不改变当前基于无人机航测遥感硬件设备的基础上，通过充分优化后置数据平差算法，引入带有数据卷积功能的多列神经网络机器学习技术，可以将基于多种测量技术的混合测量结果进行综合平差的数据精度提升2个数量级。在仿真试验中表明，在可视化结果中的长度截取误差、以及基于上述长度截取过程获得的工程量计算误差、材料费计算误差、工程总造价误差等方面，优化算法分别将数据精度提升90.7%、86.0%、87.6%、79.3%。最终认为，虽然该方案尚未被工程测量相关国家标准接纳和支持，但单纯从数据角度分析，该方案在公路规划设计领域的工程测量中表现出一定的积极意义，具有一定的推广价值。

Abstract:

A precision optimization scheme of DEM visualized model able to be suitable for the 3D modeling of road terrain is studied. By analyzing the current situation and the existing problems of road terrain 3D DEM model, it is found that the mixed measured result based on the multiple measuring techniques can upgrade 2 magnitude orders of the comprehensive balancing data accuracy on the basis of not changing the UAV air survey remote sensing hardware equipment, by fully optimizing the balancing arithmetic of the rear data and introducing the multiseriate neural network machine learning technique with the data convolution function. The simulation test shows that the optimization algorithm will upgrade the data accuracies to 90.7%, 86.0%, 87.6% and 79.3% separately in the aspects of the length intercepting error in the visual result, the engineering quantity calculation error achieved in the above length interception, the material expense calculation error and the total engineering cost error. Finally it is considered that this scheme is stilled not accepted and supported by the related national standards of engineering survey, but this scheme has a certain positive significance and has a certain popularizing value in the engineering measurement of highway planning design field from a purely statistical point of view.