中国激光, 2012, 39 (s2): s209005, 网络出版: 2012-12-24
车载三维信息采集系统的室内直线特征提取方法
Extraction Approach Based on Line Feature in Indoor Environment for Vehicle-Borne Three-Dimensional Acquisition System
机器视觉 车载三维信息采集 线性拟合 角度阈值 航位推算 同步定位及地图创建 machine vision vehicle-borne three-dimensional acquisition linear fitting angle threshold dead reckoning simultaneous localization and mapping
摘要
基于航位推算方法定位的车载测绘系统(VBMMS)可以有效地解决使用全球导航卫星系统(GNNS)定位带来的系统失锁问题。但当系统采集空间三维信息时,由于航位推算方法的误差积累,会造成较大的系统定位误差,严重影响最终的测绘精度。针对此问题,提出了利用同步定位及地图创建(SLAM)的算法对系统进行位姿矫正。从激光点云中提取出准确的特征信息是实现系统同步定位及地图创建的前提,结合室内结构化环境的特点,提出一种基于角度阈值的提取直线特征的算法。将算法应用到自主研发的车载移动测图系统中,实现了快速准确的直线特征提取,实验结果证明了该算法的有效性。
Abstract
The location of vehicle-borne mobile mapping system (VBMMS) base on dead reckoning can effectively resolve the problem of the global navigation satellite system (GNSS) outages. Whereas, due to the accumulation of dead reckoning error, the VBMMS may result in large systematic positioning errors and serious impact the final mapping accuracy when acquiring the space three-dimensional (3D) information. In this paper, the simultaneous localization and mapping (SLAM) is proposed to correct the location of VBMMS. The accurate feature extraction from the laser cloud points is the prerequisite for the robotic simultaneous localization and mapping. Considering the structure of indoor environment, a line feature extracting method based on the angle threshold is proposed. The novel method is implemented in the self-developed vehicle-borne 3D acquisition system and the experimental results demonstrate the validity of this method.
陈春朋, 胡少兴, 张爱武. 车载三维信息采集系统的室内直线特征提取方法[J]. 中国激光, 2012, 39(s2): s209005. Chen Chunpeng, Hu Shaoxing, Zhang Aiwu. Extraction Approach Based on Line Feature in Indoor Environment for Vehicle-Borne Three-Dimensional Acquisition System[J]. Chinese Journal of Lasers, 2012, 39(s2): s209005.