Assessing the accuracy of land-based mobile laser scanning data.

Abstract

Recent years have been marked by a development of many di erent surveying techniques using laser scanners, among which mobile laser scanning systems (MLS) are promising. Mobile scanning system is a complex system and its development using modular way depends on the needs. The quality of 3D data strongly depends on the exact determination of the position as well as the orientation of the laser scanner sensor during the data acquisition process. Furthermore, the navigational solution is the largest possible source of error in a MLS [8]. Di erent obstructions of GNSS (Global Navigation Satellite Systems) reception appearing in urban or sub-urban areas diminish the accuracy of position and a itude information signi cantly. Errors of position resulting from poor satellite geometry, as well as frequent expected gaps in GNSS signals directly in uence the performance of vehicle-based laser scanning system [1]. This type of disruptions impacts all the geometry of the collected point cloud. The continuity of localization can be compensated by the inertial unit. However, a unique characteristic of INS (Inertial Navigation System) is that it provides a high degree of accuracy only temporary because it gradually degrades over time. This is caused by three main sources of error: initial alignment errors for roll, pitch and yaw angles, inertial sensor errors (e.g. gyroscope drift, accelerometer shift) and computational errors [10]. In this paper a prototype mobile platform designed and developed at the laboratory of MINES ParisTech was described. In the following section a method for assessing the accuracy of mobile laser scanning data is presented. The methodology adopted by the study was to validate the received data, rather than assess the individual components of the system. To ensure real evaluation, a reference data of an accuracy of at least an order of magnitude be er than the studied system has been used.