| Résumé | Terrestrial laser scanners (TLS) measure 3D coordinates in a scene by recording the range, the azimuth angle, and elevation angle of discrete points on target surfaces. They are increasingly used in a variety of applications, including manufacturing and civil infrastructure systems. However, the error sources of these instruments are not yet adequately characterized. There is a lack of standardized test procedures [1] and detailed uncertainty budgets for TLS measurements are seldom available. Measuring curved surfaces using TLS has always proved problematic; Lichti et. al [2] describe problems with cylinder measurements. From experiments performed at the laboratories of the Dimensional Metrology Group at NIST, we know that several TLS systems are incapable of obtaining a reliable point cloud from the surface of a spherical target. The changing surface curvature, averaging of the laser spot on the surface, multiple reflections from nearby surfaces and many other factors contribute to the scanned data of the sphere making it appear either smaller (squished) or larger (flared) than the actual sphere. This not only means that the radius of the sphere is incorrectly determined but there is also an error in locating the center of the sphere.In this context, we describe here the challenges involved in measuring a simple geometry, the sphere, using TLS |
|---|
