Mätosäkerhet vid deformationsmätning med bärbar laserskanner

Sammanfattning: Mobile wearable laser scanning systems, also called personal laser scanning systems (PLS), have the potential to combine the strengths of mobile laser scanning (MLS) with usability indoors and in harsh terrain. The mobility makes surveying possible where terrestrial laser scanning (TLS) is difficult or not so resource-efficient to use. This may render PLS a both suitable and favorable alternative for certain deformation surveying. However, what measurement uncertainties that is acheivable and so how small deformations that is measureable, is yet to be clarified. The purpose of this study is therefore to investigate these subjects using a rucksack mounted PLS. A literature study is applied to outline the fundamentals of deformation surveying and thereby possible ways of controlling measurement uncertainties. Ways of georeferencing point clouds are described including the new technology Simultaneous Localization and Mapping (SLAM). Concluding is an overview of earlier work on measurement uncertainties regarding MLS, PLS and SLAM focusing on methods and results.A rucksack mounted PLS (Leica Pegasus: Backpack) is used to survey simulated deformations both out- and indoors as well as with and without control points. Rotational, horizontal and vertical displacements are tested (at an interval of 5° between 5° and 20° and 0.050 m between 0.050 m and 0.200 m, respectively) together with a nonmoving object. By optimizing the trajectory with SLAM and analyzing geometrical planes fitted into the point clouds, conclusions can be drawn regarding how small deformations that is measureable and the variability of the surveys. The results indicate possibilities to detect rotations at 5° outdoors, but the substantially fluctuating measurement uncertainties indoors show that rotations at 20° or smaller are impossible to detect. Horizontal and vertical deformations at 0.050 m can be surveyed outdoors, but the measurement uncertainties indoors exceed even the largest tested deformations for all but the vertical deformations with control points. These may be surveyed at 0,050 m. The analyzis of the nonmoving object reveals a combined 3D-uncertainty of 0.001 m outdoors, 1.49 m indoors without control points and 0.490 m indoors with control points. The results show that several factors have to be minded but also that there are possible areas of use outdoors within catastrophe analyzis, geomorphological changes in landforms, forestry and urban change detection. Indoors the results may improve by use of more advanced SLAM-algorithms along with control points, but the measurement uncertainties still imply that the main application is rough change detection.