Accurate 3D model reconstruction of cities and Earth's surface has become increasingly more accessable and undemanding due to the development of LiDAR technology. Even though the data colleted from LiDAR scanning systems is very accurate, precise and flexible, it has its shortcomings. Areas of objects that are impossible to scan from airborne sensors contain missing point data, which are regared as holes in the proccess of object reconstruction. In this thesis we propose an algorithm for detection and filling missing points in models of building walls and mountainous surface areas. Our solution assumes filling empty spaces of point clouds with newly created points. Adding new points to building walls is made based on topographical data, that is used for determining building borders and coordinates for point insertion. Adding new points to mountainous surface areas is made possible with the help of a simplified point model, that is constructed with plane projection. After determining coordinates of new points on a plane, we calculate point heights based on interpolation of neighbouring points.
|
