Abstract
It is expected that the point cloud data measured by the ground-based laser scanners and so on will be applied to result management of construction work. However, the point cloud data acquired during construction are mixed with points unnecessary for result management, including those of construction equipment and trees. Therefore, it is necessary to remove these points and extract ground surfaces. Moreover, there are cases where the ground surfaces are not measured due to occlusion with construction machines and so on. Such cases require some processing such as making complements with point cloud data measured at different timing. The existing methods for extracting the ground surfaces from point cloud data are approaches designed for airborne or ground-based laser scanners, and are suitable for measurement results of urban and forest areas. However, application of these methods to earthwork for roads or rivers may cause problems that they do not allow correct extraction of slopes as well as unlevelled ground surfaces, and that they are not capable of complementing the loss of the ground surfaces. In this work, we develop a technique for generating ground surfaces from the point cloud data acquired by measurement at the same spot, and evaluate its effectiveness.
