In-situ block size estimation using 3D terrestrial data

Abstract

For the construction of breakwaters large amounts of armourstone which are produced in dedicated quarries are needed. The better the in-situ block size distribution (IBSD) of these quarries can be determined, the more accurately the quarry production can be predicted. Currently, the IBSD prediction is based on manually acquired scanline data. This method is accompanied with a lot of disadvantages as it is an unsafe, time consuming and subjective method of data gathering in which the amount of gathered data strongly depends on the reach of the surveyor. To determine the IBSD the resulting data is inserted into Wang’s equation method, which uses the mean discontinuity set properties (orientation and principal mean spacing) as input. Taking the disadvantages of the manual scanlining method into account, 3D terrestrial laser scanning is proposed as a suitable alternative to gather the discontinuity data. This thesis investigates the applicability of laser data for IBSD determination. The research is based on the manual and laser data that was gathered during a fieldwork in the Dan-quarry in Benin. Firstly, it is investigated whether it is possible to replace the manually gathered discontinuity data by laser data by simulating a virtual scanline in the point cloud. Secondly, as the laser scanner allows to gather discontinuity data across the entire rockface, an alternative method which allows to take all the obtained discontinuity data directly into account becomes necessary. It is for this purpose that the voxel method in this thesis is developed. This approach has not been used for geotechnical purposes before and is based on the method that is used in the medical world to determine volume data from CT scans. It requires a bounding box to be subdivided into voxels of a predefined size. The position of the centre of each voxel is evaluated with respect to all discontinuity planes that intersect the bounding box. As voxels with the same position with respect to all discontinuity planes can be considered to belong to the same block, the sum of the volumes of these voxels gives the in-situ block size. When comparing the results of the two developed methods with the manually obtained results, it is found that they can represent at best a lower boundary of the IBSD. The main reason for this is that the laser data results in a lot more planes than the manual data. This decreases the IBSD drastically and allows to the conclude that although the orientation data can be easily obtained from the laser data, more effort is needed to obtain the correct spacing and the persistence of the found discontinuities. The effects caused by this are larger in case of the voxel method as all discontinuities over the full height of the face are taken into account instead of only the ones that intersect the virtual scanline. Therefore it is concluded that the virtual scanline method can be used to determine a lower boundary for the IBSD whereas the voxel method, although it has great potential to become an alternative method for IBSD determination from laser scan data, needs more research.