@phdthesis{oai:air.repo.nii.ac.jp:00005713,
 author = {TUNGOL, ZEDRICK  PAUL LIZARD},
 month = {Mar},
 note = {In any mining operation, there are a number of factors that can affect the efficiency of the day-to-day resource extraction process across all of its stages from mine (drilling, blasting, haulage) to mill (mineral processing). As such, it is up to the management, when possible, to monitor these various factors and act accordingly by making modifications to mine planning as well as tweaking its execution. One of such factors, specifically in mining operations that employ explosives and mineral processing, is the fragmentation size distribution of rock after it has been blasted.
Scaling is a critical component of fragmentation size distribution measurement using photogrammetry as it will directly determine the accuracy of the size estimation. In creating a 3D model, extrinsic data such as ground truths are needed to create a properly-scaled reconstruction of the scene. There are several methods that are used to resolve scale in photogrammetry. Most of these methods have the same basic idea in that once the exact distance between at least two different points in a scene is known, a scale factor can be applied to the 3D model. One way to do this is to include an object of known length such as scale bars in the scene. In larger applications such as aerial mapping, GCPs (Ground Control Points) are used, which are marked points of known absolute or relative coordinates.
The study aims to create a system for creating a scaled 3D model without the use of ground truth data such as GCPs (Ground Control Points) for the purpose of improving fragmentation size distribution measurement using positional data such as GNSS (Global Navigation Satellite System)-aided photogrammetry. To achieve this, the study firstly aimed to 1) investigate the effect of camera positional data and constraints on 3D model scaling accuracy; then 2) simultaneously collect image and positional data (e.g., GNSS) from scenes that are to be reconstructed using photogrammetry; then 3) use the positional data in the photogrammetry workflow to scale the resulting 3D model; then 4) observe the effect of increasing the number of datapoints (image + positional data) on the scaling accuracy of the generated 3D model and finally 5) Determine the most effective configuration in data taking and data processing to achieve acceptable scale with the least number of datapoints needed.
A preliminary experiment that was done show results that constraining camera positions to locations, relative or otherwise improves the accuracy of the generated 3D model. With this fundamental idea in mind, the study moved on to larger scale experiments that involved the actual use of GNSS positional data in conjunction with image data in 3D photogrammetry. In these further experiments, results show that the scale error decreases when more images from the same dataset are used. In conclusion two observations have been drawn from the study: firstly, constraining cameras to accurate positions in SfM will result in a properly scaled 3D model; secondly, increasing the number of georeferenced images in SfM will incrementally improve the scaling error of the reconstruction. These results lend credence to the possibility of improving the scaling aspect of 3D fragmentation measurement systems without the use of GCP or manual scales, specifically in surface mines where GNSS data is generally readily available. This shows that monitoring the fragmentation distribution can potentially be performed using just a camera and a GNSS-enabled devices, such as smartphones.},
 school = {秋田大学},
 title = {Automatic Scaling in Structure from Motion Photogrammetry for 3D Fragmentation Size Distribution Measurement},
 year = {2022}
}