Blast Hole Seeking and Dipping -- The Navigation and Perception Framework in a Mine Site Inspection Robot

Abstract: In open-pit mining, holes are drilled into the surface of the excavation site and detonated with explosives to facilitate digging. These blast holes need to be inspected internally for investigation of downhole material types and properties. Knowing these properties can lead to significant savings in material handling costs in downstream processes. Manual hole inspection is slow and expensive, with major limitations in revealing the geometric and geological properties of the holes and their contents. This has been the motivation for the development of our autonomous mine-site inspection robot - "DIPPeR". In this paper, the automation aspect of the project is explained. We present a robust blast hole seeking and detection framework that enables target-based navigation and accurate down-hole sensor positioning. The pipeline first processes point-cloud data collected by the on-board LiDAR sensors, extracting the cone-shaped volume of drill-waste above the ground. By projecting the 3D cone points into a virtual depth image, segmentation is achieved in the 2D domain, yielding a circular hole at the image centre and a collared cone face. We then identify the hole centre using a robust detection module while suppressing non-maximum candidates, ensuring precise sensor placement for down-hole inspection and avoiding collisions with the cavity wall. To enable autonomous hole-seeking, the pipeline automatically adjusts its projection parameters during robot navigation to account for variations in point sparsity and hole opening size, ensuring a consistent hole appearance in 2D images. This allows continuous tracking of the target hole as the robot approaches the goal point. We demonstrate the effectiveness of our navigation and perception system in both high-fidelity simulation environments and on-site field tests. A demonstration video is available at "https://www.youtube.com/watch?v=fRNbcBcaSqE".