Abstract
Crusher obstructions caused by broken lip shrouds and crusher delays caused by oversized material result in considerable operational downtime. These events are especially common in hard rock mines and jeopardize employee safety. This paper presents a proposed technology to detect boulders and broken lip shrouds through the application of 3D imaging and deep learning techniques. By continually monitoring the shovel bucket, the proposed technology alerts the operator when a lip shroud goes missing or a boulder is detected in the bucket so that dispatch can divert the contaminated material before it reaches the crusher.
The proposed technology uses a high-resolution 3D colour camera mounted above the shovel bucket and an in-cab monitor to delivery timely notifications to both the operator and mine management. An embedded, GPU-enabled computer in the shovel cab collects and processes the rich 3D data before pushing the results to a secure cloud platform for further analysis. It is critical that most of the collected data is processed locally to accommodate mines with limited network connectivity and bandwidth. Built on two decades of GET detection experience, the proposed technology will be supported and maintained by highly trained teams in every corner of the world.