A Portable Device for Mine Face Rock Fragmentation Analysis

Blast engineers have been taking advantage of image-based rock fragmentation analysis methods for the past 3 decades to estimate rock size distributions. These methods utilize a wide range of image segmentation techniques, from manual to fully-automated, to delineate the edges of rocks in images. Rock delineation results are then converted to physical rock sizes (and percentage passing curves) by scaling the image pixels to correspond to physical measurements. The size scaling process is critical for acquiring physical measurements, but specific to the limitations of different applications. This article starts by presenting a shovel bucket-based fragmentation analysis system, where scaling is done using the known width of the shovel bucket within the image. Next, it describes methods that use pictures captured by digital cameras with scaling reference objects in the view. Finally, the article presents a method that uses three-dimensional imaging to capture images without the need for scaling reference objects. A portable device which utilizes a three-dimensional imaging method is proposed, and compared with the other methods. The proposed solution combines the generated three-dimensional points with the delineated rock boundaries in the image to reconstruct the surface, and generate accurate rock size estimates and percentage passing rock size distributions. System operation is described and its accuracy and performance are discussed for various scenarios. It is shown that since the device does not require a physical object for scaling reference, it is free from the limitations of other methods, and can be applied for automatic rock size measurement from various sources such as mining shovels, loaders, crushers, stockpiles, haul truck beds, and conveyor belts.