Application of computer vision to automatic selective cutting with a roadheader in a potash mine

Automation of mining operations involves the use of sensing, remote monitoring and control systems in order to confront a variety of situations and environmental conditions. The basic requirement of the overall economy of the mine sometimes requires that selective cutting be performed in order to separate rich ore from waste at the cutting stage. Basically, the problems to be solved are those of modelling an uncontrolled, changing mine environment and programming the machine to cut a pattern accordingly. The authors indicate how color image segmentation, automatic image classification, camera calibration and 3D scene perception can cooperate to solve such a complex problem as selective cutting.<>