Automated characterisation of the dump materials

The determination of shear strength parameters for the spoil material plays a crucial role in the assessment of a dump’s stability; nevertheless, their evaluation necessitates laborious laboratory experiments. In order to address this issue, Simmons and McManus (2004) formulated a framework that serves as a foundation for categorising spoils according to their distinctive shear strength parameters based on observable and tactile characteristics. This framework, previously identified as the ‘BHP Mitsubishi Alliance Coal (BMAC) Spoil Shear Strength Framework’, is now recognised as the ‘CoalSpoil’ framework. The spoil characterisation provided by the CoalSpoil framework offers benefits but is not without limitations. The examination of these attributes in the field is a time-consuming task that demands expertise. Subjectivity arises in the spoil characterisation when attributes straddle the line between two categories, such as delineating matrix-supported and framework-supported fabric structures. To enhance consistency in the spoil characterisation process, minimising human intervention is crucial to mitigate bias, classification errors, and safety risks. This study seeks to achieve this aim by introducing image-assisted and automated methods for spoil characterisation.