As safety, accuracy and overall system optimisation requirements evolve, the world is rapidly moving into a computer age where equipment automation makes the most sense in every industry. The use of drones in mining environments is one way in which data pertaining to the state of a site can be remotely collected. Though capable of visualising what a miner would be able to see without the need for their physical presence, most production drones are incapable of classifying rocks or minerals with their traditional visible light camera sensors. To counter this, this paper proposes the employment of a multispectral image capturing camera mounted on a drone. Depth possessing imagery data from within the visible near-infrared range (VNIR) was captured via the Unmanned Automatic Vehicle (UAV) drone with multispectral image gathering capabilities at different flight elevations. This was an attempt to remotely identify magnetite iron sands via the UAV drone specialised in collecting five band spectral information at a minimum accuracy of +/–16 nm. Having accumulated the data, visual imagery is fed into a Spectral Angle Mapper (SAM) and a machine learning (ML) algorithm specialised in classifying spectral imagery data. This algorithm is trained and tested in order to classify the magnetite deposits, hence deducing the amount of iron present from within each image corresponding with the site capture point. With the algorithm, a high global classification accuracy of 85.7 per cent was attained. Thus, deeming the system highly viable in mining environments that are constantly aiming for risk potential elimination, by increasing the physical distance between miners and the site. This paper, therefore, confirms the initial hypothesis aimed at achieving overall system optimisation within a mine site by means of the integrated system composed of a UAV drone, multispectral imaging, SAM analysis and ML.