Development of rapid and automated remote sensing methods for ground engagement equipment enabling selective mining

采矿业正通过采用先进技术和采矿作业日益自动化来应对许多挑战（例如，复杂和品位较低的矿床、更严格的环境规定）。通过结合高分辨率多传感器数据采集系统，以及包含基于分析的决策解决方案的自动化数据处理系统，可以显著降低采矿项目的运营成本。然而，由于数据采集和测量能力的限制，选择性采矿系统（使采矿作业具有成本效益和可持续）的采用受到阻碍，从而影响了分析解决方案的稳健性。该提案的主要目标是通过开发基于实时数据采集的基于分析的决策解决方案来实现选择性挖掘。这将包括对高分辨率高光谱和激光雷达数据的采集和分析，这些数据是通过活跃矿井工作面的地面交战设备获得的。这些资料将用于描述矿石的硬度；矿石与废石破碎分析与判别。研究结果将适用于加拿大和世界各地的大量采矿作业。该项目将首先在实验室规模进行，然后进行现场实验（与行业合作伙伴合作），生产分析解决方案的原型，并使用实际数据集进行广泛的测试和验证。

The mining industry is responding to many challenges (e.g., complex and lower grade ore deposits, tighter environmental regulations) by adopting advanced technologies and increasingly automating mining operations. The operating costs of a mining project can be significantly reduced through the combined implementation of a high-resolution multi-sensor data acquisition system, along with an automated, data-processing system that incorporates analytics-based decision-making solutions. However, the adoption of selective mining systems (enabling cost-effective and sustainable mining operations) has been hampered by limitations in data acquisition and measurement capabilities, with an impact in the robustness of analytics solutions.The main goal of this proposal is to enable selective mining by developing analytics-based decision-makingsolutions based on real-time data acquisition. This will include the acquisition and analysis of high-resolutionhyperspectral and LiDAR data, obtained through ground engagement equipment in an active mine face. The information will be used for characterization of ore hardness; rock fragmentation analysis and discrimination of ore from waste rock. The results of the research will be applicable to a large number of mining operations in Canada and across the world. The project will be conducted in lab scale first, followed by field experiments (in cooperation with industry partner), producing prototype implementations of the analytical solutions, and including extensive testing and validation using real datasets.