Heap Leach Pad Surface Moisture Monitoring Using Drone-Based Aerial Images and Convolutional Neural Networks: A Case Study at the El Gallo Mine, Mexico

Abstract

An efficient metal recovery in heap leach operations relies on uniform distribution of leaching reagent solution over the heap leach pad surface. However, the current practices for heap leach pad (HLP) surface moisture monitoring often rely on manual inspection, which is labor-intensive, time-consuming, discontinuous, and intermittent. In order to complement the manual monitoring process and reduce the frequency of exposing technical manpower to the hazardous leaching reagent (e.g., dilute cyanide solution in gold leaching), this manuscript describes a case study of implementing an HLP surface moisture monitoring method based on drone-based aerial images and convolutional neural networks (CNNs). Field data collection was conducted on a gold HLP at the El Gallo mine, Mexico. A commercially available hexa-copter drone was equipped with one visible-light (RGB) camera and one thermal infrared sensor to acquire RGB and thermal images from the HLP surface. The collected data had high spatial and temporal resolutions. The high-quality aerial images were used to generate surface moisture maps of the HLP based on two CNN approaches. The generated maps provide direct visualization of the different moisture zones across the HLP surface, and such information can be used to detect potential operational issues related to distribution of reagent solution and to facilitate timely decision making in heap leach operations.