Ore characterization using dual energy X-ray transmission system for the development of sensor-based sorting application, Y.R. Zhang, M. Holuzko, and J. Alexander

The use of dual-energy X-ray transmission (DE-XRT) technology in the mining industry has been developing rapidly in the past decade. In many projects, DE-XRT images are analysed to classify ore and waste in sulphide deposits where the sulphides and non-sulphides have distinct characteristics in density. In the current practice of DE-XRT image processing, every pixel in the image gets classified into two categories, depending on the intensity of the pixel from both low and high energy X-ray response. If the pixel’s intensity is above the threshold curve generated in the high-energy and lowenergy scatter plot, it is classified as “higher density” material, and the remaining pixels are classified as “lower density” material. This paper elaborates on an alternative approach to DE-XRT image processing for in-depth characterisation of samples from a lead-zinc sulphide deposit. The detailed characterisation is achieved by providing density distribution of individual rock using DE-XRT laboratory sorting unit (COMEX) located at the University of British Columbia. Instead of classifying the X-ray images into two different classes from one threshold curve, multiple density classes are created between two threshold curves to generate the density distribution. By manipulating the two threshold curves, we can analyse images of samples obtained at different resolutions of density distribution. If the two threshold curves become infinitely close to each other, it will create a single-threshold scenario similar to the current methods being used in the industry. This method provides the ability to change the threshold curves, hence providing more flexibility on material characterisation based on the needs of the operation. Furthermore, the DE-XRT images from individual rock specimens are compared to the test results from hand-held X-ray fluorescence (XRF) analyser as well as assays of various elements for the same set of samples. The data analysis will provide information regarding ore characteristics and also the theoretical performance of the sensor-based sorting. Keywords: Dual-energy X-ray Transmission (DE-XRT), X-ray fluorescence (XRF), ore characterisation, density distribution