Challenge the practice is addressing: X-MINE Task 3.2 develops an algorithm for mineral sorting based on data fusion map and configurable user parameters. The first version is based on conventional image processing tools such as thresholding, background removal, edge detection and labelling, and it will be tested with the reference samples collected from the participating mines. Samples will be analysed and labelled by experts, and the algorithm improved based on results if needed. The second phase introduces a more developed version of the algorithm, which will be based on more sophisticated methods, such as unsupervised or supervised pattern recognition, 3D model fitting and data fusion methods (the second version of the algorithm will be reported in D3.3 Mineral-sorting algorithm test report - X-MINE sensors). This deliverable concentrates on the first version of the algorithm.
Concrete practice to achieve the expected goal: First, the deliverable describes the x-ray sorting principle and algorithms based on existing sensors. Second, it proceeds to give an overview of the sorting research prototype. It then describes the details of the test procedure including the generation of labels for reference samples, and finally the sorting results are reported and conclusions with an evaluation of the current technologies are presented.
Expected impact/goal of the practice: The practice contributes to developing current sensor technology used in mineral sorting. It introduces the baseline performance of existing state of the art x-ray sensor technology for mineral sorting, evaluates its functionality, and based on the identified gaps and flaws, suggests how the technology should be improved (the actual improvements to the technology will be published in X-MINE D3.3, which is not yet on Cordis nor X-MINE website)
Who is the target user group of the practice/intervention or implementing the practice/intervention? The target group of the practice are industry and other researchers developing new mining technologies.