Thesis DEFINICIÓN DE TECNOLOGÍAS Y SIMULACIÓN DE PROCESAMIENTO DE IMÁGENES, ASPECTOS DIFERENCIABLES Y COMPARATIVOS EN LA IDENTIFICACIÓN DE PÓRFIDO DE COBRE
Loading...
Date
2017
Journal Title
Journal ISSN
Volume Title
Program
Campus
Universidad Técnica Federico Santa María UTFSM. Casa Central Valparaíso
Abstract
Hoy en día los costos de producción de minerales, a nivel mundial, tienden al alza. Esto se debea varios factores, pero principalmente a la disminución de la concentración del mineral en losyacimientos. Este es un problema para las mineras, debido a que aumenta el consumo derecursos para la producción de una misma masa de mineral procesado que hace un par de años.Es por esto es que se debe buscar una solución para mejorar la ley del mineral previo alprocesamiento de este.Este proyecto busca realizar un acercamiento a una solución a este problema mediante unmétodo de preconcentración. Para este trabajo de memoria se identificaron tecnologías que seencuentran dentro del espectro electromagnético: visible – color y cercano al infrarrojo (NIR)con buenos resultados trabajando con el pórfido de cobre. Luego se diseñaron dos algoritmos,uno binario y otro continuo, para la identificación de la ley del cobre en imágenes obtenidasmediante las tecnologías definidas. Después se compararon mediante métricas y verificacióncruzada con 15 iteraciones. Para este último paso se utilizaron 33 rocas provistas por el grupominero Las Cenizas, las cuales venían separadas por tres rangos de ley del mineral. Finalmente,se llegó a que el algoritmo continuo resulta separar mayor cantidad de rocas con menor error yeste proyecto verifica la posibilidad de identificar y separar pórfido de cobre utilizandotecnologías NIR y RGB.
Today the mineral production costs are rising. This is due to several factors, but mainly becausethe ore concentration in deposit are decreasing, thus increasing the resources consumption forproducing the same amount of ore processed a couple years ago. Therefore, is imperative tolook for a solution to improve the ore head grade prior to processing it.This project seeks to make an approach to a solution through a preconcentration method. Toachieve this goal, two visual sensor technologies were identified and selected among a list ofthem. These selected technologies are near infrared sensors and visual – color sensors, becauseof promising results that different studies with copper porphyry showed. To verify thesetechnologies, two algorithms were design, one is binary and the other one continuous. Thesealgorithms identify the ore head grade and separate it from the others. Then, it was necessary,to compare these two, using designed metrics and k-fold cross-verification of 15 iterations. Forthis step, 33 rocks were used, provided by the mining group Las Cenizas. These rocks wereseparated in three groups by ore grade. Finally, the results showed that the continuous algorithmcould separate more rocks with less error and better metrics than the binary one, therefore thisproject open the possibility of identifying and separating copper porphyry using NIR and RGBtechnologies.
Today the mineral production costs are rising. This is due to several factors, but mainly becausethe ore concentration in deposit are decreasing, thus increasing the resources consumption forproducing the same amount of ore processed a couple years ago. Therefore, is imperative tolook for a solution to improve the ore head grade prior to processing it.This project seeks to make an approach to a solution through a preconcentration method. Toachieve this goal, two visual sensor technologies were identified and selected among a list ofthem. These selected technologies are near infrared sensors and visual – color sensors, becauseof promising results that different studies with copper porphyry showed. To verify thesetechnologies, two algorithms were design, one is binary and the other one continuous. Thesealgorithms identify the ore head grade and separate it from the others. Then, it was necessary,to compare these two, using designed metrics and k-fold cross-verification of 15 iterations. Forthis step, 33 rocks were used, provided by the mining group Las Cenizas. These rocks wereseparated in three groups by ore grade. Finally, the results showed that the continuous algorithmcould separate more rocks with less error and better metrics than the binary one, therefore thisproject open the possibility of identifying and separating copper porphyry using NIR and RGBtechnologies.
Description
Catalogado desde la version PDF de la tesis.
Keywords
CERCANO AL INFRARROJO, COBRE, NIR, PORFIDO CUPRIFERO, RGB, SELECCIONADOR DE MINERALES