Iron Ore Image Data

Published: 25 November 2024| Version 1 | DOI: 10.17632/6hp82tsb8g.1
Contributors:
Shama Firdaus,
,
,

Description

The dataset consists of Reflected Light Microscopic (RLM) Images of Iron Ores of four different grades. The zipped folder contains four sub folders, each representing and containing the raw images of a particular grade of iron ore. The grades of ores and the nomenclature used for naming the folders are - Blue Dust Ore (BDO), Hard Laminated Ore (HLO), Lateritic Ore (LO) and Soft Laminated Ore (SLO). The images are in TIF format of dimension 2048X1536. The dataset has been created for automating the task of characterization of iron ores using deep learning algorithms. The dataset can also be used for mineral composition analysis.

Files

Steps to reproduce

The iron ore subjects for the acquisition of RLM images belong to different mines of Singhbhum Craton of eastern India. The raw samples of ore are of four different grades of iron. These raw samples underwent processing to prepare them for RLM imaging. Two methods were followed for sample preparation. In first method glass slides were prepared from ores and in second method the sample surface was cut and polished to make free from scratches and to make the surface shiny for RLM. The prepared sample were then put under the Leica DM2700P digital microscope for capturing images. The images were captured under 10x and 20x magnification, in four different labels, each label representing a grade of the ore.

Institutions

Indiana Institute Of Technology, Birla Institute of Technology

Categories

Mineral Processing, Machine Learning, Image Database

Funding

Science and Engineering Research Board

CRG/2021/000918

Licence