Mineral Identification using CNN

Abstract

In this Study, a strategy for detecting minerals in a picture collection is offered. The novelty is to perform the multi-classification of mineral based on the given image using convolutional neural network. Identification and classification of minerals are the fundamental of the mining and processing of minerals [1]. The suggested technique next detects the minerals and labels them with their relevant class labels using multiple photos from the image collection. Tensor Flow, Keras, and OpenCV are used to detect these minerals. Keras is a free and open-source Python interface for artificial neural networks. Keras is a Python library that connects to the TensorFlow library. These are photos from the Training Image dataset are fed into the training model. Our system recognizes the numerous bright variations of minerals from the training dataset using a particular collection of hand specimen photographs of minerals in seven classes: diamond, bornite, chrysocolla, malachite, muscovite, pyrite, and quartz. The model is trained until the error rate becomes insignificant. The trained model is put to the test on some real-world photos.