A Novel Thermal Image-Based Metal Classification System Using Machine Learning Algorithms

Project Code :TCMAPY2015

Objective

This project presents a thermal image-based metal classification system that utilizes deep learning techniques to identify different types of metals accurately. It implements multiple architectures—CNN, MobileNet, MobileNet+LSTM, and ResNet+Transformer—to compare and achieve optimal classification performance. The dataset includes various metals such as Iron, Aluminium, Brass, Copper, Lead, and Zinc. Integrated with a Flask-based web interface, the system allows users to upload thermal images and receive instant, accurate predictions, ensuring efficiency, automation, and scalability.Through this platform, users can upload thermal images, view predictions, and analyze results with ease.

Abstract

The project presents a novel thermal image-based metal classification system that applies advanced machine learning techniques to identify different metal types using thermal imaging data. The proposed system utilizes deep learning models such as Convolutional Neural Network (CNN), MobileNet, MobileNet combined with Long Short-Term Memory (LSTM), and ResNet integrated with Transformer architectures. Each model is trained on a dataset containing various metal categories including Iron, Aluminium, Brass, Copper, Led, Zink, and a background class without metals.The implementation is developed using Flask as the backend framework, supported by a simple and responsive front-end using HTML, CSS, and JavaScript. The designed modules include Home, Register, Login, Classification, and Logout, providing an efficient and user-friendly interface for experimentation and evaluation.
Keywords: Thermal imaging, Metal classification, CNN, MobileNet, LSTM, Transformer, ResNet, Flask, Deep learning, Image analysis.

NOTE: Without the concern of our team, please don't submit to the college. This Abstract varies based on student requirements.