My bachelor's degree project covering the topic of copper ore grains classification based on thermal images. The project includes data analysis and feature extraction from footage acquired using active thermography approach. The copper grains are characterised by tracking the number of spots with lower thermal emissivity. Classification algorithm utilizes a neural network developed with Keras interface.
The neural network is a proof of concept construct, to provide practical rationale for presented method of classifying grains.
Start reading Jupyter Notebooks here.
Feel free to check out my thesis (written in Polish), which offers in-depth insight into the project and research.
