Robotics and AI
This project automates the identification and classification of rocks using a machine learning model deployed on a Raspberry Pi 4B. It enhances the speed and accuracy of rock identification for geology, mining, and educational purposes.
The project involves a custom CNN model and a pre-trained ResNet model to classify images of rocks into 53 different types. The solution includes data preprocessing, model conversion, and deployment on the Raspberry Pi for real-time identification using a camera setup.
Automated rock identification using a custom CNN model.
Real-time rock classification using a Raspberry Pi 4B.
Dataset of 53 rock types sourced and augmented via web scraping.
Machine learning models: Custom CNN and pre-trained ResNet.
Conversion of the CNN model to TensorFlow Lite for optimized Raspberry Pi deployment.
Hardware: Raspberry Pi 4B, Logitech C270 HD Webcam, Microsoft Surface Laptop Go 2.
Machine Learning Models: Custom CNN, ResNet50.
Libraries: TensorFlow, TensorFlow Lite, Python.