CubeAI

AI-powered object detection for construction cubes, assessing usability based on smoothness and cracks to ensure quality and safety on job sites.

Object Detection and Cropping of Cubes Using YOLOv10

Project Overview

This project implements an object detection pipeline using the YOLOv10 model to identify and crop images of cube objects from a dataset. The workflow includes dataset organization, model training, evaluation, and cropping of detected objects.

Features

• Dataset Organization: Automatically organizes images and labels into training, validation, and test sets.
• Model Training: Trains the YOLOv10 model on the provided dataset.
• Performance Evaluation: Calculates precision, recall, and F1-score to evaluate model performance.
• Object Cropping: Crops detect cube objects from images and save them to a specified directory.

Requirements

• Python 3.x
• ultralytics
• opencv-python
• numpy
• matplotlib
• scikit-learn
• tqdm

Installation

Install the required packages using pip:

pip install ultralytics opencv-python numpy matplotlib scikit-learn tqdm