Skip to content

About

A project for segmenting welding defects using a U-Net model. It detects and classifies 10 defect types based on labeled images. Developed as part of a data mining course, it explores the use of computer vision to improve defect detection in the welding industry.

Resources

Stars

6 stars

Watchers

1 watching

Forks

Latest commit

 

History

3 Commits

Folders and files

Repository files navigation

Welding Defects - Image Segmentation Project

Overview

This project focuses on detecting and segmenting welding defects using deep learning techniques. The goal is to automate defect identification in welding images to improve quality control in manufacturing. It has been trained from this dataset https://universe.roboflow.com/test-ulmj1/defect-detection-renamed-classes/dataset/4

Features

  • Image preprocessing and augmentation
  • Deep learning-based segmentation
  • Evaluation metrics for model performance
  • Dataset loading and visualization

Requirements

  • Python 3.x
  • PyTorch
  • OpenCV
  • NumPy
  • Matplotlib
  • Jupyter Notebook

Installation

  1. Clone the repository:
    git clone https://github.com/yourusername/welding-defects-segmentation.git
    cd welding-defects-segmentation
  2. Install dependencies:
    pip install -r requirements.txt

Usage

Run the Jupyter notebook to train and evaluate the model:

jupyter notebook

Open welding defects - Image Segmentation Project.ipynb and execute the cells.

License

This project is licensed under the MIT License.

About

A project for segmenting welding defects using a U-Net model. It detects and classifies 10 defect types based on labeled images. Developed as part of a data mining course, it explores the use of computer vision to improve defect detection in the welding industry.

Resources

Stars

6 stars

Watchers

1 watching

Forks

Releases

Packages

Contributors

Languages