nerdyoutlier/Industrial_Surface_Defect_Inspection

Industrial Surface Defect Inspection

An end-to-end computer vision pipeline for automated surface defect detection and localization on industrial components (MVTec AD metal_nut dataset).

Directory Structure

industrial-surface-defect-inspection/
│
├── data/
│   └── metal_nut/             # Downloaded dataset
│       ├── train/
│       │   └── good/
│       ├── test/
│       │   ├── good/
│       │   └── ...
│       └── ground_truth/
│           └── ...
│
├── notebooks/
│   └── 01_dataset_exploration.ipynb
│
├── src/
│   ├── __init__.py
│   ├── data_validation.py
│   ├── preprocessing.py
│   ├── classification.py
│   └── localization.py
│
├── outputs/
├── requirements.txt
└── README.md

Setup & Installation

1. Create and activate a virtual environment:

   python -m venv venv
   # On Windows:
   .\venv\Scripts\activate
   # On Linux/macOS:
   source venv/bin/activate
   

2. Install dependencies:

   pip install -r requirements.txt
   

Workflow

1. Exploration: Run notebooks/01_dataset_exploration.ipynb to analyze defect types and ground truth masks.
2. Validation: Run python src/data_validation.py to check dataset integrity.
3. Classification & Localization: Train and infer using modules under src/.