Lego Detector API

Introduction

Welcome to the Lego Detector API! This API is built using Flask, TensorFlow, and OpenCV to identify Lego bricks in a photograph and guess their types. It serves as a powerful tool for Lego enthusiasts and builders, offering a way to digitally catalog and recognize Lego pieces from images.

Project Structure

• app.py: The Flask application server.
• lego_detector.py: Module for detecting Lego bricks in images using OpenCV.
• lego_guesser.py: Module for predicting the type of Lego bricks using TensorFlow.
• model/lego_predicter.h5: Pre-trained TensorFlow model for Lego brick classification.
• requirements.txt: List of Python dependencies.

Getting Started

Prerequisites

• Python 3.x
• Flask
• TensorFlow
• OpenCV
• Other dependencies as listed in the requirements.txt file.

Installation

Clone the repository:

git clone https://github.com/thomas-rooty/lego-detector.git

Install the required Python packages:

pip install -r requirements.txt

Running the API

Start the Flask server:

python app.py

The server will start on http://127.0.0.1:5000/.

Usage

Detecting Lego Bricks

Send a POST request to /detect_legos with an image file. The API processes the image and returns bounding boxes around each detected Lego brick.

Example with curl:

curl -X POST -F "file=@path_to_your_image.jpg" http://127.0.0.1:5000/detect_legos

Guessing Lego Types

Send a POST request to /guess_lego with a list of URLs of cropped Lego brick images. The API uses a pre-trained model to predict the type of each Lego brick.

Example with curl:

curl -X POST -H "Content-Type: application/json" -d '{"urls": ["http://image_url1.jpg", "http://image_url2.jpg"]}' http://127.0.0.1:5000/guess_lego

Retrieving Processed Images

Access processed images by navigating to /output/.

API Endpoints

• GET /: Returns information about the API.
• POST /detect_legos: Accepts an image file and returns detected Lego bricks.
• POST /guess_lego: Accepts URLs of Lego brick images and returns their guessed types.
• GET /output/: Retrieves a processed image.

OpenCV steps

Note : We set a black background as a default starting point and we tell the user to take a picture of his bricks on a black mat to easily make the difference between bricks and background.

Examples may be white ! But the real AI uses a black background, there's no negative filtering here.

1. Grayscale

2. Diff between background and image + Gaussian blur to smooth out sharp edges

3. Otsu threshold

4. Find and draw contours

5. Bounding boxes

Contributing

We welcome contributions to improve the Lego Detector API! Whether it's bug fixes, feature enhancements, or documentation improvements, your help is valuable.

License

This project is licensed under the MIT License - see the LICENSE file for details.