TRAFFIC SIGN DETECTION

The goal of this project is to apply segmentation to a sample of different images. With it, the code will detect some traffic signals in the images and extract their information in order to recognize them. Different color spaces for mask generation have been used, including YCbCr, hsv and lab.

GETTING STARTED

These instructions will get you a copy of the project up and running on your local machine for testing purposes.

PREREQUISITES

The following packages must be installed in order to run the code:

• pyramid
• numpy
• imageio
• cv2
• random
• matplotlib
• skimage
• docopt
• sklearn
• plotly
• fnmatch

INSTALLING

We will use Python 3.6 as the interpreter. All the images given shall be located in the same folder (train folder for the training images and test folder for the test images).

RUNNING THE TESTS

Analyze Dataset

Execute database_analysys_main.py file. Arguments:

• output_dir: Directory where to store output masks, etc.

This execution provide with plotly a table with result, then is necessary get a free account of plot.ly

Train Model

Execute traffic_sign_detection.py file. Arguments:

• output_dir: Directory where to store output masks, etc.

The input data is stored in train folder

Get Final Masks

Execute traffic_sign_detection.py file. Arguments:

• images_dir: Directory with sample images and annotations.

• output_dir: Directory where to store output masks, etc.

• pixelMethod: Colour space used for creating the masks (YCbCr, hsv or rgb). Not required in arguments

• windowMethod: Window method used for detecting signals. Not required in arguments

AUTHORS Cara Roca, David Fuentes López, Daniel Rodríguez Orihuela, Andreu J