Traffic sign detection and recognition

Question

What is the best method for detecting and recognizing traffic signs? I look at the popular methods for detecting road signs common in recent literature, but I don’t know which way is better! I would like to use detection methods based on color and shape.

I am processing image processing using opencv in visual studio C ++.

Answer 1

Try the following: https://sites.google.com/site/mcvibot2011sep/

Answer 2

Check dlib . The example/train_object_detector.cpp * file has some information on how this can be done. It uses a technique for describing functions called Histogram Oriented Gradients .

Check the starting values ​​of the following links:

• Detecting traffic signs in Mapillary images with C ++ dlib
• Dlib 18.6 Released: Create Your Own Object Detector

---

* Note: don't just use examples to train your detector! Read the files as a guide / tutorial! These sample programs assume that you are trying to detect faces and make some improvements based on this (for example, use image mirrors for training, since the edges are symmetrical, which can be disastrous for some signs).

---

Edit: check my implementation of a road sign detector and classifier using dlib:

https://github.com/fabioperez/transito-cv