Traffic Sign Recognition for Self Driving Vehicles Using MatLab and Tensorflow

Abstract

Traffic sign recognition is one of most important aspects of creating a safe and user friendly autonomous/self-driving vehicle. The objective of this work is the development of an algorithm for the automatic recognition of traffic signs. Two major problems exist in the process of detection and recognition of traffic signals. Road signs are frequently occluded partially by other vehicles and many objects are present in traffic scenes which make the sign detection hard and pedestrians, other vehicles, buildings, and billboards may confuse the detection system with patterns similar to that of road signs. The main objective of this project is to design and construct a computer based system which can automatically detect road signs so as to provide assistance to the users or the machine so that they can take appropriate actions. The proposed approach consists of building a model using convolutional neural network by extracting traffic signs from an image using color information. We have used Convolutional Neural Networks (CNN) to classify the traffic signs and we used color based segmentation to extract/crop signs from images.