System for Detecting Traffic Sign Boards Using Machine Learning

Abstract

When the new generation of self-driving cars is fully automated, it becomes crucial for the vehicles to recognize the road signs that are prominently posted on the sides of roads in big cities. To reach level5 of automation, the systems must recognize and adhere to traffic laws. Accuracy in the technology is required when firms like Tesla, Uber, Google, Mercedes-Benz, Toyota, Ford, Audi, and others are developing autonomous and self-driven vehicles. The vehicles should be able to understand traffic signs and make judgments in accordance with them. The current system allows automobiles to recognize stop signs and road bumps in front of them. It also measures the space between the two vehicles too. Videos with at least 20 different signs will be used to create the graphics, together with local photographs and information from a public database. The photos will be categorized using Keras. The construction of a CNN model will be followed by its training and validation. The graph will be plotted using Matplotlib to determine the model's accuracy. For the interaction, a Python GUI will be constructed. The automated driving system will be able to ensure that the traffic signs are identified and are interpreted by the system with improved performance than the current capacity interaction once the traffic sign boards are recognized.