Executive Summary : | Road signs are crucial for traffic regulation and safety, serving as symbols of graphical images or text. The Ministry of Road Transport and Highways in India is responsible for designing and administering road transportation rules, regulations, and policies. With over 150 different traffic signs on Indian roads, the automatic Traffic Sign Detection and Recognition (TSDR) problem has gained international attention due to the implementation of the Advanced Driver Assistance System (ADAS). TSDR systems can learn about these signs and provide cautionary information to reduce accidents and help self-driving cars and autonomous driving systems make appropriate decisions. However, TSDR is challenging due to factors such as low visibility, low contrast, noise, and blurs. In this work, a two-step approach is proposed to address these issues. The first step involves enhancing traffic sign images using an adaptive deep learning model, while the second phase uses an object detection and recognition model to recognize traffic signs in an unconstrained environment. Limited research has been done in the Indian traffic context, so the project aims to create an Indian Traffic Sing Dataset (ITSD) in different weather and road conditions. The model will be validated with the ITSD to provide detailed information about data collection, annotations, and statistics. |