Segmentation and Detection of Potholes in Roads Using Computer Vision in Critical Embedded Systems for ADAS and Autonomous Vehicles

• Grant number: 25/03224-4
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date: May 01, 2025
• End date: December 31, 2025
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computer Systems
• Principal Investigator: Diego Renan Bruno
• Grantee: William D'Abruzzo Martins
• Host Institution: Instituto de Biociências, Letras e Ciências Exatas (IBILCE). Universidade Estadual Paulista (UNESP). Campus de São José do Rio Preto. São José do Rio Preto , SP, Brazil

Abstract

Poor road infrastructure, especially the presence of potholes, poses a significant risk to both public safety and the automotive industry, and is a critical problem in Brazil. In this context, the adoption of assistive driving technologies, focused on detecting irregularities on the road, becomes essential to mitigate these risks, allowing drivers to react more efficiently. To this end, approaches based on Deep Neural Networks have shown promise in object detection and segmentation, overcoming traditional methods that often have limitations in complex scenarios.This project proposes the development and evaluation of Deep Learning models applied to Computer Vision for this purpose, exploring architectures such as Convolutional Neural Networks (CNNs) and Vision Transformers. The goal is to overcome the limitations of existing approaches, which often rely on inadequate data sets or computationally intensive models, unfeasible for embedded devices. As a result, the developed models will be applied in a low-cost solution for real-time pothole detection, assisting drivers with Advanced Driver Assistance Systems (ADAS) and/or sending information to a decision-making module, in the case of autonomous vehicles, allowing a corrective maneuver to be performed. It is worth noting that this research is part of the project for the first autonomous truck in Latin America with the Institute of Mathematical and Computer Sciences of the University of São Paulo, called CaRINA (Intelligent Robotic Car for Autonomous Navigation), enabling the generation of new contributions.