This thesis explores the application of Artificial Intelligence (AI) techniques, with a particular focus on Neural Networks, to enhance the safety of roadside guardrails along the highways of the Turin area. Through the analysis of data collected from sensors and cameras mounted on a moving vehicle, the research aims to automate the detection of guardrails and their classification based on type, compliance with current safety regulations, structural integrity, and the presence of possible damage. The ultimate goal is to develop a monitoring database capable of identifying non-compliant or potentially critical guardrails, thus providing valuable support for preventive maintenance strategies and contributing to improved driver safety, even in the event of an accident.