The identification of atypical road conditions utilizing smart cameras, acceleration sensors, and artificial intelligence” aims to develop a system that can detect abnormal road conditions in real-time using smart cameras, acceleration sensors, and AI. The system will be designed to improve road safety by identifying and authorities about hazardous road conditions such as potholes, cracks, or other irregularities on the road surface.
The project will involve the installation of smart cameras and acceleration sensors on the roadsides or on vehicles themselves. The data collected by these devices will be analyzed by an AI algorithm or system that will use machine learning techniques to identify patterns and anomalies in the data.

Currently, road inspections are time-consuming and labor-intensive, which is a significant problem as road surfaces degrade daily due to heavy traffic. Municipalities conduct regular inspections to maintain roads efficiently. This project aims to design, build, and test an automated inspection system equipped with a camera that captures video streams from roads with and without defects. The captured data will be analyzed using the  machine learning  to train and test the network and provide recommended actions for fixing/correcting the road defects. The project’s approach includes three main tasks: Data acquisition, Data Training/Testing, and Dashboard Building and Testing.

The Data acquisition stage involves using HD cameras to capture live videos of various road defects and issues, including images from standard datasets, images from real roads, and live video recordings. The Data Training/Testing stage will utilize transfer learning to address challenges related to the scarcity of data and the lack of human labels. The  machine learning  will be used to classify road defects, and standard image processing techniques will highlight areas and guide the inspection process. Finally, the Dashboard Building and Testing stage will visualize the defects and the municipality’s recommended actions through a graphical interface.

In addition, the relevant information on the abnormal road surface is sent to the municipal road maintenance unit so that the road maintenance personnel can timely get the road damage status and repair it to ensure the safety and comfort of travel.