Artificial Intelligence Inspired Automatic Road condition Classification

Advanced sensors for control, safety, monitoring, perception, and operation are becoming more necessary as the automobile industry undergoes rapid transformation. As computing power increases, automobiles increasingly imitate mobile computers. The quality of the roads has a significant role in determining the cost and security of transportation. In order for agencies to conduct timely repairs and maintain ideal road conditions, damage identification is a crucial duty for road maintenance. Road quality is being monitored using a variety of image and video monitoring techniques. Manually assessing and analyse incoming information may be a time-consuming process. Thus, it is important to use computer vision and deep learning methods to enhance the processes for inspecting and assessing the status of control objects. Researches have rapidly adopted object detection/classification deep learning models for road damage detection thanks to the rapid advancement of deep learning methods in recent years. Municipalities may save resources and time by using automated road damage detecting technologies. This research suggests using a CNN trained on a freshly curated dataset of road damage data from Saudi Arabia to identify the degree of road damage. Cracks, potholes, depressions, and shove are the four classified classifications presented in this study's dataset. Experts analyzed data that was obtained with the help of maintenance workers, volunteer in the city of Coimbatore using smartphones. More than one deep learning algorithm was also deployed and tested on the suggested dataset. The research shown that the suggested custom CNN (RoadNet) outperformed the per-trained models in terms of accuracy.