Enhanced Road Damage Detection for Smart City Surveillance.

Monitoring Road structure periodically is necessary to ensure the safety of public driving vehicles. Damages on the road structures play a crucial part in determining the state of the road. Historically, human visual assessment methods were used up until reachable at a person’s altitude. Monitoring of high-rise building systems on the road cannot be accomplished with this archaic technique. To guarantee the health and safety of the road, a more precise prognosis of road damage is also necessary for a more accurate prediction of road injuries. The suggested study's main goal was to create a powerful damage classification model utilizing the DCNN-DL architecture. Although there are numerous more pre-trained models for damage classification, their accuracy depends on more training parameters. The proposed model seeks to offer information on road damage by sensing the damages on the road using DenseNet with high accuracy and filling a damage complaint on Google Maps using GPS components to the maintenance team. MobileNet and ResNet both had an accuracy of 96.2% and 95.4%, respectively, while DenseNet did better with an accuracy of 98.7%. This case study demonstrates how well the DCNN-DL approach works to make locating and inspect road damage in smart cities easier.