Machine Learning Based Pothole Detection and Alert System Using Arduino Onboard Accelerometer

Potholes are caused by road deterioration and weathering. They not only cause inconveniences for individuals, but also deaths attributed to automobile accidents. Roads contribute substantially to the economy and serve as a transportation platform. Potholes in roadways are one of the most significant challenges in the transportation infrastructure. Numerous studies have explored automating pothole detection with computer vision techniques, which encompass a vast array of image processing and image detection algorithms. This study's major purpose is to find an automated solution to the current pothole problem on local roads. This pothole problem has a significant impact on traffic, which promotes economic deterioration. The implemented system achieved a detection accuracy of 95%, significantly improving the real-time identification of potholes. The prototype's ability to rapidly alert authorities about the pothole's precise location resulted in faster response times for road maintenance, thereby minimizing traffic disruptions. The system's effectiveness was further validated by its successful deployment in a simulated urban environment, demonstrating its potential for widespread application in real-world scenarios.