LIVE OBJECT DETECTION OF POTHOLE AND SPEED BREAKERS BY IMPLEMENTING YOLOV3 WITH TENSORFLOW 2

Abstract

In this final project, a system to actively detect potholes and speed breakers using YOLOv3 as an object detection algorithm and its capability to provide bounding box coordinates for each detected object as well as detect multiple object classes simultaneously and Tensorflow 2 support makes it adaptable to various applications with the goal of increasing awareness of road conditions in order to improve road safety is developed. The existence of this system's development could benefit users because it has made it simpler to identify potholes or speed breakers on the road, which improves the safety of using the road using vehicles. This project was created utilizing the Object detection frameworks and Computer Visions strategy, which prioritizes real-time processing for object detection. This method is ideal for this project because it was designed to make identifying objects easier and more efficient for drivers in real-world circumstances. By combining advanced object detection frameworks and computer vision techniques, mobile applications that improve road safety for users can be developed. The device can instantaneously recognize possible risks such as speed breakers and potholes, as well as deliver safety alerts. The real-time processing capabilities ensure that drivers receive immediate feedback, allowing them to react quickly to road conditions ahead.