Effective Road Quality Mapping and Navigation
View the Paper »
Report Bug
·
Suggest Feature
Table of Contents
The proposed system addresses the issue of potholes and poor road quality in developing countries by utilizing an end-to-end architecture that combines both computer vision and sensor-based models. The product includes connected applications like an analytics dashboard for government use and a navigation application for consumers. This approach is an innovative, scalable, reliable, and can be deployed with minimal changes to existing infrastructure.
Key Features include:
- Specialized software for collecting motion sensors data from IoT devices and streaming to our cloud pipeline.
- An innovative ensemble of YOLOS and YOLOv8 models, achieving a 97.34% mAP@0.50 score for pothole detection.
- Advanced sensor-based models for road quality detection, achieving accuracies of 98.5% and 95.4%.
- An Analytics Dashboard to manage the data and insights for government officials.
- A Navigation Application for consumers to avoid potholes and poor road quality.
