Predicting Road Defects and Optimizing Traffic Light Countdown to Reduce Congestion in Indonesia

Challenge Background

Traffic management is crucial for preserving road capacity, enhancing safety, and reducing congestion in urban areas like Jakarta, Indonesia. This project aims to leverage advanced technologies and data-driven solutions to address the challenges associated with traffic management in the city.

The Problem

Jakarta faces several traffic-related issues, including congestion, accidents, and road defects. These problems lead to inefficiencies, safety hazards, and increased travel times for residents. The primary problems to tackle are: 1. Vehicle speed and category classification to enforce speed limits. 2. Traffic density classification for efficient traffic redirection. 3. Pothole detection for road maintenance and safety.

Goal of the Project

• Develop machine learning models for vehicle category classification and detection.
• Create a model for traffic density classification.
• Implement a pothole object detection system.
• Develop a user-friendly web application for real-time traffic management.
• Investigate the possibility of future developments like road lane instance segmentation, plate number recognition, and vehicle tracking.

Project Timeline

What you'll learn

1. Improved understanding of deep learning model development and evaluation.
2. Proficiency in using PyTorch, OpenCV, and other relevant libraries.
3. Experience in creating user-friendly web applications.
4. Knowledge of traffic management challenges and solutions.
5. Collaboration and project management skills through teamwork on a complex project.