The problem CityGuardian solves

CityGuardian is designed to make roads safer by automatically detecting potholes using AI and reporting them in real time. This app serves as an urban infrastructure companion that:

📱 Allows users to report potholes by capturing road images through the app.

🧠 Uses a trained YOLOv3 object detection model to identify potholes in real-time.

📍 Automatically logs pothole locations using Google Maps API.

📊 Sends reports to municipal authorities or dashboards via Firebase.

🔒 Supports secure user authentication using Firebase Auth.

This app makes it easy for commuters and city officials to spot road hazards faster, reduce vehicle damage, and plan repairs more efficiently — ultimately making transportation safer and smarter.

Challenges I ran into

Model Accuracy & Dataset Quality:
Initially, the YOLOv3 model struggled with inconsistent detection due to low-quality training data. We improved this by augmenting the dataset and fine-tuning hyperparameters.

Flutter + Python Backend Integration:
Bridging a mobile frontend with a Python-based AI backend posed some challenges. We resolved this by using REST APIs hosted on Google Cloud and Firebase Functions to smoothen communication.

Real-time Location Mapping:
Integrating Google Maps API to accurately track and display potholes required handling permissions and optimization for low-latency updates.

Despite these challenges, careful debugging and real-world testing helped us overcome hurdles and deliver a functional prototype.