An AI-powered web application that analyzes playground photos to extract structured information about playground features, age ranges, and safety considerations.
PlayGroundFinder uses computer vision and AI to automatically analyze uploaded playground photos and extract:
- Features detected: slides, swings, monkey bars, climbing structures, etc.
- Age range recommendations: estimated appropriate age groups
- Safety information: child detection, privacy concerns
- Location details: when available from image metadata
The app identifies playground features and provides structured data like age ranges and directions.
This project includes both Flask and FastAPI implementations:
- Web interface for uploading and viewing playground photos
- Template-based UI with upload forms and result display
- Image processing and AI analysis integration
- REST API for programmatic access
- Database integration with SQLAlchemy
- CRUD operations for playground data
- Schemas: Pydantic models for data validation
- Utils: Image processing and AI analysis utilities
- Models: Database models and structures
- 📸 Image Upload: Support for PNG, JPG, JPEG, GIF, WebP formats
- 🤖 AI Analysis: Uses Anthropic Claude for intelligent image analysis
- 🏗️ Feature Detection: Automatically identifies playground equipment
- 👶 Age Assessment: Estimates appropriate age ranges
- 🔒 Privacy Protection: Detects children in photos for blurring recommendations
- 💾 Data Storage: Persistent storage of analysis results
- 🌐 Dual Interface: Both web UI and REST API access
- Backend: Python, Flask, FastAPI
- AI/ML: Anthropic Claude, Instructor library
- Database: SQLAlchemy (SQLite)
- Image Processing: Pillow
- Data Validation: Pydantic
- Frontend: HTML templates with Jinja2
-
Install Dependencies:
pip install -r requirements.txt
-
Set Environment Variables:
export ANTHROPIC_API_KEY=your_api_key_here -
Run Flask App:
cd flask_app python app.pyVisit: http://127.0.0.1:5000
-
Run FastAPI App:
cd fastapi_app uvicorn fastapi_main:app --reloadVisit: http://127.0.0.1:8000
curl -X POST "http://127.0.0.1:8000/upload/" \
-F "file=@playground_photo.jpg"curl "http://127.0.0.1:8000/playground_images/"url: Image file pathtaken_at: Timestampchildren_detected: Boolean for privacyneeds_face_blurring: Privacy recommendationestimated_age_range: Min/max age valuesfeatures_detected: List of playground equipment
primary_name: Main feature namealiases: Alternative names/variations
🌐 See it in action: playgroundfinder.webflow.io
This repository contains the AI analysis backend that powers playground data extraction. The live prototype showcases the user-facing application where parents and caregivers can:
- 🗺️ Browse interactive maps of analyzed playgrounds
- 💎 Access premium content through a simple $1 subscription model
- 📍 Discover local play spaces with detailed feature information
- 📊 View structured data generated by this AI analysis system
The Webflow prototype demonstrates the complete user journey - from discovering playgrounds on a map to accessing detailed information that's been automatically extracted using the AI models in this codebase. This creates a full pipeline: AI analysis (this repo) → Structured playground data → User-friendly discovery platform (Webflow site).
This project was developed as a proof-of-concept for AI-powered playground analysis. Current status:
- ✅ Core image analysis functionality
- ✅ Web interface for uploads
- ✅ REST API for programmatic access
- ✅ Data persistence and retrieval
- 🔄 Ongoing refinements to AI accuracy
- 🔄 UI/UX improvements
- Mobile-responsive design
- Batch image processing
- Geographic mapping integration
- Advanced safety assessments
- Community features and reviews
- Export functionality for park departments
This app helps parents, caregivers, and park planners quickly assess playground suitability and features through automated image analysis.