About

🟩 Upwork Job Post: AI App to Detect Pool Border and Calculate Stone Quantities by Photo ⸻ 🧠 Project Overview: I want to build a web or mobile app that allows a user to upload a top-down photo of a swimming pool, define a real-world scale on the image, and automatically: • Detect the outer pool perimeter • Divide the contour into curved segments • Analyze the radius of curvature for each segment • Match each segment to one of 9 predefined stone types (based on curvature radius) • Calculate how many stones of each type are required to fully cover the border • Output a visual map + table with all calculations • Allow the result to be exported to PDF ⸻ 📷 How the Photo Will Be Taken: To ensure accurate results, the image will: • Be taken from above (using drone, ladder, balcony) • Show the entire pool clearly and without obstructions • Include a scale reference (e.g. 1-meter-long object or ruler placed next to the pool) • Be in good lighting, clear, without shadow or glare Example user input: • Uploads a photo • Marks 2 points on the image (e.g. ends of a 1m object) • Inputs: “this distance = 100 cm” • System then calculates stone types and quantities ⸻ 🔧 Required Features: 1. Image Upload + Scale Calibration • User uploads a pool photo • User selects two points and inputs real-world distance between them • App calculates the pixel-to-cm scale 2. Contour Detection & Shape Analysis • Automatically detect the full perimeter of the pool • Approximate the curve into segments • Calculate radius of each curved segment 3. Stone Library • Predefined 9 types of edge stones with fixed curvature radii (R30, R40, …, R110) • Match the best-fitting stone type to each segment • Straight segments use standard straight stone 4. Stone Quantity Calculator • User can enter length of a single stone (e.g. 10 cm) • App calculates: • Total perimeter • Stone types used • Quantity per type (rounded up) 5. Output & Export • Show visual drawing: pool border segmented and color-coded by stone type • Summary table: Type Radius Segment Length Quantity • Export results to PDF and/or send by email ⸻ ✅ Tech Preferences: • Frontend: React, Flutter, or Streamlit/Gradio • Backend / Logic: Python (OpenCV, Shapely, NumPy) • Optional: AI segmentation using Segment Anything (Meta) • Output formats: Web page + downloadable PDF ⸻ 📂 Deliverables: • Fully working prototype (web-based or mobile) • Source code + basic documentation • Design should be clear, modern, and easy to use ⸻ 💬 Bonus if you can: • Add support for photo correction (remove perspective distortion) • Add manual override tools (user can edit detected outline) • Build a mobile app (Flutter or React Native) ⸻ 🔎 Please include: • Examples of similar apps or tools you’ve built • How you would approach image segmentation & radius detection • Estimated timeline and budget • Any questions you need answered before starting ⸻ Thanks! Looking forward to working with you. ⸻