Autonomous Structural Intelligence System

An AI-powered system that analyzes 2D architectural floor plans using computer vision, generates 3D building models, recommends construction materials using a cost-vs-strength formula, and secures reports on the Stellar blockchain.

---

Features

• Floor Plan Analysis — Upload a 2D floor plan image (PNG/JPG). OpenCV detects walls and rooms automatically.
• 3D Visualization — Walls are extruded to 3 meters high and rendered interactively using Three.js with full orbit controls.
• Material Recommendations — An LLM (or rule-based formula) ranks construction materials (Red Brick, RCC, AAC Block, Structural Steel) using:

  Score = 0.35 × Strength + 0.30 × (1 - NormCost) + 0.20 × Durability + 0.15 × ClimateFit
  

• Stellar Blockchain Integration — Every report is SHA-256 hashed and stored on the Stellar testnet for immutable proof of authenticity.
• Report History — All analyses are persisted in SQLite and viewable with blockchain verification links.

---

Architecture

structural-ai-system/
├── artifacts/
│   ├── structural-ai/              # React + Vite frontend
│   │   ├── src/
│   │   │   ├── components/
│   │   │   │   ├── 3d/             # Three.js 3D floor plan viewer
│   │   │   │   ├── layout/         # App layout, sidebar
│   │   │   │   └── ui/             # Upload, material cards, blockchain badge
│   │   │   ├── pages/              # Dashboard, Reports, ReportDetail
│   │   │   ├── App.tsx
│   │   │   └── index.css
│   │   └── package.json
│   │
│   ├── structural-ai-backend/      # Python FastAPI backend
│   │   ├── main.py                 # FastAPI entry point
│   │   ├── requirements.txt
│   │   └── app/
│   │       ├── routes/
│   │       │   ├── analysis.py     # POST /api/analyze
│   │       │   └── reports.py      # GET /api/reports, GET /api/reports/{id}
│   │       ├── services/
│   │       │   ├── floor_plan_analyzer.py   # OpenCV wall/room detection
│   │       │   ├── material_recommender.py  # LLM + rule-based scoring
│   │       │   └── stellar_blockchain.py    # Stellar SDK integration
│   │       └── database.py         # SQLite persistence
│   │
│   └── api-server/                 # Express proxy server (TypeScript)
│       └── src/app.ts              # Routes /api/analyze → Python backend
│
└── lib/
    ├── api-spec/openapi.yaml        # OpenAPI contract (source of truth)
    ├── api-client-react/            # Generated React Query hooks
    └── api-zod/                     # Generated Zod schemas

---

API Endpoints

Method	Endpoint	Description
GET	/api/healthz	Health check
POST	/api/analyze	Upload floor plan image for analysis
GET	/api/reports	List all analysis reports
GET	/api/reports/{id}	Get a specific report

POST /api/analyze

Request (multipart/form-data):

• file — Floor plan image (PNG/JPG, max 20MB)
• budget — low, medium, or high (default: medium)
• location — Geographic/climate description (e.g., coastal, seismic zone, cold climate)

Response:

{
  "id": 1,
  "imageWidth": 800,
  "imageHeight": 600,
  "walls": [{"x1": 0.5, "y1": 0.5, "x2": 10.5, "y2": 0.5, "thickness": 0.2, "length": 10.0}],
  "rooms": [{"id": 1, "area": 25.0, "label": "Living Room", ...}],
  "totalWallLength": 48.5,
  "totalArea": 85.0,
  "recommendations": [
    {
      "material": "Red Brick",
      "costScore": 85.0,
      "strengthScore": 42.0,
      "overallScore": 72.5,
      "estimatedCostPerSqM": 1200,
      "pros": ["..."],
      "cons": ["..."],
      "reason": "..."
    }
  ],
  "model3d": {"walls": [...], "wallHeight": 3.0, "floorWidth": 10.0, "floorHeight": 7.5, "scale": 0.0125},
  "blockchainHash": "a3f8d2...",
  "blockchainTxId": "b7c9e1...",
  "stellarNetwork": "testnet-simulated",
  "createdAt": "2026-03-29T10:00:00",
  "summary": "Analyzed 800x600px floor plan. Detected 12 walls, 4 rooms..."
}

---

Material Scoring Formula

Score = w_strength × S_norm + w_cost × (1 - C_norm) + w_durability × D_norm + w_climate × Climate

Where:
  w_strength   = 0.35   (combined compressive + tensile strength, normalized 0–1)
  w_cost       = 0.30   (adjusted by budget level: low=1.5×, medium=1.0×, high=0.6×)
  w_durability = 0.20   (service life / 100 years)
  w_climate    = 0.15   (bonus for seismic/coastal/cold climate match)

Material Reference Table

Material	Compressive Strength	Tensile Strength	Cost (INR/m²)	Durability
Red Brick	7.5 MPa	0.5 MPa	₹1,200	80 years
RCC	25.0 MPa	3.5 MPa	₹2,200	100 years
AAC Block	4.0 MPa	0.8 MPa	₹1,600	70 years
Structural Steel	250 MPa	400 MPa	₹3,500	60 years

---

Stellar Blockchain

Each analysis report is:

1. Serialized to a canonical JSON object (wall count, area, top material, timestamp)
2. SHA-256 hashed to produce a 64-character hex digest
3. Stored as a memo on the Stellar testnet (via Stellar SDK or simulation)
4. Viewable on Stellar Expert

---

Tech Stack

Layer	Technology
Frontend	React 18, Vite, TypeScript, Tailwind CSS
3D Rendering	Three.js, @react-three/fiber, @react-three/drei
Charts	Recharts
Backend	Python 3.11, FastAPI, Uvicorn
Computer Vision	OpenCV (opencv-python-headless)
AI/LLM	Rule-based scoring (OpenAI optional via OPENAI_API_KEY)
Blockchain	Stellar SDK (testnet)
Database	SQLite (via Python sqlite3)
API Proxy	Express 5 + http-proxy-middleware
API Contract	OpenAPI 3.1 + Orval codegen

---

Environment Variables

Variable	Description	Default
PORT	Server port	8000
STELLAR_NETWORK	testnet or mainnet	testnet
OPENAI_API_KEY	Optional — enables LLM recommendations	None (rule-based fallback)

---

Development

# Start everything
pnpm --filter @workspace/structural-ai run dev   # Frontend (React)
cd artifacts/structural-ai-backend && python main.py  # Python API

# Run codegen after OpenAPI spec changes
pnpm --filter @workspace/api-spec run codegen