Software

Image Processing Pipeline

  1. UV camera captures frame
  2. Depth cameras provide body segmentation (person vs background) and distance
  3. UV frame mapped to display coordinates based on user position from depth data
  4. False-color overlay generated: dark areas (good coverage) vs bright areas (missed spots)
  5. Overlay rendered on OLED behind two-way mirror

MVP (Cart)

  • UV camera feed displayed on adjacent screen
  • No overlay compositing needed — mirror and screen side by side
  • OpenCV or simple Python script for contrast enhancement and false-color mapping

Production

  • Real-time overlay compositing onto two-way mirror display
  • Dynamic calibration via depth cameras — no fixed “stand here” position needed
  • Walk-by preview: depth cameras detect passersby, mirror lights up 5-second UV preview

UV Forecast Projection

  • Pull UV index forecast from NOAA API
  • Calculate LED intensity to simulate projected UV level at +2 hours
  • Display clearly labeled: “SIMULATED — projected UV at 2 hours”
  • Different visual treatment from live view (e.g., live = grayscale, projection = color heat map)

Session Management

  • Free tier: 30-second check, rate-limited via face detection hash (no identity storage, expires daily)
  • Paid tier: tap-to-pay triggers 5-minute session with projection feature
  • Summer pass: UWB beacon detection auto-activates, unlimited use

NFC

  • Static URL or alarm trigger — taps phone, creates 3-hour “Reapply sunscreen” alarm
  • No app install, no server, no state

Alerts & Monitoring

  • Accelerometer threshold → vandalism alert via cellular
  • Camera feed available remotely
  • Basic usage telemetry: session count, duration, time of day