Ball Strike Camera

Native iOS SwiftUI + AVFoundation starter app for detecting a stationary ball, switching to ready mode, and grabbing frames around the moment the ball leaves its spot.

What is implemented

• Landscape-only SwiftUI layout:
  • Top half: 2/3 live camera preview, 1/3 captured-frames panel.
  • Bottom half: five empty placeholder boxes.
• Camera preview uses .resizeAspect, so the full camera feed is scaled into the preview area instead of being cropped.
• Four opaque shutter buttons over the preview:
  • Moon: 1/1000
  • Cloud: 1/2000
  • Sun: 1/4000
  • Bright sun: 1/8000
• AVFoundation session configured for 240 fps when the device supports it.
• Lightweight Swift detector that searches each frame for a bright, low-saturation ball-like blob.
• State machine:
  1. Searching for ball
  2. Tracking with green circle overlay
  3. Ready after the ball is stationary for multiple frames
  4. Capture trigger when the ball leaves the ready spot
• Frame capture: keeps a rolling buffer and displays frames around the hit event.

Open and run

1. Open BallStrikeCamera.xcodeproj in Xcode.
2. Select your iPhone as the run target. The camera and high-FPS capture will not work correctly in Simulator.
3. Set your Apple Developer Team in Xcode under Signing & Capabilities.
4. Run on device.

Important notes

• 240 fps depends on the physical iPhone camera format. If a device does not support 240 fps at the selected resolution, the app uses the best available matching format it can find.
• The detector is intentionally simple and CPU-only. It is designed to be replaced later by a stronger Core ML / Vision detector.
• Very fast shutter speeds need a lot of light. At 1/8000, the image may be dark unless lighting is strong.
• Frame thumbnails are currently generated directly from video buffers for simple visual debugging. For production, use a more memory-efficient ring buffer or write frames to disk.

Repo layout

BallStrikeCamera/
  App/
    BallStrikeCameraApp.swift
  Camera/
    CameraController.swift
  Detection/
    BallDetector.swift
    BallObservation.swift
  Models/
    CameraPhase.swift
    PlatformImage.swift
  UI/
    ContentView.swift
    CameraPreview.swift
  Resources/
    Info.plist
    Assets.xcassets/

Next upgrades

• Replace the heuristic detector with a small Core ML model trained on ball images.
• Add manual ISO controls per shutter preset.
• Add calibration for different ball colors/backgrounds.
• Add frame-by-frame scrubber after capture.
• Export the captured frame burst as images or a slow-motion clip.