How couchpugtato/vortex works
Vortex’s main workflow captures camera images, detects AprilTags, estimates camera and robot pose from camera settings and the field layout, then publishes detections and localization for robot consumers. Its other substantial product branches are interactive simulation and failure replay, synthetic-image regression, camera calibration, Jetson acceleration, and deployment. Edges are limited to relationships shown in the excerpts, SOURCE INDEX references, or README; unsampled subsystem internals and some runtime wiring are therefore not expanded.
Main entry points: src/main.rs Vision pipeline · drive.py Interactive simulator · detector.rs AprilTag detection · src/main.rs Native deployment controller
Components 32
Vision runtime
- Vision pipeline
src/main.rs - Camera and field config
src/config.rs - Image preprocessing
src/preprocess.rs - AprilTag detection
src/detector.rs - Tag refinement
src/tag_refine.rs - Pose estimation
src/pose.rs - GPU detection
src/gpu_detector.rs - VPI acceleration
src/vpi.rs - TensorRT objects
src/yolo_detector.rs - Runtime health
src/runtime_health.rs - Runtime diagnostics
src/runtime_diagnostics.rs
Deployment
- Deployment desktop app
VortexDeploymentInterface/main.js - Deployment interface
VortexDeploymentInterface/renderer/app.js - Release staging
VortexDeploymentInterface/staging-deploy.js - Native deployment controller
deployment/src/main.rs - Release activation
deployment/src/activation.rs - Service management
deployment/src/service.rs
Simulation and validation
- Interactive simulator
chassis_sim/drive.py - Chassis routes
chassis_sim/routes.py - Ghost prediction
chassis_sim/ghost_prediction.py - Field terrain
chassis_sim/terrain.py - Failure capture and replay
chassis_sim/failure_capture.py - Synthetic image generation
simulation_tests/generate.py - Native regression replay
simulation_tests/replay/src/main.rs - Regression runner
simulation_tests/run.py
Calibration and acceleration
- Intrinsics calibration
camera_calibration/camera_intrinsics_solver.py - Distortion calibration
camera_calibration/camera_distortion_solver.py - AprilTag calibration detection
camera_calibration/detect_apriltags.py - TensorRT engine builder
model_builder/build_engine.cpp
Other components
- Operator
- Camera input
- Robot consumer
Connections 33
- Camera input → Vision pipeline: supplies frames
- Vision pipeline → Camera and field config: loads settings
src/main.rs - Vision pipeline → Image preprocessing: uses
src/main.rs - Vision pipeline → AprilTag detection: dispatches detection
src/main.rs - AprilTag detection → Tag refinement: builds patches
src/detector.rs - AprilTag detection → Pose estimation: estimates tag pose
src/detector.rs - Vision pipeline → Pose estimation: estimates robot pose
src/main.rs - Vision pipeline → GPU detection: uses optional backend
src/main.rs - Vision pipeline → VPI acceleration: uses optional acceleration
src/main.rs - Vision pipeline → TensorRT objects: dispatches object detection
src/main.rs - Vision pipeline → Runtime health: reports health
src/main.rs - Vision pipeline → Runtime diagnostics: collects diagnostics
src/main.rs - Vision pipeline → Robot consumer: publishes telemetry
- Operator → Deployment desktop app: configures deployment
- Deployment interface → Deployment desktop app: uses desktop services
- Deployment desktop app → Release staging: stages release
VortexDeploymentInterface/main.js - Release staging → Native deployment controller: uploads and invokes
VortexDeploymentInterface/staging-deploy.js - Native deployment controller → Release activation: activates release
- Native deployment controller → Service management: manages service
deployment/src/service.rs - Operator → Interactive simulator: drives simulation
- Interactive simulator → Chassis routes: loads route behavior
chassis_sim/drive.py - Interactive simulator → Ghost prediction: uses prediction
chassis_sim/drive.py - Interactive simulator → Field terrain: uses field terrain
chassis_sim/drive.py - Chassis routes → Field terrain: reads bumps
chassis_sim/routes.py - Chassis routes → Interactive simulator: uses motion step
chassis_sim/routes.py - Interactive simulator → Failure capture and replay: captures failures
chassis_sim/drive.py - Interactive simulator → Native regression replay: replays frames
- Regression runner → Synthetic image generation: generates scenarios
simulation_tests/run.py - Regression runner → Native regression replay: runs native replay
simulation_tests/run.py - Operator → Intrinsics calibration: calibrates camera
- Operator → Distortion calibration: calibrates distortion
- AprilTag calibration detection → Intrinsics calibration: provides tag observations
- TensorRT engine builder → TensorRT objects: builds inference engine