The Autonomous Computer Vision & Kinetic Tracking System for ESP32-CAM
Void Sentry is a high-performance, embedded autonomous tracking solution built for the ESP32-CAM. By utilizing optimized Frame Differencing algorithms, the system performs real-time motion analysis on the "edge," isolating moving targets and generating kinetic coordinates for hardware-based tracking without the need for external processing.
The system operates on a dual-core cycle to ensure low-latency video delivery while maintaining a high-frequency tracking loop.
- Computer Vision Pipeline: Optimized for QVGA (320x240) resolution to maximize frames-per-second (FPS) and minimize computational jitter.
- Kinetic Engine: Translates motion "blob" centroids into PWM signals for high-torque servo response.
- Visual Feedback: A real-time Augmented Reality (AR) overlay is injected directly into the MJPEG stream, providing zero-latency target telemetry.
| Feature | Implementation |
|---|---|
| Detection Algorithm | Temporal Frame Differencing (Greyscale) |
| Tracking Logic | Geometric Centroid Calculation (X, Y Coordinates) |
| Noise Reduction | Adaptive Hysteresis Thresholding |
| I/O Protocol | PWM (Pan/Tilt Servos) + HTTP MJPEG Stream |
| Resolution | 320x240 (Optimized for Edge Inference) |
- Intelligent Thresholding: Dynamically filters out sensor noise and environment fluctuations (e.g., lighting changes) to prevent false positives.
- Target Locking: Implements "stickiness" logic to maintain focus on the primary moving mass while ignoring secondary background static.
- Low-Latency Dashboard: A web-based HUD designed for mobile and desktop, featuring real-time telemetry stats and manual override controls.
- Modular Servo Control: Easily adaptable for various Pan/Tilt hardware configurations.
- Clone the repository:
git clone [https://github.com/KoganTheDev/void-sentry-esp32.git](https://github.com/KoganTheDev/void-sentry-esp32.git)