ccm-esp32-vision-node

Embedded camera node for ESP32-S3 with a modular C++ architecture for real-time computer vision. This project demonstrates professional embedded CV engineering: camera bring-up, modular pipelines, efficient frame handling, and low‑power vision processing.

Work‑in‑progress: Part of the CCM Code Embedded Vision toolkit.
Designed as a portfolio‑quality example of clean, modern ESP‑IDF C++ development.

---

🚀 Features (Current & Planned)

✅ Implemented

• ESP-IDF C++ project with multi-component architecture
• C++ CameraNode component using the esp32-camera managed component
• CvPipeline with Grayscale, ROI, Downsampling, Thresholding, and Blob Detection
• Host-Based Simulator (simulation/) for fast PC-based testing
• Persistent Settings using NVS (Non-Volatile Storage)
• Real-time FPS measurement and per-stage profiling

🛠 In Progress

• MJPEG Wi‑Fi streaming server (stream_server)
• Color blob detector (HSV)

📌 Planned

• UART debug dashboard (FPS, memory, timings)
• Example: simple color-object tracker

---

🧱 Project Architecture

For full component documentation, see:
👉 docs/architecture.md

ccm-esp32-vision-node/
│
├── components/
│   ├── camera_node/       # Camera bring-up abstraction
│   ├── cv_pipeline/       # Modular image processing pipeline (WIP)
│   ├── stream_server/     # Wi-Fi streaming endpoint (planned)
│   ├── drivers/           # Camera/sensor-specific helpers
│   └── utils/             # Logging, timers, profiling
│
├── firmware/
│   ├── main/              # App entry point (main.cpp)
│   └── CMakeLists.txt
│
├── docs/
│   ├── architecture.md
│   └── media/             # diagrams (placeholder)
│
└── examples/
    └── basic_fps_logger/  # First working demo

---

🧪 Tested / Target Environment

Developed and tested with:

• Chip: ESP32-S3
• SDK: ESP-IDF (Linux / WSL)
• Environment: Windows 10 + WSL Ubuntu
• Tools: VS Code, idf.py, CMake

Validated hardware:

• ESP32-S3 dev boards with PSRAM
• OV2640 camera modules
• OV5640 (planned, not yet validated)

---

📸 Hardware Requirements

Target boards:

• ESP32-S3 dev board with PSRAM
• Camera modules:
  • OV2640 (supported)
  • OV5640 (planned)

Recommended specs:

• 8 MB PSRAM
• QVGA or VGA frame sizes for initial demos
• Stable 5V USB power

If using boards such as Seeed XIAO ESP32‑S3 Sense, adjust only camera_config_t in camera_node.

---

🖥️ Getting Started (ESP-IDF)

1. Install ESP-IDF

https://docs.espressif.com/projects/esp-idf/en/latest/esp32s3/get-started/

2. Clone the repository

git clone https://github.com/CahillMeyer/ccm-esp32-vision-node.git
cd ccm-esp32-vision-node/firmware

3. Configure the target

idf.py set-target esp32s3
idf.py menuconfig

4. Build & flash

idf.py build
idf.py flash monitor

---

📟 Example Serial Output (FPS Logger)

I (0) cpu_start: Starting scheduler on PRO CPU.
I (0) main: Initialising camera...
I (250) camera_node: Camera init OK (320x240, PIXFORMAT_JPEG)
I (260) main: Captured frame 1 (size=32768 bytes)
I (270) main: FPS: 18.7
I (323) main: Captured frame 2 (size=32768 bytes)
I (333) main: FPS: 19.5
...

---

📊 Early Benchmarks (Placeholder)

Resolution	Pixel Format	Pipeline Stages	Avg FPS	Notes
320×240	RGB565	None (raw)	TBD	Baseline
320×240	Gray	grayscale	TBD	First CV test
320×240	Gray+TH	gray+threshold	TBD	Planned
640×480	RGB565	None	TBD	PSRAM required
320×240	Gray + Threshold	~4.5 ms	~28	Host Sim
320×240	Gray + ROI + Downsample (2x)	~1.0 ms	~90+	Host Sim (80x60 effective)

Benchmarks will be updated as the pipeline matures.

---

🗺️ Roadmap

Near-term

• Add cv_pipeline grayscale + threshold stage
• Add ROI cropping + reduction steps
• Add MJPEG Wi‑Fi streaming server
• Add diagrams under docs/media/

Mid-term

• Add blob detection demo
• Add region extraction example
• Add JSON status endpoint
• Add host-side tests

Vision

A reusable, modular Embedded Vision Node for ESP32-S3:
Setup board → configure pipeline → get structured vision output via UART or Wi‑Fi.

---

Who Is This For?

• Embedded firmware engineers exploring ESP-IDF C++ patterns
• Robotics / IoT teams evaluating low‑power embedded vision
• Engineering leads reviewing CCM Code’s architecture + code quality
• Clients needing a reference implementation for ESP32‑based smart camera nodes

---

What Works Today (v0.1.0 – Alpha)

• ESP-IDF C++ multi-component project structure
• Camera bring-up via esp32-camera
• Continuous capture loop with real-time FPS logging
• Builds and flashes cleanly from WSL → ESP32-S3
• Early example demo under examples/basic_fps_logger

---

📄 License

MIT License — see LICENSE.

---

👤 Author

Christie Cahill Meyer
Embedded Vision Engineer
GitHub: https://github.com/CahillMeyer

---

🛠️ Support & Custom Integration

Building real-time embedded vision systems is complex. This repo is a starting point, but every hardware setup is unique.

• Stuck on integration? I offer a Priority Issue Review tier on GitHub Sponsors. I'll prioritize your bug reports and config questions. Get Priority Support
• Need a custom driver? Book a Technical Quick-Sync to discuss your specific sensor or board requirements.