Skip to content

Developing Wildlife Camera Traps with Zephyr RTOS

Alex BucknallArribada Initiative

Blue TeamThe Linux Foundation21 minOpen on YouTube ↗

Details developing Wildlife Camera Traps with Zephyr RTOS and evaluates key implications for security practitioners.

Summary

### Overview In this presentation, Alex from the Arribada Initiative—a non-profit organization specializing in open-source technologies for conservation and social good—discusses developing wildlife camera traps and monitoring systems using Zephyr RTOS. The talk highlights an initiative undertaken with the government of Bermuda to investigate severe seagrass decline. Bermudian authorities had initially hypothesized that green turtles were overgrazing juvenile seagrass meadows, but protective exclosure cages failed to halt the loss. To uncover the true cause, Arribada designed low-power, field-configurable underwater time-lapse cameras equipped with environmental sensors to monitor seagrass dynamics, light levels, and water conditions.

### Key Topics - **Marine Conservation Context**: Seagrass meadows cover about 0.1% of the seafloor, sequester roughly 10% of ocean carbon, produce significant oxygen, and stabilize coastlines, yet at least 29% has vanished over the last 50 years. Understanding the failure of Bermuda's seagrass recovery required in-situ continuous monitoring. - **Modular Hardware Design**: Developing modular, inexpensive platforms based on Adafruit Feather form factors, standard interconnects (STEMMA / Qwiic), and swappable microcontrollers (ESP32, RP2040, NRF52). - **Zephyr RTOS Adoption**: Leveraging Zephyr's hardware abstraction, Devicetree overlays, sensor APIs, and shell subsystems to accelerate engineering for a small five-person team. - **Field Operations and Deployment**: Streamlining firmware delivery, remote configuration, and maintenance in challenging, low-connectivity environments.

### Technical Details - **Hardware Architecture**: The underwater camera system is built around an Adafruit Feather-compatible board and connects over I2C to RTCs and environmental sensors. Light is monitored with a Lux sensor, and water temperature is recorded using a Maxim DS18B20 sensor chosen specifically for its waterproof enclosure. To eliminate large internal RAM frame buffer requirements on the host microcontroller, the design integrates an Arducam Mega camera, which features its own internal frame buffer that the host MCU polls. - **Driver and Software Abstraction**: Arribada partnered with Arducam to develop a Zephyr driver for the Arducam Mega platform, submitted as an upstream pull request. By using Zephyr's driver abstractions, the team cleanly separates application business logic from low-level driver development, allowing external contractors to write drivers while core engineers focus on client requirements. - **Configuration and Overlays**: Systems use Zephyr Devicetree overlays as a pseudo-hardware bill of materials to quickly adapt the codebase across different microcontroller boards and sensor configurations. Interactive Zephyr shells over UART allow field operators to configure RTC timestamps and verify peripheral connections without recompiling firmware. - **CI and Flashing**: Firmware binaries are compiled in the cloud via GitHub Actions using Zephyr's Docker CI images, providing field technicians direct access to versioned builds. For non-technical operators in remote areas, UF2 bootloaders (drag-and-drop over USB) proved significantly more reliable than complex command-line flashing tools like OpenOCD.

### Challenges and Takeaways - **Vendor Support Gaps**: Ecosystem maturity varies across chipsets. The team encountered missing power management support for the ESP32-S3, forcing a switch to another platform. - **Steep Learning Curve and Manifest Drift**: Zephyr involves a steep initial learning curve, and multi-repository West manifest dependencies can exhibit pinning inconsistencies. - **Ecosystem Impact**: Ten cameras were successfully deployed in Bermuda with funding from the Darwin Initiative. Zephyr's open-source modularity enables small engineering teams to scale technology across diverse conservation projects, ranging from penguin nest cameras in Antarctica to avian satellite tracking tags in Ethiopia and seaweed monitoring in the Philippines.

Related talks