Back to Blog

Miniaturized Low-Power AI Camera Based on RK1126, Supporting Human Attribute Recognition, Facial Feature Detection, Hard Hat Detection, and More

#人工智能#数码相机

We offer a customizable AI camera hardware and software development platform.

The hardware is built around the Rockchip RV1126 processor solution, compatible with Sony IMX series image sensors, supporting PoE (Power over Ethernet) for both power supply and data transmission. Designed with a standard 38-board structure, it enables rapid, customizable enclosure design, supporting the full process from development and verification to mass production. Horus provides a complete SDK with a foundational camera platform, including system software, drivers, image processing, and AI application interfaces. It supports multiple programming languages such as Python, C, and C++, significantly accelerating the development cycle from AI algorithm prototyping to hardware deployment.

Facial Feature Recognition

Perform multi-dimensional facial feature recognition on detected subjects (e.g., on-duty personnel, field workers), such as fatigue level, alertness, smoking, yawning, and phone usage, to comprehensively assess their working status. The system can upload status data to a central server or trigger on-site alerts to notify individuals and prompt them to adjust their behavior accordingly.

Human Attribute Recognition

In high-traffic public areas (e.g., retail stores, exhibition halls, service centers), perform multi-camera, multi-position trajectory tracking. The system compares captured faces against a pre-registered facial database to identify customers versus strangers. It also analyzes and correlates additional human attributes (e.g., clothing and accessories) across different zones, enabling comprehensive customer journey analysis. This data helps optimize store layouts and improve operational efficiency.