Stealth Light: Encoding Metadata with Infrared Projection in Physical Environments

As unauthorized photography continues to pose increasing privacy and security threats in both public and restricted spaces, this work proposes a novel system that leverages information hiding through Near-Infrared (NIR) light projection to address these concerns. Building upon the principle that infrared light is invisible to the human eye but detectable by most digital imaging sensors, the system modifies commercially available projectors to emit structured NIR patterns in real-world environments. These projected patterns are imperceptible to attendees, yet encode dynamic metadata, such as timestamps, location code, and other identifiers directly into the physical scene. When a photograph is captured within the affected environment, these embedded patterns are also captured by the camera sensor, allowing for post hoc extraction of hidden information through image-processing techniques. The encoded data remain unobtrusive under normal lighting conditions and have no visual impact on the scene, ensuring that it does not interfere with human experience or aesthetics. This work presents a proof-of-concept prototype to validate the approach, including the hardware modifications, encoding strategies, and image processing pipeline necessary for decoding. We experimentally evaluated the detectability and reliability of infrared watermarks in various settings, including types of encoding patterns, distance, and lighting conditions. Furthermore, a flexible framework is proposed to guide real-world deployment in sensitive or controlled environments such as museums, theaters, laboratories, and secure facilities, where controlling or tracing unauthorized image capture is essential. By seamlessly embedding invisible, trackable information into physical spaces, this approach offers a promising new tool for digital forensics, access control, and privacy-preserving surveillance.