A Novel Approach for Detecting LED Light Blinking Patterns

With the ever-growing deployment of hundreds of access points (APs) with embedded Light Emitting Diodes (LEDs) in large facilities around the world, network operations become more frequent and the necessity to identify contextual network data becomes apparent. In this paper, we propose an approach to correctly detect a specific sequence of blinking LED lights embedded in APs under different conditions and long distances. This sequence of blinking LED lights represents bits that should be decoded correctly to determine the AP’s true identity and display the relevant data on a video feed. Many constraints emerge in the goal of detecting the message such as the limited number of bits that the LED driver can emit, in addition to fast detection at long distances. Current LED blinking identification approaches either require additional hardware attached to the detected LEDs or are limited to short distances or certain platforms. In this paper, we propose a cross-platform web-based lightweight solution that receives input from a mobile’s camera, superposes filters in real-time, and captures frame by frame at 32.04 frames per second. The system introduces a new image subtraction method for detecting messages emitted by the LED light at distances up to seven meters. After a correct detection of the LED sequence, the web application displays the AP identifier to the user in a friendly user interface.