A camera that understands indicator language: wire-free diagnostics for industrial automation

ndustrial controllers "communicate" with operators through LEDs: the color and blinking pattern of indicators reveal operating modes and fault conditions. However, modern automation modules may contain dozens of such indicators, making manual interpretation difficult. Traditional diagnostics via the controller network is not always possible either, as some equipment is isolated for security reasons, while other systems cannot be accessed without shutting down the production line.

IKSAR is testing a prototype of an intelligent computer vision solution that automatically detects industrial equipment faults by analyzing the information displayed through colored status indicators.

The solution can be used by both robots and human operators.

Simply point a camera at an indicator panel, and the algorithm instantly provides a semantic interpretation of the equipment's condition. The system identifies the required module in the image, recognizes the color of each indicator and its blinking pattern, and converts this information into a clear diagnostic result.

The solution operates entirely offline and does not require access to the controller network. A standard video stream from an ordinary camera at virtually any viewing angle is sufficient. This makes the technology suitable for AR glasses, smartphones, and fixed cameras alike.

For robots, physically connecting to diagnostic ports is extremely challenging, requiring precise mechanical alignment and dedicated connectors for every equipment type. Pointing a camera at a control panel and interpreting it the same way a human technician would is a far more natural approach. In essence, the solution gives robots the vision of an experienced diagnostician, removing one of the key barriers to autonomous maintenance of industrial automation systems.

For human operators, the same principle is already available today through AR glasses. Simply looking at a module displays the diagnostic result directly in the user's field of view. Operators keep their hands free, no physical connection to the equipment is required, and production lines can be inspected while continuously monitoring equipment status. The same approach also enables remote maintenance, where an operator only needs to show the equipment through a camera. The system is hardware-agnostic and can be configured for equipment from different manufacturers.

During testing, the solution achieved approximately 95% accuracy in recognizing indicator states.

Today, an ordinary camera is becoming a universal diagnostic tool—empowering both human operators wearing AR glasses and the autonomous industrial robots of tomorrow.