Developed is a technique for optical character recognition of seven segment displays for industrial settings. A computer vision sensor module is implemented using the proposed technique and tested in an operational factory. Static display geometry to simplify analysis is provided by optical reference pattern and distortion/perspective correction. A web-responsive dashboard application allows the user to control the operation of the computer vision unit, view the data remotely, receive a preventive maintenance warning message, and aggregate data for later analysis. The results from factory deployment showing more than 97% accuracy under multiple lighting condition when the camera was not interfered with. The proposed IoT-based system is currently being used in the factory to gravitate towards the data-driven efficiency improvement by simply augmenting the existing system.