The histogram-enhanced Hough transform applied to automated readout of analogue gauge meters using digital image processing

摘要

The paper presents a computer vision technique developed for automated reading of analogue gauge meters. The system utilizes a mobile phone digital camera for image stream capturing and the Hough transform supplemented by multiply histogram analysis for locating a pointer position and parallax effect correction. The system is capable of working with minimal user supervision and its indicator position detection uncertainty is superior to that provided by a human eye. An attempt to use the algorithm implementation for semi-automatic calibration of an analogue voltmeter is also presented.