Automatic reading recognition system for analog measuring instruments base on digital image processing

Abstract

Pointer instrument has been widely used in the electric power, chemical industry and other industrial fields, because of the advantages such as low price, high reliability, easy installation, replacement and so forth. The traditional manual reading method has its drawbacks for example process trival, big workload, lack of accuracy. This article, based on digital image processing techniques, studies the automatic reading recognition method for analog measuring instruments. After pretreatment of original image on the instrument dial, the pointer information will be split by binarization; The Hough transform is the main idea for line detection and then by the pointer straight line parameters and calculating the image pixel the angle and direction of the pointer can be got; Angle method is used to determine the reading of the meter pointer which is simple but effective. At the same time template matching method is used to recognize the range of pointer instrument automatically. The system software is designed by matlab on the platform of windows. By using MATLAB (gui) simulation, pointer meter reading recognition system can be constructed which includes control interface. © 2013 Asian Network for Scientific Information.