Research On The Automatic Reading System Of Dial Instrument Based On Machine Vision

Dial instrument is one of the main instruments in petrochemical enterprises, being used in displaying the running state of petroleum pipeline. Dial instrument can not be replaced by digital instrument in the short term. The traditional reading method is to use human eye readings. The method has a great workload, a limited accuracy. The method goes against the demand of detecting the running state of petroleum pipeline the whole day. The automatic reading system of dial instrument based on machine vision acquire the reading without touching dial instrument,is with high accuracy,can achieve automatic detecting without people participation in long distance.Paper analysis the hardware system structure based on machine vision and the operating principle, selection guide. Paper chooses the lens, acquisition card, camera and lighting system. Paper analysis and compare image filtering, image enhancement, threshold and morphology processing and so on, through analyzing contrast experiment, choose the suitable algorithms. Paper analyzed several kinds of line extraction method, gave the traditional Hough transform two constraints, increase the location accuracy. Use the dashboard scale sector characteristics, combined with pointer linear equation, Paper design an algorithm of self-adaption finding pointer rotation center. Through the circle center, extract the scribed line on both ends of the dashboard, combined with the pointer line, using a function of angle reading gets the primal indication of the meter. Then according to the relation of scale and pointer, obtain the last indication. The whole algorithm has robustness, high degree of automation, high precision, short processing time.At last, paper test the algorithms in the vc++6.0platform based on the OpenCV. Paper develop the software system, the performance is verified by many experiments.