Indication Development Of Machine Vision Inspection System Of Analog Measuring Instruments

Pressure gauges as an important tool for people understanding the world, have playedan important role in the industrial production. It has a great significance on people’s life andindustrial production no matter in the heat transmission network, natural gas transmissionnetwork or other transmission system. Considering from safety explosion-proof, some scenecannot be installed the pressure sensors, and only used mechanical pressure gauge forpressure monitoring.Existing methods for pressure gauge calibration are sensitive to the relative positionbetween the camera and the dial plate, a recognition algorithm is proposed based onmachine vision and a variety of image processing methods, which includes imageacquisition, camera video frame, the dial positioning, binarization algorithms, connectedcomponent analysis, scale extraction, long scale analysis, division value calculation, LCDdigital positioning, LCD digital recognition. Firstly, the pressure gauge which is to beidentified and precision digital meter are placed under the same pressure environment,making vertical projection for the images which using camera to take at real-time, andpositioning the pressure gauge for the first time according to the results of verticalprojection. Secondly, the pressure gauge to be identified is processed by circular rectangulartransform which is converted from Circular coordinate system to Cartesian coordinatesystem after positioning. And a binarization algorithm which has good anti-interferencecharacteristics and scale features efficiency extraction is proposed according to thecharacteristics of scale. Thirdly, digital is recognized by the method of template matchingafter extracting the number in the scale line, and pressure gauge degree can be accuratecombine the related processing result of calibration. Fourthly, the border of LCD charactersare searched by the vertical templates on the gray scale, then accurate position of LCDcharacters is positioned according to four border location information and each LCDcharacter is recognized by threading method. Finally, pressure gauge degree is contrasted tothe recognition result which has been calculated as the complete degree, and the result curveis drawn in order to recognize easily.Experimental results show that the proposed method, which has high recognition speedand high recognition precision can recognize the degree of pressure gauge effectively and can resist deflection angle between the lens and the dial effectively.