Machine Vision Based Automatic Reading Recognition For Pointer Meter And Its Application

Mechanical instrument has the advantages of simple structure,easy maintenance and low price.It is widely used in oil exploration and oilfield equipment operation monitoring.It has been an important means to obtain mechanical equipment information and realize equipment condition monitoring.Generally,acquisition of instrument information is expressed by pointer indication or deflection value.Meter display result is obtained practically by manual reading or meter reading.The reading process is easily influenced by various subjective and objective factors,such as human eye resolution,observation angle,field light and various environmental factors.Manual reading not only wastes time and energy,but cannot guarantee an accurate reading as well,which is easy to produce errors.As a result,it may lead to an inaccuracy acquisition of mechanical equipment parameters.In this paper,through digital image processing technology,a method of automatic reading recognition for pointer instrument is proposed,combining Hough transform method.It includes three main steps: firstly,image preprocessing,it is to improve the signal-to-noise ratio of the image,including scaling and image correction,grayscale,smooth de-noising,binarization and edge detection.Among them,the contour tracking method is used to segment dial area and pointer area,and noise reduction is carried out at the same time.The second is the detection of pointer straight line,including the thinning of instrument pointer by using morphological method,and the pointer position detection and pointer direction determination are realized by Hough transform method.Among them,the circumscription method is used to determine the rotation center of the instrument.Finally,the effective range of an instrument panel is determined by the relationship between the angle and the scale line.The automatic reading recognition of the pointer instrument is completed.Through the GUI platform,the experiments and results analysis are accomplished.The detection error of pointer line and coincidence error of pointer and scale line are analyzed in detail.The corresponding solutions are proposed.The experimental results show that the accuracy of reading recognition of pointer instruments by Hough transform method is significantly higher than that of manual reading recognition.The automatic reading recognition method proposed in this paper is more accurate and reliable,with high reading efficiency,time-saving and labor-saving.It has obvious practical significance for the condition monitoring of oilfield machinery.