The Application Of Machine Vision In Instrumentation Monitoring Recognition System

With the development of domestic and foreign intelligence surveillance industry and the security industry, machine vision increasingly used in many aspects of human behavior and expression recognition, PCB printed circuit detection, identification numbers and pointer instrument, product appearance inspection, logistics and goods classification, machine vision replace artificial vision into the various fields, greatly improving the efficiency of production and life of people. In the industrial field, various instruments such as sound level meters, noise dosimeters, vibration meter, pressure gauges and other play a significant role in the production, showing the value of these instruments are usually read manually, low efficiency, subjective relatively large, easy to produce misunderstanding, using machine vision to identify the instrument represents the value of great application value. Based on the above situation and problems, this paper studies the application of machine vision recognition system monitoring instrumentation.Paper introduces the significance of research and development status monitoring of machine vision recognition systems and instrumentation, and then on the various components of the machine vision made a detailed presentation and selection methods are summarized, and then were introduced this topic hardware and software platforms, and finally software design completed pointer instrument identification numbers and reading recognition, and recognition results summarized.Recognition by monitoring several industrial applications and digital pointer instrument, the subject of the German industrial Basler camera with high-quality performance of acquiring an image, by means of Microsoft’s MFC, OpenCV, pylon SDK software development tools for system design.Where the focus of this issue is the instrument identification algorithm design. The method of using machine vision system first image acquisition instrumentation, and then using digital image processing techniques for image pre-processing operations (carried out using the weighted average method graying, using the histogram equalization for image enhancement, locally adaptive binarization, treated with morphological opening and closing operation); thus the value of information extraction area according to the position shown on the digital instrument dials, length, aspect ratio, and other exterior features the outline of the instrument using the Hough transform to extract pointer circular outline, positioning the pointer area; and finally the use of digital image segmentation techniques and pointers split out using a custom template matching method to identify a character, line Hough transform to detect and calculate the indicator readings.Tested under the actual scene, recognition speed and recognition accuracy of the system are able to meet the requirements of the application, with good value.Breakthroughs and innovations made part of this project are as follows:1. strict accordance with machine vision methods proposed selection of lighting, optics, industrial surveillance cameras in the instrument identification system selection rules or guidelines;2. Basler industry through research and analysis of video storage format camera, successfully found converts Mat format YUV422 format used OpenCV method using OpenCV foundation for digital image processing;3. design a feature extraction method-based on the relative position of the object contour extraction method, partial length, aspect ratio, with an area to achieve a localization region of interest.4. The pointer instrument identification number are shown in the perspective method used to read the pointer location, i.e. according to the angle pointer coordinates of the two end points of connection with the horizontal direction, the dial to identify the minimum and maximum readings.