Research Data Processing Terminal Based On Image Recognition Pointer Instrument

Pointer instrument has been widely used in electric power, chemical and automation industry with the feature of simple structure, easy operation, etc. However, in the actual operation process, it needs to read the meter readings manually and will trigger a reading error in some special working environment, which is impeding the industry informatization. So we must complete the study of the pointer instrument readings’auto-recognition system so as to improve the reading accuracy and the work efficiency. And with the development and wide applications of image processing technology, it has become the research hotspot to combine this technology with pointer instrument automatic reading. This paper designs one pointer meter automatic reading system scheme based on image processing, which can not only improve the accuracy of reading the readings but also increase the efficiency of the readings.First of all, this thesis introduces the knowledge about the digital image processing related to the design, including the RGB model of the image, gray transformation, the image preprocessing methods, various edge detection algorithm and the theory of Hough transform. And it also realizes algorithm simulation on MATLAB software platform.Secondly, on the theoretical basis, this article proposes one general design scheme of pointer instrument’s reading automatic discriminant device based on the STM32chip. This paper introduces the hardware circuit design of this system in detail, carries out the modularization programming to the system software and shows the flowchart of every module. It also introduces the CPU circuit which takes the STM32F103RET6chip as the center, the images acquisition circuits based on OV7670camera modules and frame buffer chip AL422, the TFT screen display circuit and SD card data storage circuit, etc. On the hardware platform in a IAR development environment,the system completed with many design algorithm, the core algorithm are Sobel edge detection operator and an improved Hough transform linear detection algorithm.Finally, the correctness and feasibility of the design is tested and analyzed. After the test, the result proves that the design and use of the paper the algorithm can meet the requirements of automatic reading.