Reading Recognition Of Digital Instruments Based On Machine Vision

Digital instruments are frequently used in daily life and industrial and agricultural production,so it is imperative to improve their reading efficiency and accuracy.In this paper,a reading recognition system of digital instrument based on machine vision is proposed.The main work of this system is as follows:Firstly,in the design of image acquisition scheme,aiming at the problem of scalar map distortion caused by hardware equipment,appropriate equipment calibration and target image correction techniques are adopted to optimize the design;in order to obtain the original image of the highest quality,appropriate rectangular surface lighting and lighting scheme are selected,and image acquisition equipment that meets the requirements of the project is purchased.Then,for the collected image,the location error of the effective information caused by the inevitable deviation of the instrument placement,the image calibration method based on Hough transform is used to complete the image correction based on the information of the line segment in the target image;the image is preprocessed,and the redundancy of the information in the image is reduced by filtering,enhancing,intercepting and thinning the image.Finally,the experiment evaluates the accuracy and feasibility of the digital recognition method based on moment invariant information fusion proposed in this paper in the actual environment of the project,and proves that the algorithm designed in this paper meets the actual verification requirements of the project,and determines the whole work flow and algorithm scheme of the project.