Design And Implementation Of Image Measurement System Based On Machine Vision

With the continuous development of computer application technology,image processing technology and precision measurement technology,machine vision technology has been widely used.The research goal of machine vision is to capture one or more images,and then use computer technology to perceive geometric information such as the shape and position of surrounding scenes.It is a three-dimensional perception of the objective world.The application of machine vision technology to the measurement of spatial geometry has formed the concept of image measurement.Image measurement is based on machine vision.Image measurement is based on machine vision.The high-speed image acquisition system of highresolution camera processes the image and performs high-performance calculations to obtain the actual distance of the three-dimensional scene in the image.With the continuous improvement of camera resolution,image acquisition performance and computer processing capability,visual equipment based on monocular,binocular or multi-purpose has been widely used in the fields of autonomous driving,artificial intelligence,aerospace and other fields.There are also many applications of applying visual technology to the field of image measurement.However,the current application based on image measurement technology is specific to a certain measurement function and the equipment cost is expensive,and it is awkward and cumbersome to use.At this stage,there are few studies on the problems in daily life that are difficult to measure by conventional methods and cannot be measured by means of image measurement techniques.In this paper,a simple camera device is used to realize the planar image measurement of monocular vision and the three-dimensional image measurement of binocular stereo vision by using machine vision technology and image processing technology.Monocular vision image measurement mainly studies the method of measuring the planar object after automatically correcting the calibration reference object by correcting the image tilt distortion.The three-dimensional measurement of binocular stereo vision is mainly through the design of binocular camera system with variable baseline distance for image acquisition and measurement.The key method of camera calibration,image matching and threedimensional measurement method is studied.According to two different measurement methods,a set of image measurement system based on machine vision is designed and built,which has certain practical value.Main tasks as follows:1.Aiming at the characteristics of two-dimensional images acquired by monocular vision,the traditional image processing methods and image tilt correction are studied.Aiming at the two image tilt problems,this paper designs a general improved image correction strategy.Aiming at the problem of reference setting and calibration,this paper designs a method based on Hough transform circle detection to automatically identify the reference object and perform pixel/scale ratio calibration,which improves the automation and measurement efficiency of image measurement.2.Aiming at the three-dimensional measurement problem of binocular stereo vision,this paper designs a method of camera calibration quickly through image measurement system,which improves the efficiency of camera calibration.A variable baseline distance binocular camera system is designed to meet the measurement needs of different scenes and improve measurement accuracy.Aiming at the problems existing in traditional binocular vision research,a key method for selecting human-computer interaction and convenient use of image measurement system is designed for the problem of large amount of key point matching calculation and high mismatch rate.According to the two images obtained by binocular photographing and known information such as camera focal length and baseline distance,three-dimensional measurement of object size in space is realized.3.Designed and built a set of image measurement system based on machine vision.The main functions of monocular measurement include monocular photographing and correction,automatic identification and calibration of reference objects,measurement of plane size,etc.The main functions of binocular measurement are automatic camera calibration,binocular photography,quick selection of image matching key points,three-dimensional size measurement,etc.In addition,the local operation mode is also designed,which can measure the collected images for easy use.The general method of image tilt correction designed in this paper and the monocular vision measurement algorithm based on Hough transform circle detection to automatically identify the reference object and perform pixel/scale ratio calibration provide a new idea for monocular vision plane measurement.The manual selection of measuring points and measuring point matching methods in the binocular stereo vision image measuring system is designed to effectively reduce the matching calculation amount,thereby improving the efficiency and accuracy of binocular visual image measurement.The research on single and binocular visual image measurement promotes the development of image measurement research,making image measurement technology more convenient and suitable for more scenes.