Research On Measurement Methods Of Small Objects Based On Visual Attention

Along with the rapid development of modern manufacturing technology in the field of micro mechanical,micro parts more and more applications of microscale objects,such as various fields for measuring small size precision demand more and more high,the traditional artificial measure method is time-consuming,laborious and accuracy of measurement is not stable.Measurement method based on image processing with non-contact,high precision,high sensitivity and other unique advantages,in the field of industrial testing plays a more and more important role.But under the microscopic visual detect tiny components,there are small field of vision,high detection precision,lead to a measurement of small target,based on the study of visual attention mechanism,the feature extraction algorithm,feature matching algorithm,microscopic vision of camera calibration,machine vision field synergy measuring size,etc,on the basis of the relation between image coordinates a micro-scale object oriented view more collaborative measurement method.(1)multiple cooperating field measurement method is first collected image,secondly through visual attention mechanism to establish brightness figure significantly the great view of the low resolution images to extract interest area,respectively inspects a field of interest area to obtain high resolution images.Then through the large field of view image feature extraction and feature matching algorithm and the integrity of image match,will remain a image fusion in the large field of view image.The final image processing to get high resolution image and returns the data of need to measure.So design the camera fixed organization,3 d displacement of the mobile platform,based on the MATLAB visual attention mechanism and image processing system based on halcon of multiple cooperating field measurement method.(2)the study of visual attention mechanism in the application of the field measurement.From machine vision task displayed significant testing according to the demand of technology of visual attention is studied.First research,several kinds of visual attention mechanism is analyzed,and then puts forward a kind of visual attention scheme based on spectral residual,the screen image of wide field similar to the attention of the human eye,find the need for DE minimis field imaging area.Results show that the proposed scheme can accurately detect the need of imaging area significantly.(3)research on feature extraction and matching techniques in computer vision.Study the coordinates relationship of the different view,different feature extraction and feature matching algorithm,using the algorithm of the ORB.The ORB algorithm because of its good real-time performance,therefore has been widely used.This topic on the basis of the ORB algorithm,joined the gaussian scale space structure,enables the algorithm to the scale of transformation,still can fast detect feature points accurately.(4)Finally studied visual image processing technology in the measurement.First research,analyses the commonly used image binarization,edge detection operator and morphological image processing,image enhancement operator is proposed and then to improve the image contrast,the watershed algorithm is adopted to adhesion edge segmentation,using canny operator to extract edge detected each separate the target ball,and then to extract the target ball diameter,roundness,bad measurement and filter,finally obtain qualified micro target center coordinates of the ball.Analysis of experimental results,the experimental results show that the method has high accuracy of measurement,and can realize automatic measurement for micro-scale object.