Research On Inspection Technology Of Parts Dimension Based On Machine Vision

As the standard of industrial level,inspection technology is widely used in machinery manufacturing.In recent years,the application environment of inspection technology is poor and the precision requirement is high.Visual non-contact inspect technology could solve the problem of poor environment,such as high temperature,dust,etc.With high efficiency and high precision,it become a hotspot of businesses and universities.The paper designed the system of binocular visual inspection and aimed at aluminum block production under the condition of high temperature.In theory,the paper established a mathematic model of binocular visual system and provided the basis for calibration and 3D inspection technology.Then it inspected the image edge through image graying,filtering,segmentation algorithm.Finally,it achieved the fitting and 3D reconstruction of line or curve.The experiment grabbed images by the high precision CCD industrial cameras,used the chessboard target as a calibration object,programed to design a variety of image algorithm through Visual Studio 2010 software.And the binocular visual calibration and binocular measurement were completed.At last,a series of experiments were done and the illumination of the detection environment was proposed.The paper showed that the image processing algorithm was reasonable and the experimental results met requirement.The paper used the passive visual inspection method.The detection accuracy was improved by improving image detection algorithm.The multi view image matching was solved through direct 3D reconstruction of line or curve on the plane.The method was proved to be feasible.The successful design of detection system provided a solution for actual production application and also laid a foundation for the further optimization of visual inspection.