Research On Detection And Image Processing Technology Of Ultra - Smooth Surface Flaws

This paper mainly did research on the special conditioning equipment of the ultra-low loss gyro’s optical cavity, which is commissioned to develop by Chinese Academy of Aerospace Electronic Technology (CAAET). Its essence is to design a set of surface defects detection system of ultra-smooth surface precision components. For the image processing technology in this system, this paper conducts a study of the following aspects:We set up a CCD image acquisition and computer processing system based upon the principle of micro scattering dark field imaging. It completes the full range defect imaging of tested components’ ultra-smooth surface by 360 degree rotation and the scanning imaging of large aperture elements to be tested. As a result, we obtain the sub-aperture flaw images. This paper studies the fast translation stitching algorithm. This paper studies another stitching algorithm based on the correlation of the flaw feature of images’ overlap region. For the inverse rotation of full-aperture flaw images, a rotation algorithm which uses an arbitrary point of image as the center of rotation is studied. Finally, the full-aperture flaw images are superimposed by or operation. A series of preprocessing operations are used in the flaw images, such as noise filtering, binaryzation enhancement as well as the mathematical morphology operation. Then study the recognition algorithms respectively of linear defects, block defects and point defects. Considering of the client’s requirements of locating the defects’ area to determine the locations free from defect, this paper presents the method which uses minimum enclosing rectangle and pixel equivalent calibration to realize the positioning of the block defect region. All the image processing algorithms are simulated by MATLAB to verify the feasibility of the algorithms. This move not only can find out better algorithms through analysis and judgment, but also can lay a solid foundation for the design and optimization of the follow-up software of the surface defects detection system.