Key Technology Research In Spherical Optics Surface Defects Evaluation System

With the development of science and technology, the spherical optics develop towards the characteristics of large scale and high precision to satisfy the requirements of deep space exploration, energy access, remote sensing measurement and control. At the same time, higher demands have been put forward on the surface quality of the spherical optics, which contains surface shape, roughness and defects. Relative key indicators such as surface shape and roughness can be measured and controlled with very sophisticated scientific instruments such as digital interferometer and profiler, however digital detection of surface defects is difficult to achieve so far. Nowadays in the field of surface defects, digital defect detection of plane components is still at the initial stage and that of spherical components is faced with more problems due to the variability of optical structure parameters such as curvature and shape. Thus defect detection on the surface of spherical optics, especially the ones with large apertures, is mainly limited to the human visual approach, which holds the disadvantages of strong subjectivity, low efficiency and difficulties in location and quantitation. These problems severely restrict the development of scientific equipment of inertial confinement fusion, space optics and ultra-precision advanced optical manufacturing technology, so it is in urgent need for relative advanced technology to achieve digital detection of surface defects on large aperture spherical surface.Based on the microscopic scattering dark-field imaging method, this dissertation investigates spherical surface defect detection of different curvature, shape and aperture, and mainly study in microscopic scattering dark-field illumination, sub-aperture scanning path planning algorithm and 3D reconstruction technology for spherical surface defects. The main research contents include:The importance and necessity of the research for ultra-precision defect detection technology and system on the surface of spherical optics is discussed in combination with the modern ultra-precision machining technology for large aperture spherical optical elements and the domestic and international research progress for surface defect detection based on the machine vision method, The surface defect detection method for spherical optics is proposed based on dark-field scattering illumination on the foundation of the surface defect evaluation system (SDES) for plane optical elements, The special illumination sources and centering system for spherical optics used in this method is introduced in detail, which lays the foundation for scanning and imaging of the spherical elements.The automatic 3D scanning path planning algorithm for spherical optics is proposed, and the SOM (Same-Overlapped-area on Meridian) and SOP (Same Overlapped-area on Parallel) models are established. Based on different overlapping factor simulation for the two proposed models, the optimal SOP scheme is obtained by comparing through three different methods, which ensures the fast detection for the whole spherical aperture without any miss. At the same time, taking advantages of the mapping relation of 3D sphere imaging and combining with curvature radius of test elements, the 3D sub-aperture reconstruction algorithm based on Pin-Hole model is established, through which,2D sub-aperture plane images are reversely reconstruction to obtain 3D sub-aperture spherical image. And based on 3D reconstruction, the 3D sub-aperture straight stitching can be realized and full-aperture defect image of 3D sphere can be obtained in combination with the SOP planning scheme.The error analysis about the multi-axis linkage scanning system is done to solve the feature rupture problem encountered in the results from the 3D sub-aperture straight stitching algorithm. Based on the multi-body system theory, the perfect multi-axis linkage scanning system as well as the practical one containing error sources is established. The influence from the error terms in the simulation model on the stitching results is analyzed, and the error distribution for the system hardware is optimally estimated by simulating the practical sampling process to ensure the straight stitching accuracy.The theory model of the latitude annulus best-matched sub-aperture stitching algorithm is proposed to solve the scratch rupture problem in the spherical surface measurement for large optics. The stitching process is divided into local matching between adjacent sub-apertures in latitude annulus and global matching and adjustment of sub-apertures between different latitude annuli. The stitching precision of 3D spherical full-aperture is ensured by that of the projective images of those 3D spherical sub-aperture images.The machine-vision detection system scheme of large-aperture spherical surface defects on basis of microscopic scattering dark-field imaging is firstly proposed and principle experimental system is established on the foundation of theoretical research. The system adopted high-brightness annular zoom LED lights to illuminate the spherical surface. On precondition of avoiding the scattering light influence of spherical surface, scattering light rays induced by surface defects pass the imaging system and dark-field sub-aperture imaging with bright defect image and dark background can be obtained. Multi-axis linkage scanning system combined with optical self-collimation spherical centering system to scan sub-apertures of the spherical surface, so that sub-aperture sampling covering the full aperture can be achieved. Eventually, experiments on scanning, imaging and stitching of spherical elements are carried out and data processing procedure is elaborated in order to verify the validity of scanning model and reconstruction model and the accuracy of spherical stitching method.