Aiming at the problems of low accuracy, low positioning accuracy and low efficiency of feature points extraction in existing binocular vision positioning methods, and classical Harris corner detection algorithms depending on the selection of experiential threshold and the setting of corner points extraction, a corner detection algorithm based on PSOHarris operator to detect the double threshold is proposed in this paper, which calculates the optimal threshold by dynamic iterative optimization of the threshold. On this basis, the feature point parallax is calculated to match the image features, and the coordinates of the world coordinate system are calculated according to the coordinates of the target point. The comparison experiments show that the algorithm effectively improves the positioning accuracy and efficiency of feature point extraction.

The micro-grooving technology/micro-ruling technology is a special and efficient manufacturing technology and generally used to machine the cylindrical surface and micro grooved structures array. This paper investigates the fabrication of different types of micro structures by ruling. Experiment is conducted to machine a multi-layer micropillar array with a sharp diamond cutting tool by micro-grooving. Experimental results show that micro-grooving technology is feasible for the machining of multi-layer micropillar array with submicrometer form accuracy.

Interferometer is a powerful tool for optical surface measurement, including figure and roughness, due to its nanometer accuracy and non-contact manner. Traditional phase-shifting interferometry (PSI) is much sensitive to environmental vibration that impairs its application in measurement in workshop or on machine. Based on the iterative algorithm that is tolerant to phase-shifting error caused by vibration, two interferometers are developed to measure the optical surface figure and roughness respectively. A laser interferometer, of which the aperture size is 150mm, has been built to measure the surface figure. Practical test demonstrates that the laser interferometer achieves accuracy better than 5nm under vibration of 0.4 micron-amplitude over a large frequency range, 0-35Hz. And an interferometric microscope has been proposed to measure the surface roughness and verified to be effective. The measuring area of the microscope depends on the employed interference objective, and a typical value is about 1 squared millimeter. The error of measured roughness (Sq) under vibration, 0.4 micron-amplitude and over 0-20Hz frequency range, is less than 0.5nm. The developed method and instruments could be applied to optical surface measurement in vibration. The study relaxes the requirement of interferometers on environment and predicates an in-workshop or on-machine solution for optical surface measurement.

Surface defect is a concerned aspect of surface integrity in ultra-precision machining. The dark field microscopy method is efficient in detecting surface processing defects, and has certain development potential. However, it still has deficiencies on reliability, certainty and cross-scale adaptability. In this paper, several kinds of dark field illumination modes are compared, and high contrast and adaptable illumination modes are defined through experiments. Then a defect detection device is designed, which can detect the surface defects of opaque or transparent components by using the dome light source illumination. To enlarge the field of view (FOV), an X-Y scanning stage is used to obtain sub-area image of the surface, and a stitching method based on feature registration with SURF (Speeded Up Robust Features) is also proposed in the manuscript. Researches show that the defect detection device designed in this paper can obtain detailed, high-contrast, and wide range dark field defect images; SURF registration is insensitive to image translation, rotation, scaling and image noise, and has high calculation speed, which can relax the requirement of image stitching on positioning device and environment. This study provides an effective and low-cost solution for defects detection over large-scale surface in ultra-precision machining.

This paper mainly focuses on the impact of clearance joints on morphing wing mechanism and the way to alleviate such influences. Firstly, a new kind of multiple-closed-loop mechanism for morphing wing is proposed. The dynamic modeling approach is introduced with clearance joints in mechanism. Several analyses are conducted for the impacts of clearance joints and significant conclusions are obtained for the tolerance allocation. The improved artificial fish-swarm optimization algorithm is proposed for the optimization of alleviating the effects of clearance on morphing wing mechanism. Finally, a prototype is produced and experiment is conducted to verify the efficacy of optimization.

The breakthrough of diffraction limit in optical measurement and processing is always an important area. In near-field optics, the resolution is not limited by the diffraction limit, it can achieve ultra-high resolution imaging and surface nanostructure processing, and has important applications in various fields. For the near-field measurement and processing, the local field enhancement effect at the tip is the major factor affecting the spatial resolution. In this paper, the influence factors of the local field enhancement effect of the gold-coated nanoprobe are investigated by using finite difference time domain method, and the effects of incident light source wavelength, incident angle, tip-substrate distance and gold film thickness on the local field enhancement effect are analyzed, so as to provide theoretical guidance for improving the spatial resolution and field enhancement of near-field measurement and processing.

KDP (Potassium Dihydrogen Phosphate) crystal is widely used in inertia control fusion and other high-tech fields as a high quality non-linear optical material. Result of the surface quality correlating with the crystal optical properties, detecting surface defects on polished device is an essential part. In this paper, optical microscopy and image processing technology are used in the KDP crystal surface defects detection after MRF and surface cleaning. Firstly, the surface image is acquired by optical microscopy. Uneven illumination exists in the surface image, so the background extraction technology is presented to eliminate the impact of uneven illumination on the defect extraction. 2D maximum entropy threshold segmentation is applied to extract the defects. To identify residues and scratches defects, the features are utilized including the irregularity of residual defect edge, the linearity of scratch defect edge and the residue defect attached to the scratch defect shows the discontinuities and the curvature on straight edge. According to the features, canny operator is used to extract defects edge and improved straight recognition algorithm by freeman chain code is used to detect completed information of residues and scratches. Finally, the scratch defects are counted with width and length and using area to get a statistical result of the residue defects. Experimental results show that the method can accurately detect KDP crystal surface defects in different states after polishing and cleaning.

Silicon wafers with micro patterns were evaluated as one of the promising molds to fabricate microlens arrays for its high hardness. This study presents an efficient yet flexible manufacturing method for microlens array silicon mold by a polishing method. Unlike conventional processes for microstructures on silicon wafers such as diamond machining and photolithography, this research demonstrates a low-cost and high efficient chemical mechanical polishing (CMP) process with steel balls and diamond slurries for precision microlenses manufacturing. During the CMP process, polishing parameters for each micro cavity need to be accurately calculated and controlled to obtain microlenses with specific apertures. Therefore, a micro wear model for micro cavity CMP process was established to calculate cavity sag height with the knowledge of down force, polishing time and relative velocity between ball and silicon wafer. Several groups of microlenses polishing were then conducted under the same conditions to validate the micro wear model. Guided by the micro wear model, a 5×5 microlens array was fabricated on silicon surfaces. The shape accuracy and surface texture of the microlens arrays were evaluated by using a white light interferometer. This research demonstrates an alternative lowcost and efficient method for microstructure fabrication on silicon wafers and possible follow up optical molding processes.

With the development of integrated circuits, MEMS devices and biotechnology, people are demanding more and more for the detection of complex micro-structures. The diffraction limit of light restricts the resolution of traditional optical microscopes. Near-field optical microscopes can avoid Rayleigh criterion and break through the diffraction limit by detecting the details of objects with evanescent waves to achieve super resolution measurement. The minimum resolution of 60nm can be achieved by balancing the light throughout and aperture size of the aperture SNOM. While, the resolution of the scattering SNOM depends on the curvature radius of the tip of the probe. The resolution below 20nm can be obtained, but the signal can be extracted by a composite interference optical system and phase-locked technology. People have continued to pursue the development of near-field optical microscope with more convenient, more reliable and smaller resolution. In this paper, a surface plasmon probe with the combination of aperture and scattering is presented. The structure is shown in the attached drawings. On the basis of a commercial AFM probe, a composite probe based on the combination of surface plasmon enhancement and scattering probe focusing was formed by coating SiO 2 probe and etching a single ring. The structure of the probe was optimized, in order to achieve larger enhancement of the light field. Furthermore, the combination of aperture SNOM illumination can greatly suppress background noise and achieve higher signal-to-noise ratio. By these two technologies, interference optical system and phase-locked technology can be avoided, thus simplifying the design of SNOM instruments. The finite-difference time-domain method is utilized to simulate and calculate the field distribution of the focusing spot and optimize the microstructure of the excited surface plasmon, which provides a strong theoretical support for the probe fabrication.

It is of great importance to utilize a model to characterize the surfaces while designing an optical imaging system with freeform surfaces. For this purpose, a model with radial basis function based on surface slope (RBF-slope) for optical freeform surfaces was investigated by establishing the relationship between the shape factor and local surface slope, and improving the distributions of the basis functions for circular apertures. We performed the fitting experiment for “bumpy” paraboloids; the results demonstrated that the RBF-slope model has stronger fitting ability than conventional RBF model. A prototype of single mirror magnifier for head-worn display was designed and fabricated, in which the freeform mirror was described and characterized using the RBF-slope model. It can be proved by the design results that the RBF-slope model for optical freeform surfaces has obvious advantages in aberration balancing over conventional model. The primary experiments showed that the freeform surface was well fabricated and expected image display can be achieved.

Polarized LIDAR is defined as radar system which uses the polarization state of the laser to carry information. According to the different features, this paper divides the polarized LIDAR into non-imaging polarized LIDAR and imaging polarized LIDAR. The non-imaging polarized LIDAR can distinguish spherical particles and non-spherical particles by using the particle back polarization，which is used to study dust, haze, fog and other atmospheric phenomena. The imaging polarized LIDAR can instantly obtain the three-dimensional image of the target, which has many advantages such as high resolution, fast imaging speed and simple structure. Subsequently, the polarized LIDAR developed as an independent branch of the study, and a lot of excellent research results have been achieved. This paper mainly introduces the basic principle of polarization LIDAR, the realization way, the research progress at home and abroad.

Ultra-smooth surface is widely used in aspheric optics elements and astronomical instruments. Fabrication of ultra-smooth surfaces, especially for hard and brittle materials, has always been a challenge to the optics industry. To efficiently produce ultra-smooth surfaces, a novel disk hydrodynamic polishing (DHDP) is proposed. As a non-contact hydrodynamic polishing method for DHDP process, the polishing tool rotates at high speed under the drive of the motor. At this time, a hydrodynamic fluid film is formed between the polishing tool and the workpiece surface. In the gap of fluid film, the velocity and pressure of fluid show a regular gradient, and it drove the solid particles impact the workpiece. Then, the computational fluid dynamics(CFD), solid-liquid discrete phase model (DMP) and erosion model (EM) are combined to track the particles trajectories in the proximity of the workpiece and reveal the erosion theoretical mechanism for DHDP process. The current research has a great significance to analyze the characteristics of the complex interactions between the phases and the erosion of solid particles in DHDP process. The simulation results can also be used to quantitatively predict the local erosion depth on polishing surface.