For the LED extended light source, the bidirectional feedback mesh design method is used to design the freeform lens. Divide the energy grid of light source and target surface, and according to the feedback coefficient is established for each iteration curve, the mesh points are modified, and a new mapping relationship is established. Finally, two kinds of freeform lenses with no illumination in the center of the illumination area are designed. The Monte Carlo method is used to simulate led extended light sources with different sizes. The results show that for the LED extended light source with a size of less than 4mm×4mm, the uniformity and light energy utilization of the designed freeform lens in the lighting area can reach more than 90.5%, and it is suitable for road lighting occasions.

The harsh airborne environment has put forward high requirements for the long focal common-aperture optical system. Four key technologies are studied, reasonable and feasible solutions are given. The design of the multi-spectral common-aperture optical system developed enables the airborne optical imaging system to meet the requirements of miniaturization and lightweight. In the range of airborne ambient temperature variation, the high-gradient silicon carbide aspheric surface optical processing and detection method proposed ensures that the radius of curvature and the accuracy of surface shape of optical elements remain stable and can achieve excellent imaging quality. The application of silicon carbide reflector flexible support technology enables the optical elements and support structure to have higher structural strength and stiffness, and at the same time can ensure the dimensional stability of the optical system structure. The wavefront of the actual optical system carries a lot of errors. In this paper, the computer-aided precision mounting technology is proposed to implement the mounting of the optical system with high image quality.

The increased industrial emissions have worsened the water environment and resulted in the endangerment of marine life. Traditional water body monitoring technology was unable to meet the real-time, fast and accurate application requirements. In order to implement more effective monitoring of the water body environment, various research units have begun to use remote sensing technology to obtain water body information. Aiming at the technical requirements of water monitoring, an off-axis three-reflection optical system is designed with a focal length of 84.5mm, a diameter of 20mm, and a field of view of 36°×6°. The reflective system can meet the requirements for wide spectrum coverage of space remote sensing. The primary and third mirrors of the system adopt sixth aspheric surfaces, and secondary mirror adopts quadric surfaces and the Offner compensator, which takes the primary mirror as an reference, is designed to generate aspheric waves to perform zero compensation test on the aspheric mirror. After tolerance analysis, it has good machinability and the modulation transfer function (MTF) values of each field of view at the Nyquist frequency (71lp/mm) are all over 0.6, which is close to the diffraction limit, and shows good imaging quality. This optical system is expected to have certain reference significance for monitoring water pollution.

The G-S hologram can be used to design spatial optical filter. Due to the refractive index difference between the medium and air, the phase of spatial light will change when it passes through the holographic plate. Based on the height modulation phase of each pixel photoresist, a 2.5D gray-scale holographic spatial optical filter was designed. The photoresists with different heights were prepared by gray-scale electron beam lithography, and the patterns were etched onto the silicon wafer by dry etching. The electron beam lithography process, development parameters and etching process were optimized. The high diffraction efficiency spatial optical filter was fabricated and characterized.

Focusing of an off-axis telescope with the secondary mirror is studied based on a hexapod positioning system and a linear stage. That have compared linear positional focusing, rotating focusing and a combination of them (the hybrid method), and concluded that the linear positional focusing achieves higher MTF and lower OPD at short focusing distances, the rotating focusing performs well at medium and large focusing distances, while the hybrid method could be applied at all distances. A 800mm-diameter off-axis telescope is designed using the above methods, and the optical qualities at distances ranging from 0.6km to 10km are evaluated. Results show that the hybrid method leads to the highest imaging qualities.

The difference of the immersion depth of the planar optical element, the immersion depth of the convex spherical optical element, and the radius of curvature of the convex spherical optical element will change the shear force field at the entrance area of magnetorheological polishing. In order to study the formation mechanism of the shear force field in the entrance area of magnetorheological polishing, firstly establish the necessary fluid model in the magnetorheological polishing process and analyze the geometric characteristics of the entrance area; Secondly, by numerically calculating the different immersion depth and convexity of the planar optical element The influence of different immersion depths of spherical optical elements and different radii of curvature of convex spherical optical elements can obtain the corresponding shear force distribution. Finally, it is concluded that when the certain immersion depth of the planar optical element gradually increases, the shear force at the entrance of the polishing area gradually increases; when the certain immersion depth of the convex spherical optical element gradually increases, the shear force at the entrance of the polishing area gradually increases. Increase; the radius of curvature of the convex spherical optical element has no significant effect on the shear force, and the shear force distribution under different radii of curvature is roughly the same.

With the development of neural networks in recent years, a variety of photoelectric neural network frameworks have been successively proposed and presented strong application potentials in image and language processing and so on. A new type of optical scattering neural network based on transparent medium imaging is designed. Firstly a physical characteristic of the bidirectional propagation in scattering medium is presented, a single layer intelligent unit of optical neural network is formulated by cascading a stack of medium layer, and the nonlinear photoelectric material film is appended to act as nonlinear activation response. Then, the corresponding digital bidirectional complex-valued neural network is established according to the bidirectional scattering physical model, and the backpropagation algorithm for its digital weight update is derived. The weight parameters of a digital complex-valued neural network can be realized by adjusting the physical parameters of scattering medium. The numerical experiments about the effectiveness of the intelligent classification of proposed optical scattering neural network is verified on the MNIST data set.

Metalens has gained research interest in the field of nanophotonics. Contrary to the traditional metalens composed of high refractive index dielectric material such as TiO2, Si, or GaN, the authors propose a Pancharatnam-Berry (PB )Phase-type metalens. All the numerical simulations of the designed structure were performed through finite-difference time-domain software (FDTD Solutions). The simulation results validate that the peak focusing efficiency of the photoresist metalens in the visible light band i.e. 500~700nm is 58% while in the near-infrared band i.e. 700~900nm, the peak focusing efficiency is 52%. The results validate that the metalens made of non-high refractive index materials can also achieve better-focusing performance. It overcome the limitations of designing and processing metalens based on high-cost electron beam etching methods and is a great future for metalens. The applications of the metalens expand a broader space for development.

In order to improve the transmission capacity of the fiber and obtain better transmission performance, a six-core fiber with low crosstalk and large mode area assisted by three-layer grooves is designed, each core of the multi-core optical fiber is wrapped by three layers of low refractive index grooves, and a clapboard structure is added in the cladding. The effects of core radius, cladding index, groove depth and groove width on inter-core crosstalk and mode field area were studied by simulation with COMSOL and Rsoft software. The results show that the inter-core crosstalk can reach -55.429dB and the mode field area can reach 305.061μm2 when the wavelength is 1550nm, the core index is 1.4457 and the fiber radius is 40μm. Low crosstalk and large mode area can be obtained in multi-core optical fibers with this structure.

Medium wave infrared devices are developing from monochrome and bicolor to multispectral, metal micro nano structure is the core device to realize the function of medium wave infrared spectroscopy. In order to research the transmission spectrum characteristics of metal micro nano structures in mid infrared band and realize the modulation of transmission spectrum. Based on the principle of Surface Plasmons (SPs), Finite Difference Time Domain method (FDTD) is used to simulate the transmission spectra of thin-film metal micro-nanohole arrays. The influence of the shape and size of the holes, the period of the hole array, the thickness of the metal film and the metal material on the spectral transmittance characteristics of the incident light in the mid-infrared 3.0~5.0μm band is analyzed. By simulating different structures, it is found that the transmittance is mainly determined by the size of the hole. Changing the period of the array can modulate the transmission peak position. The transmittance increases rapidly as the thickness of the metal film decreases, and the metal material Ag is easier to transmit infrared light. Further based on the radius of the circular hole (square hole side length), the array period and the transmission peak position, the least squares method is used to fit the design relations corresponding to different structural models, at the theoretical level, the modulation of the transmission spectrum of the metal micro-nano structure in the 3.0~5.0 μm band is realized, which provides a theoretical basis for the design of multi-channel aperture filter and spectrum detection equipment.

In theory, the formula of transmission peak position of subwavelength hole array structure is used, aiming at the narrow-band single peak transmission results of the dielectric metal dielectric three-layer structure "double glass hole array" structure model with single peak narrow-band transmission characteristics required by the spectral detector device, and based on the existing processing technology for the more common upper edge over etching, double glass hole array structure The potential failure modes of residual hole bottom and over etching hole bottom are analyzed. Compared with the standard double glass hole array model, the effects of processing errors on narrow-band single peak transmission and the mechanism are analyzed. It has a certain guiding significance for the preparation of a new type of pixel level filter film system by using the abnormal transmission principle of double glass hole array structure to selectively transmit the incident light.

Transparent laser ceramics are expected to be the gain medium of new high-power-solid-state lasers instead of single crystal due to their unique superior properties. Therefore, the preparation of transparent ceramics with high optical quality, high laser efficiency and high output power has become the hot point of current research. However, sintered polycrystalline ceramics have a certain amount of grain boundaries and pore defects. The relationship between the porosity and the performance of ceramic lasers and the latest research progress in controlling the porosity of ceramic materials are interviewed. Studies have found that the porosity has a greater impact on the scattering coefficient, transmittance, laser threshold, and laser slope efficiency of laser ceramics. Controlling the ceramic preparation process and optimizing the preparation process can reduce the porosity of the laser ceramic material, thereby improving the optical performance of the laser ceramic material.

1319nm lasers, which has extensive application prospect, play a crucial role in sodium beacons, optical communication, image display and laser medical treatment. The Nd∶YAG 1319nm lasers studies in recent years were stated briefly from the following three aspects: single-wavelength continuous output lasers, pulsed lasers and quasi-continuous-wave lasers. The important applications of 1319nm laser in multiple fields was introduced. The research value and difficulties of these lasers were also presented. Finally, the future development of Nd∶YAG 1319nm lasers were discussed.

Aiming at the problem that the thread parameters can not be focused correctly in machine vision, the influence of defocus imaging on the thread angle parameters are studied by combining theory and experiment. It is proved that defocus phenomenon can cause nonnegligible error in thread parameter measurement. By analyzing the relation between the sharpness evaluation value of thread image and defocus quantity, a method for correcting the error of thread angle is proposed. Firstly, the sharpness evaluation value and defocus quantity of the image are calibrated, then the sharpness evaluation value of the image is calculated, and the defocus quantity is corrected. Finally, the length and area of the image edge of the selected area are corrected according to the defocus quantity, and the thread angle is calculated using the modified results. The experimental results show that the relative error of the modified thread angle is less than 0.0425% when the defocusing quantity is less than 0.5mm for the screw plug gauge of M20×1.5-6H.

Aiming at the calibration traceability problem of barrel’s inner diameter instrument, the structure principle of barrel’s inner diameter instrument is expounded. The principle and composition of calibration device with double dynamic measuring heads is designed. The measuring model is founded. The main factors affecting calibration result are analyzed and the affection by each factor is evaluated. Finally, the uncertainty is presented. The results show that the uncertainty of calibration device is 0.43μm, the calibration result’s uncertainty of inner diameter instrument is 2.9μm. The calibration traceability requirements can be meet and the automatic calibration is realized.

In order to simplify the three-dimensional measurement and solve the baseline limitation problem in traditional stereo vision methods, a method based on single-camera micro-angle rotation combined with neural network for three-dimensional measurement is proposed. The method gathers two-dimensional image data by rotating a single camera at a small angle. Additionally, the three-dimensional spatial coordinates are established by moving the chessboard calibration target with an electric sliding table. Based on the idea of camera linear imaging model, where the projection matrix represents the mapping relationship between image coordinates and 3D coordinates is replaced by BP neural network. It obtains the direct mapping from 2D coordinates to 3D coordinates and realizes 3D measurement under micro baseline. The experimental results show that the absolute error of the proposed method is 0.0864mm compared with the distortion measured by the traditional method. This method integrates neural network into the single-shot vision field, which can make up for the defects of traditional vision measurement methods in micro baseline scene. It has potential application value for mobile equipment, monitoring equipment and 3D information acquisition in narrow scenes.

In order to improve the automatic detection accuracy of urban buildings, a new automatic detection technique of urban buildings is proposed with combination of laser radar scanning and optical imaging technique. The proposed technique delivers the point cloud generated by laser radar scanning to the auto-encoder to down sampling and extract features, and the optical image pixels are delivered to auto-encoder to extract the spatial features, different feature sets are fused based on the graph model of features. Then, a semi-supervised classifier is designed based on convolutional neural network, the classifier is used to recognize different regions of buildings, ground and vegetation. Experimental results show that the proposed technique can detect the regions of urban buildings accurately, it also provides fundamental basis for smart cities development.

To efficiently locate the acupoint during massage, acupuncture as well as other related Chinese medicine diagnosis and treatment. Combining with the principle of monocular vision calibration and positioning, a method for positioning and conversion of traditional Chinese medical manipulator coordinates based on the conversion matrix between three-dimensional coordinate systems is proposed. Obtain the camera's internal parameters by performing camera calibration and establish the camera's three-dimensional world coordinate system firstly,establish the conversion relationship between the camera's three-dimensional world coordinate system and the robot coordinate system based on the custom label image that can be pasted on any fixed position of the manipulator by establishing a homogeneous equation through the three-dimensional coordinate points and solving the optimal homogeneous rotation matrix (R) and translation vector (T).The RT homogeneous matrix is used as the key conversion matrix of the two coordinate systems, and the three-dimensional coordinates of the coordinate points in the coordinate system agreed on the label plane can be obtained through mathematical operation of the point coordinate data and the optimal conversion matrix, which is helpful to complete the manipulator According to the spatial coordinate conversion and precise positioning of the diagnosis position. Using the six-pointed star map as the label map based on the proposed method, the effectiveness and accuracy of the proposed method are verified by simulation experiments.

In the medical field, vein imaging technology is used to assist intravenous injection and the treatment of venous diseases. The reflection-type vein imaging device has a significant advantage in clinical vein locating with the advantages of non-contact and portability. Aiming at the problem that the existing image processing methods are difficult to extract the vein patterns accurately in the reflection-type vein image because of the noises and weak vein features of the reflection-type vein image, a vein pattern extraction method(Reflection-type Imaging Vein Extractor, RIVE) based on convolutional neural network is proposed to improve the accuracy of vein pattern extraction in reflection-type vein images. First, a neural network is trained with the transmission-type vein images and labels; Then the trained network is used to extract the vein patterns in the reflection-type vein images; Finally, comparing the vein patterns extracted from the reflection-type vein images with the results of the transmission-type vein imaging and evaluating the performance of the new method based on the vein extraction rate. The experimental results demonstrate that the vein extraction rate of the RIVE can reach 63.2%, which is a 23.7% improvement compared with the traditional method. Therefore, the proposed method can extract vein patterns more accurately in reflection-type vein images, which is of great significance in clinical vein imaging technology.

In order to overcome the problem that the global relative brightness feature of the image is ignored when the current visible and infrared image fusion methods fuse the image through the energy feature of the image, resulting in the weak expression ability of the infrared target in the fused image, a fusion method by measuring the relative brightness of the image based on the Nonsubsampled shearlet transform is designed. Firstly, the low-frequency and high-frequency coefficients of the image are analyzed by NSST transform. Then, the region energy function is used to measure the energy features contained in the region image. Based on the global mean value and regional mean value of the image, the relative brightness measure model is constructed to obtain the relative brightness feature of the regional image. The energy feature and relative brightness feature of regional image are combined to fuse the low frequency coefficients. Using the frequency values of the four dimensions of row, column and diagonal of the image, a four-dimensional detail measurement factor is established, which is used to calculate the fusion high-frequency coefficients of the image detail features, and then to obtain the fusion image. Experimental results show that the proposed algorithm can not only display the details of the image better, but also better express the content of the infrared target in the image.

In the process of infrared image processing, it is particularly important to improve the contrast of the image, suppress noise and highlight the details of the image. In response to these problems, infrared image enhancement algorithm based on feature fusion is proposed. First, the image is smoothly layered based on guided filtering to obtain the basic layer and the detail layer. The basic layer uses the CLAHE algorithm to extend the range of low-frequency components, and the detail layer uses the Log operator and the Laplace operator to process separately, according to the gradient factor the weight information is used to obtain the fusion detail layer, and then Gamma correction based on the gradient factor is used to suppress the enhancement of the brighter area of the detail layer to prevent over-enhancement, after obtaining the enhanced basic layer and the detail layer, image fusion is performed according to the detailed feature information, and the experimental results are obtained . Through the analysis of the results of different infrared scenes, the proposed algorithm achieves the effect of enhancing edges and suppressing noise.