Because the structure of Brillouin optical time domain reflectometer (BOTDR) is single-ended optical signal injection and the spontaneous Brillouin scattering signal is weak, signal-to-noise ratio (SNR) of the BGS is low.Two-dimensional (2D) image Gauss filtering algorithm is proposed to process BGS image to achieve noise reduction of BOTDR system.The noise signal characteristics of BGS acquired by BOTDR based on microwave heterodyne sweep method are analyzed, 2D image Gaussian filter BGS noise reduction algorithm is designed, experiments are carried out on BOTDR temperature measurement system, and the experiment verifies the noise suppression effect and signal processing efficiency of the proposed algorithm.Experimental proof that under the condition of 100 cumulative averages of a single frequency power curve, Gaussian filter algorithm is used to reduce the noise of 2D BGS image, and the SNR of BGS is improved by about 10.7dB after noise reduction, the fluctuation range of BFS in the variable temperature zone is reduced from ±7.79MHz to ±2.61MHz, at the same time, the effective sensing distance has been increased from about 13.8km to about 24.0km, the average signal processing time of noise reduction algorithm is about 0.129s.The results show that the method can significantly improve the SNR of BOTDR system and has good real-time performance.

In order to meet the development requirements of high resolution and wide-angle mobile phone lens, a 13 megapixel wide-angle mobile phone lens was designed based on ZEMAX.The lens is composed of six plastic lenses and a near-infrared filter with the focal length of the system of 3.36mm, the field of view of 84°, the F-number of the system of 2.6, and the total length of the system and the image height of 5.4mm and 5.86mm.Image sensor OV13B10 from OmniVision is adopted, with a resolution of 4208 pixel×3120 pixel, pixel size of 1.12m×1.12m.The design results show that the modulation transfer function (MTF) of the central field of view is greater than 49% and the MTF of the field of view within 0.7 is greater than 33% at the Nyquist frequency of 223lp/mm.The maximum RMS radius in full field of view is 3.192m and the maximum distortion is -3.948%.The full field of view relative illuminance is better than 38%.The research provides a design reference for the high image quality, wide-angle mobile phone lens.

The primary mirror is the core part of the telescope, and the protection mechanism needs to be designed to maintain its performance and extend the service life of the main mirror film.On this background, according to the structural characteristics and project requirements of a 2.5-meter large aperture optical telescope, a segmented synchronous folding primary mirror protection mechanism is designed.The optimization design of the protective cover plate of composite sandwich material is emphasized, so that the stiffness and strength can meet the relevant requirements while meeting the requirements of lightweight.(Cover plate mass <2.5kg, first-order mode about >20Hz, encountered eight-level wind load and 1kg block impact Tsai-Wu <1). At the same time, through the driving design and control mode test of the folding movement of the mechanism, the multi-fan synchronous driving can be realized in the specified time, and the unidirectional folding movement time is about 52s.This design has a good protection effect on the main mirror and even the entire internal structure of the main mirror tube.This design provides a new idea and validation for the main mirror protection mechanism of large aperture telescopes in China, and has a certain reference significance for the research of other main mirror protection structures.

An electromagnetic-actuated liquid tunable focal length cylindrical lens is designed and fabricated, and its properties are characterized.The results show that the focal length of the cylindrical lens decreases gradually from the +∞ to 8.97mm when the current ranges from 0 to 1.5A, whose maximum transverse magnification is 1.6 and the system power-on response time is about 20ms, power-off relaxation time is about 25ms.As the operating current is 1.5A, its focal length of lens is 8.97mm, the corresponding focal line length is 11mm and the focal line width is 1mm.The liquid cylindrical lenses proposed in the article have potential applications in the fields of beam shaping, device scanning and 2D/3D switchable display.

In order to overcome the difficulty of traditional aspherical surfaces in the structural design of optical systems that cannot simultaneously meet the requirements of imaging quality and system volume, a free-form surface is used to optimize an off-axis two-mirror afocal system.Based on the vector aberration theory, the aberration distribution formula of the system is derived when the free-form surface is at two different positions of the aperture diaphragm and deviates from the aperture diaphragm.Based on this, the Zernike items are selectively introduced to balance the aberration, and the selection of balance parameters for different types of aberrations is discussed in detail.Using this method, a miniaturized infrared off-axis two-mirror afocal system was designed, obtain an optimization result with an entrance pupil diameter of 200mm, a wavelength range of 2.4m to 3.0m, the magnification rate of 5, the FOV of 0.4°, and the maximum wavefront RMS is 0.0682(=2.7m), the average value is 0.0291.Compared with traditional quadratic surface design scheme, the wave aberration has been significantly improved, and MTF has been significantly improved in mid to low frequencies, approaching the diffraction limit in the full field of view.

A new-type slit spatial filter is proposed, which consists of a pair of cylindrical lens, a filtering slit, and a right-angle mirror.A pair of cylindrical lens and a filtering slit are used to form a one-dimensional filter structure.The right-angle mirror is used to rotate the wavefront 90°and reflect it back to the one-dimensional filter structure for another dimension of spatial filtering, thereby achieving complete filtering of the beam.Based on the Rayleigh-Sommerfeld diffraction integral, the transmission of the beam in the new slit filter was theoretically simulated using Matlab software. Combined with experimental verification, proving the image relay and spatial filtering functions of the new slit filter structure. The new-type slit spatial filter is similar to the folded structure of the three-lens slit spatial filter, so it is nearly half shorter in length, which can further reduce the scale and cost of large laser devices.

The meridian optical focus distribution is the key to controlling the size of the distance zone, near zone and channel, when designing Progressive Addition Lenses (PALs) using the direct method.The meridian optical focus distribution function represented by a high-order polynomial cannot effectively control the rate of optical focus change, resulting in a lack of flexibility in the individual design of lenses.To solve this problem, a higher-order Bezier function is proposed for describing the desired meridian optical focus distribution.Multiple curve control points are added on the meridian line.The rate of optical focus is changed in the corresponding region by changing the position of the control points, thus realizing the personalized control of the optical focus and dispersion distribution of the lenses.Through computational simulation, manufacturing and profile measurement analysis, it can be seen that the four designed PALs can meet the wearing purposes with low aberration extremes and wide focal stability ranges, which demonstrating the high plasticity of the Bezier curves.With the ability of meridian flexibility and adjustability, the method can effectively control the design, position, size of the key area, distance area, channel size.The proposed method can be used for personalized design of progressive multifocal spectacles for anti-fatigue, office and driving scenarios.

The polarization Bidirectional Reflectance Distribution Function (p-BRDF) can describe the spatial polarization characteristics of the reflected light on an object’s surface, and has been widely used in many fields.Based on the microfacet theory, the electric field vectors of polarized incident and reflected light are decomposed and composed by the angle relationship between the microfacet and the object surface, as well as the Fresnel reflection on the microfacet, to obtain the polarization reflectivity.This is then used as the Fresnel reflection term in the Cook-Torrance BRDF model to establish a polarization microfacet BRDF model.Introducing polarization into geometric optical model is simple and has clear physical meaning.The Particle swarm optimization algorithm is used to fit the parameters in the model with the in-plane BRDF experimental data on the brass surface.The standard mean squared error of the model parameter selection is between 0.2820% and 5.7049%, and the best adjusted R2 is 0.9957.The results indicate the feasibility of using the polarization microfacet BRDF model to describe the polarization reflection characteristics.

Aiming at the problem of difficulty in selecting access points for hybrid LiFi-WiFi heterogeneous networks in complex indoor environments, a random forest model-based access point selection algorithm for hybrid LiFi-WiFi networks is proposed.The proposed network access point selection algorithm utilizes the channel characteristics of multiple networks, and constructs a training set by simulating different complex indoor environments and collecting the equivalent values of received signal strength and signal-to-noise ratio of users at different locations in different situations, so that the model can adapt to various complex indoor environments.Simulation results show that compared with the traditional network selection algorithm, the average user reachable throughput of this algorithm is improved by about 82%, especially when the indoor situation is more complex, it can be improved by 160%.The algorithm proposed can significantly reduce the number of switching times by 20% compared with other algorithms when the users are moving and the number of accessed users increases.

The rapid development of mega constellations led by Starlink has a huge impact on space security, astronomical observation and other aspects.Effective monitoring is urgently needed and optical observation is the most effective observation method.Satellite orbit model and optical reflection model are established based on the orbital information and satellite structure characteristics of all 42000 Starlink satellites.The distribution of satellite visual magnitude in different regions at different seasons and times of night was simulated.According to the imaging characteristics of Starlink satellite’s observation by photoelectric equipment, the imaging signal-to-noise ratio (SNR) model was established.The SNR distribution of Starlink satellite observation by 50mm aperture photoelectric equipment was simulated, and verified through observation experiments using a 50mm aperture image total station.The results show that the higher the latitude, the longer the Starlink satellite can be observed at night, with the longest observable time in summer and the shortest in winter.The maximum number of satellites can be observed near 40°north and south latitude, making it the optimal observation location.The closer the satellite is to the zenith position of the observation station, the higher the visual magnitude brightness, the higher the imaging SNR when the photoelectric equipment observes.

To address the problem of nonlinear distortion of high-speed transmission in the visible light communication system, a deep learning based nonlinear distortion compensation method of high-speed transmission in the visible light communication system is proposed.Firstly, a few forward and backward neighbor symbols of the target symbol in the high-speed frames are extracted by a sliding window, which construct a context symbol sequence including the target symbol.Then, the one-dimensional convolutional neural network is used to extract the space feature of the context sequence, and the bi-directional long short term memory network is used to extract the forward and backward temporal dependency. Experimental results show that, the proposed method compensates the distortion of the visible light signals, and it reduces the bit error rate of high-speed transmission in the visible light communication system.

Metalens are usually composed of rectangle, elliptical or circular micro/nano pillars arrayed with design phase.The manufacture of metalens is diffcult for the size of pillars is always subwavelength and the ratio of height to length/width (or radius) of pillar is usually bigger than 5.The machining accuracy of Direct Laser Writing (DLW) with high machining freedom and mask-free ability doesn’t satisfy the manufacture of pillar type metalens.However, the accuracy of drilling holes and the process stability of DLW is much better than pillar fabrication.Thus, this paper introduces hole type metalens, in which the hole work as the unit cell with various radius to modify the optical phase to foucus. The periphery of hole type metalens is tightly connected as one net, which shows higher structure stability.The break of phase modified cell within pillar type metalens will rarely occure within hole type metalens.The deposition of anti-reflective (AR) film on the back of holes surface of metalens will enhance the transmittance of incident light and improve the harsh environment endurance as mounting this surface to the air and holes surface into the optical system.In this paper, the finite-difference time-domain (FDTD) was applied to the analysis of imaging ability of hole type metalens, which is designed with ZnS substrate, 9m center wavelength, 1mm focal length.The simulation results show excellent focusabilty and well commercial value.

In order to avoid the threat of robot collision to construction process and safety, and stimulate the great potential of robots in prefabricated buildings, a laser sensing adaptive trajectory planning method for prefabricated building robots is proposed.The grid map of a prefabricated building is constructed, and the peripheral environment data of the robot is obtained by the laser sensor.The improved A* algorithm is used to remove redundant nodes, so that the path only includes the beginning, end and key points.The dynamic window and ant colony algorithm are used for local trajectory planning, and the target velocity is calculated to complete the global planning.The experimental results are verified in the grid map of a prefabricated building.In the static grid environment, the proposed method can plan a 38.6m path in 165 iterations and 3.16s.In a dynamic grid environment, this method can avoid dynamic obstacles in 248 iterations and 3.81s.The results show that the path of this method is smoother than that of the literature method[5, 6], and a shorter path can be planned in less time and number of iterations.

Aiming at improving the accuracy of salient object segmentation, a salient object detection method based on edge and multi-scale feature fusion was proposed.Firstly, the ResNet50 network was used to extract features from images. Then, an improved spatial attention module was utilized to enhance the representation ability of features.Next, a novel edge and multi-scale feature fusion module was proposed, which organically combines edge information with multi-scale features information, and a loss function that comprehensively considers salient object subject segmentation and edge segmentation was designed to effectively supervise the future fusion module, ensuring that the model will simultaneously focus on the detail information of salient object subject and edge during training, in order to improve the clarity of salient object subject and edge.Finally, a context enhancement module was introduced innovatively to effectively reduce information loss during multiple upsampling and downsampling processes in deep learning networks, thereby improving the accuracy of the salient object subject and edge.Compared with eight mainstream algorithms in recent years on three public datasets, this method outperforms other algorithms in both quantitative and qualitative results, verifying its effectiveness and superiority.

It is proposed to build a new type of laser self-hybrid interference system, use the laser self-mixed interference effect, and select the narrow line width and low noise line polarizing solid laser as a light source, increase the EOM electro-light modulation crystal, and introduce electro-optic orthogonal modulation and demodulation algorithms. Experimental and theoretical analysis shows that laser self-hybrid interference can achieve vibration measurement under weak feedback, and the newly constructed self-hybrid system structure is simple, convenient to operate, and easy to achieve miniaturization.The problem of migration of the jump point improves the speed of signal processing and the real time of measurement.Therefore, the laser self-hybrid interference instrument of the electro-lighting and demodulation can overcome the limitations of the target mirror on the application scenarios of the interference instrument, and realize the real-time measurement of the nano-level high-precision of the non-co-operating surface.

Aiming at the problem of weak signal and reduced measurement accuracy caused by noise interference in Brillouin Optical Time Domain Reflectometer (BOTDR), a wavelet threshold denoising method based on Brillouin gain spectrum two-dimensional image is proposed to improve the measurement accuracy of BOTDR system.Analyzed the principle of microwave frequency sweep BOTDR measurement and the characteristics of two-dimensional Brillouin gain spectrum image, designed a two-dimensional image wavelet threshold denoising algorithm flow, built a microwave scanning frequency BOTDR experimental device for denoising performance verification, and determined the optimal parameter matching scheme for two-dimensional wavelet function, degrade level, threshold regulation, and threshold function based on noise reduction performance comparison.The experimental results show that using the two-dimensional image wavelet threshold denoising algorithm, under the condition of accumulating an average of 500 times, the Brillouin frequency shift fluctuation range of the 10km sensing fiber end temperature range before and after denoising is reduced from ±4.61MHz to ±2.59MHz, and the root mean square error is reduced from 2.24MHz to 0.99MHz；Before and after noise reduction, the Brillouin frequency shift fluctuation range of the 50km sensing fiber end temperature range decreased from ±41.25MHz to ±2.83MHz, and the root mean square error decreased from 33.59MHz to 1.69MHz.Research has shown that using this noise reduction scheme can effectively improve the accuracy of BOTDR measurement.

Aiming at the measurement problems caused by different design coordinate system and measuring coordinate system reference system, a method of measuring free surface based on differential confocal normal measurement registration was proposed.The method utilizes the characteristic that the zero points of the differential confocal curves correspond accurately to the measured points to achieve precise focusing of the measured points.The inclination angle of the measured points is measured precisely by the PSD (Position Sensitive Detector) at the focus position.Based on the inclination angle of the measured points, the ICP (Iterative Closest Point) Point to Plane method is used for point cloud alignment to improve point cloud alignment accuracy and significantly improve surface measurement accuracy.According to the theoretical analysis and preliminary experiments, the accuracy of PV measured by this method can reach 65nm, and the RMS can reach 10nm.The difference between the PV and the zygo interferometer measurement results is 7nm, the difference between the RMS and the zygo interferometer measurement results is 1nm, which means the measurement results have small errors.The method provides a new technical approach for the surface profile measurement of freeform surface.

In response to the requirements for high-precision measurement of form and position errors, a form and position error measurement system has been built using position sensitive detectors (PSD), which have the advantages of high accuracy, high sensitivity, fast response, and easy integration with other systems.The system uses laser beams as reference lines and 2D position sensitive detectors as photoelectric conversion devices, Through mobile terminal software, real-time display of the position changes of sampling points relative to the reference line can be achieved, and the shape and position error evaluation results can be quickly output using the least squares method, which is faster and more stable compared to traditional measurement methods.The system measures the straightness of the linear guide rail and evaluates the uncertainty based on single point measurement and measurement system angle.The experiment shows that the standard uncertainty based on single point measurement is 5.9m, and the standard uncertainty based on measurement system is 7.4m.This verifies that the system has reliable measurement results.