Geiger mode avanlanche photon diode （GM-APD） lidar has single photon detection sensitivity， which greatly reduces the system volume and power consumption. It makes the system feasible for practical application， and has become a hot topic in recent studies. However， owing to the limitation of the pixel number， the spatial resolution is low， which makes it difficult to obtain the clear contour of the remote target， and the object detection rate is not high. To solve this problem， a detection algorithm based on multi-level processing of the intensity and range images was proposed to find the correlation between the intensity images and point clouds’ features to improve the probability of small object detection. First， the improved feature pyramid network （FPN） combines the receptive field block （RFB） and convolutional block attention module （CBAM） with the feature extraction network to enhance the selection accuracy of intensity images. Second， the intensity and range images are combined into point clouds with intensity information in the candidate regions. Finally， a dynamic graph convolution network （DGCNN） is used to perform secondary detection on the target in the candidate regions. Moreover， point cloud information is used to further select the object in the candidate regions. In the GM-APD lidar long-range vehicle dataset， the AP of the network achieves 98.8%， and it has good robustness for complex scenes， such as incomplete vehicle structure， weak echo， and strongly reflected light spot. Compared with the SSD and YOLOv5， the detection accuracy of the network improved by 3.1% and 2.5%， respectively， which is feasible for lidar dim object detection.

To achieve the automatic and rapid detection and sorting of high-brightness reflection metal cylindrical pots， as well as break through the technical problems of slow speed and low efficiency of metal pot surface defect detection， a bi-directional feature pyramid network （BiFPN） was introduced in this study based on the YOLOX network. In addition， a lightweight feature fusion network model was devised on the basis of the attention mechanism， and the lightweight design of the computing model was realized. Meanwhile， the attention mechanism module was employed to learn the channel and space of feature information， effectively alleviating the semantic gap of multi-scale features and improving the detection precision of the model. Considering the unbalanced distribution of the learning weight of the network for difficult and easy classification samples， the classification loss function regarding the attenuation factor was determined. Comparisons of the feature fusion network， classification loss function， and attention mechanism module position ablation were conducted using the metal pot cylindrical surface defect dataset. The experimental results show that the fusion attention mechanism model can effectively identify six types of defects with different shapes， the average detection precision mAP0.5 of the test set realized 90.92%， and the detection frame rate was 30.84 FPS. Thus， cylindrical surface defects of metal pots can be identified and located， rapidly as well as with high precision， by using the proposed model.

To expand the decryption method of computational holographic encrypted images for a symmetric-asymmetric hybrid encryption system that cannot be easily attacked illegally， a scheme involving the use of a neural network to restore chaotic iris phase mask computational holographic encrypted images is proposed. First， a plaintext image is encrypted and a ciphertext image of a computational hologram is generated. Next， numerous ciphertext image pairs are generated as datasets. Subsequently， they are continuously trained and tested by using them to build a neural network. Results show that the trained neural network can fit the mapping relationship from the ciphertext image to the plaintext image and that the public or private key is no longer used to decrypt the ciphertext image during decryption. Additionally， the average cross-correlation coefficient is 0.984， the average peak signal-to-noise ratio is 61.0 dB， and the average structural similarity is 0.77， which indicate better performances compared with the performances of a plaintext image recovered by the neural network. By polluting the ciphertext image with noise， a higher quality image is obtained. The purpose of decrypting a ciphertext image via a neural network is achieved， and the scheme is shown to be feasible and robust.

An offset twist-roller frictional transmission mechanism was designed to satisfy the high resolution， compact space， and low-cost requirements of a beam pointing device， given its small load and low-speed application characteristics. The mechanism exhibits the characteristic of a variable transmission ratio， which is realized by changing the twist angle of the driving and driven wheels. To optimize the design interval of the twist angle， the influence of the twist angle on the output characteristics of the mechanism was studied. Based on the analysis of the kinematic characteristics， force， and deformation of the mechanism， the rigid-flexible coupling dynamic model of the transmission was established. We studied the effect of twist angle on output angular velocity stability and angular displacement resolution by conducting a physical prototype experiment and multi-body dynamics simulation. The results indicate that the larger the twist angle， the larger is the fluctuation range of output angular velocity and more unstable is the transmission. The designed mechanism can realize stable transmission when the twist angle does not exceed 80°， in addition to achieving a low-speed micro-stepping motion of 0.005° （18ʺ）. Moreover， the output resolution is far better than that of the single-stage cylindrical gear transmission. The analysis yields a novel concept of employing an offset twist-roller frictional transmission mechanism in the transmission system of a beam pointing device.

To reduce the differences between the data scales and volume of multi-source cross-scale point cloud data， this study proposes a multi-scale decomposition method of point cloud data based on wavelet transform. This study examines the multi-scale decomposition of small-scale point cloud data with considerable attention and the application of scale decomposition in cross-scale point cloud data registration. First， the small-scale point cloud is grid modeled， and the global point cloud binary expression function is established. Subsequently， according to the theory of discrete wavelet transformation， three-dimensional wavelet decomposition of the grid point cloud is performed several times， and the low-pass characteristics of the wavelet scale function are used to retain the low-frequency information to obtain the approximate scale data of the original small-scale point cloud. The similarity with the original data is then characterized based on the surface dimension and the difference in body dimension， and the effective wavelet decomposition series is determined. Finally， the point cloud data obtained by decomposition at various levels are accurately registered with the large-scale point cloud data， and the registration relationship is applied to the original point cloud to increase the registration accuracy of the cross-scale point cloud data. The experimental results show that the multi-scale decomposition method proposed in this paper can effectively decompose the data. When applied to the multi-scale measurement of an aero-engine blade， the registration accuracy of the local cooling holes small-scale point cloud data and the overall blade structure light data of micrometry increased by 61.36%. The proposed decomposition method is applied to the multi-scale measurement of blade edge and grid parts， and the registration accuracy is increased by 48.59% and 43.86%， respectively. The proposed multi-scale decomposition method of the point cloud can effectively decompose small-scale point cloud data， and ultimately improve the registration accuracy of cross-scale data.

To reduce the effects of gravity deformation on the imaging quality of large aperture ground-based telescopes during optical tracking， a parallel adjustment mechanism with high lateral stiffness and submicron accuracy based on flexible hinges is developed. First， a system of parallel mechanism is introduced and a two-degree-of-freedom flexible hinge is designed according to specific technical indicators. Second， equivalent kinematics and stiffness models of the flexible hinge parallel mechanism are developed. Subsequently， a rigid-flexible coupling kinematics simulation system of the parallel mechanism is established， and the effects of the flexible hinge on the accuracy of the mechanism are analyzed. Finally， an experimental test system is built to verify the rationality of the flexible hinge design and the accuracy of the rigid-flexible coupling kinematics analysis of the parallel adjustment platform. Simulation and test results show that the rotational stiffness error of the flexible hinge is controlled to within 3.54%， the motion accuracy of the small displacement （micrometer/angular second） reaches the sub-micrometer level， and the motion accuracy of the large displacement （millimeter/degree） is controlled to within the micrometer level as compared with the simulation results. The lateral stiffness of the mechanism is greater than 60 N/μm and can thus meet the requirements of ground-based telescope optical imaging.

The advent of 5G technology has spurred rapid advancements in high-speed signal development， leading to increasing bandwidth requirements for signal channels. This study examines the microwave characteristics of "quasi-coaxial" interconnections in ceramic substrates and proposes a novel non-vertical interconnection structure. By optimizing the design of metallized vias between ceramic layers， the impedance at the junction of vertical vias and horizontal transmission lines is improved， enhancing high-frequency signal transmission and broadening bandwidth. The paper analyzes the effect of misalignment angle， ladder series， solder ball radius， and distance between solder balls on transmission performance. The result is a broadband， low-loss ceramic substrate interconnection. Experimental results show that the structure can operate at 55 GHz with an insertion loss less than 1.5 dB and a return loss greater than 15 dB in the DC-55 GHz band. The measured data demonstrate satisfactory transmission of 56 G/112 G NRZ and 112 G PAM4 high-speed signals without the need for preweighting and equalization.

To achieve adaptability to low frequency， wide bandwidth， high amplitude， and other vibration environments， a piezoelectric vibration harvester with excitation direction conversion （PVHEDC） is proposed； it constitutes a vibration collector and transducer. Due to the commutation structure， the vibration direction of the transducer is perpendicular to the ambient vibration direction， limiting the response amplitude. The dynamic model of the PVHEDC was established， and the influence of relevant parameters on its output characteristics was obtained via simulation and experiments. As a result， two natural frequencies were observed considering a low-frequency environment， which are the natural frequencies of the vibration collector and transducer， respectively， causing the output voltage of the PVHEDC to peak. With increasing length and proof mass of the elastic beam， fn1 gradually decreased， while fn2 remained unchanged， with the former corresponding to an essentially unchanged output voltage and the latter corresponding to an increased output voltage. Meanwhile， the bandwidth broadened. The experiment results show that when the external excitation amplitude increases to a certain value， the output voltage no longer increases， and the amplitude of the PVHEDC is effectively controlled. The achieved maximum output power is 0.4 mW for the optimal external load resistance of 540 kΩ. In practice， the above parameters influence the resonant frequency of the PVHEDC and its corresponding output voltage and limit the response amplitude， allowing adaptation to low-frequency， broadband， high-intensity， and large-amplitude working environments.

In the three-dimensional （3D） measurement of large-scale casting slabs using a multi-line structured light system， the accuracy of extracting the central stripe of the structured light is an important factor affecting the final 3D measurement accuracy. In this study， a multi-line structured light center stripe extraction method based on the optimization of the traditional gray gravity method is proposed. For linear structured light stripes projected onto the slab surface whose laser cross-section does not fit a Gaussian distribution， firstly， the environmental noise is removed by analyzing the image background difference. The boundary of linear structured light is determined according to the change in grayscale value between the linear structured light stripe and the background， and the region of interest of structured light is extracted. Then， the adaptive gray threshold of the gray gravity method is calculated according to the gray integral proportion of the light stripe in the gradient direction. The gray gravity method is used to process the region of interest and extract the approximate center point of the light stripe. Finally， the sub-pixel center coordinates of the light stripe are obtained by optimizing the extraction combined with the gravity of the roughly located center point in the neighborhood with a radius of 5 pixels. The experimental results show that this method has high precision， high speed， and strong robustness in extracting the center of the laser stripe on the surface of a casting slab with uneven reflection characteristics. The standard deviation of the 3D measurement results of the final casting slab edge trajectory points is within 2 mm.

Through silicon via （TSV） technology refers to the silicon micro-hole processing technology used to achieve vertical conduction and interconnection between chips. The three-dimensional （3D） morphology of TSV samples is commonly measured using destructive scanning electron microscope （SEM） profile imaging technology. In addition， white light interference technology has the advantage of non-destructive measurement； however， it is difficult to probe the light reaching the bottom of TSV samples with an aspect ratio higher than 6：1， and the topography is also distorted. To address this problem， a near-infrared micro-interference detection method based on aberration compensation is proposed in this paper. The method uses near-infrared broadband light as the probe light source， which can penetrate through the TSV. Moreover， an adaptive aberration compensation module in the form of a deformable mirror is introduced into the detection system to compensate for the modulation aberration that is synchronously generated by the TSV. The first step of detecting the TSV 3D topography is to set the type and quantity of the aberration that needs to be compensated by the deformable mirror. This is obtained using the finite element simulation software COMSOL Multiphysics based on the aberration modulation law for a 3D TSV high aspect ratio structure under the probe light. Subsequently， the evaluation function index threshold based on the frequency domain is used to estimate the focus state of the TSV bottom image. Furthermore， the refocusing ability of the probe light is essentially improved and a clear bottom image of the tested TSV is obtained. On this basis， the depth value of the tested TSV and its 3D topography distribution is calculated using vertical scanning interferometry. Finally， two types of TSV samples with a diameter of 10 μm and depth of 65 μm （aspect ratio 6.5）， and diameter of 10 μm and depth of 103 μm （aspect ratio 10.3）， are measured using the proposed method. The results show that the relative error of depth measurement is 1% compared with that of the high-precision SEM measurement results. Moreover， the proposed method can obtain a clear image of the TSV bottom with a high aspect ratio， and can also effectively enhance the wide spectrum interference signal and fringe contrast at the bottom compared with white light microscopic interference. Therefore， the proposed method can be used to precisely measure the 3D topography of TSVs with a higher aspect ratio.

Underwater connector is an important part of the seafloor observation. Currently， the oil-filled sealing structure is complicated and it’s hard to design in project. To solve this problem， contrast the connection mode of the land-based fiber network，we presented a novel design for the contact fiber-coupled based on underwater wet-mateable the fiber connector. First， based on the analysis of Beer-Lambert law， water absorption properties and surface tension of liquid theory， we identified the ability to form the water film thickness within 5 μm at the optical fiber end faces. This thin layer water exhibit negligible absorption of the 1 310/1 550 nm because of the absorption loss can be kept within the predefined target range. The study was conducted that the loss analyses is summarized by using optic fiber in air， water， and oil based on 1 550 nm. Finally， we designed the novel optical structure based on this idea and analyzed their transmission performance. The results showed that the summary of the transmission loss is 0.17 dB and the loss measurement in the new fiber connector is 0.23 dB. The findings show that the design of the contact fiber-coupled of underwater wet-mateable is meeting the low loss requirements. Therefor， we can directly implement in-situ wet-mate interconnects in the water rather than the silicone oil based on the surface cleaning. This paper provides a new design for the optical lines in underwater wet-mate connector. This idea is expected to break the Intellectual property that built upon the work of the sealing technique of the oil-filled.

A layered polyimide （PI）-fiber Bragg grating （FBG） concentration sensor was designed to monitor the concentration of a coolant in a nuclear cooling pipeline to solve various problems that exist in current measuring instruments， such as vulnerability to environmental interference， medium deposition， and difficulty in the distribution of multi-point monitoring. First， the design of a layered PI film-FBG concentration sensing mechanism was investigated based on a combination of the PI film water absorption characteristics， the principle of water molecule diffusion， and FBG sensing theory. Subsequently， the layered PI film-FBG concentration sensor was designed and fabricated， the concentration measurement experiment platform was built， and the concentration sensing experiment was conducted. Finally， the concentration characteristic curve of the PI film-coated FBG concentration sensor in a potassium chloride （KCl） solution was obtained， and its sensitivity and hysteresis characteristics were analyzed. The experimental results show that there is a linear relationship between the concentration and wavelength shift， and the fitting degree of the concentration characteristic curve is 0.994 2. In addition， the maximum difference between the forward and reverse stroke output is 35 pm， and the average sensitivity of the sensor is 157.6 pm/（mol·L-1）， which are 7.4 times and 49.1 times higher than those of the one-layer annular PI film-coated FBG concentration sensor and corrosive FBG concentration sensor under the same conditions， respectively. Therefore， the sensor provides a new method for measuring the coolant concentration in nuclear industrial pipelines with high sensitivity， safety， and anti-electromagnetic interference.