Pathologic diagnosis based on optical imaging suffers from the problems of cumbersome and time-consuming production of pathology slides，long manual diagnostic times，and increased risk of surgical complications. The development of rapid and quantitative auxiliary pathology diagnosis technology for cancer is of great significance. This study uses Mueller matrix polarization decomposition and Mueller matrix transformation methods，to extract the four parameters：A，b，D，and for the polarization characterization of nuclear and fibrous structures of follicular thyroid carcinoma（FTC）tissue，follicular thyroid adenoma（FTA）tissue，and normal thyroid follicular tissue. Multivariate parameter space to explore the effect of four parameters on the differentiation of FTC tissue is constructed. Statistical methods to analyze the variability of 4 polarization parameters of two polarization-sensitive structures from three tissues are used. The result shows that the A- and b-D planes are able to clearly distinguish FTC tissue from FTA tissue，and normal thyroid follicular tissue. And the parameters of A，b，D，and are able to differentiate between the follicular carcinoma tissues and the follicular adenoma tissues. The results of the study provide a reference for the study of rapid and quantitative adjuvant pathology diagnostic methods for follicular thyroid cancer based on polarized imaging.

Dark-field scattering detection is one of the main methods of non-patterned wafers detection. The incident light shapes and scattered light collection system will directly affect the detection performance of the system. Therefore，the simulation analysis of the defect scattering field and the optical path system are the basis of the system design. In this paper，a simulation system for defect detection of non-patterned wafers surface based on dark-field scattering is established by Finite-Difference Time-Domain（FDTD）method，including particle scattering simulation on the smooth wafer surface，roughness scattering simulation on the wafer surface and performance simulation of collecting optical path system. In this paper，the characteristics of scattering field under different laser incident angles，polarization states and defect sizes are analyzed. The corresponding relationship between spot diameter and detection limit is explored，and the optical path system is designed and simulated according to the scattering characteristics. The optical path system designed according to the simulation results has the characteristics of compact structure and high collection efficiency（the average collection efficiency is 77.65%），which provides a reference for the surface defect detection of non-patterned wafers.

Absolute gravimeters have significant application in many researches，in order to verify and evaluate the measurement function and performance of cold atom interferometry absolute gravimeter in field scenarios. This paper constructs a test scheme of complex field environments with multiple areas and conditions，focusing on analyzing the influence of temperature，humidity，and elevation changes on the gravimeter’s measurement precision. The experimental results show that the evaluated gravimeter exhibits high measurement accuracy in typical special environments such as high temperature，low temperature，high humidity，and high altitude. The deviation from the reference value is better than 20 Gal，and the mean square error is better than 11 Gal，which can meet the current requirements of most field measurement and mapping. It provides an experimental method for carrying out the gravimeters’ evaluation in field scene，an idea for the further development of atom interferometry gravimeters，and the support for its practical application.

Since the explosion field temperature has the characteristics of rapid change，wide range and uneven distribution，it is always a difficult problem to measure the true temperature of the two-dimensional explosion field. Therefore，a common aperture multispectral explosion field transient and two-dimensional temperature measurement system is proposed. A multi-spectral imaging system based on aperture segmentation is studied to realize the simultaneous acquisition of temperature images of four spectra. The calibration of the experimental system is completed by high-speed camera and high temperature blackbody furnace. The relationship curve between the actual gray value ratio and the theoretical radiance ratio is fitted by polynomial，and the temperature is retrieved by combining the principle of colorimetric temperature measurement. The experimental results show that the developed system can obtain the two-dimensional explosion field multi-band image in real time and measure the temperature. The measurement range is 2 057~2 818 K，and the calibration temperature measurement accuracy error is not more than 1%. The system provides a technical approach for the information collection of transient 2-D temperature measurement，and is of great significance for the study of explosion mechanism and thermal damage effect.

In order to achieve gas component identification and concentration quantitative detection，an F-P tunable filter with a wavelength range of 9~12.6 microns for gas detection is designed in this paper. The specific optical elements of the MEMS microstructure that affect its optical performance are given special attention in the design. To solve the problem of Si substrate transmission at wavelengths above 9 μm，a high-transmission film composed of ZnS and BaF2 is designed，with a transmission rate of 97% or more in the wavelength range of 9~12.6 μm. The movable reflector and fixed reflector are both made of a high-reflective film composed of PbTe and BaF2，achieving a reflection rate of up to 98% in the designed wavelength range. The design is verified by simulation using TFCalc and MATLAB software，proving the rationality of the design. The final results show that the design can provide a reference for the design of tunable filters for long-wave applications，which can be widely used in gas sensors，environmental monitoring，medical diagnostic equipment，etc.，especially in miniaturized and integrated sensor systems.

To address the mismatch between the energy bands of the first exciton absorption peak of PbS quantum dots and the traditional electron transport layer（ETL）of ZnO—particularly when the absorption peak is red-shifted to 1 550 nm—SnO2，with a deeper conduction band，is used as an alternative ETL. This adjustment ensures better alignment of the energy bands between the large-size PbS quantum dots and the functional layers，which helps to reduce dark current，minimize response time，and enhance device performance by improving carrier mobility at the interfaces. The PbS quantum dots，exhibiting a first exciton absorption peak at 1550 nm，are synthesized via a thermal injection method with controlled oleic acid content. SnO2 films are then prepared using a cost-effective sol-gel process，varying precursor concentrations. Device testing reveals that with a precursor concentration of 0.15 mol/L，the dark current is approximately 10-7 A，the photocurrent reached 10-4 A，and the rectification ratio achieved 103. These results demonstrate a low-cost，high-quality，and reliable method for fabricating large-size PbS quantum dot-based photovoltaic devices.

To meet the requirements for conciseness and high accuracy in algorithms for detecting surface defects on high-precision optical components，with scratches and pockmarks as the research targets，a method for detecting surface defects on optical components based on an improved U2-Net is proposed. First，the network’s dataset is constructed using defect information from the surface of optical components. The improved U2-Net network is used for real-time training and testing on the surface defect dataset of the components to be applied，and ultimately，the new method is compared and analyzed with the previous U2-Net network. Experimental results show that the new network model has reached 95.7% in accuracy，91.3% in similarity coefficient，91.2% in intersection over union，and 91.3% in recall rate for key performance indicators. This technology is capable of resisting interference from noise points，used for detecting and displaying scratches and pockmarks defects in images，while improving the accuracy of defect segmentation and the accuracy of detection，achieving effective identification and segmentation of surface defects on optical components.

In order to analyze the influence of different etching liquids on residual salt，etching experiments on the surfaces of polished fused silica samples are designed and carried out. The surface morphologies of different etched samples are investigated by high magnification optical microscope and profilometer，and the residual salts are detected by X-ray photoelectron spectroscopy（XPS）. The results show that only a small amount of polished defects are exposed on the surface of the sample treated with a etching liquid containing sulfuric acid，with a surface roughness of Ra 0.014 6 m、Rq 0.025 2 m. The surface element detection results are highly similar to the non-etching sample，and no F element is detected. According to the analysis，adding sulfuric acid to the etching liquid will promote the dissolution of the fluorosilicates during the reaction process，which is more couducive to obtain a smooth and flat etched surface.

In order to reduce the size of endoscopic imaging probes to accommodate more usage environments and to achieve high-resolution ultrafine endoscopic imaging，an experimental study is conducted on a scanless，lensless，flexible optical endoscopic phase imaging technique based on coherent fiber bundles（CFB）. An imaging experimental setup is established using a 1 550 nm single-frequency narrow linewidth laser，a CFB，and a high-speed short-wave infrared camera. Point-target imaging experiments are carried out. The impact of the phase error introduced by the inconsistent length of the coherent fiber bundle on the imaging and the compensation method is studied，and a theoretical modeling analysis is carried out. The experimental results show that the phase error introduced by the core length of the coherent fiber bundle is relatively stable during the observation time of up to 5 hours，and can be effectively compensated. At the 10 cm away from the distal facet of the coherent fiber bundle，the phase error compensation can be achieved within the transverse shift of the point target of ±4 mm，but with the increase of the transverse shift of the point target，the phase error compensation becomes worse or even fails completely. Theoretical analysis and simulation results show that the angular correlation of phase error compensation is related to the multimode propagation of the optical field in the coherent fiber bundle，and the use of pure single-mode transmission of the coherent fiber bundle or the fine grinding of the two end faces of the coherent fiber bundle to eliminate the core length error can suppress the angular correlation of the phase error compensation and achieve effective compensation.

To develop a compact laser with high power as well as good beam quality，a composite Nd∶YAG/YAG transparent ceramic polygonal active mirror gain medium structure is reported，and conceptual validation is conducted by researching the laser output characteristics. The composite Nd∶YAG/YAG transparent ceramics used in this work are fabricated by the transparent ceramic technique. The five sides of the polygonal active mirror are Nd∶YAG doped regions，which are also laser gain mediums. The optical properties of the composite Nd∶YAG/YAG ceramics and the laser performances of polygonal active mirror are investigated. With a pump power of 4.05 W，a maximum output power of 0.43 W is obtained，and the corresponding optical-to-optical and slope efficiencies are 10.6% and 16.9%，respectively. The corresponding beam quality factors are M2X=1.21 and M2Y=1.24. The results of this study demonstrate the feasibility of using transparent ceramic techniques to fabricate laser gain mediums with functional structures，which brings more flexibility to the design of lasers.

The service life of the electrolyte material is a key factor in the operational stability of the flow cell. In this paper，based on the Debye vector integral theory and the diffraction integral theory，the theoretical simulation experiment of the propagation characteristics of the convergent beam at the dielectric interface of the flow cell is carried out. The propagation model of the convergent beam refracting through the dielectric interface is established by MATLAB programing，the intensity distribution of the fluid cell is calculated，and the experimental device for measuring the refractive index of the end mirror is developed. The propagation behavior of the convergent beam refracted by the dielectric interface is studied，and the variation curves of reflection coefficient of the light wave are obtained，and the possible factors affecting the excitation affecting the excitation efficiency of the output beam quality are analyzed in detail. It provides a theoretical basis for the study of dye laser output efficiency and service life of electrolyte materials.

With the rapid development of ultra-short pulse laser technology，picosecond laser cleaning technology has gradually been applied in the field of cleaning and preservation of precious cultural relics. This article conducts experimental research on the cleaning of surface contaminants on stone cultural relics using picosecond laser with a wavelength of 1 064 nm. Firstly，homemade stone cultural relic samples are used，and changes in surface roughness of the samples after focusing laser and defocusing laser cleaning are measured by confocal microscope，comparing the effects of defocusing state on laser cleaning. Based on the cleaning experiments of stone cultural relic samples，exploration of optimization parameters for laser cleaning of stone cultural relics is conducted. Through experiments，the optimized parameters for picosecond laser cleaning of stone cultural relics with a wavelength of 1 064 nm are obtained as follows：power density of 2 J/cm2，spot diameter of 110 m，scanning speed of 1 000 mm/s，scanning interval of 0.01 mm，and cleaning frequency of 7 times. The results suggest that the experimental results obtained from samples of stone artifacts are equally relevant to authentic artifacts. The experimental results presented in this paper bear significant importance in the exploration of optimization parameters for laser cleaning of stone artifacts.

In the research field of fringe-pattern phase analysis，the use of digital simulation and modeling tools for simulation experiments can improve the efficiency and convenience of research. However，due to the limitations of simulation algorithms and modeling tools，the data generated by simulation experiments often deviate significantly from the actual situation. To address this issue，a framework based on neural radiance fields（NeRF）is proposed. After training on a specific three-dimensional scene，this framework can render accurate fringe-pattern phase images for any given pose. Experimental results show that the mean squared error（MSE）of the phase within the region of interest（ROI）of the newly generated view images remains at the order of 10-5. When the dataset generated by this framework is used for deep learning training，it can achieve model accuracy at the same order of magnitude as that of a real dataset. The proposed framework can effectively generate novel-view fringe-pattern phase images for use in deep learning training.

To address the limitations of traditional occluded target imaging technologies，such as low resolution and poor recognition performance，a LiDAR-based 3D point cloud reconstruction technique for occluded targets is proposed in this paper. The method collects point cloud data from multiple perspectives to acquire precise datasets of the target hidden behind the occluder. First，an adaptive density-based clustering algorithm is adopted to make point cloud segments，refining subsets by filtering out irrelevant data to yield potential target fragments. Then，an improved ICP algorithm，combining ISS feature points with a KD-Tree，aligns and merges point clouds from multiple views. Subsequently，with a vehicle as the target，experiments are conducted with 70% coverage using a camouflage net. The results demonstrate that the algorithm effectively extracts and reconstructs the target's contours，achieving an extraction accuracy exceeding 96% and a dimensional error of less than 5%.