The structure and coloration of butterfly scales play a crucial role in regulating their biological behavior. Accurate three-dimensional imaging of these structures is essential for understanding their functional mechanisms. Traditional imaging techniques usually may suffer from limitations such as sample damage， slow imaging speed， and inaccurate three-dimensional reconstruction. To address these issues， this study employed Fourier Light Field Microscopy （FLFM） to image the three-dimensional structural details of butterfly scales. This imaging method requires no slicing or mounting of samples， thereby preserving their native structural integrity， avoiding damage during imaging， allowing repeated use of the same specimen， and significantly reducing sample consumption. In principle， it can perform the three-dimension imaging in real time by taking the same advantage of single shot three-dimensional imaging and image reconstruction of light field microscopy. Moreover， compared to the conventional light field microscopy， the FLFM can increase the lateral spatial resolution thus the reconstruction quality of the three-dimensional image. We found that FLFM is particularly suitable for delicate biological samples such as butterfly wings and provides reliable technical support for studying their color and structural characteristics.

Here， the representative samples including pigmentary scales， structural color scales， and scale-free areas of different kinds of butterflies were imaged under 10X and 50X objectives with a consistent lighting condition. After introducing the optical system and the imaging mechanism of the FLFM， the three-dimensional reconstruction method based on refocusing and deconvolution algorithms is described. The experimental FLFM setup consisted of a DOIT3DMicro Fourier light-field module mounted on an Olympus IX73 inverted microscope， and the illumination was provided in both transmission and reflection modes depending on the sample type. The refocusing algorithm adopts a shift-and-sum procedure exploiting the conjugate relation between aperture stop and MLA， while the deconvolution algorithm models the system point spread function via wavefront propagation theory for high-resolution reconstruction.

The method is verified by imaging with a resolution target sample and another sample of distributed beads　the resolution target image demonstrates the lateral spatial resolution about 8 μm and the beads sample image demonstrate the depth differentiation capability around 60 μm. Finally， the imaging results for the butterfly scales demonstrate that FLFM can clearly reveal morphological characteristics， angle-dependent color variations， and spatial distribution patterns of different scale types. Imaging at 10X magnification verified the feasibility of this technology for visualizing scale structures： the Papilio wing sample shows the relative wide black and white stripes with the boundary somewhat smeared， the Morpho helenor wing sample shows the fine well-organized colorful structures and the color can change from the green blue tone to the blue violet tone， and the Haetera piera wing sample shows shrunk scales and sparsely distributed cilia. Imaging at 50X magnification further revealed significant differences in scale morphology， color formation mechanisms， and optical properties between Catopsilia pomona and Morpho helenor： the color of the Catopsilia pamona wing sample appears mainly due to the pigment structures and does not change when viewing from different angles， while the Morpho helenor wing sample show typical structured color and the hue change apparently by varying the incident light angles or the viewing perspectives.

Therefore， this study demonstrates the advantage and potential wide application prospect of the FLFM in imaging the samples with large depth-of-field. Through careful examination and comparison analysis， we observed the various fine details of the three-dimensional structures of the butterfly wing scales which is very helpful for understanding the different color formation mechanism of the scales. We expect that the FLFM not only can find important applications for studying the structured color of the butterfly wings， but also can provide the experiment support for the further exploration and investigation of biomimetic materials.

Event cameras are a novel type of visual sensor inspired by the biological retina， capable of asynchronously capturing pixel brightness changes with microsecond-level temporal resolution. Unlike traditional cameras that capture frames at fixed intervals， event cameras only trigger an event when pixel brightness exceeds a preset threshold. This unique mechanism endows event cameras with an ultra-wide dynamic range of up to 140 decibels and exceptional temporal resolution， granting them significant advantages in high-speed， high-dynamic-range scenarios such as object tracking， motion detection， and Simultaneous Localization and Mapping （SLAM）. These capabilities enable event cameras to capture rapid movements and subtle changes in dynamic scenes that conventional cameras might miss. However， event cameras have a significant limitation： they can not directly capture grayscale images in static scenes. This is because when a scene is stationary and light intensity remains constant， the camera detects no change in light intensity， failing to meet the conditions for event triggering. These characteristics restrict their application in scenarios requiring complete brightness information， such as static object or scene recognition and measurement in traditional image processing tasks. Consequently， despite their ultra-high dynamic range， temporal resolution， and exceptional performance in dynamic environments， event cameras still face challenges in applications demanding detailed， continuous brightness data acquisition under static conditions.

To address the challenge of capturing static scenes with event cameras， this paper introduces a novel approach using mechanical shutter mapping for imaging. The core innovation lies in modulating brightness variations in static scenes through the rotation of a mechanical shutter， triggering events even in the absence of light intensity changes. By adjusting the shutter's rotation speed at varying rates， discrete events are generated， which are then reconstructed into High Dynamic Range （HDR） grayscale images. This technique overcomes a significant limitation of traditional event cameras， which struggle to capture static scenes due to the lack of natural light variations. A key aspect of this method is the calibration process， which ensures that the generated event data is accurately aligned with the actual light intensity of the scene， enabling precise and accurate grayscale image reconstruction. In this setup， the mechanical shutter acts as an active driving mechanism， inducing event generation by modulating the scene's brightness. This allows the event camera to capture dynamic information in static scenes， where traditional event cameras would otherwise fail to register changes. As a result， this method significantly enhances the imaging capabilities of event cameras， broadening their potential applications in static environments with minimal or no natural light variations.

Experimental results show that the reconstructed HDR images achieve a dynamic range of up to 112 dB. This range allows the event camera to capture fine details in both dark and bright areas of the scene， overcoming the limitations of conventional cameras with fixed dynamic ranges. The calibration process further improves the accuracy of grayscale image reconstruction. Our experimental results demonstrate that shutter-based techniques successfully reconstruct grayscale images of static scenes under extreme lighting conditions. Compared to prior methods， this approach significantly improves image quality， reduces grayscale distortion， and preserves details often lost in conventional event reconstruction techniques. By generating events through controlled brightness variations， this method enhances the imaging capabilities of event cameras in static environments. Furthermore， the mechanical shutter mechanism facilitates portability， enabling adaptation to multiple event camera models and enhancing the system's versatility and practical applicability. Beyond expanding event cameras' capabilities in static scenes， this approach also provides a solution for event cameras operating under complex lighting conditions. Our shutter mapping technology offers a controlled event triggering mechanism， enabling event cameras to function in scenarios where traditional methods fail. This opensnew possibilities for event cameras in fields requiring high-quality grayscale images， such as industrial inspection.

This paper proposes and validates a shutter-mapped event camera method for grayscale imaging in high-dynamic-range static scenes. The core innovation of this approach lies in introducing a mechanical shutter as an active driving mechanism， which effectively overcomes the limitation that event cameras cannot generate event streams in static scenes due to the absence of brightness changes. By modulating the scene's brightness using controlled shutter rotations， the event camera can trigger events even in stationary conditions. Experimental results demonstrate that this method successfully reconstructs grayscale images of high-dynamic-range static scenes， achieving a dynamic range of up to 112 dB under extreme illumination conditions， all using a frameless event camera. Compared to existing methods， our approach significantly improves reconstruction quality， reduces grayscale distortion， and preserves fine details. Additionally， the proposed mechanical shutter triggering mechanism offers versatility and portability， providing a controllable active refresh mechanism that is compatible with various event camera models. In summary， the shutter mapping method extends the capabilities of event cameras in static scenes， enabling high-quality HDR grayscale image reconstruction. By inducing brightness variations through mechanical shutter actuation， event cameras can generate event streams and reconstruct HDR images， even in static scenes with extreme lighting conditions.

Ammonia （NH₃） is a fundamental chemical agent extensively utilized in industrial synthesis， agricultural fertilizer production， and large-scale refrigeration systems. However， its leakage poses severe threats to both human physiological health and the atmospheric environment， necessitating the development of high-precision， real-time concentration monitoring techniques. Currently， Tunable Diode Laser Absorption Spectroscopy （TDLAS） combined with Wavelength Modulation Spectroscopy （WMS） is the preferred method for trace gas detection due to its high selectivity and sensitivity. Nevertheless， traditional TDLAS-WMS systems typically employ a low-frequency sawtooth wave for wavelength scanning. This conventional approach faces two critical technical bottlenecks： first， the inherent non-linearity and discontinuous nature of the sawtooth waveform lead to spectral energy dispersion and significant baseline fitting inaccuracies； second， the sawtooth scanning often overlaps with low-frequency environmental noise and mechanical vibrations， which expands the noise integration bandwidth of the lock-in amplifier and consequently limits the Signal-to-Noise Ratio （SNR） in complex industrial scenarios.

To overcome these limitations， this paper proposes an optimized detection method based on dual-sinusoidal compound modulation for second-harmonic signal extraction. The core innovation lies in the simultaneous application of a low-frequency sinusoidal scanning signal （f_s） and a high-frequency sinusoidal modulation signal （f_m） to achieve precise control of the laser frequency. Unlike the abrupt transitions of a sawtooth wave， the smooth and continuous nature of the dual-sinusoidal waveform effectively eliminates time-domain non-linearities， thereby suppressing baseline drifts at the source. Theoretically， this compound modulation concentrates the signal energy more efficiently near the specific harmonic sidebands （\mathrm{2}{f}_m）， which significantly reduces the bandwidth requirements for the detection system and enhances its immunity to \mathrm{1}/f noise. By utilizing narrow-band lock-in amplification to extract the normalized second-harmonic （2f/1f） signal， the proposed method ensures a more robust and stable representation of the gas absorption line shape， providing a theoretical foundation for higher detection accuracy and lower cost.

The experimental validation was conducted using a 1 512 nm Distributed Feedback （DFB） laser and a 100 cm quartz gas cell under controlled laboratory conditions. A comprehensive platform based on a high-speed data acquisition card （NI-USB6363） and LabVIEW was established to compare the performance of the dual-sinusoidal method against the traditional sawtooth-plus-sine modulation. Precision calibration was performed using a combination of spatial cavity and fiber etalons to optimize the modulation depth. The experimental results demonstrate a substantial performance enhancement： the linear fitting degree （R^{\mathrm{2}}） between NH₃ concentration and the second-harmonic amplitude improved from 0.991 9 to 0.997 8. Quantitatively， the detection sensitivity was enhanced from 0.089 ppm to 0.053 ppm， representing a significant improvement of approximately 40.5%. Furthermore， Allan variance analysis confirms that the system achieves its optimum detection limit within a shorter integration time of 81.5 seconds， exhibiting superior long-term stability and a more concentrated Gaussian distribution of measurement data compared to the traditional scanning mode. These results quantitatively confirm that the dual-sine approach ensures higher repeatability， lower detection limits， and faster convergence to optimal precision.

In conclusion， this research successfully demonstrates that dual-sinusoidal compound modulation is a highly feasible and stable approach for high-precision trace gas detection. By circumventing the intrinsic flaws of sawtooth scanning-specifically baseline fitting errors and high sensitivity to low-frequency noise—this method provides a new technical pathway for the robust， real-time monitoring of NH₃ industrial emissions. The findings verify that the optimized system can maintain high accuracy and stability even in environments with significant physical disturbances. Future research will focus on further refining the modulation parameters and integrating advanced digital signal processing algorithms， such as adaptive filtering or artificial neural networks， to address multi-component gas interference in even more complex scenarios. Additionally， the development of integrated， FPGA-based hardware platforms and the incorporation of multipass cells will be explored to further enhance the detection limits and promote the widespread application of this technology in portable environmental monitoring and industrial safety systems.

With the rapid development of precision manufacturing， semiconductor lithography， and aerospace engineering， ultra-precision machining and metrology have placed higher demands on micro-displacement measurement systems that combine nanometer-level precision with high-bandwidth real-time performance. In dynamic， high-speed motion scenarios， such systems not only need to achieve extremely high resolution but also provide low-latency data processing at high sampling rates. As a fundamental technology in modern length metrology， laser interferometry has inherent advantages in terms of accuracy and stability. However， traditional software-based processing architectures rely on CPUs and sequential execution models， often encountering severe computational bottlenecks when processing high-speed interferometric signals. These limitations make it difficult to simultaneously meet the conflicting demands of large measurement range， high resolution， and real-time detection. To overcome these challenges， this paper proposes a single-frequency laser displacement measurement system based on FPGA， which integrates an FPGA platform and a high-performance point-by-point phase calculation algorithm. By transferring the signal processing algorithm from software to a dedicated hardware logic platform and employing parallel computing and a pipelined approach， the system significantly improves real-time performance while maintaining nanometer-level measurement accuracy.

First， this paper establishes a four-channel orthogonal laser interferometry structure and develops a real-time phase demodulation method based on the phase difference between adjacent sampling points. By introducing phase jump correction and cumulative phase compensation mechanisms， the algorithm effectively solves the problems of phase unwrapping ambiguity and displacement direction discrimination under high-speed motion conditions， thereby ensuring the continuity and robustness of displacement measurement. Meanwhile， to analyze the impact of non-ideal optical elements on nonlinear errors， this paper constructs a systematic error model using Jones matrix theory. This model quantitatively analyzes the effects of waveplate angle misalignment and Polarization Beam Splitter （PBS） splitting ratio imbalance on the interference signal， revealing their roles in introducing DC offset， amplitude mismatch， and non-orthogonality between orthogonal channels. Based on this analysis， this paper designs a vector-based error compensation algorithm to correct these defects， thereby significantly improving phase linearity and overall measurement accuracy. Based on the proposed algorithm， a complete experimental test platform was established， and a host-computer-based multi-threaded displacement measurement software was developed to evaluate the system performance. Experimental results demonstrate that， within a measurement range of 200 μm， the system achieves a statistical measurement accuracy better than 10 nm while supporting millisecond-level data update rates， thereby validating the effectiveness and practicality of the proposed method.

Based on the validated theoretical and software framework， this paper further develops an FPGA hardware system integrating a high-speed multi-channel ADC to comprehensively enhance real-time processing capabilities. By employing a fully pipelined architecture and dedicated parallel processing units， key algorithms such as error compensation and phase demodulation are transformed into hardware logic. This hardware-oriented design ensures stable and predictable computational latency， which is crucial for real-time measurement. Experimental comparisons show that the results of the FPGA-based hardware processing are in high agreement with the results of the host multi-threaded software algorithm， with a Root Mean Square Error （RMSE） of only 9.16 nm and statistical accuracy consistently maintained within 10 nm. Real-time performance evaluation further validates the advantages of the proposed system. The initial processing latency is reduced to 10.4 μs， and the displacement data update cycle reaches 100 ns. Therefore， the system can reconstruct complete micrometer-level motion trajectories in real time within 1 millisecond， representing a significant performance improvement compared to traditional CPU-based processing architectures. These performance characteristics make the system particularly suitable for applications such as high-speed motion， dynamic vibration monitoring， and real-time feedback control.

In summary， this paper proposes a laser interferometric real-time displacement measurement system by deeply integrating a point-by-point phase demodulation algorithm and FPGA-based hardware acceleration technology. This system simultaneously achieves high precision， high stability， and ultra-fast real-time performance. It provides a reliable and high-performance metrology solution for closed-loop motion control， dynamic vibration analysis， and nanometer-level precision positioning in ultra-precision engineering applications， demonstrating significant practical value and broad application prospects.

The purpose of this study is to develop a robust and accurate Phase Unwrapping （PU） algorithm that can reliably recover continuous phase information from noisy wrapped phase data while preserving structural details. In many practical measurements， noise and local phase disturbances degrade the accuracy of gradient estimation， which leads to error amplification and structural distortion during the unwrapping process. To address this challenge， this work proposes a new PU framework based on gradient consistency detection and adaptive noise correction. The main objective is to accurately identify noise-corrupted pixels in the wrapped phase map and selectively suppress their influence without altering reliable phase information. By combining localized complex-domain filtering with efficient phase reconstruction， the proposed method aims to improve both reconstruction accuracy and robustness while maintaining computational efficiency suitable for practical imaging systems.

The proposed method consists of three main stages　dual wrapped phase generation， gradient consistency-based noise detection， and integral phase reconstruction. First， two wrapped phase maps are obtained from the same group of fringe patterns through forward and reverse phase-shifting demodulation. Specifically， phase extraction is performed using five fringe images in the sequences I_{\mathrm{1}}\to I_{\mathrm{5}} and I_{\mathrm{5}}\to I_{\mathrm{1}}， respectively. These two wrapped phase maps represent the same underlying phase distribution but contain opposite phase-shift directions introduced during demodulation. Under ideal noise-free conditions， the spatial gradients of the two wrapped phase maps should be identical at each pixel location. Based on this property， the proposed algorithm detects noise by evaluating the gradient consistency between the primary wrapped phase and the auxiliary wrapped phase. The gradients of the two maps are computed and compared pixel-wise， and pixels whose gradient difference exceeds a predefined threshold are classified as noisy pixels. This detection strategy enables accurate localization of noise while avoiding unnecessary processing of reliable regions. Once the noisy pixels are identified， localized filtering is applied only to these pixels in the complex domain. The wrapped phase is first transformed into a complex representation to eliminate discontinuity issues caused by the 2π periodicity. A local mean filtering operation is then iteratively performed to suppress noise while maintaining phase continuity. To prevent excessive smoothing and loss of structural details， a gradient-based stopping criterion is introduced. The filtering process for each pixel terminates once the corrected gradient becomes sufficiently consistent with the reference gradient or when the maximum number of iterations is reached. This selective filtering mechanism effectively reduces noise while preserving reliable phase structures. After the noise suppression stage， the continuous phase is reconstructed through gradient integration. Instead of solving large optimization problems， the proposed method employs a fast integral reconstruction strategy. Starting from a reference point， the unwrapped phase is obtained by accumulating the denoised gradient values across the image. This approach avoids nonlinear optimization and large matrix operations， thereby significantly improving computational efficiency.

Extensive experiments were conducted using both simulated datasets and real measurement data to evaluate the performance of the proposed method. In the simulation experiments， wrapped phase maps with different noise levels were generated to analyze the robustness of the algorithm under controlled conditions. The proposed method was compared with several representative phase unwrapping techniques， including the Total Variation （TV） based method， the Least-Squares （LS） method， and the Transport-of-Intensity Equation （TIE） method. Quantitative evaluation was performed using multiple performance metrics， including Root Mean Square Error （RMSE）， Structural Similarity Index （SSIM）， and computational time. The results show that the proposed method consistently achieves lower RMSE values and higher SSIM scores across all noise levels. When the noise level is low， the proposed algorithm produces highly accurate phase reconstructions comparable to existing approaches. As the noise level increases， the advantages of the proposed method become more evident. Traditional LS and TIE methods exhibit significant error amplification due to inaccurate gradient estimation， while TV-based methods tend to oversmooth phase structures and blur sharp edges. In contrast， the proposed method effectively suppresses noise while preserving phase discontinuities and structural details. To further evaluate the stability of the algorithm， repeated experiments were performed under identical experimental conditions. Ten independent runs were conducted with different random noise realizations. The results show that the RMSE values obtained by the proposed method remain consistently low with minimal variation， indicating strong robustness and stability. In contrast， the TV method exhibits noticeable fluctuations in reconstruction error across different runs. Real-world experiments were also carried out using fringe images captured from physical samples. The test object consisted of stacked wafer surfaces forming a discontinuous phase structure. The proposed method successfully reconstructed the phase distribution and preserved sharp structural boundaries. Two-dimensional profile comparisons further demonstrate that the reconstructed phase closely matches the reference phase and maintains accurate edge information. Additionally， the computational time of the proposed method is significantly shorter than that of the TV-based method， demonstrating its efficiency advantage.

The experimental results demonstrate that the proposed phase unwrapping algorithm achieves high reconstruction accuracy， strong robustness， and efficient computational performance under various noise conditions. The method effectively suppresses noise while preserving structural details， providing a reliable solution for phase reconstruction in practical measurement environments.

High-efficiency solar cells based on nanostructures are a research hotspot in the photovoltaic field today. On the one hand， there is the research on new types of solar cells such as perovskite cells. On the other hand， silicon-based solar cells remain the most widely studied and applied type of cell. How to further improve the efficiency of silicon-based solar cells and reduce their cost still requires in-depth research. Silicon is an indirect bandgap semiconductor material， and its absorption of incident light in the near-infrared band is very weak. Only when the thickness of the absorption layer exceeds 200 µm can the cell achieve full absorption in a wide band （300~1 100 nm）. Reasonable light-trapping nanostructure design can enhance the light absorption performance of silicon thin-film solar cells. The morphology， size， material and spatial layout of these light-trapping nanostructures have a significant impact on the surface anti-reflection characteristics and light-trapping performance of silicon thin-film solar cells.

In this paper， a double-layer conformal grating silicon thin-film solar cell structure was designed. The influence of different grating shapes on the light-trapping performance of silicon thin-film solar cells was systematically studied by using the concept of super-quadratic subwavelength gratings. By changing the degree of the grating’s profile function， the profile shape of the grating can be adjusted. Five representative grating profile curves （n=1， 2， 4， 10， ∞） were selected. As n increases， the grating profile shape can be gradually changed from triangular （n=1）， parabolic （n=2）， to rectangular （n→∞）. The silicon thin-film solar cells with the above five grating structures were optimized by using the finite-difference time-domain method. By comparing and analyzing the short-circuit current density， absorption spectrum， and light absorption enhancement spectrum of the five grating structures and the planar structure， the light-trapping effect law of the double-layer conformal grating silicon thin-film solar cell was explored. The internal light-trapping mechanism was analyzed using the electromagnetic field intensity distribution diagrams. In addition to qualitatively analyzing the light-trapping performance of each grating structure silicon thin-film solar cell， the concept of integral absorption was also introduced to quantitatively analyze their light absorption characteristics over the entire wavelength range. Moreover， the optimization calculation of the silicon thin-film solar cells with four types of bottom metal grating materials （Ag， Au， Cu， and Al） was carried out. A comparative analysis was conducted on the short-circuit current densities and absorption spectra of four types of bottom metal grating structures， and the light absorption enhancement effect of different bottom metal grating materials was explored. The mechanism was analyzed from the perspective of optical constants.

The research results indicate that as n increases， the overall light-trapping performance of silicon thin-film solar cells decreases as the grating profile shape transforms from triangular （n = 1） to rectangular （n→∞）. Among them， the parabolic （n = 2） grating structure has superior light absorption performance in both the 300~750 nm and 750~1 100 nm bands. Because the n = 2 grating structure has excellent anti-reflection performance in the visible light band， and the LSPRs mode is excited on the surface of the bottom Ag grating in the near-infrared band， while optical waveguide mode coupling is achieved in the c-Si absorption layer. The triangular （n = 1） grating structure has better light-trapping performance in the 750~1 100 nm band， but its light absorption performance in the 300~750 nm band is poor. The results also show that the J_sc values of the silicon thin-film solar cells with the bottom metal gratings of Ag， Au， Cu and Al decrease in sequence， among which the bottom Ag grating structure has the best light-trapping performance and the bottom Al grating structure has the worst. The research results of this paper can provide theoretical guidance for the structural design of thin-film solar cells based on nano-grating structures.

Silicon nitride （Si₃N₄） photonic integrated circuits are promising for next-generation optical communications due to their low loss and wide transparency window. However， efficient coupling between submicron Si₃N₄ waveguides and standard optical fibers remains challenging due to mode-field mismatch. Existing solutions like inverse tapers or polymer claddings involve trade-offs among bandwidth， loss， and manufacturability. This work proposes a taper-straight-taper edge coupler， where the straight section acts as a mode stabilizer and reflection suppressor. The optimized device achieves sub-decibel coupling loss across the S+C+L bands with high alignment tolerance， offering a compact and fabrication-friendly solution for Si₃N₄ photonic integration.

To overcome these limitations， this work proposes and demonstrates an innovative edge coupler architecture that combines two inverse tapers with an intermediate straight waveguide section. The straight section is introduced as a mode stabilizer and reflection suppressor， effectively filtering residual higher-order modes and eliminating parasitic reflections through phase-matching interference. This configuration allows the optical mode to evolve smoothly from the tightly confined on-chip waveguide to the expanded fiber mode， while maintaining high fundamental-mode purity and minimal spectral ripple. The design principle ensures adiabatic mode conversion without relying on extreme geometries or high-aspect-ratio structures， thus preserving compatibility with standard photolithography and etching processes.

Comprehensive electromagnetic simulations based on eigenmode expansion were employed to optimize the key design parameters， including the taper lengths， transition widths， and the straight section geometry. The optimized configuration exhibits an excellent balance between low coupling loss， wide operational bandwidth， and compact device size. The inclusion of the straight section significantly enhances the stability of the mode conversion process and mitigates the oscillations typically observed in conventional two-taper couplers. The simulation results predict broadband， flat coupling performance with high tolerance to fabrication deviations and fiber alignment errors， confirming the robustness of the proposed design.

Experimental characterization was performed on fabricated devices covering the S， C， and L communication bands. The measured results show broadband low-loss performance with minimal wavelength dependence and excellent alignment tolerance. Specifically， the coupler achieves sub-decibel coupling loss per facet over a wide spectral range， with the wavelength-dependent variation remaining below one decibel across the entire operational window. The measured results agree well with the theoretical predictions， confirming that the phase-matching mechanism in the straight section effectively suppresses reflection-induced interference and improves modal stability. Furthermore， the design maintains low loss under several micrometers of lateral or vertical fiber misalignment， demonstrating strong packaging robustness suitable for large-scale manufacturing.

The proposed edge coupler achieves these results without the need for complex processes such as electron-beam lithography or high-aspect-ratio etching. Its fully planar structure and moderate feature sizes ensure high yield and compatibility with commercial foundry processes. The broadband and low-loss characteristics make it particularly advantageous for integrated systems that require efficient chip-to-fiber interfaces， such as coherent transceivers， frequency comb sources， and integrated photonic sensors.

In summary， this study introduces a broadband， low-loss， and fabrication-tolerant silicon nitride edge coupler based on a taper-straight-taper configuration. By integrating a straight single-mode section between two inverse tapers， the design simultaneously achieves mode stabilization， reflection suppression， and efficient field expansion. Both simulation and experimental results demonstrate superior performance across a wide wavelength range， with low insertion loss， minimal wavelength sensitivity， and high alignment tolerance. The proposed structure provides a compact， manufacturable， and scalable coupling solution for silicon nitride photonic integrated circuits， enabling their deployment in next-generation high-speed optical communication and broadband photonic applications.

The performance and stability of underwater wet-mate connectors are of great significance for guaranteeing the long - term reliable operation and safety of subsea engineering systems. In recent years， while some domestic research institutions and enterprises have successfully developed prototype models based on relevant principles， domestic products still face the challenge of insufficient market recognition in the competition. This is due to the lack of efficient， low-cost， and full-life-cycle system testing methods， dedicated equipment， and a comprehensive sea-trial verification system. This paper commences from the perspective of enhancing the reliability of domestically-produced underwater electric-optical hybrid wet-mate connectors. A systematic series of tests were carried out on the self-developed connector prototypes， and a set of automatic connecting test devices suitable for underwater pressure environments was successfully developed， enabling automatic cyclic mating and demating operations of the connectors. The outcomes of sea trials indicate that the self-developed connectors successfully accomplished 100 consecutive mating and demating operations at a water depth of 2 960 meters， and all performance indicators met the technical requirements of relevant foreign standards， attaining an international advanced level.

Phase unwrapping serves as a critical step in high-precision optical measurement， aiming to recover the absolute phase from its wrapped counterpart. While current deep learning-based approaches for absolute phase extraction have achieved improvements in both accuracy and computational efficiency， critical challenges persist， including excessively large network parameter counts and low accuracy in resolving discontinuous phases. To overcome these limitations， this paper introduces the Attentive Directional Encoder-Decoder Network （ADE-Net） for structured light phase unwrapping applications. The primary objective is to develop a lightweight， efficient， and robust deep learning framework that achieves consistently high unwrapping accuracy under complex conditions such as noise corruption， phase discontinuities， and aliasing artefacts， thereby providing a practical， engineering-ready solution capable for reliable development in real-world optical measurement systems.

The proposed ADE-Net is built upon a concise and efficient encoder-decoder backbone architecture. The encoder consists of four feature extraction stages， each comprising a convolutional block and a Haar Wavelet Down-sampling （HWD） module. The HWD structure preserves high-frequency directional details， such as phase jump edges， during down-sampling， providing a more complete feature foundation for subsequent processing. The decoder employs a symmetric up-sampling structure based on transposed convolutions， with skip connections to fuse multi-scale features from the encoder， ensuring effective detail recovery. To enhance feature representation， two innovative modules are integrated： the Adaptive Clockwork Long Short-Term Memory （AC-LSTM） module and the Depthwise-separable-convolution Efficient Multi-Scale Attention （DEMA） module. The AC-LSTM module is designed to capture multi-scale contextual and direction-sensitive features simultaneously. It consists of two parallel branches： an Atrous Spatial Pyramid Pooling （ASPP） branch for multi-scale context aggregation， and a bidirectional LSTM （Bi-LSTM） branch enhanced with Context Anchor Attention （CAA） to model long-range dependencies along horizontal and vertical directions with spatial awareness. The outputs of both branches are fused to produce rich， direction-aware features. The DEMA module is an improved version of the Efficient Multi-Scale Attention （EMA） mechanism. It introduces depthwise separable convolution to replace standard 3×3 convolutions， significantly reducing computational complexity while maintaining receptive field coverage. Additionally， Batch Matrix Multiply （BMM） is adopted for more efficient parallel processing of batch data. DEMA operates by grouping input features， performing adaptive height and width pooling， and applying depthwise separable convolution in parallel branches， followed by cross-branch fusion via BMM and element-wise multiplication with the original features. This design enhances the model's ability to focus on critical information across scales while optimizing computational efficiency.

To comprehensively evaluate the performance of ADE-Net， systematic experiments were conducted using both simulated and real-world datasets. For simulation， three distinct datasets were constructed， each comprising 2 000 pairs of wrapped and absolute phase images with a size of 256×256. Dataset 1 simulated noise interference by introducing Gaussian noise with signal-to-noise ratios of 0， 5， 20， 40， and 80 dB. Dataset 2 simulated phase discontinuities by randomly generating 1 to 2 rectangular jump regions with random positions， sizes， and phase values. Dataset 3 combined both noise and discontinuities to emulate aliasing effects. Additionally， the public real dataset ‘Single-input dual-output 3D shape reconstruction’ was adopted to assess generalization capability， which contains 1 500 real-scene samples acquired by a structured-light 3D measurement system. The evaluation was performed using three metrics： Normalized Root Mean Square Error （NRMSE）， Mean Absolute Error （MAE）， and Structural Similarity Index Measure （SSIM）. Benchmark comparisons were carried out against mainstream models such as U-Net， Res-UNet， Perera's Net， and TNUNet. In the comparative experiments， ADE-Net achieved the best performance across all three simulated datasets. On the discontinuous dataset， it attained an NRMSE of 4.35%， an MAE of 1.094 5 rad， and an SSIM of 0.871 4. On the random-noise dataset， the results were 4.58% NRMSE， 1.216 5 rad MAE， and 0.838 9 SSIM. On the aliased dataset， ADE-Net reached 4.34% NRMSE， 1.137 7 rad MAE， and 0.834 0 SSIM. Cross-validation on the public dataset demonstrated excellent generalization ability， with a Root Mean Square Error （RMSE） of 0.505 2， an MAE of 0.180 9 rad， and an SSIM of 0.997 1. Ablation studies further confirmed the individual performance contributions of both the AC-LSTM and DEMA modules. Moreover， while maintaining the aforementioned high accuracy， ADE-Net also exhibited remarkable computational efficiency advantages. With only 2.012 M parameters and 2.655 GFLOPs， it is significantly lighter than the compared models. The inference time per single image is only 12.71 ms， meeting real-time processing requirements without compromising accuracy， thereby achieving an effective balance between model complexity and computational performance.

In conclusion， ADE-Net presents a lightweight， efficient， and highly accurate deep learning framework for phase unwrapping in structured light 3D measurement. By integrating the AC-LSTM module for capturing direction-sensitive multiscale features and the DEMA module for computationally streamlined attention-driven feature refinement， the models achieve an ideal balance between performance and computational complexity. It demonstrates superior resilience in challenging conditions involving noise corruption， phase discontinuities， and aliasing artefacts， outperforming leading benchmark models in both accuracy and processing efficiency. Rigorous validation on real-world datasets confirms its robust generalization capability， while its fast inference speed and low resource footprint make it well-suited for high-precision phase reconstruction tasks. By addressing practical constrains in deployment， this work promotes the feasibility of structured light 3D measurement system for real-world， resource-aware applications.

Encapsulation is a critical step in semiconductor manufacturing， providing chips with a stable and protected operational environment through integration into compatible external structures. High-precision and non-destructive surface morphology measurement of transparent encapsulated devices is essential， as it enables early detection of latent defects， supports process optimization， and ensures long-term device reliability. However， conventional white light interferometric microscopy faces inherent limitations when applied to encapsulated samples. Existing optical-path compensation schemes offer only restricted compensation ranges， and even small residual mismatches between the encapsulation layer and the compensation glass can introduce monochromatic aberrations and chromatic dispersion. These residual errors degrade fringe contrast， distort the recovered phase， and can ultimately lead to failure in reconstructing surface morphology.

To address these challenges， we propose a dual-arm differential adjustable compensation strategy for white light interferometric microscopy. The method is designed to address the thickness mismatch and material dispersion when measuring encapsulated microstructures. The system incorporates two independently adjustable compensators： a plate compensator inserted into the test arm and a pair of wedge plates placed in the reference arm. The wedge plates provide a thickness adjustment range from 4.8 mm to 5.2 mm enabling the system to realize a continuously adjustable compensation range. The two compensators form a continuously adjustable configuration capable of correcting up to 400 μm of thickness deviation. This design supports stable compensation for encapsulation thicknesses from 0 mm to 5.2 mm. A 2.5× Michelson-type interferometric objective is employed in order to conveniently separate monochromatic aberrations from chromatic dispersion， thereby enabling a focused investigation of dispersion-induced phase distortions and the effectiveness of the proposed compensation strategy.

Theoretical modelling and numerical simulations were conducted to assess the feasibility and robustness of the proposed compensation strategy. The modelling incorporated the spectral characteristics of the light source， the refractive-index dispersion of the encapsulation materials， and the numerical aperture of the interferometric objective. A detailed quantitative analysis was carried out to evaluate the influence of light-source bandwidth， the intensity ratio between the two interferometric arms， and residual thickness mismatch on the achievable fringe contrast. The results highlight the necessity of incorporating adjustable wedge plates to achieve continuous and precise dispersion compensation. Based on the simulation outcomes， acceptable tolerance ranges for the residual mismatch were derived， providing practical design guidelines for implementing adjustable dispersion compensation in interferometric microscopy platforms.

Experimental validation was conducted using three representative classes of samples to verify the measurement accuracy and system stability. In addition to validating measurement accuracy， these experiments were also designed to evaluate the long-term operational stability of the compensation mechanism， especially under varying encapsulation thicknesses. First， a Bruker RM1722 roughness standard was used to evaluate surface roughness under encapsulation. The dynamic adjustment of the wedge plates in the reference arm effectively preserved high-contrast fringes. The encapsulated sample exhibited an arithmetic average roughness of （0.63 ± 0.003） nm， which closely matched the value of （0.61 ± 0.004） nm for the unpackaged sample. Second， a standard SHS-23008 step height specimen was measured beneath a 1 mm K9 cover plate to simulate the encapsulation condition. The reconstructed step height was （9.272±0.021） μm， demonstrating both the accuracy of the compensation method and the robustness of the phase retrieval process. Finally， the technique was applied to TO-packaged photodiodes to characterize the surface morphology of the encapsulated chip region. The arithmetic mean height and root mean square height of the surface were 3.855 nm and 3.892， respectively， indicating the good fabrication quality. These comprehensive results confirm that the proposed dual-arm differential adjustable interferometric method is feasible， reliable， and highly accurate for characterizing the surface morphology of transparent encapsulated microstructures. The demonstrated performance establishes the method as an effective and practical solution for high-precision， non-contact morphology evaluation of transparent encapsulated devices.