This PDF file contains the front matter associated with SPIE Proceedings Volume 12772, including the Title Page, Copyright information, Table of Contents, and Conference Committee information.

We present study of polarization properties of the narrow modes, generated in random Raman fiber lasers near the generation threshold. For this purpose, time and polarization resolved spectral measurements based on optical heterodyning technique were implemented, that allow reconstruction of the ratio of vertical and horizontal projections of the electrical field during the mode generation process. We revealed that modes have high degree of polarization, with the slow change of its state during the mode lifetime. Moreover, it appeared that each mode has its own randomly appeared state of polarization, even when the several modes are generated simultaneously.

The time-stretch imaging system is a promising method for achieving real-time imaging and low-latency cell screening. To facilitate the evaluation of time-stretch imaging systems for cell detection, we present a simulator for phase recovery in a time-stretch quantitative phase imaging (TS-QPI) system. The simulator enables the efficient evaluation of TS-QPI system, demonstrating the feasibility of accurate phase recovery for a wide range of cell screening conditions and designing the TS-QPI systems depending on the characteristics of the target cells. Furthermore, it allows synthesis of simulated phase images highly beneficial in data augmentation when training machine learning models for cell detection.

A novel signal denoising framework (EEMD-VMD-IMWOA) for Rayleigh lidar is proposed to better suppress noise in an atmospheric lidar echo signal and improve retrieval accuracy. The ensemble empirical mode decomposition (EEMD) is used to retain the intrinsic mode functions (IMFs) of signal as the low-frequency effective component. Based on the denoising ability of variational mode decomposition (VMD) under high noise signal, the IMFs with noise is further denoised by VMD to obtain high-frequency effective component, wherein the improved whale optimization algorithm (IMWOA) is used to get the optimal decomposition layer K and the quadratic penalty α of VMD. Then, the low-frequency and high-frequency effective components are reconstructed to gain denoised signal. The simulation results show that the denoising effect of EEMD-VMD-IMWOA is superior to Wavelet threshold, EEMD and VMD, especially the far-field noise interference can be suppressed. Under the condition that the temperature retrieval error is less than ± 10 K, when the integration time is only 600s, the effective retrieval altitude can reach 59.6km, which is 17.3% higher than that without denoising. Finally, the retrieval accuracy of the measured lidar signal is significantly improved by EEMD-VMD-IMWOA.

The photoacoustic sensor system based on Sagnac interference has extensive applications in perimeter security. Among them, the demand for synchronization recognition of multiple acoustic sources and integrated utilization becomes the key to expand its application field. By focusing on multi-source synchronization recognition, a lightweight optical-acoustic signal recognition algorithm that can be carried by an embedded system was studied. The collected optical-acoustic signals were extracted through preprocessing, Fast Fourier Transform (FFT), Mel filter bank (MFB), and other steps to obtain Mel spectrogram features, which were then used for optical-acoustic signal recognition using the MobilenetV3 network. The experimental results showed that this system achieved a synchronization recognition rate of 93% for six types of sounds, providing a possibility for implementing multi-sound recognition in embedded systems based on Sagnac interference photoacoustic sensor systems.

Fiber optic acoustic sensing system based on Sagnac interference is widely used in the field of perimeter security. However, the sensitivity of acoustic response varies drastically with frequency, which has become a bottleneck limiting the development of Sagnac photoacoustic sensing technology. An elastic structure is proposed to improve the sensitivity consistency of the photoacoustic sensing system based on Sagnac interference at different frequencies. The elastic structure of optical fiber pickup was optimized by finite element analysis, Solidworks and Comsol. Simulation results show that the designed titanium alloy-based elastomeric structure with a diameter of 110 mm, a height of 100 mm, and an elastomer wall thickness of 3 mm has significantly improved the consistency of the sensitivity of the acoustic response at different frequencies. This work provides a possibility for the application scenario expansion of photoacoustic sensing system based on Sagnac interference.

Talbot self-imaging has been utilized for arbitrary repetition-rate control of optical pulse trains and frequency combs, which are important for many applications. The fundamental theory of generalized Talbot self-imaging has been presented for the design of temporal and spectral phase operations, while the effects of potential phase modulation distortions and residual dispersions on repetition-rate control in experiments remain to be investigated. In this paper, numerical studies are conducted for evaluating the feasibility of yielding arbitrary repetition-rate control of optical pulse trains and frequency combs, and it provides additional insights into the optimal design conditions for practical implementation of the scheme.

Electronic information systems are highly sensitive to electromagnetic signals, rendering them vulnerable to electromagnetic weapon attacks. The presence of metal electrodes and wires in these systems prevents them from being immune to attacks from electromagnetic weapons. This study specifically focuses on enhancing the RF signal receiving front-end, comprising a lithium niobate modulation chip and a full dielectric antenna. We utilized the finite-difference time-domain (FDTD) method to simulate and optimize the chip structure of the lithium niobate waveguide. Based on the optimization results, a two-step etching process was employed to fabricate the chip. The micro-ring in the Ku-band photonic RF front-end has been experimentally measured to possess a Q-factor of 74,000, with an instantaneous bandwidth of 2.5 GHz. This research holds significant implications for safeguarding electronic information systems against the potential damage caused by electromagnetic forces.

With the development of the economy and the society, spectrum resources of higher frequencies are becoming increasingly scarce. Beneficial from the photonic technology, optoelectronic oscillators (OEOs) have the advantages in generating microwave signal with high center frequency and low phase noise. However, applications of OEO systems are limited by its bulky size. In this work, a hybrid integrated OEO is proposed and experimentally demonstrated. A high integration level is achieved by assembling all the optical and electrical chips. A compact fiber ring and a YIG filter are also well packaged. At the oscillation frequency of 10 GHz, phase noise of the proposed OEO is -115.83 dBc/Hz@10 kHz. Wideband frequency tuning from 3 GHz to 18 GHz is also realized, the phase noise is better than -110 dBc/Hz @10 kHz at the entire tuning range. This work shows the great potential of integrated OEO in a wide range of applications such as wireless communications and satellite communications.

Heterodyne dual-frequency laser interference displacement measurement has been widely applied in the measurement in precision mechanical systems. A signal modulation circuit with four-phase quadrant-square wave output was proposed to improve the sensitivity of displacement detection in the differential dual-frequency laser interference displacement measurement system. A front-end photoelectric conversion circuit, mixing circuit, waveform conversion circuit and fourway quadrantal square wave shaping circuit are designed and manufactured. Heterodyne dual-frequency laser interference displacement testing platform was built in the laboratory and the circuit test was finished. The detection speed of the system reach 3m/s, the measurement frequency range is 3.3-10MHz, and the accuracy reach 0.5ppm. The theoretical resolution of the system is a quarter of the laser vacuum wavelength, which is about 0.158 μm.

Alzheimer's disease (AD) is a neurodegenerative disorder that affects the life quality of millions of people worldwide. To diagnose new cases in a timely manner, we propose a new novelty detection technique that combines Autoencoder and Minimum Covariance Determinant (MCD). The technique consists of two steps: first, we use an Autoencoder to extract low-dimensional and discriminative features from the publicly available ADNI dataset, where we only train the Autoencoder with normal data, making the abnormal data more distinguishable in the feature space; second, based on the features of normal data, we use MCD to construct a decision boundary, and judge the degree of abnormality by the distance of the test point to the boundary. Compared with traditional methods without using Autoencoder, our technique has significant advantages in terms of accuracy and sensitivity, and can effectively deal with data imbalance problem. Experimental results show that our method can efficiently detect novel AD cases, and has a wide range of application prospects.

The high sensitivity and picosecond temporal resolution of single-photon avalanche diode (SPAD) make it the preferred single-photon detector in extreme imaging environments. Extreme imaging environments (e.g., underwater high-scattering environments) usually result in low signal-to-noise ratios of the acquired single-photon data, which leads to poor quality of image reconstruction, so it is necessary to propose a high-resolution single-photon three-dimensional reconstruction algorithm for extreme imaging environments. Principal component analysis (PCA) is widely used and robust, which is suitable for dimensionality reduction and noise reduction processing of single-photon data with sparse and noisy characteristics. Under the premise that the target data has a strong correlation with the background and random noise, the target feature extraction of the single-photon data is carried out by PCA, the principal components are used to reconstruct the original data, the relative position and size of the original data are effectively retained, the redundant information is removed, and the single-photon data is reconstructed using cross-correlation and ManiPoP algorithms to achieve high-resolution single-photon depth profile reconstruction.

Augmented reality is a visualization technology that displays information by adding virtual images to the real world. In many cases, augmented reality requires recognition of the current scene. Extracting foreground objects from a video in real-time on limited hardware, such as a smartphone, is demanding. The augmented reality for the environment must have a model in which it is clear where it is necessary to detect the object or apply a mask. One way to recognize a scene in the current context without prior information is to use semantic segmentation techniques. This article proposes a new neural network architecture for efficient semantic image segmentation in the task of building augmented reality. The developed architecture is based on the combination of Shufflenet V2 and DPC, which provides good performance due to the balance between predictive accuracy and efficiency. First, the ShuffleNet V2 neural network architecture obtains features from RGB images. The resulting feature maps are then passed to one of the Deeplab V3+ Dense Prediction Cell encoders. At the final stage, the features are decoded by bilinear interpolation to create segmentation masks. The augmented reality construction algorithm is based on the ARCore framework and the OpenGL interface for embedded systems. The proposed approach for recognizing scene objects in augmented reality uses semantic segmentation, providing real-time information. The implementation of the algorithm shows that the detected objects can be tracked in 3-D space using visual-inertial odometry without resorting to constantly updating the environment model. The frequency of object detection and semantic mask generation can be reduced, resulting in battery and processing power savings, which is critical for mobile and embedded systems. The semantic information provided by these solutions can be used in autonomous driving, robotics navigation, localization, and scene recognition in conditions of limited resources. In augmented reality, the proposed approach can remove objects from a scene, draw attention to objects, or provide scene recognition to software logic. The experiment results confirmed the high efficiency of the proposed method compared to the state-of-the-art techniques for real‐time 3-D augmented reality construction.