Abstract:A sensitive room-temperature metal-semiconductor-metal (MSM) structure is fabricated on high-resistivity silicon substrates (ρ>4 000 Ω.cm) for terahertz (THz) detection by utilizing the photoconductive effect. When radiation is absorbed by the nitrogen-rich niobium nitride, the number of free electrons and electrical conductivity increase. The detector without an attached antenna boasts a voltage responsivity of 7 058 V/W at a frequency of 310 GHz as well as small noise density of 3.5 nV/Hz0.5 for a noise equivalent power of about 0.5 pW/Hz0.5. The device fabricated by the standard silicon processing technology has large potential in high-sensitivity THz remote sensing, communication, and materials detection.

Abstract:A novel high quality-factor (Q) micro-ring resonator (MRR) structure based on the Si3N4 ridge-slot waveguide is proposed, and the MRR is pumped by orthogonally polarized bicolor pumping to generate dual-comb. We optimized the structure of MRR by the finite element method and precise dispersion engineering, which finally obtained the suitable MRR geometry with negative dispersion characteristics at 1 550 nm, having Q of 1.7×107 and the absorption loss as low as 2.6×10-5 dB/cm. The simulation model of generating dual-comb is established as coupled Lugiato-Lefever equation (LLE), which takes the higher order dispersion, cross-phase modulation (XPM), multiphoton absorption, and external pumping into account. Solved by the split-step Fourier method (SSFM) and the fourth-order Runge-Kutta (RK4) method, the numerical results show that the generated dual-comb is periodically equally spaced distribution, but with slightly different intensities in the time domain. In the frequency domain, there are 64 comb teeth with intensities higher than −100 dBm with a bandwidth of 120 nm. Particularly, in the case of bicolor orthogonal polarization pumping, a smaller amount of detuning does not greatly affect the bandwidth of the dual-comb.

Abstract:Applications for quanta and space sensing both depend on efficient low-light imaging. To precisely optimize and design image sensor pixels for these applications, it is crucial to analyze the mechanisms behind dark current generation, considering factors such as temperature, trap cross-section and trap concentration. The thresholds for these generating effects are computed using optoelectrical technology computer aided design (TCAD) simulations, and the ensuing changes in pinned photo-diode (PPD) dynamic capacitance are observed. Various generation models along with an interfacial trap model are used to compare PPD capacitance fluctuations during light and dark environments. With the use of this comparison study, current compact models of complementary metal oxide semiconductor (CMOS) image sensors can be modified to accurately capture the impacts of dark current in low-light conditions. The model developed through this study demonstrates a deviation of only 6.85% from the behavior observed in physical devices. These results not only enhance our understanding of dark current generation mechanisms but also offer practical applications by improving the performance and accuracy of image sensors.

Abstract:In this paper, the eavesdropping model based on eavesdroppers near legitimate users, and the effect of atmospheric channel correlation on the physical layer security (PLS) of the free-space optical (FSO) link are analyzed. According to the joint probability density function (PDF) and cumulative distribution function (CDF) of Gamma-Gamma (G-G) distribution, a new closed-form expression of interception probability is derived. Numerical results show that the interception probability of the FSO system depends on turbulence intensity, channel correlation and radial displacement attenuation of eavesdroppers.

Abstract:In order to further reduce the cost of manually screening suitable second harmonic signals for curve fitting when detecting methane concentration by tunable diode laser absorption spectroscopy (TDLAS) technology, as well as the influence of certain human factors on the amplitude screening of second harmonic signals, and improve the detection accuracy, a one-dimensional wide atrous convolutional neural network (1D-WACNN) method for methane concentration detection is proposed, and a real-time detection system based on TDLAS technology to acquire signal and Jetson Nano to process signal is built. The results show that the accuracy of this method is 99.96%. Compared with other methods, this method has high accuracy and is suitable for real-time detection of methane concentration.

Abstract:We introduce an all-fiber stationary phase shifter for a fiber-optic gyroscope (FOG) which simultaneously provides phase shifts of opposite signs in different cores of the dual-core optical fiber. We propose a new dual-core fiber-optic gyroscope (DCFOG) in which different cores of the dual-core optical fiber provide independent rotation rate measurements. The device enables implementation of a differential scheme, which ensures the stability of the measured phase shift. As a computer simulation result, the accuracy of the rotation rate sensing is increased by up to 10 times at typical noise levels.

Abstract:The complex refractive index dispersion (CRID) of absorbing materials is very important in many fields, especially in printing industry and medical research. However, due to their strong absorbing, CRID determination is still a challenge. In this study, without diluting treatment or the thickness information, a method is proposed to calculate the CRID of absorbing materials, based merely on the reflectance and transmittance spectra measurements. The method separates the CRID into absorbing part and transparent part based on Kramers-Kronig relations, and it also uses the common Cauchy dispersion formula and Fresnel reflection formula. The CRID of methyl-red-doped poly (methyl methacrylate) (MR-PMMA) (3% mass fraction) and hemoglobin (Hb) solutions (320 g/L) are determined over the spectral range from 400 nm to 750 nm, and the result shows good stability and consistency of the method.

Abstract:To address the challenges of inefficient manual inspections and time-consuming video monitoring for power transmission lines, this paper presents an innovative solution. It combines deep learning algorithms with visible light remote sensing images to detect defects and hazards. Deep learning offers enhanced robustness, significantly improving efficiency and accuracy. The study utilizes you only look once version 7 (YOLOv7) as a foundational framework, enhancing it with the Transformer algorithm, Triplet Attention mechanism, and smooth intersection over union (SIoU) loss function. Experimental results show a remarkable 92.3% accuracy and an 18.4 ms inference speed. This approach promises to revolutionize power transmission line maintenance, offering real-time, high-precision defect and hazard identification.

Abstract:Pulse-coupled neural network (PCNN) is a multi-parameter artificial neural network, and the characteristics of PCNN can be fully explored by analyzing different simplified networks. In this paper, the firing characteristics of non-coupled PCNN with coupled linking term are studied, the mathematical expressions of firing time and interval are summarized, and further the influence of linking weight matrix and linking weight coefficient on network characteristics is analyzed, and the constraints of parameters are given. Finally, extensive verification experiments are carried out for the phenomenon of image edge detection that occurs in the experiments, which provides theoretical support for further research and application of PCNN model.

Abstract:In order to effectively prevent and treat heart-based diseases, the study of precise segmentation of heart parts is particularly important. The heart is divided into four parts:the left and right ventricles and the left and right atria, and the left main trunk is more important, thus the left ventricular muscle (LV-MYO), which is located in the middle part of the heart, has become the object of many researches. Deep learning medical image segmentation methods become the main means of image analysis and processing at present, but the deep learning methods based on traditional convolutional neural network (CNN) are not suitable for segmenting organs with few labels and few samples like the heart, while the meta-learning methods are able to solve the above problems and achieve better results in the direction of heart segmentation. Since the LV-MYO is wrapped in the left ventricular blood pool (LV-BP), this paper proposes a new model for heart segmentation:principle component analysis network (PCA-Net). Specifically, we redesign the coding structure of Q-Net and make improvements in threshold extraction. Experimental results confirm that PCA-Net effectively improves the accuracy of segmenting LV-MYO and LV-BP sites on the CMR dataset, and is validated on another publicly available dataset, ABD, where the results outperform other state-of-the-art (SOTA) methods.