To reduce the complexity and commercial cost of radio over fiber (RoF) systems, this study proposes a modified uni-traveling carrier (MUTC) photodiode with a double Gaussian-doped absorption layer. Operating at 1 310 nm wavelength, the photodiode achieves high-speed and high-power output through precise electric field regulation, enabling direct antenna driving for signal transmission. Under a bias of -8 V, a 20 m diameter photodiode demonstrates a 3 dB bandwidth of 30.9 GHz, a responsivity of 0.82 A/W, and peak radio frequency (RF) output powers of 34.8 dBm at 15 GHz and 34.4 dBm at 20 GHz.

To address the performance degradation caused by the four-wave mixing (FWM) effect in dense wavelength division multiplexing (DWDM) systems, a method for suppressing the FWM effect is proposed. This method begins by comparing and analyzing three dispersion compensation schemes in optical networks with different data rates, identifying the separated dispersion compensation scheme as the optimal approach. The impact of the FWM effect is then evaluated in both 8-channel and 16-channel communication systems by separating dispersion compensation and gain. After multiple optimization steps, pre-chirping is applied to further suppress the FWM effect. The system performance is analyzed through numerical calculations of the channel bit error rate (BER). The results demonstrate that in both 8-channel and 16-channel systems, the FWM effect in the optical fiber is effectively suppressed as the number of dispersion and gain separations increases. With pre-chirping, the FWM effect is further reduced, achieving minimum channel BERs of 5.08×10-19 and 8.77×10-17, respectively. This study provides valuable insights and guidance for the development of next-generation high-capacity optical fiber communication systems.

Quantum cascade lasers (QCLs) have important research and application value due to their ability to directly output lasers from mid-infrared to terahertz ranges. In this paper, the design and fabrication of QCLs are studied, with a focus on the etching method of ridge waveguide. For the waveguide composed of InP cladding layer and InGaAs/InAlAs active layer, a method combining dry etching, selective wet etching and non-selective wet etching is proposed to obtain ridge structure with little side erosion as well as vertical and smooth sidewall. Based on this etching method, a QCL emitting at 8.4 m is prepared. Under pulsed operation mode at room temperature, a peak power of 10 W is achieved, with a maximum wall plug efficiency of 21.1%, which is at the leading level among long-wave infrared QCLs reported at home and abroad. An average power of 117.6 mW is obtained by increasing the duty cycle. This study has guiding significance for improving the design and fabrication of QCLs and optimizing the output performance of laser.

A wavelength-switchable thulium-doped fiber laser (TDFL) based on a double fiber-tapered filter was proposed. Firstly, a Mach-Zehnder interferometer(MZI) filter was fabricated by cascading two fiber tapers produced by tapering single-mode fiber. Subsequently, a TDFL with a pump wavelength of 793 nm and a gain medium length of 3.5 m was designed and constructed, exhibiting a laser threshold of 126 mW. In the experiment, by adjusting the polarization controller (PC), a single-wavelength laser switching output within the range of 1 830.6—1 881.7 nm was achieved, with a minimum wavelength interval of 7.1 nm. Additionally, a dual-wavelength switchable laser output within the range of 1 841.7—1 881.1 nm was realized, with minimum and maximum wavelength intervals of 4.5 nm and 35.7 nm, respectively. During the 20 min monitoring period, the laser demonstrated stable single-and dual-wavelength outputs, with power fluctuations of less than 0.52 dB and 0.91 dB, respectively. The proposed TDFL exhibits flexible wavelength switching characteristics and holds significant application potential in laser sensing and measurement fields.

In order to further reduce the solution space of non-negative matrix factorization (NMF), a NMF method based on preserving the intrinsic structure of invariant constraint and the piecewise smoothness constraint of the endmember spectrum of the hyperspectral image for hyperspectral unmixing is proposed. Firstly, a projection equation is used to describe the intrinsic structure of the hyperspectral image. Then the graph regularization is introduced to establish the relationship between hyperspectral image and abundance matrix, thereby preserving the intrinsic structure invariant of hyperspectral image. Secondly, the adaptive potential function in Markov random field model is used as a smoothing function to improve the smoothness of the endmembers spectrum. Finally, L1/2 sparse constraint is used to promote the sparsity of the abundance matrix. To validate the performance of the proposed method, an experimental analysis is conducted on two real data sets. The results confirm the superiority of the method.

To address the issues of style discrepancies between datasets and pseudo-label noise during clustering in unsupervised domain adaptation (UDA) for person re-identification (Re-ID), and to better leverage global and local information, this paper proposes a label mutual optimization-based UDA method. The approach first incorporates a feature enhancement module into the feature extractor to obtain more robust representations. A multi-branch network is then employed to separately extract global and local pedestrian features, with cross-consistency scores calculated to iteratively optimize the network and mitigate the impact of pseudo-label noise. Finally, cluster reliability evaluation criteria are applied in the clustering algorithm to further enhance network performance. Extensive experiments on multiple datasets validate the effectiveness of the proposed method.

A novel DeepLabv3++ model is proposed to address the low accuracy in identifying small targets and slow detection speed in fabric defect detection tasks. Firstly, a multi-scale lightweight backbone network is designed to extract features from defects with various shapes and sizes. Secondly, convolutional attention modules and channel spatial attention modules are introduced to capture boundary information of small targets and focus on defect regions. Additionally, two types of multi-level feature fusion (MFF) modules are added to mitigate the issue of detail information loss in decoder. Finally, the model is trained and evaluated using a fabric defect dataset collected from an industrial site. The results show that our DeepLabv3++ model outperforms other models, utilizing only 4.1 million parameters. It achieves a mean intersection over union (mIoU) of 90.01% and a mean pixel accuracy (mPA) of 95.05%, meeting the industrial site requirements for balancing detection precision and processing speed.

Acoustic emission(AE) used in non-destructive testing primarily involves analyzing elastic waves propagating within mechanical structures. The integrity of the structure can be assessed by using the information carried by the generated acoustic waves. To achieve precise localization of instrument damage, this paper proposes a method using an orthogonal fiber Bragg grating (FBG) sensor network with a symmetrical triangular arrangement for acoustic emission localization. This approach effectively reduces errors caused by the angle between the FBG axis and the direction of sound wave propagation. Based on wavelet transform, the relative time difference between the arrival of different frequency components in the Lamb wave generated by the acoustic emission source is used to analyze the acoustic emission signal. An established ellipse algorithm is then employed to locate the acoustic emission source. Experimental results demonstrate an average localization error of 2.95 cm. This study also investigates the impact of the distance between the acoustic emission source and the sensor on the received signal. This research holds significant implications for the localization of damage and defects in instruments.

High-precision extraction of fringe frequency is crucial in interferometric particle imaging (IPI) technology. In this paper, a method based on simplified joint interpolation (S-JI) algorithm is proposed to extract the fringe frequency with high speed. The algorithm obtains subpixel-accurate fringe frequency by utilizing both the real and imaginary parts of the Fourier spectrum. Without the need for calculating the frequency offset and performing frequency refinement again (frequency shift), it reduces the algorithm complexity and shortens the computation time. Simulation results show that the processing times of S-JI algorithm saves 2/3, and has the super performance in accuracy and stability of frequency estimation, compared with joint interpolation (JI) algorithm. The relative error of the measured particle diameter is less than 0.3% for standard particles of 25 m and 70 m. The results demonstrate that the method presented herein is considerably promising for its application to a high-speed particle field, such as spray, in high accurately measuring particle size.

In order to investigate the temperature and refractive index (RI) sensing performances of optical fiber surface plasmon resonance (SPR) sensors with different heterogeneous core structures, reflective optical fiber SPR sensors based on multimode fiber (MMF)-single mode fiber (SMF), MMF-no core fiber (NCF), and MMF-hollow core fiber (HCF) structures were prepared respectively. By coating a temperature-sensitive film on approximately half of the surface of the optical fiber probe, the simultaneous measurements of the refractive index and temperature can be achieved. And the temperature and refractive index sensing characteristics of the three probes were compared and analyzed. In this paper, the finite element method was used to analyze the dual-parameter sensing principle of temperature and refractive index, and the sensing characteristics of the three structures of optical fiber probes were analyzed and compared by experimental methods. The experimental results show that among the three heterogeneous optical fiber probes, the HCF-based reflective SPR sensor has the best temperature and refractive index sensing performance. Within the temperature measurement range of 30—70 ℃, the temperature sensitivity is approximately -1.6045 nm/℃, while in the refractive index range of 1.331—1.37, its refractive index sensitivity reaches up to 1 945.46 nm/RIU.

A two-dimensional principal component analysis (2DPCA) algorithm based on the L1-norm and F-norm is proposed for detecting concrete cracks in shield tunnel segments. Given the significance of addressing outlier interference in practical engineering, the L1-norm metric is adopted to reduce the sensitivity of the feature extraction algorithm to outliers. Simultaneously, the F-norm metric is used to minimize the reconstruction error, thereby enhancing the reconstruction performance and improving the accuracy of crack labeling. Tests on concrete crack images demonstrate that the proposed algorithm achieves excellent recognition and labeling performance, with a maximum recognition rate of 90.42%. Furthermore, experiments under various conditions confirm the algorithm′s strong noise resistance. Finally, the algorithm is applied to the field of face recognition, and the experimental results further validate its robustness and practical applicability. In conclusion, the strategy of employing 2DPCA algorithms shows promising applicability for concrete crack detection, and future research can focus on further refining and enhancing these algorithms.

To address the problem of low precision of the traditional inter-symbol dimming control methods in visible light communication (VLC), an efficient adaptively biased layered optical orthogonal frequency division multiplexing (ABLO-OFDM) method compatible with high-precision dimming is proposed in the paper. By introducing a hybrid pulse width modulation (PWM) signal based on both inter-symbol and intra-symbol intervals, dimming level with high precision is achieved in the proposed method. Furthermore, the signal transmission and signal detection at the receiver are immune to the dimming control, thanks to the well-designed pre-processing operation at the transmitter. Therefore, a standard OFDM receiver can be directly used for signal detection in the proposed scheme, which significantly reduces the receiver complexity and processing delay. Simulation results demonstrate that the proposed scheme achieves higher dimming precision as well as better transmission performance compared to the traditional ones.

A frequency octupling radio over fiber (ROF) system based on a dual-parallel Mach-Zehnder modulator (DPMZM) capable of overcoming the bit walk-off effect and realizing carrier reuse is proposed. A composite radio-frequency (RF) signal is used as the driving signal of the DPMZM at the central station (CS) to realize that the data is modulated on the +4th-order sideband only, and a pilot signal is inserted to realize the carrier reuse at the base station (BS). The operating principle of the designed system is theoretically analyzed and verified by simulation experiments. For the Q value is greater than 6, the transmission distance of fiber of the uplink and downlink of the system exceed 290 km and 80 km, respectively. The effect on the Q value when the main device parameters of the system deviate from the design value is analyzed. The innovation of this paper lies in the fact that the system not only overcomes the bit walk-off effect, but also solves the problems of degradation of the performance of the downlink and the poor system tunability caused by the conventional carrier reuse, and increases the transmission distances of fiber and the performance of the system considerably, and has important application prospect in ROF.

Liver cancer is a kind of malignant tumor, early screening and accurate detection is the key to improve the treatment effect and prolong the survival of patients. In view of the difficulty of accurately detecting complex and variable liver cancer using single-phase computed tomography (CT) images, in this paper, a dual-phase CT method for liver cancer detection based on fully convolutional one-stage object detection (FCOS) is proposed. Firstly, the dual-phase liver CT quadtuple network is constructed and used to match the dual-phase liver CT slices to ensure the consistency of liver position between different phases and lay a foundation for the subsequent detection of liver cancer. Secondly, FCOS network is improved to receive input of dual-phase CT images, attention-based feature fusion (AFF) module is designed and inserted, and feature fusion with mixed attention is performed at the same time to improve the accuracy of liver cancer detection. The experimental results show that the average precision (AP) of the improved algorithm on the data set in this paper reaches 78.56%, which is 4.9% higher than that of the single-phase FCOS network, showing better performance.