Haiting Xia

Phase unwrapping is essential for many phase-centric optical imaging and metrology techniques, yet remains a challenge under high speckle noise. A deep learning-based speckle phase unwrapping algorithm was developed on the basis of a physics-informed convolutional-Fourier neural network, relying on an uncertainty-aware mathematical model of speckle. To address the limitation that traditional spatial-domain CNN methods suffer from inadequacy in extracting frequency-domain interference features during phase unwrapping under speckle noise, we integrated the speckle mathematical model into the network via a parallel convolution-Fourier architecture and employed an integrated loss function with speckle noise, phase gradient, and absolute phase to significantly enhance the unwrapping performance of the network. Experimental results demonstrate that the proposed method achieves superior performance over other phase unwrapping algorithms in both simulations and digital holography experiments. By unifying spatial-domain and frequency-domain feature extraction via a physics-informed deep learning architecture, this method exhibits great potential for extension to broader information processing tasks.

Cementitious composites are heterogeneous porous materials whose pore structure plays a critical role in resistance to chloride-ion penetration. A waterborne nano-silicate-based densifier (CF-S5) was used to examine its influence on the pore structure and resistance to the chloride ion penetration of mortar. We investigated the resistance to the chloride ion penetration of mortar with added CF-S5 admixture through the Rapid Chloride Permeability Test (RCPT). We investigated the pore structure characteristics of mortar by mercury intrusion porosimetry (MIP) coupled with fractal theory and investigated the degree of hydration of the cement paste by thermogravimetric analysis (TG). Ultimately, the degree of correlation between multifractal parameters and the chloride ion migration coefficient of mortar was examined using gray relational analysis (GRA). Results indicate that the CF-S5 admixture reduces mortar porosity and the content of harmful pores while increasing pore tortuosity, thus improving the resistance to the chloride ion penetration of mortar. Multifractal analysis indicated that the CF-S5 admixture decreased the connectivity and increased the complexity of the mortar pore structure. The CF-S5 admixture did not reduce the hydration degree of cement paste at 28 d. Additionally, the multifractal parameters show a high gray relational degree with the chloride migration coefficient; therefore, they may serve as potential indicators to reflect the resistance to the chloride ion penetration of mortar.

The compound lift and thrust Vertical Take-Off and Landing (VTOL) fixed-wing Unmanned Aerial Vehicle (UAV) has generated considerable interest in configuration research due to its unique application advantages. This investigation examines the aerodynamic phenomena between the rotors and the main wings, as well as canards, during the transition phase through numerical simulations, thereby advancing the understanding of canard configurations in such UAVs. Based on a systems engineering approach, a 6 kg canard-configured compound lift and thrust VTOL fixed-wing UAV was preliminarily designed for evaluation. Computational Fluid Dynamics (CFD) methods were employed to study the aerodynamic interference under various freestream velocities and rotor speeds during the transition phase. The reliability of the CFD methodology was validated through rotor thrust experiments. Simulations were conducted with freestream velocities ranging from 3 m/s to 15 m/s and rotor speeds from 4000 to 10,000 RPM. The results indicate that the interference of the rotating rotor during the transition phase initially reduces lift, then increases lift, and finally reduces lift again for the wing, while it increases lift for the canard. This phenomenon results from the coupled influence of freestream velocity and rotor-induced flow effects.

To address the problems of low coverage rate and low detection accuracy in UAV-based aircraft skin defect detection under complex real-world conditions, this paper proposes a method combining a Greedy-based Breadth-First Search Coverage Path Planning (GB-CPP) approach with an improved YOLOv11 architecture (INN-YOLO). GB-CPP generates collision-free, near-optimal flight paths on the 3D aircraft surface using a discrete grid map. INN-YOLO enhances detection capability by reconstructing the neck with the BiFPN (Bidirectional Feature Pyramid Network) for better feature fusion, integrating the SimAM (Simple Attention Mechanism) with convolution for efficient small-target extraction, as well as employing RepVGG within the C3k2 layer to improve feature learning and speed. The model is deployed on a Jetson Nano for real-time edge inference. Results show that GB-CPP achieves 100% surface coverage with a redundancy rate not exceeding 6.74%. INN-YOLO was experimentally validated on three public datasets (10,937 images) and a self-collected dataset (1559 images), achieving mAP@0.5 scores of 42.30%, 84.10%, 56.40%, and 80.30%, representing improvements of 10.70%, 2.50%, 3.20%, and 6.70% over the baseline models, respectively. The proposed GB-CPP and INN-YOLO framework enables efficient, high-precision, and real-time UAV-based aircraft skin defect detection.

Camera channel crosstalk is an inevitable issue in dynamic real-time color digital holography (CDH) detection, significantly influencing the precision of three-dimensional deformation measurements by CDH. We propose a channel crosstalk elimination method for CDH based on a three-wavelength elementary hologram. Different interference angles of R, G, and B lasers are designed, and then the R, G, and B elementary holograms, having different frequencies and having fringes in different directions, are generated and recorded. The intensity components of the R, G, and B holograms are separated and extracted from the Fourier frequency spectra of the color elementary hologram, and then the crosstalk matrix is derived, and crosstalk elimination is achieved. The method has strong resistance to random noise and is not affected by the fringe contrast of the hologram. The effectiveness and practicability of the method were proved by both simulation and experiment.

The propeller state of unmanned aerial vehicles (UAV) is difficult to detect in real time due to trouble with laying out the sensor and multiple signal sources. To solve this problem, a fault detection method for multi-rotor UAV propellers was proposed based on a signal analysis of the built-in inertial measurement unit (IMU). Firstly, the multi-source coupled signals of the UAV flight were obtained through the ground station. Then, the picked-up signals were optimally separated according to the multi-rotor UAV propeller fault dynamics model, and signals rich in fault information were obtained. Finally, the separated signals were calculated using the symmetrized dot pattern (SDP), and then the similarity index was used to quantify the distribution of the signal in the feature plot to realize propeller fault detection. The OTSU algorithm was used to quantify the detection results, yielding a similarity of 76.2% in the z-axis direction, which is better than the values in the other two directions. The simulation and experimental analysis of the propeller failure dynamics model showed that the proposed method can effectively identify the propeller faults of multi-rotor UAVs.

Compound-wing unmanned aerial vehicles (UAVs) are highly valued for their performance. However, during the transition from vertical take-off to the cruise phase, the rotor wake can be coupled with the fixed wing. In this study, the aerodynamic effects of a DJI 9450 rotor on a NACA2415 fixed wing during transition were investigated using the computational fluid dynamics (CFD) method. The rotor-to-wing distances (R/L = 0.25, 0.5, and 0.9) were varied to analyze their impact on aerodynamic performance. The results show that increasing the distance between the front rotor and the fixed wing enhances the lift and drag of the fixed wing, while increasing the distance between the rear rotor and the fixed wing decreases the lift and drag of the fixed wing. During the rotor’s rotation, the fluctuation in the lift and drag of the fixed wing changes periodically due to the rotor wake, and the smaller the distance between the rotor and the fixed wing, the larger the fluctuation. When R/L = 0.25, the fluctuation of the fixed wing is minimized. Compound-wing UAVs with rotors mounted at R/L = 0.25 during the design stage can improve the flight stability during the transition phase in UAVs.

Weld stability is directly related to the safety and reliability of engineering, and continuous improvement of its detection technology has great research significance. This paper presents a novel method for weld stability detection based on color digital holography. A color digital holography optical path was designed to capture the holograms of welds under varying loads. Several common denoising algorithms were used to deal with speckle noise in the wrapped phase, among which the 4-f optical simulation integrated cycle-consistent generative adversarial network (4f-CycleGAN) denoising algorithm based on deep learning was more suitable for the color digital holographic detection system. The three-dimensional deformation fields of three samples (lap-jointed, butt-jointed, and defective butt-jointed aluminum alloy plates) under different loads were calculated. The center profile of the deformation field in the direction of load and holographic reconstruction images are used to determine the position of the weld. The coefficient of variation near the weld was used to evaluate the weld stability. The coefficient of variation for lap-jointed, butt-jointed and defective butt-jointed plates are 0.0335 (<0.36, good stability), 0.1240 (0.36, poor stability), respectively. The research results of this paper provide a new strategy for detecting weld stability.

The speckle noise generated during digital holographic interferometry (DHI) is unavoidable and difficult to eliminate, thus reducing its accuracy. We propose a self-supervised deep-learning speckle denoising method using a cycle-consistent generative adversarial network to mitigate the effect of speckle noise. The proposed method integrates a 4-f optical speckle noise simulation module with a parameter generator. In addition, it uses an unpaired dataset for training to overcome the difficulty in obtaining noise-free images and paired data from experiments. The proposed method was tested on both simulated and experimental data, with results showing a 6.9% performance improvement compared with a conventional method and a 2.6% performance improvement compared with unsupervised deep learning in terms of the peak signal-to-noise ratio. Thus, the proposed method exhibits superior denoising performance and potential for DHI, being particularly suitable for processing large datasets.