To address the challenges of local trajectory planning in robotic manufacturing of curved surface components, this article introduces a method grounded in multi-view stereo reconstruction and color regional growth. Initially, we apply color texture to the surface of a curved component. Subsequently, the improved PatchmatchNet model, which incorporates a high-resolution net and total variation loss, reconstructs high-quality 3D point clouds, achieving 10.42% improvement in accuracy and 10.59% improvement in completeness compared to the baseline. Then, a regional growth point cloud segmentation algorithm based on the hue-saturation-intensity color space and local color distribution similarity is used to segment the color-texture point cloud across three aerospace components with 82.2% Cat.mIoU and 77.1% Ins.mIoU. Following this, point cloud processing and curve fitting are executed on the single-color-texture point cloud, culminating in the derivation of the local processing trajectory feature line. The nonuniform rational B-splines (NURBS) curve fitting generates machining trajectories with R2 exceeding 0.94% and 30% lower root mean square error (RMSE) than polynomial and standard B-spline methods. Real machine grinding experiments validate the method's feasibility, demonstrating precise localization and effective material removal along planned trajectories. The proposed approach addresses critical challenges in local grinding of complex curved surfaces, offering significant practical value for aerospace manufacturing.

Gradient microstructural design offers a promising strategy to overcome the strength-ductility trade-off in metals with low stacking fault energy (SFE). However, its effectiveness in aluminum alloys is significantly constrained by two key factors: strain-induced precipitate dissolution and an insufficiency of deformation carriers (e.g., stacking faults or twins) due to inherently high SFE. Here, we demonstrate that ultrahigh strain rate deformation generates a gradient microstructure in a high-strength aluminum alloy. Crucially, this microstructure subsequently triggers a spontaneous nanoscale secondary precipitation process during ambient storage. This sequential process produces a synergistic effect, enhancing both the strength and ductility of the material beyond levels achievable via conventional heat treatments alone. The resulting gradient architecture, featuring a hard exterior and a soft core, contributes to improved strength while preserving superior resistance to mechanical damage. Concurrently, the nanoscale secondary precipitates effectively mitigate the detrimental precipitate-free zones typically formed during thermal aging, thereby suppressing localized plastic deformation and associated damage. These findings provide a viable pathway for optimizing the strength-toughness synergy in precipitation-strengthened alloys.

The development of additive manufacturing (AM) technique has enabled unprecedented flexibility in the design and fabrication of complex metal components. However, the AM process is prone to defects like poor surface quality, high porosity, and tensile residual stress (TRS), which adversely affect the mechanical performance and service life of the fabricated components. Laser shock peening (LSP) is an innovative surface plastic strengthening technique that can effectively introduce compressive residual stresses (CRS) and refine microstructure, thereby mitigating these shortcomings. This paper first presents the principle and development of LSP and its research advancements in AM metals. It then systematically elucidates the effects of LSP in AM metals from two perspectives: the LSP post-treatment process and the AM + LSP hybrid process. Additionally, some investigations of LSP applications on specialized AM structures are introduced. Through a comprehensive review of research progress and technical challenges, research prospects are provided on theory, equipment, processes, performance evaluation, and technical standards to clarify the theoretical support and key technologies for future engineering applications.

Extracting valuable features from a signal's Region of Interest (ROI) is critical for advanced signal analysis. However, existing methods often require domain-specific design assumptions and fixed filter structures, limiting their adaptability and generalization. To address these limitations, this paper proposes the Frequency Band Recalibration Spectrogram (FBRS), a novel time-frequency analysis method based on the wavelet packet energy distribution. FBRS adaptively optimizes filter bandwidth and dynamically generates filter distributions based solely on a signal's intrinsic frequency characteristics. By iteratively reconstructing wavelet packet components and refining frequency resolution, FBRS adaptively constructs a signal-driven time-frequency representation by dynamically recalibrating frequency bands based on intrinsic energy distribution. While this results in improved resolution in important regions, the core contribution lies in its adaptive and data-driven spectrogram generation mechanism. To verify the effectiveness of the proposed method, extensive experiments were conducted on simulated signals and three real signals from multiple practical application scenarios. The experimental results demonstrate that FBRS is theoretically innovative and exhibits strong adaptability and descriptive resolution in various signal analysis scenarios. The adaptability and flexibility of the proposed method enable it to cross different domains, providing a new perspective and a powerful tool for signal analysis. The code of the proposed method has been made public in https://github.com/Qinr1026/FrequencyBand-Recalibration-Spectrogram.

To address the challenge that a single surface strengthening technique is unable to simultaneously tackle the issues of crack initiation and propagation in the fretting fatigue failure of titanium alloys，a composite treatment involving initial laser shock peening followed by nitrogen ion implantation at 300 ℃ is employed to enhance the properties of TC6 titanium alloy. The residual stress distributions of four types of specimens，namely the untreated ones，those subjected to nitrogen ion implantation at 300 ℃，those treated with laser shock peening，and those receiving the composite treatment，are measured using an X-ray diffraction stress analyzer. The fretting fatigue life of the titanium alloy specimens is assessed on a self-designed surface-contact fretting fatigue testing rig. The fracture surfaces and wear scars at the crack initiation zones are characterized using scanning electron microscopy (SEM) and energy-dispersive X-ray spectroscopy (EDS). The results demonstrate that the composite treatment creates both a high-hardness nitrided layer and a deep gradient residual compressive stress field，with the residual compressive stress extending to a depth of approximately 1.4 mm. The average fretting fatigue life of the specimens subjected to the composite treatment reaches 2.98×105 cycles，which is 161.4%，108.4%，and 30.1% higher than that of the untreated specimens，those implanted with nitrogen ions at 300 ℃，and those treated with laser shock peening，respectively. Fracture analysis reveals that the composite strengthening transforms the damage mechanism in the fretting contact area from severe adhesive wear to predominantly abrasive wear，significantly postponing crack initiation. Meanwhile，the deep residual compressive stress effectively reduces the crack propagation rate.

Nickel-based single-crystal superalloys are widely used in aviation and industrial gas turbine blades owing to their excellent mechanical properties at high temperatures. However, under extreme service conditions, these superalloys are susceptible to fatigue, wear, corrosion, and oxidation damage, which pose significant risks to engine safety. Laser shock peening (LSP) is an advanced surface-strengthening technology that utilizes laser shock waves to induce severe plastic deformation at the surface, alter the microstructure, and introduce compressive residual stress (CRS), thereby enhancing the fatigue resistance and other properties of the alloy. Some studies have investigated the laser shock strengthening of nickel-based single-crystal superalloys for turbine blades. However, there is a notable lack of a systematic summary on this topic. This study begins by examining the technical characteristics of various LSP technologies, including traditional high-energy laser shock, low-energy laser shock without an absorption layer, warm LSP (WLSP), and femtosecond LSP (Fs-LSP). The effects of these technologies on the microstructures and properties of nickel-based single-crystal superalloys are compared, highlighting their respective advantages and disadvantages. This summary revealed that the strengthening effects of various LSP technologies on single-crystal superalloys differ. Ns-LSP and WLSP can achieve millimeter-level strengthening layer depths; LSPwC reaches several hundred micrometers; and Fs-LSP results in depths of tens of micrometers. The strengthening mechanisms of the four LSP processes for nickel-based single-crystal superalloys are distinct. Ns-LSP primarily strengthens by introducing high-density crystal defects and CRS during the shock process. WLSP improves the density, uniformity, and stability of crystal defects. LSPwC not only introduces crystal defects and CRS but also generates in-situ nano-oxide particles that further impede the dislocation motion, enhancing the strengthening effect. Fs-LSP introduces both crystal defects and CRS, and simultaneously builds surface periodic micro-nanostructures, achieving a synergistic optimization of strengthening effect and functional structure. Each of these processes has unique features and provides diverse technical paths for the performance enhancement of single-crystal alloys. Despite the different depths of strengthening layers, the core mechanism of these techniques lies in the large number of crystal defects and CRS introduced by LSP. The large plastic deformation induced by Ns-LSP and WLSP may lead to recrystallization of the single-crystal superalloy during the strengthening / service process, thereby damaging its high-temperature creep performance. Therefore, further research is needed to explore how LSP technology can achieve efficient and high-quality strengthening of single-crystal superalloys, and obtain ideal single-crystal structures. When LSPwC is used to strengthen single-crystal superalloys, the laser directly acts on the alloy surface, leading to surface remelting and oxidation, which in turn reduces the surface quality. Therefore, further research is required to improve the surface quality of the LSPwC-treated single-crystal superalloys and enhance their strengthening effect. The paper also discusses the challenges and difficulties faced in the LSP of nickel-based single-crystal superalloys and offers insights into future development trends. LSP technology has already been commercially applied in other alloy fields, and research has shown that it can improve issues such as fatigue, wear, oxidation, and corrosion in nickel-based single-crystal superalloys. To further advance this technology and expand its market penetration, it should be developed in several directions: composite strengthening process innovation, upgrading of residual stress detection equipment, innovations in equipment manufacturing technology, and the establishment of intelligent processing systems.

In this study, single point and overlapping laser impact tests were performed on Ti-6Al-4 V alloy plates, with systematic variation of plate thickness, laser power density, and pulse width parameters. Following the impact tests, the surface morphology of each specimen was characterized using a laser scanning confocal microscope. The primary objective was to investigate the influence of process parameters on the deformation behavior of the specimens and to elucidate the intrinsic relationship between single point and overlapping impact deformation phenomena. The experimental results revealed that, under single point impact conditions, the central deformation exhibited a negative correlation with plate thickness, while demonstrating positive correlations with both laser power density and pulse width. Conversely, under overlapping impact conditions, the terminal displacement of the specimen displayed a positive correlation with plate thickness, but negative correlations with laser power density and pulse width. Furthermore, the experimental datasets for central displacement under single point impact and terminal displacement of elongated narrow plates under overlapping impact were subjected to regression analysis, resulting in the development of corresponding predictive mathematical models.

To address the bottleneck of insufficient accuracy in ultrasonic probe pose tracking during non-destructive testing (NDT) of composite materials—a limitation that impedes 3D imaging—we propose a 3D ultrasonic imaging method for composites based on fusing visible light images and depth maps. A dual-stream network processes the two modalities in parallel to accurately recover the 3D coordinates of key points on the probe. We design an Edge Enhancement Module (EEM) that employs multi-directional learnable convolutions to mitigate contour blurring in depth maps; a Cross-Attention Feature Fusion Module (CA-FFM) that uses bidirectional attention to align cross-modal semantic features and improve detection performance; and a Multi-Branch Re-param Convolution (MBRConv) block that leverages branches with different kernel sizes to capture multi-scale features, thereby improving key point localisation accuracy, while re-param inference reduces computational cost. Finally, we establish a probe coordinate system from three detected key points and, combined with ultrasonic B-scan processing and spatial mapping, achieve 3D reconstruction. Experiments show that key point detection attains an mAP@0.5:0.95 of 86.3%, substantially outperforming conventional methods, with a throughput of 64.5 FPS that meets real-time requirements. On both curved composite components and composite panel, the 3D imaging error is at the sub-millimetre level, enabling high-precision visualisation of internal composite structures. This work provides a low-cost, portable 3D imaging solution for in-service aerospace inspection.

Laser shock peening (LSP) has been shown to promote a transition in the wear mechanisms of nickel-based superalloys during elevated-temperature fretting wear. However, the intrinsic relationship between LSP-induced microstructural features and the resulting wear mechanisms has not been fully elucidated, particularly for the adoption of in situ techniques. In this study, the mechanism underlying this transition is clarified in detail from a microstructural perspective through quasi-in situ experimental efforts and state-of-the-art characterization techniques. Fretting wear test results demonstrate that LSP can significantly shift the wear mechanism of GH4169 superalloy at 600 degrees C, transitioning from adhesive wear to abrasive wear. Further examination of the cross-sectional microstructure of the worn subsurface reveals that the LSP-treated sample developed a compound gradient structure consisting of an amorphous-crystalline oxide layer and a nanocrystalline grain structure on the surface. In contrast, this structure is absent in the untreated sample. The in situ formation of this compound gradient structure, coupled with the plastically deformed gradient nanostructure beneath it, results in the LSP-treated sample predominantly exhibiting an abrasive wear mechanism during fretting wear at 600 degrees C. This contrasts with the adhesive wear mechanism observed in the untreated sample. This work provides valuable insights into the fundamental understanding of plastic deformation in LSP-treated superalloys during fretting wear at elevated temperatures and offers guidance for the design of wear-resistant alloys via surface engineering.

This study investigates the fretting fatigue crack initiation in dovetail joints using crystal plasticity modeling and experimental validation. The experimental data indicate that crack initiation occurs near the center of the contact surface. Crystal plasticity finite element simulations reveal that regions with elevated cumulative crystal slip parameter and Fatemi-Socie parameter are concentrated in the central area of the contact surface, displaying a non-uniform distribution. Both the cumulative crystal slip parameter and Fatemi-Socie parameter demonstrate relatively high accuracy in predicting crack initiation locations. This is because the cumulative crystal slip parameter represents accumulated plastic damage at the microscopic scale, while the Fatemi-Socie parameter incorporates the effects of normal stress and shear strain on the slip plane. This study introduces a method for predicting crack initiation positions in dovetail joints subjected to fretting fatigue using the crystal plasticity finite element approach, providing valuable insights for the structural design and strength assessment of dovetail-shaped components.

In view of the bottleneck problems existing in the 3D ultrasonic testing of aircraft composite laminated structures—including heavy reliance on manual operation, resulting in low detection efficiency, and the inability of traditional robotic arms to adapt to the testing of complex curved surfaces due to their dependence on predefined fixed trajectories—this paper proposes an automated 3D ultrasonic testing method based on 3D vision guidance for robotic arms. Firstly, the proposed Yolo-Mask model is adopted to realize the visual recognition and segmentation of composite component regions, after which the segmentation results are mapped to the depth map and further converted into the surface point cloud of the material. Secondly, on the basis of point cloud preprocessing and trajectory point extraction, the automatic planning of the robotic arm’s scanning trajectory is achieved, which drives the robotic arm to perform precise motion and to synchronously collect spatial pose and ultrasonic testing data. Finally, 3D reconstruction is completed via a fusion algorithm, and 3D images of the material’s internal structures are generated. Experimental verification shows that the proposed method achieves a Segm-mAP of 97.4%, a detection speed of 11.7 fps, and a 3D imaging error of less than 0.1 mm, thereby realizing fully automated detection throughout the entire process. This research provides an effective solution for the non-destructive testing of aircraft composite structures.

The bending fatigue resistance of superelastic shape memory alloys (SMAs) is a key determinant for their reliable function in cyclic applications such as biomedical implants, adaptive actuators, and elastocaloric devices. However, conventional NiTi alloys exhibit limited fatigue life due to premature crack initiation and propagation under cyclic tensile loading. Here, we report a surface engineering strategy that overcomes this limitation by inducing a hierarchical surface architecture via pre-strain warm laser shock peening (pw-LSP). This architecture integrates a high-strength titanium nitride-enriched top layer, an ultrafine-grained layer with an inverse grain size gradient and a B19'-R-B2 phase gradient, and a substantial compressive residual stress exceeding 1 GPa. These features act synergistically to suppress crack nucleation and arrest propagation through a crack-tip shielding mechanism. As a result, the treated NiTi demonstrates a bending fatigue life exceeding 5 million cycles at a maximum surface tensile strain of 1.94%-representing a more than 3000-fold enhancement over untreated nanocrystalline NiTi. This work presents a robust and scalable approach for designing fatigue-resistant SMAs with broad implications for high-cycle, high-reliability applications.

ZrAlSiN/Zr multilayer coatings with different modulation ratios were designed and prepared using magnetic filtered cathode vacuum arc technology via alternating deposition of ZrAlSiN hard layers and Zr ductile layers. The microstructure and mechanical properties of ZrAlSiN/Zr coatings were systematically investigated. By combining three-point bending ultrasonic fatigue testing with in-situ scanning electron microscopy bending experiments, the fatigue performance of the coating/substrate (Ti6Al4V) system with different modulation ratios was systematically assessed. The in-situ dynamic observation captured the crack initiation, propagation, and interaction with the substrate under cyclic stress and large deformation. The results showed that the modulation ratio plays an important role in the microstructure and mechanical properties of ZrAlSiN/Zr coatings. As the modulation ratio increased, the hardness and elastic modulus of the coatings first increased and then decreased. The adhesion strength of the coatings showed a decreasing trend with increasing modulation ratio. The fatigue failure originates from the competition and synergy between 'coating crack propagation downward' and 'substrate slip-induced tearing upward' at the interface. For coatings with a small modulation ratio (1:13), cracks readily propagate through the metal layer, leading to rapid coating fracture. In contrast, a large modulation ratio (1:2.4) promotes crack deflection or confinement within the hard layer by the metal layer, thereby delaying fracture. The ZrAlSiN/Zr coating with a modulation ratio of 1:6 exhibited the best fatigue performance due to the optimal balance between crack initiation resistance and crack propagation resistance.

The evolution of corrosion products and their influence on the fretting wear behavior of GH4169 alloy in a CO2 with 0.5 % SO2 at 800 degrees C was investigated. The corrosion behavior increased the sensitivity of GH4169 alloy to the displacement amplitude of fretting wear. At low displacement amplitudes, a dense and high-hardness Cr-rich oxide film effectively isolated the contact interface and enhanced wear resistance. At high displacement amplitudes, the Cr-rich oxide film ruptured and exposed the underlying Cr-depletion zone. Discontinuous Al2O3 internal oxide particles induced microcracks at stress concentration sites. Sulfidation behavior induced interfacial failure and activated third-body wear, which continuously aggravated material degradation.

Automated Guided Vehicle (AGV) are widely used in the aviation industry. To address their insufficient visual relocalization accuracy, weak robustness, and high data dependence, this paper proposes a high-precision visual relocalization framework integrating laser-optimized 3D Gaussian Splatting (3DGS) and image retrieval. Firstly, a PEM (Prototype-based Efficient MaskFormer)+SegRefiner combined segmentation model is constructed, which optimizes PEM-generated coarse masks via discrete diffusion "denoising-correction" iteration for image segmentation (mIoU = 0.972, mAP = 0.984), providing high-quality data for 3D reconstruction. Secondly, the 3DGS model is improved by registering and fusing laser point clouds with Structure from Motion (SfM) sparse point clouds, and introducing the 3 sigma principle to constrain Gaussian distribution, solving the poor geometric consistency of traditional 3DGS (unknown view rendering PSNR = 28.6751, superior to original 3DGS and Mip-splatting). Finally, a two-stage pose estimation method is designed: coarse localization narrows the pose range via low-resolution rendered image matching; fine localization optimizes the search with a Differential Evolution-Simplex hybrid algorithm. Experiments show that the pose estimation accuracy of the proposed method is superior to that of methods such as PixLoc and NeRF-Loc. This framework forms a "segmentation-modeling-localization" technical closed-loop, which not only meets the operational accuracy requirements of AGV in the aerospace field but also expands the application scenarios of 3DGS in the field of industrial automation.

In aero-engine fastening applications, TC6 titanium alloy components are prone to fretting-induced surface damage under service conditions. Although surface treatments like shot peening have been explored to address this problem, they face challenges such as large deformation and compromised surface integrity. Here, we report a Fine Particle Peening (FPP) technique that overcomes these limitations, while significantly enhancing fretting wear resistance. This study reveals that the gradient structure introduced by FPP, characterized by near-surface equiaxed nanocrystals formed under adiabatic heating and subsurface mechanical twins generated to accommodate plastic strain, leads to a pronounced reduction in fretting wear. Specifically, the wear rate decreases by 42.00%, from 2.50 × 10−5 mm3/(N·m) to 1.45 × 10−5 mm3/(N·m). Moreover, multi-scale microstructural characterization combined with energy-dissipation analysis was employed to elucidate the intricate microstructural evolution and deformation mechanisms of TC6 alloys induced by the FPP process and subsequent tribological interactions. The excellent fretting wear performance is attributed to the formation of gradient nanocrystalline structures, gradient work-hardening layers, and stress-induced twinning, which collectively prevent matrix exfoliation caused by strain localization and facilitate stable frictional layer formation. This work provides a systematic understanding of the strengthening mechanisms associated with FPP-induced gradient structures and demonstrates how they enhance wear resistance, thereby offering guidance for the microstructural design of titanium alloys.

Fretting wear frequently occurs between turbine blade tenons and disk dovetails fabricated from TiAl alloys, leading to fatigue failure of structural components. To investigate the fretting wear damage progression and reinforcement mechanisms of TiAl alloy exposed to laser shock peening with coating (LSPwC), multiple LSPwC treatments were applied to the alloy, and cyclic interrupted fretting wear tests were performed. The findings indicate that LSPwC treatment induces the creation of a coupled gradient structure composed of an ablation-remelted layer and a hardened layer in TiAl alloys, thereby changing the wear behavior of the alloy across various fretting cycles. Firstly, the ablation-remelted layer provided lubrication and protection within the first 200 cycles, which reduced the wear volume. Subsequently, at 10,000 cycles, the ablation-remelted layer fractured rapidly due to structural defects, causing the wear volume of the specimen at this stage to increase by 22.8% compared to the original. Finally, after the remelted layer fractures, the hardened layer takes a dominant role. This layer contains numerous stacking faults and a few twin structures, which facilitated the development of a stable debris layer, transitioning the primary wear mechanism from fatigue to abrasive wear. As a result, the wear volume of the strengthened specimens decreased from 25,000 cycles onward; by the time it reached 100,000 cycles, the wear volume of the strengthened specimen is reduced by 24.4%.

This study investigates the growth behavior of thermally grown oxides (TGO) and secondary reaction zones (SRZ)—the primary interfacial degradation modes in thermal barrier coating (TBC)/single-crystal superalloy (SX) systems—during high-temperature oxidation at 900 ℃ and 1100 ℃, together with their distinct feedback effects on elemental diffusion and the evolution of residual stress in in compositionally distinct TGOs. The results indicate that, at 900 ℃, the pinning effect of high-density TCP phases retards elemental depletion in the bond coat (BC), thereby promoting the formation of Cr- and Al-enriched bilayer TGO, in which localized Cr₂₃C₆ precipitates act as stress concentrators. At 1100  ℃, enhanced outward diffusion of Al and Re within the SRZ renders the γ phase—serving as the primary diffusion pathway—dominant, thereby intensifying interdiffusion and ultimately leading to the formation of a NiAl2O4/Al2O3 dual-layer TGO. Furthermore, the nucleation and coalescence of interfacial microvoids at the Al₂O₃/BC interface constitute the primary stress-relief mechanism, resulting in TGO delamination, whereas NiAl₂O₄ act as a residual stress source that triggers cracking at the TGO/top coat (TC) interface.

Automated pick-and-place of aero-engine blades faces three key challenges: limited training samples, large scale variations among detection features, and high requirements for angle detection accuracy in robotic arm pose calculation. This study proposes a vision-guided method integrating data augmentation based on the Stable Diffusion (SD) model and improved oriented object detection. The main contributions are threefold: two in artificial intelligence methodology and one in engineering application. Firstly, a data augmentation method combining Low-Rank Adaptation (LoRA) fine-tuning with SD model is proposed to address the small sample problem, achieving a Frechet Inception Distance (FID) of 10.50 with training time under 1 h. Secondly, an improved You Only Look Once version 8 nano Oriented Bounding Boxes (YOLOv8n OBB) model is developed with three novel modules: a deformable large kernel attention module in the backbone for adaptive shape feature extraction, a context-aware multi-scale feature fusion module in the neck for handling scale variations, and a lightweight asymmetric detection head for enhancing angle prediction accuracy. Thirdly, for engineering application, a 2.5 Dimensional (2.5D) vision-guided robotic arm pose calculation method is proposed by combining oriented detection results with depth information to compute the end-effector position and euler angles. Experimental results demonstrate that the improved detection model achieves a 4.7% improvement in mean Average Precision (mAP)@50%-95% compared to the baseline YOLOv8n OBB, reaching 91.3%. Comparative experiments on both self-built and public datasets confirm that the proposed model outperforms state-ofthe-art methods. Engineering experiments validate the effectiveness of the proposed method for blade pick-and-place tasks.

Detecting microcracks in aero-engine blades is paramount for ensuring flight safety, yet conventional inspection methods exhibit significant limitations in precision, efficiency, and applicability. This paper proposes a multi-modal fusion detection network for blade microcracks—VUFNet (visible and ultraviolet multi-modal fusion detection network)—which accurately identifies blade microcracks by integrating visible and ultraviolet image features. First, the blade surface undergoes chromic acid anodisation to enhance microcrack features. Subsequently, visible and ultraviolet light illumination is applied to construct a microcrack dataset. VUFNet employs dual backbone networks to extract visible and ultraviolet features. It incorporates CSPSTR (cross-stage partial bottleneck with swin transformer) to optimise feature extraction capability, and VUFM (visible–ultraviolet multi-modal adaptive fusion module) to deeply fuse visible and ultraviolet features. Finally, MFConv (multi-branch fusion convolution) enhances detection accuracy in microcrack target regions. On the held-out test subset acquired from a single compressor-blade type under a fixed laboratory imaging protocol, VUFNet achieved mAP0.5 and mAP0.5:0.95 values of 96.8% and 85.5%, respectively, with a detection speed of 52.7 fps. These results demonstrate improved within-domain detection performance over the evaluated baselines. Ablation studies further validate the synergistic optimisation effect of VUFM, CSPSTR, and MFConv on model performance.