Aero-engine blade icing is a critical factor in aerodynamic performance, with the accurate measurement of condensed clear ice being essential. Conventional 3D reconstruction methods based on diffuse reflection are inadequate for highly transparent materials. This study proposes a novel approach for reconstructing the 3D morphology of transparent objects through inverse reconstruction of refractive light paths, enabling the measurement of clear ice formations on aircraft engine blades. First, the correspondence between refracted captured images and predetermined patterns is estimated using binocular cameras to construct the refractive light paths. Second, surface height information is recovered using a combination of binocular normal consistency and local curvature stabilisation constraints. Finally, experimental results on transparent objects with known morphology and comparative evaluations with existing methods confirm that the proposed technique achieves accurate reconstruction for transparent media with thickness greater than 0.5 mm, and yields results closely aligned with those from conventional stereo-based measurements.

Non-contact velocity measurement is critical for characterizing structures in complex flow fields. This paper proposes a hybrid velocimetry technique that integrates schlieren imaging with cross-correlation and optical flow algorithms to enhance both spatial resolution and reconstruction robustness. A Z-shaped schlieren system is employed to acquire sequential flow images. Subsequently, it combines cross-correlation matching to capture large displacements with optical flow estimation to refine small-scale turbulent structures. To validate the method, comparative evaluations were conducted using numerical simulations and hot air jet experiments. The results show that the proposed hybrid method improves the stability of velocity-structure extraction from complex schlieren image sequences while maintaining good spatial continuity. In the hot-jet experiment, quantitative validation was focused on the dominant convective velocity along the jet centerline. The velocity values obtained by the hybrid method in this region showed good agreement with Pitot-tube measurements, indicating that the proposed method can reasonably capture the main velocity trend in the core region of the heated jet.

The complexity of multi-mode industrial processes poses challenges for traditional fault diagnosis methods in achieving satisfactory results. This challenge arises from the non-stationarity caused by changes in modes, which often overlaps with temporal variations in fault generation processes. To address this issue, we propose a fault diagnosis method based on multi-mode categorization and dynamic transfer entropy graph analysis. This method can uncover the causal information changes resulting from both mode switches and fault occurrence. Firstly, modes are categorized using the criterion of maximum intra-segment similarity and minimum inter-segment similarity, and a static causal graph is constructed for each mode. Secondly, a framework for dynamic transfer entropy graph is established to extract short-term causal information transfer using transfer entropy and compare it with the static causal graph of normal modes. Time points with significant changes in causal relationships are selected as fault time points for further analysis. Finally, an anomaly score is designed to identify critical fault nodes by selecting nodes with significant causal changes before the fault time point, thereby accurately locating the fault root nodes. To validate the performance of the proposed method, a case study using the Tennessee Eastman process is provided. Simulation results demonstrate the effectiveness and feasibility of the proposed method.

Objective Aircraft maintenance scenarios present complex spatial environments characterized by densely distributed obstacles and highly varied structures, posing significant collision risks during operations such as towing or maneuvering. Accurate segmentation of three-dimensional point cloud data obtained from laser scanning is therefore essential for effective collision detection and risk mitigation. However, traditional spectral clustering methods commonly applied for segmentation are limited by their reliance solely on spatial proximity, fixed global scale parameters insensitive to local density variations, and high computational complexity. This study proposes an improved spectral clustering algorithm specifically designed to enhance segmentation precision and computational efficiency in complex aircraft maintenance scenes. Methods The proposed method integrates several targeted improvements into the spectral clustering framework. First, supervoxel-based pre-clustering is introduced to significantly reduce the point cloud data volume while preserving local geometric consistency, thus facilitating efficient subsequent processing. Second, a novel similarity measure that combines Euclidean distance and surface normal vector similarity is developed, effectively capturing local geometric characteristics and improving segmentation accuracy along object boundaries. Additionally, a locally adaptive scale parameter selection strategy based on neighborhood standard deviations is implemented to dynamically accommodate variations in point cloud density inherent in realistic maintenance environments. Finally, the Lanczos algorithm is employed to accelerate the eigen decomposition process within spectral clustering, substantially reducing computational complexity from O(N-3) to approximately O(N-2), thereby enabling efficient handling of large-scale point cloud data. Results and Discussions Experimental validation of the proposed algorithm was conducted on two datasets: the standard indoor object segmentation database (OSD) and real-world outdoor scenarios involving aircraft towing operations, characterized by complex structures such as aircraft wings, maintenance ladders, and personnel. Results demonstrated that the proposed algorithm significantly outperformed both classical segmentation approaches (K-Medoids, LCCP) and recently developed improved spectral clustering methods (Nystr & ouml;m-SC, NVASC). On the OSD dataset, the proposed algorithm achieved an average accuracy (A(ACC)) of 82.84 % and a mean intersection over union (R-mIoU) of 79.98 %. Performance further improved under realistic conditions, with A(ACC) reaching 90.70 % and R-mIoU at 88.76 % in the outdoor aircraft towing scenario. Ablation studies clearly indicated that integrating normal vector similarity and adaptive scale parameters provided substantial contributions to improved segmentation accuracy, particularly in complex regions with overlapping structures and varying densities. Conclusions The improved spectral clustering algorithm introduced in this study effectively resolves key limitations of traditional algorithms, specifically addressing geometric feature representation, adaptive scale parameterization, and computational efficiency. The algorithm demonstrates consistent accuracy and robustness across diverse experimental conditions, confirming its suitability for real-world applications in complex aircraft maintenance environments. By providing precise segmentation results, the algorithm effectively supports critical subsequent applications such as accurate aircraft structural recognition, reliable collision risk prediction, and safe path planning. These contributions collectively serve to enhance operational safety and efficiency, thereby addressing a practical need within aircraft maintenance operations.

The air conditioning system of airplanes is one of the critical airplane-mounted systems used to control the internal environments.It is highly complex in structure,having multiple closed loops and much redundancy,which causes the faults to spread among its components.The Bayesian network can be used to deduce the fault transmission path of open-looped systems,but it is unfit for close-looped structures.In order to address this issue,an enhanced model based on the Bayesian network for determining the fault transmission path of the air conditioning system in A320 aircraft in a particular structure is presented in this research.Firstly,the functional behavior and physical structure of this air conditioning system are analyzed.By using the complex network theory,a topologically directed graph of this system is constructed.Additionally,in order to quantify the relevance of sides,the intensity of side influence is established based on the network structure and message transit,which increases the measurement's accuracy.Then,through the loop-opening strategy based on the intensity of side influence,a close-looped structure is changed into an open-looped structure.In this way,the optimal Bayesian fault transmission network structure is obtained to precisely identify the path of fault transmission in a close-looped structure.Finally,a case study is conducted on the A320 air conditioning system to validate the proposed model.

This paper proposes a novel method for the three-dimensional reconstruction of aero-engine blade ice shapes under dynamic conditions. The approach employs refraction-based optical path reconstruction to recover the ice shape and compensates for blade motion by calibrating rotational axis parameters. By establishing the positional relationship between the blade’s initial and current states, the method enables pixel coordinate projection and recovers pixel offsets caused by refraction. Experimental validation is conducted using a high-speed imaging system and a transparent resin blade model, with results compared to data from a 3D scanner. The proposed method achieves accuracy comparable to conventional scanning techniques. Moreover, it supports continuous 3D measurement of icing evolution over time without requiring frame-by-frame tracking. The results demonstrate the method’s potential for accurate, efficient, and dynamic ice shape reconstruction in aero-engine applications.

Accurate reconstruction of thermophysical states in high-temperature flow systems is essential for assessing system performance and analyzing heat and mass transfer characteristics. As a non-intrusive measurement technique, background oriented schlieren (BOS) serves as a prominent tool for high-temperature high-velocity flow field reconstruction in recent years. A key limitation of the traditional BOS method lies in the reliance on the parallel paraxial optical path assumption when determining the light deflection angle, which introduces inaccuracies. And there will be a big error when reconstructing large-scale temperature field. Therefore, this paper proposes a near-field large-scale temperature field reconstruction method based on the conservation of angular momentum and accurate analysis of the light deflection angle. This method first analyzes the background image with or without flames through a cross-correlation algorithm to obtain the speckle offset at each viewing angle. Subsequently, offset data within the same elevation plane are selected, the light deflection angle is computed through rigorous application of angular momentum conservation, and the normalized refractive index difference distribution is inverted using the iRadon reconstruction algorithm. The temperature field was finally obtained by leveraging the quantitative refractive-index-temperature relationship. Experimental results confirm the effectiveness of the angular momentum conservation method, leading to improved accuracy in the reconstruction of large-scale near-field temperature distributions compared to traditional techniques.

Aircraft skin damage poses a significant threat to the safety and airworthiness of aircraft. The considerable variation in the size of damage objects introduces challenges in feature extraction and compromises the accuracy of detection models. To address these issues, this paper proposes an improved aircraft skin damage detection algorithm based on YOLOv9s. In the feature extraction stage, a lightweight Inception module is introduced to enhance the representation of multi-scale features while reducing the model's parameter count. In the feature fusion stage, the SKFusion module, incorporating the SKNet attention mechanism, is introduced to further improve the model's ability to capture damage features at different scales. Additionally, a dynamic weighted optimization strategy combining an improved Focaler loss and inner-DIoU loss function is employed to enhance the recognition of hard-to-classify samples. The proposed method was evaluated on both a self-constructed aircraft skin damage dataset and the publicly available NEU-DET dataset. Experimental results demonstrate that, compared with the original YOLOv9s, the improved model achieves a 1.6% and 3.9% increase in mAP@50 on the self-constructed and NEU-DET datasets, respectively, while reducing the model's parameters by 16.7%.

Background-oriented schlieren (BOS) is an effective noncontact temperature field reconstruction method. However, when it is used in limited near-field measurement environments, the traditional parallel light path assumption no longer holds and the location of the temperature field is hard to measured, then the adaptive measurement capability of BOS is greatly limited. Therefore, in order to expand the application scenarios of BOS to limited near-field measurement environment, a light field BOS method has been proposed in this article to improve the reconstruction accuracy of the temperature field by employing characteristic of light field camera capturing multiple directions light and fan light path model. First, the light field imaging model is combined with Zhang's calibration method to calibrate parameters of light field camera in order to obtain the internal component distance. Second, in near-field calculation, the deviation of the rays passing through temperature field is calculated by image processing, and the center of temperature field is obtained by using rays with zero deviation. Therefore, the location of temperature field can be obtained adaptively. Finally, deflection angle is solved based on the light field BOS light path model, and then refractive index field is reconstructed using Abel inverse transformation, which in turn reconstructs temperature field. The temperature field reconstruction results has been compared with traditional parallel light path schlieren method by using the measurement values of thermocouple as standard. The experimental results show that the light field BOS reconstruction error was reduced by 2.67% at maximal 1100 K of the measured temperature field, which confirmed the feasibility of the model in limited near-field measurement environment.

To address the issues of high computational complexity and insufficient real-time performance encountered in three-dimensional object detection for complex aircraft maintenance scenes,a three-dimensional object detection method which integrates visual camera and LiDAR data and based on prior information is proposed.First,the parameters of a camera and LiDAR are calibrated,the point cloud obtained by LiDAR is preprocessed to obtain an effective three-dimensional point cloud,and the YOLOv7 algorithm is used to identify aircraft fuselage targets in the camera images.Next,the depth of the target object is calculated based on its prior length using the Efficient Perspective-n-Point(EPnP)method.Finally,depth information and point cloud clustering methods are utilized to complete three-dimensional object detection and identify obstacles.Experimental results show that the proposed method can accurately detect targets from environmental point clouds,with a recognition accuracy of 94.70%.Furthermore,the processing time for one frame is 42.96 ms,which indicates good performance in terms of both recognition accuracy and real-time capability,thus satisfying the collision risk detection requirements during aircraft movement.

With the rapid development of technologies such as Virtual reality and Augmented reality (AR), projection technology has become increasingly important in both daily life and industrial production. However, since the projection surface is not always a flat plane, traditional distortion correction methods perform poorly when dealing with projections onto curved surfaces. This paper presents a projection distortion correction method for curved surfaces based on LiDAR point cloud scanning. The method utilizes LiDAR scanning to obtain three-dimensional information of the projection surface and divides the projection image into regions according to the type of surface. Distortion correction is then achieved through perspective transformations. Experimental results show that the proposed method effectively reduces visual distortion caused by variations in surface curvature and improves both the quality and consistency of the projected image.

High-precision and multi-degree-of-freedom geometric measurement holds significant importance in feature detection for large-scale equipment manufacturing. The measurement process demands the qualities of absoluteness, simultaneity, and traceability, especially in the face of attitude compensation, target monitoring, and the construction of length references. The measurement range of commonly used high-precision optical interferometry is constrained by the wavelength of light and the size of diffraction grating, thus limiting its applicability to long distances. The optical frequency comb (OFC), with an ultra-short pulse characteristic of a periodic sequence, can be traced back to a length reference so that specific points can be determined for long-distance measurements. When integrating OFC with optical interferometry, it enables the achievement of absolute high precision distance measurements. It is essential to address the design issue which demands simultaneous multi-group distance measurements to achieve multiple-degree-of-freedom expansion. In this study, we presented a technique for three-degree-of-freedom (DOF) simultaneous measurements based on dispersive interferometry using an optical frequency comb by improving the optical structure. To solve the nonlinear problem of frequency sampling in dispersive spectrum broadening, two non-even Fourier transform algorithms are improved as a method of phase calculation. By incorporating phase ω information into the non-uniform fast Fourier transform (NUFFT) method, we achieved effective calculation of non-uniform discrete Fourier transform (NUDFT). At the same time, it can reduce the mitigate mutual interference during the extraction of multiple sets of interference peaks. The experimental findings indicate that when compared with an autocollimator, there is a consistent agreement within 3 arcsec for angles up to 1000 arcsec. This absolute measurement scheme is almost not affected by time and other factors, which provides potential for angle information monitoring.

Surface defect detection in industrial manufacturing ensures product quality and prevents malfunctions. To address issues such as multi-scale damage, low contrast, and small defects on the surfaces of industrial components, we propose an efficient multi-scale feature enhancement network for improving the detection performance of industrial surface defects. First, a multi-scale extraction module is proposed to extract defect features at multiple levels to ensure sufficient semantic information for multi-scale damage and enhance the feature extraction ability of defects with different scales. Dual-orientation attention is then introduced into the detection network to establish a connection between spatial and channel dimensional information, which enables the network to focus on defect regions and filter out background noise. This alleviates the problems of low contrast and small defects. The experimental results confirm that the proposed network demonstrates superior detection performance compared to other detection algorithms across five surface defect datasets. Additionally, the parameters are reduced by 7.9%, the floating-point operations decrease by 6.7%, and the model size is reduced by 5.2%. These improvements collectively provide an efficient solution for industrial surface defect detection.

In the process of aircraft deicing, water mist interference from the vaporization of deicing fluid is one of the primary factors affecting the accuracy of optical detection. To address this issue, this study presents, a dehazing algorithm that combines an improved dark channel prior with polarimetric imaging, aimed at enhancing image visibility in high-density water mist environments. The proposed method estimates atmospheric light intensity using polarimetric information and accurately calculates the transmission rate with the dark channel prior. This approach effectively removes the interference caused by complex water mist and also avoids the traditional algorithm's dependency on sky regions. Experimental results demonstrate that the proposed algorithm exhibits superior performance in visibility enhancement and detail recovery when processing high-density water mist images, outperforming existing advanced algorithms and providing robust technical support for deicing monitoring systems.

Small defects result in the loss of essential information during defect detection, and enhancing the detection accuracy of these minor defects is a widely researched direction, especially in the development of detection network that achieve high accuracy while simultaneously maintaining low model complexity. To strike a balance between accuracy and efficiency, this paper introduces a high-efficiency aircraft fuselage defect detection model. Firstly, we restructured a teacher model consisting of parallel backbone with Aux_IoU. This redesign enhances the detection of small defects by improving the characterization capability of the feature extraction network and implementing fine-grained bounding boxes. Secondly, to decrease the complexity of teacher model while preserving high accuracy, we introduce a hybrid logical-feature distillation framework. The student model is trained to assimilate the teacher’s feature information and logits through mask generative distillation (MGD) and logical distillation, respectively. Finally, to validate the effectiveness of proposed network, we conducted experiments on Aircraft_Fuselage_DET2023 dataset. The experimental results reveal that our student model achieves enhancements of 8.3% and 2.7% over the baseline YOLOv8n on mAP50 and mAP50:95, respectively. Furthermore, it demonstrates improvements of 0.5% and 0.2% compared to the teacher model, while simultaneously reducing the number of parameters and computational complexity by 47.3% and 53.9%. In comparison to mainstream object detection algorithms, our model achieved superior performance.

Large-scale space measurement is of great significance for the assembly of large parts. In view of the problems of traditional measurement methods, such as difficult expansion of the measurement range, complicated calibration process, and poor portability, this work proposes a multicamera vision measurement network that can freely expand the measurement range based on the auto calibration method. First, cameras were fixed to auxiliary targets and every camera with a target was mounted on a precision turntable. Second, the rotation axis parameters of the camera and the target pose relative to the camera coordinate system were estimated. Finally, the extrinsic parameters between the cameras could be estimated by interactive scanning method. The experimental results show that the method in this paper simplifies the calibration method of the multicamera vision measurement network, and the measurement accuracy basically reaches the accuracy of the Zhang type calibration method under the premise that the measurement range can be expanded freely. (c) 2024 Society of Photo-Optical Instrumentation Engineers (SPIE)

To address the problem of limited field of view measurement in traditional monocular vision meas-urement systems,we propose an omnidirectional spatial monocular vision measurement method based on a two-degree-of-freedom rotary platform.First,the rotating axis parameters of the double-degree-of-freedom rotary platform are calibrated.Then,the pictures of the checkerboard calibration plate fixed with the two-de-gree-of-freedom rotary platform are captured by using an auxiliary camera.Position coordinates of the check-erboard corner points are extracted and converted to the same camera coordinate system.The direction vec-tor of the rotating axis parameters in the initial position is obtained through PCA(principal component ana-lysis)plane fitting,and the position parameter of the rotating axis parameters in the initial position is determ-ined using the method of spatial least squares circle fitting.The camera data acquired at various angles is transformed into the same coordinate system using the rotation angle of the rotary platform and the Rodrig-ues formula.This enables measurement of the target in the horizontal and vertical omnidirectional space.Fi-nally,the measurement accuracy of the proposed method is verified using a high-precision laser rangefinder.Additionally,experiments comparing the omnidirectional spatial measurement ability of the proposed meth-od with the binocular vision measurement system and wMPS measurement system are conducted.The res-ults indicate that the method achieves a measurement accuracy comparable to that of a binocular vision sys-tem.However,it also surpasses the binocular vision system in term of measurement range,making it applic-able for omnidirectional spatial measurements.

In the current methods of point cloud processing, there are still several limitations, particularly in achieving high precision and accuracy for large objects in complex environments. Existing techniques often struggle with incomplete or noisy data, leading to inaccurate contour extraction. In view of the challenges associated with the sparse and discrete nature of point clouds in complex environments, which lead to poor accuracy and stability in object contour extraction, this paper proposes a novel method for accurately extracting the contours of three-dimensional target point clouds. The method integrates high-resolution images with sparse point cloud information to address these issues. Firstly, the local characteristics of the point cloud are calculated, allowing for the selection of a contour point cloud. Next, depth information from two-dimensional images is obtained through a fuzzy mapping relationship. Finally, constraint conditions are established to derive a more accurate predicted value of the contour point cloud. Experiments demonstrate that the proposed method effectively improves the precision and accuracy of contour extraction for large objects, reducing measurement deviation by approximately 64.9% compared to using the original point cloud alone. Additionally, the method shows a more accurate completion effect on parts of the contour that are missing, underscoring its robustness and effectiveness in challenging scenarios.

Aiming at the complexity and poor adaptability of the calibration process in the traditional unmanned aerial vehicles (UAV) indoor visual positioning, this paper proposes an omnidirectional spatial tracking and localization method for indoor UAV based on the two-axis rotary table. Firstly, the position of the UAV fuselage feature points in the camera coordinate system of the turntable camera is computed by the Pespective-n-Point algorithm utilizing known position information of a plurality of feature points with pixel coordinate information in the corresponding image. Pixel coordinate information is extracted by obtaining UAV body-specific feature points from rotary table camera shots. Then, UAV localization in omnidirectional space can be obtained by using the calibrated rotary axis parameters of the rotary table and the rotation angle of the rotary table and substituting them into Rodriguez's formula to unify the UAV position information acquired by the rotary table camera at different positions into a unified coordinate system. Finally, the angle at which the rotary table should rotate is calculated from the obtained UAV pose and the spatial position of the camera optical center and the rotary axis of the rotary table. The calculated angle is fed back to the turntable as feedback information. The rotary table receiving the feedback information is rotated to a position where the UAV is located at the center of the camera image. Thereby the tracking and localization of the UAV is realized. The experimental results show that the spatial range of localization is greatly expanded with the localization accuracy reaching the level of binocular visual localization. The omnidirectional spatial tracking and localization of indoor UAV can be conveniently realized by this method.

Measurement of the combustion temperature field is an extremely important issue in industrial production. Temperature is one of the key parameters in combustion studies. With the temperature field distribution of the combustion field obtained, heat transfer, heat convection, and heat radiation can be calculated directly and efficiently. Traditional background oriented Schlieren (BOS) is an effective method for non-axisymmetric temperature field measurements, but it requires simultaneous Schlieren imaging at multiple angles for tomographic reconstruction, which will greatly limit its application. In this paper, the compressive sensing algorithm is introduced into the temperature field reconstruction, which establishes the system of equations between the deflection angle and the refractive index gradient. Then, the reconstruction of the non-axisymmetric temperature field is realized by solving the underdetermined system of equations by the method of solving the sparse solution through the compressive sensing. First, light offsets across the non-axisymmetric temperature field are measured by the under-angled BOS system and image processing method. Second, the spatial refractive index field is reconstructed by the compressive sensing BOS method proposed in this paper. Finally, the spatial temperature field is obtained. The experimental results show that by comparing the iRadon reconstruction algorithm and the compressive sensing reconstruction algorithm, the temperature field reconstructed by the compressive sensing under the condition of the under-angled sampling of projection data had a higher accuracy than that reconstructed by the tomographic reconstruction algorithm under the same condition. The average error of the temperature field was reduced from 34.6 to 29.7 K under the same measurement conditions; however, the accuracy is better maintained by using the compressive sensing algorithm under the condition of undersampling projection.