Visual inspection and quality evaluation method for small structural component of airplane

Abstract

The embodiment of the invention discloses a visual inspection and quality evaluation method for a small structural component of an airplane, and the method comprises the steps: S1, obtaining three-dimensional point cloud data, and carrying out the preprocessing of the three-dimensional point cloud data; s2, carrying out multi-angle two-dimensional image acquisition; s3, mapping the two-dimensional image texture obtained in the S2 to the surface of a three-dimensional model, and generating the three-dimensional model comprising surface information; s4, extracting geometric features and surface texture features from the three-dimensional model; s5, fusion and comparative analysis are carried out to identify potential anomalies and defects; s6, outputting specific abnormal types and positions; and S7, performing comprehensive quality evaluation according to an anomaly detection result in the step S6, and generating a quality evaluation report. According to the technical scheme provided by the embodiment of the invention, by fusing three-dimensional scanning and high-resolution two-dimensional image technologies and adopting the classification model based on machine learning, the comprehensiveness, accuracy and efficiency of detection are remarkably improved, so that the production quality and safety are effectively improved.

Description

Technical Field

[0001] The present invention relates to, but is not limited to, the field of machine vision and automated inspection technology, and in particular to a visual inspection and quality assessment method for small aircraft structural parts. Background Art

[0002] Aircraft small structural parts are an indispensable and important part in aircraft research and development and production. They are mainly used for connecting different parts during the assembly process of aircraft components, and play the role of transferring and distributing loads between different parts. They must meet extremely high quality standards to ensure aircraft quality and safety.

[0003] Traditional quality inspection methods for small aircraft structures mainly rely on manual inspection or simple mechanical measurement. These inspection methods are not only time-consuming and inefficient, but also have obvious limitations in detection accuracy and repeatability. With the development of technology, although automated inspection technologies based on two-dimensional images have been introduced, these inspection methods are often unable to cope with complex three-dimensional geometric shapes and subtle surface textures, and it is difficult to accurately identify tiny defects and internal structural abnormalities. Summary of the invention

[0004] Purpose of the present invention: In order to solve the above-mentioned technical problems, an embodiment of the present invention provides a visual inspection and quality assessment method for small aircraft structural parts, so as to solve the problem that the existing inspection method for small aircraft structural parts has obvious limitations in inspection accuracy and repeatability, and the problem that it is difficult to accurately identify tiny defects and internal structural abnormalities in three-dimensional geometric shapes and subtle surface textures.

[0005] Technical solution of the present invention: The embodiment of the present invention provides a visual inspection and quality assessment method for small aircraft structural parts, comprising:

[0006] S1: Use a 3D scanner to perform an all-round scan of the aircraft's small structural parts, obtain the 3D point cloud data of the aircraft's small structural outer surface, and pre-process the 3D point cloud data;

[0007] S2: Use high-resolution cameras to collect 2D images of small aircraft structures from multiple angles;

[0008] S3: constructing a three-dimensional model using the preprocessed three-dimensional point cloud data, and mapping the texture of the two-dimensional image obtained in S2 to the surface of the three-dimensional model to generate a three-dimensional model including surface texture information;

[0009] S4: Use 3D feature extraction algorithm to extract geometric features and surface texture information from 3D models. The extracted content includes the shape, size, surface texture information and subtle defects of small structural parts.

[0010] S5: Using multimodal data fusion, the geometric features and surface texture information extracted in S4 are fused and compared with standard samples to identify potential anomalies and defects;

[0011] S6: Analyze the fused comparison data in S5 using the pre-trained classification model and output the anomaly detection results, including the anomaly type and location of the aircraft's small structural parts;

[0012] S7: Perform a comprehensive quality assessment on the anomaly detection results of S6, generate a quality assessment report, and put forward optimization suggestions based on the detection results.

[0013] Optionally, in the visual inspection and quality assessment method for small aircraft structures as described above, a three-dimensional scanner for scanning is pre-positioned in a working area of ​​the small aircraft structure, and S1 includes:

[0014] S11: Performing initial calibration on the 3D scanner, including calibrating the position, angle and focal length of the scanner, so that the scanning area of ​​the calibrated 3D scanner covers the entire outer surface of the small structural parts of the aircraft;

[0015] S12: Controlling the 3D scanner to perform multi-view scanning on the small aircraft structure according to a predetermined path and angle, and obtaining 3D point cloud data corresponding to each view;

[0016] S13: splicing and fusing the three-dimensional point cloud data corresponding to each viewpoint obtained in S12. In S13, an iterative closest point algorithm ICP is used to align multiple groups of three-dimensional point cloud data to generate a three-dimensional point cloud data set.

[0017] S14: performing denoising processing on the spliced ​​three-dimensional point cloud data set, and removing noise points in the three-dimensional point cloud data set using a statistical filtering algorithm, wherein the statistical filtering algorithm identifies and removes abnormal points by calculating neighborhood statistical characteristics of each point;

[0018] S15: identifying and removing redundant points and isolated points in the three-dimensional point cloud data set by a density clustering algorithm; the density clustering algorithm detects and deletes redundant points with insufficient density by defining a density threshold and a minimum number of neighbors;

[0019] S16: A global least squares fitting algorithm is used to globally align the denoised and de-redundant 3D point cloud dataset.

[0020] Optionally, in the visual inspection and quality assessment method for small aircraft structures as described above, a high-resolution camera is pre-positioned in a working area of ​​the small aircraft structure, and S2 includes:

[0021] S21: performing an initial calibration on the high-resolution camera, including calibrating the position, angle, and focal length of the camera so that the camera covers all surface areas of the small structural component;

[0022] S22: Design a multi-angle shooting path, determine specific shooting points of the camera at different positions and angles, and the shooting path includes positions at least 8 different angles;

[0023] S23: Setting the image acquisition parameters of the camera, including exposure time, aperture size and sensitivity;

[0024] S24: according to the planned shooting path, control the camera to perform multi-angle image acquisition of the small structural part of the aircraft at a predetermined position and angle, obtain all-round two-dimensional image data of the small structural part of the aircraft, and shoot at least three images at each angle at each position;

[0025] S25: using an image distortion correction algorithm to correct the distortion in the collected multi-position and multi-angle images, and performing color correction and contrast enhancement;

[0026] S26: Stitch and fuse the two-dimensional images collected from multiple positions and angles, and use an image stitching algorithm to seamlessly stitch the images into a complete image covering the entire surface of the small structural part. During the stitching process, the overlapping portion between the images is no less than 30%.

[0027] Optionally, in the visual inspection and quality assessment method for small aircraft structural parts as described above, S3 includes:

[0028] S31: using the pre-processed 3D point cloud data obtained in S1, constructing a preliminary 3D model through a Delaunay triangulation algorithm, and generating a mesh model including the geometric shape of the small structural parts of the aircraft;

[0029] S32: registering the two-dimensional image obtained in S2 with the preliminary three-dimensional model, including: using a SIFT algorithm to identify corresponding feature points in the two-dimensional image and the three-dimensional model, and aligning the feature points through a camera projection matrix;

[0030] S33: performing texture mapping on the registered two-dimensional image, mapping the surface texture information in the two-dimensional image to the outer surface of the three-dimensional model; wherein the texture mapping adopts a perspective projection texture mapping algorithm, and projects the texture coordinates corresponding to each pixel of the two-dimensional image onto the surface of the three-dimensional model;

[0031] S34: performing texture fusion on the two-dimensional images acquired at multiple positions and angles, and using a weighted average fusion algorithm to fuse the textures of the overlapping areas. During the fusion process, weighted average processing is performed on the pixels of each overlapping area.

[0032] S35: Perform detail enhancement on the texture mapped and fused three-dimensional model, specifically using a high dynamic range enhancement algorithm to adjust the brightness and contrast of the texture.

[0033] Optionally, in the visual inspection and quality assessment method for small aircraft structural parts as described above, S4 includes:

[0034] S41: extracting geometric shape features from the three-dimensional model, including: fitting each surface of the three-dimensional model using a surface fitting algorithm to obtain the geometric shape of the small aircraft structure, wherein the surface fitting algorithm minimizes the error between the fitting surface and the three-dimensional point cloud data by the least square method, and the calculation formula is:

[0035]

[0036] Among them, z i is a point in the point cloud data, f(x i ,y i ) is the fitting surface function, (x i ,y i ) is the coordinate of the point cloud data, n1 is the number of points;

[0037] S42: Measuring the dimensions of the three-dimensional model, including: measuring the dimensions of each part of the small structural part of the aircraft using a distance calculation algorithm, wherein the distance calculation algorithm is used to calculate the Euclidean distance between two points in the three-dimensional model, and the calculation formula is:

[0038]

[0039] Where (x1, y1, z1) and (x2, y2, z2) are the coordinates of two points in the three-dimensional model, and d is the distance between the two points;

[0040] S43: extracting surface texture information of the three-dimensional model, including: using a normal vector analysis algorithm to calculate the normal vector of each point of the three-dimensional model, where the normal vector is defined as a vector perpendicular to the surface, and the calculation formula is:

[0041]

[0042] Where N is the normal vector, v1 and v2 are two non-parallel vectors on the surface, × represents the vector cross product, and |·| represents the modulus of the vector;

[0043] S44: Identify subtle defects in 3D models, including: using Gaussian curvature and mean curvature analysis algorithms to calculate the curvature of the 3D model surface, and identify subtle defects such as dents, protrusions and cracks on the surface of small aircraft structural parts.

[0044] Optionally, in the visual inspection and quality assessment method for small aircraft structural parts as described above, the Gaussian curvature and mean curvature analysis algorithms in S44 are respectively:

[0045]

[0046] Among them, κ1 and κ2 are the principal curvatures, and are the partial derivatives of the surface in the x and y directions respectively, K is the Gaussian curvature, and H is the mean curvature.

[0047] Optionally, in the visual inspection and quality assessment method for small aircraft structural parts as described above, S5 comprises:

[0048] S51: Obtain geometric features and surface texture information of standard samples, including shape, size, and surface texture information of small aircraft structural parts, and use them as benchmark data;

[0049] S52: Compare and analyze the geometric features extracted in S4 with the geometric features of the standard sample, and use a shape matching algorithm to evaluate the consistency of the geometric features by calculating the similarity between the two. The shape matching algorithm formula is:

[0050]

[0051] Among them, A and B represent the geometric feature point sets of small aircraft structures and standard samples respectively, n2 is the number of geometric feature points, ∥A a -B b ∥ is the geometric feature point A of the aircraft small structural part and the standard sample a and B b The Euclidean distance between

[0052] S53: Compare and analyze the surface texture information extracted in S4 with the surface texture information of the standard sample, and use a texture similarity algorithm to evaluate the consistency of the surface texture information by calculating the similarity between the surface texture information. The texture similarity algorithm formula is:

[0053]

[0054] Wherein, T1 and T2 represent the texture images of the aircraft small structure and the standard sample respectively, m is the height of the texture image, n3 is the width of the texture image, |T1(v, c)-T2(v, c)| is the absolute difference between the corresponding pixels of the aircraft small structure and the standard sample;

[0055] S54: Fusing the geometric features and the surface texture information, using a weighted fusion algorithm to combine the geometric feature similarity and the texture information similarity to form a comprehensive similarity score. The weighted fusion algorithm formula is:

[0056] Fused Similarity = w g Geometry Similarity

[0057] +w t Texture Similarity

[0058] Among them, w g and w t are the weight coefficients of geometric features and surface texture information respectively, Geometry Similarity is the geometric feature similarity, and Texture Similarity is the surface texture information similarity;

[0059] S55: Based on a preset threshold, when the comprehensive similarity score is lower than the threshold, it is marked as a potential abnormality and defect area.

[0060] Optionally, in the visual inspection and quality assessment method for small aircraft structural parts as described above, S6 includes:

[0061] S61: The pre-trained classification model is a classification model based on an extreme learning machine;

[0062] S62: Use the geometric features and surface texture information in the historical detection data and the standard samples as training data, randomly generate input weights and biases through the extreme learning machine, and determine the output weights to complete the classification model training;

[0063] S63: input the comprehensive similarity score after the fusion of multimodal data in S5 into the trained extreme learning machine classification model, analyze and process it, and generate a classification result;

[0064] S64: According to the output result of the classification model, the category to which the input data belongs is determined, including normal, surface defect, and geometric deformation. The extreme learning machine calculates the classification probability of the input data and maps it to the predefined abnormal category, and outputs the final abnormal type and abnormal feature points;

[0065] S65: Determine the position of the anomaly by analyzing the abnormal feature points output in the classification result, including: in the three-dimensional model, locate the specific position of the abnormal feature points on the small structural parts of the aircraft according to the spatial coordinates of the abnormal feature points, combined with geometric features and surface texture information, and generate corresponding marking information.

[0066] Optionally, in the visual inspection and quality assessment method for small aircraft structural parts as described above, S65 specifically includes:

[0067] S651: Determine the spatial coordinates. The spatial coordinates (x i ,y i , z i ) to extract and record the specific positions of abnormal feature points in the three-dimensional model;

[0068] S652: combining geometric features, locating the abnormal feature points according to the spatial coordinates of the abnormal feature points using geometric features, specifically including calculating the distance between the abnormal feature points and the surrounding geometric feature points;

[0069] S653: Determine the coordinates of the two-dimensional image and map the spatial coordinates of the abnormal feature points to the texture coordinates of the two-dimensional image (u d , v d ), using the perspective projection formula for conversion, the expression is:

[0070]

[0071] Among them, (F x , F y ) is the focal length of the camera, (C x , C y ) is the principal point coordinate of the camera;

[0072] S654: Combine surface texture information and use texture coordinates (u d , v d ) Extract the corresponding texture information in the two-dimensional image and calculate the similarity between the abnormal feature point o and the surrounding texture feature points. The calculation formula is:

[0073]

[0074] Among them, S oe is the similarity between the abnormal feature point o and the texture feature point e, T f (k) and T f (k) are the e-th texture values ​​of the abnormal feature point o and the texture feature point e, respectively, and n4 is the number of texture feature points;

[0075] S655: marking abnormal feature points, including: combining geometric features and surface texture information, and determining the final marking value of abnormal feature point i by weighted averaging the geometric distance and texture similarity of the abnormal feature points. The calculation formula is:

[0076] M o =W h ·min k (d op )+W t ·min k (S oe );

[0077] Among them, M o is the final label value of the abnormal feature point o, W h and W t are the weight coefficients of geometric features and surface texture information, min k (d op ) is the minimum distance between the abnormal feature point o and the geometric feature point p, min k (S oe ) is the minimum similarity between the abnormal feature point and the texture feature point;

[0078] S656: Generate tag information and set the final tag value M o The abnormality type and location are converted into descriptive information, and annotated on the 3D model and 2D image to generate label information including the spatial coordinates of the abnormal feature point o, the abnormality type, and the degree of abnormality; the label format is: label information = {coordinate: (x c ,y c , z c ), Type: Abnormal Type, Degree: M o}.

[0079] Optionally, in the visual inspection and quality assessment method for small aircraft structural parts as described above, S7 includes:

[0080] S71: Summarize the anomaly detection results, collect and organize the anomaly types and anomaly locations output by the classification model in S6, including the spatial coordinates, anomaly type, and anomaly degree of each anomaly feature point, to form a preliminary detection result data set;

[0081] S72: Comprehensively analyze the detected abnormal data, including calculating the data, distribution location and type ratio of abnormal feature points. Specifically, statistical analysis methods are used to evaluate the severity of various abnormalities and their distribution in the overall small structural parts of the aircraft. The statistical analysis formula is:

[0082]

[0083] Among them, P(type) is the severity of the corresponding type of abnormality, n5 is the number of abnormal feature points of the corresponding type, and n6 is the total number of all abnormal feature points;

[0084] S73: Develop comprehensive quality assessment standards based on industry standards and production requirements to determine the geometric tolerances, surface quality and allowable range of defects;

[0085] S74: Compare the severity of each abnormal point with the evaluation standard, and calculate the score of each abnormal feature point; wherein, for each abnormal type, the score of the abnormal feature point is calculated using the following scoring formula:

[0086]

[0087] Among them, q r The score of the rth quality assessment indicator is given by the maximum allowed value, which is the pre-set quality assessment standard;

[0088] S75: Score all outliers q r The weighted average of the parts is used to calculate the comprehensive quality assessment score of the aircraft small structural parts. The quality assessment standard formula used is:

[0089]

[0090] Among them, Q is the comprehensive quality assessment score, q r is the score of the rth quality evaluation indicator, w is the weight coefficient of the indicator, and m1 is the number of evaluation indicators;

[0091] S76: Quality assessment report generation: According to the comprehensive quality assessment standards, the test results are quantitatively scored and graded to generate a quality assessment report. The report content includes basic information of the test object, test method, test results, comprehensive assessment score, quality level and specific abnormality description;

[0092] S77: Based on the quality assessment report, make optimization suggestions for the test results, including improving design, optimizing manufacturing processes and strengthening quality control.

[0093] Beneficial effects of the invention: The embodiment of the invention provides a visual inspection and quality assessment method for small aircraft structural parts, which has the following beneficial effects:

[0094] On the one hand, by integrating 3D scanning and high-resolution 2D imaging technology, comprehensive and accurate inspection of small structural parts can be achieved. This multimodal fusion not only enhances the ability to identify surface and internal structural abnormalities of small structural parts, but also can accurately locate tiny defects and potential structural problems through detailed surface texture analysis. In addition, this method also significantly improves image quality through high dynamic range images and advanced image processing technology, ensuring the accuracy and reliability of inspection data.

[0095] On the other hand, by adopting a classification model based on machine learning, not only the efficiency and accuracy of anomaly detection are improved, but it can also automatically learn and adapt to new detection situations based on historical data. This intelligent analysis method reduces the need for manual operation, reduces detection costs, and improves the level of automation in the production process. The generation of comprehensive quality assessment reports and the proposal of optimization suggestions further help companies monitor and improve the production process, thereby improving product quality while ensuring aviation safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0096] The accompanying drawings are used to provide a further understanding of the technical solution of the present invention and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the technical solution of the present invention and do not constitute a limitation on the technical solution of the present invention.

[0097] Figure 1 A schematic flow chart of a method for visual inspection and quality assessment of small aircraft structural parts provided by an embodiment of the present invention;

[0098] Figure 2 A schematic diagram of a process for obtaining three-dimensional point cloud data in a method for visual inspection and quality assessment of small aircraft structural parts provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0099] In order to make the purpose, technical solution and advantages of the present invention more clear, the embodiments of the present invention will be described in detail with reference to the accompanying drawings. It should be noted that the embodiments and features in the embodiments of the present application can be combined with each other arbitrarily without conflict.

[0100] The above background technology has already explained the importance of small aircraft structural parts in aircraft development and the necessity of quality inspection. The existing inspection methods for small aircraft structural parts have obvious limitations in inspection accuracy and repeatability, and it is difficult to accurately identify small defects and internal structural abnormalities in three-dimensional geometric shapes and subtle surface textures.

[0101] Existing detection methods are unable to effectively integrate 3D geometric information and 2D image texture information, resulting in a lack of depth and accuracy in the comprehensive quality assessment of small aircraft structures. This technical limitation is particularly prominent in the detection of small surface defects and internal structural anomalies of small structures. For example, it is difficult to accurately capture small cracks or depressions on the complex curved surfaces of small structures using only 2D image technology, and it is impossible to effectively evaluate the details and texture of the surface material by relying solely on 3D scanning data. In addition, the inefficiency of existing detection methods in data processing and anomaly positioning has also seriously affected production efficiency and cost control.

[0102] In response to the above problems, an embodiment of the present invention provides a visual inspection and quality assessment method for small aircraft structural parts. By developing a new technology that can efficiently integrate three-dimensional scanning and two-dimensional image data, and automatically perform in-depth analysis and quality assessment, it has important practical significance for improving the manufacturing and inspection standards of small aircraft structural parts and ensuring aviation safety.

[0103] The present invention provides the following specific embodiments which can be combined with each other. The same or similar concepts or processes may not be described in detail in some embodiments.

[0104] Figure 1 The following is a flow chart of a method for visual inspection and quality assessment of small aircraft structural parts provided by an embodiment of the present invention. Figure 1 As shown, the visual inspection and quality assessment method for small aircraft structural parts provided by the embodiment of the present invention includes the following steps:

[0105] S1: Use a 3D scanner to perform an all-round scan of the aircraft's small structural parts, obtain the 3D point cloud data of the aircraft's small structural outer surface, and pre-process the 3D point cloud data;

[0106] S2: Use high-resolution cameras to collect 2D images of small aircraft structures from multiple angles;

[0107] S3: constructing a three-dimensional model using the preprocessed three-dimensional point cloud data, and mapping the texture of the two-dimensional image obtained in S2 to the surface of the three-dimensional model to generate a three-dimensional model including surface texture information;

[0108] S4: Use 3D feature extraction algorithm to extract geometric features and surface texture information from 3D models. The extracted content includes the shape, size, surface texture information and subtle defects of small structural parts.

[0109] S5: Using multimodal data fusion, the geometric features and surface texture information extracted in S4 are fused and compared with standard samples to identify potential anomalies and defects;

[0110] S6: Analyze the fused comparison data in S5 using the pre-trained classification model and output the anomaly detection results, including the anomaly type and location of the aircraft's small structural parts;

[0111] S7: Perform a comprehensive quality assessment on the anomaly detection results of S6, generate a quality assessment report, and put forward optimization suggestions based on the detection results.

[0112] In one implementation of the embodiment of the present invention, a three-dimensional scanner for scanning can be pre-positioned in the working area of ​​the small structural part of the aircraft; specifically, the three-dimensional scanner for scanning is pre-positioned in the working area of ​​the small structural part of the aircraft, and the three-dimensional scanner is initially calibrated, including calibrating the position, angle and focal length of the scanner, so that the scanning area of ​​the calibrated three-dimensional scanner covers the entire outer surface of the small structural part of the aircraft. Figure 2 A schematic diagram of a process for obtaining three-dimensional point cloud data in a method for visual inspection and quality assessment of small aircraft structural parts provided by an embodiment of the present invention. The implementation process of S1 above may include:

[0113] S11: Performing initial calibration on the 3D scanner, including calibrating the position, angle and focal length of the scanner, so that the scanning area of ​​the calibrated 3D scanner covers the entire outer surface of the small structural parts of the aircraft;

[0114] S12: Controlling the 3D scanner to perform multi-view scanning on the small aircraft structure according to a predetermined path and angle, and obtaining 3D point cloud data corresponding to each view;

[0115] S13: splicing and fusing the three-dimensional point cloud data corresponding to each perspective obtained in S12;

[0116] In this step, an iterative closest point algorithm (ICP) can be used to align multiple groups of 3D point cloud data to generate a 3D point cloud data set;

[0117] S14: performing denoising processing on the spliced ​​three-dimensional point cloud data set, and removing noise points in the three-dimensional point cloud data set using a statistical filtering algorithm, wherein the statistical filtering algorithm identifies and removes abnormal points by calculating neighborhood statistical characteristics of each point;

[0118] S15: Identify and remove redundant points and isolated points in the three-dimensional point cloud dataset through a density clustering algorithm (such as the DBSCAN algorithm);

[0119] In this step, the density clustering algorithm detects and deletes redundant points with insufficient density by defining the density threshold and the minimum number of neighbors;

[0120] S16: A global least squares fitting algorithm is used to globally align the denoised and de-redundant 3D point cloud data set to ensure the overall consistency and accuracy of the 3D point cloud data. The global least squares fitting algorithm achieves the overall alignment effect by minimizing the square error between point clouds.

[0121] In this step, the specific steps of global alignment are as follows:

[0122] Defining the error function: The global least squares fitting algorithm first defines an error function, which quantifies the difference between two sets of point cloud data. When performing global alignment, the square error function is selected to quantify the Euclidean distance between point clouds. Let the error function be E, and the expression is: Among them, p i is a point in the source point cloud, q i is the corresponding point in the target point cloud, T is a transformation function, including translation, rotation, etc.;

[0123] Finding the optimal transformation: The goal of the algorithm is to find a transformation T that minimizes the error function E. This involves solving the rotation matrix R and the translation vector t. In order to minimize the error function, R and t need to be differentiated and set to zero. Solving this optimization problem requires an iterative method, such as the iterative closest point algorithm or nonlinear minimization technique.

[0124] Iterative solution: In practical applications, in order to improve the alignment accuracy, an iterative method is used to gradually optimize R and t. In each iteration, the position of the point cloud is updated according to the current transformation estimate, and then the error function is recalculated and the transformation parameters are updated;

[0125] Finally: The iteration continues until the transformation converges or the preset number of iterations is reached, and the final R and t will be used to align all scanned point cloud data into a global coordinate.

[0126] In this implementation, by executing the above S11 to S16, it is possible to achieve all-round high-precision scanning of small aircraft structures, and perform comprehensive pre-processing of the acquired three-dimensional point cloud data. The specific implementation of steps such as positioning and calibration, multi-view scanning, point cloud data stitching and fusion, denoising, redundant point removal and point cloud data alignment ensures the high quality and high precision of the point cloud data, and provides a reliable data foundation for subsequent three-dimensional model reconstruction and feature extraction.

[0127] In an implementation of the embodiment of the present invention, a high-resolution camera may be pre-positioned in a working area of ​​a small aircraft structure; the implementation process of the above S2 may include:

[0128] S21: Initial calibration of a high-resolution camera pre-positioned in the working area of ​​a small aircraft structure, including calibration of the camera's position, angle, and focal length, so that the camera covers all surface areas of the small structure;

[0129] In this step, the camera is made to cover all surface areas of the small structure. The camera should be 10-30 cm away from the small structure, and the angle should be adjusted to between 45 and 90 degrees with the surface normal of the small structure to ensure the best imaging effect.

[0130] S22: Design a multi-angle shooting path and determine the specific shooting points of the camera at different positions and angles to ensure that all surface details of the small structural parts can be fully captured. The shooting path includes at least 8 positions at different angles to ensure that each angle is evenly spaced to cover the entire surface of the small structural parts.

[0131] S23: Setting the image acquisition parameters of the camera, including exposure time, aperture size and sensitivity;

[0132] This step requires ensuring that clear, detail-rich images can be obtained under different lighting conditions. The specific parameters can be set as follows: exposure time 1 / 100 second to 1 / 1000 second, aperture size f / 2.8 to f / 8, sensitivity ISO 100 to ISO 800.

[0133] S24: according to the planned shooting path, control the camera to perform multi-angle image acquisition of the small structural parts of the aircraft at the predetermined position and angle, obtain all-round two-dimensional image data of the small structural parts of the aircraft, and shoot at least three images at each angle at each position to ensure that the acquired images are comprehensive and clear enough;

[0134] S25: using an image distortion correction algorithm (such as a distortion correction algorithm based on a calibration plate) to correct the distortion in the collected multi-position and multi-angle images, and performing color correction and contrast enhancement to improve the quality and consistency of the images;

[0135] S26: Stitching and fusing the two-dimensional images collected at multiple positions and angles, and using an image stitching algorithm (such as the SIFT feature point matching algorithm) to seamlessly stitch the images into a complete image covering the entire surface of the small structural part. During the stitching process, the overlapping portion between the images is no less than 30% to ensure stitching quality and accuracy.

[0136] In this implementation, through the above steps S21 to S26, it is possible to achieve all-round high-resolution image acquisition of small aircraft structures, and to perform comprehensive processing of the acquired image data, and the specific implementation of steps such as positioning and calibration, multi-angle shooting path planning, image acquisition parameter setting, multi-angle image acquisition, image correction and processing, and image stitching and fusion ensures the high quality and high precision of the image data, and provides a reliable image basis for subsequent three-dimensional model reconstruction and feature extraction.

[0137] In an implementation of the embodiment of the present invention, the implementation process of the above S3 may include:

[0138] S31: using the pre-processed 3D point cloud data obtained in S1, constructing a preliminary 3D model through a Delaunay triangulation algorithm, and generating a mesh model including the geometric shape of the small structural parts of the aircraft;

[0139] S32: registering the two-dimensional image obtained in S2 with the preliminary three-dimensional model; this step includes: using the SIFT algorithm to identify corresponding feature points in the two-dimensional image and the three-dimensional model, and aligning the feature points through the camera projection matrix.

[0140] The SIFT feature point detection formula in this step is:

[0141] D(x,σ)=(L(x,kσ)-L(x,σ))·G(x,σ);

[0142] Among them, x represents a point in the image, σ represents the scale parameter, L represents the image scale space, k is the proportional coefficient, D(x, σ) represents the difference of the feature point in the scale space, L(x, σ) represents the Gaussian blur result of the image at scale σ, and G(x, σ) represents the Gaussian kernel function at scale σ.

[0143] S33: Texture mapping is performed on the registered two-dimensional image to map the surface texture information in the two-dimensional image to the outer surface of the three-dimensional model; wherein the texture mapping adopts a perspective projection texture mapping algorithm to project the texture coordinates corresponding to each two-dimensional image pixel onto the surface of the three-dimensional model; the perspective projection formula in this step is:

[0144]

[0145] Among them, u and v represent the coordinates in the two-dimensional image, x, y, and z represent the coordinates in the three-dimensional model, and f is the focal length.

[0146] S34: performing texture fusion on the two-dimensional images acquired at multiple positions and angles, and using a weighted average fusion algorithm to fuse the textures of the overlapping areas. During the fusion process, weighted average processing is performed on the pixels of each overlapping area.

[0147] The formula for weighted average processing in this step is:

[0148] Among them, I f Represents the pixel value after fusion, I i represents the pixel value of the i-th image, w i represents the weight of the i-th image, and n represents the number of images.

[0149] S35: Enhance the details of the texture mapped and fused 3D model, specifically using a high dynamic range (HDR) enhancement algorithm to adjust the brightness and contrast of the texture to ensure the details and realism of the texture.

[0150] The enhanced pixel value I in this step e From the original pixel value I o It is obtained through logarithmic transformation and contrast adjustment, and the calculation formula is:

[0151] I e =log(1+I o )·α;

[0152] Among them, I e Represents the enhanced pixel value, I o Represents the original pixel value, α represents the contrast adjustment coefficient, and log represents the natural logarithm function.

[0153] In this implementation, by executing the above steps S31 to S35, it is possible to accurately map the two-dimensional image texture to the three-dimensional model surface, generate a high-precision three-dimensional model containing surface information, and specifically implement the steps of constructing a preliminary three-dimensional model by Delaunay triangulation, performing image registration by SIFT algorithm, perspective projection texture mapping, multi-image weighted average fusion, HDR detail enhancement, etc., to ensure the high quality and high precision of the three-dimensional model, and provide a reliable data basis for subsequent feature extraction and quality assessment. This method not only improves the accuracy of three-dimensional reconstruction, but also significantly improves the details and realism of the model.

[0154] In an implementation of the embodiment of the present invention, the implementation process of the above S4 may include:

[0155] S41: extracting geometric shape features from the three-dimensional model, including: fitting each surface of the three-dimensional model using a surface fitting algorithm to obtain the geometric shape of the small aircraft structure, wherein the surface fitting algorithm minimizes the error between the fitting surface and the three-dimensional point cloud data by the least square method, and the calculation formula is:

[0156]

[0157] Among them, z i is a point in the point cloud data, f(x i ,y i ) is the fitting surface function, (x i ,y i ) is the coordinate of the point cloud data, n1 is the number of points;

[0158] S42: Measuring the dimensions of the three-dimensional model, including: measuring the dimensions of each part of the small structural part of the aircraft using a distance calculation algorithm, wherein the distance calculation algorithm is used to calculate the Euclidean distance between two points in the three-dimensional model, and the calculation formula is:

[0159]

[0160] Where (x1, y1, z1) and (x2, y2, z2) are the coordinates of two points in the three-dimensional model, and d is the distance between the two points;

[0161] S43: extracting surface texture information of the three-dimensional model, including: using a normal vector analysis algorithm to calculate the normal vector of each point of the three-dimensional model, where the normal vector is defined as a vector perpendicular to the surface, and the calculation formula is:

[0162]

[0163] Where N is the normal vector, v1 and v2 are two non-parallel vectors on the surface, × represents the vector cross product, and |·| represents the modulus of the vector;

[0164] S44: Identify subtle defects in 3D models, including: using Gaussian curvature and mean curvature analysis algorithms to calculate the curvature of the 3D model surface, and identify subtle defects such as dents, protrusions and cracks on the surface of small aircraft structural parts.

[0165] In this implementation, by executing the above steps S41 to S44, the geometric features and surface texture features of small aircraft structures can be accurately extracted from the three-dimensional model, including the shape, size, surface texture and subtle defects of small aircraft structures. The specific implementation of steps such as surface fitting algorithm, distance calculation algorithm, normal vector analysis algorithm, Gaussian curvature and mean curvature analysis algorithm ensures high quality and high precision of feature extraction, and provides a reliable data foundation for subsequent multimodal data fusion and quality assessment.

[0166] Further, in one embodiment of this implementation, the Gaussian curvature and mean curvature analysis algorithms in the above S44 are respectively:

[0167]

[0168] Among them, κ1 and κ2 are the principal curvatures, and are the partial derivatives of the surface in the x and y directions respectively, K is the Gaussian curvature, and H is the mean curvature.

[0169] In an implementation of the embodiment of the present invention, the implementation process of the above S5 may include:

[0170] S51: Obtain geometric features and surface texture information of standard samples, including shape, size, and surface texture information of small aircraft structural parts, and use them as benchmark data;

[0171] S52: Compare and analyze the geometric features extracted in S4 with the geometric features of the standard sample, and use a shape matching algorithm to evaluate the consistency of the geometric features by calculating the similarity between the two. The shape matching algorithm formula is:

[0172]

[0173] Among them, A and B represent the geometric feature point sets of small aircraft structures and standard samples respectively, n2 is the number of geometric feature points, ∥A a -B b ∥ is the geometric feature point A of the aircraft small structural part and the standard sample a and B b The Euclidean distance between

[0174] S53: Compare and analyze the surface texture information extracted in S4 with the surface texture information of the standard sample, and use a texture similarity algorithm to evaluate the consistency of the surface texture information by calculating the similarity between the surface texture information. The texture similarity algorithm formula is:

[0175]

[0176] Wherein, T1 and T2 represent the texture images of the aircraft small structure and the standard sample respectively, m is the height of the texture image, n3 is the width of the texture image, |T1(v, c)-T2(v, c)| is the absolute difference between the corresponding pixels of the aircraft small structure and the standard sample;

[0177] S54: fusing the geometric features and the surface texture information, using a weighted fusion algorithm to combine the geometric feature similarity and the texture information similarity to form a comprehensive similarity score;

[0178] The weighted fusion algorithm formula in this step is:

[0179] Fused Similarity = w g Geometry Similarity+w t Texture Similarity

[0180] Among them, w g and w t are the weight coefficients of geometric features and surface texture information respectively, Geometry Similarity is the geometric feature similarity, and Texture Similarity is the surface texture information similarity;

[0181] S55: Based on a preset threshold, when the comprehensive similarity score is lower than the threshold, it is marked as a potential abnormality and defect area.

[0182] In this implementation, by executing the above steps S51 to S55, the fusion and comparative analysis of the geometric features and surface texture information extracted from S4 and the standard sample can be achieved. The specific implementation of steps such as the shape matching algorithm, texture similarity algorithm, and multimodal data fusion algorithm ensures high quality and high accuracy of anomaly and defect identification, and provides a reliable data basis for subsequent quality assessment and optimization suggestions.

[0183] In an implementation of the embodiment of the present invention, the implementation process of the above S6 may include:

[0184] S61: The pre-trained classification model is a classification model based on an extreme learning machine (ELM), which has fast learning and good generalization performance and is suitable for processing high-dimensional feature data after multimodal fusion;

[0185] S62: Use the geometric features and surface texture information in the historical detection data and the standard samples as training data, randomly generate input weights and biases through the extreme learning machine, and determine the output weights to complete the classification model training;

[0186] The specific training process in this step includes data preprocessing, feature standardization, and classification model parameter optimization to ensure the high accuracy and stability of the classification model.

[0187] S63: Input the comprehensive similarity score after the fusion of multimodal data in S5 into the trained extreme learning machine classification model for analysis and processing. After receiving the input data, the model generates a classification result through nonlinear transformation of hidden layer nodes and linear combination of output layer.

[0188] S64: According to the output result of the classification model, the category to which the input data (geometric features and surface texture information) belongs is determined, including normal, surface defects, and geometric deformation; the extreme learning machine calculates the classification probability of the input data and maps it to the predefined abnormal category, and outputs the final abnormal type and abnormal feature points;

[0189] S65: Determine the position of the anomaly by analyzing the abnormal feature points output in the classification result, including: in the three-dimensional model, locate the specific position of the abnormal feature points on the small structural parts of the aircraft according to the spatial coordinates of the abnormal feature points, combined with geometric features and surface texture information, and generate corresponding marking information.

[0190] In this implementation, by executing the above steps S61 to S65, the fused comparison data in S5 is analyzed using the pre-trained extreme learning machine-based classification model, which can quickly and accurately output the specific anomaly type and location. The specific implementation of the steps of classification model selection, model training, feature data input, anomaly type identification and anomaly location positioning ensures high quality and high precision of anomaly detection and positioning, and provides a reliable basis for subsequent comprehensive quality evaluation and optimization suggestions.

[0191] Furthermore, in one implementation of the implementation method, the implementation process of the above S65 specifically includes:

[0192] S651: Determine the spatial coordinates. The spatial coordinates (x i ,y i , z i) to extract and record the specific positions of abnormal feature points in the three-dimensional model;

[0193] S652: Geometric feature combination: according to the spatial coordinates of the abnormal feature point, the abnormal feature point is located using the geometric feature, specifically including calculating the distance between the abnormal feature point and the surrounding geometric feature points. The calculation formula is:

[0194]

[0195] Among them, d op is the Euclidean distance between the outlier point o and the geometric feature point p, (x o ,y o , z o ) and (x p ,y p , z p ) are the spatial coordinates of the outlier points and geometric feature points respectively;

[0196] S653: Determine the coordinates of the two-dimensional image and map the spatial coordinates of the abnormal feature points to the texture coordinates of the two-dimensional image (u d , v d ), using the perspective projection formula for conversion, the expression is:

[0197]

[0198] Among them, (F x , F y ) is the focal length of the camera, (C x , C y ) is the principal point coordinate of the camera;

[0199] S654: Combine surface texture information and use texture coordinates (u d , v d ) Extract the corresponding texture information in the two-dimensional image and calculate the similarity between the abnormal feature point o and the surrounding texture feature points. The calculation formula is:

[0200]

[0201] Among them, S oe is the similarity between the abnormal feature point o and the texture feature point e, T f (k) and T f (k) are the e-th texture values ​​of the abnormal feature point o and the texture feature point e, respectively, and n4 is the number of texture feature points;

[0202] S655: marking abnormal feature points, including: combining geometric features and surface texture information, and determining the final marking value of abnormal feature point i by weighted averaging the geometric distance and texture similarity of the abnormal feature points. The calculation formula is:

[0203] M o =W h ·min k (d op )+W t ·min k (S oe );

[0204] Among them, M o is the final label value of the abnormal feature point o, W h and W t are the weight coefficients of geometric features and surface texture information, min k (d op ) is the minimum distance between the abnormal feature point o and the geometric feature point p, min k (S oe ) is the minimum similarity between the abnormal feature point and the texture feature point;

[0205] S656: Generate tag information and set the final tag value M o The abnormality type and location are converted into descriptive information, and annotated on the 3D model and 2D image to generate label information including the spatial coordinates of the abnormal feature point o, the abnormality type, and the degree of abnormality; the label format is: label information = {coordinate: (x c ,y c , z c ), Type: Abnormal Type, Degree: M o}.

[0206] In this implementation, by executing the above steps S651 to S656, the specific position of the abnormal point on the small structural part can be accurately located according to the spatial coordinates of the abnormal feature point combined with the geometric features and surface texture features, and the corresponding marking information can be generated. The specific implementation of steps such as geometric distance calculation, perspective projection conversion, texture similarity calculation, and weighted average marking ensures the high quality and high precision of abnormal point positioning and marking, providing a reliable data basis for comprehensive quality assessment.

[0207] In an implementation manner of the embodiment of the present invention, the implementation process of the above S7 may include:

[0208] S71: Summarize the anomaly detection results, collect and organize the anomaly types and anomaly locations output by the classification model in S6, including the spatial coordinates, anomaly type, and anomaly degree of each anomaly feature point, to form a preliminary detection result data set;

[0209] S72: Comprehensively analyze the detected abnormal data, including calculating the data, distribution location and type ratio of abnormal feature points. Specifically, statistical analysis methods are used to evaluate the severity of various abnormalities and their distribution in the overall small structural parts of the aircraft. The statistical analysis formula is:

[0210]

[0211] Among them, P(type) is the severity of the corresponding type of abnormality, n5 is the number of abnormal feature points of the corresponding type, and n6 is the total number of all abnormal feature points;

[0212] S73: Develop comprehensive quality assessment standards based on industry standards and production requirements to determine the geometric tolerances, surface quality and allowable range of defects;

[0213] S74: Compare the severity of each abnormal point with the evaluation standard, and calculate the score of each abnormal feature point; wherein, for each abnormal type, the score of the abnormal feature point is calculated using the following scoring formula:

[0214]

[0215] Among them, q r The score of the rth quality assessment indicator is given by the maximum allowed value, which is the pre-set quality assessment standard;

[0216] S75: Score all outliers q r The weighted average of the parts is used to calculate the comprehensive quality assessment score of the aircraft small structural parts. The quality assessment standard formula used is:

[0217]

[0218] Among them, Q is the comprehensive quality assessment score, q r is the score of the rth quality evaluation indicator, w is the weight coefficient of the indicator, and m1 is the number of evaluation indicators;

[0219] S76: Quality assessment report generation: According to the comprehensive quality assessment standards, the test results are quantitatively scored and graded to generate a quality assessment report. The report content includes basic information of the test object, test method, test results, comprehensive assessment score, quality grade and specific abnormality description.

[0220] In this step, the format of the quality assessment report is as follows:

[0221] Test object: aircraft small structural parts number;

[0222] Detection method: Combination of 3D scanning and visual inspection;

[0223] Detection results: including detailed information of various abnormal points;

[0224] Comprehensive evaluation score: Q;

[0225] Quantity level: divided into different levels according to Q;

[0226] Abnormality description: Specific abnormality type, location and degree.

[0227] S77: Based on the quality assessment report, make optimization suggestions for the test results, including improving design, optimizing manufacturing processes and strengthening quality control.

[0228] In this step, the process of formulating optimization suggestions includes, for example:

[0229] S771, identify the main abnormality types and locations, and analyze the possible causes and impacts based on the main abnormal points in the test results;

[0230] S772, formulate targeted measures, such as adjusting design parameters, improving manufacturing processes, strengthening material testing, etc.;

[0231] S773, propose specific optimization suggestions and improvement plans, including implementation steps, expected results and precautions;

[0232] In this implementation, by executing the above steps S71 to S75, a comprehensive quality assessment can be performed based on the abnormality detection results of S6, a detailed quality assessment report can be generated, and optimization suggestions based on the detection results can be put forward. The specific implementation of steps such as abnormality detection result summary, comprehensive analysis, quality assessment standard formulation, quality assessment report generation and optimization suggestion formulation ensures the high quality and high precision of quality assessment and optimization suggestions, and provides a reliable basis for improving the manufacturing quality and reliability of small aircraft structures.

[0233] The visual inspection and quality assessment method for small aircraft structural parts provided by the embodiment of the present invention has the following beneficial effects:

[0234] On the one hand, by integrating 3D scanning and high-resolution 2D imaging technology, comprehensive and accurate inspection of small structural parts can be achieved. This multimodal fusion not only enhances the ability to identify surface and internal structural abnormalities of small structural parts, but also can accurately locate tiny defects and potential structural problems through detailed surface texture analysis. In addition, this method also significantly improves image quality through high dynamic range images and advanced image processing technology, ensuring the accuracy and reliability of inspection data.

[0235] On the other hand, by adopting a classification model based on machine learning, not only the efficiency and accuracy of anomaly detection are improved, but it can also automatically learn and adapt to new detection situations based on historical data. This intelligent analysis method reduces the need for manual operation, reduces detection costs, and improves the level of automation in the production process. The generation of comprehensive quality assessment reports and the proposal of optimization suggestions further help companies monitor and improve the production process, thereby improving product quality while ensuring aviation safety.

[0236] The present invention covers any substitution, modification, equivalent method and scheme made on the essence and scope of the present invention. In order to make the public have a thorough understanding of the present invention, specific details are described in detail in the following preferred embodiments of the present invention, but those skilled in the art can fully understand the present invention without the description of these details. In addition, in order to avoid unnecessary confusion about the essence of the present invention, well-known methods, processes, procedures, components and circuits are not described in detail.

[0237] Although the embodiments disclosed in the present invention are as above, the contents are only embodiments adopted to facilitate understanding of the present invention and are not intended to limit the present invention. Any technician in the field to which the present invention belongs can make any modifications and changes in the form and details of implementation without departing from the spirit and scope disclosed in the present invention, but the patent protection scope of the present invention shall still be subject to the scope defined in the attached claims.

Claims

1. A visual inspection and quality assessment method for small aircraft structural parts, characterized in that: include; S1: Use a 3D scanner to perform an all-round scan of the aircraft's small structural parts, obtain the 3D point cloud data of the aircraft's small structural outer surface, and pre-process the 3D point cloud data; S2: Use high-resolution cameras to collect 2D images of small aircraft structures from multiple angles; S3: constructing a three-dimensional model using the preprocessed three-dimensional point cloud data, and mapping the texture of the two-dimensional image obtained in S2 to the surface of the three-dimensional model to generate a three-dimensional model including surface texture information; S4: Use 3D feature extraction algorithm to extract geometric features and surface texture information from 3D models. The extracted content includes the shape, size, surface texture information and subtle defects of small structural parts. S5: Using multimodal data fusion, the geometric features and surface texture information extracted in S4 are fused and compared with standard samples to identify potential anomalies and defects; S6: Analyze the fused comparison data in S5 using the pre-trained classification model and output the anomaly detection results, including the anomaly type and location of the aircraft's small structural parts; S7: Perform a comprehensive quality assessment on the anomaly detection results of S6, generate a quality assessment report, and put forward optimization suggestions based on the detection results.

2. The visual inspection and quality assessment method for small aircraft structural parts according to claim 1, characterized in that: The three-dimensional scanner used for scanning is pre-positioned in the working area of ​​the small aircraft structure, and the S1 includes: S11: Performing initial calibration on the 3D scanner, including calibrating the position, angle and focal length of the scanner, so that the scanning area of ​​the calibrated 3D scanner covers the entire outer surface of the small structural parts of the aircraft; S12: Controlling the 3D scanner to perform multi-view scanning on the small aircraft structure according to a predetermined path and angle, and obtaining 3D point cloud data corresponding to each view; S13: splicing and fusing the three-dimensional point cloud data corresponding to each perspective obtained in S12, and using an iterative closest point algorithm ICP in S13 to align multiple groups of three-dimensional point cloud data to generate a three-dimensional point cloud data set; S14: performing denoising processing on the spliced ​​three-dimensional point cloud data set, and removing noise points in the three-dimensional point cloud data set using a statistical filtering algorithm, wherein the statistical filtering algorithm identifies and removes abnormal points by calculating neighborhood statistical characteristics of each point; S15: identifying and removing redundant points and isolated points in the three-dimensional point cloud data set by a density clustering algorithm; the density clustering algorithm detects and deletes redundant points with insufficient density by defining a density threshold and a minimum number of neighbors; S16: A global least squares fitting algorithm is used to globally align the denoised and de-redundant 3D point cloud dataset.

3. The visual inspection and quality assessment method for small aircraft structural parts according to claim 1, characterized in that: A high-resolution camera is pre-positioned in the working area of ​​the aircraft's small structural parts. The S2 includes: S21: performing an initial calibration on the high-resolution camera, including calibrating the position, angle, and focal length of the camera so that the camera covers all surface areas of the small structural component; S22: Design a multi-angle shooting path, determine specific shooting points of the camera at different positions and angles, and the shooting path includes positions at least 8 different angles; S23: Setting the image acquisition parameters of the camera, including exposure time, aperture size and sensitivity; S24: according to the planned shooting path, control the camera to perform multi-angle image acquisition of the small structural part of the aircraft at a predetermined position and angle, obtain all-round two-dimensional image data of the small structural part of the aircraft, and shoot at least three images at each angle at each position; S25: using an image distortion correction algorithm to correct the distortion in the collected multi-position and multi-angle images, and performing color correction and contrast enhancement; S26: Stitch and fuse the two-dimensional images collected from multiple positions and angles, and use an image stitching algorithm to seamlessly stitch the images into a complete image covering the entire surface of the small structural part. During the stitching process, the overlapping portion between the images is no less than 30%.

4. The visual inspection and quality assessment method for small aircraft structural parts according to claim 1, characterized in that: The S3 includes: S31: using the pre-processed 3D point cloud data obtained in S1, constructing a preliminary 3D model through a Delaunay triangulation algorithm, and generating a mesh model including the geometric shape of the small structural parts of the aircraft; S32: registering the two-dimensional image obtained in S2 with the preliminary three-dimensional model, including: using a SIFT algorithm to identify corresponding feature points in the two-dimensional image and the three-dimensional model, and aligning the feature points through a camera projection matrix; S33: performing texture mapping on the registered two-dimensional image, mapping the surface texture information in the two-dimensional image to the outer surface of the three-dimensional model; wherein the texture mapping adopts a perspective projection texture mapping algorithm, and projects the texture coordinates corresponding to each pixel of the two-dimensional image onto the surface of the three-dimensional model; S34: performing texture fusion on the two-dimensional images acquired at multiple positions and angles, and using a weighted average fusion algorithm to fuse the textures of the overlapping areas. During the fusion process, weighted average processing is performed on the pixels of each overlapping area. S35: Perform detail enhancement on the texture mapped and fused three-dimensional model, specifically using a high dynamic range enhancement algorithm to adjust the brightness and contrast of the texture.

5. The visual inspection and quality assessment method for small aircraft structural parts according to claim 1, characterized in that: The S4 includes: S41: extracting geometric shape features from the three-dimensional model, including: fitting each surface of the three-dimensional model using a surface fitting algorithm to obtain the geometric shape of the small aircraft structure, wherein the surface fitting algorithm minimizes the error between the fitting surface and the three-dimensional point cloud data by the least square method, and the calculation formula is: Among them, z i is a point in the point cloud data, f(x i ,y i ) is the fitting surface function, (x i ,y i ) is the coordinate of the point cloud data, n1 is the number of points; S42: Measuring the dimensions of the three-dimensional model, including: measuring the dimensions of each part of the small structural part of the aircraft using a distance calculation algorithm, wherein the distance calculation algorithm is used to calculate the Euclidean distance between two points in the three-dimensional model, and the calculation formula is: Where (x1, y1, z1) and (x2, y2, z2) are the coordinates of two points in the three-dimensional model, and d is the distance between the two points; S43: extracting surface texture information of the three-dimensional model, including: using a normal vector analysis algorithm to calculate the normal vector of each point of the three-dimensional model, where the normal vector is defined as a vector perpendicular to the surface, and the calculation formula is: Where N is the normal vector, v1 and v2 are two non-parallel vectors on the surface, × represents the vector cross product, and |·| represents the modulus of the vector; S44: Identify subtle defects in 3D models, including: using Gaussian curvature and mean curvature analysis algorithms to calculate the curvature of the 3D model surface, and identify subtle defects such as dents, protrusions and cracks on the surface of small aircraft structural parts.

6. The visual inspection and quality assessment method for small aircraft structural parts according to claim 5, characterized in that: The Gaussian curvature and mean curvature analysis algorithms in S44 are respectively: Among them, κ1 and κ2 are the principal curvatures, and are the partial derivatives of the surface in the x and y directions respectively, K is the Gaussian curvature, and H is the mean curvature.

7. The visual inspection and quality assessment method for small aircraft structural parts according to claim 1, characterized in that: The S5 includes: S51: Obtain geometric features and surface texture information of standard samples, including shape, size, and surface texture information of small aircraft structural parts, and use them as benchmark data; S52: Compare and analyze the geometric features extracted in S4 with the geometric features of the standard sample, and use a shape matching algorithm to evaluate the consistency of the geometric features by calculating the similarity between the two. The shape matching algorithm formula is: Among them, A and B represent the geometric feature point sets of small aircraft structures and standard samples respectively, n2 is the number of geometric feature points, ∥A a -B b ∥ is the geometric feature point A of the aircraft small structural part and the standard sample a and B b The Euclidean distance between S53: Compare and analyze the surface texture information extracted in S4 with the surface texture information of the standard sample, and use a texture similarity algorithm to evaluate the consistency of the surface texture information by calculating the similarity between the surface texture information. The texture similarity algorithm formula is: Wherein, T1 and T2 represent the texture images of the aircraft small structure and the standard sample respectively, m is the height of the texture image, n3 is the width of the texture image, |T1(v, c)-T2(v, c)| is the absolute difference between the corresponding pixels of the aircraft small structure and the standard sample; S54: Fusing the geometric features and the surface texture information, using a weighted fusion algorithm to combine the geometric feature similarity and the texture information similarity to form a comprehensive similarity score. The weighted fusion algorithm formula is: Fused Similarity=w g ·Geometry Similarity+w t ·Texture Similarity; Among them, w g and w t are the weight coefficients of geometric features and surface texture information respectively, Geometry Similarity is the geometric feature similarity, and Texture Similarity is the surface texture information similarity; S55: Based on a preset threshold, when the comprehensive similarity score is lower than the threshold, it is marked as a potential abnormality and defect area.

8. The visual inspection and quality assessment method for small aircraft structural parts according to claim 1, characterized in that: The S6 includes: S61: The pre-trained classification model is a classification model based on an extreme learning machine; S62: Use the geometric features and surface texture information in the historical detection data and the standard samples as training data, randomly generate input weights and biases through the extreme learning machine, and determine the output weights to complete the classification model training; S63: input the comprehensive similarity score after the fusion of multimodal data in S5 into the trained extreme learning machine classification model, analyze and process it, and generate a classification result; S64: According to the output result of the classification model, the category to which the input data belongs is determined, including normal, surface defect, and geometric deformation. The extreme learning machine calculates the classification probability of the input data and maps it to the predefined abnormal category, and outputs the final abnormal type and abnormal feature points; S65: Determine the position of the anomaly by analyzing the abnormal feature points output in the classification result, including: in the three-dimensional model, locate the specific position of the abnormal feature points on the small structural parts of the aircraft according to the spatial coordinates of the abnormal feature points, combined with geometric features and surface texture information, and generate corresponding marking information.

9. The visual inspection and quality assessment method for small aircraft structural parts according to claim 8, characterized in that: The S65 specifically includes: S651: Determine the spatial coordinates. The spatial coordinates (x i ,y i , z i ) to extract and record the specific positions of abnormal feature points in the three-dimensional model; S652: combining geometric features, locating the abnormal feature points according to the spatial coordinates of the abnormal feature points using geometric features, specifically including calculating the distance between the abnormal feature points and the surrounding geometric feature points; S653: Determine the coordinates of the two-dimensional image and map the spatial coordinates of the abnormal feature points to the texture coordinates of the two-dimensional image (u d , v d ), using the perspective projection formula for conversion, the expression is: Among them, (F x , F y ) is the focal length of the camera, (C x , C y ) is the principal point coordinate of the camera; S654: Combine surface texture information and use texture coordinates (u d , v d ) Extract the corresponding texture information in the two-dimensional image and calculate the similarity between the abnormal feature point o and the surrounding texture feature points. The calculation formula is: Among them, S oe is the similarity between the abnormal feature point o and the texture feature point e, T f (k) and T f (k) are the e-th texture values ​​of the abnormal feature point o and the texture feature point e, respectively, and n4 is the number of texture feature points; S655: marking abnormal feature points, including: combining geometric features and surface texture information, and determining the final marking value of abnormal feature point i by weighted averaging the geometric distance and texture similarity of the abnormal feature points. The calculation formula is: M o =W h ·life k (d op )+W t ·life k (S oe ): Among them, M o is the final label value of the abnormal feature point o, W h and W t are the weight coefficients of geometric features and surface texture information, min k (d op ) is the minimum distance between the abnormal feature point o and the geometric feature point p, min k (S oe ) is the minimum similarity between the abnormal feature point and the texture feature point; S656: Generate tag information and set the final tag value M o The abnormality type and location are converted into descriptive information, and annotated on the three-dimensional model and two-dimensional image to generate label information including the spatial coordinates of the abnormal feature point o, the abnormality type, and the degree of abnormality; the label format is: label information = {coordinate: (x c ,y c , z c ), Type: Abnormal Type, Degree: M o }.

10. The visual inspection and quality assessment method for small aircraft structural parts according to claim 1, characterized in that: The S7 includes: S71: Summarize the anomaly detection results, collect and organize the anomaly types and anomaly locations output by the classification model in S6, including the spatial coordinates, anomaly type, and anomaly degree of each anomaly feature point, to form a preliminary detection result data set; S72: Comprehensively analyze the detected abnormal data, including calculating the data, distribution location and type ratio of abnormal feature points. Specifically, statistical analysis methods are used to evaluate the severity of various abnormalities and their distribution in the overall small structural parts of the aircraft. The statistical analysis formula is: Among them, P(type) is the severity of the corresponding type of abnormality, n5 is the number of abnormal feature points of the corresponding type, and n6 is the total number of all abnormal feature points; S73: Develop comprehensive quality assessment standards based on industry standards and production requirements to determine the geometric tolerances, surface quality and allowable range of defects; S74: Compare the severity of each abnormal point with the evaluation standard, and calculate the score of each abnormal feature point; wherein, for each abnormal type, the score of the abnormal feature point is calculated using the following scoring formula: Among them, q r The score of the rth quality assessment indicator is given by the maximum allowed value, which is the pre-set quality assessment standard; S75: Score all outliers q r The weighted average of the parts is used to calculate the comprehensive quality assessment score of the aircraft small structural parts. The quality assessment standard formula used is: Among them, Q is the comprehensive quality assessment score, q r is the score of the rth quality evaluation indicator, w is the weight coefficient of the indicator, and m1 is the number of evaluation indicators; S76: Quality assessment report generation: According to the comprehensive quality assessment standards, the test results are quantitatively scored and graded to generate a quality assessment report. The report content includes basic information of the test object, test method, test results, comprehensive assessment score, quality level and specific abnormality description; S77: Based on the quality assessment report, make optimization suggestions for the test results, including improving design, optimizing manufacturing processes and strengthening quality control.

Images