Civil aircraft flight test method and system based on model design

Abstract

The invention relates to a civil aircraft flight test method and system based on model design, and relates to the field of image processing, and the method comprises the steps: carrying out the contour feature extraction of a plurality of aircraft images, and obtaining a plurality of aircraft contours; performing point cloud fitting on the plurality of aircraft point cloud data sets according to the plurality of aircraft contours to obtain an optimal point cloud fitting result set, and constructing an aircraft three-dimensional model based on the optimal point cloud fitting result set; and in a predetermined flight test platform, testing the aircraft three-dimensional model according to a predetermined test scheme, and outputting a plurality of performance test data. According to the invention, the technical problems of insufficient modeling precision, large performance test limitation and incomplete test data in a traditional civil aircraft flight test method can be solved; through accurate three-dimensional modeling and comprehensive performance evaluation, the accuracy and scientificity of aircraft design can be remarkably improved, and the test period is shortened, so that the overall performance of the civil aircraft is improved.

Description

Technical Field

[0001] The present invention relates to the field of image processing, and in particular to a civil aircraft flight test method and system based on model design. Background Art

[0002] With the continuous development of civil aircraft technology, the design and optimization of aircraft have become increasingly complex and sophisticated. Modern civil aircraft not only need to meet increasingly stringent safety standards, but also have to meet multiple requirements such as higher fuel efficiency, maneuverability, comfort and environmental protection. In this context, traditional aircraft design and testing methods, although they have been widely used, also face many challenges.

[0003] At present, the traditional civil aircraft flight test method relies on physical prototypes and field tests to verify the performance of the aircraft. Although this method can provide certain test data, it also faces many challenges, especially in the research and development process of civil aircraft. Problems such as insufficient modeling accuracy, large limitations in performance testing, and incomplete test data have seriously affected the scientific nature of the design and the safety and reliability of the aircraft. Summary of the invention

[0004] The present invention aims to solve the technical problems of insufficient modeling accuracy, large performance test limitations and incomplete test data in traditional civil aircraft flight test methods, and provides a civil aircraft flight test method and system based on model design to solve the problems.

[0005] The technical solution of the present invention to solve the above technical problems is as follows: In a first aspect, the present invention provides a civil aircraft flight test method based on model design, comprising: performing image acquisition and point cloud acquisition on a civil aircraft to be tested to obtain a plurality of aircraft images and a plurality of aircraft point cloud data sets; performing contour feature extraction on the plurality of aircraft images to obtain a plurality of aircraft contours; performing point cloud fitting on the plurality of aircraft point cloud data sets according to the plurality of aircraft contours to obtain an optimal point cloud fitting result set, and constructing an aircraft three-dimensional model based on the optimal point cloud fitting result set; testing the aircraft three-dimensional model in a predetermined flight test platform according to a predetermined test plan, and outputting a plurality of performance test data.

[0006] Preferably, the civil aircraft flight test method based on model design also includes: acquiring multiple acquisition angles; and performing image acquisition and point cloud acquisition on the civil aircraft to be tested in sequence according to the multiple acquisition angles to obtain multiple aircraft images and multiple aircraft point cloud data sets, wherein the acquisition angles of each set of aircraft images and aircraft point cloud data sets are the same.

[0007] Preferably, the model-design-based civil aircraft flight test method further includes: performing image denoising and image enhancement on the multiple aircraft images to obtain multiple standard aircraft images; pre-training a contour feature extraction plug-in; and inputting the multiple standard aircraft images into the contour feature extraction plug-in to perform contour feature extraction, and outputting multiple aircraft contours.

[0008] Preferably, the civil aircraft flight test method based on model design also includes: constructing a contour feature extraction plug-in based on a convolutional neural network; collecting sample aircraft images to construct a sample aircraft image set based on historical modeling data of civil aircraft of the same type as the civil aircraft to be tested, and annotating the aircraft contours in different sample aircraft images to obtain a sample aircraft contour set; using sample aircraft images as input and sample aircraft contours as supervision, and using the sample aircraft image set and the sample aircraft contour set to perform supervised training on the contour feature extraction plug-in until the loss function converges, thereby obtaining a trained contour feature extraction plug-in.

[0009] Preferably, the model-design-based civil aircraft flight test method further includes: randomly selecting a first aircraft profile, and obtaining a first aircraft point cloud data set of the first aircraft profile at the same acquisition angle; performing point cloud fitting in the first aircraft point cloud data set according to the first aircraft profile to obtain a first point cloud fitting result, and calculating a first degree of fit, wherein the first degree of fit is a ratio of the number of point clouds in the first point cloud fitting result to the total number of point clouds in the first aircraft point cloud data set; performing point cloud fitting again in the first aircraft point cloud data set according to the first aircraft profile to obtain a second point cloud fitting result, and calculating a second degree of fit; iteratively performing point cloud fitting until a predetermined number of fitting times is reached, selecting a point cloud fitting result with the highest degree of fit as the optimal point cloud fitting result of the first aircraft profile, and sequentially analyzing to obtain multiple optimal point cloud fitting results of the multiple aircraft profiles, and constructing an optimal point cloud fitting result set according to the multiple optimal point cloud fitting results.

[0010] Preferably, the civil aircraft flight test method based on model design also includes: performing multi-perspective fitting on multiple optimal point cloud fitting results in the optimal point cloud fitting result set according to the multiple acquisition angles to obtain an initial aircraft three-dimensional model; and performing local repair on the initial aircraft three-dimensional model through an interpolation algorithm to obtain the aircraft three-dimensional model.

[0011] Preferably, the civil aircraft flight test method based on model design also includes: obtaining a predetermined test plan, wherein the predetermined test plan includes an aerodynamic performance test plan, a structural strength test plan, a maneuverability test plan and a power efficiency test plan; based on the aerodynamic performance test plan, the structural strength test plan, the maneuverability test plan and the power efficiency test plan, respectively testing the aircraft three-dimensional model and outputting the multiple performance test data.

[0012] In a second aspect, the present invention provides a civil aircraft flight test system based on model design, comprising: a data acquisition module, used to perform image acquisition and point cloud acquisition on the civil aircraft to be tested, so as to obtain multiple aircraft images and multiple aircraft point cloud data sets; a contour feature extraction module, used to perform contour feature extraction on the multiple aircraft images, so as to obtain multiple aircraft contours; a point cloud fitting module, used to perform point cloud fitting on the multiple aircraft point cloud data sets respectively according to the multiple aircraft contours, so as to obtain an optimal point cloud fitting result set, and to construct an aircraft three-dimensional model based on the optimal point cloud fitting result set; an aircraft testing module, used to test the aircraft three-dimensional model in a predetermined flight test platform according to a predetermined test plan, and output multiple performance test data.

[0013] The beneficial effects of the present invention are: by performing image acquisition and point cloud acquisition on the civil aircraft to be tested, multiple aircraft images and multiple aircraft point cloud data sets are obtained; contour feature extraction is performed on the multiple aircraft images to obtain multiple aircraft contours; point cloud fitting is performed on the multiple aircraft point cloud data sets according to the multiple aircraft contours to obtain an optimal point cloud fitting result set, and an aircraft three-dimensional model is constructed based on the optimal point cloud fitting result set; the aircraft three-dimensional model is tested in a predetermined flight test platform according to a predetermined test plan, and multiple performance test data are output. In other words, through precise three-dimensional modeling and comprehensive performance evaluation, the accuracy and scientificity of aircraft design can be significantly improved, the test cycle can be shortened, and the overall performance of civil aircraft can be improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 A schematic diagram of the flow of the civil aircraft flight test method based on model design provided by the present invention; Figure 2 A schematic structural diagram of a civil aircraft flight test system based on model design provided by the present invention.

[0015] In the accompanying drawings, the components represented by the reference numerals are described as follows: Data acquisition module 11, contour feature extraction module 12, point cloud fitting module 13, aircraft testing module 14. DETAILED DESCRIPTION

[0016] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0017] In the description of the present invention, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0018] In the description of the present invention, the term "for example" is used to mean "used as an example, illustration or explanation". Any embodiment described as "for example" in the present invention is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is given to enable any technician in the field to implement and use the present invention. In the following description, details are listed for the purpose of explanation. It should be understood that a person of ordinary skill in the art can recognize that the present invention can be implemented without using these specific details. In other examples, well-known structures and processes will not be elaborated in detail to avoid obscuring the description of the present invention with unnecessary details. Therefore, the present invention is not intended to be limited to the embodiments shown, but is consistent with the widest scope consistent with the principles and features disclosed in the present invention.

[0019] Embodiment 1, as Figure 1 As shown, the embodiment of the present invention provides a civil aircraft flight test method based on model design, which specifically includes the following steps: S100: performing image acquisition and point cloud acquisition on the civil aircraft to be tested, and obtaining a plurality of aircraft images and a plurality of aircraft point cloud data sets; S200: performing contour feature extraction on the plurality of aircraft images, and obtaining a plurality of aircraft contours; S300: performing point cloud fitting on the plurality of aircraft point cloud data sets respectively according to the plurality of aircraft contours, and obtaining an optimal point cloud fitting result set, and constructing an aircraft three-dimensional model based on the optimal point cloud fitting result set; S400: testing the aircraft three-dimensional model in a predetermined flight test platform according to a predetermined test plan, and outputting a plurality of performance test data.

[0020] Specifically, the present invention proposes a civil aircraft flight test method based on the combination of image acquisition, point cloud acquisition, contour feature extraction, point cloud fitting and a flight test platform. Through this method, a virtual simulation platform can be used to perform multi-scenario and multi-stage performance tests on an aircraft in a relatively short period of time, so as to improve the accuracy of aircraft design, the comprehensiveness of testing and the reliability of data.

[0021] First, perform image acquisition and point cloud acquisition on the civil aircraft to be tested, and obtain multiple aircraft image data sets and multiple aircraft point cloud data sets respectively. For example, use a high-resolution camera or video equipment to shoot the aircraft from multiple perspectives to ensure that the image data can cover the main appearance features of the aircraft; obtain the point cloud data of the aircraft through laser radar (LiDAR), stereo vision or other high-precision measurement equipment. The point cloud data set should include the external surface, internal and external structures of the aircraft, and geometric features related to aerodynamics. Then, perform contour feature extraction on the collected aircraft image data to obtain multiple aircraft contours, that is, apply image processing technology to extract the shape contour of the aircraft from the image. These contour features will serve as a guide for subsequent point cloud fitting to ensure that the point cloud fitting is highly consistent with the true shape of the aircraft.

[0022] Then, based on the extracted contour features, point cloud fitting is performed on multiple aircraft point cloud data sets. By matching the contour features, the optimal point cloud fitting result set is obtained to ensure the accuracy and precision of the fitting; according to the optimal point cloud fitting result set, surface reconstruction technology (such as Poisson reconstruction) is used to convert the point cloud data into a smooth three-dimensional surface to form a complete aircraft three-dimensional model, which will be used as a virtual test object in the subsequent flight test platform. Then, in the predetermined flight test platform, the aircraft three-dimensional model is tested according to the predetermined test plan, and multiple performance test data are output, including aerodynamic performance data (lift, drag, aerodynamic efficiency, etc.), structural strength data (stress distribution, fatigue life, etc.), control performance data (control response, overload characteristics, etc.), and power system data (thrust output, fuel consumption rate, etc.).

[0023] Furthermore, the present invention further comprises the following steps: Acquire multiple acquisition angles; and perform image acquisition and point cloud acquisition on the civil aircraft to be tested in sequence according to the multiple acquisition angles to obtain multiple aircraft images and multiple aircraft point cloud data sets, wherein the acquisition angles of each set of aircraft images and aircraft point cloud data sets are the same.

[0024] Specifically, multiple acquisition angles are obtained. The selection of acquisition angles should cover all directions of the aircraft, including multiple perspectives such as the front, side, top, bottom, and front oblique angles. These angles should cover the key external features and details of the aircraft as much as possible. Then, according to the multiple acquisition angles, the civil aircraft to be tested is imaged and point cloud acquired in turn, that is, the appearance of the aircraft is photographed by a high-resolution camera or other visual acquisition equipment according to different acquisition angles. The image at each acquisition angle will contain the complete shape of the aircraft and some details, such as important structures such as wings, tail fins, and engines; use laser radar (LiDAR) or other precision measurement equipment to collect point cloud data from each preset acquisition angle. These point cloud data will accurately capture the geometric information of the aircraft surface and provide detailed data including surface texture, contour, etc.; among them, the acquisition angles of each set of images and point cloud data must be consistent, that is, image acquisition and point cloud data acquisition must be performed at the same perspective to ensure that the image and point cloud data can correspond one to one, so as to avoid data mismatch problems caused by angle differences, thereby improving the accuracy of point cloud fitting and three-dimensional modeling.

[0025] By collecting data from multiple angles, the geometric shape of the aircraft can be captured in all directions, reducing the modeling deviation that may be caused by traditional single-angle collection. Data from each angle can supplement and improve details that cannot be obtained from other angles, thereby providing a more complete and accurate 3D modeling foundation.

[0026] Furthermore, the present invention further comprises the following steps: Perform image denoising and image enhancement on the multiple aircraft images to obtain multiple standard aircraft images; pre-train a contour feature extraction plug-in; input the multiple standard aircraft images into the contour feature extraction plug-in respectively to extract contour features, and output multiple aircraft contours.

[0027] Specifically, first, the multiple aircraft images are subjected to image denoising and image enhancement, and the collected aircraft images are processed using image denoising algorithms (such as median filtering, Gaussian filtering, convolutional neural network denoising, etc.) to remove interference caused by uneven lighting and equipment noise during the acquisition process, ensuring image clarity and detail integrity; the denoised images are enhanced using contrast enhancement, sharpening, edge enhancement and other technologies to make the contour features of the aircraft more obvious, facilitating subsequent feature extraction. Image enhancement can highlight the detailed features of the aircraft and improve the accuracy of contour extraction; after image denoising and enhancement processing, multiple standard aircraft images are obtained, which have clear edges and significant features, and can provide higher quality data for subsequent contour feature extraction.

[0028] Next, a pre-trained contour feature extraction plug-in is used. This plug-in can use deep learning models (such as convolutional neural networks (CNN)) or traditional computer vision algorithms (such as Canny edge detection, Sobel operator, etc.) to extract contour features in images. Then, multiple standard aircraft images are sequentially input into the pre-trained contour feature extraction plug-in for processing. The plug-in will automatically detect and extract aircraft contour features in the image based on the pre-learned feature patterns, and output multiple aircraft contours. That is, for each input standard aircraft image, the extraction plug-in will output its corresponding contour features. These contour features can clearly reflect the external form of the aircraft and provide accurate contour information for subsequent point cloud fitting and 3D modeling.

[0029] This solution makes the aircraft image clearer through image denoising and image enhancement, which facilitates the subsequent contour feature extraction; at the same time, the pre-trained contour feature extraction plug-in can automatically and efficiently extract the aircraft's contour information, providing high-quality input data for subsequent point cloud fitting and three-dimensional modeling. Through these optimization measures, not only the accuracy and efficiency of modeling are improved, but also more reliable technical support is provided for the entire flight test process.

[0030] Furthermore, the present invention further comprises the following steps: A contour feature extraction plug-in is constructed based on a convolutional neural network; sample aircraft images are collected based on historical modeling data of civil aircraft of the same type as the civil aircraft to be tested to construct a sample aircraft image set, and aircraft contours in different sample aircraft images are labeled to obtain a sample aircraft contour set; sample aircraft images are used as input and sample aircraft contours are used as supervision, and the sample aircraft image set and the sample aircraft contour set are used to perform supervised training on the contour feature extraction plug-in until the loss function converges, thereby obtaining a trained contour feature extraction plug-in.

[0031] Specifically, first, multiple aircraft image samples are collected from historical modeling data of civil aircraft of the same type as the civil aircraft to be tested. The images should contain the appearance of the aircraft at different angles and under different conditions to ensure the diversity and comprehensiveness of the image samples. These images will become the input basis of the training data. Then, the aircraft contours in the collected sample aircraft images are manually annotated. The annotation content includes the external contours of the aircraft, important feature points, etc. The specific position and shape of the aircraft contour are marked in each image, and a sample aircraft contour set is obtained.

[0032] Next, a contour feature extraction plug-in is constructed based on a convolutional neural network (CNN). CNN is a classic deep learning method for processing image data with powerful feature extraction capabilities. The network structure includes multiple convolutional layers, pooling layers, and fully connected layers, which are used to automatically extract the contour features of the aircraft from the input image. Then, supervised training is performed using a set of labeled sample aircraft images and a corresponding set of sample aircraft contours. The sample images are used as inputs and the sample contours are used as target outputs. Through training, the network can automatically extract accurate contour features from the image, select a suitable loss function (such as cross entropy loss, mean square error loss, etc.) to measure the difference between the predicted contour and the labeled contour, and continuously optimize the network parameters until the loss function converges, indicating that the network has learned the optimal feature extraction method. When the loss function in the training process reaches a convergence state, the trained contour feature extraction plug-in can be applied to new aircraft images for contour extraction. The plug-in can automatically identify the contour of the aircraft and output its features, which is suitable for subsequent 3D modeling and performance testing.

[0033] Through the convolutional neural network-based contour feature extraction plug-in combined with a large number of labeled sample image sets for supervised training, the automatic and accurate extraction of aircraft contours is achieved. Through optimization during the training process, the plug-in can adapt to aircraft images under a variety of different angles and lighting conditions, significantly improving the accuracy and efficiency of modeling, while providing accurate contour feature data for subsequent flight tests.

[0034] Furthermore, the present invention further comprises the following steps: A first aircraft contour is randomly selected, and a first aircraft point cloud data set of the first aircraft contour at the same acquisition angle is obtained; point cloud fitting is performed in the first aircraft point cloud data set according to the first aircraft contour to obtain a first point cloud fitting result, and a first degree of fit is calculated, wherein the first degree of fit is the ratio of the number of point clouds in the first point cloud fitting result to the total number of point clouds in the first aircraft point cloud data set; point cloud fitting is performed again in the first aircraft point cloud data set according to the first aircraft contour to obtain a second point cloud fitting result, and a second degree of fit is calculated; point cloud fitting is iteratively performed until a predetermined number of fitting times is reached, a point cloud fitting result with the highest degree of fit is selected as the optimal point cloud fitting result of the first aircraft contour, and multiple optimal point cloud fitting results of the multiple aircraft contours are analyzed in turn to obtain, and an optimal point cloud fitting result set is constructed according to the multiple optimal point cloud fitting results.

[0035] Specifically, any aircraft outline is randomly selected from multiple aircraft outlines as a first aircraft outline, and a first aircraft point cloud data set of the first aircraft outline at the same acquisition angle is obtained; then, point cloud fitting is performed in the first aircraft point cloud data set according to the first aircraft outline. This process randomly selects points from the point cloud data set and fits the outline, and fits these points into a geometric model through a certain algorithm to obtain a first point cloud fitting result, and a first degree of fit is calculated, wherein the first degree of fit is the ratio of the number of point clouds in the first point cloud fitting result to the total number of point clouds in the first aircraft point cloud data set, that is, the degree of fit refers to the ratio of the number of point clouds matching the outline in the fitted point cloud result to the total number of point clouds in the first aircraft point cloud data set, and a higher degree of fit indicates a better fitting effect.

[0036] After the first fitting, the point cloud fitting is performed again according to the position of the first aircraft contour in the point cloud data set to obtain the second point cloud fitting result, and the second fitting degree is calculated; the above steps are repeated to perform multiple point cloud fittings, and the fitting degree of each fitting is calculated. The fitting result is updated after each fitting, and the change in the fitting degree is recorded. When the predetermined maximum number of fittings is reached, the fitting process is stopped. Then, the point cloud fitting result with the highest fitting degree is selected as the optimal point cloud fitting result of the first aircraft contour, and multiple optimal point cloud fitting results of the multiple aircraft contours are analyzed in turn, and an optimal point cloud fitting result set is constructed based on the multiple optimal point cloud fitting results.

[0037] Through multiple point cloud fitting and fitting degree calculations, the point cloud fitting results can be automatically optimized to ensure the most accurate aircraft surface data. This method not only improves the accuracy of aircraft 3D modeling, but also can adapt to and obtain the best modeling data under different acquisition conditions, further optimizing subsequent flight tests and performance evaluations.

[0038] Furthermore, the present invention further comprises the following steps: Perform multi-view fitting on multiple optimal point cloud fitting results in the optimal point cloud fitting result set according to the multiple acquisition angles to obtain an initial aircraft three-dimensional model; and perform local repair on the initial aircraft three-dimensional model through an interpolation algorithm to obtain the aircraft three-dimensional model.

[0039] Specifically, according to the multiple acquisition angles, the multiple optimal point cloud fitting results in the optimal point cloud fitting result set are multi-view fitting, that is, the data at different angles are combined, and the best point cloud fusion method is calculated by the algorithm to obtain a preliminary aircraft three-dimensional model; finally, an initial aircraft three-dimensional model is obtained, which is based on the integrated point cloud data fitted at different acquisition angles, and preliminarily shows the overall shape of the aircraft. Then, the initial aircraft three-dimensional model is partially repaired by the interpolation algorithm. The interpolation algorithm can fill the missing data points in the missing or incomplete areas according to the existing point cloud data, so as to improve the details of the model. The repaired part is usually the hole or uneven area caused by the angle limitation or occlusion during the point cloud data acquisition, and the aircraft three-dimensional model is obtained. Through multi-view point cloud fitting and interpolation algorithm local repair, the three-dimensional model of the aircraft is accurately constructed. The model not only accurately restores the shape of the aircraft, but also fills the missing data that may appear in the acquisition process through interpolation repair, making the entire three-dimensional model more complete and refined. This provides accurate and reliable data support for subsequent flight tests, performance evaluation and simulation analysis.

[0040] Furthermore, the present invention further comprises the following steps: Obtain a predetermined test plan, wherein the predetermined test plan includes an aerodynamic performance test plan, a structural strength test plan, a maneuverability test plan, and a power efficiency test plan; based on the aerodynamic performance test plan, the structural strength test plan, the maneuverability test plan, and the power efficiency test plan, respectively test the three-dimensional model of the aircraft and output the multiple performance test data.

[0041] Specifically, a predetermined test plan is obtained, wherein the predetermined test plan includes an aerodynamic performance test plan, a structural strength test plan, a maneuverability test plan and a power efficiency test plan. The aerodynamic performance test plan mainly evaluates the aerodynamic characteristics of the aircraft under different flight conditions, such as lift, drag, aerodynamic stability, aerodynamic shape, etc. It ensures that the aircraft has optimal aerodynamic efficiency and stability during flight and adapts to various flight environments; the structural strength test plan is used to evaluate the compression, tension, bending and shear resistance of various components of the aircraft under different loads. This test ensures that the structure of the aircraft is not damaged under high load to ensure flight safety; the maneuverability test plan evaluates the flight control performance of the aircraft by testing the response of the aircraft under different control inputs, including response time, sensitivity, controllability and flight stability, etc., to ensure that the pilot can efficiently and safely control the aircraft; the power efficiency test plan is used to evaluate factors such as the aircraft's engine performance, fuel consumption, thrust / fuel ratio, etc., to ensure that the aircraft has an efficient power system, reduce operating costs and improve flight performance.

[0042] Then, based on the aerodynamic performance test scheme, structural strength test scheme, maneuverability test scheme and power efficiency test scheme, the three-dimensional model of the aircraft is tested respectively, and the multiple performance test data are output. Each test stage generates corresponding performance test data, including aerodynamic performance data, structural strength data, maneuverability data and power efficiency data, which will provide a scientific basis for the design and optimization of the aircraft and ensure that the various indicators of the aircraft meet the airworthiness requirements.

[0043] The civil aircraft flight method based on model design provided by the embodiment of the present invention has at least the following technical effects: By performing image acquisition and point cloud acquisition on the civil aircraft to be tested, multiple aircraft images and multiple aircraft point cloud data sets are obtained; contour feature extraction is performed on the multiple aircraft images to obtain multiple aircraft contours; point cloud fitting is performed on the multiple aircraft point cloud data sets according to the multiple aircraft contours to obtain an optimal point cloud fitting result set, and an aircraft three-dimensional model is constructed based on the optimal point cloud fitting result set; the aircraft three-dimensional model is tested in a predetermined flight test platform according to a predetermined test plan, and multiple performance test data are output. In other words, through precise three-dimensional modeling and comprehensive performance evaluation, the accuracy and scientificity of aircraft design can be significantly improved, the test cycle can be shortened, and the overall performance of civil aircraft can be improved.

[0044] Embodiment 2, as Figure 2 As shown, based on the same inventive concept as the civil aircraft flight test method based on model design provided in Example 1, an embodiment of the present invention also provides a civil aircraft flight test system based on model design, including: a data acquisition module 11, used to perform image acquisition and point cloud acquisition on the civil aircraft to be tested, and obtain multiple aircraft images and multiple aircraft point cloud data sets; a contour feature extraction module 12, used to extract contour features on the multiple aircraft images, and obtain multiple aircraft contours; a point cloud fitting module 13, used to perform point cloud fitting on the multiple aircraft point cloud data sets according to the multiple aircraft contours, respectively, to obtain an optimal point cloud fitting result set, and to construct an aircraft three-dimensional model based on the optimal point cloud fitting result set; an aircraft testing module 14, used to test the aircraft three-dimensional model according to a predetermined test plan in a predetermined flight test platform, and output multiple performance test data.

[0045] Furthermore, the model-designed civil aircraft flight test system is also used to: obtain multiple acquisition angles; and perform image acquisition and point cloud acquisition on the civil aircraft to be tested in sequence according to the multiple acquisition angles to obtain multiple aircraft images and multiple aircraft point cloud data sets, wherein the acquisition angles of each set of aircraft images and aircraft point cloud data sets are the same.

[0046] Furthermore, the model-designed civil aircraft flight test system is also used to: perform image denoising and image enhancement on the multiple aircraft images to obtain multiple standard aircraft images; pre-train a contour feature extraction plug-in; input the multiple standard aircraft images into the contour feature extraction plug-in respectively to perform contour feature extraction, and output multiple aircraft contours.

[0047] Furthermore, the model-designed civil aircraft flight test system is also used to: construct a contour feature extraction plug-in based on a convolutional neural network; collect sample aircraft images to construct a sample aircraft image set based on historical modeling data of civil aircraft of the same type as the civil aircraft to be tested, and annotate the aircraft contours in different sample aircraft images to obtain a sample aircraft contour set; use sample aircraft images as input and sample aircraft contours as supervision to perform supervised training on the contour feature extraction plug-in using the sample aircraft image set and the sample aircraft contour set until the loss function converges, thereby obtaining a trained contour feature extraction plug-in.

[0048] Furthermore, the model-designed civil aircraft flight test system is also used to: randomly select a first aircraft contour, and obtain a first aircraft point cloud data set of the first aircraft contour at the same acquisition angle; perform point cloud fitting in the first aircraft point cloud data set according to the first aircraft contour to obtain a first point cloud fitting result, and calculate a first degree of fit, wherein the first degree of fit is a ratio of the number of point clouds in the first point cloud fitting result to the total number of point clouds in the first aircraft point cloud data set; perform point cloud fitting again in the first aircraft point cloud data set according to the first aircraft contour to obtain a second point cloud fitting result, and calculate a second degree of fit; iteratively perform point cloud fitting until a predetermined number of fitting times is reached, select the point cloud fitting result with the highest degree of fit as the optimal point cloud fitting result of the first aircraft contour, and sequentially analyze to obtain multiple optimal point cloud fitting results of the multiple aircraft contours, and construct an optimal point cloud fitting result set according to the multiple optimal point cloud fitting results.

[0049] Furthermore, the model-designed civil aircraft flight test system is also used to: perform multi-perspective fitting on multiple optimal point cloud fitting results in the optimal point cloud fitting result set according to the multiple acquisition angles to obtain an initial aircraft three-dimensional model; and perform local repair on the initial aircraft three-dimensional model through an interpolation algorithm to obtain the aircraft three-dimensional model.

[0050] Furthermore, the civil aircraft flight test system based on model design is also used to: obtain a predetermined test plan, wherein the predetermined test plan includes an aerodynamic performance test plan, a structural strength test plan, a maneuverability test plan and a power efficiency test plan; based on the aerodynamic performance test plan, the structural strength test plan, the maneuverability test plan and the power efficiency test plan, respectively test the aircraft three-dimensional model and output the multiple performance test data.

[0051] Although preferred embodiments of the present invention have been described, additional changes and modifications may occur to these embodiments once those skilled in the art understand the basic inventive concepts.

[0052] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention belong to the scope of the present invention and its equivalent technologies, the present invention is also intended to include these changes and variations.

Claims

1. A civil aircraft flight test method based on model design, characterized in that: Methods include: Perform image acquisition and point cloud acquisition on the civil aircraft to be tested, and obtain multiple aircraft images and multiple aircraft point cloud data sets; Extracting contour features from the multiple aircraft images to obtain multiple aircraft contours; performing point cloud fitting on the plurality of aircraft point cloud data sets respectively according to the plurality of aircraft contours to obtain an optimal point cloud fitting result set, and constructing an aircraft three-dimensional model based on the optimal point cloud fitting result set; The three-dimensional model of the aircraft is tested in a predetermined flight test platform according to a predetermined test plan, and a plurality of performance test data are output.

2. The civil aircraft flight test method based on model design according to claim 1, characterized in that: Perform image acquisition and point cloud acquisition on the civil aircraft to be tested, and obtain multiple aircraft images and multiple aircraft point cloud data sets, including: Get multiple acquisition angles; According to the multiple acquisition angles, image acquisition and point cloud acquisition are performed on the civil aircraft to be tested in sequence to obtain multiple aircraft images and multiple aircraft point cloud data sets, wherein the acquisition angle of each set of aircraft images and aircraft point cloud data sets is the same.

3. The civil aircraft flight test method based on model design according to claim 2, characterized in that: Extracting contour features from the plurality of aircraft images to obtain a plurality of aircraft contours includes: Performing image denoising and image enhancement on the multiple aircraft images to obtain multiple standard aircraft images; Pre-trained contour feature extraction plugin; The multiple standard aircraft images are respectively input into the contour feature extraction plug-in to extract contour features, and multiple aircraft contours are output.

4. The civil aircraft flight test method based on model design according to claim 3, characterized in that: Pre-trained contour feature extraction plugin, including: Build a contour feature extraction plug-in based on convolutional neural network; According to historical modeling data of civil aircraft of the same type as the civil aircraft to be tested, sample aircraft images are collected to construct a sample aircraft image set, and aircraft contours in different sample aircraft images are annotated to obtain a sample aircraft contour set; Taking sample aircraft images as input and sample aircraft contours as supervision, the contour feature extraction plug-in is supervised and trained using the sample aircraft image set and the sample aircraft contour set until the loss function converges, thereby obtaining a trained contour feature extraction plug-in.

5. The civil aircraft flight test method based on model design according to claim 2, characterized in that: Performing point cloud fitting on the multiple aircraft point cloud data sets according to the multiple aircraft contours to obtain an optimal point cloud fitting result set includes: Randomly selecting a first aircraft outline, and obtaining a first aircraft point cloud data set of the first aircraft outline at the same acquisition angle; performing point cloud fitting in the first aircraft point cloud dataset according to the first aircraft outline to obtain a first point cloud fitting result, and calculating a first fitting degree, wherein the first fitting degree is a ratio of the number of point clouds in the first point cloud fitting result to the total number of point clouds in the first aircraft point cloud dataset; performing point cloud fitting again in the first aircraft point cloud data set according to the first aircraft outline to obtain a second point cloud fitting result, and calculating a second fitting degree; The point cloud fitting is iterated until a predetermined number of fitting times is reached, and the point cloud fitting result with the highest fitting degree is selected as the optimal point cloud fitting result of the first aircraft contour, and multiple optimal point cloud fitting results of the multiple aircraft contours are analyzed in turn, and an optimal point cloud fitting result set is constructed according to the multiple optimal point cloud fitting results.

6. The civil aircraft flight test method based on model design according to claim 2, characterized in that: Constructing an aircraft three-dimensional model based on the optimal point cloud fitting result set includes: Performing multi-view fitting on a plurality of optimal point cloud fitting results in the optimal point cloud fitting result set according to the plurality of acquisition angles to obtain an initial aircraft three-dimensional model; The initial three-dimensional aircraft model is partially repaired by an interpolation algorithm to obtain the three-dimensional aircraft model.

7. The civil aircraft flight test method based on model design according to claim 1, characterized in that: The three-dimensional model of the aircraft is tested according to a predetermined test plan, and a plurality of performance test data are output, including: Obtaining a predetermined test plan, wherein the predetermined test plan includes an aerodynamic performance test plan, a structural strength test plan, a maneuverability test plan, and a power efficiency test plan; Based on the aerodynamic performance test scheme, the structural strength test scheme, the maneuverability test scheme and the power efficiency test scheme, the three-dimensional model of the aircraft is tested respectively, and the plurality of performance test data are output.

8. The civil aircraft flight test system based on model design is characterized by: The steps for implementing the civil aircraft flight test method based on model design as described in any one of claims 1 to 7 include: A data acquisition module is used to acquire images and point clouds of the civil aircraft to be tested, and obtain multiple aircraft images and multiple aircraft point cloud data sets; A contour feature extraction module, used to extract contour features from the plurality of aircraft images to obtain a plurality of aircraft contours; a point cloud fitting module, configured to perform point cloud fitting on the plurality of aircraft point cloud data sets according to the plurality of aircraft contours, obtain an optimal point cloud fitting result set, and construct an aircraft three-dimensional model based on the optimal point cloud fitting result set; The aircraft test module is used to test the three-dimensional model of the aircraft in a predetermined flight test platform according to a predetermined test plan and output a plurality of performance test data.

Images