Real-time measurement system and method for leading edge slat of civil aircraft

Abstract

The invention belongs to the technical field of lappet and slat testing, and particularly relates to a civil aircraft leading edge slat real-time measuring system and method. The measuring system comprises a leading edge slat mark point, a main wing mark point, a plurality of high-resolution cameras, a synchronous trigger controller, an airborne high-definition video acquisition recorder, a time code distributor and a direct-current power converter. The leading edge slat mark points are arranged on the upper surface of the closed contact position of a leading edge slat and a main wing, and the main wing mark points are arranged on the upper surface of the main wing and correspond to the leading edge slat mark points one to one in position. The plurality of high-resolution cameras are respectively in communication connection with the synchronous trigger controller, the airborne high-definition video acquisition recorder, the time code distributor and the direct-current power supply converter, and the synchronous trigger controller is used for generating synchronous trigger control signal instructions of the plurality of high-resolution cameras; the airborne high-definition video acquisition recorder is used for acquiring image information of each high-resolution camera, and the time code distributor is used for unifying the time of the plurality of high-resolution cameras; the direct current power supply converter is used for providing 12VDC power supply for each high-resolution camera; every two high-resolution cameras form a group to form intersection measurement on the leading edge slat mark point and the main wing mark point, and the view field covers the whole wing. Compared with single-station measurement, the intersection measurement is higher in precision, and the three-dimensional deformation of the slat leading edge can be obtained.

Description

Technical Field

[0001] The present invention belongs to the technical field of flap and slat testing, and in particular relates to a real-time measurement system and method for leading edge slats of a civil aircraft. Background Art

[0002] Large aircraft will produce aeroelastic deformation during flight, which mainly includes bending and twisting of the wing, and relative deformation of the active surface and fixed surface (such as the wing and flaps). The maximum deformation of the wing of a large aircraft usually reaches several meters. If it exceeds the design expectation, it may have a negative impact on the lift and drag characteristics of the aircraft, and more seriously, it may even affect the flight safety of the aircraft. Therefore, how to obtain the actual shape of a large aircraft during flight in order to ensure the design performance of the aircraft is a higher requirement for refined aircraft design to refined flight test technology. At present, there is a lack of methods to measure such deformation in China. This refined flight test technology can help aircraft design units more accurately locate and solve actual model problems.

[0003] In the post-certification drag reduction optimization work of the ARJ aircraft, a series of engineering studies have proved that there is an uncoordinated deformation between the slats and the main wing of the ARJ aircraft in the cruising state, and the deformation cannot be ignored, which will bring a large increase in drag in the cruising state. Therefore, it is necessary to accurately measure the relative displacement of the slats and the main wing. Summary of the invention

[0004] Purpose of the invention: The purpose of the invention is to establish a real-time measurement system and method for the leading edge slat of a civil aircraft, so as to achieve accurate measurement of the relative displacement between the slat and the main wing of the civil aircraft.

[0005] Technical solution of the invention: According to the first aspect of the invention, a real-time measurement system for leading edge slats of civil aircraft is proposed, comprising leading edge slat markers, main wing markers, multiple high-resolution cameras, a synchronous trigger controller, an airborne high-definition video acquisition recorder, a time code distributor, and a DC power converter;

[0006] The leading edge slat mark point is set on the upper surface where the leading edge slat and the main wing are in contact, and the main wing mark point is set on the upper surface of the main wing, corresponding to the position of the leading edge slat mark point one by one; multiple high-resolution cameras are respectively connected to the synchronous trigger controller, the airborne high-definition video acquisition recorder, the time code distributor, and the DC power converter. The synchronous trigger controller is used to generate synchronous trigger control signal instructions for multiple high-resolution cameras, the airborne high-definition video acquisition recorder is used to collect image information of each high-resolution camera, and the time code distributor is used to unify the time of multiple high-resolution cameras; the DC power converter is used to provide 12VDC power for each high-resolution camera; two high-resolution cameras form a group to form an intersection measurement of the leading edge slat mark point and the main wing mark point, and the field of view covers the entire wing. Compared with single-station measurement, the intersection measurement has higher accuracy and can obtain the three-dimensional deformation of the leading edge of the slat.

[0007] In a possible embodiment, a plurality of high-resolution cameras are mounted at the observation windows of the aircraft.

[0008] In a possible embodiment, the leading edge slat marking points and the main wing marking points are evenly spaced along the wingspan direction.

[0009] In a possible embodiment, the leading edge slat marking point and the main wing marking point adopt the BMW logo.

[0010] In a possible embodiment, a total station is used to calibrate multiple high-resolution cameras respectively.

[0011] According to a second aspect of the present invention, a method for real-time measurement of a leading edge slat of a civil aircraft is provided, which uses the above-mentioned real-time measurement system for a leading edge slat of a civil aircraft, and comprises the following steps:

[0012] Step 1: Conduct ground tests to establish a calibration control field for system calibration. Use the calibration plate to calibrate the internal orientation elements of the high-resolution camera; use the total station to calibrate the external orientation elements of the high-resolution camera. Use the total station to obtain the three-dimensional coordinates of the marker points A1-A8, B1-B8 and the image coordinates of the marker points A1-A8, B1-B8 in the calibration image to perform rear intersection calculations to obtain the external orientation elements of each high-resolution camera in the installed state.

[0013] Step 2: During the aircraft cruising process, each high-resolution camera is synchronously triggered to continuously collect and record video data of the aircraft's leading edge slats;

[0014] Step 3: Download the recorded video data and extract the landmarks. First, preprocess the image based on the fourth-order partial differential equation to reduce the noise in the image and enhance the edge features of the landmarks. Then, segment the landmarks based on the multi-threshold iterative image segmentation algorithm to reduce the algorithm complexity increased by the blind search of the single threshold segmentation algorithm, reduce the segmentation time, and improve the accuracy. Finally, improve the KCF algorithm, extract the HOG features and texture features of the image, design the scale filter, optimize the learning rate, and track multiple targets simultaneously to achieve rapid tracking of multiple landmarks and obtain the image coordinates of the landmarks.

[0015] Step 4: Combine the calibration parameters of the high-resolution cameras, intersect and calculate the relative movement distance of the leading edge slats of each group of high-resolution cameras, and splice the measurement results of each group of high-resolution cameras to obtain the relative movement distance of the entire leading edge slat;

[0016] Step 5: Compare the intersection calculation results with the total station measurement results to verify the measurement accuracy.

[0017] In one possible embodiment, in step 1, in order to ensure the final measurement accuracy and the effectiveness of the solution, it is necessary to calibrate the internal and external elements of each high-resolution camera during the calibration process, wherein the internal elements mainly refer to the accurate focal length, image principal point deviation, lens distortion and other parameters of the high-resolution camera; the external elements mainly refer to the position and attitude parameters of the high-resolution camera coordinate system relative to the measurement coordinate system after the high-resolution camera is installed. The specific process of calibrating the internal and external parameters of the high-resolution camera is as follows:

[0018] Use the calibration plate to calibrate the internal orientation elements of high-resolution cameras. Synchronously trigger multiple groups of high-resolution cameras to shoot calibration plates at different angles. Based on the principle of bundle adjustment, calculate the internal orientation elements and lens distortion parameters of each high-resolution camera in the installed state.

[0019] The total station is used to calibrate the exterior orientation elements of the high-resolution camera. Since the target is slender, a single high-resolution camera covers nearly 5m along the wingspan and less than 0.5m in the vertical wingspan direction. Therefore, in order to improve the measurement accuracy, the high-resolution camera is calibrated in different areas.

[0020] Firstly, a total station is used to establish a three-dimensional coordinate system with the center of the wing root as the origin, the X-axis along the wing direction, the Z-axis perpendicular to the wing upward as the wing, and the Y-axis direction satisfying the right-hand rule. The total station is used to measure the three-dimensional coordinates of the marker points A1-A8 and B1-B8 when the leading edge slat is fully opened and fully closed. While the total station is measuring, each high-resolution camera is synchronously triggered to capture images when the leading edge slat is fully opened and fully closed. Afterwards, the image coordinates of the marker points A1-A8 and B1-B8 are extracted from the captured images. Finally, the rear intersection calculation is performed based on the three-dimensional coordinates and image coordinates of the marker points A1-A8 and B1-B8 to obtain the exterior orientation elements of each high-resolution camera in the installed state.

[0021] In a possible embodiment, in step 3, the specific process of extracting the landmark points is as follows:

[0022] Due to the limited installation angle of the high-resolution camera in the observation window, the marker points pasted on the wing are greatly distorted, and the integrity of the geometric edges of the marker points cannot be guaranteed. Therefore, the image is first preprocessed based on the fourth-order partial differential equation to reduce the noise in the image and enhance the edge features of the marker points; then the marker points are segmented based on the multi-threshold iterative image segmentation algorithm to reduce the algorithm complexity increased by the single threshold segmentation algorithm due to blind search, reduce the segmentation time, and improve the accuracy. Finally, the KCF algorithm is improved, the HOG features and texture features of the image are extracted at the same time, and a scale filter is designed to make the algorithm have scale adaptive ability. The occlusion factor is designed and the learning rate of the filter model is optimized. The single target tracking algorithm is optimized to a multi-target simultaneous tracking algorithm to achieve rapid tracking of multiple markers and obtain the pixel coordinates of the marker points A1-A8 and B1-B8.

[0023] In a possible embodiment, in step 4, the specific process of calculating the relative displacement includes:

[0024] Combined with the calibration results of the high-resolution camera, the forward intersection calculation is performed based on the pixel coordinates of the marker points A1-A8 and B1-B8 and the internal and external parameters of the high-resolution camera to obtain the three-dimensional coordinates of the marker points A1-A8 and B1-B8, and the distance between the marker points A1-A8 and B1-B8 is calculated respectively, and the moving distance of the leading edge slat can be obtained; finally, the measurement results of multiple groups of high-resolution cameras are spliced ​​to obtain the relative moving distance of the entire leading edge slat.

[0025] Beneficial effects of the present invention:

[0026] The present invention proposes a real-time measurement method for the leading edge slats of civil aircraft, which makes up for the shortcomings of contact-type flap and slat deformation measurement. The method is derived from the engineering practice of ARJ aircraft flap and slat measurement, and is also a summary and improvement of engineering practice. It has strong guidance and operability and is of great significance. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 A schematic diagram of a real-time measurement system for leading edge slats of a civil aircraft according to a preferred embodiment of the present invention;

[0028] Figure 2 A schematic diagram of equipment and signal connections in a preferred embodiment of the present invention;

[0029] Figure 3 This is a landmark point extraction result diagram of a preferred embodiment of the present invention;

[0030] Figure 4 A flowchart of an improved KCF algorithm according to a preferred embodiment of the present invention;

[0031] Figure 5 A trend diagram of the control points of the leading edge slat closing process according to a preferred embodiment of the present invention;

[0032] Figure 6 This is a diagram showing the accuracy comparison results of a preferred embodiment of the present invention. DETAILED DESCRIPTION

[0033] The present invention proposes a real-time measurement system and method for the leading edge slat of a civil aircraft, which can accurately measure the relative displacement between the slat and the main wing of the civil aircraft from four aspects: image acquisition, landmark point extraction, intersection calculation and data splicing, and accuracy verification.

[0034] 1. Image acquisition

[0035] like Figure 1 As shown, without changing the aerodynamic shape of the aircraft, multiple marking points are evenly pasted on the slats and the main wing to cover the entire wing; they are respectively recorded as leading edge slat marking points 200 (B1-B8) and main wing marking points 100 (A1-A8); a high-resolution camera 300 is used, and the focal length and field of view of the high-resolution camera are determined; multiple high-resolution cameras are set up at the aircraft observation window, the field of view is adjusted, and two high-resolution cameras are grouped together to form an intersection measurement, and the field of view covers the entire wing, forming a relay measurement mode;

[0036] A real-time measurement system for leading edge slats of civil aircraft is mainly composed of 6 high-resolution cameras 300, a synchronous trigger controller, an airborne high-definition video acquisition recorder, a time code distributor and a DC power converter. Figure 2 As shown, it mainly realizes the video collection and recording, video data processing and analysis of the deformation of the gap between the slat and the main wing during the aircraft ground taxiing and cruising configuration.

[0037] (II) Landmark point extraction

[0038] The landmark point extraction process consists of three parts: designing an image preprocessing algorithm based on a fourth-order partial differential equation, an image segmentation algorithm based on multi-threshold iteration, and a multi-target tracking algorithm based on KCF. Figure 3 shown.

[0039] (1) Image preprocessing algorithm

[0040] Since the installation angle of the high-resolution camera in the observation window is limited, the marker points pasted on the wing are greatly distorted and the integrity of the geometric edges of the marker points cannot be guaranteed. Therefore, the image needs to be preprocessed. An image preprocessing algorithm based on the fourth-order partial differential equation is designed. This algorithm retains edge features while removing noise, and has high processing efficiency and good stability.

[0041] Edge detection function:

[0042]

[0043] Denoising model:

[0044]

[0045] (2) Image segmentation algorithm

[0046] There are multiple landmarks on the wing that need to be segmented at the same time, and the single threshold segmentation algorithm cannot handle it well. Therefore, a multi-threshold iteration-based image segmentation algorithm is designed to reduce the algorithm complexity increased by the blind search of the single threshold segmentation algorithm, reduce the segmentation time, and improve the accuracy. Segmentation steps:

[0047] 1) Image equalization: Increase the grayscale level of the image to make the slight grayscale changes in a large area appear, which is convenient for image segmentation;

[0048] 2) Binary mask: Generate a binary mask for the image inside the region to improve the uneven grayscale;

[0049] 3) Iterative threshold segmentation: Segmentation is performed based on the binary mask, reducing the blind search threshold and the number of iterations.

[0050] (3) Multi-target tracking algorithm

[0051] The KCF algorithm uses circulant matrices for dense sampling and uses Fourier transform to transfer calculations from the time domain to the frequency domain, reducing computational complexity, updating the filter model in real time during tracking, and improving algorithm robustness. In order to track the target more accurately in the relative deformation measurement of flaps and slats, the KCF algorithm is improved:

[0052] 1) Multi-feature fusion: extract the HOG features and texture features of the image, and use the PCA algorithm to reduce the dimension. The algorithm accuracy increased by 3.62% and the success rate increased by 6.7%;

[0053] The formula for extracting the HOG features of an image is:

[0054] Among them, C(i,j) represents the feature vector of each cell, N -1,-1 (i,j) represents the normalized eigenvector.

[0055] Formula for extracting texture features of an image:

[0056] Among them, CON represents image contrast, ASM represents image energy, IDM represents image inverse moment, ENT represents image entropy, and COR represents image correlation.

[0057] 2) Scale Adaptation: The scale filter enables the algorithm to have scale adaptive capabilities, and the algorithm accuracy rate increases by 6.34% and the success rate increases by 38.89%;

[0058] Scaling filter:

[0059] Where f is the input image, h is the filter template, g is the output response, λ is the regularization parameter, η is the learning rate, and A t and B t is the factor of the current frame, A t-1 and B t-1 is the factor of the previous frame.

[0060] 3) Optimize learning rate: Design the occlusion factor and optimize the learning rate of the filter model, the algorithm accuracy increased by 1.36% and the success rate increased by 1.06%;

[0061] Occlusion Factor:

[0062] Learning Rate:

[0063] Among them, y max is the maximum value of the output response, y min is the minimum value of the output response, σ 2 is the variance of the output response, n is the number of output responses, is the mean of the output responses.

[0064] 4) Multi-target tracking: It mainly detects the targets in each frame, assigns an ID to each target, and keeps the ID of each target unchanged during the target movement. For each tracked individual, it is assumed that the prediction box with the largest IoU ratio is the same target as it. For multi-target tracking, the improved algorithm calculates the IoU ratio of all detector prediction boxes and tracker observation boxes, and the optimal pairing result has the largest sum of IoU ratios.

[0065] The improved KCF algorithm process is as follows Figure 4 As shown, the tracking steps are as follows:

[0066] a) Initialize the filter parameters. Set the feature dimension of the position filter to 28, the number of scales of the scale filter to 17, the number of scales after interpolation to 33, the regularization term λ of the position filter to 0.0001, the regularization term λ of the scale filter to 0.01, the learning rate η when not occluded to 0.005, the learning rate η when occluded to 0, and the size of the candidate box to 2.5 times the size of the tracking box;

[0067] b) assigning IDs to multiple targets, reading visible light images and infrared images and determining whether it is the first frame, if so, jumping to step c), otherwise jumping to step d);

[0068] c) Detect the target on the first frame and select the target to be tracked; then extract the image multi-image and use principal component analysis to reduce the dimension and fuse the features; finally initialize the position filter and scale filter;

[0069] d) extracting the features of the candidate samples in the same manner as the method of extracting the training features in step 3, and then using the position filter and the scale filter to detect the target and obtain the output response;

[0070] e) Calculate the occlusion factor using the output response. If the occlusion factor ρ occ Greater than the maximum occlusion factor ρ max 0.15 times of the occlusion factor ρ, the target is not occluded, and the position and scale of the maximum response are the position and scale of the tracked target. Then, the features of the visible light image and infrared image are re-extracted and trained, and the position filter and scale filter are updated. occ Less than or equal to the maximum occlusion factor ρ max If the value is 0.15 times of , the target is blocked. At this time, the update of the position filter and the scale filter is stopped, and the target position is obtained using the target detection algorithm.

[0071] f) After obtaining the position and scale of the tracking target, determine whether it is the last frame image. If it is the last frame, end the current tracking; if not, jump to step b).

[0072] (III) Intersection calculation and data splicing

[0073] Based on the pixel coordinates of the marker point and the internal and external parameters of the high-resolution camera, the three-dimensional coordinates of the marker point are obtained, and the distance of the leading edge slat movement can be obtained by difference calculation. Among them, the binocular stereo intersection measurement method is to observe the same object from two perspectives to obtain the perceived images under different perspectives, and obtain the three-dimensional shape information of the object surface by calculating and analyzing the parallax of the corresponding image points in different images of the same three-dimensional point.

[0074] In the present invention, the control field range selected is point A1 to point A8, point B1 to point B8. The pixel coordinates in the closing process are intersected forward, and the three-dimensional coordinates of each point in the closing process can be calculated, so as to obtain the change trend of the distance between each point in the closing process of the leading edge slat, such as Figure 5 The variation trend of the relay measurement is consistent with the variation trend during the movement of the leading edge slats, meeting the model mission requirements.

[0075] (IV) Accuracy Verification

[0076] The relative movement distance of the leading edge slats of the ARJ aircraft when they are opened and closed is calculated by intersection, and the measurement results of multiple sets of high-resolution cameras are spliced ​​and compared with the results of total station measurement. Figure 6 As shown, it can be seen that the intersection measurement error is within ±5mm, which meets the model mission requirements.

Claims

1. A real-time measurement system for leading edge slats of civil aircraft, characterized in that: It includes a leading edge slat mark point, a main wing mark point, a plurality of high-resolution cameras, a synchronous trigger controller, an airborne high-definition video acquisition recorder, a time code distributor, and a DC power converter; the leading edge slat mark point is set on the upper surface of the leading edge slat and the main wing at the closed contact point, and the main wing mark point is set on the upper surface of the main wing, corresponding to the position of the leading edge slat mark point one by one; the plurality of high-resolution cameras are respectively connected to the synchronous trigger controller, the airborne high-definition video acquisition recorder, the time code distributor, and the DC power converter for communication; the synchronous trigger controller is used to generate synchronous trigger control signal instructions for the plurality of high-resolution cameras, the airborne high-definition video acquisition recorder is used to collect image information of each high-resolution camera, and the time code distributor is used to unify the time of the plurality of high-resolution cameras; The DC power converter is used to provide 12VDC power for each high-resolution camera; two high-resolution cameras form a group to form intersection measurements of the leading edge slat mark points and the main wing mark points, and the field of view covers the entire wing.

2. A civil aircraft leading edge slat real-time measurement system according to claim 1, characterized in that: Several high-resolution cameras are installed at the aircraft observation windows.

3. A civil aircraft leading edge slat real-time measurement system according to claim 1, characterized in that: The leading edge slat marking points and the main wing marking points are evenly spaced along the wingspan direction.

4. A civil aircraft leading edge slat real-time measurement system according to claim 1, characterized in that: The leading edge slat marking points and main wing marking points adopt the BMW logo.

5. A civil aircraft leading edge slat real-time measurement system according to claim 1, characterized in that: A total station is used to calibrate multiple high-resolution cameras.

6. A real-time measurement method for leading edge slats of a civil aircraft, characterized in that: A civil aircraft leading edge slat real-time measurement system according to any one of claims 1 to 5 is used, comprising the following steps: Step 1: Conduct ground tests to establish a calibration control field for system calibration; use the calibration plate to calibrate the internal orientation elements of the high-resolution camera; use the total station to calibrate the external orientation elements of the high-resolution camera, obtain the three-dimensional coordinates of the main wing marker points A1-A8 and the leading edge slat marker points B1-B8 through the total station, and perform rear intersection calculations on the image coordinates of the main wing marker points A1-A8 and the leading edge slat marker points B1-B8 in the calibration image to obtain the external orientation elements of each high-resolution camera in the installed state; Step 2: During the aircraft cruising process, each high-resolution camera is synchronously triggered to continuously collect and record video data of the aircraft's leading edge slats; Step 3: Download the recorded video data and extract the landmarks; Step 4: Combine the calibration parameters of the high-resolution cameras, intersect and calculate the relative movement distance of the leading edge slats of each group of high-resolution cameras, and splice the measurement results of each group of high-resolution cameras to obtain the relative movement distance of the entire leading edge slat; Step 5: Compare the intersection calculation results with the total station measurement results to verify the measurement accuracy.

7. A method for real-time measurement of leading edge slats of a civil aircraft according to claim 6, characterized in that: In step 1, in order to ensure the final measurement accuracy and the effectiveness of the solution, it is necessary to calibrate the internal and external orientation elements of each high-resolution camera during the calibration process, wherein the internal orientation elements mainly refer to the accurate focal length, image principal point deviation, and lens distortion parameters of the high-resolution camera; the external orientation elements mainly refer to the position and attitude parameters of the high-resolution camera coordinate system relative to the measurement coordinate system after the high-resolution camera is installed.

8. A method for real-time measurement of leading edge slats of a civil aircraft according to claim 6, characterized in that: In step 1 and step 3, the specific process of extracting the landmark points is as follows: Firstly, the image is preprocessed based on the fourth-order partial differential equation to reduce the noise in the image and enhance the edge features of the landmark points. Then, the landmark points are segmented based on the multi-threshold iterative image segmentation algorithm to reduce the algorithm complexity increased by the blind search of the single threshold segmentation algorithm, reduce the segmentation time and improve the accuracy. Finally, the KCF algorithm is improved, the HOG features and texture features of the image are extracted at the same time, and a scale filter is designed to make the algorithm scale adaptive. The occlusion factor is designed and the learning rate of the filter model is optimized. The single target tracking algorithm is optimized to a multi-target simultaneous tracking algorithm to achieve rapid tracking of multiple landmark points, and the pixel coordinates of the main wing landmark points A1-A8 and the leading edge slat landmark points B1-B8 are obtained.

9. A method for real-time measurement of leading edge slats of a civil aircraft according to claim 6, characterized in that: In step 4, the specific process of calculating the relative displacement includes: Combined with the calibration results of the high-resolution camera, the forward intersection calculation is performed based on the pixel coordinates of the main wing marker points A1-A8 and the leading edge slat marker points B1-B8 and the internal and external parameters of the high-resolution camera to obtain the three-dimensional coordinates of the main wing marker points A1-A8 and the leading edge slat marker points B1-B8, and the distances between the main wing marker points A1-A8 and the leading edge slat marker points B1-B8 are calculated respectively to obtain the moving distance of the leading edge slat; finally, the measurement results of multiple groups of high-resolution cameras are spliced ​​to obtain the relative moving distance of the entire leading edge slat.

Images