Track identification and shock wave visualization device and method for thrown object in wind tunnel

Abstract

The invention discloses a device and a method for track identification and shock wave visualization of a thrown object in a wind tunnel, and belongs to the technical field of aerodynamics test. In order to realize accurate identification of a throwing track and combined visualization of shock waves, the device comprises a double-layer support formed by connecting square guide rails through a wrench sliding block, the double-layer support is arranged at the position, corresponding to an optical window, of the outer side of a side wall plate of a wind tunnel test section, and a first imaging device and a second imaging device are arranged on the upper layer of the double-layer support; a direct-current high-power LED light source is arranged in the middle of the lower layer of the double-layer bracket; the computer is connected with the first imaging device and the second imaging device. The digital pulse signal delay generator is connected with the first imaging device and the second imaging device. Random speckle backgrounds are arranged on the inner side of the wind tunnel test section side wall plate corresponding to the inner wall of the wind tunnel test section; the putting device is connected with the model and is arranged in a wind tunnel test section for putting the model. The comprehensive analysis capability of the wind tunnel experiment is improved by combining a visualization means.

Description

Technical Field

[0001] The invention belongs to the technical field of aerodynamic test and experiment, and in particular relates to a device and method for trajectory recognition and shock wave visualization of an object dropped in a wind tunnel. Background Art

[0002] In the field of modern aviation, the accuracy of object delivery directly affects the operational effect and system performance. The identification of delivery trajectories is crucial, especially in complex aerial environments. Accurate trajectory identification can help predict the stability and flight characteristics during delivery. Wind tunnel testing is an important means to study the aerodynamic characteristics of delivery. By simulating the delivery process, the performance and potential problems of objects in complex flight environments can be effectively evaluated.

[0003] In wind tunnel tests, the shock wave effect accompanying the dropping of objects has an important impact on flight stability and drop accuracy. Therefore, combining trajectory recognition with shock wave visualization can more comprehensively analyze the flow field changes encountered by objects during the drop process, thereby providing an important basis for the design, optimization and performance improvement of the aviation drop system.

[0004] Traditional methods for identifying delivery trajectories mainly rely on feature recognition technology, which analyzes the features of the object's shape or motion trajectory by extracting them. However, when the object rotates or changes its posture significantly, traditional methods are prone to feature loss, resulting in unstable trajectory tracking and a significant decrease in accuracy. On the other hand, although traditional shock wave observation methods, such as the Schlieren method, can effectively capture the dynamics of shock waves, their application in wind tunnel tests is limited. The Schlieren method has strict requirements on the adjustment of the optical path, and the equipment is large and occupies a large space, which makes it difficult to flexibly arrange it in a wind tunnel environment with limited space, especially in complex flow fields or small spaces. It is difficult to provide continuous and high-quality shock wave observations. Therefore, existing methods have significant limitations in delivery trajectory identification and shock wave observation, and innovative technologies are needed to break through these bottlenecks. Summary of the invention

[0005] The problem to be solved by the present invention is to realize accurate identification of delivery trajectory and joint visualization of shock wave, and propose a device and method for trajectory identification and shock wave visualization of delivered objects in a wind tunnel.

[0006] To achieve the above object, the present invention is implemented through the following technical solutions: A device for trajectory recognition and shock wave visualization of an object dropped in a wind tunnel, comprising a first imaging device, a second imaging device, an L-shaped block, a wrench slider, a DC high-power LED light source, a computer, a digital pulse signal delay generator, a square guide rail, and a random speckle background; The square guide rails are connected to form a double-layer bracket through a wrench slider, and are placed on the outer side of the side wall plate of the wind tunnel test section corresponding to the optical window position. The first imaging device and the second imaging device are placed on the upper layer of the double-layer bracket, a DC high-power LED light source is placed in the middle of the lower layer of the double-layer bracket, and L-shaped blocks are connected to both ends of the lower layer of the double-layer bracket; The computer is connected to the first imaging device and the second imaging device respectively, and the digital pulse signal delay generator is connected to the first imaging device and the second imaging device respectively; A random speckle background is arranged on the inner side of the side wall plate of the wind tunnel test section corresponding to the inner wall of the wind tunnel test section; The launching device is connected to the model and placed in the wind tunnel test section for launching the model.

[0007] Furthermore, a plurality of feature points are irregularly arranged on the surface of the model, ensuring that 6-10 complete feature points are recognized in the field of view of the first imaging device and the second imaging device.

[0008] Furthermore, the duty cycle of the spots in the random speckle background is 40%-50%, the spots are 3-4 pixels in the image, and the colors of the spots and the blank space are black and white.

[0009] Furthermore, the first imaging device and the second imaging device are composed of a camera and a lens.

[0010] A method for identifying the trajectory of an object placed in a wind tunnel and visualizing shock waves is implemented based on the device for identifying the trajectory of an object placed in a wind tunnel and visualizing shock waves, and includes the following steps: S1. Model preparation, irregularly arranging feature points on the model surface; S2. Perform three-dimensional mapping of the center of the feature point in the model coordinate system to obtain a three-dimensional coordinate sequence of the feature points arranged in step S1 in the model coordinate system; S3. Arrange random speckle background on the inner wall of the test section; S4. Build a device for identifying trajectory of objects and visualizing shock waves in a wind tunnel, and adjust the focal length and exposure time of the camera lens of the first imaging device and the second imaging device; S5. Obtain parameters of the first imaging device and the second imaging device, place a checkerboard calibration plate at different positions in the test area, and solve the internal and external parameters of the camera based on Zhang's calibration method; S6. Wind tunnel coordinate system calibration: the wind tunnel coordinate system calibration is realized by using a checkerboard calibration plate, and the three-dimensional coordinates under the camera coordinates are converted to the wind tunnel coordinate system; S7. The first imaging device and the second imaging device take random speckle background images when there is no wind and the model is not installed to obtain a background image with light noise removed , , then taking images of the model during the wind tunnel test to obtain an image sequence captured by the first imaging device and an image sequence captured by the second imaging device; S8. Draw the model delivery trajectory based on the image sequence acquired by the first imaging device and the image sequence acquired by the second imaging device obtained in step S7; S9. Solve the shock wave stereogram and qualitatively observe the refractive index change in the two-dimensional shock wave phenomenon of the model; S10, visualizing the model delivery trajectory obtained in step S8 and the shock wave shaping diagram obtained in step S9.

[0011] Furthermore, the specific implementation method of step S6 includes the following steps: Place the checkerboard calibration plate horizontally on the side wall of the wind tunnel, and ensure that the x direction of the checkerboard is parallel to the airflow direction, and the y direction of the checkerboard is parallel to the vertical direction of the ground; calculate the difference of the three-dimensional coordinates of the corner points in the x direction of the checkerboard, and normalize them to obtain the airflow axis vector , calculate the difference of the three-dimensional coordinates of the intersection points in the y direction of the chessboard, and normalize them to obtain the yaw axis vector ; then based on , and the pitch axis vector are orthogonal to each other, and the pitch axis vector is obtained Finally, we get the axis transformation matrix from the camera coordinate system to the wind tunnel coordinate system .

[0012] Furthermore, the specific implementation method of step S7 includes the following steps: S7.1. When there is no wind and the model is not installed, take a random speckle background image, collect 10 random speckle background images and take the average to eliminate light noise, and obtain the background image collected by the first imaging device without light noise , the background image with light noise removed collected by the second imaging device ; S7.2. During the wind tunnel test, the test conditions are first reached, and then the computer sends a trigger signal to the digital pulse signal delay generator to trigger the first imaging device and the second imaging device to collect images. After a half-second interval, a trigger signal is sent to the delivery device to trigger the model to start delivery. The first imaging device and the second imaging device collect all process photos of the model from the beginning to the end of delivery, and obtain the image sequence collected by the first imaging device. , a sequence of images acquired by a second imaging device ,in, is the image data of the maximum time acquired by the first imaging device, The image data of the maximum time acquired by the second imaging device.

[0013] Furthermore, the specific implementation method of step S8 includes the following steps: S8.1. Image preprocessing: Preprocess the image collected in step S7 by binarizing it using the Otsu threshold method to generate a black and white image; S8.2. Contour detection and screening: The contours in the image processed in step S8.1 are extracted by the Suzuki contour tracking algorithm, and the roundness of the contours is calculated to screen out contours that are close to a circle; the roundness formula is: ; in, is the roundness of the identified contour, is the contour area, is the perimeter of the contour, if Less than a given contour threshold , then the recognized contour is ignored; S8.3. Ellipse fitting and affine transformation: Perform ellipse fitting on the selected contours to obtain the center coordinates of the ellipse , major axis a, minor axis b and rotation angle, the elliptical area is converted into a perfect circle for standardization, and the affine transformation matrix M is calculated based on the four vertices of the fitted ellipse and the four vertices of the standard circle; S8.4. Region transformation and secondary binarization: According to the center coordinates of the ellipse , and cut the area by three times the length of the major axis a and the minor axis b to include the local image where the encoding point is located ; According to the affine transformation matrix M Perform affine transformation to obtain the transformed image , enlarge the image to the standard size to obtain the transformed image of the standard size ;Will Convert to grayscale and binary images again; S8.5. Code point detection: make The center coordinates of , set the sampling radius to: first sampling radius , the second sampling radius , the third sampling radius is ,in, is the base radius, for 1 / 3 of the radius of the largest inscribed circle; At each sampling radius, uniformly select sampling points, the coordinate calculation formula of the jth sampling point is: ; ; Where j represents the sampling point index; at each sampling point Read the pixel value and then judge Whether it is a code point, the judgment rule is Are all sample points of the radius white? Are all sample points of the radius black? The number of sampling points of the radius>2; if the above rules are met, the extracted area is considered Contains the code point, and obtains the area containing the code point; S8.6. Encoder decoding: Extracting coding information from the area confirmed to contain the coding point and generating a unique identifier for the coding point; N points are sampled based on the third sampling radius area to extract the coding information. When sampling, the image is rotated at different angles to eliminate the influence of the angle on the coding. The coordinate calculation formula of the kth sampling point used to extract the coding information is: ; ; Where k represents the sampling point index used to extract the coding information. is the rotation angle, which is used to change the starting angle of the sampling point circle by circle; The encoding result of each circle is rotated and normalized, and then the encoding results of all circles are averaged. The average code is binarized to obtain the final binary code, which is converted into a decimal digital index as the unique identifier of the code point; S8.7. Draw the model delivery trajectory based on encoder decoding; For the image captured by the first imaging device at time t , the image captured by the second imaging device at time t ,in , identify the sequence of image coordinates of the center of the code point of the first imaging device as , the image coordinate sequence of the center of the code point of the second imaging device is ; based on and The three-dimensional coordinate sequence of n coded points in the camera coordinate system is obtained by using the camera internal and external parameters obtained in step S6. , and the code point sequence corresponding to n code points ; According to the axis transformation matrix from the camera coordinate system to the world coordinate system calculated in step S6 , get the coordinate sequence of n coded points in the world coordinate system ; By comparing The identification sequence corresponding to the code point on the model in step S2 , and , Coordinate sequence of points of the same name in the model coordinate system ,in for A subset of based on and , calculate the rotation and translation matrix from the model coordinate system to the world coordinate system through the Genro-Kabsch algorithm ; Assume that the center point of the model in the model coordinate system is , then the coordinates of the point in the world coordinate system at time t are ; right Repeat the above steps for the image collected at any time, and you can get the coordinates of the center point of the model in the world coordinate system, and sort them out to get the coordinate sequence on the time axis , draw the trajectory according to the time axis to obtain the model delivery trajectory.

[0014] Furthermore, the specific implementation method of step S9 includes the following steps: S9.1. Automatic extraction of model masks for any , the sequence of image coordinates of the coded points on the model calculated by S8 , calculate its mean coordinates ,by Centered on ,Width Extract image regions , The choice ensures Fully inclusive model; right Perform binarization, use dilation to fill holes in the target area, and combine erosion to remove small noise points; Use the Canny operator to perform edge detection to obtain all contours, and use area screening to retain the largest contour as the mask of the model. , set the model position value to 0 and the other values ​​to 1; S9.2. Calculate the divergence of each pixel of the image; The displacement field of each pixel position of the image at time t is expressed as , with the expression: ; in, is a cross-correlation function used to measure the similarity between two windows, and the normalized cross-correlation method is preferably selected as the cross-correlation method; based on Find the divergence of the current pixel ; S9.3. Create an adaptive divergence-preserving filter to smooth the noise: The weight adjustment is introduced into the divergence calculation to adaptively smooth the high noise area. The formula is: ; in, is the weight function, is the divergence after weight adjustment; Weight function Defined as: ; in, represents the local vector gradient strength, is the local variance of the vector field intensity, It is a hyperparameter that controls weight sensitivity, usually between 0.5 and 2.0; ; Then a smoothness constraint is added to further reduce the noise by minimizing the following expression: ; in, represents the scale factor; The optimized divergence is ; Repeat the above steps until convergence and get the optimal ; S9.4. Process each pixel according to steps S9.1 to S9.3 to obtain the optimized divergence of each pixel. The divergence map of the current image can be obtained by sorting. , combining this image with the one in step S9.1 Get the final shock wave qualitative diagram of the bomb model at time t , the expression is: .

[0015] Further, the specific implementation method of step S10 is to replace the Combine the internal and external parameters of the camera and back-project them to In the image, we get the coordinates of the center point of the model. , draw a circle with a radius of 5 pixels with the image coordinates of the model center point as the center, and display the corresponding 3D coordinates above the circle in real time , superimpose the image coordinate information of the drawn circle and the model center point onto the displacement divergence image Up, yes The above steps are repeated for the images collected at any time, and then the processed images are displayed dynamically to achieve the simultaneous visualization of the two-dimensional shock wave characterization and the model trajectory.

[0016] Beneficial effects of the present invention: The method for identifying the trajectory of an object dropped in a wind tunnel and visualizing shock waves described in the present invention is based on the basic theories of binocular vision three-dimensional reconstruction and image correlation. By arranging asymmetric coding points on the surface of the object model, the problem of missing marking points caused by rotation of the object model during the dropping process is solved, ensuring that a sufficient number of marking points can be identified in any rotation state. This not only improves the robustness of marking point recognition, but also realizes real-time conversion between the wind tunnel coordinate system and the model coordinate system through the correspondence between the coding points in the wind tunnel coordinate system and the model coordinate system. This real-time conversion capability further supports the accurate real-time calculation of the three-dimensional coordinates of any point of the object model in the wind tunnel coordinate system, thereby improving the real-time and accuracy of the experimental measurement, meeting the needs of complex dynamic test scenarios, and providing reliable data support for studying the dynamic characteristics and aerodynamic parameters of the object model.

[0017] The method for identifying the trajectory of an object dropped in a wind tunnel and visualizing shock waves described in the present invention can obtain the two-dimensional shock wave distribution of the object model in real time based on the background schlieren method. Compared with the traditional schlieren method, the present invention does not require complex optical path arrangement, greatly reduces the complexity of the equipment and the installation cost, simplifies the experimental operation process, and significantly improves the efficiency and repeatability of the experiment.

[0018] The method for identifying the trajectory of an object dropped in a wind tunnel and visualizing shock waves described in the present invention proposes an adaptive divergence-fidelity filter field that can optimize the results of background schlieren, clearly display the boundaries of shock waves, and effectively reduce background noise.

[0019] The present invention discloses a method for identifying the trajectory of an object in a wind tunnel and visualizing shock waves. The combination of trajectory identification and shock wave visualization realizes the synchronous measurement and intuitive display of the motion state of the object model and the characteristics of the flow field around it, providing strong technical support for in-depth research on the aerodynamic performance, shock wave behavior, and the interaction between the object model and the shock wave. This combined visualization method improves the comprehensive analysis capability of wind tunnel experiments and provides a more comprehensive and efficient solution for aerodynamic research and engineering applications. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 It is a structural schematic diagram of a device for identifying trajectory of objects dropped into a wind tunnel and visualizing shock waves according to the present invention; Figure 2A schematic diagram of irregularly arranging characteristic points on the surface of the model of the present invention; Figure 3 The present invention is a flow chart of a method for identifying the trajectory of an object dropped into a wind tunnel and visualizing shock waves. DETAILED DESCRIPTION

[0021] In order to make the purpose, technical solution and advantages of the present invention more clear, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention, that is, the specific embodiments described are only part of the embodiments of the present invention, rather than all of the specific embodiments. The components of the specific embodiments of the present invention described and shown in the drawings herein can be arranged and designed in various different configurations, and the present invention can also have other embodiments.

[0022] Therefore, the following detailed description of the specific embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents the selected specific embodiments of the present invention. Based on the specific embodiments of the present invention, all other specific embodiments obtained by those skilled in the art without making creative work are within the scope of protection of the present invention.

[0023] In order to further understand the content, features and effects of the present invention, the following specific implementation methods are given as examples, and the attached Figure 1 -Attached Figure 3 The detailed instructions are as follows:

[0024] Example 1 A device for trajectory recognition and shock wave visualization of an object dropped in a wind tunnel, comprising a first imaging device 4, a second imaging device 5, an L-shaped block 6, a wrench slider 7, a DC high-power LED light source 8, a computer 9, a digital pulse signal delay generator 10, a square guide rail 11, and a random speckle background 12; The square guide rail 11 is connected to form a double-layer bracket through a wrench slider 7, and is placed on the outer side of the wind tunnel test section side wall plate 13 corresponding to the position of the optical window 1. The first imaging device 4 and the second imaging device 5 are placed on the upper layer of the double-layer bracket, and a DC high-power LED light source 8 is placed in the middle of the lower layer of the double-layer bracket. Both ends of the lower layer of the double-layer bracket are connected to L-shaped blocks 6; The computer 9 is connected to the first imaging device 4 and the second imaging device 5 respectively, and the digital pulse signal delay generator 10 is connected to the first imaging device 4 and the second imaging device 5 respectively; A random speckle background 12 is arranged on the inner side of the wind tunnel test section side wall plate 13 corresponding to the inner wall of the wind tunnel test section; The launching device 3 is connected to the model 2 and is placed in the wind tunnel test section for launching the model 2.

[0025] Furthermore, a plurality of feature points are irregularly arranged on the surface of the model 2 , ensuring that 6-10 complete feature points are recognized in the field of view of the first imaging device 4 and the second imaging device 5 .

[0026] Furthermore, the duty cycle of the spots in the random speckle background 12 is 40%-50%, the spots are 3-4 pixels in the image, and the colors of the spots and the blank space are black and white.

[0027] Furthermore, the first imaging device 4 and the second imaging device 5 are composed of a camera and a lens.

[0028] Example 2 A method for identifying the trajectory of an object dropped in a wind tunnel and visualizing shock waves is implemented based on the device for identifying the trajectory of an object dropped in a wind tunnel and visualizing shock waves described in Example 1, and includes the following steps: S1. Model preparation, irregularly arranging feature points on the model surface; Furthermore, the feature points include one or a combination of annular coded points, marking points of different colors and special markings. No matter how the model rolls, at least 6 complete feature points can always be identified in the field of view.

[0029] S2. Perform three-dimensional mapping of the center of the feature point in the model coordinate system to obtain a three-dimensional coordinate sequence of the feature points arranged in step S1 in the model coordinate system; Furthermore, the specific implementation method of step S2 includes the following steps: select a high-precision coordinate measuring machine (CMM) and strictly implement the calibration procedure to ensure that the equipment parameters are set accurately to meet the measurement accuracy requirements. Secondly, use a professional fixture or fixture to firmly mount the blade on the measurement platform to ensure that it remains absolutely still during the measurement process to avoid any displacement or vibration affecting the measurement results. Then, according to the predetermined measurement strategy, use the measurement probe to contact the positioning reference points on the blade surface in turn, and accurately record the coordinate data of all reference points in three-dimensional space. , and then record the identification sequence corresponding to each code point , kmax is the total number of feature points.

[0030] S3. Arrange random speckle background on the inner wall of the test section; Furthermore, the specific implementation method of step S3 is to realize the background by spraying paint and stickers. Preferably, matte white primer is first sprayed on the inner wall, and after the white primer is dry, a hollow low-viscosity sticker is placed on the white primer, and matte black paint is sprayed. After it is naturally dried, the low-viscosity hollow sticker is torn off to realize the preparation of the random spot background; S4. Build a device for identifying trajectory of objects and visualizing shock waves in a wind tunnel, and adjust the focal length and exposure time of the camera lens of the first imaging device and the second imaging device; Furthermore, the specific implementation method of step S4 includes the following steps: Install the first imaging device, the second imaging device and the DC high-power LED light source outside the optical window on the side wall of the test section through guide rails, sliders and other components, and determine the focal length of the camera lens according to the field of view and the model movement speed. and camera exposure time E. Among them, the camera focal length It can be determined according to the following formula: ; Among them, FOV represents the field of view, H represents the chip size of the high-speed camera, is the distance from the camera lens to the object being photographed.

[0031] To avoid motion blur, the exposure time The dragging of the model should not exceed one pixel, that is, the movement length of the model should not exceed 1 pixel. It is the theoretical maximum speed when the missile model is released; S5. Obtain parameters of the first imaging device and the second imaging device, place a checkerboard calibration plate at different positions in the test area, and solve the internal and external parameters of the camera based on Zhang's calibration method; Furthermore, the internal parameters in step S5 are ; The external parameters are ;in is the focal length in the u and v directions, is the radial distortion parameter, are the coordinates of the image principal point; R and t are the rotation matrix and translation vector from the world coordinate system to the camera coordinate system, respectively.

[0032] S6. Wind tunnel coordinate system calibration: the wind tunnel coordinate system calibration is realized by using a checkerboard calibration plate, and the three-dimensional coordinates under the camera coordinates are converted to the wind tunnel coordinate system; Furthermore, the specific implementation method of step S6 includes the following steps: Place the checkerboard calibration plate horizontally on the side wall of the wind tunnel, and ensure that the x direction of the checkerboard is parallel to the airflow direction, and the y direction of the checkerboard is parallel to the vertical direction of the ground; calculate the difference of the three-dimensional coordinates of the corner points in the x direction of the checkerboard, and normalize them to obtain the airflow axis vector , calculate the difference of the three-dimensional coordinates of the intersection points in the y direction of the chessboard, and normalize them to obtain the yaw axis vector ; then based on , and the pitch axis vector are orthogonal to each other, and the pitch axis vector is obtained Finally, we get the axis transformation matrix from the camera coordinate system to the wind tunnel coordinate system .

[0033] S7. The first imaging device and the second imaging device take random speckle background images when there is no wind and the model is not installed to obtain a background image with light noise removed , , then taking images of the model during the wind tunnel test to obtain an image sequence captured by the first imaging device and an image sequence captured by the second imaging device; S7.1. When there is no wind and the model is not installed, take a random speckle background image, collect 10 random speckle background images and take the average to eliminate light noise, and obtain the background image collected by the first imaging device without light noise , the background image with light noise removed collected by the second imaging device ; S7.2. During the wind tunnel test, the test conditions are first reached, and then the computer sends a trigger signal to the digital pulse signal delay generator to trigger the first imaging device and the second imaging device to collect images. After a half-second interval, a trigger signal is sent to the delivery device to trigger the model to start delivery. The first imaging device and the second imaging device collect all process photos of the model from the beginning to the end of delivery, and obtain the image sequence collected by the first imaging device. , a sequence of images acquired by a second imaging device ,in, is the image data of the maximum time acquired by the first imaging device, The image data of the maximum time acquired by the second imaging device.

[0034] S8. Draw the model delivery trajectory based on the image sequence acquired by the first imaging device and the image sequence acquired by the second imaging device obtained in step S7; Furthermore, the specific implementation method of step S8 includes the following steps:

[0035] S8.1. Image preprocessing: Preprocess the image collected in step S7 by binarizing it using the Otsu threshold method to generate a black and white image; S8.2. Contour detection and screening: The contours in the image processed in step S8.1 are extracted by the Suzuki contour tracking algorithm, and the roundness of the contours is calculated to screen out contours that are close to a circle; the roundness formula is: ; in, is the roundness of the identified contour, is the contour area, is the perimeter of the contour, if Less than a given contour threshold , then the recognized contour is ignored; S8.3. Ellipse fitting and affine transformation: Perform ellipse fitting on the selected contours to obtain the center coordinates of the ellipse , major axis a, minor axis b and rotation angle, the elliptical area is converted into a perfect circle for standardization, and the affine transformation matrix M is calculated based on the four vertices of the fitted ellipse and the four vertices of the standard circle; S8.4. Region transformation and secondary binarization: According to the center coordinates of the ellipse , and cut the area by three times the length of the major axis a and the minor axis b to include the local image where the encoding point is located ; According to the affine transformation matrix M Perform affine transformation to obtain the transformed image , enlarge the image to the standard size to obtain the transformed image of the standard size ;Will Convert to grayscale and binary images again; S8.5. Code point detection: make The center coordinates of , set the sampling radius to: first sampling radius , the second sampling radius , the third sampling radius is ,in, is the base radius, for 1 / 3 of the radius of the largest inscribed circle; At each sampling radius, uniformly select sampling points, the coordinate calculation formula of the jth sampling point is: ; ; Where j represents the sampling point index; at each sampling point Read the pixel value and then judge Whether it is a code point, the judgment rule is Are all sample points of the radius white? Are all sample points of the radius black? The number of sampling points of the radius>2; if the above rules are met, the extracted area is considered Contains the code point, and obtains the area containing the code point; S8.6. Encoder decoding: Extracting coding information from the area confirmed to contain the coding point and generating a unique identifier for the coding point; N points are sampled based on the third sampling radius area to extract the coding information. When sampling, the image is rotated at different angles to eliminate the influence of the angle on the coding. The coordinate calculation formula of the kth sampling point used to extract the coding information is: ; ; Where k represents the sampling point index used to extract the coding information. is the rotation angle, which is used to change the starting angle of the sampling point circle by circle; The encoding result of each circle is rotated and normalized, and then the encoding results of all circles are averaged. The average code is binarized to obtain the final binary code, which is converted into a decimal digital index as the unique identifier of the code point; Furthermore, the encoding results of each circle are rotationally normalized to eliminate the influence of angles on the encoding results. The encoding results of all circles are averaged to reduce the influence of noise on the encoding results. The average value of each position represents whether it is "black" or "white" at that position. The mean code is binarized. The threshold is set to 0.5. If the mean value of a position is greater than 0.5, the position is "black" (coded as 1), otherwise it is "white" (coded as 0). The final binary code is obtained. The binary code can be converted into a decimal digital index as the unique identifier of the code point; S8.7. Draw the model delivery trajectory based on encoder decoding; For the image captured by the first imaging device at time t , the image captured by the second imaging device at time t ,in , identify the sequence of image coordinates of the center of the code point of the first imaging device as , the image coordinate sequence of the center of the code point of the second imaging device is ; based on and The three-dimensional coordinate sequence of n coded points in the camera coordinate system is obtained by using the camera internal and external parameters obtained in step S6. , and the code point sequence corresponding to n code points ; According to the axis transformation matrix from the camera coordinate system to the world coordinate system calculated in step S6 , get the coordinate sequence of n coded points in the world coordinate system ; By comparing The identification sequence corresponding to the code point on the model in step S2 , and , Coordinate sequence of points of the same name in the model coordinate system ,in for A subset of based on and , calculate the rotation and translation matrix from the model coordinate system to the world coordinate system through the Genro-Kabsch algorithm ; Assume that the center point of the model in the model coordinate system is , then the coordinates of the point in the world coordinate system at time t are ; right Repeat the above steps for the image collected at any time, and you can get the coordinates of the center point of the model in the world coordinate system, and sort them out to get the coordinate sequence on the time axis , draw the trajectory according to the time axis to obtain the model delivery trajectory.

[0036] S9. Solve the shock wave stereogram and qualitatively observe the refractive index change in the two-dimensional shock wave phenomenon of the model; Furthermore, the specific implementation method of step S9 includes the following steps: S9.1. Automatic extraction of model masks for any , the sequence of image coordinates of the coded points on the model calculated by S8 , calculate its mean coordinates ,by Centered on ,Width Extract image regions , The choice ensures Fully inclusive model; right Perform binarization, use dilation to fill holes in the target area, and combine erosion to remove small noise points; Use the Canny operator to perform edge detection to obtain all contours, and use area screening to retain the largest contour as the mask of the model. , set the model position value to 0 and the other values ​​to 1; S9.2. Calculate the divergence of each pixel of the image; The displacement field of each pixel position of the image at time t is expressed as , with the expression: ; in, is a cross-correlation function used to measure the similarity between two windows, and the normalized cross-correlation method is preferably selected as the cross-correlation method; based on Find the divergence of the current pixel ; S9.3. Create an adaptive divergence-preserving filter to smooth the noise: By introducing divergence smoothing regularization and region-dependent weight function, the robustness of the signal region to noise is enhanced during the divergence calculation process, while the boundary details of high divergence regions (such as shock wave regions) are retained.

[0037] Weight adjustment is introduced into the divergence calculation to adaptively smooth high noise areas. The formula is: ; in, is the weight function, is the divergence after weight adjustment; Weight function Defined as: ; in, represents the local vector gradient strength, is the local variance of the vector field intensity, It is a hyperparameter that controls weight sensitivity, usually between 0.5 and 2.0; ; Then a smoothness constraint is added to further reduce the noise by minimizing the following expression: ; in, represents the scale factor; The square error of the first term of this formula is to make the optimized divergence as close as possible to the original divergence result, avoiding over-smoothing or deviation from the real physical information, ensuring that the optimization will not destroy the shock wave details. The second term introduces smoothing regularization, the purpose is to reduce the violent fluctuations of the optimized divergence field to suppress high-frequency noise. Make the divergence field smoother, in line with the real object separation phenomenon, especially the background area. The optimized divergence is ; In general, the adaptive divergence-preserving filter seeks an optimal divergence field so that it is as close to the real divergence as possible while having less noise and smooth visual effects.

[0038] Repeat the above steps until convergence and get the optimal ; S9.4. Process each pixel according to steps S9.1 to S9.3 to obtain the optimized divergence of each pixel. The divergence map of the current image can be obtained by sorting. , combining this image with the one in step S9.1 Get the final shock wave qualitative diagram of the bomb model at time t , the expression is: .

[0039] S10, visualizing the model delivery trajectory obtained in step S8 and the shock wave shaping diagram obtained in step S9.

[0040] Further, the specific implementation method of step S10 is to replace the Combine the internal and external parameters of the camera and back-project them to In the image, we get the coordinates of the center point of the model. , draw a circle with a radius of 5 pixels with the image coordinates of the model center point as the center, and display the corresponding 3D coordinates above the circle in real time , superimpose the image coordinate information of the drawn circle and the model center point onto the displacement divergence image Up, yes The above steps are repeated for the images collected at any time, and then the processed images are displayed dynamically to achieve the simultaneous visualization of the two-dimensional shock wave characterization and the model trajectory.

[0041] It should be noted that relational terms such as "first" and "second" are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0042] Although the present application has been described above with reference to specific embodiments, various modifications may be made thereto and parts thereof may be replaced with equivalents without departing from the scope of the present application. In particular, as long as there is no structural conflict, the various features in the specific embodiments disclosed in the present application may be used in combination with each other in any manner, and the fact that these combinations are not exhaustively described in this specification is only for the sake of omitting space and saving resources. Therefore, the present application is not limited to the specific embodiments disclosed herein, but includes all technical solutions falling within the scope of the claims.

Claims

1. A device for identifying trajectory of objects dropped in a wind tunnel and visualizing shock waves, characterized in that: It comprises a first imaging device (4), a second imaging device (5), an L-shaped block (6), a wrench slider (7), a DC high-power LED light source (8), a computer (9), a digital pulse signal delay generator (10), a square guide rail (11), and a random speckle background (12); The square guide rail (11) is connected to form a double-layer bracket via a spanner slider (7), and is placed on the outer side of the wind tunnel test section side wall plate (13) at a position corresponding to the optical window (1); the first imaging device (4) and the second imaging device (5) are placed on the upper layer of the double-layer bracket, a DC high-power LED light source (8) is placed in the middle of the lower layer of the double-layer bracket, and L-shaped blocks (6) are connected to both ends of the lower layer of the double-layer bracket; The computer (9) is connected to the first imaging device (4) and the second imaging device (5), respectively; and the digital pulse signal delay generator (10) is connected to the first imaging device (4) and the second imaging device (5), respectively; A random speckle background (12) is arranged on the inner side of the wind tunnel test section side wall plate (13) corresponding to the inner wall of the wind tunnel test section; The launching device (3) is connected to the model (2) and is placed in the wind tunnel test section for launching the model (2).

2. The device for identifying trajectory of objects dropped into a wind tunnel and visualizing shock waves according to claim 1, characterized in that: The model (2) has a plurality of feature points arranged irregularly on its surface, ensuring that 6 to 10 complete feature points can be identified in the field of view of the first imaging device (4) and the second imaging device (5).

3. The device for identifying trajectory of objects dropped into a wind tunnel and visualizing shock waves according to claim 2, characterized in that: The duty cycle of the spots in the random speckle background (12) is 40%-50%, and the spots are 3-4 pixels in the image, and the colors of the spots and the blank space are black and white.

4. The device for identifying trajectory of objects dropped into a wind tunnel and visualizing shock waves according to claim 2, characterized in that: The first imaging device (4) and the second imaging device (5) are composed of a camera and a lens.

5. A method for identifying the trajectory of an object placed in a wind tunnel and visualizing shock waves, implemented by the device for identifying the trajectory of an object placed in a wind tunnel and visualizing shock waves according to any one of claims 1 to 4, characterized in that: The steps include: S1. Model preparation, irregularly arranging feature points on the model surface; S2. Perform three-dimensional mapping of the center of the feature point in the model coordinate system to obtain a three-dimensional coordinate sequence of the feature points arranged in step S1 in the model coordinate system; S3. Arrange random speckle background on the inner wall of the test section; S4. Build a device for identifying trajectory of objects and visualizing shock waves in a wind tunnel, and adjust the focal length and exposure time of the camera lens of the first imaging device and the second imaging device; S5. Obtain parameters of the first imaging device and the second imaging device, place a checkerboard calibration plate at different positions in the test area, and solve the internal and external parameters of the camera based on Zhang's calibration method; S6. Wind tunnel coordinate system calibration: the wind tunnel coordinate system calibration is realized by using a checkerboard calibration plate, and the three-dimensional coordinates under the camera coordinates are converted to the wind tunnel coordinate system; S7. The first imaging device and the second imaging device take random speckle background images when there is no wind and the model is not installed to obtain a background image with light noise removed , , then taking images of the model during the wind tunnel test to obtain an image sequence captured by the first imaging device and an image sequence captured by the second imaging device; S8. Draw the model delivery trajectory based on the image sequence acquired by the first imaging device and the image sequence acquired by the second imaging device obtained in step S7; S9. Solve the shock wave stereogram and qualitatively observe the refractive index change in the two-dimensional shock wave phenomenon of the model; S10, visualizing the model delivery trajectory obtained in step S8 and the shock wave shaping diagram obtained in step S9.

6. The method for identifying trajectory of an object dropped into a wind tunnel and visualizing shock waves according to claim 5, characterized in that: The specific implementation method of step S6 includes the following steps: Place the checkerboard calibration plate horizontally on the side wall of the wind tunnel, and ensure that the x direction of the checkerboard is parallel to the airflow direction, and the y direction of the checkerboard is parallel to the vertical direction of the ground; calculate the difference of the three-dimensional coordinates of the corner points in the x direction of the checkerboard, and normalize them to obtain the airflow axis vector , calculate the difference of the three-dimensional coordinates of the intersection points in the y direction of the chessboard, and normalize them to obtain the yaw axis vector ; then based on , and the pitch axis vector are orthogonal to each other, and the pitch axis vector is obtained Finally, we get the axis transformation matrix from the camera coordinate system to the wind tunnel coordinate system .

7. The method for identifying trajectory of an object dropped into a wind tunnel and visualizing shock waves according to claim 6, characterized in that: The specific implementation method of step S7 includes the following steps: S7.

1. When there is no wind and the model is not installed, take a random speckle background image, collect 10 random speckle background images and take the average to eliminate light noise, and obtain the background image collected by the first imaging device without light noise , the background image with light noise removed collected by the second imaging device ; S7.

2. During the wind tunnel test, the test conditions are first reached, and then the computer sends a trigger signal to the digital pulse signal delay generator to trigger the first imaging device and the second imaging device to collect images. After a half-second interval, a trigger signal is sent to the delivery device to trigger the model to start delivery. The first imaging device and the second imaging device collect all process photos of the model from the beginning to the end of delivery, and obtain the image sequence collected by the first imaging device. , a sequence of images acquired by a second imaging device ,in, is the image data of the maximum time acquired by the first imaging device, The image data of the maximum time acquired by the second imaging device.

8. The method for identifying trajectory of an object dropped into a wind tunnel and visualizing shock waves according to claim 7, characterized in that: The specific implementation method of step S8 includes the following steps: S8.

1. Image preprocessing: Preprocess the image collected in step S7 by binarizing it using the Otsu threshold method to generate a black and white image; S8.

2. Contour detection and screening: The contours in the image processed in step S8.1 are extracted by the Suzuki contour tracking algorithm, and the roundness of the contours is calculated to screen out contours that are close to a circle; the roundness formula is: ； in, is the roundness of the identified contour, is the contour area, is the perimeter of the contour, if Less than a given contour threshold , then the recognized contour is ignored; S8.

3. Ellipse fitting and affine transformation: Perform ellipse fitting on the selected contours to obtain the center coordinates of the ellipse , major axis a, minor axis b and rotation angle, the elliptical area is converted into a perfect circle for standardization, and the affine transformation matrix M is calculated based on the four vertices of the fitted ellipse and the four vertices of the standard circle; S8.

4. Region transformation and secondary binarization: According to the center coordinates of the ellipse , and cut the area by three times the length of the major axis a and the minor axis b to include the local image where the encoding point is located ; According to the affine transformation matrix M Perform affine transformation to obtain the transformed image , enlarge the image to the standard size to obtain the transformed image of the standard size ;Will Convert to grayscale and binary images again; S8.

5. Code point detection: make The center coordinates of , set the sampling radius to: first sampling radius , the second sampling radius , the third sampling radius is ,in, is the base radius, for 1 / 3 of the radius of the largest inscribed circle; At each sampling radius, uniformly select sampling points, the coordinate calculation formula of the jth sampling point is: ； ； Where j represents the sampling point index; at each sampling point Read the pixel value and then judge Whether it is a code point, the judgment rule is Are all sample points of the radius white? Are all sample points of the radius black? The number of sampling points of the radius>2; if the above rules are met, the extracted area is considered Contains the code point, and obtains the area containing the code point; S8.

6. Encoder decoding: Extracting coding information from the area confirmed to contain the coding point and generating a unique identifier for the coding point; N points are sampled based on the third sampling radius area to extract the coding information. When sampling, the image is rotated at different angles to eliminate the influence of the angle on the coding. The coordinate calculation formula of the kth sampling point used to extract the coding information is: ； ； Where k represents the sampling point index used to extract the coding information. is the rotation angle, which is used to change the starting angle of the sampling point circle by circle; The encoding result of each circle is rotated and normalized, and then the encoding results of all circles are averaged. The average code is binarized to obtain the final binary code, which is converted into a decimal digital index as the unique identifier of the code point; S8.

7. Draw the model delivery trajectory based on encoder decoding; For the image captured by the first imaging device at time t , the image captured by the second imaging device at time t ,in , identify the sequence of image coordinates of the center of the code point of the first imaging device as , the image coordinate sequence of the center of the code point of the second imaging device is ; based on and The three-dimensional coordinate sequence of n coded points in the camera coordinate system is obtained by using the camera internal and external parameters obtained in step S6. , and the code point sequence corresponding to n code points ; According to the axis transformation matrix from the camera coordinate system to the world coordinate system calculated in step S6 , get the coordinate sequence of n coded points in the world coordinate system ; By comparing The identification sequence corresponding to the code point on the model in step S2 , and , Coordinate sequence of points of the same name in the model coordinate system ,in for A subset of based on and , calculate the rotation and translation matrix from the model coordinate system to the world coordinate system through the Genro-Kabsch algorithm ; Assume that the center point of the model in the model coordinate system is , then the coordinates of the point in the world coordinate system at time t are ; right Repeat the above steps for the image collected at any time, and you can get the coordinates of the center point of the model in the world coordinate system, and sort them out to get the coordinate sequence on the time axis , draw the trajectory according to the time axis to obtain the model delivery trajectory.

9. The method for identifying trajectory of an object dropped into a wind tunnel and visualizing shock waves according to claim 8, characterized in that: The specific implementation method of step S9 includes the following steps: S9.

1. Automatic extraction of model masks for any , the sequence of image coordinates of the coded points on the model calculated by S8 , calculate its mean coordinates ,by Centered on ,Width Extract image regions , The choice ensures Fully inclusive model; right Perform binarization, use dilation to fill holes in the target area, and combine erosion to remove small noise points; Use the Canny operator to perform edge detection to obtain all contours, and use area screening to retain the largest contour as the mask of the model. , set the model position value to 0 and the other values ​​to 1; S9.

2. Calculate the divergence of each pixel of the image; The displacement field of each pixel position of the image at time t is expressed as , with the expression: ； in, is a cross-correlation function used to measure the similarity between two windows, and the normalized cross-correlation method is preferably selected as the cross-correlation method; based on Find the divergence of the current pixel ; S9.

3. Create an adaptive divergence-preserving filter to smooth the noise: The weight adjustment is introduced into the divergence calculation to adaptively smooth the high noise area. The formula is: ； in, is the weight function, is the divergence after weight adjustment; Weight function Defined as: ； in, represents the local vector gradient strength, is the local variance of the vector field intensity, It is a hyperparameter that controls weight sensitivity, usually between 0.5 and 2.0; ； Then a smoothness constraint is added to further reduce the noise by minimizing the following expression: ； in, represents the scale factor; The optimized divergence is ; Repeat the above steps until convergence and get the optimal ; S9.

4. Process each pixel according to steps S9.1 to S9.3 to obtain the optimized divergence of each pixel. The divergence map of the current image can be obtained by sorting. , combining this image with the one in step S9.1 Get the final shock wave qualitative diagram of the bomb model at time t , the expression is: 。 10. The method for identifying trajectory of an object dropped into a wind tunnel and visualizing shock waves according to claim 9, characterized in that: The specific implementation method of step S10 is to convert the Combine the internal and external parameters of the camera and back-project them to In the image, we get the coordinates of the center point of the model. , draw a circle with a radius of 5 pixels with the image coordinates of the model center point as the center, and display the corresponding 3D coordinates above the circle in real time , superimpose the image coordinate information of the drawn circle and the model center point onto the displacement divergence image Up, yes The above steps are repeated for the images collected at any time, and then the processed images are displayed dynamically to achieve the simultaneous visualization of the two-dimensional shock wave characterization and the model trajectory.

Images