Novel high-time-resolution two-dimensional velocity profile measurement method

Abstract

The invention discloses a novel high-time-resolution two-dimensional velocity profile measurement method which is used for calculating a two-dimensional velocity profile of air where aircraft wings are located based on a line-scan digital camera recording technology. The method comprises the following steps: firstly, obtaining two sequential images of motion of spherical particles in a flow field through two rows of photosensitive elements which are fixed at intervals in a linear array camera; then, the vertical displacement amount and the time offset amount of the particles are accurately calculated through particle imaging recognition and a cross-correlation algorithm, and then the horizontal speed and the vertical speed are obtained; finally, results obtained in the two time sequence images are used for cross validation to eliminate wrong speed values, and the measurement precision is further improved. The method provides a measurement solution which is light in weight, small in size, low in cost and high in precision for calculating the speed profile of the air where the aircraft wings are located.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of air flow velocity measurement of the environment where the wing of an aircraft is located, and more particularly to a high-precision measurement method of high time resolution two-dimensional velocity profile. BACKGROUND

[0002] Currently, for the measurement of air flow velocity, commonly used measurement methods include particle image velocimetry (PIV), hot wire anemometer, etc. The basic principle of PIV is to scatter tracer particles in the fluid and illuminate them with a sheet laser, and then record the movement of the tracer particles with a camera to calculate the velocity distribution of the fluid. This method can obtain the size and direction of the velocity in a whole measurement plane. The high precision of PIV measurement mainly depends on high-performance high-speed cameras and high-energy, high-frequency lasers matched therewith. For example, Particle Image Velocimetry, Third Edition, Markus Raffel, Christian E. Willert, et al., June 2017, pages 8-15. This whole system is not only expensive (generally in the order of millions of yuan), but also bulky, making it difficult to develop into a lightweight aircraft on-board device. The hot wire anemometer method mainly uses a small powered thin wire. After being powered, the wire can generate a certain amount of heat. The faster the flowing air, the more heat it carries away, so a larger current or voltage is needed to heat the wire. After calibration, the size of the current or voltage can be monitored to infer the size of the flow field velocity. This measurement method has many shortcomings. First, the hot wire is very thin and is easily damaged during use. Moreover, temperature calibration needs to be performed constantly during measurement, which is inconvenient to use. Second, the hot wire method can only measure the absolute value of the velocity, and cannot distinguish the direction of the velocity. Third, the hot wire method can only measure the velocity value at a single location. To measure multiple locations, multiple probes need to be arranged to form an array or measurements need to be taken at different times and different locations, which limits the spatial resolution of the method.

[0003] Cameras for image acquisition can be categorized into area scan cameras and line scan cameras based on the size of their image sensors. A common area scan camera uses a rectangular array of image sensors, allowing for the capture of a single, complete image frame per shot (i.e., two-dimensional imaging in spatial scale). A line scan camera, on the other hand, uses only one or a few rows of image sensors. Its imaging relies on scanning; it can only capture a small portion of the object at a time (i.e., one-dimensional imaging in spatial scale), but with relative movement between the object and the camera, the entire object can be scanned and captured. Finally, the images captured at different times are stitched together to form a complete image frame. Compared to area scan cameras, line scan cameras offer higher spatial resolution and faster sampling rates. Summary of the Invention

[0004] This invention addresses the design requirements of confined space, lightweight miniaturization, and low cost for aircraft wings, as well as the inability to easily replace components. The invention replaces the image sensor with a linear array camera. Additionally, it utilizes spherical particle recording technology to calculate the airflow velocity around the aircraft wing. In the velocity measurement system, a novel high temporal resolution two-dimensional velocity profile measurement method is proposed, incorporating a linear array camera. By embedding velocity profile measurement into the velocity distribution measurement system (computer and software, or the aircraft wing's image processing unit), a sampling frequency of 20,000 to 30,000 Hz can be easily achieved without sacrificing spatial resolution, thereby obtaining high temporal resolution image information.

[0005] The high temporal resolution two-dimensional velocity profile measurement method of this invention first obtains two time-series images of the motion of spherical particles in a flow field using two fixed-spaced photosensitive elements in a linear array camera. Then, particle identification and cross-correlation algorithms are used to accurately calculate the particle displacement, thereby obtaining the horizontal and vertical velocities. Finally, the results obtained from the two time-series images are cross-validated to eliminate erroneous velocity values, further improving the accuracy of the velocity profile measurement. This invention provides a lightweight, miniaturized, low-cost, and high-precision velocity profile measurement solution for studying the flow velocity in the air around an aircraft wing.

[0006] This invention discloses a novel high-temporal-resolution two-dimensional velocity profile measurement method, based on spherical particle recording technology, for calculating the airflow velocity around an aircraft wing. The specific steps include:

[0007] Step 1: Acquisition of velocity profile measurement and imaging information;

[0008] The velocity profile measurement computer is used to receive image information acquired by the linear scan camera;

[0009] At least obtain the image plane Two images, such as the first image Second imaging ;

[0010] Step 2, set up a window that contains elliptical particles;

[0011] Identify the For each elliptical particle, then determine the small window containing that elliptical particle. .

[0012] Step 3: Set the size of the matching particle window;

[0013] The estimated Particles in The approximate position is determined; then, using this approximate position as the center, the location is determined. Large window of particles ;

[0014] Step 4: Calculate the delay cross-correlation of the imaging;

[0015] Will and Perform cross-correlation calculations to obtain the first image. To the second imaging Two-dimensional correlation coefficient ; correlation coefficient Two-dimensional Gaussian fitting was performed on the peak position to determine the particle origin from the first imaging. To the second imaging The displacements are respectively the first time delay First distance delay And the parameters characterizing the goodness of fit, namely the first fitting parameters. .

[0016] Step 5: Swap the roles of the two images;

[0017] (A) Identify the For each elliptical particle, then determine the small window containing that elliptical particle. .

[0018] (B) Predicting the above Particles in The approximate position is determined; then, using this approximate position as the center, the location is determined. Large window of particles ;

[0019] (C) will and Perform cross-correlation calculations to obtain the second image. To the first image Two-dimensional correlation coefficient ;

[0020] (D) Through the analysis of Two-dimensional Gaussian fitting is performed on the peak position to determine the second time delay. Second distance delay And the parameters characterizing the goodness of fit, namely the second fitting parameters. .

[0021] Step 6, particle matching for large and small windows;

[0022] Remove bad data points;

[0023] Situations that produce bad data points include:

[0024] (1) Particle matching error, that is, the particles in the large window do not correspond to the particles in the small window;

[0025] (2) No particles can be matched, that is, there are no particles in the large window.

[0026] Regarding the first scenario: assuming the first image... Particles and second imaging If the particles in the equation can be correctly matched, then there exist two time delays with equal absolute values ​​and two distance delays with equal absolute values.

[0027] If a mismatch occurs, the sum of the two time delays will deviate from 0, and the sum of the two distance delays will deviate from 0.

[0028] For the second case: if there are no particles in the large window, the obtained correlation coefficient has no peak or the peak is very small. In this case, the effect of two-dimensional Gaussian fitting is not good, so these two images are discarded.

[0029] Step 7: Calculate the horizontal and vertical velocities of the velocity profile;

[0030] After removing bad data points, if particles Still exists in the first imaging In the middle, the second image Neutron That is, with particles Matching particles, at this point, calculation and The average time delay of the particles is denoted as ,and ;

[0031] calculate and The average distance delay of the particles is denoted as ,and ;

[0032] calculate and Horizontal velocity of spherical particles and vertical velocity .

[0033] The technical advantages of the method of this invention compared to traditional methods are as follows:

[0034] (1) This invention provides a low-cost and easily miniaturized speed measurement technology for measuring the airflow velocity in the environment where the aircraft wing is located. Simply replacing the image sensor with a line scan camera, which has only one (or several) rows of pixels, allows for a size of 8000×1 (e.g., Hikvision's line scan camera). Therefore, the spatial resolution is higher. Furthermore, since each image captured by a line scan camera is only one row, the data volume is small. Without sacrificing spatial resolution, a sampling frequency of 20,000 to 30,000 Hz can be achieved, resulting in a very high temporal resolution. Traditional PIV measurement systems, which combine high-speed cameras and high-frequency lasers, often cost tens or even millions of dollars; moreover, the entire system requires additional cooling and other features, making it bulky. Replacing the high-speed camera in a traditional PIV measurement system with a line scan camera costs only 20,000 to 30,000 dollars, and the required laser can be a continuous laser, bringing the total cost of the measurement system to less than 100,000 dollars. Therefore, this invention uses an LED light source, which is not only low-cost but also easy to integrate, forming a miniaturized airborne device.

[0035] (2) A novel velocity measurement method suitable for the airflow environment of aircraft wings is proposed, utilizing... The axis represents the vertical direction of particle motion. The axis represents the imaging plane at the measurement time of the linear scan camera. Two elliptical images are formed within the image, and a two-dimensional velocity profile measurement algorithm is used under time delay and distance delay conditions.

[0036] (3) Imaging plane Elliptic imaging within the image provides high temporal and spatial resolution. High-speed cameras struggle to achieve both high temporal and spatial resolution simultaneously, as each shot is a complete frame. High spatial resolution, however, means a large data volume per frame. This poses significant challenges for data transmission and storage, hindering high frame rates (i.e., low temporal resolution). For example, a typical high-speed camera pixel array is 2048×2048, corresponding to a frame rate of only a few kilohertz. Only by reducing spatial resolution (i.e., decreasing the data volume per frame) can the frame rate of a high-speed camera be increased to tens of thousands of kilohertz. In contrast, this invention utilizes a line scan camera, obtaining images with only one (or several) rows of pixels, with a size of up to 8000×1 (e.g., Hikvision's line scan camera), resulting in higher spatial resolution. Furthermore, the smaller image data volume avoids latency during image reading and writing. Attached Figure Description

[0037] Figure 1 This is a flowchart of the novel high-temporal-resolution two-dimensional velocity profile measurement method of the present invention.

[0038] Figure 2 This is a schematic diagram of a two-dimensional velocity profile measurement scenario. In the diagram: 1-laser, 2-light source, 3-fluid domain, 4-illuminated area plane, 41-illuminated area 1, 42-illuminated area 2, 5-line scan camera, 51-the first photosensitive element in the line scan camera, i.e., the recorded image A (…). The second image sensor in the 52-line scan camera, namely the one that records image B ( ).

[0039] Figure 3 This is a schematic diagram illustrating the principle of velocity profile measurement in this invention.

[0040] Figure 4 This is a more detailed flowchart of the principle of step 2 of the present invention.

[0041] Figure 5 This is a more detailed flowchart of the principles of steps 3 and 4 of the present invention.

[0042] Figure 6 This is a graph showing the two-dimensional velocity calculation test results after processing by the method of this invention.

[0043] Figure 7 It is to intercept Figure 6 Enlarged view of part of the image. Detailed Implementation

[0044] The present invention will now be described in further detail with reference to the accompanying drawings and embodiments. The examples of the parameters listed are merely preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

[0045] See Figure 2 As shown, during the velocity measurement process in fluid dynamics, all spherical particles are in the illuminated area. It moves in a plane following the fluid. The axis represents the horizontal direction of particle motion. The axis represents the vertical direction of the particle's motion. An arbitrary spherical particle is denoted as... A set of spherical particles is denoted as The The entire motion will be recorded by the line scan camera 5. During the measurement process, the two photosensitive elements in the line scan camera 5 measure synchronously, that is, at the first moment... Both image sensors simultaneously begin capturing and outputting the first column of images, continuing until the final Nth moment. After the two image sensors capture and output the final Nth column of images, the entire imaging process ends. During the measurement process, the measurement time of the line scan camera 5 is recorded as... , This is the first measurement time. For the first At each measurement time, This is the last measurement moment.

[0046] In this invention, the velocity profile measurement system is as follows: Figure 2 The diagram shows a system consisting of a velocity profile measurement computer, a laser 1, a fluid domain 3, and a linear array camera 5. The velocity profile measurement computer refers to a device relying on computer hardware and software; the software includes at least the novel high temporal resolution two-dimensional velocity profile measurement method of this invention (installed in the computer hardware as software). The spherical particle is imaged as an elliptical particle during the velocity profile measurement calculation process, such as... Figure 2 , Figure 4 and Figure 5 Laser 1 is used to generate sheet light source 2 and illuminate the fluid domain 3. Illumination area plane 4 has a first illumination area 41 and a second illumination area 42. Linear scan camera 5 has at least a first photosensitive element 51 and a second photosensitive element 52. The spherical particles used in fluid domain 3 are used to obtain the motion state of spherical particles in the fluid domain (i.e., air), and the velocity of the fluid is obtained from the velocity of the spherical particles.

[0047] See Figure 2 and Figure 3 As shown, in this invention, the area of ​​illumination... The plane corresponds to the imaging plane of the line scan camera 5, that is... Image plane, The axis represents the vertical direction of particle motion. The axis represents the measurement time when the line scan camera 5 captures each column of images.

[0048] Linear scan camera 5 is used to acquire imaging information of spherical particles in a fluid domain. This imaging information is obtained by scanning and capturing different parts of the particle at different times using the first photosensitive element 51 and the second photosensitive element 52. See also Figure 3 The present invention uses one of the particles Let's take an example to explain the imaging principle of the two image sensors in a line scan camera. In reality, the motion of spherical particles is determined by the actual flow field. For simplicity, we'll assume the particles are spherical here. Following the flow field, it moves from the upper left to the lower right, passing through the imaging range of the first photosensitive element 51 and the second photosensitive element 52. (Particles) At any moment It was captured by the first image sensor 51 in the line scan camera 5, and then continued to move until the time... Time particles It has moved out of the shooting range of the first photosensitive element 51. Combining the image information captured by the first photosensitive element 51 in chronological order yields the particle image. First imaging An ellipse in the image plane forms an image. Then the particles... Continue moving, enter the imaging range of the second image sensor 52, and repeat the same process in the second imaging... An ellipse is formed in the image plane to create an image. First imaging. Image plane and second imaging Both elliptical particles in the image plane are imaged by the same spherical particle. The resulting image consists of two elliptical particle pairs that can be successfully matched in the algorithm.

[0049] Throughout the entire process of the linear array camera, the spherical particle collection Different parts of the particles at different times The image information is captured by the first photosensitive element 51 and the second photosensitive element 52 of the line scan camera 5. Combining the image information captured by the first photosensitive element 51 in chronological order yields the image of all particles. exist An elliptical image is formed in the image plane, denoted as the first image. ,like Figure 2 As shown in the lower left image, combining the image information captured by the second photosensitive element 52 in chronological order yields the particle image. exist An ellipse forming an image in the image plane is denoted as the second image. ,like Figure 2 As shown in the image in the bottom right corner.

[0050] Because the two photosensitive elements of the line scan camera 5 are spaced apart by a small distance (the distance is the distance between a series of pixels), and this distance is along the horizontal direction of particle movement, it is denoted as... The distance between the first photosensitive element 51 and the second photosensitive element 52 In this case, after any particle leaves the shooting range of the first photosensitive element 51, it needs to move for a while before it can be captured by the second photosensitive element 52. That is, there is a short time delay, denoted as . Meanwhile, during this period The particle also moves a distance along the Y-axis, denoted as the distance delay. .

[0051] First image obtained directly using line scan camera 5 Second imaging Compared to velocity profile measurement computers, a second imaging technique emerged. Neutron position and first imaging The particle's position is offset along both the Y-axis and T-axis. These offsets are then calculated. and This is the purpose of the present invention. By adjusting the offset ( and The correction can easily achieve a resolution of 20,000 to 30,000 Hz without sacrificing spatial resolution, thus obtaining high temporal resolution image information. On the other hand, it only requires replacing the original image sensor in the narrow space of the aircraft wing with a small linear array camera. The third aspect is that software for high temporal resolution two-dimensional velocity profile measurement method has been added to the aircraft's image processing.

[0052] Adopting such Figure 1 The procedure shown is for the first imaging Second imaging The steps involved in processing high-temporal-resolution two-dimensional velocity profile measurement methods are as follows:

[0053] Step 1: Acquisition of velocity profile measurement and imaging information;

[0054] The velocity profile measurement computer receives image information acquired by the linear array camera 5.

[0055] For ease of explanation, let's select two images; image plane The first image in Second imaging ;

[0056] The imaging plane of the line scan camera 5 is denoted as , The axis represents the vertical direction of particle motion. The axis represents the measurement time during the linear array camera measurement. The spatial plane of motion of the spherical particle is denoted as... The X-axis represents the horizontal direction of particle motion. Spherical particles in the plane were recorded An elliptical particle image will be displayed after the plane is turned on.

[0057] Step 2, set up a window that contains elliptical particles;

[0058] In order to identify the For each elliptical particle in the array, a small window is set to determine the area containing that elliptical particle, denoted as . See details for the steps.Figure 4 .

[0059] Figure 4 yes Figure 1 Step 2: Get the small window The schematic diagram. First, the image plane... The first image in Binarization is performed, treating connected bright pixels as a cluster of particles. Then, the centroid positions of the connected regions of these particles are found. This is used as the particle's position. Then, a minimal ellipse is used to cover the connected region, and the major axis *a*, minor axis *b*, and the angle between the major axis and the horizontal line are calculated. This ellipse has two uses: (i) utilizing and ellipse parameters, in (i) Obtain a rectangular window containing the particle image (i.e., the original image, not the binarized image); (ii) Use the ellipse to obtain a particle mask, that is, set the pixels inside the ellipse to 1 and the pixels outside the ellipse to 0. The purpose of setting the mask is: when the particle concentration is high, the obtained rectangular window may capture incomplete images of other particles in addition to the complete particle image, which will affect the calculation results. By setting the mask, these interfering images can be removed. Therefore, multiplying the rectangular window by the mask at the end will give the final window. .

[0060] Step 3: Set the size of the matching particle window;

[0061] from Figure 3 As can be seen from the imaging principle diagram, under ideal conditions, One of the many elliptical particle images should be related to the stated The particles in the image are matched. From these two matched elliptical particle image pairs, the first time delay can be accurately calculated. and first distance delay Therefore, it is necessary to The middle determines that it includes the above. The large window of particles that match the particles in the middle is denoted as .

[0062] Sure Large window of particles Two elements are needed: the length and width of the large window and the center position of the large window.

[0063] In this invention, the size of the large window is W times larger than the size of the small window. W should not be too large; generally, 2 to 5 times is suitable to avoid excessively increasing the computational load. The large window... The center position can be seen from the small window The position is estimated from the elliptical particle information. The method for estimating the position is described in [link to method]. Figure 5 .

[0064] Step 4: Calculate the delay cross-correlation of the imaging;

[0065] Will and Perform cross-correlation calculations to obtain the first image. To the second imaging Two-dimensional correlation coefficient ; correlation coefficient Two-dimensional Gaussian fitting was performed on the peak position to determine the particle origin from the first imaging. To the second imaging The displacements are respectively the first time delay First distance delay And the parameters characterizing the goodness of fit, namely the first fitting parameters. .

[0066] Figure 5 yes Figure 1 The diagram illustrates the calculation principle of combining Step 3 and Step 4. The method of this invention requires identifying the specific parameters for the image plane. The first image in The position of the center of mass of a certain particle It is in the second imaging What is its approximate location? (Estimation) Large window of particles The principle of the central position is as follows. First, the present invention can be derived from... The horizontal velocity of the elliptical particle can be roughly estimated from its size information. and vertical velocity .from Figure 2 As can be seen from the imaging principle of the linear scan camera in the image, a particle appears to move at a time interval. The particle moved horizontally a distance equal to the diameter of a particle. The particle can directly observe the properties of the ellipse, including its major axis a, minor axis b, and the angle of inclination of the major axis. The particle diameter was calculated. Therefore, the horizontal velocity of the particle can be estimated. The vertical velocity is Secondly, the movement of the particle from the first photosensitive element 51 to the second photosensitive element 52 (i.e., horizontal movement) can be further estimated. Time required for distance And the distance the particle travels in the Y direction during this time. Finally, combining this particle in The position of the centroid in It is possible to estimate the particle's position. The approximate location corresponding to the middle is ( , Then, this can be used as the center position of the large window to obtain the large window. .

[0067] Will and Perform cross-correlation calculations to obtain the first image. To the second imaging Two-dimensional correlation coefficient As shown in the figure, this two-dimensional correlation coefficient There is a peak value. In this invention, a two-dimensional Gaussian function is used to fit the correlation coefficient. The peak value in the image is used to obtain the particle from the first imaging. To the second imaging The precise displacement amounts are respectively the first time delay First distance delay And the parameters characterizing the goodness of fit, namely the first fitting parameters. .

[0068] Step 5: Swap the roles of the two images;

[0069] (A) Identify the For each elliptical particle, then determine the small window containing that elliptical particle. .

[0070] (B) Predicting the above Particles in The approximate position is determined; then, using this approximate position as the center, the location is determined. Large window of particles ;

[0071] (C) will and Perform cross-correlation calculations to obtain the second image. To the first image Two-dimensional correlation coefficient ;

[0072] (D) Through the analysis of Two-dimensional Gaussian fitting is performed on the peak position to determine the second time delay. Second distance delay And the parameters characterizing the goodness of fit, namely the second fitting parameters. .

[0073] Step 5 of this invention is interchangeability. and The role is to repeat Steps 2 through 4 to obtain particles from the surface. arrive The amount of displacement.

[0074] Step 6, particle matching for large and small windows;

[0075] Remove bad data points;

[0076] Situations that produce bad data points include:

[0077] (1) Particle matching error, that is, the particles in the large window do not correspond to the particles in the small window;

[0078] (2) No particles can be matched, that is, there are no particles in the large window.

[0079] Regarding the first scenario: assuming the first image... Particles and second imaging If the particles in the array can be correctly matched, then there exist two absolute values ​​of time delay and two absolute values ​​of distance delay that are equal. Numerically, this is represented by equal time delays. Equal distance delay Due to the acquisition of the second image In the first imaging Delay one The acquired image information, therefore, when applying numerical representation, is the first imaging... To the second imaging A gap of one (e.g., a positive 1 second), while the second image To the first image A gap of one (e.g., negative 1 second), therefore the second image Compared to the first imaging The "-" sign was used.

[0080] If a mismatch occurs, the sum of the two time delays will deviate from 0, and the sum of the two distance delays will also deviate from 0. Numerically, this is expressed as the sum of the time delays being... The sum of distance delays is Therefore, 3 can be used under the condition of mismatch. The principle (i.e., the three sigma principle) is used to identify unmatched particles that deviate too much from 0, and the corresponding unmatched particle data results are discarded as bad data points.

[0081] The 3 The principle is based on the characteristics of a normal distribution, which is used to remove outliers of mismatched particles in two images.

[0082] For the second scenario: if there are no particles in the large window, the obtained correlation coefficient will have no peak or a very small peak. In this case, the two-dimensional Gaussian fitting will not perform well, and the first parameter... and / or the second parameter It is very small, so this parameter can be used to discard two images.

[0083] Step 7: Calculate the horizontal and vertical velocities of the velocity profile;

[0084] Indicates the first imaging The identifier collected within the window is One of the particles. Indicates the second imaging The identifier collected within the window is A particle. With particles The particles were collected by different photosensitive elements of a linear array camera at different measurement times, hence the two particle identification numbers.

[0085] After removing bad data points, if particles Still exists in the first imaging In the middle, and the second imaging Neutron It is a particle Matching particles, at this point, calculation and The average time delay of the particles is denoted as ,and ;

[0086] calculate and The average distance delay of the particles is denoted as ,and ;

[0087] calculate and Horizontal velocity of spherical particles and vertical velocity .

[0088] With particles and particles Examples are provided to help distinguish the numbering of particles in different images.

[0089] In this invention, using Step 1 to Step 7, the velocity profile measurement computer will complete all the images acquired by the line scan camera 5. Example

[0090] Figure 6 andFigure 7 These are simulation test results of the high temporal resolution two-dimensional velocity profile measurement method according to the present invention. Figure 7 It is to intercept Figure 6 Enlarged view of the area within the dashed box. Within the 1024×1024 pixel illuminated area. Particles are randomly arranged in a plane under the following conditions: particle position is randomly set, particle diameter is randomly set within 3 to 6 pixels, and particle center brightness is randomly set within 150 to 200 pixels, then the brightness decreases Gaussian along the periphery. After setting, these particles move at a uniform horizontal velocity. pixel / s, vertical velocity pixel / s In-plane motion. Then, the pixel positions of columns 820 and 824 are selected to simulate the shooting positions of the first image sensor 51 and the second image sensor 52 of the line scan camera 5. The image information extracted from these two positions is then arranged in chronological order to obtain... Two images of the image plane. The images were captured at a frequency of 1000 Hz for 1 second. Finally, the velocity of the flow field was obtained from these two images using a high-time-resolution two-dimensional velocity profile measurement method, such as... Figure 6 The blue arrow indicates the speed calculated using the method of this invention, compared to a known reference speed. (Right now Figure 6 The gray arrow in the image matches perfectly. According to the formula... The calculations show that the error in the horizontal velocity is 0.008 pixels / pixel, and the error in the vertical velocity is 0.027 pixels / pixel. These results preliminarily confirm the feasibility of the above algorithm.

Claims

1. A novel high-temporal-resolution two-dimensional velocity profile measurement method, characterized in that: First, two time-series images of the motion of spherical particles in the flow field are obtained using two fixed-spaced photosensitive elements in a linear array camera. Then, particle identification and cross-correlation algorithms are used to accurately calculate the displacement of the particles, thereby obtaining the horizontal and vertical velocities. Finally, the results obtained from the two time-series images are cross-validated to eliminate erroneous velocity values, further improving the accuracy of velocity profile measurement.

2. The high temporal resolution two-dimensional velocity profile measurement method according to claim 1 is based on spherical particle recording technology for calculating the airflow velocity in the air surrounding an aircraft wing, characterized in that... It includes the following steps: Step 1: Acquisition of velocity profile measurement and imaging information; The velocity profile measurement computer is used to receive image information acquired by the linear scan camera; At least obtain the image plane Two images, such as the first image Second imaging ; Step 2, set up a window that contains elliptical particles; Identify the For each elliptical particle, then determine the small window containing that elliptical particle. ; Step 3: Set the size of the matching particle window; The estimated Particles in The approximate position is determined; then, using this approximate position as the center, the location is determined. Large window of particles ; Step 4: Calculate the delay cross-correlation of the imaging; Will and Perform cross-correlation calculations to obtain the first image. To the second imaging Two-dimensional correlation coefficient ; correlation coefficient Two-dimensional Gaussian fitting was performed on the peak position to determine the particle origin from the first imaging. To the second imaging The displacements are respectively the first time delay First distance delay And the parameters characterizing the goodness of fit, namely the first fitting parameters. ; Step 5: Swap the roles of the two images; (A) Identify the For each elliptical particle, then determine the small window containing that elliptical particle. ; (B) Predicting the above Particles in The approximate position is determined; then, using this approximate position as the center, the location is determined. Large window of particles ; (C) will and Perform cross-correlation calculations to obtain the second image. To the first image Two-dimensional correlation coefficient ; (D) Through the analysis of Two-dimensional Gaussian fitting is performed on the peak position to determine the second time delay. Second distance delay And the parameters characterizing the goodness of fit, namely the second fitting parameters. ; Step 6, particle matching for large and small windows; Remove bad data points; Situations that produce bad data points include: (1) Particle matching error, that is, the particles in the large window do not correspond to the particles in the small window; (2) No particles can be matched, that is, there are no particles in the large window; Regarding the first scenario: assuming the first image... Particles and second imaging If the particles in the equation can be correctly matched, then there exist two time delays with equal absolute values ​​and two distance delays with equal absolute values. If a mismatch occurs, the sum of the two time delays will deviate from 0, and the sum of the two distance delays will also deviate from 0. For the second case: if there are no particles in the large window, the obtained correlation coefficient has no peak or the peak is very small. In this case, the two-dimensional Gaussian fitting effect is not good, so these two images are removed. Step 7: Calculate the horizontal and vertical velocities of the velocity profile; After removing bad data points, if particles Still exists in the first imaging In the middle, the second image Neutron That is, with particles Matching particles, at this point, calculation and The average time delay of the particles is denoted as ,and ; Indicates the first imaging The identifier collected within the window is One particle; Indicates the second imaging The identifier collected within the window is One particle; calculate and The average distance delay of the particles is denoted as ,and ; calculate and Horizontal velocity of spherical particles and vertical velocity .

3. The high temporal resolution two-dimensional velocity profile measurement method according to claim 2, characterized in that: Get the small window in Step 2 The principle is to first divide the image plane The first image in Binarization is performed, clustering connected bright pixels as a single particle; then, the centroid of the connected region of the particle is found and used as the particle's position; next, the connected region is covered with a minimal ellipse, and the major axis a, minor axis b, and the angle between the major axis and the horizontal line of the ellipse are calculated. By setting a mask to remove interfering images, and finally multiplying the rectangular window by the mask, the final small window can be obtained. .

4. The high temporal resolution two-dimensional velocity profile measurement method according to claim 2, characterized in that: Estimate Large window of particles The principle of the center position is that, firstly, from The horizontal velocity of the elliptical particle can be roughly estimated from its size information. and vertical velocity Secondly, the time required for the particle to move from the first photosensitive element to the second photosensitive element was estimated. And the distance the particle travels in the Y direction during this time. Finally, combining this particle in The position of the centroid in The particle was calculated to be in The approximate location corresponding to the middle is ( , ); then, this can be used as the center position of the large window to obtain the large window. .

5. A velocity profile measurement system for implementing the high time resolution two-dimensional velocity profile measurement method according to any one of claims 1-4, characterized in that: The velocity profile measurement system consists of a velocity profile measurement computer, a laser (1), a fluid domain (3), and a line array camera (5). The velocity profile measurement computer refers to a device that relies on computer hardware and software and is equipped with a new high-time-resolution two-dimensional velocity profile measurement method. The spherical particles are imaged as elliptical particles during the velocity profile measurement calculation process. The laser (1) is used to generate a sheet light source (2) and illuminate the fluid domain (3). The illumination area plane (4) is provided with an illumination first area (41) and an illumination second area (42). The line array camera (5) has at least a first photosensitive element (51) and a second photosensitive element (52). The spherical particles used in the fluid domain (3) are used to obtain the motion state of the spherical particles in the fluid domain, and the fluid movement speed is obtained by the motion speed of the spherical particles.

Images