Wavefront Distortion Detection Method Based on the Similarity of Continuous Image Sequences in a High-Speed Flow Field

Abstract

The present invention relates to a method for detecting wavefront distortion based on the similarity of continuous image sequences in a high-speed flow field, belonging to the field of data processing, and includes: decomposing the collected video data into a sequence of pictures; preprocessing each picture, obtaining the centroid of each light spot in each picture to obtain the centroid coordinate sequence of each picture, taking the centroid coordinates of the first picture as the calibration reference centroid, and simultaneously obtaining the initial restoration matrix; performing centroid matching on two adjacent pictures before and after to obtain the slope vector and the correction parameter, and updating the offset reference centroid coordinate sequence; substituting the slope vector and the correction parameter into the restoration equation to obtain the Zernike coefficients and complete the wavefront restoration; repeating the above steps to realize the wavefront restoration of the entire video data. The present invention can better face complex situations such as large tilt distortion and missing light spot data, realize the detection and restoration of the wavefront, and is of great significance for improving the effective detection of the distorted complex wavefront by the Hartmann wavefront sensor.

Description

Technical Field

[0001] The present invention provides a method for detecting wavefront distortion based on the similarity of continuous image sequences in a high-speed flow field, belonging to the technical field of data processing for beam wavefront measurement. Background Art

[0002] The Hartmann wavefront sensor is a precision optical measurement device. It uses a microlens array to divide the incident light wavefront into multiple sub-wavefronts, and determines the shape of the wavefront by measuring the position changes of the sub-spots formed by these sub-wavefronts on the focal plane. This technology can monitor wavefront distortion in real time and perform correction in dynamic optical systems, featuring high precision and high real-time performance. The non-contact measurement method of the Hartmann wavefront sensor avoids physical interference with optical elements, making it particularly important in optical element testing. In addition, due to its strong adaptability, the Hartmann wavefront sensor can be applied to various wavelengths and different types of optical systems.

[0003] The origin of the Hartmann wavefront sensor can be traced back to the early 20th century and was first proposed by the German physicist Johann Karl Friedrich Hartmann. Hartmann published a paper on wavefront measurement in 1910, describing a method of using a microlens array to measure the wavefront of starlight. This original Hartmann wavefront sensor was designed for the astronomy field to measure and correct the wavefront distortion of starlight caused by atmospheric turbulence. Over time, the design and application of the Hartmann wavefront sensor have been continuously improved and extended. In the 1970s, with the development of computer technology, the data processing ability of the wavefront sensor was significantly improved, enabling the Hartmann wavefront sensor to be more widely used in optical testing and adaptive optical systems. By the 1990s, with the development of microelectromechanical system (MEMS) technology, the miniaturization and integration of the Hartmann wavefront sensor became possible, further promoting its application in the field of precision optical measurement.

[0004] The application fields of the Hartmann wavefront sensor are very extensive, including but not limited to atmospheric turbulence correction in adaptive optical systems, vision correction in ophthalmology, and starlight wavefront distortion correction in astronomical telescope systems. These applications demonstrate the importance of the Hartmann wavefront sensor in improving imaging quality and optimizing the performance of optical systems. With the optimization of algorithms and the progress of hardware technology, the measurement speed and accuracy of the Hartmann wavefront sensor have been significantly improved, enabling it to better adapt to high-speed dynamic environments and complex optical measurement requirements.

[0005] At present, the development directions of Hartmann wavefront sensors mainly include: improving the dynamic range, enhancing the ability to process missing data, and improving the detection accuracy, etc. Among them, the dynamic range and the ability to process missing data are particularly important when facing complex distorted wavefronts. However, the existing detection methods hardly take both aspects into account. Therefore, there is an urgent need for a wavefront distortion detection algorithm based on the similarity of continuous image sequences, which can take into account both the dynamic range and the ability to process missing data. Summary of the Invention

[0006] Aiming at the deficiencies of the existing technology, the present invention provides a wavefront distortion detection method based on the similarity of continuous image sequences in a high-speed flow field, which can better face complex situations such as large tilt distortion and missing spot data, realize the detection and restoration of the wavefront, and is of great significance for improving the effective detection of distorted complex wavefronts by Hartmann wavefront sensors.

[0007] The technical solution of the present invention is as follows:

[0008] A wavefront distortion detection method based on the similarity of continuous image sequences in a high-speed flow field, comprising the following steps:

[0009] (1) Decompose the video data collected by the CCD into a picture sequence to obtain the original data;

[0010] (2) Preprocess each picture, calculate the centroid of each light spot in each picture to obtain the centroid coordinate sequence of each picture, use the centroid coordinates of the first picture as the calibration reference centroid, and at the same time obtain the initial restoration matrix E;

[0011] (3) Perform centroid matching on two adjacent pictures before and after to obtain the slope vector and the correction parameter, and update the offset reference centroid coordinate sequence;

[0012] (4) Substitute the slope vector and the correction parameter into the restoration equation to calculate the Zernike coefficients and complete the wavefront restoration;

[0013] (5) Match the next picture in the picture sequence with the updated offset reference centroid coordinate sequence, that is, repeat steps (3) to (4) to realize the wavefront restoration of the entire video data.

[0014] Preferably, step (2) includes the following steps:

[0015] (2.1) Perform image binarization and morphological operations to obtain the approximate positions of the light spots;

[0016] The morphological operation refers to opening operation and closing operation. The approximate position of the light spot is one of the morphological operations, which can directly obtain the centroid of each light spot (each independent region in morphology);

[0017] (2.2) Use the centroid formula to obtain the exact position of the light spot (a picture contains multiple light spots, and the centroid of the light spot corresponding to each sub-aperture needs to be calculated). The centroid formula is:

[0018]

[0019] where (X c , Y c ) are the centroid coordinates of the light spot, M represents the number of pixels in the sub-aperture area, I i is the pixel value of the i-th pixel, and (x i , y i ) represent the pixel coordinates within the sub-aperture; after finding the centroid of the light spot corresponding to all sub-apertures in the first picture, the calibrated reference centroid can be obtained;

[0020] (2.3) By comparing the centroid coordinates of the latter picture with the calibrated reference centroid coordinates, the slope vector G can be obtained. The slope vector G contains the horizontal light spot offset G x and the vertical light spot offset G x of each sub-aperture, which is expressed as:

[0021]

[0022] where (X t , Y t ) represent the centroid coordinates obtained from the latter picture, and f is the focal length of the sub-aperture;

[0023] (2.4) Use Φ(x, y) to represent the incident wavefront phase. The wavefront phase can be expanded by Zernike polynomials as:

[0024]

[0025] where Z m (x, y) is the m-th Zernike polynomial, a m is its coefficient. When m = 1, it is the average value of the wavefront phase, which is usually ignored and recorded as 0;

[0026] Take the derivative of both sides of the Zernike polynomial. Then the wavefront restoration equation is expressed in matrix form as:

[0027] G = EA

[0028] where G represents the slope vector, with a size of 2k × 1; A = {a2 a3…a m} represents the coefficient vector; E represents the restoration matrix,

[0029]

[0030] The size of the restoration matrix is 2k×(M - 1), and each term represents the average wavefront slope when the corresponding Zernike coefficient is 1 on the corresponding sub-aperture. k is the number of sub-apertures; (Z xM ) k represents the average wavefront slope in the horizontal direction at the k-th sub-aperture in the M-th Zernike polynomial, (Z yM ) k represents the average wavefront slope in the vertical direction at the k-th sub-aperture in the M-th Zernike polynomial;

[0031] (2.5) Calculate the generalized inverse matrix E + of E, then the coefficient vector A can be obtained. Substituting it into the Zernike polynomial, the wavefront phase Φ(x, y) can be calculated.

[0032] Preferably, step (3) includes the following steps:

[0033] (3.1) Use the centroid of the previous picture as the offset reference centroid and match it with the centroid of each subsequent picture (each centroid of the previous picture is matched with each centroid of the subsequent picture, that is, any centroid is taken from the previous picture and the subsequent picture for pairwise matching). The initial value of the offset reference centroid is the calibrated reference centroid of the first picture, and the coordinate difference is expressed as:

[0034] D x = X H - X P

[0035] D y = Y H - Y P

[0036] where, (X H , Y H ) represents the centroid coordinates of the subsequent picture, and (X P , Y P ) represents the offset reference centroid coordinates, that is, the centroid used as a reference in the matching step;

[0037] The spot distance Dist can be expressed as:

[0038]

[0039] (3.2) If Dist < DSpots / 3, it is considered that the two light spots at this time are the spots of the same sub-aperture. Then, the spot offset can be calculated using the centroid of the subsequent frame and the calibrated reference centroid corresponding to the matched offset reference centroid, where DSpots represents the average spot distance of the calibrated reference centroid and is a fixed value;

[0040] Calculate the slope vector, and replace the centroid coordinates of the latter image with the centroid of the offset reference centroid at this sub-aperture, that is, use the centroid of the latter image as the new offset reference centroid;

[0041] (3.3) If Dist ≥ DSpots / 3, it is considered that the offset reference centroid fails to match the centroid of the latter image, then it is considered that the data of this sub-aperture is missing, and mark its sub-aperture serial number, that is, the correction parameter.

[0042] The offset reference centroid refers to: for the first frame for which restoration calculation is performed, its offset reference centroid comes from the result of calibration calculation. Due to the spot flickering, not every reference centroid can find its corresponding centroid of the latter frame in the restored frame. At this time, the reference centroid needs to be used to update the missing centroid of the latter frame and used as the offset reference centroid for the next frame for which restoration calculation is performed, while the spots that can be matched follow the centroid of the latter frame.

[0043] Preferably, step (4) includes the following steps:

[0044] (4.1) The slope vector G calculated through step (3) only includes the successfully matched centroids; the centroids that are not successfully matched retain the vector I composed of their serial numbers, and delete the data in the restoration matrix E through the vector I to obtain E′;

[0045] (4.2) By solving the equation

[0046] G = E′A

[0047] Obtain the coefficient vector A;

[0048] (4.3) Substitute the obtained coefficient vector A into the wavefront phase expression:

[0049]

[0050] That is, the restored wavefront phase is obtained.

[0051] For the parts not elaborated in the present invention, reference can be made to the prior art.

[0052] The beneficial effects of the present invention are:

[0053] 1. The present invention realizes the dynamic wavefront data processing of large tilt and complex distortion, realizes the dynamic wavefront detection in this case, and this algorithm improves the detection dynamic range of the Hartmann sensor.

[0054] 2. The present invention solves the problem of wavefront detection under the conditions of large tilt, complex distortion and spot missing, and improves the processing performance of the Hartmann wavefront sensor.

[0055] 3. The method proposed by the present invention has good scalability. During operation, parameters such as beam quality index and stability index are involved in the calculation, and corresponding functions can be extended according to requirements. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] The accompanying drawings forming a part of this application are used to provide a further understanding of this application. The illustrative embodiments of this application and their descriptions are used to explain this application and do not constitute an improper limitation to this application.

[0057] Figure 1 It is the overall flowchart of the wavefront distortion detection method based on the similarity of continuous image sequences in a high-speed flow field in the present invention;

[0058] Figure 2 It is the centroid matching flowchart in step (3) of the present invention;

[0059] Figure 3 It is a schematic diagram of the centroid matching process, where (a) is the offset reference centroid and (b) is a certain centroid matching process;

[0060] Figure 4 It is the calibrated reference centroid distribution;

[0061] Figure 5 It is the offset centroid distribution;

[0062] Figure 6 It is according to Figure 4 and Figure 5 The calculated Zernike coefficient distribution;

[0063] Figure 7 It is according to Figure 4 Figure 5 The calculated wavefront restoration diagram. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0064] In order to enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of this specification. However, it is not limited thereto. For those parts not elaborated in the present invention, the conventional techniques in the art are adopted.

[0065] Embodiment 1

[0066] A wavefront distortion detection method based on the similarity of continuous image sequences in a high-speed flow field, as Figure 1 , includes the following steps:

[0067] (1) Decompose the video data collected by the CCD into a picture sequence to obtain the original data;

[0068] (2) Preprocess each image, calculate the centroid of each light spot in each image to obtain the centroid coordinate sequence of each image, use the centroid coordinates of the first image as the calibration reference centroid, and simultaneously obtain the initial restoration matrix E;

[0069] (3) Use two adjacent images before and after for centroid matching to obtain the slope vector and correction parameters, and update the offset reference centroid coordinate sequence;

[0070] (4) Substitute the slope vector and correction parameters into the restoration equation to calculate the Zernike coefficients and complete the wavefront restoration;

[0071] (5) Match the next image in the image sequence with the updated offset reference centroid coordinate sequence, that is, repeat steps (3) to (4) to achieve the wavefront restoration of the entire video data.

[0072] Embodiment 2

[0073] A method for detecting wavefront distortion based on the similarity of continuous image sequences in a high-speed flow field, as described in Embodiment 1, except that step (2) includes the following steps:

[0074] (2.1) Perform image binarization and morphological operations to obtain the approximate positions of the light spots;

[0075] The morphological operations refer to opening operation and closing operation. The approximate position of the light spot is one of the morphological operations, and it can directly obtain the centroid of each light spot (each independent region in morphology);

[0076] (2.2) Use the centroid formula to calculate the accurate positions of the light spots (a single image includes multiple light spots, and the centroid of the light spot corresponding to each sub-aperture needs to be calculated). The centroid formula:

[0077]

[0078] where, (X c , Y c ) are the centroid coordinates of the light spot, M represents the number of pixels in the sub-aperture region, I i is the pixel value of the i-th pixel, and (x i , y i ) represent the pixel coordinates within the sub-aperture; after calculating the centroids of the light spots corresponding to all sub-apertures in the first image, the calibration reference centroid can be obtained;

[0079] (2.3) By comparing the centroid coordinates of the next image with the calibration reference centroid coordinates, the slope vector G can be obtained. The slope vector G contains the horizontal light spot offset G x and the vertical light spot offset G x of each sub-aperture, which is expressed as:

[0080]

[0081] Among them, (X t , Y t ) represents the centroid coordinates obtained from the latter image, and f is the focal length of the sub-aperture;

[0082] (2.4) Use Φ(x, y) to represent the incident wavefront phase, and the wavefront phase can be expanded by the Zernike polynomial as:

[0083]

[0084] Among them, Z m (x, y) is the m-th Zernike polynomial, a m is its coefficient. When m = 1, it is the average value of the wavefront phase, which is usually ignored and recorded as 0;

[0085] Derive both sides of the Zernike polynomial. Then the wavefront restoration equation is expressed in matrix form as:

[0086] G = EA

[0087] Among them, G represents the slope vector, with a size of 2k × 1; A = {a2 a3... a m} represents the coefficient vector; E represents the restoration matrix,

[0088]

[0089] The size of the restoration matrix is 2k × (M - 1), and each term of it represents the average wavefront slope corresponding to the Zernike coefficient of 1 on the corresponding sub-aperture. k is the number of sub-apertures; (Z xM ) k represents the horizontal average wavefront slope at the k-th sub-aperture in the M-th Zernike polynomial, and (Z yM ) k represents the vertical average wavefront slope at the k-th sub-aperture in the M-th Zernike polynomial;

[0090] (2.5) Find the generalized inverse matrix E + of E, then the coefficient vector A can be obtained, and substituting it into the Zernike polynomial can calculate the wavefront phase Φ(x, y).

[0091] Example 3

[0092] A method for detecting wavefront distortion based on the similarity of continuous image sequences in a high-speed flow field, as described in Example 2. The difference is that step (3) includes the following steps:

[0093] (3.1) Use the centroid of the previous image as the offset reference centroid and match it with each centroid of the subsequent image (each centroid of the previous image is matched with each centroid of the subsequent image, that is, any centroid is randomly selected from the previous image and the subsequent image for pairwise matching). The initial value of the offset reference centroid is the calibrated reference centroid of the first image, and the coordinate difference is expressed as:

[0094] D x =X H -X P

[0095] D y =Y H -Y P

[0096] Among them, (X H ,Y H ) represents the centroid coordinates of the subsequent image, and (X P ,Y P ) represents the offset reference centroid coordinates, that is, the centroid used as a reference in the matching step;

[0097] The spot distance Dist can be expressed as:

[0098]

[0099] (3.2) If Dist < DSpots / 3, it is considered that the two light spots at this time are the spots of the same sub-aperture. Then, the spot offset can be calculated by using the centroid of the subsequent frame and the calibrated reference centroid corresponding to the matched offset reference centroid, where DSpots represents the average spot distance of the calibrated reference centroid and is a fixed value;

[0100] Calculate the slope vector and replace the centroid of the offset reference centroid at this sub-aperture with the centroid coordinates of the subsequent image, that is, use the centroid of the subsequent image as the new offset reference centroid;

[0101] (3.3) If Dist ≥ DSpots / 3, it is considered that the offset reference centroid fails to match the centroid of the subsequent image, then it is considered that the data of this sub-aperture is missing, and its sub-aperture serial number is marked, that is, the correction parameter.

[0102] The offset reference centroid refers to: for the first frame for restoration calculation, its offset reference centroid comes from the result of calibration calculation. Due to spot flickering, not every reference centroid can find the corresponding centroid of the subsequent frame in the restored frame. At this time, the reference centroid needs to be used to replace and update the missing centroid of the subsequent frame, which is used as the offset reference centroid for the next frame for restoration calculation, while the matched spots continue to use the centroid of the subsequent frame.

[0103] Example 4

[0104] A method for detecting wavefront distortion based on the similarity of continuous image sequences in a high-speed flow field. As described in Embodiment 3, the difference is that step (4) includes the following steps:

[0105] (4.1) Through the slope vector G calculated in step (3), only the centroids that have successfully matched are included; the centroids that have not successfully matched retain the vector I composed of their serial numbers. By deleting the data in the restoration matrix E through the vector I, E′ is obtained;

[0106] For example: for a certain frame undergoing restoration calculation, if it is found through centroid matching that the first spot data of this frame is missing, then the corresponding restoration matrix removes the data of the first sub-aperture:

[0107]

[0108] Its size is changed to 2(k - 1)×(M - 1). And the size of G becomes 2(k - 1)×1.

[0109] (4.2) By solving the equation

[0110] G = E′A

[0111] The coefficient vector A is obtained;

[0112] (4.3) Substitute the obtained coefficient vector A into the wavefront phase expression:

[0113]

[0114] That is, the restored wavefront phase is obtained.

[0115] Embodiment 5

[0116] A method for detecting wavefront distortion based on the similarity of continuous image sequences in a high-speed flow field, as Figure 1 shown in the flowchart. This method first decomposes the dynamic spot distribution video into a picture sequence, and each picture records different spot position distributions; after preprocessing, the centroid coordinate sequence of each picture is obtained, and the centroid of the first picture is used as the calibration reference centroid to calculate the initial restoration matrix.

[0117] Then, centroid matching operation is performed, and its process is as Figure 2 shown, and the process schematic diagram is as Figure 3 shown, where the green dots represent the offset reference centroid distribution (actual situation as Figure 4 ), the red dots represent the offset centroid distribution (actual situation as Figure 5 ), the black area represents the sub-aperture range, and sub-apertures 8 and 11 represent the situation of missing spots. Finally, the slope vector calculated from the matched light spots and the correction parameters of the unmatched light spots are obtained, and the offset reference centroid is updated according to the matching situation for matching the offset centroid of the next frame.

[0118] Bring the obtained slope vector and correction parameters into the restoration equation to obtain the Zernike coefficients. The Zernike coefficients obtained can be used to perform wavefront restoration of the current offset centroid frame (as Figure 6 shown), repeat steps (3)-(4) to achieve wavefront restoration of the entire video data, as Figure 7 .

[0119] 1. Restoration matrix and slope vector:

[0120] Each term of the restoration matrix represents the average wavefront slope when the corresponding Zernike coefficient is 1 on the corresponding sub-aperture.

[0121]

[0122] Among them, M is the selected number of Zernike terms, which can be changed according to requirements. In this embodiment, M is 21. As Figure 6 shown, the abscissa is the number of Zernike terms, and the ordinate is the Zernike coefficient; k is the number of sub-apertures.

[0123] Each term of the slope vector represents the wavefront slope in different directions on the corresponding sub-aperture.

[0124]

[0125] When a spot is missing in a certain sub-aperture, delete its elements in the restoration matrix and slope vector, so that the effective data of the remaining sub-apertures can participate in the calculation of the Zernike coefficients with the correct matrix dimension.

[0126] 2. Average spot distance DSpots

[0127] As Figure 3 (a) shows the calibrated reference spot distribution (green dots). The light spots are evenly distributed at the center of the sub-apertures. If the size of the sub-aperture is a, then the spot distance is also a. However, in actual situations, considering that the reference beam for generating the calibrated reference centroid is not a strictly plane wave, and the sub-aperture sizes cannot be guaranteed to be exactly the same in actual processing, so the method of calculation is adopted to obtain DSpots to reflect the sub-aperture size.

[0128] As Figure 3 (b) shows the offset spot distribution (red dots). In the figure, the distance between the red dots and the green dots is exaggerated for illustration purposes. In actual situations, the distance is very small. Therefore, by judging whether the distance between them is less than one-third of the sub-aperture size, that is, DSpots / 3, it can be determined whether the two points are located in the same sub-aperture. The present invention adopts the threshold DSpots / 3 to further prevent the situation of overlapping light spots to find the most matching light spots.

[0129] The above are the preferred embodiments of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the principle of the present invention, several improvements and modifications can be made, and these improvements and modifications should also be regarded as the protection scope of the present invention.

Claims

1. A method for detecting wavefront distortion based on the similarity of continuous image sequences in a high-speed flow field, characterized in that, including the following steps; (1) Decompose the video data collected by the CCD into a sequence of pictures to obtain the original data; (2) Preprocess each picture, calculate the centroid of each light spot in each picture to obtain the centroid coordinate sequence of each picture, use the centroid coordinates of the first picture as the calibration reference centroid, and at the same time obtain the initial restoration matrix E; (3) Perform centroid matching on two adjacent pictures before and after to obtain the slope vector and correction parameters, and update the offset reference centroid coordinate sequence; (4) Substitute the slope vector and correction parameters into the restoration equation to calculate the Zernike coefficients and complete the wavefront restoration; (5) Match the next picture in the picture sequence with the updated offset reference centroid coordinate sequence, that is, repeat steps (3) to (4) to achieve the wavefront restoration of the entire video data.

2. The method for detecting wavefront distortion based on the similarity of continuous image sequences in a high-speed flow field according to claim 1, wherein, Step (2) includes the following steps: (2.1) Perform image binarization and morphological operations to obtain the approximate position of the light spot; (2.2) Use the centroid formula to calculate the exact position of the light spot. The centroid formula: Among them, (X c , Y c ) is the centroid coordinate of the light spot, M represents the number of pixels in the sub-aperture region, I i is the pixel value of the i-th pixel, (x i , y i ) represents the pixel coordinates within the sub-aperture; after obtaining the centroid of the light spot corresponding to all sub-apertures in the first picture, the calibration reference centroid can be obtained; (2.3) The slope vector G is obtained by comparing the centroid coordinates of the latter image with the calibrated reference centroid coordinates. The slope vector G contains the horizontal spot offset G of each sub-aperture x and the vertical spot offset G x , which is expressed as: Among them, (X t , Y t ) represents the centroid coordinates obtained from the latter image, and f is the focal length of the sub-aperture; (2.4) Use Φ(x, y) to represent the incident wavefront phase, and the wavefront phase is expanded by the Zernike polynomial as: Among them, where Z m (x, y) is the m-th Zernike polynomial, and a m is its coefficient. When m = 1, it is the average value of the wavefront phase and is ignored and denoted as 0; Take the derivative of both sides of the Zernike polynomial, then the wavefront restoration equation is expressed in matrix form as: G = EA where G represents the slope vector; A = {a2 a3…a m} represents the coefficient vector; E represents the restoration matrix, (Z xM ) k represents the horizontal average wavefront slope at the k-th sub-aperture of the M-th Zernike polynomial, (Z yM ) k represents the vertical average wavefront slope at the k-th sub-aperture of the M-th Zernike polynomial; (2.5) Obtain the generalized inverse matrix \(E^+\) of \(E\). + Then the coefficient vector \(A\) can be obtained. Substituting it into the Zernike polynomial, the wavefront phase \(\varPhi(x,y)\) can be calculated.

3. The method for detecting wavefront distortion based on the similarity of continuous image sequences in a high-speed flow field according to claim 2, characterized in that Step (3) includes the following steps: (3.1) Use the centroid of the previous picture as the offset reference centroid and match it with the centroid of each subsequent picture. The initial value of the offset reference centroid is the calibration reference centroid of the first picture, and the coordinate difference is expressed as: D x = X H - X P D y = Y H -Y P Among them, (X H , Y H ) represents the centroid coordinates of the subsequent image, and (X P , Y P ) represents the offset reference centroid coordinates; The light spot distance Dist is expressed as: (3.2) If Dist < DSpots / 3, it is considered that the two light spots at this time are the light spots of the same sub-aperture, where DSpots represents the average light spot distance of the calibration reference centroid; Calculate the slope vector, and replace the centroid of the offset reference centroid at this sub-aperture with the centroid coordinates of the subsequent picture, that is, use the centroid of the subsequent picture as the new offset reference centroid; (3.3) If Dist ≥ DSpots / 3, it is considered that the offset reference centroid fails to match the centroid of the subsequent picture, then it is considered that the data of this sub-aperture is missing, and mark its sub-aperture number, that is, the correction parameter.

4. The wavefront distortion detection method based on the similarity of continuous image sequences in a high-speed flow field according to claim 3, wherein Step (4) includes the following steps: (4.1) The slope vector G calculated in step (3) only includes the successfully matched centroids; the centroids that are not successfully matched retain the vector I composed of their numbers. Delete the data in the restoration matrix E through the vector I to obtain E'; (4.2) By solving the equation G = E'A Obtain the coefficient vector A; (4.3) Substitute the obtained coefficient vector A into the wavefront phase expression: That is, the restored wavefront phase is obtained.

Images