Three-dimensional reconstruction method for complex cavity shape of hydrofoil based on binocular orthogonal view angle

Abstract

The application discloses a three-dimensional reconstruction method of complex cavity shape of a hydrofoil based on binocular orthogonal view angles, and belongs to the field of cavity interface capturing in fluid measurement. The application obtains time sequence images of two view angles of the cavity by using two high-speed cameras. Initial boundaries are extracted based on a boundary detection method, image noise points are removed by a connected region labeling method, and continuous and clear cavity interfaces are obtained by a morphological closing operation method. Then, according to actual shape characteristics of the cavity, the three-dimensional space structure is reconstructed by using a normal ellipse, a rotating ellipse and a half-ellipse hypothesis to reconstruct cavity sections in each flow direction. The application can accurately reconstruct the three-dimensional and dynamic cavity structure by using relatively simple binocular experiment steps.

Description

Technical Field

[0001] The present invention relates to capturing cavitation interfaces in fluid measurement, and more specifically, to a three-dimensional reconstruction method for complex cavitation morphology on hydrofoils based on binocular orthogonal perspectives. Based on time-resolved cavitation images from two perspectives, this method uses an ellipse fitting method, combined with cylindrical, double-helical, and semicircular cavitation interface assumptions, to reconstruct cavitation under different operating conditions near the hydrofoil tip. Background Art

[0002] Cavitation is a complex fluid dynamics process in multiphase flows, widely present in hydraulic machinery such as ship propellers. Cavitation is influenced by its own instability and the nonuniform pressure field of the flow field during its evolution, resulting in highly unsteady and three-dimensional characteristics. The volume fluctuations of cavitation induce localized radiated noise and vibrations, and their spatial distribution influences the three-dimensional vortex structure in the flow field, inducing turbulence. Therefore, accurately estimating the three-dimensional morphology of cavitation is of great significance.

[0003] Traditional bubble 3D reconstruction methods are mainly aimed at bubble flows in free fields. These bubbles are mostly circular and elliptical, which are very different from the bubbles generated by propellers or hydrofoils in terms of shape and scale. Traditional bubble reconstruction methods mostly use basic ellipse fitting algorithms. However, this algorithm is not accurate enough for fitting slender ribbon-like double-spiral tip vortex cavitation and sheet cavitation. The cavitation images from two perspectives cannot accurately obtain the long and short axis information of the cavitation. At the same time, some binocular reconstruction methods based on computer vision generally require clear texture information at the gas-liquid interface. For cavitation conditions with relatively stable interfaces, this will greatly increase the requirements for the light source.

[0004] For example, patent application CN117058335A discloses a method, apparatus, electronic device, and storage medium for reconstructing the three-dimensional morphology of bubbles. While this method mentions bubble slicing and reconstruction based on ellipse fitting, it does not consider ellipse rotation. In practice, the sides of the ellipse's circumscribed rectangle extracted from two perspectives are not necessarily parallel to the major and minor axes of the actual ellipse, which can underestimate the actual cavitation volume. Summary of the Invention

[0005] To overcome the shortcomings of existing technologies, this paper provides a 3D reconstruction method for complex cavitation morphology on hydrofoils based on binocular orthogonal perspectives. By capturing the temporally evolving cavitation images from two perspectives and combining hypotheses of cylindrical, double-helical, and semicircular cavitation interfaces, the method can determine the pulsation of the actual cavitation diameter, area, and volume.

[0006] To achieve the above object, the present invention adopts the following technical solutions:

[0007] The present invention provides a three-dimensional reconstruction method of complex cavitation morphology of a hydrofoil based on binocular orthogonal viewing angles, which comprises the following steps:

[0008] S1: Arrange a CMOS high-speed camera at two orthogonal viewing angles of the flow field condition to be measured, and control the two CMOS high-speed cameras to synchronously capture cavitation images of the flow field condition to achieve synchronous acquisition of time-series images; the flow field condition to be measured is the vortex cavitation and surface sheet cavitation formed by the hydrofoil,

[0009] S2: Surface sheet cavitation and downstream vortex cavitation are partitioned according to the location of the cavitation, and the vortex cavitation is classified into columnar vortex cavitation and double helical vortex cavitation according to the difference between the maximum and minimum diameters of the vortex cavitation in a single view;

[0010] S3: Physically calibrate the CMOS high-speed camera and establish the two-dimensional coordinate systems of the xy plane and xz plane respectively with the hydrofoil tip as the origin in the cavitation image;

[0011] S4: Use the boundary detection method based on the Canny algorithm to locate the gas-liquid interface of the collected cavitation image, set the filter parameters and algorithm threshold according to the partition, and obtain the boundary binary image;

[0012] S5: Calculate and mark the connected areas of the boundary binary image, delete the connected area objects with less than a set number of pixels in the image, perform morphological closing operations on the images in each partition, and obtain the cavitation contour;

[0013] S6: According to the classification results of S2, the corresponding reconstruction model is selected for the cavitation slice in the yz plane, and the cavitation contour slice is reconstructed;

[0014] S7: The cavitation contour slices obtained in S6 are superimposed layer by layer along the flow direction to finally reconstruct the three-dimensional shape of the cavitation.

[0015] According to a preferred embodiment of the present invention, in S1, a signal synchronizer is used to control two CMOS high-speed cameras to achieve synchronous acquisition of time-series images. After acquisition, all images are transmitted to the processing end for storage and processing.

[0016] According to a preferred embodiment of the present invention, the S2 includes classifying a portion of the hydrofoil surface as a sheet cavitation, whose flow cross section can be regarded as a semi-elliptical sheet structure; the portion separated from the hydrofoil is a vortex cavitation; according to the maximum diameter D of the cavitation in a single view max and minimum diameter D min The difference between vortex cavitation and cavitation cavitation is classified as follows: (D max -D min ) / D min Vortex cavitations with < 30% were classified as columnar vortex cavitations, otherwise they were classified as double helical vortex cavitations.

[0017] According to a preferred embodiment of the present invention, S3 includes: placing calibration reference objects near the cavitation generation position and focusing the CMOS high-speed camera, respectively capturing calibration images of two perspectives, and calculating the image magnifications of the two perspectives; then establishing two-dimensional coordinate systems of the xy plane and the xz plane in the image with the hydrofoil tip as the origin, wherein the x direction is the free flow direction, the y direction is the vertical direction, and the z direction is the span direction of the hydrofoil.

[0018] According to a preferred embodiment of the present invention, S4 includes: first, using a Gaussian smoothing operator to perform a convolution operation on the image to remove noise, and setting different filter standard deviations for the sheet cavitation and vortex cavitation regions to extract boundary information; using a derivative operator to calculate the amplitude and direction of the image gradient; performing non-maximum suppression on the global image gradient amplitude, comparing the gradient amplitude in the gradient direction, retaining the local maximum value of the gradient amplitude as the boundary, and removing non-boundary pixels; using a dual threshold algorithm to detect and connect edges, setting two high and low thresholds according to the contrast of the image, retaining strong edges with gradient amplitudes greater than the high threshold and weak edges with gradient amplitudes between the high and low thresholds and connected to the strong edges, and removing other edges.

[0019] According to a preferred embodiment of the present invention, when performing morphological closing operation in S5, square structuring elements are used to erode and close the vortex cavitation area; and disk-shaped structuring elements are used to process the sheet cavitation area due to the large curvature of the boundary.

[0020] According to a preferred embodiment of the present invention, in S6, for the cylindrical vortex cavitation region, the true ellipse assumption is used for fitting. For the double-helical cavitation, the streamwise slices of the cavitation are still assumed to be elliptical, and these elliptical slices rotate along the stream at a constant angular velocity, corresponding to two helical lines. For the attached sheet cavitation region, a semi-ellipse geometric model with a constant deflection angle α is used to fit its cross-sectional slices.

[0021] According to a preferred embodiment of the present invention, when performing layer-by-layer superposition in S7, a Gaussian function is used to smooth the diameter and center position of the cavitation bubble in the transition region between the sheet cavitation and the vortex cavitation.

[0022] Compared with the prior art, the present invention has the following beneficial effects:

[0023] (1) The vortex-cavitation reconstruction model in the present invention fully considers the three-dimensional spatial evolution law of the cavitation itself, overcomes the limitation of the existing technology that cannot reconstruct the distorted structure and the error caused by underestimating the actual cavitation diameter, and has stronger adaptability and accuracy for typical vortex-cavitation structures such as double helix morphology.

[0024] (2) The proposed multi-regional cavitation physical model fully considers the structural characteristics of the attached sheet cavitation on the hydrofoil surface and its transition relationship with the downstream vortex cavitation, effectively overcoming the information loss problem in the reconstruction of complex cavitation morphologies in existing technologies. At the same time, the regional setting can effectively remove the interference of impurity particles and free bubbles in the water and restore some occluded areas in the original image. Compared with the existing bubble orthogonal reconstruction technology, this method can achieve higher-precision three-dimensional reconstruction and ensure the continuity between cavitation structures when facing complex working conditions with multiple cavitation morphologies.

[0025] (3) The cavitation reconstruction method of the present invention has low requirements for illumination, requiring only information from the two outer boundaries. Therefore, it is suitable for scenes with uneven illumination distribution, effectively overcoming the high reliance on surface texture information in existing parallel binocular reconstruction, and has greater adaptability in complex or harsh environments. Furthermore, the low illumination requirement enables this method to be used for high-temporal-resolution dynamic cavitation 3D structure reconstruction, enabling the acquisition of real-time bubble volume information, significantly increasing the spatiotemporal resolution of information such as cavitation diameter, area, and volume. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 A schematic diagram of a process for reconstructing a three-dimensional morphology of a cavitation interface based on binocular orthogonal viewing angles provided by an embodiment of the present invention;

[0027] Figure 2 An experimental arrangement diagram based on binocular orthogonal viewing angles provided by an embodiment of the present invention;

[0028] Figure 3 Two original images of cavitation bubbles from different perspectives provided by the embodiment of the present invention; top view (left) and side view (right);

[0029] Figure 4 The gas-liquid boundary obtained after pre-processing of the cavitation shadow image provided by the embodiment of the present invention; wherein, top view (left) and side view (right);

[0030] Figure 5 The gas-liquid boundary after boundary continuity processing provided by an embodiment of the present invention; wherein, top view (left), side view (right);

[0031] Figure 6 The cylindrical vortex cavitation reconstruction principle provided by the embodiment of the present invention;

[0032] Figure 7 The major and minor axis calculation principle of the double helical vortex cavitation slice provided in the embodiment of the present invention;

[0033] Figure 8 The calculation principle of the attached sheet cavitation slice provided by the embodiment of the present invention;

[0034] Figure 9 The three-dimensional cavitation reconstruction results provided by the embodiment of the present invention; among them, the cylindrical vortex cavitation (top) and the double helical cavitation (bottom);

[0035] Figure 10 Comparison of the actual image and reconstruction result of the cylindrical vortex cavitation provided by the embodiment of the present invention; (a) side view, (b) top view;

[0036] Figure 11 Comparison of the actual image and reconstruction result of the double helical vortex cavitation provided by the embodiment of the present invention; (a) side view, (b) top view;

[0037] Figure 12 Comparison between the actual image and the reconstructed result of the attachment sheet vacuoles in the double helix condition provided in an embodiment of the present invention; wherein, (a) is a side view, and (b) is a top view. DETAILED DESCRIPTION

[0038] The method and technology of the present invention will be further described below through specific examples. It should be pointed out that the examples described below are intended to facilitate understanding of the present invention and do not have any limiting effect on it.

[0039] The method proposed in this paper is mainly used to reconstruct the three-dimensional morphology of cavitation bubbles, and is particularly suitable for reconstructing cavitation bubbles under different working conditions near the tip of a hydrofoil. A hydrofoil is an airfoil moving in water, such as the blade of a ship's propeller.

[0040] To facilitate description of the technical process and effects of the present invention, the following examples simulate actual hydrofoil operating conditions by constructing a simulation experimental platform. This provides a detailed description of the implementation process of the present invention's 3D cavitation morphology reconstruction and illustrates the processing results of each step. It should be noted that the present invention is applicable not only to the 3D morphology reconstruction of cavitation generated during the operation of actual hydrofoils (ship propeller blades), but also to the 3D morphology reconstruction of vortex and sheet cavitation in hydrofoil fluid dynamics experimental platforms or industrial simulation platforms of any simulation scale.

[0041] like Figure 1 As shown, the embodiment of the present invention discloses a method for reconstructing the three-dimensional morphology of cavitation bubbles from a binocular orthogonal perspective, and the method steps include:

[0042] (1) Obtaining time-series images of cavitation from two perspectives

[0043] According to the actual working conditions of the ship propeller blades, a corresponding hydrofoil simulation experimental platform is built. The layout diagram is shown in the figure. Figure 2As shown, the system includes a hydrofoil model 2-1, a water tunnel test section 2-2, a CMOS high-speed camera 2-3, an LED light 2-4, a signal synchronizer 2-5, and a PC 2-6. The hydrofoil model 2-1 is mounted in the central viewing area of ​​the water tunnel test section 2-2 at a 9° angle of attack. Two CMOS high-speed cameras 2-3, each equipped with a 105mm focal length lens, are positioned above and to the side of the hydrofoil at orthogonal viewing angles. A continuous light source from LED 2-4 illuminates the area near the hydrofoil where cavitation is pre-generated. Calibration targets are placed for physical calibration of the two CMOS high-speed cameras 2-3. During the experiment, the water tunnel is operated and the internal pressure is adjusted to the target cavitation number. The cavitation is allowed to fully develop. After the target cavitation is generated, the pressure is stabilized and the cavitation state is observed. The camera exposure time and lens focus are adjusted to ensure clarity of the gas-liquid interface. After the cavitation bubble stabilizes, a trigger signal is input to signal synchronizer 2-6 via a command from PC 2-5. Signal synchronizer 2-6 then controls CMOS high-speed camera 2-3 to capture the cavitation bubble image. Signal synchronizer 2-6 controls two CMOS high-speed cameras 2-3 to synchronize the acquisition of time-series images. After acquisition, all images are transferred to PC 2-5 for storage.

[0044] (2) Cavitation partitioning and vortex cavitation classification

[0045] In the current experiment, cavitation is divided according to its location. The part on the hydrofoil surface is classified as sheet cavitation, whose flow direction cross section can be regarded as a semi-elliptical sheet structure. The part separated from the hydrofoil is vortex cavitation, which appears as a cylindrical or double helical shape. According to the maximum diameter of the cavitation in a single view (D max ) and minimum diameter (D min ) differences to classify vortex cavitation: (D max -D min ) / D min Vortex cavitations with a value less than 30% are classified as columnar vortex cavitations; vortex cavitations with a value greater than or equal to this threshold are classified as double helical vortex cavitations, which appear as twisted ribbons.

[0046] (3) Physical calibration of CMOS high-speed cameras

[0047] Calibration reference objects were placed near the cavitation generation site in the water tunnel and a CMOS high-speed camera was focused. Calibration images were captured from two perspectives, and the image magnifications for the two perspectives were calculated. Two-dimensional coordinate systems were then established in the (x, y) and (x, z) planes, respectively, with the hydrofoil tip as the origin. The x-direction corresponds to the free stream, the y-direction is vertical, and the z-direction is the spanwise direction of the hydrofoil.

[0048] (4) Cavitation shadow image pre-processing

[0049] After adjusting the target cavitation, the cavitation image is acquired, for example Figure 3 The instantaneous images of the two perspectives are shown. The boundary detection method based on the Canny algorithm is used to locate the gas-liquid interface and obtain the binary image of the cavitation boundary. The principle is as follows:

[0050] First, a Gaussian smoothing operator is convolved with the image to remove noise. The vortex and sheet cavitation regions are processed separately, with different filter standard deviations set based on the image contrast and clarity of the gas-liquid interface. A smaller standard deviation (0.05-0.1) is used for the sheet cavitation boundary region on the hydrofoil surface to preserve more detail. A larger standard deviation (0.1-0.3) is used for the vortex cavitation boundary region on the hydrofoil surface to remove interference from free bubbles surrounding the cavitation.

[0051] Calculate the magnitude of the image gradient using the derivative operator and direction .in, 、 are the grayscale gradients in the x and y directions of the image, respectively. The amplitude reflects the strength of the image boundary, and the gradient direction is perpendicular to the edge direction.

[0052] The non-maximum suppression is performed on the global image gradient amplitude in the gradient direction, that is, the local maximum of the image gradient amplitude in this direction is considered as the boundary, and other non-boundary pixels are eliminated.

[0053] A dual-threshold algorithm is used to detect and connect edges. This involves determining the relationship between the image gradient magnitude at each location and a set threshold. Two thresholds are set based on image contrast. In this experiment, the high threshold is set to 0.25 (25% of the maximum gradient), and the low threshold is set to 0.1 (10% of the maximum gradient). Strong edges with gradient magnitudes greater than the high threshold and weak edges with gradient magnitudes between the two thresholds that are connected to strong edges are retained, while all other edges are discarded. Figure 4 The following is an example of a cylindrical vortex cavitation image processed through the above steps.

[0054] (5) Boundary consistency processing

[0055] Calculate the connected areas of the boundary binary image and mark them, delete the connected area objects with less than M pixels in the image, take M as 20 for the sheet cavitation area, and take M as 40 for the vortex cavitation area. This step can remove the interference of some tiny bubbles and hydrofoil surface information in the original image, but there are still some discontinuous or discontinuous areas in the boundary image, and the boundary contours of some areas are not smooth enough. The image needs to be further processed by morphological closing operations. In this experiment, for the vortex cavitation area, a square structure element with a side length of 40 pixels is used for corrosion and closure. The sheet cavitation area has a large curvature at the boundary, so a circular structure element with a radius of 30 pixels is used for processing. Taking the columnar cavitation as an example, the boundary images of the two perspectives are obtained as follows Figure 5 shown.

[0056] At this time, some boundaries are still not closed at the transition position between the attachment sheet cavitation and the vortex cavitation. The Gaussian weighted moving average filter is used to smooth the obtained cavitation center coordinates and diameter distribution data vectors along the flow direction to ensure the continuity of the boundary.

[0057] (6) Reconstruction of vacuolar slice shape

[0058] According to the classification results in step (2), the corresponding reconstruction model is selected for the cavitation slice in the yz plane, and the cavitation contour slice is reconstructed; 6-1) Columnar vortex cavitation

[0059] For cylindrical vortex cavitation regions, a proper ellipse assumption was used for fitting. The cavitation boundary pixels corresponding to the same flow direction position in the top-view and side-view bubble images were taken as the major and minor axes of the ellipse slice, respectively. Assuming the major and minor axes are perpendicular to the y and z axes, respectively, the ellipse slice at that flow direction position was reconstructed, with a flow direction thickness of 1 pixel. Figure 6 The ellipse slices at each flow direction position obtained by using the assumption of a regular ellipse are given and stacked to form a cylindrical cavity structure. The ellipse equation used is:

[0060]

[0061] Where z0 and y0 are the spanwise and vertical coordinates of the ellipse center, respectively, which can be obtained from the upper and lower boundaries of the cavitation bubble [y1, y2] and [z1, z2] in the two views:

[0062]

[0063] a, b are the major and minor axes, corresponding to the diameter information in the two views:

[0064]

[0065] 6-2) Double helical vortex cavitation

[0066] For the double helical cavitation, it is still assumed that the shape of the flow direction slice of the cavitation is elliptical, and these elliptical slices rotate along the flow direction at a constant angular velocity, corresponding to two helical lines. The method for reconstructing the major and minor axes of the slices is shown in Figure 7 Assume that the major and minor axes a and b of the ellipse are unknown quantities, and the angle between the major axis and the z-axis is β, then the equation of the ellipse becomes:

[0067]

[0068] z0 and y0 are the spanwise and vertical coordinates of the ellipse center, respectively, which can be obtained from the equations of the two horizontal and vertical tangent lines:

[0069]

[0070] p and q are the differences between the upper and lower boundaries of the cavitation bubble obtained in the side view and top view, respectively:

[0071]

[0072] Combining the above formulas, we can get:

[0073]

[0074] The a and b values ​​of each slice can be obtained and the tilted elliptical slice can be reconstructed.

[0075] 6-3) Reconstruction of attached vacuoles

[0076] For the cavitation region of the attached sheet, a semi-elliptical geometric model with a constant deflection angle α is used to fit its cross-sectional slice, as shown in Figure 8 The value of α is consistent with the hydrofoil angle of attack. In this experiment, α = 9°. The major and minor axes a and b of the semi-ellipse are also calculated using the above formula:

[0077]

[0078] Where p and q are the differences between the upper and lower boundaries of the cavitation bubble obtained in the side view and top view, respectively;

[0079] For the spanwise and vertical coordinates z0 and y0 of the ellipse center, to simplify the calculation, the lower right corner point (z2, y1) can be considered to coincide with the endpoint of the semi-ellipse, and the following formula can be used for calculation:

[0080]

[0081] The above formula can be used to obtain the corresponding elliptical slice. Assuming that the part of the cavitation bubble on the hydrofoil surface is a semi-elliptical structure, a straight line passing through the center of the ellipse (z0, y0) is needed to clip the ellipse, that is, to retain the area of ​​the following formula:

[0082]

[0083] When the cavitation bubble begins to separate from the hydrofoil, its shape gradually transitions from a semi-ellipse to an ellipse. For this transition region, a vertically inclined straight line is used to clip the ellipse, that is, the region with the following formula is retained:

[0084]

[0085] where h(x) is a function of the flow position x. Let the flow direction in the transition region be [x1, x2] (x2 is downstream of x1). h(x) is calculated as follows:

[0086]

[0087] where p(x2) represents the difference between the upper and lower boundaries of the cavitation bubble obtained in the side view of the slice at the flow direction position x2.

[0088] (7) Three-dimensional cavitation reconstruction

[0089] The slices of each region obtained by the above method are stacked layer by layer along the flow direction to finally reconstruct the three-dimensional shape of the cavitation. In the transition area between the sheet cavitation and the vortex cavitation, the cavitation diameter and center position are smoothed using a Gaussian function. The three-dimensional structure of the cavitation is finally obtained as shown in the figure. Figure 9 shown.

[0090] The reconstruction results of the cylindrical vortex cavitation show that the vortex cavitation surface has a distinct wavy structure under this operating condition. Its trajectory deflects significantly in the spanwise direction immediately after leaving the hydrofoil, then flattens downstream. At this point, the area of ​​the attached sheet cavitation on the hydrofoil surface is relatively small. However, for the double-helical cavitation sheet, the cavitation area is significantly larger, and the cavitation thickness increases. The vortex cavitation portion that leaves the hydrofoil undergoes significant twisting, resembling a twisted "ribbon." Overall, the reconstruction results for both cavitation types agree well with the results from the high-speed cavitation images.

[0091] The local vortex cavitation images of the two views were further selected for comparison with the two-dimensional field of view intercepted from the three-dimensional reconstruction results.

[0092] Figure 10 The comparison results of the cylindrical vortex cavitation show that the size and trajectory of the reconstructed cavitation are in good agreement with the actual shooting results, and the wavy distribution of the gas-liquid interface is also clearly identified, indicating that the cross-section assumption of the regular ellipse is consistent with the actual physical characteristics of the cylindrical cavitation, and the reconstruction result is good.

[0093] Figure 11Comparison results for a double-helical vortex cavitation are shown. The cavitation trajectory and diameter distribution at various flow locations also agree well with the original images obtained by high-speed photography. The cavitation distortion morphology, wavelength, and node distribution are highly consistent. This demonstrates that the elliptical rotation hypothesis is consistent with the physical structure of the double-helical cavitation, resulting in a good reconstruction.

[0094] Figure 12 Comparison was performed on a double-helix mode with a large attached sheet cavitation. The reconstructed cavitation morphology is roughly identical to the actual sheet cavitation. In the top view, the air-liquid interface at the upper boundary of the cavitation is low in contrast. In the transition region where the sheet cavitation separates from the hydrofoil and connects to the vortex cavitation, the interface is unclear and highly reflective, introducing a small amount of error into the reconstruction. Nevertheless, judging by the degree of contour matching, the current method reconstructs the attached sheet cavitation structure well.

[0095] To further evaluate the error of the current reconstruction method, we took into account the time-dependent disturbances of the cavitation itself. Here, we calculated the repeatability of the vortex cavitation volume in the interval [20 mm to 70 mm] from the hydrofoil tip. We also calculated the repeatability of the volume and surface area of ​​the attached cavitation in the interval [-13.0 mm to 15.5 mm]. The mean, standard deviation, and repeatability of the volume (surface area) are calculated as follows:

[0096]

[0097]

[0098]

[0099] Where N is the number of sample groups, which is 20, and 25 samples are selected in each group. The mean volume (surface area) value calculated for each group of samples, is the average volume (surface area) of all samples, S is the sample standard deviation, is the uncertainty of volume (surface area), i.e., the test repeatability. The calculated repeatability results are shown in Table 1. The volume repeatability of the cylindrical cavities and double-helix cavities is 4.1% and 1.6%, respectively. The volume and surface area repeatability of the attached sheet cavities is 1.2% and 0.75%, respectively. The measurement repeatability is better than 5%.

[0100] Table 1 - 3D reconstruction repetition rate statistics

[0101]

[0102] The above-described embodiments merely illustrate several implementations of the present invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. Persons skilled in the art will readily appreciate that variations and modifications may be made without departing from the scope of the present invention, all of which fall within the scope of protection of the present invention.

Claims

1. A three-dimensional reconstruction method for complex cavitation morphology of hydrofoils based on binocular orthogonal viewing angles, characterized in that: The steps include: S1: Arrange a CMOS high-speed camera at two orthogonal viewing angles of the flow field condition to be measured, and control the two CMOS high-speed cameras to synchronously capture cavitation images of the flow field condition to be measured to achieve synchronous acquisition of time-series images; the flow field condition to be measured is the vortex cavitation and surface sheet cavitation formed by the hydrofoil, S2: Surface sheet cavitation and downstream vortex cavitation are partitioned according to the location of the cavitation, and vortex cavitation is classified into columnar vortex cavitation and double helical vortex cavitation according to the difference between the maximum and minimum diameters of the vortex cavitation in a single view; S3: Physically calibrate the CMOS high-speed camera and establish the two-dimensional coordinate systems of the xy plane and xz plane respectively with the hydrofoil tip as the origin in the cavitation image; S4: Use the boundary detection method based on the Canny algorithm to locate the gas-liquid interface of the collected cavitation image, set the filter parameters and algorithm threshold according to the partition, and obtain the boundary binary image; S5: Calculate and mark the connected areas of the boundary binary image, delete the connected area objects with less than a set number of pixels in the image, perform morphological closing operations on the images in each partition, and obtain the cavitation contour; S6: According to the classification results of S2, the corresponding reconstruction model is selected for the cavitation slice in the yz plane, and the cavitation contour slice is reconstructed; S7: The cavitation contour slices obtained in S6 are superimposed layer by layer along the flow direction to finally reconstruct the three-dimensional shape of the cavitation.

2. The three-dimensional reconstruction method of complex cavitation morphology of hydrofoils based on binocular orthogonal viewing angles according to claim 1 is characterized in that: In S1, a signal synchronizer is used to control two CMOS high-speed cameras to achieve synchronous acquisition of time-series images. After acquisition, all images are transmitted to the processing end for storage and processing.

3. The three-dimensional reconstruction method of complex cavitation morphology of hydrofoils based on binocular orthogonal viewing angles according to claim 1 is characterized in that: The S2 includes classifying the part of the hydrofoil surface as a lamellar cavitation, whose flow cross section is regarded as a semi-elliptical lamellar structure; the part separated from the hydrofoil is a vortex cavitation; according to the maximum diameter D of the cavitation in a single view max and minimum diameter D min The difference between vortex cavitation and cavitation cavitation is classified as follows: (D max -D min ) / D min Vortex cavitations with < 30% were classified as columnar vortex cavitations, otherwise they were classified as double helical vortex cavitations.

4. The three-dimensional reconstruction method of complex cavitation morphology of hydrofoils based on binocular orthogonal viewing angles according to claim 1 is characterized in that: The S3 includes: placing calibration reference objects near the cavitation generation location and focusing the CMOS high-speed camera to respectively capture calibration images from two perspectives, and calculating the image magnifications of the two perspectives; then establishing two-dimensional coordinate systems in the xy plane and the xz plane with the hydrofoil tip as the origin in the image, where the x direction is the free flow direction, the y direction is the vertical direction, and the z direction is the span direction of the hydrofoil.

5. The three-dimensional reconstruction method of complex cavitation morphology of hydrofoils based on binocular orthogonal viewing angles according to claim 1 is characterized in that: The S4 includes: Firstly, a Gaussian smoothing operator is used to perform convolution operation on the image to remove noise. Different filter standard deviations are set for the sheet cavitation and vortex cavitation regions to extract boundary information. Use the derivative operator to calculate the magnitude and direction of the image gradient; Perform non-maximum suppression on the global image gradient amplitude, compare the gradient amplitude in the gradient direction, retain the local maximum of the gradient amplitude as the boundary, and remove non-boundary pixels; A dual-threshold algorithm is used to detect and connect edges. Two high and low thresholds are set according to the contrast of the image. Strong edges with gradient amplitudes greater than the high threshold and weak edges with gradient amplitudes between the high and low thresholds and connected to strong edges are retained, and other edges are eliminated.

6. The three-dimensional reconstruction method of complex cavitation morphology of hydrofoils based on binocular orthogonal viewing angles according to claim 1 is characterized in that: When performing the morphological closing operation in S5, a square structuring element is used to erode and close the vortex cavitation region; and a disk-shaped structuring element is used to process the sheet cavitation region due to the large curvature of the boundary.

7. The three-dimensional reconstruction method of complex cavitation morphology of hydrofoils based on binocular orthogonal viewing angles according to claim 1 is characterized in that: In S6, for the cylindrical vortex cavitation region, the correct ellipse assumption is used for fitting, including: The cavitation boundary pixels corresponding to the same flow direction position in the top view image and the side view bubble image are taken as the major axis and minor axis of the elliptical slice, respectively. Assuming that the major and minor axes are perpendicular to the y and z coordinate axes, respectively, the elliptical slice at the flow direction position is reconstructed, and its flow direction thickness is 1 pixel. The elliptical slices at each flow direction position obtained by using the regular ellipse assumption are stacked to form a cylindrical cavity structure. The ellipse equation is: ； Where z0 and y0 are the spanwise and vertical coordinates of the ellipse center, respectively. Based on the upper and lower boundaries of the cavitation [y1, y2], [z1, z2] in the two views, we get: ； a, b are the major and minor axes, corresponding to the diameter information in the two views: 。 8. The three-dimensional reconstruction method of complex cavitation morphology of hydrofoils based on binocular orthogonal viewing angles according to claim 1 is characterized in that: In S6, for the double helical cavitation, it is still assumed that the flow direction slice shape of the cavitation is elliptical, and these elliptical slices rotate along the flow direction at a constant angular velocity, corresponding to two helical lines, including: Assuming that the major and minor axes a and b of the ellipse are unknown, and the angle between the major axis and the z-axis is β, the equation of the ellipse becomes: ； Where z0 and y0 are the spanwise and vertical coordinates of the ellipse center, respectively. From the equations of two horizontal and two vertical tangent lines we get: ； p and q are the differences between the upper and lower boundaries of the cavitation bubble obtained in the side view and top view, respectively: ； The combined formulas yield: ； That is, the a and b values ​​of each slice are obtained and the tilted elliptical slices are reconstructed.

9. The three-dimensional reconstruction method of complex cavitation morphology of hydrofoils based on binocular orthogonal viewing angles according to claim 1 is characterized in that: In S6, for the attached lamellar cavitation region, a semi-elliptical geometric model with a constant deflection angle α is used to fit its cross-sectional slices, including: The value of α is consistent with the angle of attack of the hydrofoil. For the major and minor axes a and b of the semi-ellipse, we have: ； p and q are the differences between the upper and lower boundaries of the cavitation bubble obtained in the side view and top view, respectively. The spanwise and vertical coordinates z0 and y0 of the ellipse center are calculated as follows by assuming that the lower right corner (z2, y1) coincides with the endpoint of the semi-ellipse: ； The corresponding elliptical slice can be obtained by using the above formula; Assuming that the part of the lamellar cavitation on the hydrofoil surface is a semi-elliptical structure, a straight line passing through the center of the ellipse (z0, y0) is needed to clip the ellipse, that is, to retain the area of ​​the following formula: ； When the cavitation bubble begins to separate from the hydrofoil, its shape gradually transitions from a semi-ellipse to an ellipse. For this transition region, an inclined straight line moving along the vertical linear direction is used to clip the ellipse, that is, the region with the following formula is retained: ； where h(x) is a function of the flow position x. Let the flow direction interval in the transition region be [x1, x2], with x2 downstream of x1. h(x) is calculated as follows: ； where p(x2) represents the difference between the upper and lower boundaries of the cavitation bubble obtained in the side view of the slice at the flow direction position x2.

10. The three-dimensional reconstruction method of complex cavitation morphology of hydrofoils based on binocular orthogonal viewing angles according to claim 1 is characterized in that: When performing layer-by-layer superposition in S7, the Gaussian function is used to smooth the cavitation diameter and center position in the transition region between the sheet cavitation and the vortex cavitation.

Images