Liquid film flow test method, apparatus, and storage medium

Abstract

The application discloses a liquid film flow test method, device and storage medium, and relates to the technical field of optical measurement. The application first acquires a grid reference image and a pure color reference image in a liquid film free state, and establishes a benchmark relationship between grid offset and liquid film height. Then, the grid image containing grid deformation information and the pure color image containing pixel intensity information are synchronously acquired in a liquid film flow state. The liquid film height of each measurement grid center point is calculated based on the grid deformation information and the benchmark relationship. Further, the linear relationship model of the pixel intensity difference and the liquid film height in the local area is constructed by combining the grid center point height and the pixel intensity difference. Finally, the linear relationship model is used to solve the liquid film height of all pixel points in the measurement range, and the pixel-level liquid film height three-dimensional distribution field is generated through splicing processing. In this way, the accuracy and integrity of the liquid film flow pattern and thickness test are greatly improved.

Description

Technical Field

[0001] This application relates to the field of optical measurement technology, specifically to a liquid film flow testing method, apparatus, and storage medium. Background Technology

[0002] Digital Image Projection (DIP) is a widely used non-contact optical liquid film flow testing technique. Its basic principle is that the liquid film structure causes the grid image projected onto the experimental surface to shift. A cross-correlation algorithm is used to search for the grid shift, and then the local liquid film thickness is calculated based on the relationship between the calibrated grid shift and the local height. Finally, the three-dimensional discrete distribution field of the liquid film height is synthesized based on the local height calculated from each image, thus obtaining the three-dimensional structure of the liquid film.

[0003] However, the spatial measurement accuracy of DIP technology is limited by the size of the projected grid. Grids that are too small are not only unclear, but if the grid size is smaller than the offset caused by the liquid film height, it can also lead to search errors during cross-correlation calculations, causing significant errors. Therefore, the spatial resolution of liquid film testing using DIP technology can only reach a maximum of about millimeters. For relatively simple surface structures, this limitation may not be a major problem, but for moving liquid films with extremely complex surface structures and numerous fine waveforms, grid-based resolution will lose many detailed structural features, preventing more refined analysis.

[0004] Since DIP technology was initially used to measure solid surfaces, relying primarily on diffuse reflection of light, it's necessary to add low-concentration working fluids, such as nano-sized titanium dioxide, to liquid surfaces due to their high light transmittance. This reduces light transmittance and increases diffuse reflection. Different liquid film thicknesses result in varying titanium dioxide content, leading to different diffuse reflection intensities and corresponding pixel intensities in the acquired images. However, this difference is influenced by both light intensity and particle concentration. Titanium dioxide is insoluble in water, making it difficult to ensure overall uniformity of its concentration. Therefore, the relationship between pixel intensity and height for the entire experiment cannot be directly obtained through pre-test calibration. Summary of the Invention

[0005] The purpose of this application is to provide a liquid film flow testing method, device, and storage medium to solve the problems of traditional DIP technology, which makes it difficult to directly and accurately establish the correlation between pixel intensity and liquid film height, and is constrained by grid resolution to generate a continuous and complete three-dimensional distribution field of liquid film height.

[0006] To achieve the above objectives, the first aspect of this application provides a liquid film flow testing method, comprising: Acquire a mesh reference image and a solid color reference image in the state without liquid film, and establish a baseline relationship between mesh offset and liquid film height; Under liquid film flow conditions, simultaneously acquire grid images containing grid deformation information and solid color images containing pixel intensity information; Based on the grid deformation information and the reference relationship, the liquid film height at the center point of each measurement grid is calculated; Based on the liquid film height and pixel intensity difference at the center point of each measurement grid, a linear relationship model between the pixel intensity difference and liquid film height in a local area with the grid center as the reference is established. Using the linear relationship model, the liquid film height of each pixel in all local areas within the measurement range is calculated and stitched together to generate a pixel-level three-dimensional distribution field of liquid film height.

[0007] A second aspect of this application provides a liquid film flow testing device, comprising: The acquisition module is used to acquire the mesh reference image and the solid color reference image in the state without liquid film, and to establish the reference relationship between the mesh offset and the liquid film height. The acquisition module is used to simultaneously acquire a grid image containing grid deformation information and a solid color image containing pixel intensity information under liquid film flow conditions. The calculation module is used to calculate the liquid film height at the center point of each measurement grid based on the grid deformation information and the reference relationship; The modeling module is used to establish a linear relationship model between the pixel intensity difference and the liquid film height in a local area with the grid center as the reference, based on the liquid film height and pixel intensity difference at the center point of each of the measurement grids. The generation module is used to calculate the liquid film height of each pixel in all local areas within the measurement range using the linear relationship model, and to perform stitching processing to generate a pixel-level three-dimensional distribution field of liquid film height.

[0008] A third aspect of this application provides a computer-readable storage medium storing a program that can be loaded by a processor and executed using the liquid film flow test method described above.

[0009] The beneficial effects of this application are: This application acquires grid reference images and solid color reference images under liquid film-free conditions and establishes a benchmark relationship between grid offset and liquid film height, providing a calibration basis for subsequent quantitative calculations. Then, under liquid film flow conditions, grid images containing grid deformation information and solid color images containing pixel intensity information are simultaneously acquired to ensure spatiotemporal matching of deformation and light intensity information. Based on the grid deformation information and benchmark relationship, the liquid film height at the center point of each measurement grid is calculated, forming a discrete grid benchmark height. Furthermore, a linear relationship model between pixel intensity difference and liquid film height within a local region is constructed by combining the grid center point height and pixel intensity difference, overcoming the limitations of traditional DIP (Digital Intensity Perimeter) which relies solely on grid nodes and whose resolution is constrained by grid size. Finally, the linear relationship model is used to solve for the liquid film height of all pixels within the measurement range, and the results are stitched together to generate a pixel-level three-dimensional distribution field of liquid film height. This achieves an improvement from grid-level measurement to pixel-level full-field measurement, effectively overcoming measurement deviations caused by uneven particle concentration and illumination distribution, and obtaining a continuous, complete, and spatially higher-resolution three-dimensional distribution field of liquid film height, significantly improving the accuracy and completeness of liquid film flow morphology and thickness testing.

[0010] Other features and advantages of this application will be described in detail in the following detailed description section. Attached Figure Description

[0011] Figure 1 This is a schematic flowchart of a liquid film flow testing method provided in an embodiment of this application; Figure 2 This is a schematic diagram of a timing control pulse for a synchronizer provided in an embodiment of this application; Figure 3 This is a schematic diagram illustrating a method of dividing a region using a grid image, as provided in an embodiment of this application. Figure 4 This is a schematic diagram of the fitting relationship curve between liquid film thickness and light intensity difference provided in an embodiment of this application; Figure 5 This is a schematic diagram illustrating the construction of a local instantaneous linear relationship mathematical model for a projected grid image provided in an embodiment of this application; Figure 6 This is a schematic diagram illustrating the construction of a local instantaneous linear relationship mathematical model for a projected solid color image provided in an embodiment of this application; Figure 7 This is a schematic diagram comparing the average thickness of the liquid film measured by traditional DIP technology and DIP+I technology under the same working conditions, as provided in the embodiments of this application. Figure 8 This is a schematic diagram of interpolation correction deviation for missing positions in conventional DIP liquid film thickness measurement provided in an embodiment of this application; Figure 9This is a schematic diagram of the structure of a liquid film flow testing device provided in the embodiments of this application. Detailed Implementation

[0012] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0013] In the description of this application, it should be understood that the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the stated features. In the description of this application, "a plurality of" means two or more, unless otherwise explicitly specified. Details are set forth in the following description for illustrative purposes. It should be understood that those skilled in the art will recognize that this application can be implemented without using these specific details. In other instances, well-known structures and processes will not be described in detail to avoid unnecessarily obscuring the description of this application. Therefore, this application is not intended to be limited to the embodiments shown, but rather to be consistent with the broadest scope of the principles and features disclosed herein.

[0014] Figure 1 This is a schematic flowchart of a liquid film flow testing method provided in an embodiment of this application. Figure 1 As shown, this liquid film flow test method may include steps 101-105, which will be described in detail below.

[0015] Step 101: Obtain the mesh reference image and solid color reference image in the state without liquid film, and establish the reference relationship between mesh offset and liquid film height.

[0016] The grid reference image is a baseline image obtained by projecting a standard grid under liquid film-free conditions. The solid color reference image is a background baseline image acquired by projecting a uniform solid color light source under liquid film-free conditions. The grid offset is the pixel offset distance of the grid nodes relative to the baseline position after the liquid film deforms. The baseline relationship is the quantitative correspondence between the pre-calibrated offset values ​​and the actual liquid film height.

[0017] As an example, in the initial stage of testing, a free-flowing liquid film is maintained within the test area. A DIP projection device can project a standard grid pattern and a uniform solid-color light source, respectively, while a high-speed camera acquires the corresponding grid reference image and solid-color reference image. Through calibration experiments, the standard liquid film of known thickness is varied, and the grid offset corresponding to different film thicknesses is recorded. A stable quantitative benchmark relationship between the two is established through fitting, providing a calibration basis for subsequent conversion of actual liquid film height. Completing the benchmark data acquisition and calibration work in the early stage of testing can eliminate initial systematic errors caused by hardware equipment and shooting environment, build a conversion bridge between grid deformation data and actual liquid film thickness, and ensure that subsequent measurement data has a unified metrological standard.

[0018] Step 102: Under the condition of liquid film flow, simultaneously acquire a grid image containing grid deformation information and a solid color image containing pixel intensity information.

[0019] Mesh deformation information refers to image deformation data resulting from positional shifts and morphological changes in the projected mesh under the refraction of the liquid film. Pixel intensity information refers to the grayscale brightness values ​​of each pixel in the image after a pure color light source is transmitted and reflected through the liquid film. Synchronous acquisition refers to the acquisition of two types of images under the same operating conditions and at the same time through timing control.

[0020] As an example, the liquid film conveying device can be turned on to form a stable flowing liquid film of the working fluid. With the help of the synchronous control equipment, the working sequence of the projector and the camera can be coordinated. The grid pattern and the solid color light source are projected alternately according to the set acquisition frequency. During the corresponding projection time period, the grid image carrying the grid deformation characteristics and the solid color image that can reflect the light intensity change are captured respectively, ensuring that the spatiotemporal dimensions of the two types of images are completely matched.

[0021] Synchronous acquisition of deformation data and light intensity data can reduce the error caused by time-sharing acquisition, and completely retain the morphological changes and optical characteristics changes during the liquid film flow process, providing complete original image data for joint calculation of dual-layer data.

[0022] Step 103: Calculate the liquid film height at the center point of each measurement grid based on the grid deformation information and the reference relationship.

[0023] The measurement grid consists of independent computational sub-regions obtained by uniformly dividing the acquired image, with the grid center point being the core localization pixel of a single measurement grid. Spatial cross-correlation matching is a commonly used operation for image deformation comparison, used to accurately identify grid displacement.

[0024] As an example, the real-time acquired flow state grid image can be divided into several measurement grids of uniform size. Each measured grid area is compared with a reference grid area without liquid film to accurately extract the actual offset of each grid center point. By retrieving the pre-established reference conversion relationship and substituting the offset values ​​one by one, the accurate liquid film height values ​​corresponding to the center points of all measurement grids can be obtained.

[0025] By relying on mature grid deformation measurement methods, the height of the liquid film at discrete points is accurately obtained, and the basic reference points for the thickness of the liquid film across the entire field are determined, providing reliable core reference data for subsequent full-domain pixel-level thickness extrapolation.

[0026] Step 104: Based on the liquid film height and pixel intensity difference at the center point of each measurement grid, establish a linear relationship model between the pixel intensity difference and liquid film height in a local area with the grid center as the reference.

[0027] The pixel intensity difference is the difference between the grayscale value of a measured solid-color image pixel and the grayscale value of a solid-color image without a liquid film background. The local region refers to the image area covered by a single measurement grid. The linear relationship model is a quantitative linear calculation formula for the change in light intensity difference with liquid film height within a single grid area.

[0028] As an example, the measured solid-color image and the background solid-color image can be uniformly converted to grayscale, and the distribution data of the pixel intensity difference across the entire field can be calculated pixel by pixel. The light intensity difference corresponding to the center point of each grid is extracted, and a benchmark data pair is constructed by combining it with the measured liquid film height at the center point. The benchmark data across the entire field is fitted to obtain the fitting relationship, and then the linear coefficients specific to each grid are obtained by differentiation. Based on this, a linear correspondence model between light intensity difference and liquid film height that is only applicable to the local range of a single grid can be built.

[0029] This step abandons the traditional global unified calibration method and adopts a grid-based local modeling approach, which effectively reduces measurement interference caused by uneven distribution of working fluid particles and differences in on-site lighting and darkness. This allows the correspondence between light intensity and liquid film thickness to closely match the actual local working conditions, significantly improving the accuracy of thickness estimation.

[0030] Step 105: Using a linear relationship model, calculate the liquid film height of each pixel in all local areas within the measurement range, and perform stitching to generate a pixel-level three-dimensional distribution field of liquid film height.

[0031] Pixel-level liquid film height refers to the liquid film thickness value obtained by solving with a single pixel of the image as the smallest unit of measurement. Stitching is an integration method that fuses the boundaries and corrects errors in the local thickness field obtained from the mesh-based solution. The three-dimensional distribution field is a visualized data field that can intuitively represent the spatial distribution of the liquid film thickness across the entire test area.

[0032] As an example, a single measurement grid can be used as an independent computational unit. The intensity difference of all pixels within the grid is substituted into the local linear relationship model of the corresponding grid to solve for the liquid film height of all pixels within the grid, forming multiple independent local height fields. Then, the overlapping boundaries of adjacent height fields are identified, and methods such as weighted smoothing, interpolation correction, and replacement of outlier values ​​are used to eliminate grid splicing faults and data deviations. All local height fields are then integrated and fused to finally generate a continuous and complete pixel-level three-dimensional distribution field of liquid film height within the measurement range.

[0033] This step breaks through the measurement resolution limitations imposed by traditional grid sizes, improving the accuracy of liquid film thickness measurement from the grid level to the pixel level. It can accurately capture subtle flow fluctuations and thickness gradients in the liquid film. After integration and correction, continuous distributed data is obtained across the entire domain, which can completely restore the true flow morphology of the liquid film, greatly improving the precision and data integrity of liquid film flow testing.

[0034] This application embodiment acquires grid reference images and solid color reference images under liquid film-free conditions and establishes a benchmark relationship between grid offset and liquid film height, providing a calibration basis for subsequent quantitative calculations. Then, under liquid film flow conditions, grid images containing grid deformation information and solid color images containing pixel intensity information are simultaneously acquired to ensure spatiotemporal matching of deformation and light intensity information. Based on the grid deformation information and benchmark relationship, the liquid film height at the center point of each measurement grid is calculated, forming a discrete grid benchmark height. Furthermore, a linear relationship model between pixel intensity difference and liquid film height within a local area is constructed by combining the grid center point height and pixel intensity difference, overcoming the limitations of traditional DIP which relies solely on grid nodes and whose resolution is constrained by grid size. Finally, the linear relationship model is used to solve for the liquid film height of all pixels within the measurement range, and the results are stitched together to generate a pixel-level three-dimensional distribution field of liquid film height. Thus, it achieves an improvement from grid-level measurement to pixel-level full-field measurement, effectively overcoming measurement deviations caused by uneven particle concentration and illumination distribution, and obtaining a continuous, complete, and spatially higher-resolution three-dimensional distribution field of liquid film height, significantly improving the accuracy and completeness of liquid film flow morphology and thickness testing.

[0035] In traditional liquid film testing, the difficulty in synchronously acquiring grid deformation data and light intensity data leads to spatiotemporal information misalignment. Therefore, in step 102, a synchronizer can be used to perform time-series synchronization control on the digital image projector and the high-speed camera, enabling the digital image projector to alternately project grid images and solid color images at a preset frequency. Time-series synchronization control refers to uniformly managing the projection timing of the digital image projector and the exposure timing of the high-speed camera through the synchronizer, ensuring that the projection and shooting actions are matched in timing and phase. The grid image refers to a structured image formed by projecting a standard grid pattern, used to characterize the grid deformation displacement information caused by the liquid film. The solid color image refers to a brightness image formed by projecting a uniform solid color light source, used to reflect the pixel intensity information corresponding to changes in liquid film transmittance. The preset frequency is the switching acquisition frequency at which the projector alternately projects grid images and solid color images.

[0036] As an example, the synchronizer can employ a multi-channel digital delay or pulse generator, controlling the digital image projector and the high-speed camera separately via dual pulse signals. Both the digital image projector and the high-speed camera are set to pulse rising edge triggering, ensuring that the image projection timing matches the camera exposure timing. This achieves precise matching between the image projection timing and the camera exposure timing, effectively eliminating timing offset problems caused by equipment start-up and shutdown time differences. During testing, driven by the synchronizer's pulse signals, the digital image projector alternately and cyclically projects grid images and solid color images at a preset frequency. The grid image can include horizontal one-dimensional grid images and vertical one-dimensional grid images, recording grid deformation offset information caused by the liquid film in both horizontal and vertical dimensions. The solid color image can include horizontal solid color images and vertical solid color images, used to obtain pixel intensity change information corresponding to the light transmittance characteristics of the liquid film under different projection directions.

[0037] Simultaneously, the high-speed camera synchronously triggers operation following the pulse signal, executing a time-division multiplexing acquisition strategy. It acquires grid deformation information during the time period when the projector projects a grid image, and acquires pixel intensity information during the time period when the projector projects a solid color image. In this embodiment, the preset acquisition frequency is set higher than the characteristic frequency of the liquid film flow itself, ensuring that the high-speed camera's sampling rate can cover the entire process of changes in the liquid film surface fluctuations and transient flow structures. This avoids the loss of subtle flow characteristics due to excessively low sampling frequencies, ensuring that the acquired grid deformation information and pixel intensity information correspond to the same instantaneous liquid film flow condition, thus achieving a high degree of spatiotemporal consistency between the two types of data.

[0038] Taking the BNC Model577 multi-channel digital delay or pulse generator as an example of a synchronizer, the timing pulse control logic can be referenced. Figure 2 , Figure 2This is a schematic diagram of a synchro timing control pulse provided in an embodiment of this application. Both the digital image projector (hereinafter referred to as the projector) and the high-speed camera (hereinafter referred to as the camera) are triggered by the rising edge of a pulse signal. Synchronizer channel 1 outputs pulse signal A to control the projector, and channel 2 outputs pulse signal B to control the high-speed camera. The projector's trigger frequency is set to 400Hz, and the high-speed camera's trigger frequency is set to 200Hz, ensuring that the camera can completely acquire a set of grid images and solid color images in a single exposure. The projector projects images in a fixed sequence: horizontal one-dimensional grid lines → vertical one-dimensional grid lines → horizontal solid color image → vertical solid color image. The pre-exposure delay for a single image is set to 300μs, and the exposure time is set to 1200μs. The high-speed camera's single-frame exposure time is set to 4500μs, perfectly matching the projector's projection timing. In the liquid-free state, through the above-mentioned synchronization control logic, the projector projects a DIP grid reference image and a solid color reference image, and the high-speed camera is simultaneously controlled to acquire the DIP grid reference image and pixel solid color reference image in the liquid-free state, serving as the basis for subsequent calculations.

[0039] By employing a high-precision pulse timing synchronization and time-division acquisition mechanism, the geometric deformation information and optical intensity information of the liquid film can be acquired simultaneously without adding multiple sets of test optical paths. This effectively avoids the operational misalignment and data deviation caused by the dynamic fluctuations of the liquid film in traditional step-by-step and time-division acquisition methods. It completely preserves the morphological deformation characteristics and optical intensity change characteristics of the liquid film during dynamic flow, providing complete and highly matched original image data for subsequent joint calculations.

[0040] In step 103, after acquiring the grid image under the liquid film flow state, the acquired global grid image can be uniformly partitioned into multiple sub-regions. The complete measurement image area is divided into multiple sub-regions of the same size that do not overlap. Each sub-region uniquely corresponds to a measurement grid, realizing the grid-based decomposition of the entire measurement area and laying the foundation for fine-grained partitioning calculation.

[0041] For each sub-region, a spatial cross-correlation algorithm is used to perform image matching operations, calculating the cross-correlation coefficient distribution between the grid image and the grid reference image within the sub-region. The spatial cross-correlation algorithm is a sub-pixel image matching algorithm that compares the grayscale distribution similarity between the measured image and the reference image to output the cross-correlation coefficient distribution. This is used to accurately identify the overall offset displacement of the image region and features anti-interference and high accuracy. For example, the grid sub-region image acquired in real-time under liquid film conditions is compared point-by-point with the corresponding grid reference sub-region image under non-liquid film conditions to obtain the cross-correlation coefficient distribution map for each sub-region. This distribution map can intuitively reflect the overall offset pattern of the measured grid relative to the reference grid.

[0042] The pixel displacement vector corresponding to the peak position in the cross-correlation coefficient distribution is determined as the offset of the measurement grid. The cross-correlation coefficient is used to characterize the matching similarity between the measured sub-region grid image and the liquid-free grid reference image. The higher the coefficient value, the higher the matching degree between the two image regions, and the peak position is the optimal matching position. For example, in the cross-correlation coefficient distribution of each sub-region, the pixel position corresponding to the maximum coefficient is extracted, and this peak position is the optimal matching position between the two grid images. The pixel displacement vector is the positional offset of the measured grid relative to the reference grid after the grid is deformed by the liquid film refraction. It contains lateral and longitudinal offset information and is the core original parameter for calculating the liquid film height. By accurately determining the pixel offset vector corresponding to this position as the actual grid offset of the current measurement grid, the matching error caused by image noise and slight light and shadow interference is effectively avoided, ensuring the accuracy of displacement acquisition.

[0043] Finally, based on the offset of the measurement grid and the reference relationship, the liquid film height at the center point of the measurement grid is calculated. The reference relationship is a pre-calibrated quantitative correspondence between the grid offset and the liquid film height, providing a unified calibration basis for the quantitative calculation of the liquid film height. After obtaining the precise offset of each measurement grid, the pre-established reference calibration relationship between the grid offset and the liquid film height is called. The grid pixel offset is substituted into the calibration relationship formula, and the liquid film height corresponding to the center point of each measurement grid is calculated one by one. Finally, the precise liquid film height values ​​of all measurement grid center points in the entire area are obtained, completing the thickness acquisition of discrete reference points.

[0044] By employing gridded partitioned matching calculations, full coverage and seamless computation of liquid film deformation information across the entire domain are achieved. Leveraging the high-precision matching characteristics of spatial cross-correlation algorithms, the accuracy of grid offset recognition is significantly improved, effectively resisting interference from ambient light and image noise. Simultaneously, using calibrated baseline relationships as the conversion basis, the liquid film height at the center point of each grid is accurately obtained. This provides a large number of uniform and reliable discrete height reference points for subsequent local linear relationship modeling, solving the problems of scarce reference points and insufficient data support in traditional testing methods. This provides core and accurate foundational data support for breaking through grid resolution limitations and achieving pixel-level full-field extrapolation of liquid film height.

[0045] Before step 104, calibration preprocessing can be performed. First, the solid color image and the solid color reference image are converted to grayscale to obtain the measured grayscale image and the background grayscale image. Grayscale conversion transforms the solid color image into a single-channel grayscale image, ensuring that only pixel brightness features are retained, eliminating color interference, and standardizing the light intensity calculation. The measured grayscale image is the brightness distribution map obtained after grayscale conversion of the solid color image acquired under liquid film flow conditions, carrying the light intensity information corresponding to the light transmission attenuation of the liquid film. The background grayscale image is the baseline brightness distribution map obtained after grayscale conversion of the solid color reference image under no-liquid-film conditions, used to characterize the original illumination background without a liquid film. By performing grayscale conversion on the solid color image acquired under liquid film flow conditions and the solid color reference image under no-liquid-film conditions, redundant color information is removed, and the images are uniformly converted into single-channel grayscale images, generating the measured grayscale image and the background grayscale image respectively. This ensures that subsequent light intensity calculations rely solely on brightness dimension data, reducing color noise interference.

[0046] Figure 3 This is a schematic diagram illustrating a method of dividing regions using a grid image, as provided in an embodiment of this application. Figure 3 As shown, the left side is a projected grid image under liquid film flow conditions, with the offset of grid nodes ("cross center offset") characterizing the geometric deformation caused by the liquid film. The right side is a projected solid color image under the corresponding operating condition, with the change in pixel grayscale ("cross center light intensity change") characterizing the optical intensity attenuation caused by the liquid film. Both were acquired synchronously, providing fundamental data for subsequent construction of a local linear model by combining grid offset and light intensity change, and for achieving pixel-level liquid film height measurement.

[0047] Next, the difference between the measured grayscale image and the background grayscale image is calculated according to pixel position to obtain a pixel intensity difference distribution map within a local area. The pixel intensity difference is the difference between the measured grayscale and the background grayscale at the same pixel position, which can intuitively reflect the degree of light intensity attenuation caused by changes in liquid film thickness. The processed measured grayscale image and the background grayscale image are then compared according to the one-to-one correspondence of pixels, and the light intensity change is extracted pixel by pixel to generate a pixel intensity difference distribution map covering the entire measurement area, accurately preserving the differences in light intensity attenuation caused by liquid film transmission and refraction at each location.

[0048] Then, the pixel intensity difference is extracted from the pixel intensity difference distribution map of the center point of the measurement grid within the local area. Combined with the liquid film height of the center point of the measurement grid, a single-point benchmark pair consisting of the pixel intensity difference and the liquid film height within the local area is constructed. The single-point benchmark pair is a one-to-one correspondence data sample consisting of "known liquid film height - corresponding pixel intensity difference" for the same center point of the measurement grid, serving as the basic calibration unit for local modeling. In the pixel intensity difference distribution map, the pixel intensity difference corresponding to each center point of the measurement grid is accurately extracted. Combined with the calculated liquid film height of the corresponding center point of the grid, multiple sets of "liquid film height - pixel intensity difference" single-point benchmark pairs are constructed one-to-one, providing accurate sample data for subsequent fitting modeling. For example, assuming the liquid film height of the center point of the measurement grid is... Calculate the pixel intensity difference at the center point of the grid. : .in, The pixel brightness at the center point of the grid in the liquid film state. This represents the pixel intensity at the corresponding location in the reference image.

[0049] Next, the liquid film height at the center points of all measurement grids within the measurement range was measured. Difference with corresponding pixel intensity By performing a global fit, the fitting relationship between the liquid film height and the pixel intensity difference is obtained. The fitting formula obtained by global fitting is based on the reference data of all grid center points in the entire field, and is a formula for the overall variation law of height-light intensity. Single-point reference pairs of all measurement grid center points within the measurement range are collected, and global fitting is performed on the difference data of height and light intensity in the entire field. The overall fitting formula of liquid film height and pixel intensity difference is obtained by solving the problem, and the quantitative correlation law between the two is initially established.

[0050] Figure 4 This is a schematic diagram of a fitting curve showing the relationship between the liquid film thickness and the difference in light intensity provided in an embodiment of this application. Figure 4 As shown in the figure, the red scatter points (Exp) represent the baseline data pairs for all grid center points within the measurement range. The horizontal axis represents the pixel intensity difference between the grid center points. The vertical axis represents the corresponding liquid film height. The green fitted curve is the global relationship obtained by fitting these scattered points; the example in the figure is in exponential form. .

[0051] Finally, by differentiating the fitted relation, the derivative values ​​at the center points of each measurement grid are obtained. The derivative value is used as the local linear coefficient for the corresponding measurement grid. The local linear coefficient is a slope parameter specific to a single grid obtained by differentiating the global fitting formula, representing the instantaneous rate of change of liquid film height with light intensity difference within a single local region. This derivative value characterizes the instantaneous slope of height change with light intensity within a small local area, and is defined as the local linear coefficient for the corresponding measurement grid, enabling differentiated parameter calibration for different grid regions. By transforming discrete reference points into differentiable global relationships and customizing a unique linear coefficient for each measurement grid, the problem of traditional global calibration models being unable to adapt to local illumination or uneven particle concentration is solved.

[0052] By preprocessing image grayscale, calculating pixel differences, and performing global fitting and differentiation, the purification of the full-field reference data and the local parameter differentiation calibration were completed. This method abandons the traditional unified parameter calibration method and fully adapts to the actual working conditions of uneven illumination and uneven distribution of working particles in different local areas. It provides exclusive coefficient parameters for subsequent precise mesh modeling, greatly weakens the systematic errors caused by environmental and working material interference, and ensures the fit and accuracy of local modeling.

[0053] In step 104, within a local region of a single measurement grid, based on the premise of uniform working fluid particle concentration and projected light intensity, a linear relationship model is constructed between the liquid film height and pixel intensity difference of any pixel within the local region, using a single-point reference pair as a benchmark. This linear relationship model is a linear calculation formula established using a single measurement grid as an independent unit, relying on the grid center point reference pair and a specific local linear coefficient, and only adapts to the height-light intensity variation law of the current local region of the grid. In this way, it can be ensured that the height and light intensity difference within the local region exhibit a linear variation law, satisfying the linear modeling conditions.

[0054] Then, traversing each of the measurement grids within the measurement range, the steps of parameter extraction, global fitting and differentiation, and local linear relationship model construction are repeated to complete the local instantaneous linear relationship calibration of all the measurement grids within the measurement range. Local instantaneous linear calibration is a calibration method that completes the entire process of parameter extraction, coefficient solution, and model construction grid by grid, enabling each grid to independently adapt to its own local flow conditions.

[0055] As an example, a linear relationship model can satisfy the following formula: ; in, Coordinates within a local area The height of the liquid film corresponding to each pixel. The height of the liquid film at the center point of the grid. These are the local linear coefficients of the grid, obtained by fitting and differentiating from the grid center. coordinates The difference in pixel intensity. Among them, , This represents the intensity of the pixel in the liquid film state. This represents the intensity of the corresponding pixel in the reference image.

[0056] Figure 5 This is a schematic diagram illustrating the construction of a local instantaneous linear relationship mathematical model for a projected grid image provided in an embodiment of this application. For example... Figure 5 As shown, the left side is the reference surface without a liquid film, and the right side is the liquid film surface. The center of the red crosshair is the center point of the grid measured by DIP. The right side is Its liquid film height is The pixel intensity difference is In this way, a unique anchor point can be found for each grid. Alternatively, a known height and corresponding light intensity difference can be provided as a reference for the local model.

[0057] Figure 6 This is a schematic diagram illustrating the construction of a local instantaneous linear relationship mathematical model for a projected solid color image provided in an embodiment of this application. For example... Figure 6 As shown, the left side is the reference plane, and the right side is the liquid film plane. The measurement is performed on any pixel within the grid. There is a corresponding difference in light intensity. Based on the difference between the liquid film height and pixel intensity at the anchor point, combined with Figure 5 Obtain local linear coefficients A linear model of the measurement grid is constructed. Within a very small region of a single grid, assuming uniform working fluid concentration and illumination, the liquid film height is linearly related to the light intensity difference. Through anchor point calibration, the light intensity difference of all pixels within the grid can be converted into the corresponding liquid film height.

[0058] By employing local instantaneous modeling with a sub-mesh, this method overcomes the limitations of traditional global unified models that cannot adapt to local operating conditions, fully aligning with actual measurement scenarios involving uneven local illumination and particle concentration during liquid film flow. Relying on a combination of precise reference height at the grid center point and locally specific linear coefficients, it achieves accurate estimation of single-pixel height, completely breaking through the bottleneck of traditional DIP technology, which relies solely on grid nodes for solution and is limited by grid size resolution. This provides core algorithmic support for the subsequent generation of pixel-level, high-precision, and fully covered liquid film height fields.

[0059] In step 105, each measurement grid can be used as an independent calculation unit. All pixels within the measurement grid are traversed, and the pixel intensity difference between each pixel is substituted into the local linear relationship model of the corresponding measurement grid to calculate the liquid film height of all pixels within the measurement grid, forming multiple local height fields based on the measurement grid. A local height field is a set of liquid film height data for all pixels within a given grid region, calculated using its specific local linear relationship model, with each measurement grid as the calculation unit.

[0060] Then, overlapping boundary regions of adjacent local height fields are identified. A bilinear weighting method is used to perform a weighted average smoothing of the liquid film height of pixels within these overlapping boundary regions, eliminating abrupt height changes at the grid boundaries. The bilinear weighting method is a smoothing algorithm that uses the distance from a pixel to the center points of the two grids as weights to perform a weighted average of the height values ​​calculated by the two grid models within the overlapping boundary regions. The distance from the pixel to the center points of the two measured grids is inversely proportional; the closer the grid is, the higher the weight of its calculated height value, thus eliminating abrupt height changes at the grid boundaries and ensuring a continuous transition.

[0061] Boundary consistency verification is performed on the liquid film height field within the smoothed measurement range. Boundary consistency verification involves checking the liquid film height gradient at the boundaries of adjacent grids to determine if there are any abrupt changes exceeding a preset threshold, thus assessing data continuity. If the height gradient at the boundaries of adjacent grids exceeds the preset threshold, a quadratic linear interpolation correction is performed on the overlapping boundary region, using the liquid film heights at the center points of the two measurement grids as a reference. This quadratic linear interpolation correction method eliminates discontinuities by re-interpolating the boundary region using the liquid film heights at the center points of two adjacent grids when boundary gradients are abnormal.

[0062] Next, outliers exceeding the set range for liquid film thickness are replaced with the median of the pixel neighborhood to eliminate single-point errors caused by random noise. Invalid values ​​in extremely thin liquid film regions are completed based on the local linear relationship of the surrounding measurement grid. For example, based on the local linear relationship model of the surrounding measurement grid, the data is completed by combining the light intensity difference of the corresponding pixel, reducing data loss caused by weak signals.

[0063] Finally, all local height fields are merged, and all smoothed, verified and corrected local height fields are merged to form a continuous, complete and uninterrupted pixel-level three-dimensional distribution field of liquid film height within the measurement range.

[0064] By employing localized grid modeling and batch computation, the measurement resolution of liquid film height is improved from the traditional grid level to the pixel level, enabling the capture of subtle thickness variations and transient structures in liquid film flow, significantly enhancing the spatial resolution of the measurement. Through bilinear weighting and quadratic linear interpolation correction, abrupt height changes between adjacent grids are effectively eliminated, ensuring the continuity and smoothness of the entire field data and reducing the "grid step effect" caused by traditional block-based computation. Boundary consistency verification, neighborhood median replacement, and local linear relationship completion effectively handle noise, outliers, and invalid values, improving data stability and reliability, and obtaining complete full-field data even under extremely thin liquid film conditions. The resulting three-dimensional distribution field of liquid film height accurately recreates the flow morphology and thickness distribution of the liquid film, providing high-precision, high-resolution experimental data support for liquid film flow mechanism analysis and numerical model verification.

[0065] This application's embodiments overcome the bottleneck of traditional DIP technology being limited by the size of the projection grid. By using local calibration of the DIP grid and linear inversion of pixel intensity, the spatial resolution of liquid film thickness measurement is improved to the pixel level. This can completely capture the fine waveforms and complex surface structures of the moving liquid film, solving the problem of traditional DIP technology losing the fine structure of the liquid film.

[0066] Furthermore, the embodiments of this application employ a local instantaneous linear calibration method instead of the traditional global calibration, which effectively offsets the measurement errors caused by uneven working fluid particle concentration and uneven illumination distribution. The standard deviation of the measured average liquid film thickness is significantly reduced, and the measurement random error for the average liquid film thickness is significantly reduced. Figure 7 This is a schematic diagram comparing the average liquid film thickness measured by traditional DIP technology and DIP+I technology under the same operating conditions, as provided in an embodiment of this application. Figure 7 As shown, the horizontal axis represents measurement time in t / s, reflecting the dynamic flow process of the liquid film. The vertical axis represents H, indicating the average thickness of the liquid film in mm. Green scatter points represent measurement data from traditional DIP technology, while red scatter points represent measurement data from the DIP+I technology of this embodiment. The colored dashed line represents the mean ± standard deviation range, indicating the dispersion of the data. Traditional DIP technology is affected by uneven working fluid particle concentration and uneven distribution of illumination, resulting in highly dispersed data points, large standard deviation, and drastic fluctuations in measurement results. The DIP+I technology of this embodiment employs local instantaneous linear calibration, establishing a unique light intensity-height relationship for each grid, effectively offsetting the interference caused by local operating condition differences. As can be clearly seen from the figure, the data points of DIP+I are highly concentrated, closely following the mean line, and the standard deviation is much smaller than that of traditional DIP, significantly reducing the random error of the measurement and greatly improving the stability of the results.

[0067] In addition, the average liquid film thickness measured in the embodiments of this application solves the problem of excessive thickness caused by interpolation filling of thin liquid film regions in traditional DIP technology, which greatly improves the accuracy of liquid film thickness measurement and provides reliable measurement data support for the refined analysis of liquid film flow characteristics, transport mechanism and instability mechanism. Figure 8 This is a schematic diagram illustrating the interpolation correction bias for missing locations in a conventional DIP liquid film thickness measurement, as provided in an embodiment of this application. Figure 8 As shown, the black curve (actual liquid film structure) represents the true thickness distribution of the liquid film, exhibiting peaks and troughs in a typical undulating pattern. The blue-green bars (DIP-I measurement results) are the measurement results of the technology in this application embodiment, completely reproducing the continuous thickness change of the liquid film from peak to trough with pixel-level resolution, closely matching the actual curve. The purple grid bars (DIP measurement results) are discrete measurement points of traditional DIP technology, acquiring data only at grid nodes. The dark purple thick bars (DIP interpolation results) are the results of traditional DIP interpolation to fill in the data between nodes, significantly deviating from the actual thickness in the trough region, showing a noticeable "artificial height" phenomenon. Traditional DIP, limited by grid size, cannot measure the liquid film thickness between nodes and can only estimate it through interpolation. This method incorrectly inflates the thickness in thin liquid film (trough) regions, causing a systematically large error. The technology in this application embodiment overcomes the grid resolution limitation, directly achieving pixel-level continuous measurement without interpolation. Therefore, it can accurately reproduce the thickness of thin liquid film regions such as liquid film troughs, avoiding the deviation caused by interpolation and greatly improving the accuracy of measurement.

[0068] Figure 9 This is a schematic diagram of the structure of a liquid film flow testing device provided in an embodiment of this application. Figure 9 As shown, the liquid film flow testing device 900 may include an acquisition module 901, a data collection module 902, a calculation module 903, a modeling module 904, and a generation module 905.

[0069] The acquisition module 901 is used to acquire the grid reference image and the solid color reference image in the state without liquid film, and to establish the reference relationship between the grid offset and the liquid film height.

[0070] The acquisition module 902 is used to simultaneously acquire a grid image containing grid deformation information and a solid color image containing pixel intensity information under liquid film flow conditions. The calculation module 903 is used to calculate the liquid film height at the center point of each measurement grid based on grid deformation information and reference relationship.

[0071] Modeling module 904 is used to establish a linear relationship model between pixel intensity difference and liquid film height in a local area with the grid center as the reference, based on the liquid film height and pixel intensity difference at the center point of each measurement grid.

[0072] The generation module 905 is used to calculate the liquid film height of each pixel in all local areas within the measurement range using a linear relationship model, and then perform stitching to generate a pixel-level three-dimensional distribution field of liquid film height.

[0073] This application also provides a computer-readable storage medium storing a program that can be loaded by a processor and executed by any of the liquid film flow testing methods in this application.

[0074] Those skilled in the art will understand that all or part of the functions of the various methods in the above embodiments can be implemented by hardware or by computer programs. When all or part of the functions in the above embodiments are implemented by computer programs, the program can be stored in a computer-readable storage medium, which may include: read-only memory, random access memory, disk, optical disk, hard disk, etc., and the program is executed by a computer to achieve the above functions. For example, the program can be stored in the memory of a device, and when the program in the memory is executed by the processor, all or part of the above functions can be achieved. In addition, when all or part of the functions in the above embodiments are implemented by computer programs, the program can also be stored in a server, another computer, disk, optical disk, flash drive, or external hard drive, etc., and can be downloaded or copied to the memory of a local device, or the system of the local device can be updated. When the program in the memory is executed by the processor, all or part of the functions in the above embodiments can be achieved.

[0075] The above examples illustrate this application only to aid understanding and are not intended to limit its scope. Those skilled in the art to which this application pertains can make various simple deductions, modifications, or substitutions based on the ideas presented.

Claims

1. A method for testing liquid film flow, characterized in that, include: Acquire a mesh reference image and a solid color reference image in the state without liquid film, and establish a baseline relationship between mesh offset and liquid film height; Under liquid film flow conditions, simultaneously acquire grid images containing grid deformation information and solid color images containing pixel intensity information; Based on the grid deformation information and the reference relationship, the liquid film height at the center point of each measurement grid is calculated; Based on the liquid film height and pixel intensity difference at the center point of each measurement grid, a linear relationship model between the pixel intensity difference and liquid film height in a local area with the grid center as the reference is established. Using the linear relationship model, the liquid film height of each pixel in all local areas within the measurement range is calculated and stitched together to generate a pixel-level three-dimensional distribution field of liquid film height.

2. The liquid film flow testing method according to claim 1, characterized in that, The simultaneous acquisition of a grid image containing grid deformation information and a solid color image containing pixel intensity information under liquid film flow conditions includes: A synchronizer is used to perform time synchronization control between the digital image projector and the high-speed camera, so that the digital image projector alternately projects grid images and solid color images at a preset frequency; The high-speed camera is controlled to be triggered synchronously with the digital image projector. During the time period when the digital image projector projects a grid image, grid deformation information is collected, and during the time period when the projector projects a solid color image, pixel intensity information is collected. The preset frequency is higher than the characteristic frequency of liquid film flow to ensure the spatiotemporal consistency between the grid deformation information and the pixel intensity information collected at the same time.

3. The liquid film flow testing method according to claim 2, characterized in that, The synchronizer uses a multi-channel digital delay or pulse generator to control the digital image projector and the high-speed camera respectively through dual pulse signals. Both the digital image projector and the high-speed camera are set to be triggered by the rising edge of the pulse, so that the image projection timing matches the camera exposure timing.

4. The liquid film flow testing method according to claim 1, characterized in that, The calculation of the liquid film height at the center point of each measurement grid based on the grid deformation information and the reference relationship includes: The grid image is divided into multiple sub-regions, and each sub-region corresponds to a measurement grid. For each sub-region, the cross-correlation coefficient distribution between the grid image and the grid reference image within the sub-region is calculated using a spatial cross-correlation algorithm; The pixel displacement vector corresponding to the peak position in the cross-correlation coefficient distribution is determined as the offset of the measurement grid. The liquid film height at the center point of the measurement grid is calculated based on the offset of the measurement grid and the reference relationship.

5. The liquid film flow testing method according to claim 1, characterized in that, Before the step of establishing a linear relationship model between pixel intensity difference and liquid film height within a local region based on the grid center, the method further includes: The solid color image and the solid color reference image are respectively converted to grayscale to obtain the measured grayscale image and the background grayscale image; The difference between the measured grayscale image and the background grayscale image is calculated according to the pixel position to obtain the pixel intensity difference distribution map in the local area; Extract the pixel intensity difference from the pixel intensity difference distribution map of the center point of the measurement grid in the local area, and combine it with the liquid film height of the center point of the measurement grid to construct a single-point reference pair composed of the pixel intensity difference and the liquid film height in the local area; Globally fit the liquid film height and the corresponding pixel intensity difference at all center points of the measurement grid within the measurement range to obtain the fitting relationship between the liquid film height and the pixel intensity difference; The derivative of the fitted relation is obtained to obtain the derivative value of the center point of each measurement grid, and the derivative value is used as the local linear coefficient of the corresponding measurement grid.

6. The liquid film flow testing method according to claim 5, characterized in that, The establishment of a linear relationship model between pixel intensity difference and liquid film height within a local region based on the grid center includes: Within a local area of ​​a single measurement grid, based on the premise that the concentration of working particles and the intensity of projected light are uniform, a linear relationship model between the liquid film height and the pixel intensity difference of any pixel point within the local area is constructed using the single-point reference pair as a reference. By traversing each of the measurement grids within the measurement range, the steps of parameter extraction, global fitting and differentiation, and local linear relationship model construction are repeatedly performed to complete the local instantaneous linear relationship calibration of all the measurement grids within the measurement range; The linear relationship model satisfies the following formula: ； in, Coordinates within a local area The height of the liquid film corresponding to each pixel. The height of the liquid film at the center point of the grid. These are the local linearity coefficients of the grid. coordinates The pixel intensity is different.

7. The liquid film flow testing method according to claim 1, characterized in that, The step of calculating the liquid film height of each pixel within all the local regions using the linear relationship model includes: Using a single measurement grid as an independent calculation unit, the pixel intensity difference of each pixel point within the measurement grid is substituted into the local linear relationship model of the corresponding measurement grid to calculate the liquid film height of all pixels within the measurement grid, thus forming multiple local height fields with the measurement grid as the unit.

8. The liquid film flow testing method according to claim 7, characterized in that, The splicing process includes: Identify overlapping boundary regions of adjacent local height fields, and use a bilinear weighting method to perform weighted average smoothing on the liquid film height of the pixels in the overlapping boundary regions to eliminate height abrupt changes at the grid boundaries. The distances from the pixels in the overlapping boundary regions to the center points of the two measurement grids are inversely proportional. Boundary consistency verification is performed on the liquid film height field within the measurement range after smoothing. If the height gradient at the boundary of adjacent grids exceeds a preset threshold, the overlapping boundary region is corrected by quadratic linear interpolation based on the liquid film height at the center points of the two measurement grids. Outlier values ​​that exceed the set range of liquid film thickness are replaced with the median of the pixel neighborhood, and invalid values ​​in extremely thin liquid film regions are completed based on the local linear relationship of the surrounding measurement grid. All local height fields are merged to form a continuous and complete pixel-level three-dimensional distribution field of liquid film height within the measurement range.

9. A liquid film flow testing device, characterized in that, include: The acquisition module is used to acquire the mesh reference image and the solid color reference image in the state without liquid film, and to establish the reference relationship between the mesh offset and the liquid film height. The acquisition module is used to simultaneously acquire a grid image containing grid deformation information and a solid color image containing pixel intensity information under liquid film flow conditions. The calculation module is used to calculate the liquid film height at the center point of each measurement grid based on the grid deformation information and the reference relationship; The modeling module is used to establish a linear relationship model between the pixel intensity difference and the liquid film height in a local area with the grid center as the reference, based on the liquid film height and pixel intensity difference at the center point of each of the measurement grids. The generation module is used to calculate the liquid film height of each pixel in all local areas within the measurement range using the linear relationship model, and to perform stitching processing to generate a pixel-level three-dimensional distribution field of liquid film height.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a program that can be loaded by a processor and executed as described in any one of claims 1 to 8 for testing liquid film flow.

Images