A method for automatically extracting foam aging kinetics parameters based on a sequence of microscope images

Abstract

This invention belongs to the field of petroleum extraction technology, specifically relating to an automatic extraction method for foam aging kinetic parameters based on microscope image sequences. It includes the following steps: Step 1, correcting the spatial physical drift between microscope image sequences acquired during foam aging; Step 2, separating bubbles from the matrix in the microscope images and assigning a unique label to each bubble; Step 3, automatically extracting the morphological parameters of each bubble in the microscope images and simultaneously calculating the summary statistics for the entire image; Step 4, automatically fitting log-normal and Weibull distributions to the bubble diameter data at each time point, outputting the distribution parameters and the goodness of fit R. 2 Step 5: Fit three aging kinetic models; Step 6: Output foam aging kinetic parameters. This invention integrates image alignment, bubble segmentation, multi-parameter measurement, distribution fitting, and kinetic models into a fully automated workflow, requiring no manual intervention throughout the process.

Description

Technical Field

[0001] This invention belongs to the field of petroleum extraction technology, specifically relating to an automatic extraction method for foam aging kinetic parameters based on microscope image sequences. Background Technology

[0002] Foam is a thermodynamically unstable dispersion system formed by gas dispersed in a liquid phase. Its unique structure enables it to play a crucial role in numerous fields. In oil extraction, foam enhanced oil recovery (FOR) is an important technology. Foam selectively blocks high-permeability channels in the porous media of the reservoir through the Jamin effect, forcing the displacing fluid to redirect to low-permeability areas, thereby significantly improving sweep efficiency. The displacement effect of foam directly depends on its stability under formation conditions—the longer the foam remains in the porous media and the stronger its ability to maintain mobility control, the better the oil displacement effect. Therefore, accurately characterizing the aging behavior of foam and obtaining quantitative aging kinetic parameters are core prerequisites for foam flooding scheme design and surfactant formulation screening.

[0003] Foam aging is a complex process involving multiple coupled mechanisms, primarily including three: Ostwald coarsening (diffusion-driven coarsening), bubble coalescence (liquid film rupture and merging), and liquid film drainage (gravity-driven liquid phase outflow). These three mechanisms do not occur in isolation but are coupled with each other: drainage accelerates liquid film thinning, promoting bubble coalescence; coarsening alters bubble size distribution, affecting capillary pressure differential, and thus further influencing the coarsening rate. Therefore, it is necessary to extract multiple sets of kinetic parameters simultaneously to fully understand the aging behavior of foam.

[0004] Currently, obtaining the above dynamic parameters mainly relies on the following three types of methods: Background Technique 1: Manual Image Analysis. This is the most traditional method in foam aging research. Researchers first prepare foam samples and take photos at different aging time points under an optical microscope. Then, in image processing software such as ImageJ, they manually measure the diameter by clicking on the edge of each bubble with a mouse. Typically, 20-50 bubbles are measured at each time point, taking about 30-60 minutes. After measurement, the data is manually entered into Excel or Origin software, and plotted manually. R (t) curve and fit R 3 -t linear relationship to obtain the coarsening rate constant k CMagrabi et al. (see Jameson G J. Bubble size distribution and coarsening of aqueous foams. Chemical Engineering Science, 1999, 54(18): 4007-4022) systematically studied the bubble size distribution and coarsening behavior of water-based foams using this method. Cheng and Lemlich (see Errors in the measurement of bubble size distribution infoam. Industrial & Engineering Chemistry Fundamentals, 1983, 22(1): 105-109) specifically analyzed the systematic error in bubble size measurement in this method, pointing out that the difference between different operators caused by manual measurement can reach 20%. Gido et al. (see Foam bubble size measured using image analysis before and after passage through a porous medium. Journal of Dispersion Science and Technology, 1989, 10(6): 785-793) studied the bubble size change of foam before and after passing through a porous medium using a similar method.

[0005] Background Technology 2: Semi-automatic Image Segmentation Method. To overcome the inefficiency of manual measurement, researchers have developed various automatic image analysis tools. Sadr-Kazemi and Cilliers (see An image processing algorithm for measurement of flotation froth bubble size and shape distributions. Minerals Engineering, 1997, 10(10): 1075-1083) proposed an image processing algorithm for the bubble size and shape distribution of flotation foam, using watershed segmentation to automatically identify bubble boundaries. Jahedsaravani et al. (see An image segmentation algorithm for measurement of flotation froth bubble size distributions. Measurement, 2017, 111: 29-37) improved the image segmentation algorithm for flotation foam, increasing segmentation accuracy. Kracht et al. (see A stochastic approach for measuring bubble size distribution via image analysis. International Journal of Mineral Processing, 2013, 121: 6-11) proposed a stochastic method to measure bubble diameter distribution, reducing segmentation bias. Mesa and others developed the open-source software Bubble Analyser, which can automatically identify bubbles in images and measure their size. It is currently one of the more representative bubble image analysis tools.

[0006] Background Technology 3: Macroscopic Parameter Measurement Method. In engineering, the graduated cylinder method is commonly used to measure the decay of foam volume over time, the balance weighing method is used to measure the change in liquid content, or the conductivity method is used to indirectly estimate the gas content to characterize the macroscopic stability of foam. Yu and Zhou (see Experimental and modeling study of foam coarsening kinetics inporous media. Frontiers in Energy Research, 2022, 10: 1012728) studied the foam coarsening kinetics in porous media in microfluidic chips and proposed a mass transfer model by visually observing the change in bubble area, but image analysis is still mainly based on manual annotation. Thorat and Bruining (see Foam Flow Experiments. I. Estimation of the Bubble generation-Coalescence Function. Transport in PorousMedia, 2016, 112(1): 53-76) indirectly estimated the bubble generation-coalescence function using pressure drop data through core displacement experiments, but could not obtain information at the single bubble level.

[0007] The aforementioned background technologies share the following common drawbacks: Drawback 1: Discontinuous workflow and extremely low efficiency. From capturing images and measuring bubble size to fitting a kinetic model, each step is completed in different software such as ImageJ, Excel, and Origin, requiring repeated manual data transfer. While semi-automatic segmentation tools (such as Bubble Analyser) can automatically identify bubbles, they only output size data, which researchers still need to manually import into other software for theoretical model fitting. The entire process of analyzing a single sample takes 4-8 hours, severely limiting experimental throughput.

[0008] Defect 2: Incomplete parameters. Existing tools only provide bubble size distribution and do not automatically fit multiple sets of kinetic model parameters, such as LSW coarsening rate constant, coalescence kinetic parameters, and gas content decay parameters. Researchers need to have expertise in nonlinear least squares fitting to complete this manually, and different researchers use different fitting methods (linear regression, nonlinear fitting, different initial value settings), making it difficult to directly compare parameter values ​​reported in the literature.

[0009] Deficiency 3: Lack of distribution analysis. Polydispersion. Systemic time evolution tracking is almost never performed in existing methods (see Barik TK, Roy A. Statistical distribution of bubble size in wetfoam. Chemical Engineering Science, 2009, 64(9): 2039-2043). Log-normal / Weibull statistical fitting of bubble size distribution requires specialized software and statistical knowledge, which is not covered in most foam-related studies. This results in a lack of reliable statistical distribution input parameters for foam numerical simulations (such as Population Balance Model), which can only assume uniform distribution or use empirical values, reducing the reliability of simulation predictions.

[0010] Defect 4: Shape evolution is ignored. During aging, bubbles gradually evolve from near-spherical shapes to polygonal structures (Plateau regularity). Changes in shape factors (roundness, eccentricity) contain information about liquid film drainage and structural reorganization. Although Maurdev et al. (see Bubble motion measurements during foam drainage and coarsening. Journal of Colloid and Interface Science, 2006, 300(2): 735-743) measured bubble motion during drainage and coarsening, they did not systematically track the quantitative evolution of shape factors. Existing tools do not include automatic extraction functions for shape factors.

[0011] Defect 5: Non-reproducibility. Cheng and Lemlich clearly pointed out that the bubble diameter measured manually by different operators can vary by up to 20%. In addition, different fitting methods and software choices may lead to significantly different kinetic parameters when the same experimental data is analyzed by different researchers, which seriously affects the comparability and engineering usability of the results. Summary of the Invention

[0012] The purpose of this invention is to provide an automatic extraction method for foam aging kinetic parameters based on microscope image sequences, overcoming the defects of the above-mentioned background technology, such as workflow breaks, incomplete parameters, missing distribution analysis, neglect of shape evolution, and non-reproducibility, and realizing one-click fully automatic output from the original foam microscope image to a complete set of aging theoretical parameters.

[0013] To facilitate understanding of the technical solution of this patent, the three mechanisms of foam aging are first introduced and explained: (1) Ostwald coarsening (diffusion-driven coarsening). According to the Laplace equation (inΔP For capillary pressure difference, For gas-liquid interfacial tension, R (where is the bubble radius). The gas pressure inside the smaller bubble is higher than that inside the larger bubble. Gas molecules diffuse through the liquid film from the high-pressure (small bubble) to the low-pressure (large bubble), causing the smaller bubble to gradually shrink and disappear, while the larger bubble continues to grow. This process follows the Lifshitz-Slyozov-Wagner (LSW) theory, and is characterized by a linear increase in the cube of the average bubble radius over time. ,in k C The roughening rate constant is a core parameter for measuring the foam's resistance to roughening. k C The larger the foam, the faster it coarsens and the worse its stability.

[0014] (2) Bubble coalescence (liquid film rupture and merging). The liquid film between adjacent bubbles ruptures when the liquid discharge thins to a critical thickness, causing two bubbles to merge into a larger bubble. Coalescence kinetics are usually described by rate equations: ,in N The number of bubbles, k A Let be the aggregation rate constant. α This represents the reaction order. α =1 indicates that the bubble breaks up independently and randomly (first-order dynamics). α >1 indicates co-aggregation. Reaction order. α The value directly reflects the essence of the aggregation mechanism.

[0015] (3) Liquid film drainage (gravity-driven liquid phase outflow). Under the action of gravity, the liquid phase precipitates out from the foam structure, with a gas content of The percentage of gas phase in the total foam volume gradually increases. The decline in gas content can be described by a model: ,in To balance the gas content, k D This represents the decay rate constant. Changes in air content directly affect the apparent viscosity and flowability control of the foam.

[0016] Based on this, the present invention adopts the following technical solution: This invention provides an automatic extraction method for foam aging kinetic parameters based on microscope image sequences, comprising the following steps: Step 1: Correcting the spatial physical drift (image alignment) between microscope image sequences acquired during foam aging. Specifically, for the microscope image sequences acquired during foam aging, SIFT (Scale Invariant Feature Transform) feature point matching combined with ECC (Enhanced Correlation Coefficient) affine transformation is used for image registration to eliminate positional drift between time series and ensure pixel-level alignment of images at different time points.

[0017] Step 2: Bubble Segmentation. A watershed segmentation algorithm guided by a matrix reference map is used to separate bubbles from the matrix in the microscope image, assigning a unique label to each bubble. Specifically, this includes: (a) extracting matrix regions as seed points based on the color annotations in the matrix reference map; (b) detecting bubble boundaries within the bubble region using brightness gradients; and (c) using distance transform combined with the watershed algorithm to separate adhering bubbles. The segmentation results are visualized and verified using different color labels.

[0018] Step 3: Bubble-by-bubble multi-parameter measurement. For the segmented label matrix, automatically extract the following morphological parameters for each bubble: area, equivalent radius (circle equivalent), equivalent diameter, ellipse fitting parameters (major axis, minor axis, eccentricity), and roundness (…). (1) The perfect circle is defined as the centroid coordinates. Simultaneously, the summary statistics for the entire image are calculated, including: number of bubbles N, average radius R, and gas content. Polydispersion Average roundness and average eccentricity.

[0019] Step 4: Bubble diameter distribution fitting. For the bubble diameter data at each time point, automatically fit a log-normal distribution and a Weibull distribution, and output the distribution parameters (log-normal: geometric mean radius). Logarithmic standard deviation Weibull: Shape parameter k, scale parameter ) and goodness of fit R 2 R is calculated by comparing the empirical cumulative distribution with the theoretical cumulative distribution. 2 .

[0020] In the log-normal distribution, μ can be written as the logarithmic characteristic size parameter of the bubble diameter distribution, i.e., μ = ln ,in Let μ be the average bubble radius at time t. The reason for taking the logarithm of the average radius is that the bubble radius remains positive throughout the foam aging process, and coarsening and aggregation typically cause the bubble diameter distribution to exhibit a right-skewed, long-tailed characteristic. Using a logarithmic scale transforms the multiplicative size changes caused by the growth of large bubbles and the disappearance of small bubbles into a more stable additive change, facilitating the characterization of the distribution center shifting towards larger sizes over time. Therefore, an increase in μ over time indicates an overall rightward shift in the bubble diameter distribution, reflecting the increase in characteristic size caused by bubble coarsening and aggregation. (The log-normal distribution...) σ The standard deviation is logarithmic, which mainly reflects the broadening of the bubble diameter distribution on a logarithmic scale. σ As time increases, the size difference between small and large bubbles widens, indicating increased polydispersity. This corresponds to the aging process where small bubbles gradually shrink or disappear due to the Laplace pressure difference, while large bubbles continue to grow. The scale parameter λ in the Weibull distribution characterizes the characteristic size of the bubble diameter distribution; an increase in λ over time also indicates an increase in the size of representative bubbles in the system. The shape parameter k characterizes the distribution shape and concentration; a larger k indicates a more concentrated bubble diameter, while a smaller k usually indicates a longer distribution tail and an increased proportion of large bubbles. Therefore, μ, σ The temporal evolution of k and λ is not simply a mathematical fitting result, but rather a quantitative reflection of the physical processes of bubble diameter growth, distribution broadening, and the increase in the proportion of large bubbles during foam aging, from four perspectives: distribution center, distribution width, distribution shape, and characteristic scale. The goodness of fit R0 2 The calculation is performed by comparing the empirical cumulative distribution function with the theoretical cumulative distribution function because this step evaluates the fit of the bubble diameter distribution, rather than a simple yx curve regression; the empirical cumulative distribution F is obtained by sorting the measured bubble diameters. emp ( R i Then, the theoretical cumulative distribution F is calculated using log-normal or Weibull parameters. fit ( R i ), then you can use R 2 Characterizes the degree of agreement between the theoretical distribution and the overall morphology of the measured bubble diameter distribution.

[0021] Step 5: Multi-model kinetic fitting. Using time series data, three classical aging kinetic models are fitted simultaneously: (a) LSW coarsening law The coarsening rate constant is extracted by nonlinear least squares fitting. k C and initial radius R0; where, t For aging time, R ( t )for t The average radius of the bubble at time t.

[0022] (b) Merging dynamics equations Try different initial values ​​and choose R. 2 The optimal fitting result is used to extract the aggregation rate constant. k A and reaction order In the formula, N This represents the number of bubbles.

[0023] (c) Gas content decline model Extract equilibrium gas content and gas content decay rate constant k D In the formula, The initial gas content, for t The gas content at any given time.

[0024] Step Six: Output foam aging kinetic parameters, including: coarsening rate constant. k C Initial radius R0, aggregation rate constant k A Reaction order equilibrium gas content Attenuation rate constant k D Initial polydispersity PDI0, final polydispersity PDI end Log-standard deviation of the log-normal distribution The range of the shape parameter k of the Weibull distribution.

[0025] The output method for foam aging kinetic parameters is as follows: A summary report and visualization charts (including bubble segmentation verification charts) containing all foam aging kinetic parameters are automatically generated. Linear verification plot, N(t) clustering fitting plot, bubble diameter distribution evolution plot, PDI evolution plot, distribution parameter evolution plot, shape factor evolution plot, and detailed data table for each bubble.

[0026] The key technical point of this invention is: 1. Image alignment, bubble segmentation, multi-parameter measurement, distribution fitting, and dynamic model fitting are integrated into a fully automated workflow, requiring no manual intervention from raw image input to theoretical parameter output.

[0027] 2. Simultaneously fit three complementary aging kinetic models (LSW coarsening, coalescence kinetics, and gas content decay), extract multi-dimensional theoretical parameters from the same set of image data, and establish a complete aging kinetic picture.

[0028] 3. Automatically perform log-normal and Weibull fitting of bubble diameter distribution, and output R0. 2 Quantitative verification provides reliable statistical distribution input parameters for bubble numerical simulations (population equilibrium models, etc.).

[0029] 4. The system tracks the time evolution of parameters such as polydispersity index (PDI), roundness, and eccentricity, incorporating the distribution broadening and shape evolution information that is ignored in existing technologies into the aging kinetic analysis.

[0030] By adopting the above technical solution, the beneficial effects of the present invention are as follows: 1. Significantly improved efficiency. Traditional manual methods take 4-8 hours per sample, while this method is fully automated. Images from 10 time points can be processed from input to output of all parameters in just a few minutes, improving efficiency by 1-2 orders of magnitude.

[0031] 2. Complete parameter system. This method automatically outputs more than 15 theoretical parameters in a single run, including LSW coarsening rate constant, coalescence rate constant and reaction order, gas content decay parameter, polydispersity evolution, log-normal and Weibull distribution parameters, shape factor, etc. Existing tools (such as Bubble Analyser) only output static bubble diameter distribution and cannot obtain these kinetic parameters.

[0032] 3. Significantly enhanced statistical representativeness. This method automatically analyzes 50-100 bubbles at each time point, far exceeding the 20-50 bubbles analyzed by manual methods, and the log-normal and Weibull fitting results are significantly improved. 2 All values ​​are greater than 0.96, which significantly improves the statistical significance and the reliability of the distribution description.

[0033] 4. Reproducible results. The fully automated process eliminates the randomness of manual operation, yielding nearly consistent calculation results and overcoming the 20% operator-to-operator variability in bubble diameter measurement pointed out by Cheng and Lemlich.

[0034] 5. Revealing a complete dynamic picture. This method is the first to simultaneously extract time series of multi-dimensional parameters such as coarsening, coalescence, gas content decay, distribution broadening, and shape evolution from the same set of image data, providing a complete dynamic picture that existing technologies cannot provide for foam aging research. Attached Figure Description

[0035] Figure 1 This is a bubble segmentation result diagram for Example 1 when the aging time is 660 min; Figure 2 The graph shows the verification results of the LSW coarsening law in Example 1. Figure 3This is a graph showing the fitting results of the aggregation kinetics in Example 1; Figure 4 This is a diagram showing the evolution of bubble area distribution in Example 1; Figure 5 This is the evolution diagram of polydispersity index (PDI) in Example 1; Figure 6 This is a diagram showing the evolution of bubble diameter distribution parameters in Example 1; Figure 7 This is the bubble shape factor evolution diagram in Example 1. Detailed Implementation

[0036] Example 1 Static aging experiment of CO2 foam oil generated by a compound surfactant system in a porous medium: Experimental conditions: After CO2 foam oil was generated under the action of the compound surfactant system, it was injected into a microscopic model with a porous medium and allowed to age statically at room temperature. Images were taken using an optical microscope at different time points (t = 0, 30, 60, 180, 300, 360, 480, 540, 600, 660 minutes). Image size: 787 × 590 pixels, scale bar: 50 μm / 8 pixels (i.e., 6.25). (m / pixel), a total of 10 time points image sequences were collected.

[0037] Step 1: For the microscopic image sequence acquired during the foam aging process, image registration is performed using SIFT (Scale Invariant Feature Transform) feature point matching combined with ECC (Enhanced Correlation Coefficient) affine transformation to eliminate positional drift between time series and ensure pixel-level alignment of images at different time points. In this embodiment, SIFT+ECC registration and alignment are performed on 10 original images to eliminate minor displacements during shooting. The aligned images are then uniformly 787×590 pixels.

[0038] Step 2: Bubble Segmentation: (a) Extract matrix regions as seed points based on the labeled colors in the matrix reference image; (b) Detect bubble boundaries within the bubble region using brightness gradients; (c) Separate adhering bubbles using distance transform combined with the watershed algorithm. This embodiment uses a watershed algorithm guided by a red matrix reference image for bubble segmentation. Figure 1 The segmentation results at t = 660 min show a total of 46 bubbles identified, each labeled with a different color, with clear boundaries and no omissions.

[0039] Step 3: Bubble-by-bubble multi-parameter measurement: For the segmented label matrix, automatically extract the following morphological parameters for each bubble: area, equivalent radius (circle equivalent), equivalent diameter, ellipse fitting parameters (major axis, minor axis, eccentricity), and roundness (…). (1) The perfect circle is defined as the centroid coordinates. Simultaneously, the summary statistics for the entire image are calculated, including: number of bubbles N, average radius R, and gas content. Polydispersion The average roundness and average eccentricity are summarized in Table 1.

[0040] Table 1 Summary of measurement parameters at each time point

[0041] Step 4: Bubble diameter distribution fitting: For the bubble diameter data at each time point, automatically fit a log-normal distribution and a Weibull distribution, and output the distribution parameters (log-normal: geometric mean radius). Logarithmic standard deviation Weibull: Shape parameter k, scale parameter ) and goodness of fit R 2 R is calculated by comparing the empirical cumulative distribution with the theoretical cumulative distribution. 2 Bubble-by-bubble measurement and distribution fitting were performed on the label matrix at 10 time points. The output results are shown in Table 2 and... Figure 6 As shown.

[0042] Step 5: Fit three dynamic models using statistical data from 10 time points.

[0043] (1) LSW coarsening law Initial radius R0 = 101.17 coarsening rate constant k C =2666.41 m 3 / min, R 2 = 0.8793. R 3 ( t ) shows a good linear relationship with time (e.g. Figure 2 As shown in the figure, Ostwald coarsening is one of the main aging mechanisms of the system.

[0044] (2) Aggregation Dynamics Initial bubble number N0 = 108.6, aggregation rate constant. k A = 1.82e -8 / min, reaction order = 3.74, R 2 = 0.9600 (e.g.) Figure 3 (As shown). >1 indicates that aggregation has a synergistic effect, and densely populated bubble areas merge faster.

[0045] (3) Gas content decay model Extract equilibrium gas content and gas content decay rate constant k D ; , k D = 0.0463 / min. Where, The initial gas content is 0.20. for t The gas content at any given time.

[0046] Figure 4 The comparison of bubble area probability density distribution at four time points t = 0, 60, 180, and 600 min shows that the distribution shifts towards larger sizes over time, while the distribution width increases, reflecting the bubble growth and polydispersity caused by coarsening.

[0047] Polydispersity Index (PDI) = The increase from 0.484 to 0.541 (an increase of approximately 12%) is a characteristic feature of Ostwald coarsening (e.g. Figure 5 (As shown). The continuous increase in PDI means that the bubble diameter distribution is getting wider and wider, and the capillary pressure difference between small bubbles and large bubbles is getting bigger and bigger, which further accelerates coarsening and forms a positive feedback.

[0048] Log-normal and Weibull distribution fitting R0 2 All > 0.96 (log-normal mean R) 2 = 0.964, Weibull average R 2 = 0.987). The log-normal distribution's log-standard deviation increased from 0.540 to 0.661, and the Weibull distribution's shape parameter decreased from 2.19 to 1.95, both reflecting a broadening trend in the distribution (e.g., Figure 6 (As shown). These distribution parameters can be directly used as inputs to the population equilibrium model.

[0049] The average roundness of the bubbles decreased from 0.76 to 0.63, indicating that the bubbles gradually evolved from near-spherical shapes to polygonal structures, conforming to Plateau's rule (triangular intersection angle 120°). The eccentricity remained relatively stable between 0.61 and 0.72, indicating that the flattening of the bubbles did not change significantly (e.g., Figure 7 (As shown).

[0050] Step Six: Output foam aging kinetic parameters. For details on the output method of foam aging kinetic parameters in this embodiment, please refer to [link to relevant documentation]. Figures 1-7The complete set of dynamic parameters extracted in this example is summarized in Table 2.

[0051] Table 2 Summary of Foam Aging Kinetic Parameters

Claims

1. An automatic extraction method for foam aging kinetic parameters based on microscope image sequences, characterized in that, Includes the following steps: Step 1: Correct the spatial physical drift between the microscopic image sequences acquired during the foam aging process; Step 2: Using a watershed segmentation algorithm guided by a matrix reference image, the bubbles in the microscope image are separated from the matrix, and each bubble is assigned a unique label. Step 3: Automatically extract the morphological parameters of each bubble in the microscope image from the segmented label matrix of Step 2, and calculate the summary statistics of the entire image at the same time. Step four, for each time point of bubble diameter data, automatically fit log-normal distribution and Weibull distribution, output distribution parameters and goodness of fit R 2 ; Step 5: Using time series data, simultaneously fit three aging kinetic models: (a) LSW coarsening law The coarsening rate constant is extracted by nonlinear least squares fitting. k C and the initial radius R0, where, t For aging time, R ( t )for t The average radius of the bubble at time t; (b) Merging dynamics equations Try different initial values ​​and choose R. 2 The optimal fitting result is used to extract the aggregation rate constant. k A and reaction order In the formula, N The number of bubbles; (c) Gas content decline model Extract equilibrium gas content and decay rate constant k D In the formula, The initial gas content, for t The gas content at any given time; Step 6: Output foam aging kinetic parameters.

2. The automatic extraction method for foam aging kinetic parameters based on microscope image sequences according to claim 1, characterized in that, In step six, the foam aging kinetic parameters include: coarsening rate constant. k C Initial radius R0, aggregation rate constant k A Reaction order equilibrium gas content Attenuation rate constant k D Initial polydispersity PDI0, final polydispersity PDI end Log-standard deviation of the log-normal distribution The range of the shape parameter k of the Weibull distribution.

3. The automatic extraction method for foam aging kinetic parameters based on microscope image sequences according to claim 1, characterized in that, In step six, the output method for the foam aging kinetic parameters is as follows: Automatically generates a summary report containing all foam aging kinetic parameters, visualization charts, and detailed data tables for each bubble. The visualization charts include bubble segmentation verification diagrams, etc. R 3 (t) linear verification plot, N(t) clustering fitting plot, bubble diameter distribution evolution plot, PDI evolution plot, distribution parameter evolution plot and shape factor evolution plot.

4. The automatic extraction method for foam aging kinetic parameters based on microscope image sequences according to claim 1, characterized in that, Step one specifically involves: using SIFT feature point matching combined with ECC affine transformation to perform image registration on the microscope image sequence collected during the foam aging process, eliminating positional drift between time series, and ensuring pixel-level alignment of microscope images at different time points.

5. The automatic extraction method for foam aging kinetic parameters based on microscope image sequences according to claim 1, characterized in that, Step two, the watershed segmentation algorithm guided by the matrix reference map, specifically includes: (a) Extract matrix regions as seed points based on the labeled colors in the matrix reference map; (b) Detect bubble boundaries within the foam region using brightness gradients; (c) Separate adhering bubbles using distance transformation combined with the watershed algorithm.

6. The automatic extraction method for foam aging kinetic parameters based on microscope image sequences according to claim 1, characterized in that, In step three, the morphological parameters include: area, equivalent radius, equivalent diameter, ellipse fitting parameters, roundness, and centroid coordinates.

7. The automatic extraction method for foam aging kinetic parameters based on microscope image sequences according to claim 6, characterized in that, Ellipse fitting parameters include major axis, minor axis, and eccentricity.

8. The automatic extraction method for foam aging kinetic parameters based on microscope image sequences according to claim 1, characterized in that, In step three, the summarized statistics include: number of bubbles N, average radius R, and gas content. Polydispersion Average roundness and average eccentricity.

9. The automatic extraction method for foam aging kinetic parameters based on microscope image sequences according to claim 1, characterized in that, In step four, the distribution parameters of the log-normal distribution are: geometric mean radius μ and log-standard deviation. , where μ = ln The distribution parameters of the Weibull distribution are: shape parameter k, scale parameter... .

10. The method for automatically extracting foam aging kinetic parameters based on microscope image sequences according to claim 1, characterized in that, In step four, R is calculated by comparing the empirical cumulative distribution with the theoretical cumulative distribution 2 .

Images