A parameter-adjustable porous medium micp process microscopic observation system and method

Abstract

The application discloses a parameter-adjustable porous medium MICP process microscopic observation system and method, and relates to the technical field of water conservancy projects.The system comprises a microfluidic chip, a syringe pump, a waste liquid collecting container, a microscopic observation module and an image processing module; the microscopic observation module collects MICP process images of a random porous medium structure in the microfluidic chip; the image processing module performs image data collection and processing on the MICP process images, and identifies particles, pores and calcium carbonate deposits in the MICP process images; and the random porous medium structure is prepared by etching the chip based on a generated random porous medium structure pattern.The system can accurately regulate microscopic parameters such as particle sizes and porosities of the random porous medium structure according to actual requirements of the MICP process; and the particles, the pores and the calcium carbonate deposits can be accurately identified through contrast enhancement, edge identification and threshold segmentation processing on the MICP process images.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of water conservancy engineering, and particularly relates to a porous medium MICP process microscopic observation system and method with adjustable parameters. BACKGROUND

[0002] Microbially induced carbonate precipitation (MICP) technology is a green and environmentally friendly technology that utilizes microbial metabolic activity to promote the precipitation of carbonate minerals. MICP technology has broad application prospects in the fields of water conservancy engineering, environmental protection, and building materials. For example, in the field of soil reinforcement, MICP technology can enhance the strength and stability of soil through calcium carbonate cementation; in the field of wastewater treatment, MICP technology can remove heavy metal ions in water through the adsorption and precipitation of calcium carbonate.

[0003] Traditional MICP research mainly focuses on the macroscopic properties of solidified samples, such as quantifying the improvement of the compressive performance of MICP-treated soil or materials through unconfined compressive strength tests, and evaluating the effect of MICP treatment technology on reducing the permeability of soil or materials through permeability tests. Although macroscopic experimental methods can intuitively reflect the strengthening effect of calcium carbonate precipitation on the performance of materials or soil, they are difficult to reveal the underlying microscopic mechanisms and action mechanisms. For example, it is difficult to accurately grasp and control the deposition position, shape, and distribution of carbonate minerals in porous media, which greatly limits the optimization and application of MICP technology.

[0004] Therefore, it is particularly important to provide an experimental device and method for observing and analyzing the MICP process in a porous medium structure at a microscale. Currently, there are such devices and related research. For example, the document "Microfluidic chip test research on spatial and temporal evolution of microbial mineralization deposition" (He Xiang, Liu Hanlong, Han Fei, Ma Guoliang, Zhao Chang, Chu Jian, Xiao Yang; Chinese Journal of Geotechnical Engineering, DOI: 10.11779 / CJGE202110012) discloses a conceptual microfluidic chip, bacteria and reaction liquid are injected from two separate branch pipes at one end of the chip and mixed in the main pipe; the main pipe is divided into three regions, namely the position close to the injection port, the middle and the position close to the flow outlet. A row of sand particles is placed in the middle of each region, and the distance between adjacent sand particles is 30-120 μm; the main pipe is 944 μm wide and 120 μm deep, and the test sand is standard sand whose main component is silicon dioxide (particle size is 120-210 μm). For another example, the document "Visual research on microbial reinforcement based on microfluidic chip technology" (He Xiang, Ma Guoliang, Wang Yang, Zhao Chang, Liu Hanlong, Chu Jian, Xiao Yang; Chinese Journal of Geotechnical Engineering, DOI: 10.11779 / CJGE202006003) also discloses a microfluidic chip, the pipe is made of polydimethylsiloxane (PDMS) inverted mold curing, the bonding layer is cured PDMS sheet, and the glass slide is high-transmittance glass. The mold plate used for the pipe is a silicon wafer processed by photoetching technology. The pipe is a Y-shaped structure, the two side branch pipes are sample inlets; the width is 437 μm, the main pipe width is 944 μm, the end is a sample outlet, and the pipe depth is 120 μm. The test sand is standard sand with a particle size of 50-100 μm. The sand particles are filled into the pipe by a curved tip tweezers. The sand-filled pipe, bonding layer and glass slide are bonded after being treated by a plasma cleaning machine to form a microfluidic chip.

[0005] However, although the above microfluidic chips all design porous media such as sand particles on the glass slide, this structure can observe and analyze the MICP reaction process at a microscale by using the chip, but the microstructure of the porous medium is mostly fixed and single in scale, the structure parameters of the porous medium are uncontrollable, the arbitrary design and adjustment of the porous medium cannot be realized, and the influence of different particle sizes, porosities and other structure parameters on the MICP process cannot be studied. SUMMARY

[0006] The present application provides a parameter-adjustable porous medium MICP process micro-observation system and method, which can not only observe and analyze the MICP process of the porous medium structure at a microscale, but also control the structure parameters of the porous medium, study the influence of different particle sizes, porosities and other structure parameters on the MICP process, and meanwhile, is matched with a feature edge detection threshold value segmentation image post-processing technology based on contrast enhancement, so that the MICP process reaction and deposition process can be captured and quantitatively analyzed in real time. The specific implementation is as follows.

[0007] The application provides a parameter-adjustable micromechanism observation method for a porous medium MICP process, comprising the following steps:

[0008] A random porous medium structure pattern is designed to obtain a microfluidic chip containing the random porous medium structure;

[0009] An original image of the random porous medium structure on the microfluidic chip is acquired, and the original image of the random porous medium structure is processed to obtain a processed image of the random porous medium structure; the processing process comprises mirror flipping, matrix splicing, grayscale processing, first global contrast reduction, first local contrast enhancement, edge recognition detection, first morphological processing, center original image extraction, second morphological processing, noise reduction processing, edge smoothing processing and binary processing;

[0010] After the MICP process is started, an original image of the MICP process at each time is acquired, the processed image of the random porous medium structure is matched and overlaid on the original image of the MICP process to obtain an original overlay image;

[0011] The original overlay image is subjected to second global contrast reduction, second local contrast enhancement and block division Otsu image threshold segmentation processing to obtain a final processed image of the MICP process.

[0012] Further, the random porous medium structure pattern is generated in the following manner:

[0013] The particle size range, pore ratio, throat diameter and length limit, particle fitting edge number, structure boundary size and randomness control parameter of the random porous medium structure are determined;

[0014] The design region is randomly divided according to the structure boundary size, particle size range and pore ratio to obtain a plurality of grid structures of arbitrary shapes, the maximum diameter and the minimum diameter of each grid structure are within the set particle size range, and each grid structure corresponds to one particle; the center point of each grid structure is set as the center of the particle, and the pores between adjacent particles are determined according to the particle size range and throat size; a polygon is generated according to the particle fitting edge number, and each edge of the polygon is spliced into a closed and edge-smoothed particle shape by an inscribed circular arc;

[0015] The throat is designed according to the throat diameter and length limit, and each pore is connected to at least one throat;

[0016] The simulated annealing algorithm is used to make the generated particles uniformly distributed, and the diameter and length of the throat are adjusted to obtain the shape and distribution results of the particles, pores and throats, and the random porous medium structure image is drawn.

[0017] Further, the method for processing the random porous medium structure original image to obtain a random porous medium structure processed image is:

[0018] The random porous medium structure original image is processed by mirror flipping, matrix splicing, and grayscale processing.

[0019] The first global contrast reduction processing is performed by using global histogram equalization, and the first local contrast enhancement processing is performed by using limited contrast adaptive histogram equalization; the first local contrast enhancement processing is performed at least once.

[0020] Edge recognition detection is performed on objects in the image, and the first morphological processing is performed on edges of all detected objects; the first morphological processing includes first erosion processing, first inflation processing, and filling processing; a sub-image with the same size as the random porous medium structure original image is extracted by cutting a center part of the image;

[0021] The second morphological processing is performed on the sub-image, and the second morphological processing includes second erosion processing and second inflation processing.

[0022] Denoising processing, edge smoothing processing, and binarization processing are performed to obtain a binarization image.

[0023] Further, the method for the first inflation processing and the second inflation processing is: a first circular structural element with a corresponding radius is determined according to an actual situation, and a corresponding pixel is added or expanded around each pixel of the image subjected to edge recognition detection by using the first circular structural element.

[0024] Further, the method for the first erosion processing and the second erosion processing is: a second circular structural element with a corresponding radius is determined according to an actual situation, and a corresponding pixel is removed or contracted around each pixel of the image subjected to the first inflation processing or the second inflation processing by using the second circular structural element.

[0025] Further, the second global contrast reduction processing is performed by using global histogram equalization, and the second local contrast enhancement processing is performed by using limited contrast adaptive histogram equalization; the second local contrast enhancement processing is performed at least once.

[0026] Further, the method for performing the block division Otsu image threshold segmentation processing on the image subjected to the second local contrast enhancement is:

[0027] The image subjected to the second local contrast enhancement is divided into a plurality of sub-regions according to predetermined row and column numbers;

[0028] The length and width of each sub-region are calculated, and image segmentation is performed on each sub-region by using two nested for loops to traverse each sub-region;

[0029] The starting row and column and the ending row and column of the current sub-region are calculated, and the image information of the current sub-region is extracted;

[0030] The Otsu threshold value of the current sub-region is calculated, binary image segmentation is performed using the Otsu threshold value, and the pixel value of the binary image is converted to the range of [0, 255];

[0031] The segmentation results of all sub-regions are assigned to complete the segmentation of the entire image, and the MICP process processing image is obtained.

[0032] Further, the method for matching and covering the random porous medium structure processing image to the MICP process original image is: extracting the white pixel points in the random porous medium structure processing image, determining the coordinates of the white pixel points, adjusting the gray value of the corresponding position of the white pixel points in the MICP process original image to 255, so that the particles with unclear edges or blurred particles in the MICP process original image are replaced by the particles with clear edges in the random porous medium structure processing image, and an original covering image is obtained.

[0033] In the above-mentioned parameter-adjustable porous medium MICP process microscopic observation method, the final MICP process processing image obtained is used for subsequent identification and analysis of particles, pores and calcium carbonate precipitation in the MICP process.

[0034] The application also provides a parameter-adjustable porous medium MICP process microscopic observation system for performing any one of the above-mentioned parameter-adjustable porous medium MICP process microscopic observation methods; comprising a microfluidic chip, a syringe pump, a waste liquid collection container, a microscopic observation module, and an image processing module.

[0035] The syringe pump injects bacterial liquid and cementing liquid into the microfluidic chip;

[0036] The waste liquid collection container recovers waste liquid flowing out of the microfluidic chip;

[0037] The microscopic observation module collects the random porous medium structure original image of the random porous medium structure part in the microfluidic chip and the MICP process original image at each time point;

[0038] The image processing module processes the random porous medium structure original image to obtain a random porous medium structure processed image; an original coverage image is obtained, and the original coverage image is processed to obtain a final MICP process processed image.

[0039] Further, one end of the microfluidic chip is provided with a No. 1 liquid inlet channel and a No. 2 liquid inlet channel symmetrically distributed in a V shape, the middle part is provided with the random porous medium structure, and the other end is provided with a liquid outlet channel; the No. 1 liquid inlet channel and the No. 2 liquid inlet channel are respectively connected with a syringe pump, and the liquid outlet channel is connected with the waste liquid collector.

[0040] The random porous medium structure is prepared by etching the chip surface based on the generated random porous medium structure pattern.

[0041] Optionally, the chip is etched in an ultraviolet lithography manner.

[0042] Specifically, the operation steps of the above-mentioned parameter-adjustable porous medium MICP process microscopic observation system are as follows: first, the prepared microfluidic chip is fixed and installed on the objective table of a microscopic observation module (for example, an inverted optical microscope), a syringe pump (for example, a double-channel microsyringe pump) is connected with the No. 1 liquid inlet channel and the No. 2 liquid inlet channel through a catheter, the liquid outlet channel is connected to the waste liquid collector (for example, a waste liquid collection dish) through a catheter, and an image processing module (an image data acquisition computer) is connected with the microscopic observation module.

[0043] Then, according to a predetermined rate, the bacterial liquid and the cementing liquid are injected at the same time, respectively, after the cementing is completed, the MICP process image of the random porous medium structure is observed through the microscopic observation module, the MICP process image is transmitted to the image processing module, and the image processing module analyzes and processes the MICP process image.

[0044] Compared with the prior art, the present application has the following advantages:

[0045] 1. The parameter-adjustable porous medium MICP process microscopic observation system provided by the present application can accurately regulate the micro parameters such as particle size and porosity of the random porous medium structure according to actual needs; one or a plurality of random porous medium structures with different micro parameters can be arranged on the same microfluidic chip, so that the behavior of the MICP process under different microstructure conditions can be systematically studied on the same test platform. Under the condition that the micro parameters are determined, the shape and position of each particle of the random porous medium structure are randomly generated, which can meet various test requirements and is not unique, and the chip can be mass-produced with high efficiency.

[0046] 2. The microfluidic chip structure provided by the application realizes sufficient mixing of injected bacteria liquid and cementing liquid before entering the reaction area, improves the uniformity and repeatability of the reaction, effectively reduces the error caused by uneven fluid distribution in the test process, and ensures the stability and accuracy of the MICP cementing effect.

[0047] 3. In the application, the image data of the precipitation process are collected and recorded by the microscopic observation module (inverted optical microscope) and the image processing module (image data acquisition computer), so that the MICP process in the porous medium structure is captured in real time. The image processing module can effectively solve the problems of blurred particle and precipitation boundary recognition and many noises at the edge of the field of view caused by precipitation adsorption, and significantly enhance the image quality based on the threshold segmentation image post-processing technology of contrast enhancement feature edge detection. The particles, pores and calcium carbonate precipitates can be accurately identified by contrast enhancement, edge recognition and threshold segmentation processing of the MICP process image.

[0048] 4. The application integrates the microfluidic chip with random porous medium structure with real-time image acquisition and data processing technology, and constructs a multidisciplinary micro-MICP reaction observation and analysis experimental platform, which can be widely applied in the fields of water conservancy engineering, building materials and environmental remediation, and promotes the optimization of MICP technology and its popularization and application in related fields. BRIEF DESCRIPTION OF DRAWINGS

[0049] Figure 1 The structure diagram of the microfluidic chip in the parameter-adjustable porous medium MICP process micro-observation system provided by the application.

[0050] Figure 2 The preparation process flowchart of the microfluidic chip.

[0051] Figure 3 The structure diagram of the parameter-adjustable porous medium MICP process micro-observation system provided by the application.

[0052] Figure 4 The image generation process diagram of the random porous medium structure on the microfluidic chip in the application.

[0053] Figure 5 The software interface diagram of the random porous medium structure automatically generated after inputting related parameters.

[0054] Figure 6 The method flowchart of MICP process image data acquisition and processing.

[0055] Figure 7For the original image to be flipped up and down and left and right, the matrix spliced to form a spliced image.

[0056] Figures 8-10 Respectively, the original image, the image after contrast enhancement and the image after contrast reduction result.

[0057] Figure 11 For Figure 7 Image result after overall contrast reduction processing.

[0058] Figure 12 And 13 For the local and overall effects of CLAHE processed images.

[0059] Figure 14 For Canny edge detection contrast effect diagram.

[0060] Figure 15 For Canny edge detection single threshold diagram.

[0061] Figure 16 For the extraction of the center original image.

[0062] Figure 17 For the filling, erosion and dilation processing of the image.

[0063] Figure 18 For the contrast diagram after edge erosion processing and dilation processing.

[0064] Figure 19 For the image denoising and edge smoothing processing in sequence. In the figure, Original Image is the particle structure image after erosion and dilation, Denoised Image is the image after median filter denoising, and SmoothedImage is the image after edge smoothing processing.

[0065] Figure 20 , 21 Respectively, the image after binarization processing and image overlay.

[0066] Figure 22 , 23 And 24 are the images after global contrast reduction, local contrast enhancement and block division Otsu image segmentation.

[0067] In the figure, 1 is a first liquid inlet flow channel; 2 is a second liquid inlet flow channel; 3 is a buffer flow channel; 4 is a random porous medium structure; 5 is a liquid outlet flow channel; 6 is a syringe pump; 7 is an image processing module; 8 is a microscopic observation module; 9 is a catheter; 10 is a waste liquid collector. DETAILED DESCRIPTION

[0068] The technical solutions of the present application will be described clearly and completely below. Obviously, the described embodiments are only some of the embodiments of the present application, but not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative efforts fall within the scope of the present application.

[0069] The present application provides a parameter-adjustable porous medium MICP process microscopic observation system, which comprises a microfluidic chip, a syringe pump 6, a waste liquid collection container 10, a microscopic observation module 8, and an image processing module 7.

[0070] The syringe pump 6 injects bacterial liquid and cementing liquid into the microfluidic chip.

[0071] The waste liquid collection container 10 recovers waste liquid flowing out of the microfluidic chip.

[0072] The microscopic observation module 8 collects original images of the random porous medium structure in the microfluidic chip and original images of the MICP process at each time point.

[0073] The image processing module 7 processes the original images of the random porous medium structure to obtain processed images of the random porous medium structure, obtains an original coverage image, and processes the original coverage image to obtain a final processed image of the MICP process.

[0074] One end of the microfluidic chip is provided with a No. 1 liquid inlet channel 1 and a No. 2 liquid inlet channel 2 symmetrically distributed in a "V" shape. The No. 1 liquid inlet channel 1 and the No. 2 liquid inlet channel 2 are located on the upper surface of the microfluidic chip and are used to inject bacterial liquid and cementing liquid respectively and simultaneously and accurately. Alternatively, the No. 1 liquid inlet channel 1 and the No. 2 liquid inlet channel 2 are both narrow and wide, which facilitates the mixing and reaction of the two kinds of liquid.

[0075] The middle part of the microfluidic chip is provided with the random porous medium structure 4, which is used to simulate the microstructure of the particle system under different conditions and can be set with different structural parameters such as particle size and porosity according to actual needs.

[0076] The other end of the microfluidic chip is provided with a liquid outlet channel 5. The No. 1 liquid inlet channel and the No. 2 liquid inlet channel are connected with the syringe pump respectively. The liquid outlet channel is connected with the waste liquid collector and is used to discharge and collect waste liquid after reaction.

[0077] A buffer channel 3 is arranged between the intersection of the No. 1 liquid inlet channel 1 and the No. 2 liquid inlet channel 2 and the random porous medium structure 4, which facilitates the mixing and reaction of the bacterial liquid and the cementing liquid after entering the chip.

[0078] The random porous medium structure is prepared by first generating a random porous medium structure pattern and then etching the chip according to the random porous medium structure image.

[0079] In the above parameter-adjustable porous medium MICP process microscopic observation system, the microscopic observation module is configured to capture random porous medium structure original images and MICP process original images of the surface of the random porous medium structure, and a common microscope can be selected; and the image processing module is generally configured to use a computer or the like to process the random porous medium structure original images and the MICP process original images.

[0080] Optionally, the generation and preparation technology of the random porous medium structure in the microfluidic chip can be realized by using a common etching method in the industry. In some embodiments of the present application, the random porous medium structure can be obtained by using ultraviolet lithography.

[0081] The microfluidic chip with the random porous medium structure is prepared by etching the surface of the chip based on the generated random porous medium structure pattern. Optionally, the chip is etched by using ultraviolet lithography. As an embodiment, the specific preparation method is as follows:

[0082] (1) To meet the experimental requirements, the random porous medium structure pattern with one or more porosities and particle sizes is designed and generated by considering the fluid dynamics characteristics, porosity and distribution uniformity.

[0083] (2) According to the designed random porous medium structure pattern, a mask is made by using a high-resolution photomask.

[0084] The mask uses a quartz substrate, and a layer of photoresist resistant to the lithography pattern is coated on the surface. The designed pattern is transferred by using a precision exposure technology.

[0085] (3) A high-purity silicon wafer is selected as the substrate, and the silicon wafer is cleaned with a Piranha solution (a mixture of H2SO4 and H2O2) to remove surface impurities. A layer of photoresist (SU-8) is spin-coated on the surface of the silicon wafer.

[0086] (4) Soft baking is performed on the photolithography machine to remove the solvent in the photoresist and improve the adhesion and film quality of the photoresist.

[0087] (5) The prepared mask is aligned with the photoresist-coated silicon wafer, and ultraviolet light with a wavelength of 365 nm is used for exposure to transfer the designed pattern to the photoresist.

[0088] (6) After exposure, the developer (PGMEA) is used for development to remove the substrate material at the position not protected by the photoresist, and the required microstructure pattern is formed.

[0089] (7) According to the design requirements, the silicon wafer is etched by plasma etching method to form the microfluidic channel and the random porous medium structure 4.

[0090] (8) 25 g of PDMS base material and 2.5 g of crosslinking agent (mass ratio 10:1) are added to a mixing container, and a motorized mechanical stirrer is used to stir at medium speed for 10 min to ensure uniform mixing of the mixture.

[0091] (9) The mixed PDMS liquid is transferred to a measuring cup to avoid introducing additional air bubbles, and is placed in a vacuum degassing machine, and a negative pressure of-25 kPa is applied for continuous degassing until there are no obvious air bubbles in the mixed liquid.

[0092] (10) The degassed PDMS mixture is evenly poured onto the pre-prepared silicon wafer mold to ensure that the entire mold surface is covered.

[0093] At the same time, part of the PDMS mixture is poured on 6-7 glass slides to form a spare PDMS film.

[0094] (11) The culture dish containing PDMS and glass slides are placed in a drying oven and heated to solidify at 75°C for 3 hours; after solidification, the microfluidic chip is carefully peeled off from the solidified PDMS block to avoid damaging the microstructure.

[0095] (12) A high-precision puncher device is used to make a first liquid inlet, a second liquid inlet and a liquid outlet corresponding to a first liquid inlet, a second liquid inlet and a liquid outlet at the predetermined position of the PDMS chip; then the chip is placed in a plasma cleaning machine and treated with oxygen plasma at a power of 100 W for 30 s;

[0096] (13) The PDMS chip treated by plasma is quickly pressed against the glass slide and left to stand at room temperature for a few minutes to achieve firm adhesion.

[0097] (14) Finally, the integrity and accuracy of the microfluidic channel and the porous structure are checked using a microscope or scanning electron microscope (SEM), and the chip is cut into 15 mm x 3 mm. At this point, the microfluidic chip with a random porous medium structure is prepared and subsequent test tests can be carried out.

[0098] The use method of the above-mentioned adjustable porous medium MICP process micro-observation system is:

[0099] First, install the microfluidic chip, syringe pump 6, waste collection container 10, microscopic observation module 8, image processing module 7, and place the microfluidic chip on the specific position of the microscopic observation module 8; load the prepared bacteria liquid and cementing liquid into the two syringe pumps 6 respectively, connect the No. 1 liquid inlet channel 1 and the No. 2 liquid inlet channel 2 of the microfluidic chip with the two syringe pumps 6 respectively through two conduits 9, and connect the liquid outlet channel 5 with the waste collection dish 10 through another conduit 9, and finally connect the microscopic observation module 8 with the image processing module 7, thus the parameter controllable random porous medium structure MICP process microscopic observation system is connected.

[0100] Then, open the two syringe pumps 6, and simultaneously inject the bacteria liquid and the cementing liquid into the microfluidic chip from the No. 1 liquid inlet channel 1 and the No. 2 liquid inlet channel 2 respectively, and the bacteria liquid and the cementing liquid are first mixed and reacted in the buffer channel 3, and then enter the random porous medium structure 4 to start the MICP reaction and deposition process. The waste liquid after the reaction is discharged to the waste liquid collection container 10 through the liquid outlet channel 5.

[0101] Third, during the MICP process, the reaction and deposition in the random porous medium structure 4 can be observed through the inverted optical microscope (microscopic observation module 8), and the cementing condition at different times can be captured through the image processing module 7. For the obtained cementing image, the image processing module 7 performs image data acquisition and processing on the original image of the random porous medium structure and the original image of the MICP process at each time, and identifies the particles, pores and calcium carbonate deposits.

[0102] The parameter controllable porous medium MICP process microscopic observation system and method provided by the application can effectively solve the problems of blurred boundary recognition of particles and deposits caused by deposit adsorption, and many noises at the edge of the field of view, significantly enhance the image quality, accurately identify the particles, pores and calcium carbonate deposits, and quantitatively analyze the plugging and distribution process of the calcium carbonate deposits, thereby providing a way for in-depth exploration of the microscopic mechanism and action mechanism of the MICP process.

[0103] I. Generation of random porous medium structure image

[0104] Before preparing the microfluidic chip with the random porous medium structure, in order to etch to obtain the random porous medium structure meeting the actual observation requirements, it is necessary to design a random porous medium structure image.

[0105] 1. Parameter confirmation

[0106] The particle size range, pore ratio, throat diameter and length limit, particle fitting edge number, structure boundary size and randomness control parameters in the random porous medium structure are determined.

[0107] Particle size range: Set the minimum and maximum size of the particles to determine the size of the particles in the random porous media structure.

[0108] Pore fraction: Set the total area or volume of the pores as a proportion of the overall structure, affecting the pore density.

[0109] Throat diameter and length constraints: Define the diameter and length of the channels between adjacent pores, affecting the fluid flow performance.

[0110] Particle fitting edge number: Control the smoothness of the fitted particles through polygon fitting techniques. Lower edge numbers result in sharper shapes, while higher edge numbers generate smoother particles.

[0111] Structure boundary dimensions: Define the length and width of the boundary of the porous structure.

[0112] Randomness control parameters (random seed): Set the same or different random seeds to control the repeatability of the random process while enhancing the diversity of the structure.

[0113] 2. Generation of random porous media structures

[0114] Based on the above parameters input by the user, the image of the random porous media structure that meets the requirements is gradually constructed. The detailed generation process includes the following steps:

[0115] (1) Initialization and parameter preparation

[0116] The following parameters are obtained from the user input interface.

[0117] Particle size range: Minimum size (d min) and maximum size (d max) of the particles.

[0118] Pore fraction: Define the total volume or area of the pore region as a proportion of the entire structure (represented as porosity φ).

[0119] Throat diameter and length constraints: Set the limiting diameter (d throat) and length (l throat) of the channels between pores.

[0120] Particle fitting edge number: Set the number of edges of the fitted polygon of the particles to control the smoothness of the particles.

[0121] Structure boundary dimensions: Length (L) and width (W) of the boundary of the porous structure.

[0122] Randomness control parameters: Choose whether to enable the random seed to control the randomness of the generated structure.

[0123] (2) Preliminary layout and particle generation

[0124] Mesh initialization and size control: Based on the structure boundary size (L, W), particle size range, and porosity ratio, the design area is randomly divided into several arbitrary-shaped mesh structures, each with a maximum diameter (d max ) and a minimum diameter (d min ) within the set particle size range, and the generated size should meet the porosity ratio requirement; each mesh structure corresponds to one particle.

[0125] Particle position initialization: The center point of each mesh structure is set as the center of the particle, ensuring sufficient spacing between them to prevent overlap. This spacing can be determined based on the particle size range and throat diameter and length restrictions, ensuring a certain porosity between each particle for subsequent generation of pores and throats.

[0126] According to the particle size range (i.e., the minimum and maximum size of the set particles) and throat size, the porosity between adjacent particles is determined.

[0127] Particle shape generation: Generate a fitted polygon (polygon) according to the particle fitting edge number parameter; then, draw the inscribed circle arc of each edge of the polygon, and splice the inscribed circle arc of each edge of the polygon to form a closed shape, and convert these polygons into edge-smoothed particle shapes.

[0128] (3) Generation of pores and throats

[0129] Porosity calculation: According to the user input porosity, adjust the number and size of the generated particles. Ensure that the total area or volume of the pores in the entire region accounts for the required value.

[0130] Throat size control: According to the diameter (d throat) and length (l throat) restrictions of the throat, design the connecting channels (i.e., throats) between particles. The diameter of the throat should be suitable for fluid passage and connect adjacent pores. The length of the throat is achieved by adjusting the spacing between two adjacent particles, ensuring the flow performance of the throat.

[0131] Pore and throat connection: Connect all particles through throats to form a connected porous network. Connection can be adjusted by setting a connection factor to ensure that each pore has at least one throat connected to other pores.

[0132] The connection factor is a parameter that measures the connectivity between pores and throats. It defines the average number of throats connected to each pore, directly affecting the flow characteristics of fluid in the porous medium. The specific value of the connection factor depends on the permeability requirements of the target structure and the required flow uniformity.

[0133] The connection factor is calculated as follows: First, count the initial generated pores and measure the distance between them. Then, according to the throat diameter and length limit, determine how many adjacent pores each pore can be connected to. To ensure the connectivity of the structure, set the minimum connection factor CF min , that is, ensure that each pore is connected to at least CF min adjacent pores.

[0134] The actual setting of the connection factor should avoid the case of too high or too low connection factor in local area, in order to ensure the overall flow uniformity. Too high connection factor will increase the number of throats, making the fluid flow path complex, which may affect the directionality of the channel, and too low connection factor may cause some pores to be in isolated state, affecting the overall connectivity.

[0135] (4) Optimize the structure layout

[0136] Pore distribution optimization: use simulated annealing algorithm to fine-tune the generated particles, to ensure the uniformity of the overall pore structure. The optimization goal is to ensure that the distance between pores meets the diameter requirement of the throat; make the pore distribution more uniform, avoid the appearance of dense and sparse areas.

[0137] Simulated annealing algorithm is an optimization algorithm based on the physical annealing process, widely used in random structure optimization problems. In the design process of this porous medium structure, the main goal of the simulated annealing algorithm is to optimize the distribution of pores, making them more uniform, and ensuring that the pore spacing meets the diameter and length requirements of the throat.

[0138] The main optimization process of simulated annealing algorithm is as follows:

[0139] Take the randomly generated initial porous structure layout as the initial state. Establish an energy evaluation function, that is, define a target function E(S) to quantify the "energy" of the current structure, reflecting the non-uniformity of the structure and the degree of violation of the design constraints. The target function includes three parts: overlap penalty term, uniformity penalty term, and throat deviation penalty term. The overlap penalty term (E 重叠惩罚 ) refers to any two particles, if the center distance is lower than the set minimum distance (determined by the respective radius and the required throat diameter), a larger penalty is given. The uniformity penalty term (E 均匀性惩罚 ) is to measure the deviation of the distance between each particle and its adjacent particles from the ideal spacing, to ensure the uniformity of the overall layout, and if the deviation is large, a larger penalty is given. The throat deviation penalty term (E 喉道偏差惩罚 ) is to increase the additional penalty if the actual size (diameter or length) of the channel (throat) formed between the particles deviates too much from the user set value.

[0140] ;

[0141] where w1, w2, w3 are weight coefficients to balance the importance of different penalty terms.

[0142] The energy function comprehensively considers the deviation of particle spacing from the target spacing, the deviation of throat size from the target size, the presence of particle overlap, and the uniformity of pore distribution.

[0143] Subsequently, initial temperature, temperature decay coefficient a, and termination conditions are set to allow the system to explore various layout schemes with a high probability and avoid falling into local optimal solutions. In each iteration of the algorithm, a small perturbation is randomly applied to the particle positions, and the energy function value corresponding to the perturbed layout is calculated. If the new state energy is lower, the state is accepted as the current layout. If the new state energy is higher, it is accepted with a certain probability (which gradually decreases as the temperature decreases), thereby avoiding falling into local optimal solutions.

[0144] As the iteration continues, the temperature parameter gradually decreases, and the perturbation amplitude decreases accordingly. The system tends to be stable for better states. Finally, when the temperature parameter decreases to the preset termination temperature or the maximum number of iterations is reached, the algorithm terminates, obtaining an optimized porous structure layout that satisfies the uniformity of pore distribution, the absence of particle overlap, and the throat size meeting the design requirements. Through this process, the simulated annealing algorithm can efficiently complete the layout optimization of the porous structure, ensuring that the structure layout not only meets the design indicators but also significantly improves the fluid transport performance of the microfluidic chip.

[0145] Throat optimization: Adjust the diameter and length of the throat using the simulated annealing algorithm to ensure they meet the user's set restrictions and ensure the fluid channel is unobstructed.

[0146] Finally, the shape and distribution of particles, pores, and throats are obtained.

[0147] (5) Generate the final random porous medium structure pattern

[0148] Pattern rendering: Based on the optimized particle layout and structure, draw the final random porous medium structure pattern and generate the corresponding 2D pattern.

[0149] Structure saving and exporting: Save the generated random porous medium structure pattern in the file format specified by the user. Generate SVG files through the svgwrite library or DXF files through the ezdxf library. Suitable for subsequent import into CAD software or L-edit tools for chip image conversion. The generated structure can be stored in various file formats, including.svg or.dxf, suitable for general CAD design software.

[0150] Design the throat according to the throat diameter and length restrictions, connecting each pore to at least one throat.

[0151] II. Method for data acquisition and processing of original image of random porous medium structure

[0152] In the process of microscopic observation of MICP process by using the parameter-adjustable porous medium MICP process microscopic observation system, the image processing module carries out image data acquisition and processing on the MICP process image. The method includes five parts, namely image preprocessing, image enhancement, edge detection, first morphological processing, denoising processing, edge smoothing processing and binarization processing.

[0153] Based on the original image with poor quality, the image contrast is improved by multiple image enhancement processing, the image quality is enhanced by image morphology and denoising processing, and high-quality binarization of the image is realized by image matching and partition threshold segmentation processing.

[0154] 1. Image preprocessing

[0155] The preprocessing of the MICP process image mainly includes mirror flipping up and down or left and right, matrix splicing and other processing of the original color image of random porous medium structure, and then converting it into a gray image. The input format of the original image of random porous medium structure can be selected from common images such as tif, tiff and bmp.

[0156] Usually, the particles at the edge of the image can not be completely recognized when the canny edge is recognized because only half or part of the boundary is recognized. In order to solve the problem that the edge of the solid particles at the image boundary is difficult to detect and recognize, the original image of random porous medium structure is first flipped up and down and left and right, and the flipped image is matrix spliced to form a 3x3 spliced image. As shown in Figure 7 After mirror flipping, the particles at the edge of the image can be accurately recognized.

[0157] Then, the obtained color spliced image is grayed to convert it into a gray image to simplify the image processing process.

[0158] A gray image is an image containing only gray values, and the gray value represents the brightness of the pixel. The gray scale formula can convert the RGB value of each pixel in a color image into a gray value, thereby obtaining a gray image. The commonly used gray scale method is the weighted average value method, and the calculation formula is:

[0159] ;

[0160] In the formula, I gray is the pixel value after graying; I R , I G and I BRed, green, and blue channels of the original image corresponding to the pixel value of the pixel.

[0161] 2. Image enhancement

[0162] The process of image enhancement mainly includes global contrast reduction processing (first global contrast reduction) on the image after grayscale processing, and several times of local contrast enhancement processing (first local contrast enhancement). Global contrast reduction processing can improve the visibility of the overall edge of the image, and make the overall gray value of the image more uniform and stable. Several times of local contrast enhancement processing can redistribute the gray levels of the image, so that the brightness value in the image is more uniform, thereby improving the visual quality of the image. Through the above series of image enhancement processing, the subsequent particle solid edge detection is more accurate, and the edge misjudgment is reduced.

[0163] The degree of global contrast reduction and the number of local enhancement times and other parameters can be adjusted according to actual needs.

[0164] (1) First global contrast reduction processing (using global histogram equalization)

[0165] When the input picture is unevenly distributed due to shooting light problems, it will seriously affect the subsequent picture processing effect. By using the image enhancement technology of global histogram equalization, the gray level distribution of the entire image can be adjusted (increased or decreased) to adjust the contrast of the image.

[0166] Using the global histogram adjustment function, the following formula is used to adjust the gray value of each pixel in each image:

[0167] ;

[0168] In the formula, x is the pixel value in the input image; γ is the optional gamma correction parameter, and the default value is 1; low in , high in are the gray scale ranges in the input image; low out , high out are the gray scale ranges in the output image.

[0169] Figures 8-10 The original image, the image after global contrast enhancement, and the image after contrast reduction are respectively shown. Figure 7 The image after global contrast reduction is shown in Figure 11 .

[0170] (2) First local contrast enhancement (using limited contrast adaptive histogram equalization, CLAHE)

[0171] Contrast Limited Adaptive Histogram Equalization (CLAHE) is an image enhancement algorithm that is mainly used to improve the brightness and contrast of local images, and better handle the differences between different regions. This method can significantly improve the quality of the image, making the details more clearly visible.

[0172] CLAHE divides the image into different blocks or windows and performs independent histogram equalization on each block, which can better adapt to the local contrast changes in the image. Finally, the equalized regions are recombined into a complete image using interpolation.

[0173] For each block in the image, the following histogram equalization formula is used to calculate the mapping function:

[0174] ;

[0175] In the formula, L is the total number of gray levels in the block; CDFi(r) is the local cumulative distribution function; r is the original gray value; i represents the current block. The pixels within each block are remapped to new gray levels through this function, thereby achieving local contrast equalization.

[0176] Figure 12 and 13 The effect of CLAHE processing the image is shown. From Figure 12 we can see that the input image's gray scale distribution chart shows that the pixel value is limited to the range of 125-160, and after processing by the CLAHE algorithm, the image's gray scale histogram shows that the pixels are evenly distributed in the range of 0-255, achieving local contrast enhancement of the image. From Figure 13 we can see that the brightness values of the images after the first, second and third enhancement are more uniform, thereby improving the visual quality of the image.

[0177] 3. Edge detection

[0178] Edge detection and morphological processing mainly implement edge detection, dilation processing, filling operation and center image extraction on the enhanced image.

[0179] Canny edge detection algorithm is a commonly used image processing algorithm, which is used to detect the edges of solid particles or other objects in the image; it is widely used in image processing tasks such as target detection, image segmentation and feature extraction. The main goal of Canny algorithm is to detect edges as accurately as possible while minimizing noise and unnecessary edges.

[0180] First, a threshold range (low threshold and high threshold) is artificially defined, which will be used in the subsequent Canny edge detection. The threshold range is generally an empirical value, and further fine-tuning is made according to the effect of edge recognition.

[0181] Then, a Gaussian filter is applied to the input image to reduce the effect of noise, and the calculation formula is:

[0182] ;

[0183] In the formula, G(x, y) represents the value of the Gaussian filter at coordinates (x, y); σ is the standard deviation of the filter.

[0184] Then, the Sobel operator is used to calculate the gradient amplitude and gradient direction of each pixel point in the image. The gradient amplitude represents the edge strength in the image, and the gradient direction represents the direction of the edge. The Sobel operator formula is:

[0185] ;

[0186] The calculation formula of the gradient amplitude M(x, y) and the gradient direction θ(x, y) is:

[0187] ;

[0188] In the formula, G x and G y are the gradient values in the x and y directions, respectively.

[0189] In the gradient image, only the pixels with local maximum gradient amplitude are retained to refine the edges and eliminate non-maximum points. Through the high threshold and low threshold of the set threshold range, the pixels are divided into strong edges, weak edges and non-edges. Only when the pixel value is greater than the high threshold, it is considered as a strong edge. The pixels between the low threshold and the high threshold are considered as weak edges, which may be real edges; non-edge pixels are excluded.

[0190] The above operation process can be calculated by the non-maximum suppression formula and the double threshold processing formula:

[0191] ;

[0192] ;

[0193] In the formula, M L is the low threshold of the gradient strength; M H is the high threshold of the gradient strength, δ x and δ y are the offset values of the gradient direction.

[0194] Finally, based on the connectivity analysis and iterative algorithm (Canny edge detection algorithm), the strong edge pixels and the adjacent weak edge pixels are connected to form a complete edge.

[0195] The enhanced image is subjected to edge detection using the Canny edge detection algorithm. The detection process is provided with a loop structure. If the threshold value is set as a range, the edge detection can be performed according to different threshold values according to the threshold interval, and the contrast effect is given to facilitate the selection of the best threshold value. If the threshold value is set as a fixed value, the edge detection of the corresponding threshold value is directly performed according to the threshold value, and the result is output. Through the above steps, the Canny edge detection algorithm can effectively detect the edges in the image and has a certain robustness to noise.

[0196] As shown in Figure 14 , the effect comparison of Canny edge detection under different threshold conditions can be seen. As shown in Figure 15 , the best Canny edge detection effect diagram can be seen.

[0197] 4. First morphological processing

[0198] (1) First erosion processing

[0199] In contrast to the filling processing, the erosion processing reduces the area of the image by removing or shrinking pixels from around each pixel in the image. It can make the edges more refined, the objects more elongated, and it can be used to remove noise, separate overlapping objects, or segment connected regions. The erosion processing also uses a structuring element to define the shape and size of the erosion. The schematic diagram of filling, erosion and dilation processing of the image is shown in Figure 17 .

[0200] The structuring element of the erosion processing can be selected in different shapes and sizes according to actual needs. In this embodiment, a circular structuring element with a radius of 10 is created as the structuring element for erosion processing. The specific process is as follows (the process of dilation is opposite to it):

[0201] ① A circular structuring element is constructed according to the preset radius (for example, 10 pixels).

[0202] A two-dimensional matrix is usually created, in which the pixel value located in the circle (satisfying (x-c x ) 2 +(y-c y ) 2 ≤10 2 , where c x and c y are the centroid coordinates of the circle) is set to 1, and the other positions are set to 0.

[0203] This matrix actually describes a circular region with a diameter of 20 pixels, which will serve as the kernel for subsequent operations.

[0204] ② Process the neighborhood of each pixel (usually foreground pixels) in the input image.

[0205] Align the center of the structuring element (i.e., the center of the circle) to the image pixel being currently processed. Take out a neighborhood region on the image that is the same size as the structuring element.

[0206] ③ For the current neighborhood region, check the image values corresponding to all pixels within the structuring element:

[0207] If all positions in the structuring element that are 1 in the image satisfy a certain condition (e.g., are all foreground values in the image), then retain the foreground status of the current center pixel;

[0208] If one or more pixels in the region covered by the structuring element do not satisfy the condition (e.g., contain background pixels), then set the current center pixel to a background pixel.

[0209] Through the above process, the value of each pixel is determined by the minimum value in its surrounding neighborhood (which can be visualized as a local "AND" operation in the image), so that the overall foreground region is "eroded" or reduced.

[0210] The edges of the eroded image become more contracted, and small noise points are easily removed. The adhesion between objects may also be separated.

[0211] (2) First dilation processing

[0212] Dilation processing is a special case of filling operation, which increases the area of the image by adding or expanding pixels around each pixel in the image. Dilation processing is similar to filling operation, but usually emphasizes expanding the edges of the object.

[0213] Dilation processing also uses a structuring element to define the shape and size of the dilation. Specifically, after detecting the edges of the image, a circular structuring element with a radius of 2 is created to dilate the edge-detected image, thickening the edges and making them more complete and clear. The method of dilation is the same as the principle of the above erosion processing method.

[0214] The specific process is as follows:

[0215] ① Construct a circular structuring element according to a predetermined radius (e.g., 2 pixels).

[0216] A two-dimensional matrix is usually created, where the elements inside the circle (satisfying (x-c x ) 2 +(y-c y )2 ≤10 2 , where c x and c y The pixel value of the circle's centroid coordinates is set to 1, and the values ​​of other positions are set to 0.

[0217] This matrix is ​​actually a circular region with a diameter of 4 pixels, which will serve as the kernel for subsequent dilation operations.

[0218] ② Process the neighborhood of each pixel (usually a foreground pixel) in the input image.

[0219] Align the center (i.e., the center of the circle) of the structuring element with the currently processed image pixel, and extract a neighborhood region of the same size as the structuring element from the image.

[0220] ③ In the current neighborhood region, check the image values ​​corresponding to all positions within the structuring element:

[0221] If at least one pixel in the area covered by the structuring element is a foreground value, then the current center pixel is set as the foreground pixel; otherwise, it is either left as is or set as the background pixel.

[0222] This process can be viewed as a local "OR" operation, where the value of each pixel is determined by the maximum value in its neighborhood, thus causing the overall foreground area to expand outward.

[0223] After the above operations, the edges that were originally detected in the image will become thicker, and small breaks or holes will be filled, making the overall object boundary appear more complete and continuous.

[0224] (3) Filling process

[0225] Fill processing is used to detect and fill background regions in an image that are completely surrounded by foreground elements, thus forming closed object outlines. This operation utilizes image connectivity analysis to determine which background regions constitute "holes." The specific process is as follows:

[0226] ① Starting from the image edge, perform connectivity analysis on all pixels based on their adjacency (e.g., using 8-connectivity) to mark the regions connected to the external background.

[0227] ② For those areas that are not marked as the external background, i.e. "holes" that are completely surrounded by the foreground, fill all their pixel values ​​with the foreground values.

[0228] This operation is equivalent to filling the background areas detected as "holes" in the image, thereby forming a continuous, closed object outline and providing complete shape information for subsequent image processing.

[0229] 5. Extract the original image from the center.

[0230] The purpose of extracting the center original image is to extract the center region in the image after edge detection, expansion, filling and other operations, so as to subsequent processing or analysis.

[0231] The principle of extracting the center original image is to extract the original size image of the center part from the image after the filling processing, and to obtain the image of the center position by indexing a sub-image with the same size as the input image from the spliced image.

[0232] The method of extracting the center image is to calculate the center coordinates and cut out a sub-image with the same size as the input image (i.e. the original image of the random porous medium structure) at the center coordinate position in the original spliced image, that is, the center image, as shown in Figure 16 .

[0233] 6、Second morphological processing

[0234] In order to eliminate the sawtooth of the particle edge and other difficult-to-handle noise points, the extracted center image needs to be sequentially subjected to a second erosion processing and a second expansion processing.

[0235] (1) Second erosion processing

[0236] In contrast to the filling processing, the erosion processing reduces the area of the image by removing or shrinking pixels from around each pixel in the image. It can make the edges more refined, the objects more elongated, and it can be used to remove noise, separate overlapping objects, or segment connected regions. The erosion processing also uses a structure element to define the shape and size of the erosion. The schematic diagram of filling, erosion and expansion processing of the image is shown in Figure 17 .

[0237] In this embodiment, a circular structure element with a radius of 10 is created as the structure element for erosion processing. The structure element for erosion processing can be selected in different shapes and sizes according to actual needs.

[0238] In the erosion operation, a circular structure element with a radius of 10 is used, and the specific process is as follows (the process of expansion is opposite to it):

[0239] ① According to the pre-set radius (for example, 10 pixels), a circular structure element is constructed.

[0240] A two-dimensional matrix is usually created, and the pixel value located in the circle (satisfying (satisfying (x-c x ) 2 +(y-c y ) 2 ≤10 2 , where c x and c y refer to the centroid coordinates of the circle) is set to 1, and the other positions are set to 0.

[0241] This matrix actually describes a circular area with a diameter of 20 pixels, which will serve as the kernel in subsequent operations.

[0242] ② Process the neighborhood of each pixel (usually foreground pixels) in the input image. Align the center of the structuring element (i.e. the center of the circle) to the image pixel currently being processed. Take out a neighborhood region on the image that is the same size as the structuring element.

[0243] ③ For the current neighborhood region, check the image values corresponding to all pixels within the structuring element:

[0244] If all positions in the structuring element that are 1 in the image satisfy a certain condition (e.g. are all foreground values in a binary image), then the foreground status of the current center pixel is preserved;

[0245] If one or more pixels in the region covered by the structuring element do not satisfy the condition (e.g. contain background pixels), then the current center pixel is set to a background pixel.

[0246] ④ Through the above process, the value of each pixel is determined by the minimum value in its surrounding neighborhood (in a binary image, it can be regarded as a local "AND" operation), so that the overall foreground region is "eroded" or reduced.

[0247] The edges of the eroded image become more contracted, and small noise points are easily removed. The adhesion between objects may also be separated.

[0248] (2) Second dilation processing

[0249] In this embodiment, a circular structuring element with a radius of 8 is created as the structuring element for dilation processing. The structuring element for dilation processing can be selected in different shapes and sizes according to actual needs, and the process of dilation processing is opposite to the process of erosion described above.

[0250] Erosion processing can reduce the particles in the image and eliminate small noise. While dilation processing can expand the particles in the image and fill the holes. The selection of the respective structuring elements and the adjustment of the radius will affect the erosion and dilation effects, which need to be adjusted according to the actual application scenario.

[0251] The comparison of edge erosion processing and dilation processing is shown in Figure 18 , where the left image is the image after erosion processing, and the right image is the image after dilation processing.

[0252] 7. Denoising and edge smoothing processing

[0253] The image processed by the above corrosion and expansion is denoised and edge smoothed, and the steps include converting the image to double type, median filter denoising, and applying Gaussian filter for edge smoothing. Specifically:

[0254] ①In order to more flexibly process the gray value range of the image in subsequent processing, the image is converted to double type.

[0255] ②The denoising process adopts median filter denoising, and the size of the median filter is set to [3, 3], indicating that median filtering is performed in a 3x3 neighborhood.

[0256] ③Then a 5x5 Gaussian filter with a standard deviation of 1 is created to perform edge smoothing on the image after median filter denoising.

[0257] The schematic diagram of the above series of processing is shown in Figure 19 ; In the figure, Original Image is the particle structure image after corrosion and expansion, Denoised Image is the image after median filter denoising, and Smoothed Image is the image after edge smoothing.

[0258] The denoising and edge smoothing are to facilitate better application of subsequent image analysis, feature extraction or other algorithms. The specific steps of the processing can be adjusted according to actual needs and image characteristics.

[0259] 8. Binaryzation processing

[0260] The image processed by edge smoothing is binarized. The purpose of binarization is to convert the gray value of the image to one of two values (usually 0 and 1) to facilitate subsequent threshold segmentation or feature extraction operations.

[0261] First, set the binarization threshold, usually between 0 and 1, to more clearly separate the particles and pores in the image. The selection of the threshold may need to be adjusted according to the specific application scenario to obtain the best segmentation effect.

[0262] Then the image is binarized.

[0263] As shown in Figure 20 , the pore and solid parts are clearly distinguished.

[0264] Three, obtain the original image of the MICP process, match the processed image of the random porous medium structure to obtain the original coverage image, and further process to obtain the final MICP process processed image.

[0265] The step mainly includes image covering, second global contrast reduction, second local contrast enhancement, block division Otsu image threshold segmentation processing.

[0266] Firstly, the MICP process original image at each time is obtained, one of the MICP process original images is selected, the random porous medium structure processing image with clear particle edges processed by binaryzation is matched and covered to the MICP process original image with unclear edges or blurred particles to obtain an original covering image, so as to reduce the difficulty of subsequent image segmentation processing.

[0267] Then, the original covering image is subjected to second global contrast reduction and several times of second local contrast enhancement.

[0268] Finally, the image is subjected to region division, and Otsu method is used for image threshold segmentation processing in each sub-region, so that the final segmentation image can be output. The parameters such as contrast reduction degree, local enhancement times and Otsu threshold can be selected and determined according to actual requirements.

[0269] The specific steps are as follows:

[0270] 1. Image covering

[0271] The image after binaryzation is a pore structure image with clear particle edges. Through image covering, the particle solid part in the image with unclear edges or blurred particles can be replaced by the processed particle solid, which can greatly simplify the difficulty of subsequent image segmentation processing.

[0272] Image covering is mainly achieved by replacing the gray value in image B (i.e. the MICP process original image that needs to be processed, which is the original image containing particles and calcium carbonate precipitate collected by the microscopic observation module and the image processing module) with the white pixel in image A (i.e. the random porous medium structure processing image obtained after binaryzation, which is the image with clear particle structure only). The clear particle structure in image A is covered on image B, so that clear particles can also be obtained in image B, and it is easy to distinguish particles, pores and calcium carbonate precipitate later.

[0273] The specific steps are as follows:

[0274] (1) Firstly, two images are read, the random porous medium structure processing image is assigned to image A, and is converted into 8-bit unsigned integer type to ensure consistency in subsequent processing. Image B (i.e. the MICP process original image) is read, and the file path can be adjusted according to the actual storage location.

[0275] (2) Extract the white pixel points in image A, find the row and column coordinates of the white pixel points in the binary image. Convert image B to a grayscale image, use a for loop to traverse the coordinates of the white pixel points, and set the grayscale value of the corresponding position in image B to 255, realizing the merging or superimposition of the two images. The overlaid image is as shown in Figure 21 .

[0276] 2. Second global contrast reduction

[0277] Perform a second global contrast reduction (global histogram equalization) on the overlaid image, which helps to make the image more suitable for specific application scenarios, improve the visibility of the overall edges of the image, and make the overall grayscale value of the image more uniform and stable. The obtained image is as shown in Figure 22 .

[0278] 3. Second local contrast enhancement

[0279] Perform 3 times of local contrast enhancement (CLAHE) on the image after the second global contrast reduction using the CLAHE method, which is used to improve the contrast of the image. This multiple enhancement method can further enhance the contrast of the image, but attention should be paid to the possible introduction of over-enhancement or noise. In actual application, the number of local contrast enhancement can be adjusted according to the actual effect. The image after three times of second local contrast enhancement is as shown in Figure 23 .

[0280] 4. Block division Otsu image threshold segmentation processing

[0281] The Otsu threshold segmentation algorithm aims to split the image into foreground and background; by calculating the grayscale histogram of the image, an optimal threshold is found to divide the image into two parts, so that the intra-class variance after segmentation is minimized or the inter-class variance is maximized. The intra-class variance measures the difference in grayscale values of pixels within the same region, and the inter-class variance measures the difference between the two regions. The advantage of Otsu threshold segmentation processing is that it can automatically determine the threshold without the need for users to provide pre-parameters.

[0282] Generally, the main steps of block division Otsu image threshold segmentation processing are as follows: first, perform grayscale processing on the input image to obtain a grayscale histogram; then, normalize the histogram (to calculate the relative frequency of each grayscale level), calculate the intra-class variance and inter-class variance for each possible threshold (from 0 to 255); according to the relationship between the intra-class variance and the inter-class variance, select an optimal threshold that minimizes the intra-class variance or maximizes the inter-class variance; after traversing all possible thresholds, select the optimal threshold to divide the image into two parts, forming the foreground and background.

[0283] Since the effect of directly performing Otsu image segmentation is difficult to meet the actual demand, the image can be first regionally segmented into several sub-regions, and then the Otsu method is applied to each sub-region for image segmentation. The specific steps are as follows:

[0284] ① Specify the number of rows and columns of image division, for example, divide the image into 3 rows and 3 columns of sub-regions, and obtain the size of the image.

[0285] ② Calculate the size of the row and column of each sub-region. Loop on the original image, and perform image segmentation on each sub-region; use two nested for loops to traverse each sub-region.

[0286] ③ Calculate the starting row and column and the ending row and column of the current sub-region, and extract the image information of the current sub-region.

[0287] ④ Calculate the Otsu threshold value of the current sub-region, and use the Otsu threshold value to perform binary image segmentation, and convert the pixel value of the binary image to the range of [0, 255].

[0288] ⑤ Assign the segmentation result of the sub-region, and finally obtain the segmentation result of the entire image, and finally process the image by MICP. As shown in Figure 24 , the left image is the image before Otsu segmentation, and the right image is the image after segmentation.

[0289] Such image segmentation operation is usually used to divide the image into sub-regions with similar features for further processing or analysis. The specific segmentation effect may be affected by the Otsu method and the size of the sub-region, and can be adjusted according to the actual demand.

[0290] The above specific embodiments describe the implementation of the present application in detail, but the present application is not limited to the specific details in the above embodiments. Within the scope of the claims and technical concepts of the present application, the technical solutions of the present application can be modified and changed in many ways, and these simple modifications all belong to the protection scope of the present application.

Claims

1. A method for microscopic observation of a parameter-adjustable porous medium MICP process, characterized in that, The method comprises the following steps: designing a random porous medium structure pattern, preparing a microfluidic chip containing the random porous medium structure; obtaining a random porous medium structure original image on the microfluidic chip; processing the random porous medium structure original image to obtain a random porous medium structure processed image; the processing process comprises mirror flipping, matrix splicing, grayscale processing, first global contrast reduction, first local contrast enhancement, edge recognition detection, first morphological processing, center original image extraction, second morphological processing, noise reduction processing, edge smoothing processing and binary processing; after starting the MICP process, obtaining MICP process original images at each time, matching and covering the random porous medium structure processed image to the MICP process original images to obtain original overlay images; performing second global contrast reduction, second local contrast enhancement and block division Otsu image threshold segmentation processing on the original overlay images to obtain final MICP process processed images.

2. The parameter-adjustable porous media MICP process microscopic observation method according to claim 1, characterized in that, The random porous medium structure pattern is generated in the following manner: determining the particle size range, pore ratio, throat diameter and length limit, particle fitting edge number, structure boundary size and randomness control parameters of the random porous medium structure; randomly dividing a design region according to the structure boundary size, particle size range and pore ratio to obtain a plurality of grid structures of arbitrary shapes, the maximum diameter and minimum diameter of each grid structure being within the set particle size range, and each grid structure corresponding to one particle; the center point of each grid structure is set as the center of the particle, and the pores between adjacent particles are determined according to the particle size range and throat size; a polygon is generated according to the particle fitting edge number, and each edge of the polygon is spliced into a closed and edge-smoothed particle shape by an inscribed circular arc; designing a throat according to the throat diameter and length limit, and each pore is connected to at least one throat; using the simulated annealing algorithm to make the generated particles uniformly distributed, and adjusting the diameter and length of the throat to obtain the final shape and distribution of the particles, pores and throats, and drawing the random porous medium structure image.

3. The parameter-adjustable porous media MICP process microscopic observation method according to claim 1, characterized in that, The method for processing the random porous medium structure original image to obtain a random porous medium structure processed image is as follows: the random porous medium structure original image is subjected to mirror flipping, matrix splicing and grayscale processing; the first global contrast reduction processing is performed by using global histogram equalization, and the first local contrast enhancement processing is performed by using limited contrast adaptive histogram equalization; the first local contrast enhancement processing is performed at least once; the edge recognition detection is performed on the objects in the image, and the first morphological processing is performed on the edges of all the detected objects; the first morphological processing comprises first erosion processing, first inflation processing and filling processing; a sub-image with the same size as the random porous medium structure original image is obtained by cutting and extracting from the center part of the image; The second morphological processing is performed on the sub-image, and the second morphological processing includes a second erosion processing and a second expansion processing; The noise reduction processing, the edge smoothing processing and the binarization processing are performed to obtain a binarized image.

4. The parameter-adjustable porous media MICP process micro-observation method according to claim 3, characterized in that, The first expansion processing and the second expansion processing are performed by: determining a first circular structural element with a corresponding radius according to actual conditions, and adding or expanding corresponding pixels around each pixel of the image detected through the edge recognition.

5. The parameter-adjustable porous media MICP process micro-observation method according to claim 3, characterized in that, The first erosion processing and the second erosion processing are performed by: determining a second circular structural element with a corresponding radius according to actual conditions, and removing or contracting corresponding pixels around each pixel of the image processed through the first expansion processing or the second expansion processing.

6. The parameter-adjustable porous media MICP process micro-observation method according to claim 1, characterized in that, The second global contrast reduction processing is performed in a global histogram equalization manner, and the second local contrast enhancement processing is performed in a limited contrast adaptive histogram equalization manner; the second local contrast enhancement processing is performed at least once.

7. The parameter-adjustable porous media MICP process microscopic observation method according to claim 1, characterized in that, The method for performing the block division Otsu image threshold segmentation processing on the image processed through the second local contrast enhancement processing is as follows: The image processed through the second local contrast enhancement processing is divided into a plurality of sub-regions according to predetermined row numbers and column numbers. The length and the width of each sub-region are calculated, and the image segmentation is performed on each sub-region on the original image by using two nested for loops to traverse each sub-region. The image information of the current sub-region is extracted by calculating the starting row and column and the ending row and column of the current sub-region. The Otsu threshold of the current sub-region is calculated, the binarized image is segmented by using the Otsu threshold, and the pixel value of the binarized image is converted to the range of [0, 255]. The segmentation results of all sub-regions are assigned to complete the segmentation of the entire image, and the image processed in the MICP process is obtained.

8. The parameter-adjustable porous media MICP process micro-observation method according to claim 1, characterized in that, The method for matching and covering the random porous medium structure processed image to the MICP process original image is as follows: the white pixel points in the random porous medium structure processed image are extracted, the coordinates of the white pixel points are determined, the gray value of the corresponding position of the white pixel points in the MICP process original image is adjusted to 255, so that the particles with unclear edges or fuzzy particles in the MICP process original image are replaced by the particles with clear edges in the random porous medium structure processed image, and an original covered image is obtained.

9. A parameter-adjustable porous media MICP process micro-observation system, characterized in that, The method is used for performing the parameter-adjustable porous medium MICP process microscopic observation method in any one of claims 1-8; the method comprises a microfluidic chip, a syringe pump, a waste liquid collection container, a microscopic observation module, and an image processing module. The syringe pump injects bacterial liquid and cementing liquid into the microfluidic chip. The waste liquid collection container recovers waste liquid flowing out of the microfluidic chip. The microscopic observation module collects the random porous medium structure original image of the random porous medium structure part in the microfluidic chip and the MICP process original image at each time point. The method is used for performing the parameter-adjustable porous medium MICP process microscopic observation method in any one of claims 1-8; the method comprises a microfluidic chip, a syringe pump, a waste liquid collection container, a microscopic observation module, and an image processing module. The syringe pump injects bacterial liquid and cementing liquid into the microfluidic chip. The waste liquid collection container recovers waste liquid flowing out of the microfluidic chip. The microscopic observation module collects the random porous medium structure original image of the random porous medium structure part in the microfluidic chip and the MICP process original image at each time point. The image processing module processes the random porous medium structure original image to obtain a random porous medium structure processed image; an original overlay image is obtained, and the original overlay image is processed to obtain a final MICP process processed image.

10. The parameter-adjustable porous media MICP process micro-observation system according to claim 9, wherein, One end of the microfluidic chip is provided with a "V" shaped symmetrically distributed first liquid inlet channel and a second liquid inlet channel, a middle part is provided with the random porous medium structure, and the other end is provided with a liquid outlet channel; the first liquid inlet channel and the second liquid inlet channel are connected with the injection pump respectively, and the liquid outlet channel is connected with the waste liquid collector.

Images