Method for monitoring rapid water flow in porous media using X-ray projection

Abstract

The application discloses a method for monitoring rapid water flow in a porous medium by X-ray projection, and aims at solving the technical problem that an existing industrial CT device is difficult to detect rapid flow.The method for monitoring rapid water flow in a porous medium comprises the following steps: 1, dynamic water flow test of a porous medium material; 2, 2D projection image is acquired by using an industrial CT machine; 3, x coordinates of a sample edge in the 2D projection image at different rotation angles are extracted, and the x coordinate values are subjected to sine curve fitting; 4, projection image gray correction; 5, difference domains of adjacent 2D gray images are calculated; 6, boundary noise points are removed; 7, steps 4-6 are repeated to obtain difference domains of N-1 2D projection images; 8, test piece region cutting; 9, dynamic behavior identification and analysis.The application proposes a 2D projection image dynamic flow tracking and corresponding image processing method on the basis of the industrial CT device, realizes rapid monitoring of an industrial CT dynamic process, and reduces noise of difference domain images.

Description

Technical Field

[0001] This invention belongs to the field of porous media seepage testing, specifically relating to a method for monitoring the rapid flow of moisture in porous media using X-ray projection. Background Technology

[0002] Industrial CT equipment is widely used for non-destructive testing of the microstructure of porous media materials, serving as a crucial technology for studying the design methods and performance optimization of these materials. Existing CT equipment can be categorized into two types based on image acquisition modes: one uses a fixed X-ray source and receiving screen, employing a rotating sample to acquire 2D projection images from different angles, which are then used to reconstruct a 3D model—this is the most common approach used in current industrial CT equipment. The other type uses a fixed sample, rotating the X-ray source and receiving screen to acquire 2D projection images from different angles, and then reconstructing a 3D model. This latter method requires higher equipment stability and is more costly, and is only adopted by some research institutions. Regardless of the CT principle, the number of exposures for scanning the 2D image of the sample reaches 1000–2000 times, with a single exposure duration of 0.5–2 seconds, resulting in a scan cycle of 0.5–2 hours. Such a long scan cycle necessitates that the specimen be quasi-static, making it impossible to monitor dynamic processes on a timescale of seconds.

[0003] In fact, rapidly exposed 2D projection images already encompass some information about the sample within that exposure time. The patent titled "Experimental Apparatus for Determining the Oil-Water Contact Angle on Rock Surfaces Using X-ray Projection Method" (Patent No. ZL201710637061.2) uses X-ray projection images to identify the oil-water interface state on the rock surface. By analyzing the orientation of the line connecting the centers of oil and water droplets at different rotation angles, a 2D projection image perpendicular to the X-ray direction is extracted, thereby extracting the oil-water contact angle. This method is a typical example of information extraction using 2D projection images. The exposure time of 2D projection images is typically between 0.5 and 2 seconds, and it covers the internal structural information of the sample at that time point. If the dynamic processes within the 2D projection image information can be identified through analysis, the accuracy of industrial CT in identifying dynamic processes can be reduced to the second level. Summary of the Invention

[0004] This invention addresses the technical challenge of the generally long scanning cycles of existing industrial CT equipment, which make it difficult to detect rapid flow. It proposes a method for monitoring the rapid flow of moisture in porous media using X-ray projection.

[0005] The present invention utilizes X-ray projection to monitor the rapid flow of moisture in porous media, and is implemented according to the following steps:

[0006] Step 1: Experimental Design of Dynamic Moisture Flow in Porous Media Materials

[0007] The porous media material sample is loaded into the container, the X-ray source is aimed at the porous media material sample, and a projection screen is set on the opposite side of the X-ray source to display the projected image. A water injection device is set on the upper part of the porous media material sample, and water is injected into the surface of the porous media material through the water injection device to carry out in-situ dynamic flow test.

[0008] Step 2: Acquisition of 2D Projected Images

[0009] Using an industrial CT scanner, the rotation angle step Δα and exposure time t were determined based on the time scale requirements of the in-situ dynamic flow test. p Then the time resolution of the 2D projected image is t. p Total scan time T c See equation (1);

[0010]

[0011] Thus, a total of N = 360 / Δα 2D projection images (IMG) are obtained;

[0012] Step 3: Tracking the specimen area in the projected image

[0013] Using the diameter span direction of the porous media material sample as the x-axis and the height direction as the y-axis, the x-coordinates of the edge of the porous media material sample in 2D projection images with different rotation angles are extracted. The least squares method is used to fit the x-coordinate values ​​with a sine curve to determine the sine expression of the control points of the porous media material sample area. Based on the diameter span and height of the porous media material sample, the position area of ​​the porous media material sample on each 2D projection image is determined.

[0014] Step 4: Grayscale Correction of Projected Image

[0015] Analysis of the i-th 2D projection image (IMG) i And the (i+1)th 2D projection image IMG i+1 Extract the 2D projection image IMG respectively i And 2D projection image IMG i+1 The gray value G of the air region at the same point in the middle a,i and G a,i+1 In the 2D projection image IMG i And 2D projection image IMG i+1 Select the same point in the container area and extract the grayscale value G. b，i and G b,i+1 According to the following formula (2), the air-marker two-point grayscale correction is used to correct the 2D projection image IMG. i+1 All grayscale values ​​are adjusted to obtain the grayscale-corrected 2D projected image IMG'. i+1 ;

[0016]

[0017] Among them G j,k For IMG i+1 The gray value G' corresponding to coordinates (j, k) in the projected image. j,k For the IMG after grayscale correction i+1 The gray value corresponding to coordinates (j, k) in the projected image;

[0018] Step 5: Calculation of Gray-Scale Difference Domain

[0019] The difference domain (DIF) of adjacent 2D grayscale images is calculated by grayscale subtraction. i The calculation formula is shown in equation (3);

[0020] DIF i =IMG' i+1 -IMG i (3)

[0021] Among them, IMG' i+1 Let IMG be the grayscale matrix of the (i+1)th projection image after grayscale correction. i Let be the grayscale matrix of the i-th projection image;

[0022] Step Six: Boundary Noise Removal

[0023] Formula (4) is used to remove boundary noise, assuming that the distribution of boundary noise is similar in three adjacent projection images;

[0024]

[0025] DIF i For the (i+1)th and ith grayscale images, DIF i-1 For the difference domain between the i-th and (i-1)-th grayscale images, DIF i+1 Let DIF' be the difference domain between the (i+2)th and (i+1)th grayscale images. i The difference domain of the image after removing boundary noise;

[0026] Step 7: Repeat steps 4 to 6 to obtain the difference domain DIF' of N-1 2D projection images;

[0027] Step 8: Cutting the specimen area

[0028] Based on the control points of the porous medium material sample area in step three, the difference domain DIF' of the 2D projection image is used to cut the sample area and extract the target area;

[0029] Step Nine: Dynamic Behavior Recognition and Analysis

[0030] The method of monitoring rapid water flow in porous media is completed by analyzing the gray-scale feature distribution of DIF' in the target area and extracting the region and features where dynamic behavior occurs.

[0031] The method for monitoring rapid water flow in porous media using X-ray projection of the present invention has the following beneficial effects:

[0032] (1) Traditional industrial CT scanning cycles are as long as 0.5 to 2 hours, which means that CT scanning is usually carried out under quasi-static conditions and cannot monitor the dynamic flow of porous media with high speed. Based on industrial CT equipment, this invention proposes a 2D projection image dynamic flow tracking and corresponding image processing method, which realizes rapid monitoring of dynamic processes in industrial CT and makes the time resolution accurate to 1 second, providing a technical approach for monitoring the dynamic flow of porous media.

[0033] (2) This invention employs methods such as same-point tracking, air-marker two-point grayscale correction, and boundary denoising to reduce noise in the difference domain image and improve image quality. Compared with before processing, the noise amplitude is reduced by 80%. Attached Figure Description

[0034] Figure 1 This is a schematic diagram of the porous media dynamic moisture flow testing device in the embodiment, wherein 1-porous media material, 2-water tank, 3-constant flow control valve, 4-X-ray source, and 5-projection screen;

[0035] Figure 2 This is an X-ray 2D projection image from the embodiment;

[0036] Figure 3 This is a grayscale extraction image of the air-marker two points of the projected image in the embodiment, where red dots represent air and yellow dots represent markers;

[0037] Figure 4 The difference domain of the 2D grayscale image in the embodiment;

[0038] Figure 5 The difference domain image DIF' after boundary noise removal in the embodiment 850 and a map of the region of interest;

[0039] Figure 6 The images shown are a pseudo-color map of the dynamic process distribution (left), a threshold segmentation map (middle), and a difference domain distribution map (right) in the embodiment.

[0040] Figure 7 This is a distribution diagram of the average difference domain across all projected images in the example. Detailed Implementation

[0041] Specific Implementation Method 1: This implementation method for monitoring rapid moisture flow in porous media using X-ray projection is carried out according to the following steps:

[0042] Step 1: Experimental Design of Dynamic Moisture Flow in Porous Media Materials

[0043] The porous media material sample is loaded into the container, the X-ray source is aimed at the porous media material sample, and a projection screen is set on the opposite side of the X-ray source to display the projected image. A water injection device is set on the upper part of the porous media material sample, and water is injected into the surface of the porous media material through the water injection device to carry out in-situ dynamic flow test.

[0044] Step 2: Acquisition of 2D Projected Images

[0045] Using an industrial CT scanner, the rotation angle step Δα and exposure time t were determined based on the time scale requirements of the in-situ dynamic flow test. p Then the time resolution of the 2D projected image is t. p Total scan time T c See equation (1);

[0046]

[0047] Thus, a total of N = 360 / Δα 2D projection images (IMG) are obtained;

[0048] Step 3: Tracking the specimen area in the projected image

[0049] Using the diameter span direction of the porous media material sample as the x-axis and the height direction as the y-axis, the x-coordinates of the edge of the porous media material sample in 2D projection images with different rotation angles are extracted. The least squares method is used to fit the x-coordinate values ​​with a sine curve to determine the sine expression of the control points of the porous media material sample area. Based on the diameter span and height of the porous media material sample, the position area of ​​the porous media material sample on each 2D projection image is determined.

[0050] Step 4: Grayscale Correction of Projected Image

[0051] Analysis of the i-th 2D projection image (IMG) i And the (i+1)th 2D projection image IMG i+1 Extract the 2D projection image IMG respectively i And 2D projection image IMG i+1 The gray value G of the air region at the same point in the middle a,i and G a,i+1 In the 2D projection image IMG i And 2D projection image IMG i+1 Select the same point in the container area and extract the grayscale value G. b，i and Gb,i+1 According to the following formula (2), the air-marker two-point grayscale correction is used to correct the 2D projection image IMG. i+1 All grayscale values ​​are adjusted to obtain the grayscale-corrected 2D projected image IMG'. i+1 ;

[0052]

[0053] Among them G j,k For IMG i+1 The gray value G' corresponding to coordinates (j, k) in the projected image. j,k For the IMG after grayscale correction i+1 The gray value corresponding to coordinates (j, k) in the projected image;

[0054] Step 5: Calculation of Gray-Scale Difference Domain

[0055] The difference domain (DIF) of adjacent 2D grayscale images is calculated by grayscale subtraction. i The calculation formula is shown in equation (3);

[0056] DIF i =IMG' i+1 -IMG i (3)

[0057] Among them, IMG' i+1 Let IMG be the grayscale matrix of the (i+1)th projection image after grayscale correction. i Let be the grayscale matrix of the i-th projection image;

[0058] Step Six: Boundary Noise Removal

[0059] Formula (4) is used to remove boundary noise, assuming that the distribution of boundary noise is similar in three adjacent projection images;

[0060]

[0061] DIF i For the (i+1)th and ith grayscale images, DIF i-1 For the difference domain between the i-th and (i-1)-th grayscale images, DIF i+1 Let DIF' be the difference domain between the (i+2)th and (i+1)th grayscale images. i The difference domain of the image after removing boundary noise;

[0062] Step 7: Repeat steps 4 to 6 to obtain the difference domain DIF' of N-1 2D projection images;

[0063] Step 8: Cutting the specimen area

[0064] Based on the control points of the porous medium material sample area in step three, the difference domain DIF' of the 2D projection image is used to cut the sample area and extract the target area;

[0065] Step Nine: Dynamic Behavior Recognition and Analysis

[0066] The method of monitoring rapid water flow in porous media is completed by analyzing the gray-scale feature distribution of DIF' in the target area and extracting the region and features where dynamic behavior occurs.

[0067] Specific Implementation Method Two: This implementation method differs from Specific Implementation Method One in that the container described in step one is made of glass.

[0068] Specific Implementation Method 3: This implementation method differs from Specific Implementation Method 1 or 2 in that the top of the container containing the porous medium material in step 1 is open, and a constant flow control valve is installed on the outlet of the water injection device.

[0069] Specific Implementation Method Four: This implementation method differs from one of the specific implementation methods one to three in that the porous media material sample mentioned in step one is made of sand, gravel, (porous) rock, asphalt concrete, or cement concrete.

[0070] Specific Implementation Method 5: This implementation method differs from Specific Implementation Methods 1 to 4 in that the container containing the porous medium material in step 1 is placed on a rotating table with a rotation speed of 0.3 to 0.6 degrees per second.

[0071] Specific Implementation Method Six: This implementation method differs from Specific Implementation Methods One to Five in that step two uses a voltage of 200kV and a current of 110μA as the radiation intensity of the industrial CT machine.

[0072] Specific Implementation Method Seven: This implementation method differs from Specific Implementation Methods One through Six in that the rotation angle step size Δα in step two is generally 0.18 to 0.72 degrees, and the exposure time t... p The time is typically 0.5 to 2 seconds.

[0073] Specific Implementation Method Eight: This implementation method differs from Specific Implementation Methods One to Seven in that step three extracts the x-coordinate of the sample edge from 20 to 40 2D projection images with different rotation angles.

[0074] Specific Implementation Method Nine: This implementation method differs from Specific Implementation Methods One to Eight in that a pseudo-color image is applied to DIF' in step nine to analyze the spatial distribution and corresponding time points of areas with large grayscale differences, so as to realize the spatial and temporal description of the dynamic flow occurring in the porous medium.

[0075] Specific Implementation Method 10: This implementation method differs from Specific Implementation Methods 1 to 9 in that step 9 extracts the area of ​​the critical grayscale difference region through threshold segmentation to achieve quantitative evaluation of water flow.

[0076] Example: This example demonstrates a method for monitoring rapid water flow in porous media using X-ray projection, implemented according to the following steps:

[0077] Step 1: Experimental Design of Dynamic Moisture Flow in Porous Media

[0078] A sand and gravel material with a particle size range of 2.36–4.75 mm was placed in a cylindrical glass bottle and vibrated on a vibration table for 10 minutes to ensure that the sand and gravel material did not undergo significant displacement during the scanning process. A constant flow rate of 1 mL / min was calibrated using a constant flow control valve, which was placed above the sand and gravel material sample to create a constant flow rate. The structural diagram of the porous media dynamic moisture flow testing device is shown below. Figure 1 As shown;

[0079] Step 2: Acquiring 2D Projected Images

[0080] Using 200kV voltage and 110μA current as the radiation intensity of the industrial CT scanner, with a rotation step size Δα of 0.36° and an exposure time of 1s, and performing two consecutive scans, the total scanning time was 33 minutes, resulting in 2000 2D projection images {IMG}. 2000}, 2D projection image see Figure 2 As shown;

[0081] Step 3: Tracking the specimen area in the projected image

[0082] The 2D projection images were extracted at rotation angles of 0°, 18°, 36°, 54°, ..., 342°, and 360°, respectively. These rotation angles formed an arithmetic sequence. The x-coordinates of the porous media sample edges were extracted from the 2D projection images at different rotation angles. (See figure) Figure 2 As shown in Table 1, the x-coordinates corresponding to different rotation angles are shown in Table 1.

[0083] Table 1. x-coordinates of specimen edges in 2D projection images

[0084]

[0085] The least squares method was used to fit the x-coordinate values ​​to a sine curve to determine the sine expression of the control points of the sample area, as shown in Equation 5.

[0086]

[0087] Where P(β) is the x-coordinate of the control point in the specimen region in the projected image at angle β;

[0088] The diameter of the porous media material sample is 580 pixels, and the height spans the entire image by 1000 pixels. The coordinates of the sample's position region on each 2D projection image are shown in Equation 6.

[0089] REG(x, y)=[P(β)~(P(β)+580), 0~1000] (6)

[0090] Step 4: Grayscale Correction of Projected Image

[0091] With the 850th 2D projection image from IMG 850 And the 851st 2D projection image (IMG) 851 For example, see Figure 3 As shown, the grayscale value G of the air region at the same point in the projection image is extracted respectively. a，850 =252.93 and G a，851 =248.85, extract the grayscale value G from the same point in the test fixture area (glass bottle). b，850 =107.41 and G b，851 =118.35, substituting the above grayscale value into formula (2), we get the following formula (7), for IMG 851 All grayscale values ​​were corrected using air-marker two-point grayscale correction to obtain the grayscale-corrected 2D projection image IMG'. 851 ;

[0092]

[0093] Step 5: Calculation of Gray-Scale Difference Domain

[0094] The difference domain (DIF) of adjacent 2D grayscale images is calculated by grayscale subtraction. 850 The calculation formula is shown in equation (8);

[0095] DIF 850 =IMG' 851 -IMG 850 (8)

[0096] Among them, IMG' i+1 Let IMG be the grayscale matrix of the (i+1)th projection image after grayscale correction. i The grayscale matrix of the i-th projection image is shown in the calculation results. Figure 4 As shown.

[0097] Step Six: Boundary Noise Removal

[0098] Since each angular step Δα corresponds to a new projection angle during the rotation of the specimen, boundary noise inevitably arises when calculating the grayscale domain of adjacent projection images due to the change in position caused by the angular rotation. This embodiment uses formula (9) to remove boundary noise, which assumes that the distribution positions of boundary noise are similar in three adjacent projection images.

[0099]

[0100] DIF 850 For the difference domain between the 851st and 850th grayscale images, DIF 849 For the difference domain of the 850th and 849th grayscale images, DIF 851 For the difference domain between the 852nd and 851st grayscale images, DIF' i The difference domain of the image after removing boundary noise is calculated, and the results are shown in Figure 5 middle.

[0101] Step 7: Repeat steps 4 to 6 to obtain the difference domain DIF' of 1999 2D projection images.

[0102] Step 8: Cutting the specimen area

[0103] Based on the specimen region control points (Equation 6) in step three, the specimen region is segmented using the difference domain DIF' of the 2D projection image to extract the target (region of interest), see [link to relevant documentation]. Figure 5 As shown in the box.

[0104] Step Nine: Dynamic Behavior Recognition and Analysis

[0105] Apply a pseudo-color map to the difference domain of the 2D projected image, see Figure 6 As shown in the figure, the red and yellow areas are the areas where dynamic flow processes occur. It can be seen that in the 850th image, water flows in the lower part of the specimen and fills part of the void structure.

[0106] This embodiment uses sand and gravel as the target porous medium material, and drips it into the sand and gravel material at a constant rate of 1 mL / min. A Phoenix v|tome|x S240 industrial CT scanner is used as the testing equipment to illustrate the specific implementation of the present invention.

[0107] Using a critical difference value of 80 as the critical threshold for judging the dynamic process occurrence domain, the two-dimensional projection region caused by the water flow is extracted, see... Figure 6 As shown, the area is 4755 pixels. 2 The location is concentrated in the lower part of the specimen, and the location of dynamic flow can also be seen from the distribution of the difference domain along the height of the specimen.

[0108] Analyze the average difference values ​​of the difference domain images throughout the entire scanning process, see Figure 7 As shown, dynamic flow occurred in the 850th, 1000th, 1312th, and 1498th scan images during the entire scanning process, corresponding to time points of 850s, 1000s, 1312s, and 1498s, respectively. Figure 7 The projected image shows that the dynamic flow behavior gradually moves upward.

[0109] In summary, the innovation of this invention in using X-ray projection to monitor the rapid flow of moisture in porous media lies in:

[0110] 1) By analyzing the grayscale differences in X-ray projection images, dynamic water flow in porous media is extracted, and the time resolution is reduced from 1 hour for CT imaging to 1 second for projection images.

[0111] Traditional industrial X-ray CT imaging equipment requires acquiring a sufficient number of two-dimensional projection images of the object under test to reconstruct a three-dimensional model. Whether using sample rotation or X-ray source rotation, this process requires acquiring projection images from different angles at specific time steps. The scanning process can last from 0.5 to 2 hours depending on the required precision, thus it is typically a quasi-static scan. This invention utilizes continuously and rapidly exposed 2D projection images from CT scans as data analysis objects. By comparing the differences between 2D projection images from adjacent exposure times during dynamic flow, it identifies the region where dynamic flow occurs within each exposure time. This allows for the identification accuracy of CT dynamic processes down to 1 second (exposure time), enabling rapid tracking of dynamic behavior in porous media.

[0112] 2) By employing image point tracking and air-marker two-point grayscale correction method, the grayscale distribution of different projected images was corrected, and the noise of the projected images was reduced.

[0113] The positional patterns of sample markers in the projected images were analyzed, and the coordinates of the control points of the specimen were fitted with a sine curve to identify the sample area range in the two-dimensional projected images at different time points. A gray-scale correction method of air-marker two-point gray-scale was proposed to correct the gray-scale distribution of the sample area in adjacent projected images, avoiding gray-scale noise caused by the difference in X-ray intensity in different exposures. The gray-scale differences between three consecutive projected images were analyzed to eliminate boundary noise caused by the edge effect during specimen rotation.

[0114] 3) A dynamic water flow monitoring and morphological evaluation method based on grayscale difference domain analysis of projected images is proposed.

[0115] After image denoising, the grayscale differences between two-dimensional projection images at adjacent exposure times are compared and calculated sequentially. A pseudo-color image is used to mark the distribution of grayscale differences, thereby realizing the spatial distribution recognition of the dynamic flow process. The critical threshold segmentation method is used to segment and extract the region where the dynamic flow process occurs. The area of ​​the region is calculated to obtain the flow characteristics of the dynamic flow. The distribution of grayscale differences between adjacent two-dimensional projection images along the scanning time is calculated. Each difference peak represents the time point where the dynamic process occurs, thereby obtaining the distribution of water dynamic flow in the time domain.

Claims

1. A method for monitoring rapid water flow in a porous medium using X-ray projections, characterized in that The method for monitoring the rapid flow of water in a porous medium is realized according to the following steps: Step one, dynamic water flow test design of the porous medium material The porous medium material sample is loaded into a container, an X-ray source is aligned with the porous medium material sample, a projection screen is arranged on the opposite side of the X-ray source to display the projection image, and a water injection device is arranged on the upper part of the porous medium material sample to inject water to the surface layer of the porous medium material for in-situ dynamic flow test; Step two, 2D projection image acquisition The rotation angle step size Δα and the exposure time t are determined according to the time scale requirement of the in-situ dynamic flow test by using an industrial CT machine p The time resolution of the 2D projection image is t p The total scanning time T is c as shown in formula (1) Thus, a total of N = 360 / Δα 2D projection images IMG are obtained; Step three, sample area tracking of the projection image The diameter span direction of the porous medium material sample is taken as the x-axis, and the height direction is taken as the y-axis, the x-coordinate of the edge of the porous medium material sample in the 2D projection image at different rotation angles is extracted, a least square method is used to perform sinusoidal curve fitting on the x-coordinate value, a control point sinusoidal expression of the porous medium material sample area is determined, and the position area of the porous medium material sample on each 2D projection image is determined according to the diameter span and height of the porous medium material sample; Step four, projection image gray scale correction Analysis of the i-th 2D projection image (IMG) i And the (i+1)th 2D projection image IMG i+1 Extract the 2D projection image IMG respectively i And 2D projection image IMG i+1 The gray value G of the air region at the same point in the middle a,i and G a,i+1 In the 2D projection image IMG i And 2D projection image IMG i+1 Select the same point in the container area and extract the grayscale value G. b,i and G b,i+1 According to the following formula (2), the air-marker two-point grayscale correction is used to correct the 2D projection image IMG. i+1 All grayscale values ​​are adjusted to obtain the grayscale-corrected 2D projected image IMG'. i+1 ; where G j,k is IMG i+1 G' is the gray value corresponding to the coordinate (j, k) in the projection image j,k is IMG' i+1 G' is the gray value corresponding to the coordinate (j, k) in the projection image Step five, gray scale difference domain calculation DIF is calculated by subtracting the gray scale of the adjacent 2D gray scale image i The calculation formula is shown in equation (3). DIF i = IMG i+1 - IMG i (3) where IMG i+1 is the (i+1)th projection image gray matrix after gray correction, IMG i is the ith projection image gray matrix; Step six, boundary noise point removal The formula (4) is used to remove the boundary noise points, and it is assumed that the distribution positions of the boundary noise points are similar in the adjacent three projection images; where DIF i is the difference field of the i+1th and the ith gray scale image, DIF i-1 is the difference field of the ith and the i-1th gray scale image, DIF i+1 is the difference field of the i+2th and the i+1th gray scale image, DIF i is the difference field of the image after the boundary noise removal; Step seven, repeating steps four to six to obtain the difference domain DIF' of N-1 2D projection images; Step eight, sample area cutting According to the sample area control points in step three, the difference domain DIF' of the 2D projection image is cut to extract the target area; Step nine, dynamic behavior identification and analysis The gray scale feature distribution of DIF' in the target area is analyzed, the dynamic behavior occurrence area and features are extracted, and thus the method for monitoring the rapid flow of water in a porous medium is completed.

2. The method for monitoring rapid water flow in a porous medium using X-ray projections of claim 1, wherein The material of the container in step one is glass.

3. The method for monitoring rapid water flow in a porous medium using X-ray projections of claim 1, wherein The container with the porous medium material in step one is open at the top, and a constant flow control valve is arranged on the water outlet of the water injection device.

4. The method for monitoring rapid water flow in a porous medium using X-ray projections of claim 1, wherein The material of the porous medium material sample in step one is sandstone, rock, asphalt concrete or cement concrete.

5. The method for monitoring rapid water flow in a porous medium using X-ray projections of claim 1, wherein The container with the porous medium material in step one is placed on a rotating table, and the rotating speed of the rotating table is 0.3-0.6 degrees per second.

6. The method for monitoring rapid water flow in a porous medium using X-ray projections of claim 1, wherein In step two, 200kV voltage and 110μA current are used as the ray intensity of the industrial CT machine.

7. The method for monitoring rapid water flow in a porous medium using X-ray projections of claim 1, wherein The rotation angle step size Δα in step two is 0.18-0.72 degrees, and the exposure time t p is 0.5-2 seconds.

8. The method for monitoring rapid water flow in a porous medium using X-ray projections of claim 1, wherein In step three, 20-40 x-coordinates of the sample edge in the 2D projection image at different rotation angles are extracted.

9. The method for monitoring rapid flow of water in a porous medium using X-ray projections of claim 1, wherein In step nine, a pseudo-color map is applied to DIF', the spatial distribution and corresponding time point of the area with large gray scale difference are analyzed to realize the spatial and temporal description of the dynamic flow in the porous medium.

10. The method for monitoring rapid water flow in a porous medium using X-ray projections of claim 1, wherein In step nine, the critical gray scale difference domain area is extracted by threshold segmentation to realize the quantitative evaluation of water flow.

Images