Image processing and analysis method for observing soil cracking under constant temperature conditions of wetting-drying cycles

Abstract

The image processing and analysis method for observing soil cracking under dry-wet cycle constant temperature conditions relates to the technical fields of geotechnical engineering test and digital image processing, and constructs a complete process of "sample preparation-parameter setting-calibration-constant temperature drying-quantitative spraying wetting-image and working condition synchronous acquisition-standardized pretreatment-fracture network identification-geometric and topological parameter calculation-data verification-report generation-closed loop control", and the whole process does not need manual intervention to ensure the data synchronicity and traceability in different cycle stages; meanwhile, the improved OTSU threshold, morphological closing operation and connected domain screening are introduced in the fracture network identification to ensure that the small cracks and complex fracture networks can be more stably identified; finally, the abnormal value checking, repeatability checking and closed loop control mechanism are established to form a closed loop of test data quality control and equipment operation control, and the test reliability, automation level and result repeatability are improved.

Description

Technical Field

[0001] This invention relates to the field of geotechnical engineering testing and digital image processing technology, and mainly to an image processing and analysis method for observing soil cracking under dry-wet cycle constant temperature conditions. Background Technology

[0002] Soil is prone to shrinkage, expansion, and cracking under cyclical wet-dry cycles, such as rainfall-evaporation, groundwater level fluctuations, and irrigation-drying. The formation and propagation of cracks alter the soil's structural integrity, strength, permeability, and deformation characteristics, thereby inducing engineering problems such as roadbed settlement, slope instability, foundation pit deformation, and dam leakage. Therefore, accurately observing the cracking process of soil under constant temperature conditions of wet-dry cycles and quantitatively describing the development patterns of crack networks are of great significance for revealing the cracking mechanism of soil, optimizing engineering design, and carrying out disaster prevention and control.

[0003] Existing soil cracking test equipment typically treats temperature control, spraying and wetting, weighing and monitoring, and image capture as relatively independent steps, and the test process still requires a lot of manual intervention. Manual weighing, manual spraying, and manual imaging are not only inefficient, but also prone to introducing operational errors, making it difficult to ensure the synchronization and traceability of data at different cycle stages.

[0004] Existing image processing methods mostly rely on general-purpose software for manual or semi-automatic processing. Common workflows include grayscale conversion, fixed-threshold binarization, and simple area statistics. These methods are prone to problems such as missed crack detection, false crack detection, crack contour breakage, and inconsistent parameter calculations when there are fluctuations in lighting, water spray, slight camera vibration, color differences on soil sample surfaces, and interference from biochar particles.

[0005] Existing technologies typically have limited dimensions for characterizing cracks, focusing primarily on crack area or crack ratio. They lack unified quantification of topological features such as crack length, width distribution, endpoints, intersections, angles, and connectivity. Furthermore, they lack outlier removal, repeatability testing, and closed-loop control mechanisms, resulting in room for improvement in the reliability, comparability, and automation level of experimental results. Summary of the Invention

[0006] To address the aforementioned technical problems, the purpose of this invention is to propose an image processing and analysis method for observing soil cracking under constant temperature conditions of wet-dry cycles.

[0007] The objective of this invention is achieved through the following technical solution. The image processing and analysis method for observing soil cracking under constant temperature and dry-wet cycling conditions proposed in this invention includes the following steps:

[0008] S1, Prepare and cure soil samples;

[0009] S2, through the control and analysis system, preset the test temperature, number of dry and wet cycles, target dry quality, target wet quality, spray water volume, image acquisition interval and crack termination threshold;

[0010] S3, place the soil sample on the weighing module and adjust the focus of the image acquisition module to take the initial image, and use the standard ruler to complete the calibration of the actual length of a single pixel and the actual length of the physical ruler;

[0011] S4, the soil sample is dried at a constant temperature in the constant temperature chamber, the weighing module collects the mass in real time, and the image acquisition module collects crack images at preset intervals.

[0012] S5, when the soil sample quality reaches the preset drying target quality, the control analysis system turns off the heating or reduces the heating power and starts the automatic spraying component until the soil sample quality reaches the preset moistening target quality.

[0013] S6 performs effective region cropping, Gaussian filtering, grayscale conversion, improved OTSU threshold segmentation, morphological closing operation, connected component filtering, skeleton extraction and topology modeling on each frame of the crack image.

[0014] S7. Calculate the fracture area, fracture ratio, total fracture length, average width, width standard deviation, number of nodes, intersection angle, and connectivity rate according to the calibration ratio in step 3.

[0015] S8 performs outlier verification, repeatability verification, and original data tracing on the calculation results, and generates an experimental report;

[0016] S9. Based on the quality threshold, crack threshold, or number of cycles, determine whether to continue the test, proceed to the next wet-dry cycle, or terminate the test.

[0017] Furthermore, in step S3, the actual length of a single pixel is obtained by using the correspondence between the actual length of the standard ruler and the image pixel length. The actual length of a single pixel is expressed as:

[0018] Where s is the actual length of a single pixel. The actual length of the standard ruler. This represents the pixel length of the standard ruler in the image.

[0019] Furthermore, the grayscale conversion in step S6 adopts a weighted transformation model, specifically expressed as follows:

[0020]

[0021] Wherein, R, G, and B are the pixel values ​​of the red, green, and blue channels of the original color image, respectively.

[0022] Furthermore, the improved OTSU adaptive threshold binarization in step S6, based on maximizing inter-class variance, introduces an intra-class variance constraint term. ,

[0023] Select the grayscale threshold that maximizes the value. As the optimal segmentation threshold;

[0024] in, For candidate thresholds, For inter-class variance, , The within-class variances for the background class and the crack class, respectively. These are constraint coefficients;

[0025] Generate binary crack image: ;

[0026] Where x: horizontal pixel index, representing the position of the pixel on the horizontal axis of the image; y: vertical pixel index, representing the position of the pixel on the vertical axis of the image.

[0027] Furthermore, the morphological closing operation in step S6:

[0028]

[0029] in, This is the binary crack image after the closing operation. For structural elements, For expansion operations, For erosion calculation.

[0030] Furthermore, the crack area in step S7: Crack ratio: ,

[0031] in, The area of ​​the crack. The fracture ratio, This represents the number of pixels in the crack region. The number of pixels in the effective area. This is the actual length of a single pixel.

[0032] The total length of the crack:

[0033] in, The total length of the crack. This represents the number of adjacent skeleton pixel pairs in the horizontal or vertical direction. This represents the number of diagonally adjacent skeleton pixel pairs.

[0034] The average width of the crack: , Width standard deviation: ;

[0035] in, This represents the number of pixels in the skeleton.

[0036] The angle of the crack: Fractal connectivity: ;

[0037] in, , Let be the fitted direction vector of the two cracks at the intersection point; The number of fractures connected to the main fracture. This represents the total number of independent fractures.

[0038] Furthermore, outlier verification in step S8 includes calculating the mean of time-series data with the same parameter. and standard deviation ,when The corresponding data is marked as an anomaly, and the original image and operating condition data are retained for traceability.

[0039] Furthermore, the repeatability check in step S8 includes randomly selecting 10% of the time-series images, performing preprocessing and parameter calculation three times on the same image, and checking for relative deviations. If the percentage is greater than 1%, then the preprocessing and parameter calculation for that group of images will be re-executed; among which, , and These represent the maximum, minimum, and average values ​​of the repeated calculation results, respectively.

[0040] Based on the foregoing technical solution, the present invention has the following beneficial effects:

[0041] 1. High degree of automation: Temperature control, spraying and humidification, quality acquisition, image capture, image processing, parameter calculation, data verification and report generation are all completed automatically by the control and analysis system, which significantly reduces manual intervention and labor intensity.

[0042] 2. Stable test conditions: The constant temperature chamber has high temperature control accuracy, the spray water volume can be quantitatively controlled, and the weighing module provides real-time feedback on soil sample quality, which can improve the consistency and repeatability of wet-dry cycle conditions.

[0043] 3. High image recognition accuracy: Through noise reduction, illumination compensation, improved threshold segmentation, morphological optimization, connected component screening and skeleton extraction, it can more stably identify micro-cracks and complex crack networks, and improve robustness to biochar particles, primary pores and light fluctuations in soil samples.

[0044] 4. Complete parameter system: It outputs both crack geometric parameters and topological parameters, which can describe the soil cracking process from multiple dimensions such as cracking degree, crack size, width distribution and connectivity structure.

[0045] 5. High data reliability: Through outlier verification, repeatability verification, and original data traceability mechanisms, the credibility, verifiability, and comparability of experimental results are improved.

[0046] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in conjunction with the accompanying drawings. Attached Figure Description

[0047] Figure 1 This is an overall flowchart of the image processing and analysis method for observing soil cracking under constant temperature and dry-wet cycle conditions according to the present invention.

[0048] Figure 2 This is an image processing flowchart of the image processing and analysis method for observing soil cracking under constant temperature and dry-wet cycle conditions according to the present invention.

[0049] Figure 3 This is a binary image of cracks generated in the image processing flow of the image processing and analysis method for observing soil cracking under dry-wet cycle constant temperature conditions according to the present invention.

[0050] Figure 4 This is a schematic diagram of the experimental apparatus for the image processing and analysis method of soil cracking under constant temperature and dry-wet cycle conditions according to the present invention.

[0051] Figure 5 This is a logical relationship diagram of the control analysis system and data analysis module of the image processing and analysis method for observing soil cracking under dry-wet cycle constant temperature conditions according to the present invention.

[0052] Figure 6 This is the data synchronization and closed-loop control logic diagram of the image processing and analysis method for observing soil cracking under dry-wet cycle constant temperature conditions according to the present invention. Detailed Implementation

[0053] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings:

[0054] Image processing and analysis method for soil cracking observed under wet-dry cycle constant temperature conditions, including the following steps:

[0055] S1: Prepare and cure soil samples

[0056] Prepare soil samples of specified size and density according to the test standards. In this embodiment, the soil sample is one of the following: ordinary soil sample, biochar-mixed soil sample, or solidified soil sample, or other improved soil sample. The soil sample shape is a cylindrical soil sample with a diameter of 100 mm and a height of 50 mm, a cubic soil sample with a side length of 150 mm, or other soil samples of the size that match the engineering test requirements. After the soil sample is prepared, it is placed in a constant temperature and humidity environment for curing for no less than 24 hours.

[0057] S2, through the control and analysis system, preset the test temperature, number of dry and wet cycles, target dry quality, target wet quality, spray water volume, image acquisition interval and crack termination threshold;

[0058] S3, place the soil sample on the weighing module and adjust the focus of the image acquisition module; take the initial image and use the standard ruler to calibrate the actual length of a single pixel against the actual length of the physical ruler;

[0059] Specifically: Place the standard ruler and the soil sample observation surface on the same imaging plane, and obtain the actual length of a single pixel by matching the actual length of the standard ruler with the pixel length of the image. and the actual area of ​​a single pixel The calibration accuracy is no less than 0.01 mm / pixel; the actual length of a single pixel is expressed as: Actual area of ​​a single pixel .

[0060] Where s is the actual length of a single pixel. The actual length of the standard ruler. This represents the pixel length of the standard ruler in the image.

[0061] S4, the soil sample is dried at a constant temperature in a constant temperature chamber, the weighing module collects the mass in real time, and the image acquisition module collects crack images at preset intervals; the image acquisition interval during the drying stage is 1-60 minutes, and the acquisition interval can be automatically shortened or extended according to the crack development rate.

[0062] S5, when the soil sample quality reaches the preset drying target quality, the control analysis system turns off the heating or reduces the heating power and starts the automatic spraying component until the soil sample quality reaches the preset wetting target quality; the image acquisition interval of the wetting stage is 0.5-10min, and the acquisition interval can be automatically shortened or extended according to the crack development rate.

[0063] S6 sequentially performs effective region cropping, Gaussian filtering, grayscale conversion, improved OTSU threshold segmentation, morphological closing operation, connected component filtering, skeleton extraction, and topology modeling on each frame of the crack image.

[0064] Specifically: the grayscale conversion uses a weighted transformation model. To preserve the grayscale difference between the fissures and the soil matrix or biochar particles; where R, G, and B are the pixel values ​​of the red, green, and blue channels of the original color image, respectively;

[0065] The improved OTSU adaptive threshold binarization, based on maximizing between-class variance, introduces a within-class variance constraint term and selects a term that makes the variance maximized. Choose the grayscale threshold that maximizes the grayscale value. As the optimal segmentation threshold; where... For candidate thresholds, For inter-class variance, , The within-class variances for the background class and the crack class, respectively. For constraint coefficients, The optimal segmentation threshold is [value].

[0066] like Figure 3 As shown, a binary crack image is first generated:

[0067] ;

[0068] Where x: horizontal pixel index, representing the position of the pixel on the horizontal axis of the image; y: vertical pixel index, representing the position of the pixel on the vertical axis of the image;

[0069] Morphological operation closing operation:

[0070] in, This is the binary crack image after the closing operation. For structural elements, For expansion operations, For erosion calculation;

[0071] Connected domain screening: Connected domain analysis uses the 8-neighbor labeling rule to identify candidate fracture regions, and combines morphological features such as area, aspect ratio, roundness and perimeter to remove original pores, surface particles and isolated noise points; non-fracture connected domains with an area less than 0.1 mm² and an aspect ratio less than 3 are preferred for removal.

[0072] Skeleton extraction: The Zhang-Suen parallel thinning algorithm is used to extract the binary crack region into a crack centerline with a width of one pixel, and the crack endpoints, intersections and branch points are identified based on the neighborhood relationship of the skeleton pixels.

[0073] S7 calculates the fracture area, fracture ratio, total fracture length, average width, width standard deviation, number of nodes, intersection angle, and connectivity based on the calibration ratio; the binary fracture image includes the following parameters: fracture area, fracture ratio.

[0074] The area of ​​the crack: Crack ratio: ,

[0075] in, The area of ​​the crack. The fracture ratio, This represents the number of pixels in the crack region. The number of pixels in the effective area. This is the actual length of a single pixel.

[0076] Total crack length:

[0077] in, The total length of the crack. This represents the number of adjacent skeleton pixel pairs in the horizontal or vertical direction. This represents the number of diagonally adjacent skeleton pixel pairs.

[0078] Average crack width: , Width standard deviation: ;

[0079] in, Number of skeleton pixels

[0080] Crack intersection angle: Fractal connectivity: ;

[0081] in, , The fitted direction vectors of the two cracks at the intersection point; The number of fractures connected to the main fracture. This represents the total number of independent fractures.

[0082] S8 performs outlier verification, repeatability verification, and original data tracing on the calculation results, and generates an experimental report;

[0083] The outlier detection includes calculating the mean value of time-series data with the same parameter. and standard deviation ,when The corresponding data will be marked as an outlier. The i-th observation in the time series data sequence representing the same fracture parameter is used, and the corresponding original image and working condition data are retained for traceability; repeatability verification includes randomly selecting 10% of the time series images, performing preprocessing and parameter calculation three times on the same image, and if the relative deviation is... Greater than 1% (where δ is the relative deviation, , and If the maximum, minimum, and average values ​​of the repeated calculation results are respectively the maximum, minimum, and average values, then the preprocessing and parameter calculation of that set of images will be re-executed; the test report includes basic test information, parameter settings, temperature and quality process data, fracture parameter table, fracture rate and width evolution curves, fracture network map, abnormal data records, and conclusions and suggestions, and supports exporting to Word, Excel, PDF, and JPG formats.

[0084] S9. Based on the quality threshold, crack threshold, or number of cycles, determine whether to continue the test, proceed to the next wet-dry cycle, or terminate the test.

[0085] Closed-loop control includes a quality threshold closed loop and a fracture threshold closed loop. The quality threshold closed loop controls the switching between drying and spraying processes based on the real-time quality control of the soil sample. The fracture threshold closed loop determines whether to terminate the test based on the fracture rate, average width, connectivity, or number of cycles. Specifically, as shown... Figure 6 As shown, data such as images, quality, temperature, and spray volume are synchronized through a unified timestamp, and the continuation or termination of the experiment is controlled by a threshold.

[0086] For a better explanation of this plan, please refer to [link / reference]. Figure 4 The present invention also provides an apparatus for image processing and analysis of soil cracking under constant temperature and dry conditions, comprising a hardware layer: a constant temperature chamber, an automatic spraying assembly, a weighing module, an image acquisition module, and a data acquisition card, and a software layer: a control and analysis system;

[0087] The constant temperature chamber has an insulated chamber body, heating elements, and temperature sensors to create a preset constant temperature environment inside the chamber. The insulation material of the constant temperature chamber is polyurethane insulation board or equivalent insulation material. The heating elements are heating wires, heating plates, or hot air circulation components that are evenly laid on the inner wall or in the interlayer of the chamber. The temperature control accuracy is ±0.1℃, and the temperature distribution uniformity error inside the chamber does not exceed 1℃ (the control and analysis system adjusts the start and stop of the heating elements according to the real-time feedback from the temperature sensor to keep the temperature inside the chamber stable within the preset range).

[0088] The automatic spraying assembly is positioned above the soil sample and includes a water source, a water pump, a water pipeline, a flow meter, and a misting nozzle. It is used to uniformly wet the surface of the soil sample according to a preset spraying water volume. The spraying water volume of the automatic spraying assembly is adjustable from 10 to 100 ml / min. The flow meter is used to provide real-time feedback on the spraying water volume, with a measurement accuracy of ±0.1 ml / min. The nozzle is a misting nozzle, a multi-hole nozzle, or a linear spraying nozzle. The spraying range covers the effective surface area of ​​the soil sample, achieving uniform wetting of the soil sample surface and avoiding misjudgment of crack images caused by local water accumulation.

[0089] The weighing module is located at the bottom of the constant temperature chamber and carries the soil sample, and is used to collect the soil sample mass in real time. The weighing module is a high-precision electronic scale or weighing sensor array with a measurement accuracy of ±0.01g and a sampling frequency of not less than 1Hz. The weighing data is stored synchronously with temperature data, spray flow data and image data.

[0090] The image acquisition module is located inside the constant temperature chamber and is used to take timed or continuous images of surface cracks in soil samples. The image acquisition module is a high-definition camera with a resolution of no less than 1920×1080 pixels and a focal length adjustment range of 10-50cm. The shooting methods include timed shooting and continuous recording, and its optical axis is kept perpendicular to the soil sample observation surface, or is kept equivalently perpendicular after perspective correction.

[0091] Please see Figure 5 The data acquisition card is connected to a temperature sensor, a weighing module, a flow meter, and an image acquisition module, respectively, to generate multi-source data with a unified timestamp;

[0092] The control and analysis system is used to complete parameter setting, equipment linkage control, image preprocessing, crack identification, skeleton extraction, crack parameter calculation, data verification, visualization archiving, report generation, and closed-loop control; it also includes an illumination compensation module, an image distortion correction module, and a data access control module.

[0093] The visualization archive is used to automatically generate curves showing the changes in fracture rate, total fracture length, average width, number of nodes, intersection angle distribution, and connectivity over time or number of wet-dry cycles, and to generate a fracture network map; the report generation is used to integrate basic test information, equipment parameters, operating data, original images, processed images, fracture parameter tables, abnormal data records, and conclusions and recommendations into a standardized test report.

[0094] To further explain the present invention, an embodiment is provided: a wet-dry cycle splitting experiment of a common soil sample.

[0095] 1. Device Assembly: A constant temperature chamber is made of polyurethane insulation board with dimensions of 80cm×50cm×60cm. Heating wires are evenly laid on the inner wall of the chamber with a spacing of 8cm. A temperature sensor is installed in the middle of the chamber, an atomizing nozzle and a high-definition camera are installed on the top of the chamber, and a high-precision electronic scale is placed in the center of the bottom of the chamber. The temperature sensor, electronic scale, flow meter and camera are connected to the computer through a data acquisition card.

[0096] 2. Parameter settings: The test temperature is set to 25℃, and the temperature control accuracy is ±0.1℃; the spray water volume is set to 50ml / min; the target mass for drying is set to 90% of the initial mass, and the target mass for wetting is set to 98% of the initial mass; the image acquisition interval during the drying stage is 10min, the image acquisition interval during the wetting stage is 2min, and the preset number of dry and wet cycles is 10.

[0097] 3. Test Run: Place the cured soil sample on an electronic scale, take initial images, and complete dimensional calibration. The system starts constant temperature drying, records temperature and mass in real time, and takes images at preset intervals. When the soil sample mass reaches the target dry mass, the system automatically stops heating and starts spraying; when the soil sample mass reaches the target moistening mass, spraying stops, completing one cycle.

[0098] 4. Image Analysis: For each frame of image, the following steps are performed sequentially: effective region cropping, Gaussian filtering, standard grayscale conversion, improved OTSU threshold segmentation, morphological closing operation, connected component filtering, skeleton extraction, and parameter calculation. The output parameters include the gap ratio, total gap length, average width, number of nodes, and intersection angle.

[0099] 5. Result Output: The system analyzes the crack parameters. Outlier and repeatability checks are performed, and curves showing the change in crack rate and average width with the number of cycles are generated (visual archive), and test reports are exported (report generation).

[0100] The present invention also provides another embodiment, a wet-dry cycle splitting experiment of biochar-mixed soil sample (comparative):

[0101] 1. Using the apparatus and process of the first embodiment (1-5 of the first embodiment), multiple groups of soil samples with different biochar content are set up, such as 0%, 5%, 10% and 15%; each group of soil samples uses the same preparation method and curing conditions to ensure that the initial moisture content, density and size are consistent.

[0102] 2. In response to the significant color difference of surface particles in biochar-mixed soil samples, the control analysis system enabled illumination compensation and improved OTSU threshold segmentation. In the connected component screening, area, aspect ratio, and roundness were combined to eliminate pseudo-cracks caused by biochar particles or primary pores.

[0103] 3. By comparing the evolution curves of crack ratio, average width, total crack length and connectivity under different biochar content, the influence of biochar content on the degree of soil cracking and crack network structure can be obtained.

[0104] 4. If the fracture rate of a certain group of soil samples reaches the preset threshold or the maximum fracture width exceeds the set threshold, the control and analysis system will automatically terminate the test for that group and generate an independent report; test groups that do not reach the threshold will continue to complete the preset number of cycles.

[0105] Of course, in other embodiments of the present invention, tests with soil samples of different temperatures and sizes can be conducted:

[0106] 1. Set different temperature gradients such as 20℃, 25℃, and 30℃ within the temperature control range of the constant temperature chamber, and conduct comparative tests using the same soil sample size and number of wet-dry cycles to analyze the effect of temperature on drying rate and crack development speed.

[0107] 2. For cylindrical soil samples with a diameter of 100mm and a height of 50mm and cubic soil samples with a side length of 150mm, the camera focal length and effective area cropping range were adjusted respectively, and the pixel-physical size calibration was re-performed to ensure that the parameter calculations of soil samples of different sizes are comparable.

[0108] 3. The system can archive and compare multiple sets of test data in parallel, and output comparison charts of crack parameters under different temperatures, sizes and modified materials.

[0109] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Any other modifications or equivalent substitutions made by those skilled in the art to the technical solutions of the present invention, as long as they do not depart from the design and scope of the technical solutions of the present invention, should be covered within the scope of the claims of the present invention.

Claims

1. A method for image processing and analysis of soil cracking observed under constant temperature and dry-wet cycling conditions, characterized by: Includes the following steps: S1, Prepare and cure soil samples; S2, through the control and analysis system, preset the test temperature, number of dry and wet cycles, target dry quality, target wet quality, spray water volume, image acquisition interval and crack termination threshold; S3, place the soil sample on the weighing module and adjust the focus of the image acquisition module to take the initial image, and use the standard ruler to complete the calibration of the actual length of a single pixel and the actual length of the physical ruler; S4, the soil sample is dried at a constant temperature in the constant temperature chamber, the weighing module collects the mass in real time, and the image acquisition module collects crack images at preset intervals. S5, when the soil sample quality reaches the preset drying target quality, the control analysis system turns off the heating or reduces the heating power and starts the automatic spraying component until the soil sample quality reaches the preset moistening target quality. S6 performs effective region cropping, Gaussian filtering, grayscale conversion, improved OTSU threshold segmentation, morphological closing operation, connected component filtering, skeleton extraction and topology modeling on each frame of the crack image. S7. Calculate the fracture area, fracture ratio, total fracture length, average width, width standard deviation, number of nodes, intersection angle, and connectivity rate according to the calibration ratio in step 3. S8 performs outlier verification, repeatability verification, and original data tracing on the calculation results, and generates an experimental report; S9. Based on the quality threshold, crack threshold, or number of cycles, determine whether to continue the test, proceed to the next wet-dry cycle, or terminate the test.

2. The image processing and analysis method for observing soil cracking under constant temperature and dry-wet cycling conditions according to claim 1, characterized in that: In step S3, the actual length of a single pixel is obtained by using the correspondence between the actual length of the standard ruler and the pixel length of the image. The actual length of a single pixel is expressed as: Where s is the actual length of a single pixel. The actual length of the standard ruler. This represents the pixel length of the standard ruler in the image.

3. The image processing and analysis method for observing soil cracking under constant temperature and dry-wet cycling conditions according to claim 1, characterized in that: The grayscale conversion in step S6 adopts a weighted transformation model, specifically expressed as follows: Wherein, R, G, and B are the pixel values ​​of the red, green, and blue channels of the original color image, respectively.

4. The image processing and analysis method for observing soil cracking under constant temperature and dry-wet cycling conditions according to claim 1, characterized in that: The improved OTSU adaptive threshold binarization in step S6 is based on maximizing the inter-class variance and introduces an intra-class variance constraint term. , Select the grayscale threshold that maximizes the value. As the optimal segmentation threshold; in, For candidate thresholds, For inter-class variance, , The within-class variances for the background class and the crack class, respectively. These are constraint coefficients; Generate binary crack image: ; Where x: horizontal pixel index, representing the position of the pixel on the horizontal axis of the image; y: vertical pixel index, representing the position of the pixel on the vertical axis of the image.

5. The image processing and analysis method for observing soil cracking under constant temperature and dry-wet cycling conditions according to claim 1, characterized in that: Morphological closing operation in step S6: in, This is the binary crack image after the closing operation. For structural elements, For expansion operations, For erosion calculation.

6. The image processing and analysis method for observing soil cracking under constant temperature and dry-wet cycling conditions according to claim 1, characterized in that: The crack area in step S7: Crack ratio: , in, The area of ​​the crack. The fracture ratio, This represents the number of pixels in the crack region. The number of pixels in the effective area. This is the actual length of a single pixel; The total length of the crack: in, The total length of the crack. This represents the number of adjacent skeleton pixel pairs in the horizontal or vertical direction. This represents the number of diagonally adjacent skeleton pixel pairs. The average width of the crack: , Width standard deviation: ; in, This represents the number of pixels in the skeleton. The angle of the crack: Fractal connectivity: ; in, , Let be the fitted direction vector of the two cracks at the intersection point. The number of fractures connected to the main fracture. This represents the total number of independent fractures.

7. The image processing and analysis method for observing soil cracking under constant temperature conditions of wet-dry cycling according to claim 1, characterized in that: Outlier verification in step S8 includes calculating the mean of time series data with the same parameter. and standard deviation ,when The corresponding data is marked as an anomaly, and the original image and operating condition data are retained for traceability.

8. The image processing and analysis method for observing soil cracking under constant temperature and dry-wet cycling conditions according to claim 1, characterized in that: The repeatability check in step S8 involves randomly selecting 10% of the time-series images, performing preprocessing and parameter calculation three times on the same image, and checking for relative deviations. If the percentage is greater than 1%, then the preprocessing and parameter calculation for that group of images will be re-executed; among which, , and These represent the maximum, minimum, and average values ​​of the repeated calculation results, respectively.

Images