A method for identifying through-cracks in concrete

Abstract

This invention relates to the field of image processing technology, specifically disclosing a method for identifying through-cracks in concrete. The method involves acquiring and registering reference thermal image sequences from opposite sides of a structural member, identifying the area to be inspected based on local temperature changes, and dividing it into inspection segments. Different humidity levels are alternated at varying time intervals for each inspection segment. Isothermal gases are input from both sides of the member, and forward and reverse thermal image sequences are acquired. The acquired thermal images are compared frame-by-frame with the reference thermal images to extract the temperature transition sequence and position. Based on the input time interval, transition time interval, and the correspondence between the two sides, bidirectional through-crack segments are determined and connected. This invention can distinguish between temperature anomalies caused by differences in moisture content and temperature anomalies caused by the through-crack path, reducing the interference of uneven moisture content distribution on crack penetration judgment.

Description

Technical Field

[0001] This invention relates to the field of image processing technology, and more specifically to a method for identifying through-cracks in concrete. Background Technology

[0002] Infrared thermal imaging detection identifies temperature anomalies by recording surface temperature changes of concrete components under natural temperature differences or external thermal excitation, utilizing the difference in heat conduction between cracked areas and intact concrete. Engineering inspections typically acquire continuous thermal images, analyze the shape, temperature difference, and extent of temperature anomaly areas, and combine this with thermal image information from opposite sides of the component to determine the location and extension state of cracks. It is suitable for crack screening of concrete components such as walls and slabs.

[0003] The surface temperature of concrete is simultaneously affected by internal moisture content, heating conditions, and environmental changes. Due to differences in heat capacity and moisture phase change, locally moist areas may develop continuous, banded temperature anomalies during heating and cooling processes, with morphology similar to the temperature anomalies corresponding to cracks. Cracks may also become pathways for moisture migration, causing the thermal responses generated by cracks and moisture content differences to overlap. When judging the crack penetration status based on the continuous morphology and temperature difference changes of the temperature anomaly area, the judgment result is easily influenced by the distribution of the moist area. Summary of the Invention

[0004] The purpose of this invention is to provide a method for identifying through-cracks in concrete, thereby solving the aforementioned technical problems.

[0005] The objective of this invention can be achieved through the following technical solutions: A method for identifying through-cracks in concrete includes the following steps: S1. Acquire and register reference thermal image sequences of the first and second detection surfaces on opposite sides of the concrete component; S2. Connect adjacent pixels in the reference thermal image sequence whose temperature change difference exceeds the preset limit. Based on the registration result, pair the temperature anomaly regions corresponding to the positions on both sides of the surface as the areas to be inspected, and divide them into inspection segments along their extension direction. S3. In the order of alternating high and low humidity of the test segments, isothermal gas is introduced from the first detection surface, and a forward thermal image sequence of the second detection surface is collected. S4. In the order of alternating high and low humidity, isothermal gas is introduced from the second detection surface, and the reverse thermal image sequence of the first detection surface is collected. S5. For each frame in the forward and reverse thermal image sequences, select thermal images with the same acquisition time in the corresponding reference thermal image sequence, perform pixel-by-pixel temperature difference, and extract the temperature transition order and temperature transition position. S6. Match the alternation order of high and low humidity and the order of temperature transition, match the temperature transition position according to the registration result, and determine the bidirectional through segment; S7. Connect adjacent bidirectional through segments and generate through crack identification results.

[0006] As a further aspect of the present invention, the process of acquiring and registering the reference thermal image sequence is as follows: Multiple thermal markers are set at corresponding edge positions of the first and second detection surfaces, and the thermal markers are heated sequentially according to different heating durations. Simultaneously acquire thermal images of the first and second detection surfaces, and record the start time of temperature rise and the start time of temperature fall in the area where each thermal marker is located; The regions where the start times of temperature rise and temperature fall are paired are determined based on the center pixel coordinates of the paired thermal marker regions, and the rotation, scaling and translation amounts are determined based on the center pixel coordinates of the paired thermal marker regions. Register the thermal images of the first and second detection surfaces according to the rotation, scaling and translation amounts, stop heating the thermal marker, apply the same surface temperature excitation to the first and second detection surfaces, and arrange the continuously acquired thermal images into a reference thermal image sequence according to the acquisition time.

[0007] As a further aspect of the present invention: the process of dividing the inspection segments is as follows: Two thermal images with an average surface temperature difference not exceeding a preset pairing limit and opposite temperature change directions are selected from the reference thermal image sequences of the first and second detection surfaces, respectively. Calculate the temperature difference between each pixel and its adjacent pixels in each frame of thermal image, and subtract the temperature differences of pixels in the same position in two frames of thermal images to obtain the temperature change difference value; Connect adjacent pixels whose temperature change difference exceeds a preset limit to form a temperature anomaly region on the first detection surface and the second detection surface. According to the registration results, the temperature anomaly areas corresponding to each other at the boundary positions are paired, and the paired temperature anomaly areas are recorded as the areas to be inspected. The areas to be inspected are then continuously divided into inspection segments along the extension direction of the areas to be inspected, and the order of each inspection segment is recorded.

[0008] As a further aspect of the present invention: the process of acquiring a forward thermal image sequence is as follows: Each inspection segment in the first inspection surface is sealed separately, and an air supply port is set within the sealing range corresponding to each inspection segment; The periods of gas supply with humidity higher than the ambient humidity and the periods of gas supply with humidity lower than the ambient humidity are divided into multiple continuous periods. The continuous periods are combined according to the alternation of high and low humidity, and the time interval between adjacent humidity transition times in each combination is calculated. Select combinations with different time intervals, and determine the selected combinations as the humidity alternation order for each test segment, so as to keep the total duration of isothermal gas received by each test segment that is higher or lower than the humidity of the test environment the same. A surface temperature excitation corresponding to the reference thermal image sequence is applied to the second detection surface. At the start of the surface temperature excitation, isothermal gas is synchronously input into each inspection segment in alternating order of high and low humidity. The humidity transition time is recorded, and thermal images of the second detection surface are continuously acquired to form a forward thermal image sequence.

[0009] As a further aspect of the present invention: the process of acquiring the reverse thermal imaging sequence is as follows: After completing the forward thermal imaging sequence acquisition, stop the surface temperature excitation, release the seal of each inspection segment in the first detection surface, and input isothermal gas with the same humidity as the detection environment into the first and second detection surfaces; The thermal images of the first and second detection surfaces are continuously acquired. When the pixel-by-pixel temperature difference between the acquired thermal image and the first frame of the corresponding reference thermal image sequence does not exceed the preset reset limit, the input of isothermal gas is stopped. Arrange the inspection segments in reverse chronological order according to the alternating order of high and low humidity used during the forward data acquisition process, and seal each inspection segment on the second detection surface. A surface temperature excitation corresponding to the reference thermal image sequence is applied to the first detection surface. At the start of the surface temperature excitation, isothermal gas is synchronously input into each inspection segment in reverse order of humidity high and low. The humidity transition time is recorded, and thermal images of the first detection surface are continuously acquired to form a reverse thermal image sequence.

[0010] As a further aspect of the present invention: the process of extracting the temperature transition sequence and temperature transition position is as follows: Calculate the elapsed time from the start of surface temperature excitation to the acquisition time of each frame of the forward and reverse thermal imaging sequences, respectively; Based on the elapsed time, each frame of thermal image in the forward thermal image sequence and the reverse thermal image sequence is subtracted pixel by pixel from the thermal image in the corresponding reference thermal image sequence with the same elapsed time to obtain a temperature difference image. Divide the temperature difference image according to the recorded humidity change time, compare the pixels in the same position in adjacent temperature difference images in turn, and record the direction of temperature change of each pixel; Pixels whose temperature change direction changes after a humidity transition and continues until the next humidity transition are retained. The retained pixels are arranged according to the time of the change in temperature change direction to obtain the temperature transition order. Adjacent pixels with the same change time are connected, and the coordinate range of the connected pixels is recorded as the temperature transition position.

[0011] As a further aspect of the present invention: the process of determining the bidirectional through-segmentation is as follows: Calculate the input time interval between adjacent humidity transition times in the alternation sequence of high and low humidity in each test segment, and calculate the transition time interval between adjacent temperature transition times in the forward thermal image sequence and the reverse thermal image sequence. The turning time intervals corresponding to the forward thermal imaging sequence are compared sequentially with the input time intervals corresponding to each test segment, and the turning time intervals corresponding to the reverse thermal imaging sequence are compared sequentially with the input time intervals arranged in reverse order. The test segment records the difference between the input time interval and the transition time interval at the corresponding position, which does not exceed the preset time limit. According to the registration results, the forward temperature transition position is mapped to the first detection surface, and the reverse temperature transition position is mapped to the second detection surface. The inspection segment where the mapped temperature transition position is located is determined. If an inspection segment satisfies both forward and reverse matching, and its temperature transition positions are all located inside the segment, then the segment is determined to be a bidirectional through segment.

[0012] As a further aspect of the present invention: the process of generating the identification result of the penetrating crack is as follows: According to the registration results, the temperature transition positions in the forward thermal image sequence are mapped to the first detection surface, and the temperature transition positions in the reverse thermal image sequence are mapped to the second detection surface. In the first detection surface, the mapped forward temperature transition position and the temperature transition position extracted from the reverse thermal image sequence are paired; in the second detection surface, the temperature transition position extracted from the forward thermal image sequence and the mapped reverse temperature transition position are paired. If the distance between pixels that are paired on both the first and second detection surfaces does not exceed a preset limit, a segmented crack region is formed. Connect the segmented crack regions containing adjacent pixels at the segment boundaries according to the order of the test segments to form a continuous through region. Connect the outer edge pixels of the continuous through region to form the crack boundary, and record the continuous through region and crack boundary as the through crack identification result.

[0013] The beneficial effects of this invention compared to the prior art are as follows: This invention distinguishes between naturally formed temperature anomalies in concrete components and active thermal responses with a defined time sequence, making it difficult for continuous temperature anomalies caused solely by differences in moisture content to meet pre-set time variation relationships. By cross-checking the response results on opposite sides of the component, the interference of unilateral dampness, localized heating, and changes in ambient temperature on the continuity judgment is reduced. The continuity state is characterized by both time interval and spatial location, reducing the impact of differences in initial surface temperature and response delay on the judgment results. Crack regions and crack boundaries are formed in continuous segments to obtain the extension location and continuous range of penetrating cracks on the component surface. Attached Figure Description

[0014] The invention will now be further described with reference to the accompanying drawings.

[0015] Figure 1 This is a flowchart illustrating a method for identifying through-cracks in concrete according to the present invention. Figure 2 This is a schematic diagram of the process for extracting the temperature transition sequence and temperature transition position according to the present invention. Detailed Implementation

[0016] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0017] Please see Figures 1-2 As shown, this invention provides a method for identifying through-cracks in concrete, comprising the following steps: S1. Acquire and register reference thermal image sequences of the first and second detection surfaces on opposite sides of the concrete component; In a preferred embodiment of the present invention, the process of acquiring and registering a reference thermal image sequence is as follows: Two opposing surfaces of the same concrete member are designated as the first and second detection surfaces, respectively. Multiple physical locations are selected along the edge of the concrete member, and each physical location is projected onto the first and second detection surfaces. Thermal markers are then affixed to the projected locations. The thermal markers are sheet-like objects that can be independently heated and form a continuous temperature region in the thermal image. Two thermal markers corresponding to each physical location form a group. Thermal markers in the same group are connected to the same heating control terminal, and heating is started and stopped synchronously. Different groups of thermal markers are heated sequentially, with each group using a different heating duration. The start and stop times of heating for each group are recorded.

[0018] While heating the thermal marker, thermal images of the first and second detection surfaces are continuously acquired using an infrared thermal imager, with the acquisition clocks of the two infrared thermal imagers synchronized. The acquisition time is recorded for each frame of thermal image. Based on the attachment position of the thermal marker, a corresponding pixel range is selected within each frame of thermal image, and the average temperature of each pixel within that range is calculated. The acquisition time when the average temperature first exceeds that of the previous frame and continues to rise in subsequent thermal images is recorded as the start time of temperature rise; the acquisition time when the average temperature first falls below that of the previous frame after heating stops and continues to decrease in subsequent thermal images is recorded as the start time of temperature fall. The areas of the thermal markers on the first and second detection surfaces where the start times of both temperature rise and temperature fall are the same are paired.

[0019] In each paired thermal marker region, pixels whose temperature continuously increased during heating were extracted, and the average abscissa and ordinate of the extracted pixels were calculated. These two averages were used to form the center pixel coordinates of the thermal marker region. Multiple paired regions with different center pixel coordinate intervals were selected, and the center pixel coordinates in the first detection surface and the center pixel coordinates in the second detection surface were connected respectively. The rotation amount was determined based on the angular difference between the directions of the connecting lines on both sides, and the scaling amount was determined based on the ratio between the lengths of the connecting lines on both sides. The center pixel coordinates of the second detection surface were transformed according to the rotation and scaling amounts, and the coordinate difference between the transformed center pixel coordinates and the corresponding center pixel coordinates in the first detection surface was calculated. This coordinate difference was determined as the translation amount.

[0020] The pixel coordinates of each frame of thermal image in the second detection surface are transformed by rotation, scaling, and translation, and the transformed pixel coordinates are mapped to the thermal image coordinates of the first detection surface. When the transformed pixel coordinates are between adjacent pixels, the temperature at the transformation position is determined based on the temperature and coordinate distance of the adjacent pixels. After thermal image registration is completed, heating of the thermal marker is stopped, and the same heat source is placed on both the first and second detection surfaces. Surface temperature excitation is applied according to the same start time, heating duration, and heat source state. Thermal images of the two detection surfaces are continuously acquired, and the thermal images of the first detection surface and the second detection surface after coordinate transformation are arranged according to the acquisition time to form a reference thermal image sequence corresponding to the two detection surfaces.

[0021] It is understood that the first and second detection surfaces are parallel planes or can be approximated as planes. When acquiring thermal images, the position and shooting direction of the infrared thermal imager are fixed, and the coordinate directions of the two thermal images are unified according to the height and length directions of the concrete component. Each group of thermal markers forms a corresponding temperature rise start time and temperature fall start time according to different heating durations, and the regions where the thermal markers are located are paired based on the time correspondence. After stopping heating the thermal markers, thermal images are continuously acquired. When the temperature difference between the region where the thermal marker is located and the adjacent region returns to the preset range, surface temperature excitation is applied again. Surface temperature excitation includes a heating process and a cooling process after heating stops. The heating process and cooling process are completed by continuously acquiring the average temperature of the detection surface based on the thermal image sequence.

[0022] S2. Connect adjacent pixels in the reference thermal image sequence whose temperature change difference exceeds the preset limit. Based on the registration result, pair the temperature anomaly regions corresponding to the positions on both sides of the surface as the areas to be inspected, and divide them into inspection segments along their extension direction. In a preferred embodiment of the present invention, the process of dividing the inspection segments is as follows: Reference thermal image sequences of the first and second detection surfaces are read separately, and each frame of thermal image is processed sequentially according to the acquisition time. The average surface temperature of the corresponding thermal image is obtained by summing the temperatures of all pixels belonging to the detection surface in a frame and dividing by the number of pixels. The average surface temperature at adjacent acquisition times is compared; if the average surface temperature at a later acquisition time is higher than that at a previous acquisition time, the corresponding thermal image is recorded as a rising thermal image; if the average surface temperature at a later acquisition time is lower than that at a previous acquisition time, the corresponding thermal image is recorded as a falling thermal image. A preset pairing limit represents the upper limit of the allowable difference in average surface temperature between two frames of thermal images. The average surface temperature difference between each frame of rising thermal images and each frame of falling thermal images is calculated separately. The frame of rising thermal images and the frame of falling thermal images with the smallest average surface temperature difference not exceeding the preset pairing limit are selected to form two frames of thermal images for the corresponding detection surface.

[0023] For each selected thermal image frame, the temperature of each pixel is read sequentially according to its pixel coordinates, and pixels above, below, to the left, and to the right of the current pixel are identified as adjacent pixels. Pixels located at the edge of the thermal image are selected only as adjacent pixels within the thermal image area. The average temperature of adjacent pixels is calculated, and then the difference between the current pixel temperature and the average temperature of adjacent pixels is calculated. The calculation result is recorded as the local temperature difference of the current pixel in the corresponding thermal image. According to the registered pixel coordinates, the local temperature difference of each pixel in the heated thermal image is subtracted from the local temperature difference of pixels at the same position in the cooled thermal image, and the absolute value of the difference is taken to obtain the temperature change difference of the corresponding pixel. When the average surface temperatures are similar, the temperature change difference represents the local temperature distribution change at the same location during the heating and cooling processes.

[0024] Starting with a pixel whose temperature change difference exceeds a preset limit, examine the adjacent pixels above, below, to the left, and to the right. When the temperature change difference of adjacent pixels also exceeds the preset limit, add the adjacent pixel to the same pixel set, and continue examining the adjacent pixels of the newly added pixel until there are no more adjacent pixels that can be added. Record the continuous range covered by each pixel set as the temperature anomaly region, and record the pixels adjacent to the unadded pixels on the periphery of the pixel set as the boundary pixels of the temperature anomaly region. Obtain the temperature anomaly regions of the first detection surface and the second detection surface using the same process.

[0025] Based on the rotation, scaling, and translation amounts used in thermal image registration, the boundary pixels of the temperature anomaly region in the second detection surface are transformed to the coordinate range of the first detection surface. The two boundary pixels with the largest distance are selected in both the transformed temperature anomaly region and the temperature anomaly region on the first detection surface, and the direction between these two boundary pixels is recorded as the boundary extension direction of the corresponding temperature anomaly region. The angle difference between the two boundary extension directions is calculated, and the boundary pixels of the two temperature anomaly regions are projected onto the boundary extension direction of the temperature anomaly region in the first detection surface. When the two projection ranges overlap, for the temperature anomaly region with fewer boundary pixels, the nearest boundary pixel in the other temperature anomaly region is searched pixel by pixel, and the number of boundary pixels whose nearest distance does not exceed a preset position offset limit is counted. When the angle difference does not exceed a preset direction limit, and the proportion of the counted boundary pixels to the number of boundary pixels participating in the search is not less than a preset boundary pairing ratio, the two temperature anomaly regions are paired. When one temperature anomaly region meets the pairing conditions with multiple temperature anomaly regions, the temperature anomaly region with the largest boundary pairing ratio is selected; when the boundary pairing ratios are the same, the temperature anomaly region with the smallest average distance between paired boundary pixels is selected. Temperature anomaly regions that have already been paired do not participate in other pairings. The coverage area of ​​the paired region on the two detection surfaces is recorded as the area to be inspected. The inspection segments are then divided sequentially along the boundary extension direction of the area to be inspected on the first detection surface. Based on the inverse conversion relationship of rotation, scaling, and translation, the boundary positions of each inspection segment are transferred to the second detection surface, and inspection segments with the same arrangement order are formed in the paired temperature anomaly area on the second detection surface.

[0026] Understandably, after forming a temperature anomaly region, its morphology is examined according to the boundary extension direction. Temperature anomaly regions extending continuously along the surface of the concrete component are designated as inspection areas, while those appearing as scattered blocks or covering large areas are not directly recorded as inspection areas. The inspection areas on the first and second inspection surfaces are established with a positional correspondence based on the registration results. The preset position offset limit is determined based on the difference in opening positions of the confirmed through-cracks on the two inspection surfaces and the coordinate residual after registration of the thermal markers. The preset direction limit is determined based on the angular difference between the boundary extension directions of the same through-crack on the two inspection surfaces. The preset boundary pairing ratio is determined based on the proportion of pixels on both sides of the same through-crack boundary in the registered thermal image of the test specimen that can establish a positional correspondence to the total number of pixels participating in the pairing.

[0027] When determining bidirectional through-segmentation, simultaneously verify whether the temperature inflection point is located within the corresponding inspection area; when generating through-crack identification results, only connect segmented crack areas belonging to the same inspection area and continuous at the segment boundary. Temperature anomaly areas corresponding to known construction joints, component splicing boundaries, and reserved holes are recorded separately and not included in through-crack identification.

[0028] S3. In the order of alternating high and low humidity of the test segments, isothermal gas is introduced from the first detection surface, and a forward thermal image sequence of the second detection surface is collected. In a preferred embodiment of the present invention, the process of acquiring a forward thermal image sequence is as follows: Based on the inspection segment boundaries recorded in the preceding steps, sealing covers are installed segment by segment on the first inspection surface, ensuring that each sealing cover covers only one inspection segment. Sealing material is continuously applied along the inspection segment boundaries at the contact points between the sealing cover and the first inspection surface. The surface area covered inside the sealing cover is defined as the sealing range of the corresponding inspection segment. Each sealing cover is equipped with a gas supply port, and each gas supply port is connected to an independent gas supply pipe. These independent gas supply pipes are connected to a high-humidity gas output terminal and a low-humidity gas output terminal via switching valves, thereby switching the input gas according to the alternating high and low humidity sequence corresponding to each inspection segment.

[0029] The humidity of the ambient air is measured at the detection location, and the measurement result is recorded as the ambient humidity. The gas entering the high-humidity gas output terminal is humidified so that the humidity of the output gas is higher than the ambient humidity; the gas entering the low-humidity gas output terminal is dehumidified so that the humidity of the output gas is lower than the ambient humidity. The gas temperatures at both gas output terminals are adjusted until they match the initial temperature of the first detection surface. Isothermal gas indicates that the high-humidity and low-humidity gases have the same temperature, and the input gas temperature remains unchanged when switching gas humidity levels.

[0030] Multiple gas supply duration periods are pre-set, and the gas supply duration periods above and below the detection ambient humidity are combined according to the alternating humidity levels. Following the order of the duration periods in each combination, the elapsed duration periods are sequentially added from the start of the gas supply to obtain the humidity transition time corresponding to each humidity change. The time interval between adjacent humidity transition times is calculated sequentially, and the results are arranged in chronological order to form the time interval arrangement for each combination. The time interval arrangements of each combination are compared, and combinations with the same time interval arrangement are deleted. The total gas supply duration above and below the detection ambient humidity is then calculated separately for each of the remaining combinations. Combinations where the total gas supply duration above and below the detection ambient humidity are the same across all combinations are selected. These selected combinations are then allocated one by one according to the order of the test segments, forming the humidity alternation order for each test segment. The total gas supply duration above and below the detection ambient humidity are set independently, and it is not required that the two total gas supply durations in the same combination be equal.

[0031] On the second inspection surface, surface temperature excitation is applied according to the heat source location, heating start time, and heating process used when acquiring the reference thermal image sequence, and the imaging position of the infrared thermal imager is fixed. When the surface temperature excitation begins, the gas supply pipes of each inspection segment are simultaneously turned on. Each switching valve is sequentially connected to the high-humidity gas output terminal and the low-humidity gas output terminal according to the corresponding humidity alternation sequence, and the moment when the switching valve completes the gas output terminal switching is recorded as the humidity conversion time. The infrared thermal imager continuously acquires thermal images of the second inspection surface from the start of the surface temperature excitation until each inspection segment completes all gas supply periods in the corresponding humidity alternation sequence. The acquired thermal images are arranged sequentially according to the acquisition time, and the acquisition time of each frame of thermal image is associated with the recorded humidity conversion time to form a forward thermal image sequence.

[0032] Understandably, each inspection segment is equipped with a mutually isolating sealed cover, with each cover covering one inspection segment and connected to an independent gas supply pipe. The gas supply pipes are connected to humidity-adjustable gas output devices, ensuring consistent gas temperature, supply pressure, and flow rate across all supply pipes, while maintaining a sealed gap between adjacent sealed covers. During forward inspection, isothermal gas is input from the first inspection surface, while the second inspection surface remains in communication with the inspection environment. During reverse inspection, isothermal gas is input from the second inspection surface, while the first inspection surface remains in communication with the inspection environment. The supply pressure is set to a level that allows isothermal gas to enter the crack from the supply port without altering the crack's opening state. Before inspection, the edges of the sealed covers are checked, and thermal imaging begins when no gas leakage is detected at the edges.

[0033] S4. In the order of alternating high and low humidity, isothermal gas is introduced from the second detection surface, and the reverse thermal image sequence of the first detection surface is collected. In a preferred embodiment of the present invention, the process of acquiring a reverse thermal imaging sequence is as follows: After completing all gas supply periods and thermal image acquisition in the forward thermal imaging sequence, shut off the heat source of the second detection surface and stop inputting isothermal gas (humidity higher or lower than the detection environment) to the first detection surface. Remove the sealing covers corresponding to each inspection segment on the first detection surface and clean the sealing material between the sealing cover and the first detection surface. According to the arrangement order of the inspection segments, install a gas supply cover on one inspection segment of the first detection surface, and install a gas collection cover on the same sequence of inspection segments on the second detection surface, with a humidity detection component installed at the exhaust port of the gas collection cover. Adjust the humidity of the input gas to be the same as the detection environment humidity, and adjust the input gas temperature to match the initial temperature of the first detection surface. Then, input isothermal gas from the gas supply cover on the first detection surface. After completing the gas input in the current direction, swap the positions of the gas supply cover and the gas collection cover, adjust the input gas temperature to the initial temperature of the second detection surface, and then input isothermal gas with the same humidity as the detection environment from the second detection surface to the first detection surface. Process each inspection segment sequentially according to the same procedure.

[0034] During each isothermal gas input period, the gas humidity at the exhaust port of the gas collection hood is read according to the preset humidity acquisition interval, and the difference between the gas humidity and the ambient humidity is calculated. When the gas humidity difference does not exceed the preset humidity reset limit, and the difference between adjacent humidity readings within a consecutive preset reset observation period does not exceed the preset humidity stability limit, the isothermal gas input in the current gas supply direction is stopped. During isothermal gas input, the shooting positions of the two infrared thermal imagers are fixed, and thermal images of the first and second detection surfaces are acquired synchronously and continuously. For each acquisition moment, the current thermal image of the two detection surfaces and the first frame of the corresponding reference thermal image sequence are read respectively, and the temperature difference between pixels at the same position is calculated according to pixel coordinates and the absolute value is taken. When the gas humidity acquired by each inspection segment in both gas supply directions meets the preset humidity reset limit and the preset humidity stability limit, and the temperature difference of all pixels on the two detection surfaces does not exceed the preset reset limit within a consecutive preset reset observation period, the isothermal gas input is stopped and the gas supply hood and gas collection hood are removed. If any condition is not met, isothermal gas is re-inputted into the corresponding inspection segment according to the two gas supply directions until all conditions are met.

[0035] The humidity alternation sequence used by each test segment during the forward data acquisition process is read separately. Each humidity alternation sequence includes gas humidity states arranged chronologically and their corresponding durations. The last gas humidity state and its corresponding duration in the forward data acquisition process are placed at the top, and then the remaining gas humidity states and their corresponding durations are arranged in reverse chronological order, forming a reversed humidity alternation sequence. Reversing the order changes both the position of the gas humidity states and the position of the durations, but does not change the combination relationship between each gas humidity state and its corresponding duration. The reversed humidity alternation sequence is then reassigned to the corresponding test segments according to their arrangement order.

[0036] Sealing covers are installed along the boundaries of each inspection segment on the second inspection surface, with each cover covering one inspection segment. The gas supply ports of the sealing covers are connected to independent gas supply pipes. The temperatures of the high-humidity gas and the low-humidity gas are adjusted to match the initial temperature of the second inspection surface. Surface temperature excitation is applied to the first inspection surface according to the heat source location, heating start time, and heating process used when acquiring the reference thermal image sequence. When the surface temperature excitation begins, the gas supply pipes of each inspection segment on the second inspection surface are simultaneously turned on, and the input gas is switched in reverse order of alternating high and low humidity. The moment each gas humidity state switch is completed is recorded as the humidity conversion moment, while thermal images of the first inspection surface are continuously acquired. After all gas supply periods are completed, the acquired thermal images are arranged according to the acquisition time, and each frame of thermal image is associated with the humidity conversion moment to form a reverse thermal image sequence.

[0037] Understandably, the preset humidity reset limit represents the maximum permissible difference between the gas humidity at the exhaust port of the gas collection hood and the humidity of the detection environment, while the preset humidity stability limit represents the maximum permissible difference between adjacent humidity readings. The preset humidity reset limit is determined based on the measurement error of the humidity detection component and the fluctuation range of the detection environment humidity. The preset humidity stability limit is determined based on the difference in adjacent humidity readings formed after isothermal gas with the same humidity as the detection environment passes through a penetrating crack during the pre-inspection process. The preset reset observation period is determined based on the longest time elapsed from when isothermal gas is input from one detection surface to when the humidity of the gas collection hood on another detection surface begins to change until the humidity reading stabilizes. After resetting each inspection segment and removing the gas supply hood and gas collection hood, each inspection segment in the second detection surface is then sealed, and a reverse thermal imaging sequence is acquired.

[0038] S5. For each frame in the forward and reverse thermal image sequences, select thermal images with the same acquisition time in the corresponding reference thermal image sequence, perform pixel-by-pixel temperature difference, and extract the temperature transition order and temperature transition position. In a preferred embodiment of the present invention, the process of extracting the temperature transition sequence and temperature transition position is as follows: Read the forward thermal image sequence, the reverse thermal image sequence, the reference thermal image sequence of the first detection surface, and the reference thermal image sequence of the second detection surface, and read the acquisition time of each frame of thermal image in each sequence. Record the start time of surface temperature excitation as the time starting point, and calculate the time length from the time starting point to the acquisition time of each frame of thermal image in the forward and reverse thermal image sequences, respectively, and record the calculation results as the elapsed time of the corresponding thermal image. The elapsed time of the reference thermal image sequence is recorded in the same way. For each frame of thermal image in the forward thermal image sequence, read the thermal image with the same elapsed time from the reference thermal image sequence of the second detection surface; for each frame of thermal image in the reverse thermal image sequence, read the thermal image with the same elapsed time from the reference thermal image sequence of the first detection surface.

[0039] Pixels at the same position in both the detected thermal image and the corresponding reference thermal image are read one by one according to their pixel coordinates. The pixel temperature in the detected thermal image is subtracted from the pixel temperature in the reference thermal image, and the difference is recorded in the same pixel coordinates to obtain a temperature difference image. Pixel values ​​in the temperature difference image retain their positive and negative directions. When a pixel value is greater than the pixel value at the previous acquisition time, the direction of temperature change between adjacent acquisition times is recorded as upward; when a pixel value is less than the pixel value at the previous acquisition time, the direction of temperature change is recorded as downward. The forward and reverse thermal image sequences are converted into temperature difference images arranged according to acquisition times using the above process.

[0040] Read the humidity transition times recorded during both forward and reverse acquisition processes, and classify the temperature difference images acquired between adjacent humidity transition times into the same gas supply period. For each humidity transition, first read the temperature change direction of the same pixel in the last two frames of temperature difference images before the humidity transition, and then read the temperature difference images after the humidity transition sequentially according to the acquisition time. When the temperature change direction of a pixel after the humidity transition is opposite to the temperature change direction before the humidity transition, and the opposite temperature change direction continues until the next humidity transition time, retain the corresponding pixel, and record the thermal image acquisition time when the temperature change direction first changes as the temperature inflection time of the corresponding pixel.

[0041] The retained pixels are arranged in chronological order of temperature transition moments, and the arrangement result is recorded as the temperature transition order. The temperature transition order consists of each temperature transition moment and its corresponding pixel. When the same temperature transition moment contains multiple pixels, starting from each pixel, the adjacent pixels above, below, to the left, and to the right are checked, and adjacent pixels with the same temperature transition moment are connected to form a continuous pixel range. The x and y coordinates of each connected pixel within the continuous pixel range are recorded one by one, and the coordinates of all connected pixels are combined to form the coordinate range of the corresponding continuous pixel range. This coordinate range is recorded as the temperature transition position of the corresponding temperature transition. The temperature transition position does not include the coordinates of pixels not connected to the corresponding continuous pixel range. When the same temperature transition moment forms multiple non-adjacent continuous pixel ranges, the temperature transition position corresponding to each continuous pixel range is recorded separately.

[0042] S6. Match the alternation order of high and low humidity and the order of temperature transition, match the temperature transition position according to the registration result, and determine the bidirectional through segment; In a preferred embodiment of the present invention, the process of determining the bidirectional through-segment is as follows: The humidity alternation sequence for each test segment is read according to the order of the test segments, and the time of each humidity transition is recorded. For the same test segment, the previous humidity transition time is subtracted from the next humidity transition time to obtain the input time interval between adjacent humidity transition times. The input time intervals are arranged according to the order in which the humidity transitions occur, forming the input time interval arrangement for the corresponding test segment. Forward acquisition uses the input time interval arrangement recorded during the forward input process, while reverse acquisition uses the input time interval arrangement recorded after reversing the humidity alternation sequence.

[0043] The temperature transition sequence corresponding to the forward and reverse thermal imaging sequences is read separately. The inspection segment containing each temperature transition position is determined according to the inspection segment boundary on each inspection surface, and the temperature transition position is assigned to the corresponding air supply period according to the humidity transition time. For temperature transition positions in adjacent air supply periods within the same inspection segment, the number of pixels with the same or adjacent coordinates between each temperature transition position in the previous air supply period and each temperature transition position in the subsequent air supply period is counted. The subsequent temperature transition position with the most corresponding pixels is determined as the corresponding position; if the number of corresponding pixels is the same, the temperature transition position with the smallest center pixel coordinate spacing is selected. A subsequent temperature transition position that has already established a correspondence with a previous temperature transition position will not establish a correspondence with other temperature transition positions in the same air supply period. When there are no pixels with the same or adjacent coordinates, the subsequent association of the corresponding temperature transition position ends. The above process is repeated according to the order of the air supply periods to obtain the temperature transition time corresponding to each continuous pixel range within the same inspection segment. The difference between adjacent temperature transition times is calculated to obtain the transition time interval, and the transition time intervals are arranged according to the order in which the temperature transitions occur. The transition time intervals in the forward thermal imaging sequence are compared with the forward input time intervals of each test segment, and the transition time intervals in the reverse thermal imaging sequence are compared with the reverse input time intervals of each test segment. During the comparison, the number of time intervals included in each arrangement is first checked, and then the difference between the input time interval and the transition time interval is calculated one by one according to the arrangement position.

[0044] The preset time limit represents the maximum allowable difference between the input time interval and the transition time interval within the same arrangement position. When two arrangements contain the same number of time intervals, and the time interval difference corresponding to each arrangement position does not exceed the preset time limit, the corresponding temperature transition sequence is matched with the corresponding test segment. The test segments and corresponding temperature transition positions obtained from the matching of the forward thermal imaging sequence and the reverse thermal imaging sequence are recorded respectively.

[0045] The forward temperature transition position is located on the second detection surface. Following the rotation, scaling, and translation amounts used in thermal image registration, the forward temperature transition position is transformed to the coordinate range of the first detection surface. The reverse temperature transition position is located on the first detection surface. Following the reverse coordinate transformation order, the translation amount is first removed, then the scaling and rotation amounts are restored, transforming the reverse temperature transition position to the coordinate range of the second detection surface. The transformed temperature transition positions are then compared with the boundary coordinates of each inspection segment. When forward and reverse matching correspond to the same inspection segment, and the temperature transition positions in both directions are within the boundary range of the corresponding inspection segment, the corresponding inspection segment is recorded as a bidirectional through segment.

[0046] S7. Connect adjacent bidirectional through segments and generate through crack identification results.

[0047] In a preferred embodiment of the present invention, the process of generating the identification result of a penetrating crack is as follows: Read the forward and reverse temperature transition positions corresponding to each bidirectional through-segment. The forward temperature transition position is the position directly extracted from the second detection surface, and the reverse temperature transition position is the position directly extracted from the first detection surface. Map the forward temperature transition positions to the first detection surface according to the rotation, scaling, and translation amounts used in thermal image registration; map the reverse temperature transition positions to the second detection surface according to the reverse coordinate transformation order. After mapping, both the first and second detection surfaces contain one directly extracted temperature transition position and one temperature transition position mapped from the relative detection surface.

[0048] In the first detection surface, pixels within the mapped positive temperature transition positions are read sequentially, and the nearest pixel is found from the directly extracted negative temperature transition positions. In the second detection surface, pixels within the directly extracted positive temperature transition positions are read sequentially, and the nearest pixel is found from the mapped negative temperature transition positions. A preset position limit represents the maximum allowable coordinate distance between corresponding pixels in two detection directions. When the positional distance between two pixels does not exceed the preset position limit, the two pixels are recorded as paired pixels.

[0049] Pixels involved in pairing are extracted from both the first and second detection surfaces. Starting from each paired pixel, adjacent pixels located above, below, to the left, and to the right are examined. When adjacent pixels also belong to the paired pixels, they are added to the same continuous pixel range until no more adjacent pixels can be added. The continuous pixel range formed within the same bidirectional through segment is recorded as the segmented crack region of the corresponding detection surface, and the segmented crack region is associated with the arrangement position of the inspection segments.

[0050] The segmented crack regions in adjacent bidirectional through-segments are read sequentially according to the order of the test segments. When two segmented crack regions extend to a common segment boundary, and there are adjacent pixels on both sides of the common segment boundary, the two segmented crack regions are connected to form a continuous through-segment region. The same process is continued to check subsequent adjacent bidirectional through-segments until there are no more segmented crack regions that can be connected. For each continuous through-segment region, the adjacent pixels of each pixel in the region are checked. Pixels with at least one adjacent pixel located outside the continuous through-segment region are recorded as outer edge pixels, and the outer edge pixels are connected according to their adjacency to form the crack boundary. The pixel coordinates of the continuous through-segment region, the crack boundary coordinates, and the arrangement position of the corresponding test segments are recorded together as the through-crack identification result.

[0051] Understandably, before introducing isothermal gas with alternating high and low humidity, the surface of the area to be inspected should be checked for any sealing coatings, accumulated water, or coverings. Furthermore, it should be checked whether gas enters the area during the gas supply process. If the crack is completely sealed by moisture, precipitates, or filling material, preventing isothermal gas from entering, the absence of a temperature transition does not automatically indicate that the crack is not penetrated. Instead, the corresponding inspection segment is recorded as an incomplete active inspection segment. For inspection segments where isothermal gas input has been completed and a temperature transition has occurred, penetration assessment continues according to the forward and reverse alternation of high and low humidity, the temperature transition sequence, and the location of the temperature transition.

[0052] Understandably, the temperature of the input isothermal gas is adjusted to match the initial temperature of the corresponding detection surface. When isothermal gas with humidity higher than the detection environment enters the crack, water vapor is adsorbed on the crack wall and near the crack exit; when isothermal gas with humidity lower than the detection environment enters the crack, moisture on the crack wall and near the crack exit undergoes desorption and evaporation. The heat exchange generated during adsorption, desorption, and evaporation is superimposed on the temperature change caused by surface temperature excitation and reflected in the direction of temperature change of pixels near the crack exit relative to the reference thermal image sequence. Before detection, test specimens with material composition, thickness, and surface condition corresponding to the concrete component to be tested are selected, and through cracks that have been confirmed by observation from both sides are set in the test specimens. Pre-inspection conditions are arranged according to the heat source location, surface temperature excitation process, isothermal gas temperature, gas supply pressure, gas supply flow rate, and thermal image acquisition interval used in the formal inspection. Forward pre-inspection is performed using the humidity alternation sequence to be used in each inspection segment, and gas supply begins according to the initial humidity state recorded in each humidity alternation sequence. After completing the forward pre-inspection, the test specimen is processed according to the reset procedure, and then a reverse pre-inspection is performed from the relative detection surface in the reverse order of alternating high and low humidity. The humidity transition time is recorded for each of the forward and reverse pre-inspections, and the temperature change direction at the corresponding position on the other side of the crack is compared frame by frame in adjacent temperature difference images. The acquisition time when the temperature change direction first changes and remains unchanged until the next humidity transition time is recorded as the temperature transition time. If the temperature change direction does not change before the next humidity transition, the corresponding duration is extended and the pre-inspection is repeated. The difference between the input time interval and the corresponding transition time interval in both the forward and reverse pre-inspections is calculated. The maximum difference formed by each arrangement position in the repeated pre-inspections is added to a thermal image acquisition interval to obtain the preset time limit. The duration in which both the forward and reverse pre-inspections form a complete temperature transition sequence is selected as the gas supply duration in the corresponding high and low humidity alternation sequence.

[0053] This invention transforms crack identification from static temperature anomaly recognition to proactive response source verification. While differences in moisture content can create significant temperature variations, these variations passively change with the overall heating and cooling process and cannot reliably reproduce artificially set time-varying characteristics. Through-cracks, on the other hand, have continuous paths connecting opposite sides of a component; humidity changes input on one side can be transmitted to the other, creating a corresponding thermal response over time. By cross-checking the time sequence, location, and continuous range of the positive and negative responses, it is possible to distinguish between temperature anomalies caused by localized moisture content and those transmitted through the through-crack path, thus resolving the problem of moisture content differences being misjudged as through-cracks.

[0054] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the present invention should still fall within the scope of the present invention.

Claims

1. A method for identifying through-cracks in concrete, characterized in that, Includes the following steps: S1. Acquire and register reference thermal image sequences of the first and second detection surfaces on opposite sides of the concrete component; S2. Connect adjacent pixels in the reference thermal image sequence whose temperature change difference exceeds the preset limit. Based on the registration result, pair the temperature anomaly regions corresponding to the positions on both sides of the surface as the areas to be inspected, and divide them into inspection segments along their extension direction. S3. In the order of alternating high and low humidity of the test segments, isothermal gas is introduced from the first detection surface, and a forward thermal image sequence of the second detection surface is collected. S4. In the order of alternating high and low humidity, isothermal gas is introduced from the second detection surface, and the reverse thermal image sequence of the first detection surface is collected. S5. For each frame in the forward and reverse thermal image sequences, select thermal images with the same acquisition time in the corresponding reference thermal image sequence, perform pixel-by-pixel temperature difference, and extract the temperature transition order and temperature transition position. S6. Match the alternation order of high and low humidity and the order of temperature transition, match the temperature transition position according to the registration result, and determine the bidirectional through segment; S7. Connect adjacent bidirectional through segments and generate through crack identification results.

2. The method for identifying through-cracks in concrete according to claim 1, characterized in that, The process of acquiring and registering reference thermal image sequences is as follows: Multiple thermal markers are set at corresponding edge positions of the first and second detection surfaces, and the thermal markers are heated sequentially according to different heating durations. Simultaneously acquire thermal images of the first and second detection surfaces, and record the start time of temperature rise and the start time of temperature fall in the area where each thermal marker is located; The regions where the start times of temperature rise and temperature fall are paired are determined based on the center pixel coordinates of the paired thermal marker regions, and the rotation, scaling and translation amounts are determined based on the center pixel coordinates of the paired thermal marker regions. Register the thermal images of the first and second detection surfaces according to the rotation, scaling and translation amounts, stop heating the thermal marker, apply the same surface temperature excitation to the first and second detection surfaces, and arrange the continuously acquired thermal images into a reference thermal image sequence according to the acquisition time.

3. The method for identifying through-cracks in concrete according to claim 2, characterized in that, The process of dividing the test segments is as follows: Two thermal images with an average surface temperature difference not exceeding a preset pairing limit and opposite temperature change directions are selected from the reference thermal image sequences of the first and second detection surfaces, respectively. Calculate the temperature difference between each pixel and its adjacent pixels in each frame of thermal image, and subtract the temperature differences of pixels in the same position in two frames of thermal images to obtain the temperature change difference value; Connect adjacent pixels whose temperature change difference exceeds a preset limit to form a temperature anomaly region on the first detection surface and the second detection surface. According to the registration results, the temperature anomaly areas corresponding to each other at the boundary positions are paired. The paired temperature anomaly areas are recorded as the areas to be inspected. The areas to be inspected are continuously divided into inspection segments along the extension direction of the areas to be inspected, and the order of each inspection segment is recorded.

4. The method for identifying through-cracks in concrete according to claim 1, characterized in that, The process of acquiring a forward thermal image sequence is as follows: Each inspection segment in the first inspection surface is sealed separately, and an air supply port is set within the sealing range corresponding to each inspection segment; The periods of gas supply with humidity higher than the ambient humidity and the periods of gas supply with humidity lower than the ambient humidity are divided into multiple continuous periods. The continuous periods are combined according to the alternation of high and low humidity, and the time interval between adjacent humidity transition times in each combination is calculated. Select combinations with different time intervals, and determine the selected combinations as the humidity alternation order for each test segment, so as to keep the total duration of isothermal gas received by each test segment that is higher or lower than the humidity of the test environment the same. A surface temperature excitation corresponding to the reference thermal image sequence is applied to the second detection surface. At the start of the surface temperature excitation, isothermal gas is synchronously input into each inspection segment in alternating order of high and low humidity. The humidity transition time is recorded, and thermal images of the second detection surface are continuously acquired to form a forward thermal image sequence.

5. The method for identifying through-cracks in concrete according to claim 4, characterized in that, The process of acquiring reverse thermal imaging sequences is as follows: After completing the forward thermal imaging sequence acquisition, stop the surface temperature excitation, release the seal of each inspection segment in the first detection surface, and input isothermal gas with the same humidity as the detection environment into the first and second detection surfaces; The thermal images of the first and second detection surfaces are continuously acquired. When the pixel-by-pixel temperature difference between the acquired thermal image and the first frame of the corresponding reference thermal image sequence does not exceed the preset reset limit, the input of isothermal gas is stopped. Arrange the inspection segments in reverse chronological order according to the alternating order of high and low humidity used during the forward data acquisition process, and seal each inspection segment on the second detection surface. A surface temperature excitation corresponding to the reference thermal image sequence is applied to the first detection surface. At the start of the surface temperature excitation, isothermal gas is synchronously input into each inspection segment in reverse order of humidity high and low. The humidity transition time is recorded, and thermal images of the first detection surface are continuously acquired to form a reverse thermal image sequence.

6. The method for identifying through-cracks in concrete according to claim 1, characterized in that, The process of extracting the temperature transition sequence and temperature transition position is as follows: Calculate the elapsed time from the start of surface temperature excitation to the acquisition time of each frame of the forward and reverse thermal imaging sequences, respectively; Based on the elapsed time, each frame of thermal image in the forward thermal image sequence and the reverse thermal image sequence is subtracted pixel by pixel from the thermal image in the corresponding reference thermal image sequence with the same elapsed time to obtain a temperature difference image. Divide the temperature difference image according to the recorded humidity change time, compare the pixels in the same position in adjacent temperature difference images in turn, and record the temperature change direction of each pixel; Pixels whose temperature change direction changes after a humidity transition and continues until the next humidity transition are retained. The retained pixels are arranged according to the time of the change in temperature change direction to obtain the temperature transition order. Adjacent pixels with the same change time are connected, and the coordinate range of the connected pixels is recorded as the temperature transition position.

7. The method for identifying through-cracks in concrete according to claim 1, characterized in that, The process of determining the bidirectional through-segment is as follows: Calculate the input time interval between adjacent humidity transition times in the alternation sequence of high and low humidity in each test segment, and calculate the transition time interval between adjacent temperature transition times in the forward thermal image sequence and the reverse thermal image sequence. The turning time intervals corresponding to the forward thermal imaging sequence are compared sequentially with the input time intervals corresponding to each test segment, and the turning time intervals corresponding to the reverse thermal imaging sequence are compared sequentially with the input time intervals arranged in reverse order. The test segment records the difference between the input time interval and the transition time interval at the corresponding position, which does not exceed the preset time limit. According to the registration results, the forward temperature transition position is mapped to the first detection surface, and the reverse temperature transition position is mapped to the second detection surface. The inspection segment where the mapped temperature transition position is located is determined. If an inspection segment satisfies both forward and reverse matching, and its temperature transition positions are all located inside the segment, then the segment is determined to be a bidirectional through segment.

8. The method for identifying through-cracks in concrete according to claim 1, characterized in that, The process of generating the identification results of penetrating cracks is as follows: According to the registration results, the temperature transition positions in the forward thermal image sequence are mapped to the first detection surface, and the temperature transition positions in the reverse thermal image sequence are mapped to the second detection surface. In the first detection surface, the mapped forward temperature transition position and the temperature transition position extracted from the reverse thermal image sequence are paired; in the second detection surface, the temperature transition position extracted from the forward thermal image sequence and the mapped reverse temperature transition position are paired. If the distance between pixels that are paired on both the first and second detection surfaces does not exceed a preset limit, a segmented crack region is formed. Connect the segmented crack regions containing adjacent pixels at the segment boundaries according to the order of the test segments to form a continuous through region. Connect the outer edge pixels of the continuous through region to form the crack boundary, and record the continuous through region and crack boundary as the through crack identification result.

Images