Building crack width measurement method based on image analysis

Abstract

The invention relates to the technical field of crack image analysis, in particular to a building crack width measuring method based on image analysis, which comprises the following steps of: acquiring an autofluorescence attenuation image of a building surface under ultraviolet excitation, and generating a crack response coefficient matrix according to a material aging degree; driving micro-area pulse thermal imaging based on the crack response coefficient matrix, and extracting a heat conduction abnormal boundary as a crack physical edge; surface attachment interference is eliminated through an edge acoustic impedance matching model, and the real width value of the crack is calculated under a material intrinsic coordinate system. According to the method, the normal width of the crack is measured through uniform point selection in the main axis direction, the measured value is subjected to weighted fusion in combination with the local acoustic impedance matching weight, the obtained width result can better reflect the actual structural integrity and elastic continuity of the crack, geometric quantities are provided, and the local damage degree is reflected; the method is suitable for subsequent structure analysis and safety evaluation.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of crack image analysis, and particularly relates to a building crack width measurement method based on image analysis. BACKGROUND

[0002] Due to environmental stress changes, material aging, foundation settlement, structural load fluctuations and other factors, the outer facade structure of a building often has cracks of different degrees. These cracks not only affect the durability and safety of the building, but also cause damage to the anti-seepage and corrosion resistance performance. Therefore, accurately measuring the width of the cracks is of great significance for structural health assessment, development of post-maintenance strategies and quality responsibility determination.

[0003] Existing building crack width measurement methods mainly include manual visual inspection, caliper card, contact displacement sensor and visual image recognition. Among them, the manual method relies on personnel experience and has poor repeatability; the contact measurement needs to contact the crack edge, which is not suitable for high-altitude, outer wall or fragile surface areas; the visual image method has the advantages of non-contact and high efficiency, but it is easy to misjudge under the influence of light changes, surface pollution or attachments, and lacks a structural judgment mechanism to distinguish between "active cracks" and "historical cracks", which makes it difficult to support structural safety decisions. SUMMARY

[0004] The present application provides a building crack width measurement method based on image analysis, a new crack measurement method that combines perception and structural response characteristics, which can accurately extract the crack boundary under non-contact conditions, eliminate surface interference, and obtain real width values with structural significance in the material intrinsic direction, to improve the accuracy and engineering adaptability of building facade crack detection.

[0005] A building crack width measurement method based on image analysis, comprising the following steps:

[0006] S1, collecting self-fluorescence decay images of the building surface under ultraviolet excitation, and generating a crack response coefficient matrix according to the material aging degree;

[0007] S2, driving micro-pulse thermal imaging based on the crack response coefficient matrix, and extracting the thermal conduction abnormal boundary as the physical edge of the crack;

[0008] S3, eliminating surface attachment interference through an edge acoustic impedance matching model, and calculating the real width value of the crack in the material intrinsic coordinate system.

[0009] Optionally, the S1 specifically comprises:

[0010] S11, vertically irradiating the building surface with an ultraviolet LED array with a wavelength of 365±5nm, synchronously triggering an industrial camera equipped with a time-resolved imaging module, and continuously capturing 8-12 frames of fluorescence decay images;

[0011] S12, performing double exponential fitting on the decay image sequence to extract the fluorescence lifetime parameter τ(x, y) of each pixel point;

[0012] S13, mapping τ(x, y) to carbonation depth index D(x, y) according to a pre-stored material aging grade database, the material aging grade database being established through an accelerated aging experiment and including a τ-D correspondence relationship of concrete water-cement ratio and carbonation age;

[0013] S14, constructing a crack response coefficient matrix K(x, y) based on the carbonation depth index.

[0014] Optionally, the crack response coefficient matrix is represented as: wherein erf(·) is a Gaussian error function, D(x, y) is the carbonation depth at the pixel point (x, y), and δ is a material-dependent sensitivity factor.

[0015] Optionally, the S2 specifically includes:

[0016] S21, generating an energy density distribution map E(x, y) according to the crack response coefficient matrix K(x, y);

[0017] S22, applying pulse directional heating to the region where K(x, y) > 0.5 by using a laser micro-heating array, and synchronously starting an infrared thermal imager to record a surface temperature field T(x, y, t);

[0018] S23, establishing a two-dimensional heat conduction characteristic equation to calculate a theoretical temperature change rate;

[0019] S24, extracting the deviation degree of the actual temperature change rate from the theoretical value;

[0020] S25, performing a morphological closing operation on the abnormal boundary points to generate a continuous crack physical edge.

[0021] Optionally, in the S21, the energy density distribution map is generated by mapping according to the crack response coefficient matrix, combining the basic energy density and the square value of the response coefficient, and is used for controlling the region adaptation of the heating power, so that the region with higher response intensity obtains higher energy excitation.

[0022] Optionally, the two-dimensional heat conduction characteristic equation calculates the theoretical temperature change rate of each pixel and is represented as:

[0023]

[0024] wherein α is a material thermal conductivity, is a two-dimensional Laplace operator of the temperature field, β is a convection loss factor, T amb is an environmental background temperature, is a temperature change rate.

[0025] Optionally, the S3 specifically comprises:

[0026] S31, non-contact acoustic wave excitation and detection: linear scanning excitation is performed on both sides of the physical edge of the crack by using a pulsed laser, surface vibration response is captured by a laser Doppler vibrometer, and an acoustic wave propagation velocity field is generated;

[0027] S32, acoustic impedance inversion calculation: surface impedance distribution is constructed based on acoustic wave propagation characteristics;

[0028] S33, based on the surface impedance distribution, for the crack edge pixel points that exist and significantly deviate from the substrate impedance, according to the boundary outlier criterion, it is identified as an adherent interference point;

[0029] S34, according to the overall geometric trend of the crack edge and the local texture direction, a main direction vector of the material is constructed, and all crack edge coordinates are converted into an intrinsic coordinate system with the main direction vector as the axis;

[0030] S35, in the intrinsic coordinate system, a plurality of equally spaced measurement points are arranged along the crack direction, the apparent width value of the crack in the normal direction of each measurement point is extracted, a weighted factor based on the local acoustic impedance deviation is constructed, and the width values of the measurement points are weighted and fused, and the real average width result of the crack is output as the final real crack width value.

[0031] Optionally, in the S33, for the adherent interference point, the acoustic impedance gradient information of the adjacent effective area is used for interpolation repair to re-fit the real position of the crack edge.

[0032] Optionally, the surface acoustic impedance distribution is represented as:

[0033]

[0034] Wherein, ρ0 is the nominal density of the material, V(x, y) represents the acoustic wave propagation velocity of the corresponding point, α(x, y) represents the frequency-dependent acoustic wave attenuation coefficient, f is the main frequency of the excited acoustic wave, and Z(x, y) represents the surface acoustic impedance distribution.

[0035] Optionally, the final real crack width value W true is calculated as:

[0036] Wherein, d(k) represents the normal apparent width of the crack at the kth measurement point, Z c (k) represents the matching weight constructed based on the difference value of the local acoustic impedance and the substrate impedance, the closer to the substrate impedance, the greater the weight, and p represents the total number of measurement points.

[0037] The beneficial effects of the present application are as follows:

[0038] The present application realizes quantitative differentiation of the aging degree and crack activity of building surface materials by constructing a crack response coefficient matrix based on the ultraviolet excitation fluorescence decay lifetime, utilizing the corresponding relationship between the carbonization depth and the fluorescence lifetime, and on this basis, the energy distribution in the pulse thermography process is adaptively regulated in space according to the crack response coefficient, so that the heating excitation is focused on the high-response area, the invalid heating range is reduced, the energy utilization rate is improved, and the clarity of the thermal anomaly profile extraction is improved, thereby providing thermal support for subsequent accurate identification of the crack edge.

[0039] The present application introduces a two-dimensional heat conduction theory model and a dynamic temperature field deviation criterion based on thermography, realizes dynamic extraction of the physical boundary of the crack, and obtains the sound wave propagation speed through non-contact laser ultrasonic measurement, and combines the material acoustic parameters to perform impedance inversion, for the edge points disturbed by the surface attachments, an acoustic impedance anomaly discrimination mechanism and neighborhood interpolation correction are adopted to eliminate the influence of non-structural impurities on the boundary determination, effectively improving the continuity of the edge profile and the boundary determination accuracy, and solving the problem that the traditional thermography or optical imaging is easily disturbed by the cover.

[0040] The present application constructs an intrinsic coordinate system according to the main direction of the material fibers or the structural texture, eliminates the deviation caused by anisotropy to the width direction selection by rotating the coordinate system, uniformly selects points to measure the crack normal width in the principal axis direction, and combines the local acoustic impedance matching weight to weight and fuse the measurement values, so that the obtained width result can better reflect the actual structural integrity and elastic continuity of the crack, not only provides the geometric quantity, but also reflects the local damage degree, and is suitable for subsequent structure analysis and safety evaluation. BRIEF DESCRIPTION OF DRAWINGS

[0041] In order to more clearly illustrate the technical solutions in the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description only illustrate the present application, and for those skilled in the art, other drawings can also be obtained without creative labor on the basis of these drawings.

[0042] Fig. 1 The present application is a measurement method flowchart of the embodiment.

[0043] Fig. 2 The present application is a calculation of the crack true width value diagram of the embodiment. DETAILED DESCRIPTION

[0044] The application will be described in detail below with reference to the drawings and specific embodiments. It should be noted that in order to make the embodiments more detailed, the following embodiments are the best, preferred embodiments, and other alternative ways can also be implemented by those skilled in the art for some known technologies; and the drawings are only used to describe the embodiments in more detail, and are not intended to specifically limit the application.

[0045] It should be noted that in the specification, "one embodiment", "embodiment", "exemplary embodiment", "some embodiments" and the like indicate that the described embodiment can include a specific feature, structure or property, but not necessarily every embodiment includes the specific feature, structure or property. In addition, when a specific feature, structure or property is described in combination with an embodiment, it should be within the knowledge of those skilled in the related art to implement such a feature, structure or property in combination with other embodiments, whether or not it is explicitly described.

[0046] Generally, the terms can be understood at least in part from the context of their usage. For example, depending at least in part upon the context, the term "one or more" as used herein can be used to describe any feature, structure, or characteristic in the singular or can be used to describe combinations of features, structures, or characteristics, in the plural, depending at least in part upon the context. Further, the term "based on" can be understood as not necessarily requiring exclusively derived from that which is stated, but rather can allow for being based on additional factors not mentioned, depending at least in part upon the context.

[0047] As shown in Figs. 1-2 An image analysis-based building crack width measurement method, comprising the following steps:

[0048] S1, collecting self-fluorescence decay images of the building surface under ultraviolet excitation, and generating a crack response coefficient matrix according to the material aging degree;

[0049] S2, driving micro-pulse thermal imaging based on the crack response coefficient matrix, and extracting the thermal conduction abnormal boundary as the physical edge of the crack;

[0050] S3, eliminating the interference of surface attachments by an edge acoustic impedance matching model, and calculating the real width value of the crack in the material intrinsic coordinate system.

[0051] The self-fluorescence characteristics of silicate components in building materials (concrete / mortar) under ultraviolet excitation are utilized, and the decay time constant is strongly related to the carbonation degree of the material. By establishing a mapping relationship between the decay time constant and the crack development stage, fresh cracks and historical cracks are distinguished, and the problem of false measurement caused by the inability of traditional algorithms to judge the activity state of the crack is solved.

[0052] Micro-area pulse heating is applied to the identified active crack area, and the heat wave conduction process is captured by an infrared thermal imager. The real crack forms a sudden change in thermal resistance due to the internal cavity, and its heat conduction front speed is lower than that of the intact area. By solving the heat conduction equation to invert the physical boundary position, the visual pseudo-edge interference is overcome.

[0053] S1 specifically includes:

[0054] S11 uses a UV LED array with a wavelength of 365±5nm to vertically illuminate the building surface, synchronously triggering a high-speed camera equipped with a time-resolved imaging module to continuously capture 8 to 12 frames of fluorescence decay image sequences with a time resolution of 20ns.

[0055] S12, perform a double exponential fit on the decay image sequence to extract the fluorescence lifetime parameter τ(x,y) of each pixel, which is expressed as:

[0056]

[0057] Among them, τ1 and τ2 represent the fluorescence lifetime components of the bulk phase and crack phase of the material, respectively. A1 and A2 are the excitation amplitude coefficients of the corresponding lifetime components, respectively. A1 is the contribution of the "bulk phase of the wood material (uncracked area)" in the fluorescence intensity decay, and A2 is the contribution of the crack phase (air gap or penetration layer). The sum of the two, A1+A2, is the total initial excitation intensity.

[0058] Obtained by fitting a multi-frame fluorescence image sequence and estimated using the nonlinear least squares method:

[0059]

[0060] The fitting goal is to minimize the residual error between the time series I(t) of each pixel in the measured image and the model fitting.

[0061] S13, mapping the fluorescence lifetime parameter τ(x,y) to the carbonation depth index D(x,y) based on a pre-stored material aging grade database. The database is established based on accelerated aging experiments and covers the τ-D correspondence relationship for concrete with a water-cement ratio w / c = 0.35 to 0.55 and a carbonation age of 1 to 30 years;

[0062] S14, construct the crack response coefficient matrix K(x,y) based on the carbonization depth index, which is expressed as:

[0063]

[0064] Where erf(·) is the Gaussian error function, D(x,y) is the carbonation depth at the pixel point (x,y), and δ is the material-dependent sensitivity factor. For C30-C40 ordinary concrete, δ = 1.0-1.5 mm.

[0065] Wherein, the experience judgment standard is as follows:

[0066] When D(x, y) > 2.5mm, K(x, y) < 0.05, the region is judged as historical crack;

[0067] When D(x, y) < 0.8mm, K(x, y) > 0.85, the region is judged as active crack region.

[0068] Gaussian error function is used in the application to map carbonation depth D(x, y) to crack response coefficient K(x, y), which is a common S-shaped nonlinear mapping function, and the standard definition is as follows:

[0069]

[0070] Z is a real number, erf(z) represents the standard normal distribution error cumulative distribution from 0 to z, the function value range is (-1, 1), and it is S-shaped curve, which is commonly used for smoothing threshold judgment.

[0071] Table 1 numerical property comparison table

[0072]

[0073]

[0074] As can be seen from table 1, when z > 2.5, erf is almost close to 1, supporting the engineering criterion that "carbonation depth exceeding 2.5mm is historical crack".

[0075] Adopted: Corresponding to the process of gradually deepening carbonation depth, let Smoothly transition from 1 to 0, embodying "the fresher and shallower the carbonation crack response is higher", and "the more serious the carbonation area response is weaker", thereby supporting the differentiation mechanism of "active crack" and "historical crack".

[0076] Table 2 material aging grade database table

[0077]

[0078] The material aging grade database is established by accelerated carbonation aging experiment on concrete specimens with different water-cement ratios w / c ∈ [0.35, 0.55], and the fluorescence lifetime parameters τ of each specimen are collected at different ages (1 year-30 years), and the carbonation depth D is measured by acid washing method, the database realizes accurate mapping of any τ(x, y) to D(x, y) by using bilinear interpolation method, which is used to support the construction of crack response coefficient K(x, y).

[0079] S2 specifically includes:

[0080] S21, based on the crack response coefficient matrix K(x,y), generate the corresponding energy density distribution map E(x,y), which is expressed as: E(x,y) = E base ·[1+η·K(x,y) 2 ], where E base =20W / cm 2 represents the basic energy density, η is the active area gain coefficient, and η=3.0.

[0081] S22: Use a laser micro-area heating array to perform targeted pulse heating on the area where K(x,y)>0.5, with a pulse width of 50ms. Simultaneously, start an infrared thermal imager to continuously collect the surface temperature field T(x,y,t) sequence at a sampling frequency of 200Hz.

[0082] S23, establish a two-dimensional heat conduction characteristic equation and calculate the theoretical temperature change rate of each pixel Expressed as:

[0083]

[0084] Where α is the thermal conductivity of the material, is the two-dimensional Laplace operator of the temperature field (spatial heat diffusion term), β is the convection loss factor, T amb is the ambient background temperature.

[0085] S24, calculate the relative deviation Δ(x, y) between the actual measured temperature change rate and the theoretical value, which is defined as:

[0086]

[0087] When Δ(x, y)>30%, the pixel is marked as a thermal conduction abnormality boundary point.

[0088] S25: Perform a morphological closing operation on all thermal conductivity anomaly boundary points, using a circular structuring element with a radius of 0.5 mm for boundary closure. It is recommended to perform an opening operation followed by a closing operation to improve the coherence and structural integrity of the crack boundary, eliminate isolated anomalies caused by material porosity or local defects, and ultimately generate a continuous crack physical edge map.

[0089] Table 3 Heat conduction abnormality criteria

[0090]

[0091]

[0092] The method for implementing directional pulse heating with a laser micro-area heating array is as follows:

[0093] 1. System composition:

[0094] Laser array module: uses a semiconductor laser array with adjustable power, working in continuous or pulsed mode, with wavelengths in the near-infrared band (808nm, 980nm) or the mid-infrared band (1470nm), suitable for shallow thermal excitation of concrete and masonry building materials;

[0095] Spot size: The spot size of a single beam after focusing is approximately 0.5 mm to 1.5 mm, which is used to cover the single-pixel hot zone in the crack response matrix K(x,y).

[0096] Scanning and positioning unit: equipped with a two-dimensional electrically controlled deflection mirror (MEMS micro-vibration mirror) or an XY stage to align the heating point with the target coordinates and perform dynamic energy adjustment according to the pre-generated energy distribution map E(x,y).

[0097] Pulse controller: supports laser pulse width setting, setting the pulse width to 50ms, and the pulse repetition period and power amplitude are adjusted through the energy mapping function (i.e., power is controlled by K(x,y));

[0098] The temperature field T(x, y, t) is collected in real time and fed back to the control module to verify whether the temperature rise response reaches the set heating target value (such as a maximum temperature rise of 10 to 20°C).

[0099] 2. Implementation steps:

[0100] (1) Target area determination: According to the crack response coefficient matrix K(x,y), select the pixel points that satisfy K(x,y)>0.5 and generate the corresponding coordinate set

[0101] (2) Calculation of energy control parameters: For each target point, calculate its energy density E(x i ,y i ):

[0102] E(x i ,y i )=E base ·[1+η·K(x i ,y i ) 2 ];

[0103] Mapping energy density to laser power P i and pulse duration Δt i .

[0104] (3) Laser path and power scheduling: Control the laser array to scan and heat point by point or block by block in sequence. If the device supports parallel control (independent control of multiple laser beams), multiple heating areas can be excited in parallel, and the platform can be kept stationary or stably positioned during heating.

[0105] (4) Pulse heating execution: for each target point, a single pulse is emitted with a pulse width of 50 ms, and after each pulse execution, infrared temperature rise monitoring is immediately performed to record T(x, y, t), and if the desired temperature rise is not reached, a "supplementary heating pulse" strategy can be executed (up to 2-3 times).

[0106] S3 specifically comprises:

[0107] S31, non-contact acoustic wave excitation and detection: a laser with a wavelength of 1064 nm and a pulse energy of 2 mJ is used to perform high-precision line scanning excitation within a range of 0.5 mm on both sides of the aforementioned crack physical edge, and the laser spot diameter is 50 μm. A laser Doppler vibrometer is used synchronously to record the vibration response of the crack area surface at a sampling frequency of 10 MHz, and the acoustic wave propagation velocity field V(x, y) is reconstructed.

[0108] S32, acoustic impedance inversion calculation: according to the acoustic wave propagation velocity and the material density distribution, the surface acoustic impedance distribution Z(x, y) of each pixel on the building surface is calculated, which is expressed as:

[0109]

[0110] wherein ρ0 is the nominal density of the material, and the value of ordinary concrete is 2400 kg / m 3 , V(x, y) represents the acoustic wave propagation velocity of the corresponding point, α(x, y) represents the frequency-dependent acoustic wave attenuation coefficient, f is the main frequency of the excited acoustic wave, and the value is 1-2 MHz, the unit of the denominator term is kg / (m 3 ·s), and after being multiplied by the unit of α, which is Np / m, a unitless quantity is obtained, ensuring that the entire square term is dimensionless, so that the inside of the entire term is the same dimension as 1. In order to correct the influence of frequency-dependent acoustic wave propagation loss on the surface acoustic impedance, a normalized attenuation correction is introduced based on the standard impedance calculation.

[0111] The attenuation coefficient α(x, y) is determined by measuring the energy attenuation under pulse excitation: after laser acoustic excitation, the response intensity of the acoustic wave along the propagation path is recorded, and the local area attenuation index is extracted by fitting the spatial exponential decay trend of the response intensity. The extracted index term in the fitting expression is the local attenuation coefficient α(x, y) corresponding to the main frequency f.

[0112] The acoustic wave propagation velocity V(x, y) is calculated by continuously exciting laser points along a scanning line with a known spacing in the crack edge area, synchronously recording the surface vibration response time generated by the acoustic wave propagating into adjacent positions at each excitation point, and calculating the propagation velocity V(x, y) according to the time difference of the acoustic wave propagation and the spacing between the excitation points.

[0113] S33, interference suppression mechanism for each crack edge point (x i ,yi ), if the following condition is met:

[0114] |Z(x i ,y i )-Z0|>0.25Z0, it is determined that the point is a heterogeneous adherent or surface covering layer interference, and needs to be corrected. Wherein Z0 is the nominal acoustic impedance of the substrate, for the points meeting the condition, the acoustic impedance gradient of the effective edge points around is used for interpolation repair to restore the true material boundary shape.

[0115] S34, intrinsic coordinate system mapping: in order to eliminate the error influence of material anisotropy on crack width measurement, based on the geometric trend of crack edge and the analysis result of surface texture direction, a local principal direction angle θ is constructed, the edge point coordinates are rotated and mapped from the image coordinate system to the material intrinsic coordinate system, and the conversion relationship is as follows:

[0116]

[0117] Wherein,.x ′ ,y ′ / is the edge point coordinate in the image coordinate system, and (X, Y) is the edge point coordinate in the material intrinsic coordinate system.

[0118] S35, true crack width calculation: 30 measurement points are arranged along the crack principal axis direction at equal intervals in the intrinsic coordinate system, and the apparent width value d(k) in the normal direction is extracted at each point, and the true crack width W c (k) is calculated in combination with the acoustic impedance matching weight Z true , and the expression is as follows:

[0119] Wherein d(k) represents the normal apparent width of the crack at the kth measurement point, indicates the matching weight constructed based on the difference value of local acoustic impedance and substrate impedance, and the closer to the substrate impedance, the greater the weight.

[0120] The present application covers any substitution, modification, equivalent method and scheme made on the essence and scope of the present application. In order to make the public have a thorough understanding of the present application, specific details are described in the following preferred embodiments of the present application, and the present application can also be completely understood without the description of these details for those skilled in the art. In addition, in order to avoid unnecessary confusion to the essence of the present application, well-known methods, processes, procedures, elements and circuits are not described in detail.

[0121] The above is only the preferred embodiment of the present application, and it should be pointed out that for ordinary skilled in the art, without departing from the principle of the present application, a number of improvements and refinements can be made, and these improvements and refinements should be regarded as the protection scope of the present application.

Claims

1. A method for measuring building crack width based on image analysis, characterized in that: The following steps are involved: S1. Collect the autofluorescence attenuation image of the building surface under ultraviolet excitation and generate the crack response coefficient matrix according to the degree of material aging; S2, based on the crack response coefficient matrix, driving micro-area pulse thermal imaging, extracting the thermal conductivity anomaly boundary as the physical edge of the crack; S3. Eliminate the interference of surface attachments through the edge acoustic impedance matching model and calculate the true crack width value in the material intrinsic coordinate system.

2. The method for measuring building crack width based on image analysis according to claim 1, characterized in that: Said S1 specifically includes: S11, uses a UV LED array to vertically illuminate the building surface, synchronously triggering an industrial camera equipped with a time-resolved imaging module to continuously capture multiple frames of fluorescence decay images; S12, performing a double exponential fitting process on the decay image sequence to extract the fluorescence lifetime parameter τ(x,y) of each pixel; S13, mapping τ(x, y) to a carbonation depth index D(x, y) according to a pre-stored material aging grade database, wherein the material aging grade database is established through accelerated aging experiments and includes a τ-D correspondence between concrete water-cement ratio and carbonation age; S14, constructing a crack response coefficient matrix K(x,y) based on the carbonization depth index.

3. The method for measuring building crack width based on image analysis according to claim 2, wherein: The crack response coefficient matrix is ​​expressed as: where erf(·) is the Gaussian error function, D(x,y) is the carbonization depth at the pixel point (x,y), and δ is the material-dependent sensitivity factor.

4. The method for measuring building crack width based on image analysis according to claim 1, characterized in that: The S2 specifically includes: S21, generating an energy density distribution map E(x,y) according to the crack response coefficient matrix K(x,y); S22, using a laser micro-area heating array to apply pulsed directional heating to the area where K(x,y)>0.5, and synchronously starting an infrared thermal imager to record the surface temperature field T(x,y,t); S23, establish a two-dimensional heat conduction characteristic equation to calculate the theoretical temperature change rate; S24, extracting the deviation between the actual temperature change rate and the theoretical value; S25, performing a morphological closing operation on the abnormal boundary points to generate a continuous crack physical edge.

5. The method for measuring building crack width based on image analysis according to claim 4, characterized in that: In S21, based on the crack response coefficient matrix, combined with the basic energy density and the square value of the response coefficient, a spatially distributed energy density distribution diagram is generated through mapping to control the regional adaptation of the heating power so that regions with higher response intensity receive higher energy excitation.

6. The method for measuring building crack width based on image analysis according to claim 5, characterized in that: The two-dimensional heat conduction characteristic equation is used to calculate the theoretical temperature change rate of each pixel as follows: Where α is the thermal conductivity of the material, is the two-dimensional Laplace operator of the temperature field, β is the convection loss factor, T amb is the ambient background temperature, is the rate of temperature change.

7. The method for measuring building crack width based on image analysis according to claim 1, characterized in that: The S3 specifically includes: S31, non-contact acoustic wave excitation and detection: a pulsed laser is used to perform line scanning excitation on both sides of the physical edge of the crack, and a laser Doppler vibrometer is used to capture the surface vibration response to generate an acoustic wave propagation velocity field; S32, Acoustic Impedance Inversion Calculation: Construct surface impedance distribution based on acoustic wave propagation characteristics; S33, based on the surface impedance distribution, for crack edge pixels that deviate significantly from the substrate impedance, identify them as attachment interference points according to the boundary outlier criterion; S34, constructing a principal direction vector of the material according to the overall geometric orientation of the crack edge and the local texture direction, and transforming all crack edge coordinates into an intrinsic coordinate system with the principal direction vector as the axis; S35: In the intrinsic coordinate system, multiple equally spaced measurement points are set along the crack strike direction. The apparent width of the crack is extracted in the normal direction of each measurement point. At the same time, a weighting factor based on the local acoustic impedance deviation is constructed. The width values ​​of each measurement point are weighted and fused to output the actual average width of the crack as the final actual width of the crack.

8. The method for measuring building crack width based on image analysis according to claim 7, characterized in that: In the above S33, for the attachment interference point, the acoustic impedance gradient information of the adjacent effective area is used to perform interpolation repair to re-fit the actual position of the crack edge.

9. The method for measuring building crack width based on image analysis according to claim 7, characterized in that: The surface acoustic impedance distribution is expressed as: Where ρ0 is the nominal density of the material, V(x,y) represents the acoustic wave propagation velocity at the corresponding point, α(x,y) represents the frequency-dependent acoustic wave attenuation coefficient, f is the dominant frequency of the excited acoustic wave, and Z(x,y) represents the surface acoustic impedance distribution.

10. The method for measuring building crack width based on image analysis according to claim 9, characterized in that: The final crack true width value W true Calculated as: Where d(k) represents the normal apparent width of the crack at the kth measurement point, Z c (k) represents the matching weight based on the difference between the local acoustic impedance and the substrate impedance. The closer to the substrate impedance, the greater the weight. p represents the total number of measurement points.

Images