A method for measuring structural displacement under high temperature conditions based on monocular vision

Abstract

The invention discloses a method for measuring structural displacement under high-temperature working conditions based on monocular vision, specifically relating to the field of image analysis from bitmap to non-bitmap based on monocular vision. The method includes extracting a local grayscale feature set and statistically analyzing grayscale distribution parameters based on a sequence of continuous frame grayscale images collected by a monocular vision system, completing the determination and screening of preliminary connected regions, establishing a regional degradation probability map, and constructing a standard Markov random field graph structure. Based on the standard Markov random field graph structure, the method extracts a grayscale gradient feature set and establishes a local gradient perturbation vector field to generate a perturbation potential energy map. By constructing a variant Markov random field model based on local grayscale features and mutual information entropy regulation, combined with dynamic degradation characteristic extraction and benchmark point time series trajectory analysis, the method realizes structural information recovery and displacement measurement of locally degraded regions under high-temperature working conditions, overcoming the problems of image connectivity loss and structural tracking failure caused by optical window degradation.

Description

Technical Field

[0001] The present invention relates to the technical field of image analysis from bitmap to non-bitmap based on monocular vision, and more specifically, to a method for measuring structural displacement under high-temperature working conditions based on monocular vision. Background Art

[0002] Monocular vision systems are widely used in steel structure displacement monitoring under high-temperature conditions. They rely on stable image acquisition to ensure the continuity and accuracy of displacement measurements. Typically, these systems use high-temperature-resistant optical windows, such as quartz glass, to protect the camera lens to withstand the intense radiant heat flux in the environment. However, under long-term high-temperature exposure conditions, the optical window surface inevitably undergoes localized carbonization, fogging, and charring, resulting in decreased optical transmittance and increased transmission non-uniformity, forming randomly distributed localized degraded areas. This degradation effect dynamically expands over time, causing non-uniform transmittance distortion during image acquisition.

[0003] As a result, the bitmap (i.e., binary contour feature) information corresponding to the degraded area in the image is broken or lost, and the original edge contour is partially obscured. Traditional image segmentation processing methods based on global threshold or edge detection cannot effectively extract the complete boundary. At the same time, the non-bitmap grayscale field and local texture features in the degraded area are easily obscured by carbonization, resulting in grayscale drift and texture blur, which seriously interferes with the image structure information extracted based on grayscale or morphological operators.

[0004] Especially under monocular vision conditions, the lack of multi-view redundant information compensation and local optical degradation directly lead to a decrease in structural integrity in the image processing stage, making it impossible to support the reliable calculation of subsequent displacements.

[0005] Furthermore, imaging anomalies caused by optical window degradation are not static but evolve dynamically with environmental thermal disturbances, sediment formation rates, and window aging, resulting in a "degradation drift" phenomenon with non-uniform spatial and temporal expansion. This phenomenon not only destroys the regional connectivity of static images but also makes it difficult for fixed templates or static segmentation strategies to adapt to local changes, ultimately causing the failure of the reference datum in the displacement measurement process.

[0006] In summary, under high-temperature conditions, existing monocular vision systems suffer from the problem of local image information loss and non-image texture blurring caused by optical window surface degradation, which seriously limits the accuracy and reliability of steel structure displacement measurement based on image processing, becoming the core problem restricting the application performance of monocular vision in high-temperature environments. Summary of the Invention

[0007] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a method for measuring structural displacement under high-temperature conditions based on monocular vision. By constructing a variational Markov random field model based on local grayscale features and mutual information entropy adjustment, combined with dynamic degradation characteristic extraction and benchmark point time-series trajectory analysis, structural information recovery and displacement measurement of locally degraded areas under high-temperature conditions are achieved, overcoming the problems of image connectivity loss and structural tracking failure caused by optical window degradation.

[0008] To achieve the above-mentioned object, the present invention provides the following technical solutions: a method for measuring structural displacement under high-temperature conditions based on monocular vision, comprising a monocular vision system for collecting image data;

[0009] S1. Based on the continuous frame grayscale image sequence collected by the monocular vision system, extract the local grayscale feature set and calculate the grayscale distribution parameters, complete the preliminary determination and screening of connected regions, establish the regional degradation probability map, and build the standard Markov random field graph structure;

[0010] S2. Based on the standard Markov random field graph structure, the grayscale gradient feature set is extracted and the local gradient perturbation vector field is established to generate the perturbation potential energy map. The mutual information entropy adjustment matrix is ​​superimposed to construct the coupling potential energy structure. The temporal coherence constraint matrix is ​​constructed by combining the grayscale temporal change sequence. The neighborhood transition probability is jointly updated to form a variant Markov random field model.

[0011] S3. Apply the variant Markov random field model to the current frame grayscale image, solve the local region state configuration to generate a segmentation map, perform morphological closing operations to generate a morphological structure recovery map, extract connected regions and filter the valid boundary set, calculate the coordinates of the boundary center points to form the initial reference point set for displacement measurement, combine the continuous frame grayscale image sequence to track the reference point set features frame by frame, and establish the reference point time series trajectory matrix;

[0012] S4. Calculate the displacement vector sequence based on the reference point time-series trajectory matrix, smooth the process to generate a smooth displacement vector field, combine the mutual information entropy change trend of the variant Markov random field model to determine the local scale drift, and correct the displacement vector field to output the final structural displacement measurement result.

[0013] In a preferred embodiment, in S1, a sequence of continuous frame grayscale images is collected by a monocular vision system, and a fixed area sliding window partition is performed to form a local grayscale region group of continuous frames. The local grayscale region group of continuous frames is divided according to the neighborhood relationship of pixel points, and the grayscale distribution vector of each local region is calculated to form a local grayscale feature set;

[0014] Based on the local grayscale feature set, the mean and second-order moment of the grayscale distribution within the statistical region are calculated to generate a local grayscale statistical matrix. The local grayscale statistical matrix is ​​used to perform pixel-level connectivity judgment, extract connected regions, and form a preliminary connected region set. The coefficient of variation of the neighborhood grayscale of the preliminary connected region set is calculated, and regions with a coefficient of variation greater than a preset coefficient of variation threshold are screened to construct a regional degradation probability map.

[0015] Based on the regional degradation probability graph, the neighborhood consistency rule is set to form the initial neighborhood dependency structure and generate the standard Markov random field graph structure.

[0016] In a preferred embodiment, in S2, based on the standard Markov random field graph structure, the grayscale first-order gradient vector of the local area is extracted to form a grayscale gradient feature set, the grayscale gradient feature set is jointly modeled with the neighborhood grayscale difference matrix to generate a local gradient perturbation vector field, based on the local gradient perturbation vector field, a local perturbation energy function is established, the neighborhood potential energy function is updated, and a perturbation potential energy map is formed;

[0017] The regional mutual information entropy distribution is introduced, the interaction relationship between the local regional mutual information entropy and the neighborhood is counted, and the mutual information entropy adjustment matrix is ​​constructed. The perturbation potential energy map and the mutual information entropy adjustment matrix are superimposed to form a coupled potential energy structure, and the coupled potential energy neighborhood update rule is defined.

[0018] The grayscale distribution vector groups extracted from the local grayscale region groups of continuous frames are arranged in chronological order to generate a local region grayscale temporal change sequence, and a regional temporal coherence constraint matrix is ​​constructed based on the change sequence;

[0019] Taking the coupled potential energy structure of the neighborhood and the temporal coherence constraint matrix as joint constraints, the neighborhood transition probability is reconstructed to generate a mutational Markov random field model.

[0020] In a preferred embodiment, in S3, the variant Markov random field model is applied to the grayscale image of the current frame, the local region state configuration is solved based on the neighborhood state transition probability matrix, and a local region segmentation map is generated. On the local region segmentation map, a morphological closing operation is performed based on the grayscale statistical matrix to generate a morphological structure recovery map;

[0021] The connected areas in the morphological structure restoration graph are extracted, and the effective boundary set is screened based on the boundary connectivity to form a boundary restoration vector group. The coordinates of the boundary center point are calculated through the boundary restoration vector group to form the initial reference point set for displacement measurement;

[0022] Based on the initial reference point set of displacement measurement and combined with the continuous frame grayscale image sequence, the local grayscale features and connectivity relationship of the reference point set are tracked frame by frame to establish the reference point temporal trajectory matrix.

[0023] In a preferred embodiment, in S4, based on the reference point time series trajectory matrix, the reference point displacement vector group is calculated frame by frame to form a displacement vector sequence, local time series smoothing is performed on the displacement vector sequence, local non-physical jump trajectories are eliminated, and a smoothed displacement vector field is output. The smoothed displacement vector field is used to determine the local scale drift based on the mutual information entropy change trend of the variant Markov random field model in the local area, and a local scale drift matrix is ​​constructed;

[0024] Based on the local scale drift matrix, the displacement vector field is reversely corrected to form a corrected displacement measurement result set, which is then weighted averaged by the regional confidence weight matrix to form the final structural displacement measurement output result.

[0025] In a preferred embodiment, in S1, it is defined Indicates the Frame The local grayscale statistical feature vector extracted from the position is constructed:

[0026] ;

[0027] in:

[0028] ;

[0029] ;

[0030] ;

[0031] in is the local grayscale mean; is the local grayscale standard deviation; is the local grayscale skewness; Indicates the In the local grayscale area group of continuous frames, Extracted from the center Pixel gray value; is the number of pixels in the sliding window; is the frame number of the continuous frame grayscale image sequence; is the pixel coordinate of the center of the sliding window in the image;

[0032] definition Represents the set of connected regions after screening, and extracts connected regions and screens the coefficient of variation;

[0033] ;

[0034] in Indicates the first connected subsets; pixels express Pixel coordinates within ; Represents pixels Neighborhood pixel coordinates of ; Represents pixels The neighborhood set of ; is the Euclidean distance of the local statistical eigenvector; is the skewness difference adjustment factor; The energy threshold preset for connectivity determination; For the Frame Pixels Gray value of express The average grayscale; is the preset grayscale variation energy threshold; For the Frames in Pixels The local grayscale skewness corresponding to the position; For the Frames in Pixels The local grayscale skewness corresponding to the position; Indicates the Frames in Pixels Local grayscale statistical feature vector extracted from the position; Indicates the Frames in Pixels Local grayscale statistical feature vector extracted from the position;

[0035] definition is the standard Markov random field graph structure, which is composed of and composition;

[0036] ;

[0037] in:

[0038] ;

[0039] ;

[0040] in is the pixel coordinate The degradation probability value at ; for Grayscale variation energy; In the above formula, it is used to control the decay rate of the degradation probability Gaussian map; is the neighborhood degradation probability difference adjustment factor; is the pixel coordinate Neighborhood consistency potential value at ; For the Frame in pixel coordinates The gradient vector of For the Frames in Pixels The gradient vector of is the neighborhood gradient weight adjustment factor; is the pixel coordinate The neighborhood pixel set of is the natural exponential function; is the two-norm; Pixels The degradation probability value at .

[0041] In a preferred embodiment, in S2, it is defined For the Frame Pixels Grayscale first-order gradient vector:

[0042] ;

[0043] definition is the local gradient perturbation vector field:

[0044] ;

[0045] ;

[0046] ;

[0047] in For the Frame in pixel coordinates The gradient vector of is the pixel coordinate Laplace enhanced gradient vector at ; Pixels Laplace enhanced gradient vector at ; is the second-order derivative of the pixel; is the pixel coordinate The neighborhood pixel set of is the local perturbation energy function; is the exponential decay adjustment factor of the local perturbation energy function; is the nonlinear amplification order of the disturbance energy;

[0048] definition For the Frame pixel coordinates The mutual information entropy value of is:

[0049] ;

[0050] definition Indicated in pixel coordinates The neighborhood mutual information entropy interaction matrix centered on is used to quantify the mutual information entropy difference between local pixels and neighboring pixels and their degree of coordinated change;

[0051] ;

[0052] definition is the coupled potential energy field:

[0053] ;

[0054] in is the pixel coordinate Gray value is The probability distribution of is a grayscale value set; For the Frame Pixels The mutual information entropy value of is the mutual information second-order synergy term adjustment factor; Adjust the weights for mutual information entropy;

[0055] definition Indicates the Frame and adjacent frame pixel coordinates A continuous grayscale vector sequence:

[0056] ;

[0057] definition is the pixel coordinate Grayscale temporal coherence measure of:

[0058] ;

[0059] definition Represents pixels The migration probability is:

[0060] ;

[0061] in For the Frame pixel coordinates The local grayscale vector of ; For the Frame pixel coordinates The local grayscale vector of ; For the Frame pixel coordinates The local grayscale vector of ; Number the timeframe; Indicates the Frame in pixel coordinates Extracted local grayscale vector; is the timing variation constraint adjustment factor; Constrain the nonlinear order for temporal coherence; Represents the neighborhood pixel coordinates traversed in the normalized denominator; for Grayscale temporal coherence measure of ; For the frame The coupling potential energy value of the position.

[0062] In a preferred embodiment, in S3, define For the Frame pixel coordinates The local state label of ;

[0063] ;

[0064] ;

[0065] ;

[0066] in is a state label set; is the candidate state label; Pixels The current status label of is the state consistency constraint adjustment factor; is the Kronecker function, which is the indicator function. and If they are equal, the value is 1, otherwise it is 0; is the grayscale feature similarity constraint adjustment factor; For the Pixel coordinates in the frame Grayscale statistical feature vector of ; For the Pixels in frame Grayscale statistical feature vector of ; Represents the morphological closing operation based on the grayscale statistical matrix; It is the image structure restored after morphological closing operation; Indicates the first connected subsets; is the preset connected area threshold; is a set of connected regions; express The pixel area;

[0067] definition Recover the set of vectors for the boundary:

[0068] ;

[0069] Calculate the boundary center point set and generate the initial reference point set :

[0070] ;

[0071] definition As the reference point Trajectory matrix on the time axis:

[0072] ;

[0073] in for The boundary set of is the boundary pixel coordinate; is the number of boundary pixels; Represents the neighborhood matching function based on grayscale feature vector; is the number of consecutive frames; As the reference point In the The spatial coordinate position in the frame image; As the reference point In the The grayscale statistical feature vector corresponding to the new coordinate position determined by the neighborhood matching of the frame; As the reference point In the The grayscale statistical feature vector corresponding to the new coordinate position determined by neighborhood matching.

[0074] In a preferred embodiment, in S4, define For the Frame reference point The displacement vector:

[0075] ;

[0076] definition is the smoothed displacement vector:

[0077] ;

[0078] in is the width of the smoothing time window; Number the timeframe; As the reference point In the The spatial coordinate position in the frame image;

[0079] definition As the reference point Local drift measure of :

[0080] ;

[0081] definition is the corrected displacement vector:

[0082] ;

[0083] in is the sign of the discrete rate of change on adjacent frame numbers; For the Frame at reference point current The mutual information entropy value of is the drift correction adjustment factor;

[0084] definition As the reference point The confidence weight of :

[0085] ;

[0086] Definition The final structural displacement measurement output of the frame is :

[0087] ;

[0088] in is the weight matrix adjustment factor.

[0089] Technical effects and advantages of the present invention:

[0090] 1. By introducing an image analysis mechanism from bitmap to non-bitmap, combined with a variational Markov random field model and a grayscale-mutual information entropy joint modeling method, this method can extract the structural features and grayscale information of degraded areas under conditions of local optical degradation caused by carbonization, fogging, and coking of the optical window, achieve dynamic recovery of local structures, and improve the continuity and reliability of monocular vision displacement measurement under high-temperature conditions. This provides an effective solution to the problems of local structural fracture and benchmark failure caused by optical degradation in existing monocular vision systems.

[0091] 2. By combining the statistical characteristics of local grayscale distribution vectors and second-order moments with connectivity determination and degradation probability map generation, this method effectively identifies regions with local grayscale variation characteristics, avoids misjudgment and information loss caused by global threshold segmentation methods under high-temperature degradation conditions, and enhances the image preprocessing stage's adaptability to local degradation disturbances.

[0092] 3. By constructing a gradient perturbation vector field and a mutual information entropy adjustment matrix, coupling them to form a potential energy structure. Combined with local grayscale temporal continuity constraints, this method captures and models the dynamic trends of image degradation regions, overcomes the spatial-temporal non-uniform expansion problem caused by degradation drift, and improves the stability and continuity of region segmentation and benchmark extraction.

[0093] 4. Based on a combined processing method of morphological closing operations and grayscale statistical features, this method repairs boundary breaks and detail occlusion in degraded image environments, ensuring complete extraction of connected regions. Furthermore, a stable reference point set is calculated using a boundary recovery vector group, providing highly reliable reference information for displacement measurement and reducing displacement calculation errors caused by degradation.

[0094] 5. By constructing the reference point timing trajectory matrix and the local drift matrix, combined with the drift detection and dynamic correction mechanism of the mutual information entropy change rate, non-physical jumps caused by local optical degradation are eliminated, forming a smooth and physically consistent displacement vector field, thereby improving the stability of displacement measurement;

[0095] 6. By introducing a regional confidence weight matrix and adaptively weighting the displacement measurement results based on local drift stability, the impact of local abnormal drift reference points on the overall measurement accuracy is avoided, and the structural deformation trend in high-temperature and complex environments is reflected, thereby improving the reliable measurement capability of the monocular vision system under harsh working conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0096] Figure 1 The figure is a flow chart of the method steps of the present invention. DETAILED DESCRIPTION

[0097] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0098] Refer to the instruction manual Figure 1 , a method for measuring structural displacement under high temperature conditions based on monocular vision according to an embodiment of the present invention includes a monocular vision system for collecting image data;

[0099] S1. Based on the continuous frame grayscale image sequence collected by the monocular vision system, extract the local grayscale feature set and calculate the grayscale distribution parameters, complete the preliminary determination and screening of connected regions, establish the regional degradation probability map, and build the standard Markov random field graph structure;

[0100] S2. Based on the standard Markov random field graph structure, the grayscale gradient feature set is extracted and the local gradient perturbation vector field is established to generate the perturbation potential energy map. The mutual information entropy adjustment matrix is ​​superimposed to construct the coupling potential energy structure. The temporal coherence constraint matrix is ​​constructed by combining the grayscale temporal change sequence. The neighborhood transition probability is jointly updated to form a variant Markov random field model.

[0101] S3. Apply the variant Markov random field model to the current frame grayscale image, solve the local region state configuration to generate a segmentation map, perform morphological closing operations to generate a morphological structure recovery map, extract connected regions and filter the valid boundary set, calculate the coordinates of the boundary center points to form the initial reference point set for displacement measurement, combine the continuous frame grayscale image sequence to track the reference point set features frame by frame, and establish the reference point time series trajectory matrix;

[0102] S4. Calculate the displacement vector sequence based on the reference point time-series trajectory matrix, smooth the generated smooth displacement vector field, determine the local scale drift based on the mutual information entropy change trend of the variant Markov random field model, and correct the displacement vector field to output the final structural displacement measurement result;

[0103] The monocular vision system is an image acquisition device based on a single imaging sensor, which is used to obtain two-dimensional image data from a single perspective.

[0104] In S1, a sequence of continuous frame grayscale images is collected through a monocular vision system, and a fixed area sliding window is performed to form a continuous frame local grayscale region group. The continuous frame local grayscale region group is divided according to the pixel neighborhood relationship, and the grayscale distribution vector of each local region is calculated to form a local grayscale feature set;

[0105] Based on the local grayscale feature set, the mean and second-order moment of the grayscale distribution within the statistical region are calculated to generate a local grayscale statistical matrix. The local grayscale statistical matrix is ​​used to perform pixel-level connectivity judgment, extract connected regions, and form a preliminary connected region set. The coefficient of variation of the neighborhood grayscale of the preliminary connected region set is calculated, and regions with a coefficient of variation greater than a preset coefficient of variation threshold are screened to construct a regional degradation probability map.

[0106] Based on the regional degradation probability graph, the neighborhood consistency rule is set to form the initial neighborhood dependency structure and generate the standard Markov random field graph structure.

[0107] In S2, based on the standard Markov random field graph structure, the first-order grayscale gradient vector of the local area is extracted to form a grayscale gradient feature set. The grayscale gradient feature set is jointly modeled with the neighborhood grayscale difference matrix to generate a local gradient perturbation vector field. Based on the local gradient perturbation vector field, a local perturbation energy function is established, and the neighborhood potential energy function is updated to form a perturbation potential energy map.

[0108] The regional mutual information entropy distribution is introduced, the interaction relationship between the local regional mutual information entropy and the neighborhood is counted, and the mutual information entropy adjustment matrix is ​​constructed. The perturbation potential energy map and the mutual information entropy adjustment matrix are superimposed to form a coupled potential energy structure, and the coupled potential energy neighborhood update rule is defined.

[0109] The grayscale distribution vector groups extracted from the local grayscale region groups of continuous frames are arranged in chronological order to generate a local region grayscale temporal change sequence, and a regional temporal coherence constraint matrix is ​​constructed based on the change sequence;

[0110] Taking the coupled potential energy structure of the neighborhood and the temporal coherence constraint matrix as joint constraints, the neighborhood transition probability is reconstructed to generate a mutational Markov random field model.

[0111] In S3, the variant Markov random field model is applied to the grayscale image of the current frame, the local region state configuration is solved based on the neighborhood state transition probability matrix, and a local region segmentation map is generated. On the local region segmentation map, a morphological closing operation is performed based on the grayscale statistical matrix to generate a morphological structure recovery map;

[0112] The connected areas in the morphological structure restoration graph are extracted, and the effective boundary set is screened based on the boundary connectivity to form a boundary restoration vector group. The coordinates of the boundary center point are calculated through the boundary restoration vector group to form the initial reference point set for displacement measurement;

[0113] Based on the initial reference point set of displacement measurement and combined with the continuous frame grayscale image sequence, the local grayscale features and connectivity relationship of the reference point set are tracked frame by frame to establish the reference point temporal trajectory matrix.

[0114] In S4, based on the reference point temporal trajectory matrix, the reference point displacement vector group is calculated frame by frame to form a displacement vector sequence. Local temporal smoothing is performed on the displacement vector sequence to eliminate local non-physical jump trajectories and output a smoothed displacement vector field. The smoothed displacement vector field is used to determine the local scale drift based on the mutual information entropy change trend of the local region's variant Markov random field model, and a local scale drift matrix is ​​constructed.

[0115] Based on the local scale drift matrix, the displacement vector field is reversely corrected to form a corrected displacement measurement result set, which is then weighted averaged by the regional confidence weight matrix to form the final structural displacement measurement output result.

[0116] It should be noted that in the formula structure involved in this solution, dimensionless terms can serve as proportionality or structural adjustment factors. When combined with quantities with units, they only play a numerical scaling role and do not introduce new physical dimensions. Therefore, they will not change or confuse the overall unit system of expression. This combination of "dimensionless terms and units" can be understood as a composite structural expression commonly used in mathematical and physical modeling, conforming to the principle of dimensional consistency and having a clear physical interpretation basis.

[0117] Secondly, in the formula structure of this scheme, if multiple variables with different physical units are involved, including but not limited to time, mass or energy variables, their joint appearance is to express the collaborative modeling relationship of multiple physical mechanisms. Each variable can be formed into a unified structure through function mapping, ratio combination or normalization adjustment. The units and meanings are clear, and the overall expression conforms to the principle of dimensional consistency and the common formula of engineering modeling.

[0118] Any constants, weights, adjustment factors, threshold parameters, and proportional coefficients involved in this solution are all adjustable control parameters for different application environments. Their values ​​depend on the target device configuration, data input characteristics, and performance optimization goals. During the implementation phase, they are set within a reasonable range through model verification, performance constraints, or engineering calibration. Although these parameters do not have preset unique values, they have clear adjustment logic and calculation paths and are part of the deterministic setting process in engineering implementation. The purpose of such setting is to ensure that the solution is both universally adaptable, reproducible, and operable, without affecting its technical clarity and feasibility.

[0119] In S1, define Indicates the Frame The local grayscale statistical feature vector extracted from the position is constructed:

[0120] ;

[0121] in:

[0122] ;

[0123] ;

[0124] ;

[0125] in is the local grayscale mean, in the above formula the local grayscale mean represents the arithmetic average of the pixel grayscale values ​​within the window; is the local grayscale standard deviation, in the above formula, the local grayscale standard deviation represents the square root of the sum of the squares of the grayscale values ​​of pixels in the window that deviate from the mean; is the local grayscale skewness, in the above formula, the local grayscale skewness represents the third-order central moment of the grayscale value in the window about the mean; Indicates the In the local grayscale area group of continuous frames, Extracted from the center Pixel gray value; is the number of pixels in the sliding window; is the frame number of the continuous frame grayscale image sequence; is the pixel coordinate of the center of the sliding window in the image;

[0126] definition Represents the set of connected regions after screening, and extracts connected regions and screens the coefficient of variation;

[0127] ;

[0128] in Indicates the first connected subsets; pixels express Pixel coordinates within ; Represents pixels Neighborhood pixel coordinates of ; Represents pixels The neighborhood set of ; is the Euclidean distance of the local statistical eigenvector; is the skewness difference adjustment factor, which is used to weigh the impact of skewness. In practical applications, According to the asymmetric sensitivity requirements of the local grayscale distribution, it can be selected in the range of [0.1, 10] through cross-validation. The larger the value, the more likely it is to enhance the weight of the local grayscale skewness, and the smaller the value, the more likely it is to balance the influence of the grayscale mean and variance. The skewness itself is the third-order central moment, and its dimension is more volatile than the variance. Taking 0.1 as the lower limit can avoid the skewness effect being smoothed out by numerical precision when it is less than 0.1. Taking 10 as the upper limit is because the skewness in the local area of ​​the natural image usually does not exceed [-5, 5]. Exceeding 10 times will lead to the loss of neighborhood gradient stability and amplification of noise errors. Therefore, 0.1 to 10 can ensure the effective expression of skewness features and avoid value loss under the actual grayscale distribution scale. The energy threshold preset for connectivity determination; For the Frame Pixels Gray value of express The average grayscale; is the preset grayscale variation energy threshold, which is used to filter out areas with high degradation levels; For the Frames in Pixels The local grayscale skewness corresponding to the position; For the Frames in Pixels The local grayscale skewness corresponding to the position; Indicates the Frames in Pixels Local grayscale statistical feature vector extracted from the position; Indicates the Frames in Pixels Local grayscale statistical feature vector extracted from the position;

[0129] definition is the standard Markov random field graph structure, which is composed of and composition;

[0130] ;

[0131] in:

[0132] ;

[0133] ;

[0134] in is the pixel coordinate The degradation probability value at ; for Grayscale variation energy, which represents the overall fluctuation intensity of the pixel grayscale value relative to the mean in a local area, is used to measure the intensity of grayscale changes and structural complexity within the region in practical applications. In the above formula, it is used to control the decay rate of the degradation probability Gaussian map; is the neighborhood degradation probability difference adjustment factor; in addition and It is a parameter set artificially based on empirical rules and task characteristics, which can be selected through experimental verification to adapt to different image degradation characteristics. The larger the degradation probability, the higher the discrimination. The smaller the neighborhood, the smoother the potential energy changes, and the two trade off between adjustment accuracy and robustness; is the pixel coordinate Neighborhood consistency potential value at ; For the Frame in pixel coordinates The gradient vector of For the Frames in Pixels The gradient vector of is the neighborhood gradient weight adjustment factor, It is a parameter set artificially based on experience, which is used to control the influence weight of the gradient difference in the neighborhood on the potential energy function. The larger the neighborhood gradient difference, the stronger the suppression, which helps to enhance the connectivity of the smooth area. The smaller it is, the more sensitive the edge details are and the better the structural resolution is. is the pixel coordinate The neighborhood pixel set of It is a natural exponential function, that is, an exponential operation with the natural constant e as the base, which is used to describe the exponential decay or growth process; is the two-norm, that is, the Euclidean norm of the vector, which is used to measure the Euclidean distance between two vectors; Pixels The degradation probability value at .

[0135] In S2, define For the Frame Pixels Grayscale first-order gradient vector:

[0136] ;

[0137] definition is the local gradient perturbation vector field:

[0138] ;

[0139] ;

[0140] ;

[0141] in For the Frame in pixel coordinates The gradient vector of is the pixel coordinate The Laplace enhanced gradient vector at ; Pixels The Laplace enhanced gradient vector at ; is the second-order derivative of the pixel, Laplace operator; is the pixel coordinate The neighborhood pixel set of is the local perturbation energy function. In the above formula, the local perturbation energy function is used to reflect the severity of the local perturbation; The exponential decay adjustment factor of the local perturbation energy function is used to control the influence of the perturbation intensity on the energy value. The larger the value, the faster the perturbation energy decays with the gradient, and the smaller the value, the slower the decay, which enhances the sensitivity of the local perturbation to the energy field. is the nonlinear amplification order of the disturbance energy, It is used to control the amplification amplitude of the disturbance intensity. The larger the value, the more obvious the nonlinear enhancement of the local disturbance is. The smaller the value, the weaker the enhancement effect is. It adjusts the sensitivity of the disturbance energy to large gradient changes.

[0142] definition For the Frame pixel coordinates The mutual information entropy value of is:

[0143] ;

[0144] definition Indicated in pixel coordinates The neighborhood mutual information entropy interaction matrix centered on is used to quantify the mutual information entropy difference between local pixels and neighboring pixels and their degree of coordinated change;

[0145] ;

[0146] definition is the coupled potential energy field:

[0147] ;

[0148] in is the pixel coordinate Gray value is The probability distribution of is a grayscale value set; For the Frame Pixels The mutual information entropy value of is the mutual information second-order synergy term adjustment factor; is the mutual information entropy adjustment weight, where the mutual information entropy adjustment weight is used to control the influence of the mutual information entropy on the coupling potential energy field. The larger the value, the more significant the contribution of neighborhood information difference to the potential energy. In practical applications, it can be adaptively set according to the complexity of local grayscale changes or determined through empirical parameter optimization;

[0149] definition Indicates the Pixel coordinates of the frame and the adjacent frames A continuous grayscale vector sequence:

[0150] ;

[0151] definition is the pixel coordinate Grayscale temporal coherence measure of:

[0152] ;

[0153] definition Represents pixels The migration probability is:

[0154] ;

[0155] in For the Frame pixel coordinates The local grayscale vector of ; For the Frame pixel coordinates The local grayscale vector of ; For the Frame pixel coordinates The local grayscale vector of ; is the time frame number, in In the formula The value range is { , , }; Indicates the Frame in pixel coordinates Extracted local grayscale vector; is the temporal change constraint adjustment factor, which is used to control the impact of temporal coherence on the transition probability; The nonlinear order of temporal coherence constraint is used to adjust the amplification or compression effect of the coherence influence; Represents the neighborhood pixel coordinates traversed in the normalized denominator, which is used to normalize all neighborhood transition probabilities; for Grayscale temporal coherence measure of ; For the frame The coupling potential energy value of the position.

[0156] In S3, define For the Frame pixel coordinates The local state label of ;

[0157] ;

[0158] ;

[0159] ;

[0160] in is a state label set; is the candidate state label; Pixels The current status label of The state consistency constraint adjustment factor is used to balance the impact of the consistency of the state labels of the neighborhood pixels on the overall energy minimization. The larger the value, the more consistent the state in the neighborhood. It can be determined according to the segmentation smoothness requirements or through cross-validation. is the Kronecker function, which is the indicator function. and If they are equal, the value is 1, otherwise it is 0; is the grayscale feature similarity constraint adjustment factor, which is used to control the impact of the grayscale statistical feature differences of neighboring pixels on energy minimization. The larger the value, the more consistent the grayscale features in the segmented area. It can be set according to the image texture complexity or experimental experience. For the Pixel coordinates in the frame Grayscale statistical feature vector of ; For the Pixels in frame Grayscale statistical feature vector of ; Represents the morphological closing operation based on the grayscale statistical matrix; It is the image structure restored after morphological closing operation; Indicates the first connected subsets; is the preset connected area threshold; is a set of connected regions; express The pixel area, defined as the number of pixels contained in the region, is used to quantify the region size;

[0161] definition Recover the set of vectors for the boundary:

[0162] ;

[0163] Calculate the boundary center point set and generate the initial reference point set :

[0164] ;

[0165] definition As the reference point Trajectory matrix on the time axis:

[0166] ;

[0167] in for The boundary set of is the boundary pixel coordinate; is the number of boundary pixels; Represents a neighborhood matching function based on grayscale feature vectors. In the above formula, the neighborhood matching function of grayscale feature vectors is used to measure the similarity of grayscale feature vectors of two pixel positions in consecutive frames. By minimizing the feature difference, the correspondence between the reference points in the time series is determined, supporting reference point tracking. is the number of consecutive frames; As the reference point In the The spatial coordinate position in the frame image is used to record the trajectory information of the reference point changing over time; As the reference point In the The grayscale statistical feature vector corresponding to the new coordinate position determined by the neighborhood matching of the frame; As the reference point In the The grayscale statistical feature vector corresponding to the new coordinate position determined by neighborhood matching.

[0168] In S4, define For the Frame reference point The displacement vector:

[0169] ;

[0170] definition is the smoothed displacement vector:

[0171] ;

[0172] in is the width of the smoothing time window; is the time frame number, in In the formula The value range is { , }; As the reference point In the The spatial coordinate position in the frame image;

[0173] definition As the reference point Local drift measure of :

[0174] ;

[0175] definition is the corrected displacement vector:

[0176] ;

[0177] in It is the discrete change rate symbol of adjacent frame numbers, which is used to measure the change rate of mutual information entropy in the direction of frame numbers; For the Frame at reference point current The mutual information entropy value is used to describe the uncertainty of the grayscale distribution of the local image; is the drift correction adjustment factor, which is used to control the influence of the local drift metric on the displacement vector correction;

[0178] definition As the reference point The confidence weight of :

[0179] ;

[0180] Definition The final structural displacement measurement output of the frame is :

[0181] ;

[0182] in is the weight matrix adjustment factor. In practical applications, the weight matrix adjustment factor is used to control right The degree of influence of , which determines the drift sensitivity and weight decay rate, The larger the value, the stronger the drift metric's inhibitory effect on the weight, the faster the weight decreases, and the ability to eliminate local drift-sensitive points is enhanced; the smaller the value, the weaker the impact of drift on the weight, allowing a greater drift tolerance. In practical applications, it can be selected through cross-validation, including but not limited to adjustment within the range of 0.01-1, and manual adjustment based on the image noise level and structural stability.

[0183] It should be noted that this solution is designed to address the key issues faced by monocular vision systems when applied to steel structure displacement measurement under high-temperature conditions. In high-temperature environments, carbonization, fogging, and coking inevitably occur on the surface of the optical window, leading to local optical degradation during the imaging process. This degradation effect causes image breakage and non-image blurring in the captured image, seriously interfering with structural recognition and displacement measurement based on traditional image processing methods. To this end, this solution is based on monocular vision image analysis technology from image to non-image, and overcomes the impact of optical degradation by constructing a dynamically degraded and adaptive image processing and displacement measurement link.

[0184] First, during the image acquisition phase, a monocular vision system is used to acquire a sequence of continuous frame grayscale images, and the image is divided into fixed-region sliding windows. For local regions, the grayscale distribution vector is statistically analyzed to form a local grayscale feature set, and a local grayscale statistical matrix is ​​constructed through the mean and second-order moment. Pixel-level connectivity is determined with the help of grayscale distribution characteristics, and connected regions with a coefficient of variation higher than a set threshold are screened to construct a regional degradation probability map. By introducing the neighborhood consistency rule, a standard Markov random field graph structure is established to describe the probabilistic correlation between local regions. The purpose of this part is to extract regional degradation information through local statistical characteristics, provide a reliable foundation for subsequent dynamic model construction, avoid misjudgment caused by local grayscale fluctuations, and improve the anti-degradation robustness of image analysis.

[0185] On this basis, we further rely on the standard Markov random field graph structure to extract the local grayscale first-order gradient vector, combine it with the neighborhood grayscale difference matrix for joint modeling, and generate a local gradient perturbation vector field; for local perturbations, we introduce an exponential decay adjustment factor and a nonlinear enhancement mechanism to construct a local perturbation energy function, thereby forming a perturbation potential energy map; in order to enhance the adaptability to complex grayscale changes, the regional mutual information entropy adjustment matrix is ​​superimposed to construct a coupled potential energy structure; by defining the coupled potential energy neighborhood update rule and combining the continuous frame grayscale time series change sequence, the regional temporal coherence constraint matrix is ​​generated; finally, the dual constraints of coupled potential energy and temporal coherence are combined to reconstruct the neighborhood transfer probability and form a variant Markov random field model; this part integrates spatial perturbation information with temporal coherence information, which not only improves the sensitivity to dynamic changes of degradation, but also effectively suppresses the loss of structural information caused by local degradation drift, and constructs a dynamic adaptive regional division basis;

[0186] Based on the above-mentioned variational Markov random field model, it is applied to the current frame grayscale image, and the local region state configuration is solved by minimizing the local energy function to generate a local region segmentation map. Subsequently, a morphological closing operation is performed based on the grayscale statistical matrix to repair local fractures and missing details caused by degradation, and a morphological structure recovery map is generated. Through connected region extraction and boundary connectivity screening, an effective boundary recovery vector group is obtained, and the coordinates of each boundary center point are calculated to form an initial reference point set for displacement measurement; based on the initial reference point set, combined with the continuous frame grayscale image sequence, the local grayscale features and connectivity relationships of the reference point set are tracked frame by frame to establish a reference point time series trajectory matrix; through multi-level structure recovery and reference point screening, it is ensured that under optical degradation conditions, the reference information points reflecting the real physical structure changes can still be stably extracted and tracked, laying the foundation for the subsequent displacement vector field construction;

[0187] Finally, based on the reference point time-series trajectory matrix, the displacement vector group is calculated frame by frame to form a displacement vector sequence; in view of the local non-physical jump trajectories that may appear under high-temperature conditions, a local time-series smoothing processing strategy is designed to effectively eliminate abnormal motion vectors and generate a smooth displacement vector field; further combined with the mutual information entropy change trend in the variant Markov random field model, the local scale drift situation is determined, and a local scale drift matrix is ​​constructed. Based on this, the displacement vector field is reversely corrected to form a corrected displacement measurement result set; by defining a regional confidence weight matrix, a weighted average is taken based on the local drift stability, and finally a stable and reliable structural displacement measurement result is output; this process ensures the accuracy and robustness of the displacement measurement results under complex degradation conditions through a three-level mechanism of time-series smoothing, drift judgment and confidence weighting;

[0188] Overall, this solution overcomes the problems of local occlusion and feature loss caused by optical degradation in monocular vision displacement measurement under high-temperature degradation conditions by designing a multi-stage, multi-level image analysis and processing chain from image information extraction, non-image feature recovery, temporal continuity modeling, local drift judgment to confidence weighting. The introduction of the standard Markov random field graph and its mutation model, combined with grayscale gradient perturbation analysis and mutual information entropy constraint mechanism, enhances the adaptability to complex grayscale changes and improves the stability of structure recognition and tracking.

[0189] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for measuring structural displacement under high-temperature conditions based on monocular vision, comprising a monocular vision system for collecting image data, characterized in that: S1. Based on the continuous frame grayscale image sequence collected by the monocular vision system, extract the local grayscale feature set and calculate the grayscale distribution parameters, complete the preliminary determination and screening of connected regions, establish the regional degradation probability map, and build the standard Markov random field graph structure; S2. Based on the standard Markov random field graph structure, the grayscale gradient feature set is extracted and the local gradient perturbation vector field is established to generate the perturbation potential energy map. The mutual information entropy adjustment matrix is ​​superimposed to construct the coupling potential energy structure. The temporal coherence constraint matrix is ​​constructed by combining the grayscale temporal change sequence. The neighborhood transition probability is jointly updated to form a variant Markov random field model. S3. Apply the variant Markov random field model to the current frame grayscale image, solve the local region state configuration to generate a segmentation map, perform morphological closing operations to generate a morphological structure recovery map, extract connected regions and filter the valid boundary set, calculate the coordinates of the boundary center points to form the initial reference point set for displacement measurement, combine the continuous frame grayscale image sequence to track the reference point set features frame by frame, and establish the reference point time series trajectory matrix; S4. Calculate the displacement vector sequence based on the reference point time-series trajectory matrix, smooth the process to generate a smooth displacement vector field, combine the mutual information entropy change trend of the variant Markov random field model to determine the local scale drift, and correct the displacement vector field to output the final structural displacement measurement result.

2. The method for measuring structural displacement under high temperature conditions based on monocular vision according to claim 1, characterized in that: In S1, a sequence of continuous frame grayscale images is collected through a monocular vision system, and a fixed area sliding window is performed to form a continuous frame local grayscale region group. The continuous frame local grayscale region group is divided according to the pixel neighborhood relationship, and the grayscale distribution vector of each local region is calculated to form a local grayscale feature set; Based on the local grayscale feature set, the mean and second-order moment of the grayscale distribution within the statistical region are calculated to generate a local grayscale statistical matrix. The local grayscale statistical matrix is ​​used to perform pixel-level connectivity judgment, extract connected regions, and form a preliminary connected region set. The coefficient of variation of the neighborhood grayscale of the preliminary connected region set is calculated, and regions with a coefficient of variation greater than a preset coefficient of variation threshold are screened to construct a regional degradation probability map. Based on the regional degradation probability graph, the neighborhood consistency rule is set to form the initial neighborhood dependency structure and generate the standard Markov random field graph structure.

3. The method for measuring structural displacement under high temperature conditions based on monocular vision according to claim 2, characterized in that: In S2, based on the standard Markov random field graph structure, the first-order grayscale gradient vector of the local area is extracted to form a grayscale gradient feature set. The grayscale gradient feature set is jointly modeled with the neighborhood grayscale difference matrix to generate a local gradient perturbation vector field. Based on the local gradient perturbation vector field, a local perturbation energy function is established, and the neighborhood potential energy function is updated to form a perturbation potential energy map. The regional mutual information entropy distribution is introduced, the interaction relationship between the local regional mutual information entropy and the neighborhood is counted, and the mutual information entropy adjustment matrix is ​​constructed. The perturbation potential energy map and the mutual information entropy adjustment matrix are superimposed to form a coupled potential energy structure, and the coupled potential energy neighborhood update rule is defined. The grayscale distribution vector groups extracted from the local grayscale region groups of continuous frames are arranged in chronological order to generate a local region grayscale temporal change sequence, and a regional temporal coherence constraint matrix is ​​constructed based on the change sequence; Taking the coupled potential energy structure of the neighborhood and the temporal coherence constraint matrix as joint constraints, the neighborhood transition probability is reconstructed to generate a mutational Markov random field model.

4. The method for measuring structural displacement under high temperature conditions based on monocular vision according to claim 3, characterized in that: In S3, the variant Markov random field model is applied to the grayscale image of the current frame, the local region state configuration is solved based on the neighborhood state transition probability matrix, and a local region segmentation map is generated. On the local region segmentation map, a morphological closing operation is performed based on the grayscale statistical matrix to generate a morphological structure recovery map; The connected areas in the morphological structure restoration graph are extracted, and the effective boundary set is screened based on the boundary connectivity to form a boundary restoration vector group. The coordinates of the boundary center point are calculated through the boundary restoration vector group to form the initial reference point set for displacement measurement; Based on the initial reference point set of displacement measurement and combined with the continuous frame grayscale image sequence, the local grayscale features and connectivity relationship of the reference point set are tracked frame by frame to establish the reference point temporal trajectory matrix.

5. The method for measuring structural displacement under high temperature conditions based on monocular vision according to claim 4, characterized in that: In S4, based on the reference point temporal trajectory matrix, the reference point displacement vector group is calculated frame by frame to form a displacement vector sequence. Local temporal smoothing is performed on the displacement vector sequence to eliminate local non-physical jump trajectories and output a smoothed displacement vector field. The smoothed displacement vector field is used to determine the local scale drift based on the mutual information entropy change trend of the local region's variant Markov random field model, and a local scale drift matrix is ​​constructed. Based on the local scale drift matrix, the displacement vector field is reversely corrected to form a corrected displacement measurement result set, which is then weighted averaged by the regional confidence weight matrix to form the final structural displacement measurement output result.

6. The method for measuring structural displacement under high temperature conditions based on monocular vision according to claim 5, characterized in that: In S1, define Indicates the Frame The local grayscale statistical feature vector extracted from the position is constructed: ； in: ； ； ； in is the local grayscale mean; is the local grayscale standard deviation; is the local grayscale skewness; Indicates the In the local grayscale area group of continuous frames, Extracted from the center Pixel gray value; is the number of pixels in the sliding window; is the frame number of the continuous frame grayscale image sequence; is the pixel coordinate of the center of the sliding window in the image; definition Represents the set of connected regions after screening, and extracts connected regions and screens the coefficient of variation; ； in Indicates the first connected subsets; pixels express Pixel coordinates within ; Represents pixels Neighborhood pixel coordinates of ; Represents pixels The neighborhood set of ; is the Euclidean distance of the local statistical eigenvector; is the skewness difference adjustment factor; The energy threshold preset for connectivity determination; For the Frame Pixels Gray value of express The average grayscale; is the preset grayscale variation energy threshold; For the Frames in Pixels The local grayscale skewness corresponding to the position; For the Frames in Pixels The local grayscale skewness corresponding to the position; Indicates the Frames in Pixels Local grayscale statistical feature vector extracted from the position; Indicates the Frames in Pixels Local grayscale statistical feature vector extracted from the position; definition is the standard Markov random field graph structure, which is composed of and composition; ； in: ； ； in is the pixel coordinate The degradation probability value at ; for Grayscale variation energy; In the above formula, it is used to control the decay rate of the degradation probability Gaussian map; is the neighborhood degradation probability difference adjustment factor; is the pixel coordinate Neighborhood consistency potential value at ; For the Frame in pixel coordinates The gradient vector of For the Frames in Pixels The gradient vector of is the neighborhood gradient weight adjustment factor; is the pixel coordinate The neighborhood pixel set of is the natural exponential function; is the two-norm; Pixels The degradation probability value at .

7. The method for measuring structural displacement under high temperature conditions based on monocular vision according to claim 6, characterized in that: In S2, define For the Frame Pixels Grayscale first-order gradient vector: ； definition is the local gradient perturbation vector field: ； ； ； in For the Frame in pixel coordinates The gradient vector of is the pixel coordinate Laplace enhanced gradient vector at ; Pixels Laplace enhanced gradient vector at ; is the second-order derivative of the pixel; is the pixel coordinate The neighborhood pixel set of is the local perturbation energy function; is the exponential decay adjustment factor of the local perturbation energy function; is the nonlinear amplification order of the disturbance energy; definition For the Frame pixel coordinates The mutual information entropy value of is: ； definition Indicated in pixel coordinates The neighborhood mutual information entropy interaction matrix centered on is used to quantify the mutual information entropy difference between local pixels and neighboring pixels and their degree of coordinated change; ； definition is the coupled potential energy field: ； in is the pixel coordinate Gray value is The probability distribution of is a grayscale value set; For the Frame Pixels The mutual information entropy value of is the mutual information second-order synergy term adjustment factor; Adjust the weights for mutual information entropy; definition Indicates the Frame and adjacent frame pixel coordinates A continuous grayscale vector sequence: ； definition is the pixel coordinate Grayscale temporal coherence measure of: ； definition Represents pixels The migration probability is: ； in For the Frame pixel coordinates The local grayscale vector of ; For the Frame pixel coordinates The local grayscale vector of ; For the Frame pixel coordinates The local grayscale vector of ; Number the timeframe; Indicates the Frame in pixel coordinates Extracted local grayscale vector; is the timing variation constraint adjustment factor; Constrain the nonlinear order for temporal coherence; Represents the neighborhood pixel coordinates traversed in the normalized denominator; for Grayscale temporal coherence measure of ; For the frame The coupling potential energy value of the position.

8. The method for measuring structural displacement under high temperature conditions based on monocular vision according to claim 7, characterized in that: In S3, define For the Frame pixel coordinates The local state label of ; ； ； ； in is a state label set; is the candidate state label; Pixels The current status label of is the state consistency constraint adjustment factor; is the Kronecker function, which is the indicator function. and If they are equal, the value is 1, otherwise it is 0; is the grayscale feature similarity constraint adjustment factor; For the Pixel coordinates in the frame Grayscale statistical feature vector of ; For the Pixels in frame Grayscale statistical feature vector of ; Represents the morphological closing operation based on the grayscale statistical matrix; It is the image structure restored after morphological closing operation; Indicates the first connected subsets; is the preset connected area threshold; is a set of connected regions; express The pixel area; definition Recover the set of vectors for the boundary: ； Calculate the boundary center point set and generate the initial reference point set : ； definition As the reference point Trajectory matrix on the time axis: ； in for The boundary set of is the boundary pixel coordinate; is the number of boundary pixels; Represents the neighborhood matching function based on grayscale feature vector; is the number of consecutive frames; As the reference point In the The spatial coordinate position in the frame image; As the reference point In the The grayscale statistical feature vector corresponding to the new coordinate position determined by the neighborhood matching of the frame; As the reference point In the The grayscale statistical feature vector corresponding to the new coordinate position determined by neighborhood matching.

9. The method for measuring structural displacement under high temperature conditions based on monocular vision according to claim 8, characterized in that: In S4, define For the Frame reference point The displacement vector: ； definition is the smoothed displacement vector: ； in is the width of the smoothing time window; Number the timeframe; As the reference point In the The spatial coordinate position in the frame image; definition As the reference point Local drift measure of : ； definition is the corrected displacement vector: ； in is the sign of the discrete rate of change on adjacent frame numbers; For the Frame at reference point current The mutual information entropy value of is the drift correction adjustment factor; definition As the reference point The confidence weight of : ； Definition The final structural displacement measurement output of the frame is : ； in is the weight matrix adjustment factor.

Images