Wall detection instrument wood detection intelligent marking system

Abstract

The present application relates to the technical field of wood detection, and discloses a wall detector wood detection intelligent marking system, which comprises a first scanning unit, an analysis unit, a second scanning unit, a comparison unit, a calculation unit and a marking unit. The first scanning unit is used for constructing a multi-dimensional discrimination baseline containing a dielectric distribution entropy matrix, a regional dielectric fluctuation confidence interval and an interlayer dielectric transition coefficient, which is suitable for multi-layer walls. The analysis unit is used for screening suspicious areas and excluding normal interlayer interference. The second scanning unit is used for fine scanning at multiple angles, calculating anisotropic speckle parameters, and capturing the structure and orientation difference of wood structures and repair materials. The dynamic threshold of the comparison unit is used for distinguishing material types. The dynamic safety threshold of the calculation unit is combined with the accurate construction boundary of the marking unit, and different marking is performed according to the confidence level, so as to effectively exclude the interference of cement, gypsum and other repair materials, ensure the accurate positioning of original wood structures such as Ming and Qing wooden dowels and wood tie bars in ancient houses, and avoid construction damage.

Description

Technical Field

[0001] This invention relates to the field of wood detection technology, and more specifically to a smart marking system for wood detection using a wall detector. Background Technology

[0002] Currently, most wall detectors use electromagnetic waves to detect targets. They identify targets by pre-setting the density or conductivity range of wood and common wall materials. During operation, the detector moves close to the wall. When the wood structure parameters detected by the sensor fall into the wood threshold range, it is determined that wood is present, and the location is indicated by light or sound. Then, the staff marks it to facilitate later construction positioning.

[0003] However, existing detection markers often have significant flaws when it comes to buildings like ancient houses. Specifically, the walls of ancient houses are mostly multi-layered composite structures, and some walls are repaired later with cement, plaster, etc. The physical properties, especially the density, of these repair materials from different eras are very similar to the wooden structure inside the ancient house (such as Ming and Qing dynasty wooden tenons and wooden bracing). When the wall detector is used, because the density of the wooden structure and the repair materials are very similar, it is easy to misjudge the repair materials as wood, resulting in the deviation of the marker position and easily damaging the original wooden structure of the ancient house. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a smart marking system for wood detection in wall detectors, which solves the aforementioned problems.

[0005] The above-mentioned technical objective of the present invention is achieved through the following technical solution:

[0006] The intelligent marking system for wood detection in wall detectors includes:

[0007] The first scanning unit is used to scan the target area according to the first scanning mode of the wall detector, acquire the first echo signal data of the target area, analyze the distribution characteristics of the preprocessed first echo signal data, and obtain the discrimination baseline. The target area is: the wall of the old house.

[0008] The analysis unit is used to analyze the first echo signal data and the discrimination baseline, screen out regions whose dielectric properties deviate from the discrimination baseline, and generate a map of suspicious regions.

[0009] The second scanning unit is used to scan each suspicious point in the suspicious area map based on the second scanning mode of the wall detector to obtain the second echo signal data; the coherence difference and attenuation characteristics of the preprocessed second echo signal data are calculated to obtain the anisotropic speckle parameters. The second echo signal data is multi-angle echo signal data.

[0010] The comparison unit is used to perform cluster analysis on the anisotropic speckle parameters of multiple suspicious points to obtain the first dynamic threshold and the second dynamic threshold. The anisotropic speckle parameters of the suspicious point are compared with the first dynamic threshold and the second dynamic threshold respectively to generate the baseline correction coefficient and the confidence level of the woody traits.

[0011] The calculation unit is used to calculate the confidence level of all wood properties and generate the dynamic safety threshold of the target area;

[0012] The marking unit is used to identify suspicious points that are determined to have a loose fibrous structure and whose woody properties have a confidence level exceeding the dynamic safety threshold, and to designate them as key avoidance points; the spatial coordinates of all key avoidance points are analyzed to generate construction safety boundaries and mark them.

[0013] Furthermore, the distribution characteristics of the preprocessed first echo signal data are analyzed to obtain the discrimination baseline, including:

[0014] The dielectric distribution entropy matrix of the wall material is obtained by extracting the first echo signal data after preprocessing.

[0015] Based on the dielectric distribution entropy matrix of wall material, general material regions without structural anomalies are screened within the target area, and the dielectric property fluctuation range of these general material regions is calculated to obtain the confidence interval of regional dielectric fluctuation.

[0016] The dielectric property variation law of the multilayer structure in the target region is analyzed to obtain the interlayer dielectric transition coefficient;

[0017] The dielectric distribution entropy matrix of wall materials, the confidence interval of regional dielectric fluctuations, and the interlayer dielectric transition coefficient are integrated to generate a multi-dimensional dielectric discrimination baseline.

[0018] Furthermore, by analyzing the first echo signal data and the discrimination baseline, regions whose dielectric properties deviate from the discrimination baseline are identified, and a map of suspected regions is generated, including:

[0019] Based on the preprocessed first echo signal data, the dielectric value of each scanning point is calculated point by point to exceed the confidence interval of the dielectric fluctuation in the region, and the dielectric over-limit of a single point is obtained.

[0020] Based on the interlayer dielectric transition coefficient, analyze the rate of change of dielectric properties of adjacent scanning points to generate the dielectric gradient residual of adjacent points;

[0021] Based on the dielectric distribution entropy matrix of the wall material, the dielectric distribution entropy of each scanning point in the local area of ​​10×10mm and the target area is calculated to generate the local-global entropy discrepancy.

[0022] The composite anomaly value is obtained by fusing the single-point dielectric out-of-bounds value, the neighboring point dielectric variation residual, and the local-to-global entropy divergence.

[0023] Analyze the distribution of composite anomaly quantization values ​​of all scan points, take the average of the top 5% of values ​​as the anomaly identification threshold, screen out scan points whose composite anomaly quantization values ​​exceed the anomaly identification threshold, record their three-dimensional spatial coordinates, and generate a list of suspicious point coordinates;

[0024] The coordinates in the list of suspicious points are mapped according to the actual scale of the wall, and the composite anomaly quantization value of each suspicious point is used as the deviation to generate a suspicious area map.

[0025] Furthermore, the coherence difference and attenuation characteristics of the preprocessed second echo signal data are calculated to obtain anisotropic speckle parameters, including:

[0026] For the preprocessed second echo signal, the signal group is divided according to the scanning angle, and the angular timing offset of each group of signals in the same propagation depth segment is calculated.

[0027] The fluctuation threshold is obtained by calculating the confidence interval of regional dielectric fluctuation and the interlayer dielectric transition coefficient;

[0028] Based on the angular time offset, depth segment signals with time fluctuations exceeding the fluctuation threshold are filtered out, and the interlayer interference stripping coefficient is calculated for the depth segment signals.

[0029] The attenuation compensation calibration of the second echo signal intensity is performed based on the interlayer interference stripping coefficient to obtain the calibrated second echo signal set.

[0030] Based on the calibrated second echo signal set, the coherence difference at different depths under the same angle and the attenuation characteristics at the same depth under different angles are analyzed to obtain the reverse coupling degree.

[0031] Furthermore, the coherence difference and attenuation characteristics of the preprocessed second echo signal data are calculated to obtain anisotropic speckle parameters, including:

[0032] A dynamic analysis window is constructed based on the reverse coupling degree. Within the window, the degree of dispersion of the coherence-attenuation correlation curve at each scanning point is calculated to obtain the dynamic structure reference deviation value.

[0033] Based on the dynamic structural reference deviation value, the directional distribution characteristics of the reverse coupling degree at different angles are analyzed to obtain the fiber orientation vector value;

[0034] The dynamic structural reference deviation value and fiber orientation vector value are integrated to generate anisotropic speckle parameters.

[0035] Furthermore, cluster analysis was performed on the anisotropic speckle parameters of multiple suspicious points to obtain the first dynamic threshold and the second dynamic threshold, including:

[0036] The anisotropic speckle parameters of all suspicious points and the interlayer dielectric transition coefficients corresponding to each suspicious point are calculated to generate interlayer penetration correction values ​​for different wall layers.

[0037] Based on the interlayer penetration correction value, the structural homogeneity correlation characteristics and fiber orientation correlation characteristics of each suspicious point are analyzed to generate a material property correlation matrix.

[0038] Cluster analysis was performed on the correlation matrix of material properties to obtain the cluster density gradient values;

[0039] Based on the cluster density gradient value, two clusters are determined, and the inter-cluster transition fit coefficient of the parameters at the boundary of the two clusters is calculated.

[0040] The inter-cluster transition adaptation coefficients are calibrated based on the confidence interval of regional dielectric fluctuations to obtain the calibrated adaptation coefficients.

[0041] Based on the calibrated adaptation coefficients, the first dynamic threshold and the second dynamic threshold are determined.

[0042] Furthermore, based on the anisotropic speckle parameters of the suspected point, the first dynamic threshold and the second dynamic threshold are compared respectively to generate baseline correction coefficients and woody trait confidence scores, including:

[0043] When the comprehensive value of the anisotropic speckle parameters is less than or equal to the first dynamic threshold, the internal material of the suspicious point is determined to be a dense homogeneous body. The dielectric properties of the suspicious point are calculated to determine the deviation from the discrimination baseline, and a baseline correction coefficient is generated.

[0044] When the comprehensive value of the anisotropic speckle parameters is greater than or equal to the second dynamic threshold, the internal material of the suspicious point is determined to be a loose fibrous structure. The anisotropic speckle parameters, the second dynamic threshold, and the second echo signal of the suspicious point are analyzed to generate the confidence level of the woody properties.

[0045] Furthermore, the confidence levels for all woody traits are calculated to generate a dynamic safety threshold for the target area, including:

[0046] The confidence scores of the wood properties of all suspicious points identified as loose fibrous structures and the interlayer dielectric transition coefficients corresponding to each suspicious point are calculated according to different wall layers to obtain the interlayer confidence weighted value.

[0047] Based on the inter-layer confidence weighted value, the distribution range of all inter-layer confidence weighted values ​​is determined, and a dynamic safety threshold is generated.

[0048] Furthermore, the spatial coordinates of all critical avoidance points are analyzed to generate and mark construction safety boundaries, including:

[0049] Based on the three-dimensional spatial coordinates of all key avoidance points and the corresponding interlayer dielectric transition coefficients, the vertical projection deviation of coordinates in different wall layers is calculated to obtain the interlayer projection deviation value of coordinates.

[0050] The coordinates of all critical avoidance points are corrected based on the inter-coordinate projection deviation value, and the boundary extension redundancy is calculated by combining the confidence interval of regional dielectric fluctuation.

[0051] Based on the corrected coordinates of the critical avoidance points and the boundary expansion redundancy, an initial construction safety boundary is generated.

[0052] Calculate the minimum distance from all critical avoidance points within the initial construction safety boundary to the boundary to obtain the boundary coverage verification value.

[0053] Furthermore, the spatial coordinates of all critical avoidance points are analyzed to generate and mark construction safety boundaries, which also includes:

[0054] When the boundary coverage verification value meets the preset requirements, the final construction safety boundary is generated;

[0055] Based on the confidence level of the woody characteristics of the key avoidance points, all key avoidance points are differentiated and labeled.

[0056] In summary, the present invention has the following main beneficial effects:

[0057] The first scanning unit constructs a dielectric distribution entropy matrix of the wall material using high-precision sampling, accurately capturing subtle differences in dielectric properties at each wall surface. Then, it analyzes the dielectric variation patterns of each layer using interlayer dielectric transition coefficients to adapt to multi-layer wall structures. Simultaneously, it screens out abnormal areas to establish regional dielectric fluctuation confidence intervals, clarifying the normal characteristic range of general-purpose materials. The analysis unit further integrates single-point dielectric out-of-bounds values, adjacent-point dielectric variation residuals, and local-to-global entropy discrepancies to generate composite anomaly values. The top 5% of highly anomaly points are then selected as suspicious points. This process overcomes the limitations of traditional fixed thresholds, effectively eliminating interference from repair materials such as cement and plaster, generating accurate suspicious area maps, and avoiding misclassification of repair materials as wooden structures during the coarse scanning stage.

[0058] The second scanning unit employs multi-angle fixed-point scanning, dividing the signal depth segment according to the time delay-layer correspondence, calculating the angular time sequence offset, and determining the fluctuation threshold by combining the regional dielectric fluctuation confidence interval and the interlayer dielectric transition coefficient. Abnormal signals exceeding normal interlayer fluctuations are screened out, and the second echo signal intensity is compensated and calibrated by the interlayer interference stripping coefficient to eliminate the interference of different wall layers on the signal, restoring the true signal characteristics of the raw materials. Based on the calibration signal, the reverse coupling degree and dynamic structural reference deviation value are further calculated, and the fiber orientation vector value is extracted (wood structures have clear fiber orientations, while repair materials have no directional fibers). Finally, these are integrated into anisotropic speckle parameters. The comparison unit uses cluster analysis to divide the parameters into dense homogeneous bodies (repair materials) and loose fibrous structures (wood structures). The credibility of the wood structure is quantified by combining the wood property confidence measure to avoid misjudgment caused by similar physical properties.

[0059] This system calculates the interlayer confidence weight for each layer by weighting the confidence level of the wood properties of suspected loose fibrous points with the interlayer dielectric transition coefficient. A dynamic safety threshold is then generated by subtracting three times the standard deviation from the mean. Simultaneously, the marking unit further optimizes positioning accuracy: the three-dimensional coordinates of key avoidance points are corrected using interlayer projection deviation values ​​to eliminate depth positioning errors caused by interlayer dielectric differences. Boundary expansion redundancy is calculated by combining the regional dielectric fluctuation confidence interval to ensure that the construction boundary covers all high-risk points. Boundary coverage verification is then used to prevent key avoidance points from exceeding the boundary. Finally, key avoidance points are differentiated and marked according to the wood property confidence level, ensuring precise and controllable construction boundaries. This effectively avoids damage to the original wooden structure of the ancient house due to marking deviations and facilitates subsequent construction positioning. Attached Figure Description

[0060] Figure 1 This is a schematic diagram of the intelligent marking system for wood detection in the wall detector of the present invention. Detailed Implementation

[0061] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0062] refer to Figure 1 The intelligent marking system for wood detection in wall detectors includes:

[0063] The first scanning unit is used to scan the target area according to the first scanning mode of the wall detector, acquire the first echo signal data of the target area, analyze the distribution characteristics of the preprocessed first echo signal data, and obtain the discrimination baseline. The target area is the wall of the old house. The first scanning mode is: the wall detector moves and scans the target area at a constant speed, and transmits electromagnetic waves with a lower power than the second scanning mode and a wider beamwidth than the second scanning mode.

[0064] The analysis unit is used to analyze the first echo signal data and the discrimination baseline, screen out regions whose dielectric properties deviate from the discrimination baseline, and generate a map of suspicious regions.

[0065] The second scanning unit is used to scan each suspicious point in the suspicious area map based on the second scanning mode of the wall detector to obtain second echo signal data; the coherence difference and attenuation characteristics of the preprocessed second echo signal data are calculated to obtain anisotropic speckle parameters. The second echo signal data is multi-angle echo signal data. The second scanning mode is: the wall detector performs fixed-point scanning on the suspicious point, transmits electromagnetic waves with higher power and narrower beamwidth than the first scanning mode, and scans the suspicious point at at least three different angles;

[0066] The comparison unit is used to perform cluster analysis on the anisotropic speckle parameters of multiple suspicious points to obtain the first dynamic threshold and the second dynamic threshold. The anisotropic speckle parameters of the suspicious point are compared with the first dynamic threshold and the second dynamic threshold respectively to generate the baseline correction coefficient and the confidence level of the woody traits.

[0067] The calculation unit is used to calculate the confidence level of all wood properties and generate the dynamic safety threshold of the target area;

[0068] The marking unit is used to identify suspicious points that are determined to have a loose fibrous structure and whose woody properties have a confidence level exceeding the dynamic safety threshold, and to designate them as key avoidance points; the spatial coordinates of all key avoidance points are analyzed to generate construction safety boundaries and mark them.

[0069] By establishing a multi-dimensional dielectric discrimination baseline through a wide-beam coarse scan in the first scanning mode, and adapting the interlayer dielectric transition coefficient to the multi-layer structure of the ancient house wall, and then performing multi-angle scanning through a narrow-beam scan in the second scanning mode to obtain anisotropic speckle parameters, and generating dynamic thresholds through cluster analysis, repair materials with similar physical properties and wooden structures can be accurately distinguished. At the same time, through dual screening of wood property confidence and dynamic safety threshold, key avoidance points can be accurately located and construction safety boundaries can be generated, which greatly reduces the probability of repair materials being misjudged as wooden structures and avoids damage to the original wooden tenons, wooden tie bars and other structures of the ancient house during construction.

[0070] In one embodiment, the distribution characteristics of the preprocessed first echo signal data are analyzed to obtain a discrimination baseline, including:

[0071] The preprocessed first echo signal data is extracted to obtain the dielectric distribution entropy matrix of the wall material. Specifically, this involves: when the wall detector moves at a constant speed to scan the wall of the ancient house, sampling points are set according to the horizontal (X-axis) and vertical (Y-axis) directions of the wall surface. Each 1cm along the X-axis is a sampling point, and each 1cm along the Y-axis is a sampling point. Each spatial coordinate corresponds to a set of preprocessed first echo signal time-domain amplitudes. The time-domain amplitudes of all spatial coordinates are arranged according to the spatial coordinates to form a two-dimensional signal matrix (rows correspond to the X-axis, columns correspond to the Y-axis, and matrix elements are sequences of time-domain amplitudes at corresponding positions). The overall amplitude range of all elements (time-domain amplitudes at each spatial position) in the signal matrix is ​​divided into equal intervals from the minimum to the maximum value. The signal matrix is ​​divided into 16 continuous intervals until all amplitudes are covered. 32 sampling points (corresponding to 32 continuous spatial coordinates on the wall, such as a 2×16 two-dimensional window consisting of 16 points on the X-axis and 2 points on the Y-axis) are used as a sliding window. Within the two-dimensional space of the signal matrix, the window traverses the signal matrix in steps of one sampling point (moving one spatial coordinate along the X-axis or Y-axis each time), ensuring that the signal in each local area of ​​the wall is covered by the window. Within each window, the frequency of signal occurrence in each interval is divided by the total number of sampling points in the window to obtain the probability of each interval. The probability of each interval is then multiplied by the natural logarithm of that interval probability, and the result is summed and the negative value is taken; this is the dielectric entropy value at the center of the window. By traversing all areas, the dielectric entropy matrix of the wall material for the entire region can be generated.

[0072] Based on the dielectric distribution entropy matrix of wall materials, regions of general materials without structural anomalies are screened within the target area. The dielectric property fluctuation range of these general material regions is calculated to obtain the confidence interval of regional dielectric fluctuation. Specifically, this includes: calculating the variance of each dielectric distribution entropy value in the dielectric distribution entropy matrix within its 3×3 neighborhood; setting a threshold of 30% of the global entropy variance, and screening all regions with neighborhood variances below the threshold, which are general material regions without anomalies; calculating the mean and standard deviation of all dielectric distribution entropy values ​​in the general material regions, adding 3 to the mean and multiplying by the standard deviation to obtain the upper limit of the interval, and subtracting 3 from the mean and multiplying by the standard deviation to obtain the lower limit of the interval. The interval formed by the upper and lower limits is the confidence interval of regional dielectric fluctuation.

[0073] The dielectric property variation law of the multilayer structure in the target area is analyzed to obtain the interlayer dielectric transition coefficient. Specifically, the signal delay range of 0-1ns corresponds to the lime surface layer, the signal delay range of 1-5ns corresponds to the mortar layer, and the signal delay range of 5-10ns corresponds to the brick masonry base layer. From the dielectric distribution entropy matrix, the dielectric distribution entropy values ​​of each wall layer (brick masonry base layer, mortar bonding layer, lime surface layer) are extracted along the direction (perpendicular to the wall surface) according to the wall depth (the wall depth is obtained by multiplying the propagation speed of electromagnetic waves in the medium by the echo signal delay and then dividing by 2). Each layer has an independent entropy value sequence. In the entropy value sequence of the corresponding layer, select 10 consecutive sampling points along the X-axis (each sampling point is 1 cm apart). Each sampling point corresponds to one dielectric distribution entropy value. Calculate the difference between the dielectric distribution entropy values ​​of adjacent sampling points. Then add the 9 differences (out of a total of 10 sampling points, there are only 9 differences between adjacent sampling points) and divide by 9 to obtain the average change. Divide the average change by 1 (representing 1 cm) to obtain the slope. Divide the average absolute value of the slope of two adjacent layers by the average value of the dielectric distribution entropy values ​​of these two layers to obtain the interlayer dielectric transition coefficient between the two layers.

[0074] The dielectric distribution entropy matrix of the wall material, the confidence interval for regional dielectric fluctuations, and the interlayer dielectric transition coefficient are integrated to generate a multi-dimensional dielectric discrimination baseline. Specifically, this includes setting the weights of the wall material dielectric distribution entropy matrix to 0.4, the regional dielectric fluctuation confidence interval to 0.35, and the interlayer dielectric transition coefficient to 0.25. For each spatial point in the wall material dielectric distribution entropy matrix, if the dielectric distribution entropy value at that point is within the regional dielectric fluctuation confidence interval, then the dielectric distribution entropy value is multiplied by 0.4 and used as the actual contribution value. If the dielectric entropy value at a point is within the confidence interval for dielectric fluctuations in the region, calculate the deviation of the dielectric entropy value at that point: If the dielectric entropy value at that point is greater than the upper limit of the confidence interval for dielectric fluctuations in the region, subtract the upper limit from the dielectric entropy value at that point to obtain the absolute deviation; if the dielectric entropy value at that point is less than the lower limit of the confidence interval for dielectric fluctuations in the region, subtract the dielectric entropy value at that point from the lower limit of the interval to obtain the absolute deviation; then subtract the lower limit from the upper limit of the interval to obtain the width; divide the absolute deviation by the width to obtain the value of the dielectric entropy value at that point. The deviation of a point is calculated; multiplying the deviation by 0.4 and then by (1 - deviation) yields the actual contribution value of that point; multiplying the actual contribution value by the mean of the regional dielectric fluctuation confidence interval by 0.35 and the interlayer dielectric transition coefficient by 0.25 gives the multidimensional dielectric value of each point; arranging the multidimensional dielectric values ​​of all spatial points on the wall surface according to the X-axis row and Y-axis column positions forms a two-dimensional matrix, which is the discrimination baseline; the dielectric distribution entropy matrix of the wall material directly corresponds to the sampling points per 1cm of the wall surface. Dielectric properties can accurately reflect whether there is a wooden structure in a local area, and are the core basis for anomaly location, so the weight is relatively high at 0.4; while the confidence interval of regional dielectric fluctuation represents the normal dielectric range of general materials and is the benchmark for judging whether the points of the dielectric distribution entropy matrix of wall materials are abnormal. Without this benchmark, it is impossible to distinguish between normal fluctuation and wooden structure anomalies, so the weight is relatively low at 0.35; the interlayer dielectric transition coefficient is only used to correct the dielectric changes of multi-layer walls. The core of ancient house detection is to find the wooden structure rather than to identify the layers, so it only plays an auxiliary correction role, so the weight is the lowest at 0.25.

[0075] By constructing a two-dimensional signal matrix and combining it with a sliding window to calculate the dielectric distribution entropy, subtle differences in the dielectric properties of local materials can be accurately captured. This overcomes the shortcomings of traditional detection methods, which have weak ability to distinguish materials with similar densities. By screening out areas without anomalies, the confidence interval for dielectric fluctuations is determined. At the same time, the interlayer dielectric transition coefficient is introduced to adapt to the structural characteristics of multi-layer composite walls in ancient houses, avoiding misjudgments caused by changes in interlayer dielectric. The baseline for judgment can comprehensively reflect the local characteristics of materials, the overall fluctuation range, and the interlayer transition law, thereby improving the accuracy of distinguishing between repair materials with similar physical properties and wooden structures.

[0076] By calculating the deviation of points where the dielectric distribution entropy value exceeds the normal range and dynamically correcting the contribution value, the judgment baseline can adapt to local anomalies in the wall, avoiding misjudgment of repair materials due to accidental proximity to wood parameters.

[0077] In one embodiment, the analysis of the first echo signal data and the discrimination baseline, the screening of regions whose dielectric properties deviate from the discrimination baseline, and the generation of a suspicious region map include:

[0078] Based on the preprocessed first echo signal data, the dielectric value of each scan point is calculated point by point to determine the value exceeding the regional dielectric fluctuation confidence interval, thus obtaining the single-point dielectric overshoot. Specifically, this involves: finding the dielectric distribution entropy value corresponding to each scan point (i.e., the dielectric distribution entropy value corresponding to that scan point in the wall material dielectric distribution entropy matrix) from the preprocessed first echo signal data; if the dielectric distribution entropy value > the upper limit of the regional dielectric fluctuation confidence interval, subtracting the upper limit of the regional dielectric fluctuation confidence interval from the dielectric distribution entropy value to obtain the single-point dielectric overshoot; if the dielectric distribution entropy value < the lower limit of the regional dielectric fluctuation confidence interval, subtracting the dielectric distribution entropy value from the regional dielectric fluctuation confidence interval to obtain the single-point dielectric overshoot; if the lower limit of the regional dielectric fluctuation confidence interval ≤ the dielectric distribution entropy value ≤ the upper limit of the regional dielectric fluctuation confidence interval, then the single-point dielectric overshoot is 0.

[0079] Based on the interlayer dielectric transition coefficient, the dielectric property change rate of adjacent scan points is analyzed to generate the dielectric grade residual of adjacent points. Specifically, this includes: for adjacent scan points, subtracting the dielectric distribution entropy value of the previous scan point from the dielectric distribution entropy value of the subsequent scan point, and dividing the calculation result by 1 (1 represents a spacing of 1 cm) to obtain the actual change rate; calculating the absolute value of the difference between the actual change rate and the interlayer dielectric transition coefficient, and using the absolute value as the dielectric grade residual of adjacent points.

[0080] Based on the dielectric entropy matrix of the wall material, the dielectric entropy of each scanning point in a 10×10mm local region and the target region is calculated to generate the local-to-global entropy discrepancy. Specifically, this includes: 1×1 sampling points corresponding to the 10×10mm local region (1 sampling point for every 1cm along the X and Y axes, 10mm = 1cm); calculating the mean of the dielectric entropy values ​​of all sampling points in the local region, and using the mean as the local average entropy; dividing the sum of the dielectric entropy values ​​of all sampling points in the target region by the total number of sampling points to obtain the global average entropy; and calculating the absolute value of the difference between the local average entropy and the global average entropy, using the absolute value as the local-to-global entropy discrepancy.

[0081] The composite anomaly value is obtained by fusing the single-point dielectric out-of-bounds value, the neighboring point dielectric variation residual, and the local-to-global entropy discrepancy. Specifically, the weights are set as follows: the single-point dielectric out-of-bounds value, the neighboring point dielectric variation residual, and the local-to-global entropy discrepancy measure are set to 0.4, 0.3, and 0.3 respectively; the single-point dielectric out-of-bounds value, the neighboring point dielectric variation residual, and the local-to-global entropy discrepancy measure of each scan point are multiplied by their respective weights and then summed to obtain the composite anomaly value of that scan point; where the single-point dielectric out-of-bounds value is 0.4, the neighboring point dielectric variation residual, and the local-to-global entropy discrepancy measure are set to 0.3. The out-of-bounds quantity directly reflects whether the entropy value of a single scan point exceeds the confidence interval of general materials. It is the core basis for distinguishing between timber structures and general materials. The abnormal signal is the most direct, so its weight is relatively high at 0.4. The dielectric variation residual of adjacent points is used to eliminate misjudgments caused by normal dielectric changes between layers. It only plays a corrective role, so its weight is relatively low at 0.3. The local-to-global entropy discrepancy verifies whether anomalies are concentrated (timber structures are mostly locally distributed) through the difference between local and global entropy. It serves as an auxiliary verification, so its weight is relatively low at 0.3.

[0082] The distribution of composite anomaly quantization values ​​of all scan points was analyzed. The average of the top 5% of values ​​was used as the anomaly identification threshold. Scan points whose composite anomaly quantization values ​​exceeded the threshold were selected, and their three-dimensional spatial coordinates were recorded to generate a list of suspicious point coordinates. Specifically, the composite anomaly quantization values ​​of all scan points were sorted in descending order. For the top 5% of values ​​after sorting, these composite anomaly quantization values ​​were added together and divided by the number of values ​​to obtain the anomaly identification threshold. The composite anomaly values ​​of each scan point were compared with the anomaly identification threshold. Scan points that exceeded the threshold were identified as suspicious points. The three-dimensional coordinates of all suspicious points (X and Y are wall coordinates, and Z is the depth coordinate, which is the wall depth) were recorded and compiled into a list of suspicious point coordinates.

[0083] The coordinates in the list of suspicious points are mapped according to the actual scale of the wall. The composite anomaly quantization value of each suspicious point is used as the deviation to generate a suspicious area map. Specifically, since each 1cm of the X and Y axes of the sampling points corresponds to 1cm of the actual size of the wall, the X and Y coordinate values ​​in the list of suspicious points are directly used as the actual X and Y coordinates of the wall to complete the mapping. Suspicious points are marked at the corresponding mapped coordinates, and the composite anomaly quantization value of each point is marked as the deviation next to the corresponding suspicious point to form a suspicious area map.

[0084] By calculating the dielectric excess at each point, the scanning points where the dielectric properties exceed the normal range of general materials are identified. Then, combined with the interlayer dielectric transition coefficient, the dielectric variation residuals of adjacent points are calculated. This effectively eliminates interference from normal dielectric changes between layers of multi-layer walls, preventing the natural transition of repair materials between layers from being misjudged as abnormal wood structure. It also solves the problem of misjudgment caused by the similarity in density between the two materials. This improves the accuracy of distinguishing materials with similar physical properties.

[0085] By using precise coordinate mapping, the marking deviation of the detection of wooden structures in ancient houses is reduced. Combined with the recording of the location of suspicious points in three-dimensional coordinates, it can be adapted to the multi-layered composite walls of ancient houses. It can clearly identify whether the suspicious points are located in the lime layer, mortar layer or brick base layer, avoiding the mislabeling of repair materials in different layers as wooden structures. At the same time, the composite anomaly value is marked as the deviation degree in the map, which intuitively presents the degree of abnormality of the suspicious points and reduces the invalid detection of normal areas. The resulting map of suspicious areas allows staff to accurately locate high-risk points and avoid damage to the original wooden tenon and mortise joints, wooden tie bars and other structures of ancient houses due to marking deviation.

[0086] In one embodiment, the coherence difference and attenuation characteristics of the preprocessed second echo signal data are calculated to obtain anisotropic speckle parameters, including:

[0087] For the preprocessed second echo signal, the signal groups are divided according to the scanning angle, and the angular timing offset of each group of signals within the same propagation depth segment is calculated. Specifically, the signal groups are divided according to the actual scanning angle (three angles) of the second scanning mode, with each angle corresponding to one group of second echo signals; the same propagation signal depth segment is set according to the time delay between each layer (signal delay of 0-1ns corresponds to the lime surface layer, signal delay of 1-5ns corresponds to the mortar layer, and signal delay of 5-10ns corresponds to the brick masonry base layer); for each group of signals, the peak occurrence time of the echo signal within the signal depth segment is calculated; then the difference in peak occurrence time between signal groups at different angles within the same depth segment is calculated, and this difference is the angular timing offset of each group of signals within the same propagation depth segment.

[0088] The fluctuation threshold is obtained by calculating the confidence interval of regional dielectric fluctuation and the interlayer dielectric transition coefficient. Specifically, this includes: subtracting the lower limit from the upper limit of the confidence interval of regional dielectric fluctuation to obtain the width; adding the interlayer dielectric transition coefficients of all layers and dividing by the number of transition coefficients to obtain the average transition coefficient; and multiplying the width by the average transition coefficient to obtain the fluctuation threshold.

[0089] Based on the angular time-series offset, depth segment signals with time-series fluctuations exceeding the fluctuation threshold are filtered out, and the interlayer interference stripping coefficient is calculated for each depth segment signal. Specifically, this involves: for each signal depth segment, comparing its angular time-series offset with the fluctuation threshold one by one; retaining the signal depth segment if the angular time-series offset exceeds the fluctuation threshold, and not retaining the signal if the angular time-series offset does not exceed the fluctuation threshold; for the retained signal depth segments, calculating the mean signal intensity at different angles; then calculating the absolute value of the difference between the signal intensity at each angle and the mean; finally, dividing the absolute value by the mean to obtain the interlayer interference stripping coefficient for that angle.

[0090] The attenuation compensation calibration of the second echo signal intensity is performed based on the interlayer interference stripping coefficient to obtain the calibrated second echo signal set. Specifically, for the original signal intensity of the second echo in the retained depth segment at each angle, the original signal intensity is multiplied by (1 + the interlayer interference stripping coefficient at that angle) to obtain the compensated signal intensity of that depth segment at that angle. Finally, the compensated signals of all angles and all retained signal depth segments are integrated to obtain the calibrated second echo signal set.

[0091] Based on the calibrated second echo signal set, the coherence difference at different depths under the same angle and the attenuation characteristics at the same depth under different angles are analyzed to obtain the reverse coupling degree. Specifically, this includes: dividing the calibration signals of two adjacent depth segments under the same angle into equal-length sub-segments with 10 consecutive sampling points to ensure that the sub-segments of the two adjacent depth segments correspond one-to-one with the same spatial position; for each pair of corresponding sub-segments, calculating the sum of squares of the amplitude differences at corresponding positions in their respective signal amplitude sequences, and taking the reciprocal of the sum of squares as the local coherence value of the pair of sub-segments; summing the local coherence values ​​of all sub-segments and dividing by the total number of sub-segments to obtain the correlation coefficient between the two depth segments; subtracting the correlation coefficient from 1 to obtain the coherence difference value at different depths under the same angle; calculating the mean value of the calibration signal intensity at each angle under the same depth segment, and dividing the difference between the calibration signal intensity at each angle and the mean value by the mean value to obtain the attenuation rate; taking the standard deviation of the attenuation rate at each angle as the attenuation characteristic value at the same depth under different angles; and multiplying the coherence difference value and the attenuation characteristic value to obtain the reverse coupling degree.

[0092] The second echo signal group is divided into three angles. The angular time sequence offset of the same depth segment is calculated by combining the time delay-layer correspondence. Then, the fluctuation threshold is determined based on the regional dielectric fluctuation confidence interval and the average interlayer transition coefficient. Abnormal depth segments that exceed normal interlayer fluctuations are screened out to avoid misjudging the natural dielectric changes of multi-layer walls as timber structure signals. The signal strength is compensated by the interlayer interference stripping coefficient to eliminate the interference of different wall layers on the signal and restore the true signal characteristics of the raw materials. The reverse coupling degree is calculated by the coherence difference and attenuation characteristics to accurately capture the subtle differences in signal correlation and attenuation law between timber structure and repair materials, thereby improving the accuracy of distinguishing between the two.

[0093] In one embodiment, the calculation of the coherence difference and attenuation characteristics of the preprocessed second echo signal data to obtain anisotropic speckle parameters further includes:

[0094] A dynamic analysis window is constructed based on the inverse coupling degree. Within the window, the dispersion of the coherence-attenuation correlation curve for each scan point is calculated to obtain the dynamic structure reference deviation value. Specifically, this includes: for the inverse coupling degree of all scan points, finding the range between its maximum and minimum values ​​as the inverse coupling degree range; sorting the inverse coupling degrees of all scan points, using the 1 / 3 and 2 / 3 quantiles of the sorted data as dividing nodes to divide the inverse coupling degree range into three intervals, each containing an equal number of scan points; each interval corresponds to one dynamic analysis window, covering 5×5 sampling points; within each dynamic analysis window, the coherence difference value of each scan point is plotted on the horizontal axis, and the attenuation characteristic value on the vertical axis; for the coherence difference value and attenuation characteristic value of each scan point; calculating the mean of all coherence difference values, using this as the first mean, and calculating the mean of the attenuation characteristic values... This is used as the second mean; calculate the difference between each first mean and the first mean, and calculate the difference between each second mean and the second mean; multiply the difference between the first and second means at each scan point by the difference between the second and second means, and then add the products of all scan points to get the numerator; calculate the difference between the coherence difference value and the first mean, and then calculate the difference between the attenuation characteristic value and the second mean, multiply these two differences at each scan point, and then add the products of all scan points to get the denominator; divide the numerator by the denominator to get the slope; subtract the slope multiplied by the first mean from the second mean to get the intercept; the straight line determined by attenuation characteristic value = slope multiplied by coherence difference value + intercept is the coherence-attenuation correlation curve; calculate the vertical distance from each scan point to the coherence-attenuation correlation curve, and multiply the mean of the vertical distances of all scan points by the mean of the reverse coupling degree within the window to get the dynamic structure reference deviation value;

[0095] Based on the dynamic structural reference deviation value, the directional distribution characteristics of the reverse coupling degree at different angles are analyzed to obtain the fiber orientation vector value. Specifically, this includes: dividing the reverse coupling degree corresponding to the three scanning angles by the dynamic structural reference deviation value of each angle to obtain the directional feature weight of each angle; using the scanning angle as the polar angle and the directional feature weight as the polar radius, marking the three vector points corresponding to the three angles on the polar coordinate diagram according to the polar angle corresponding to the scanning angle and the polar radius corresponding to the directional feature weight; adding the directional feature weights of the three point vectors to obtain the composite vector; the polar angle of the composite vector is the fiber orientation angle, and dividing the polar radius of the composite vector by the composite vector yields the orientation confidence. The two constitute the fiber orientation vector value (fiber orientation angle, orientation confidence).

[0096] The dynamic structural reference deviation value and fiber orientation vector value are integrated to generate anisotropic speckle parameters. Specifically, this includes: setting the weight of the dynamic structural reference deviation value to 0.5 and the weight of the orientation confidence value in the fiber orientation vector value to 0.5; multiplying the dynamic structural reference deviation value and the orientation confidence value by their respective weights and then adding them together to obtain a comprehensive value of structural homogeneity-orientation confidence; integrating this comprehensive value with the fiber orientation angle in the fiber orientation vector value to form a binary parameter of "comprehensive value, fiber orientation angle", which is the anisotropic speckle parameter representing the internal structural homogeneity and fiber orientation of the suspicious point.

[0097] A 5×5 dynamic analysis window was constructed using reverse coupling, and the dynamic structural baseline deviation was calculated using the dispersion of the coherence-attenuation correlation curve. This quantified the difference in signal correlation between the loose fiber structure of wood and the dense homogeneous structure of the repair material, initially distinguishing the two from the dimension of structural homogeneity. Then, fiber orientation vector values ​​were extracted by synthesizing vectors using polar coordinates. Utilizing the unique fiber orientation characteristics of wood structures (the repair material has no directional fibers), and combining orientation confidence, the wood structure signal was further locked. Finally, structural homogeneity and orientation confidence were integrated into a binary parameter. The resulting anisotropic speckle parameter can accurately distinguish material types from both structural and orientation dimensions, significantly reducing the misjudgment rate.

[0098] In one embodiment, cluster analysis is performed on the anisotropic speckle parameters of multiple suspicious points to obtain a first dynamic threshold and a second dynamic threshold, including:

[0099] The anisotropic speckle parameters and interlayer dielectric transition coefficients corresponding to all suspicious points are calculated to generate interlayer penetration correction values ​​for different wall layers (brick masonry base layer, mortar bonding layer, and lime finish layer). Specifically, the suspicious points are classified according to each wall layer (brick masonry base layer, mortar bonding layer, and lime finish layer). For each suspicious point in each category, the comprehensive value of its anisotropic speckle parameters is multiplied by the interlayer dielectric transition coefficient of that layer to obtain the single-point interlayer penetration correction value. The single-point correction values ​​of all points in the same layer are added together and then divided by the total number of suspicious points in that layer to obtain the interlayer penetration correction value of the corresponding wall layer.

[0100] Based on the interlayer penetration correction value, the structural homogeneity correlation characteristics and fiber orientation correlation characteristics of each suspicious point are analyzed to generate a material property correlation matrix. Specifically, for each suspicious point, the comprehensive value of its anisotropic speckle parameter is divided by the interlayer penetration correction value of the layer in which it is located to obtain the structural homogeneity correlation characteristics; the fiber orientation angle of the anisotropic speckle parameter of the suspicious point is multiplied by the interlayer penetration correction value of the layer in which it is located to obtain the fiber orientation correlation characteristics; then, the sampling points on the wall surface are sorted according to the coordinates of the X-axis and Y-axis, with each coordinate corresponding to a suspicious point, and the matrix elements are the "structural homogeneity correlation characteristics and fiber orientation correlation characteristics" of that point. The resulting two-dimensional matrix is ​​the material property correlation matrix.

[0101] Cluster analysis is performed on the material property correlation matrix to obtain the cluster density gradient value. Specifically, this includes: calculating the difference between the structural homogeneity correlation characteristics and the fiber orientation correlation characteristics of adjacent suspicious points in the material property correlation matrix, and calculating the mean of the two differences; for each suspicious point, suspicious points with a difference exceeding twice the mean are defined as parametric coherence mutation points, and these parametric coherence mutation points are used as the initial cluster centers; with each initial cluster center as the core, two sampling points are extended to the X-axis and two to the Y-axis (since each sampling point is 1cm, extending two points is 2cm), forming a 5×5 square neighborhood; the total number of suspicious points in the neighborhood is used as the cluster density; the difference between the cluster density of the initial cluster center and the cluster density of the neighborhood edge points (the actual number of suspicious points within the outermost ring of the neighborhood (the four sides of 5cm×5cm)) is calculated, and the difference is divided by the neighborhood radius (2, representing 2cm), which is the cluster density gradient value.

[0102] Based on the cluster density gradient values, two clusters are identified, and the inter-cluster transition fit coefficient of the parameters at the boundary between the two clusters is calculated. Specifically, this includes: dividing the sum of the cluster density gradient values ​​of all suspicious points by the total number of cluster density gradient values ​​to obtain the global mean; selecting two regions whose cluster density gradient values ​​are in the top 20% and whose adjacent point gradient values ​​are all higher than the global mean; these two regions are the two significant clusters; the point with the largest cluster density gradient value between the two significant clusters is taken as the boundary point; calculating the mean of the structural homogeneity association characteristics and the mean of the fiber orientation association characteristics of the boundary point; then calculating the mean of the structural homogeneity association characteristics and the mean of the fiber orientation association characteristics of the core region of the significant cluster (the 5 points with the highest neighborhood density); for the structure For the uniformity association feature, the mean of the structural uniformity association feature at the boundary point is divided by the mean of the structural uniformity association feature corresponding to the two significant clusters to obtain two structural uniformity association feature values. The mean of the two structural uniformity association feature values ​​is calculated and used as the first fitting association value. For the fiber orientation association feature, the mean of the fiber orientation association feature at the boundary point is divided by the mean of the fiber orientation association feature corresponding to the two significant clusters to obtain two fiber orientation association feature values. The mean of the two fiber orientation association feature values ​​is calculated and used as the second fitting association value. The mean of the first fitting association value and the second fitting association value is calculated, and this mean is the inter-cluster transition fitting coefficient of the parameter at the boundary of the two clusters.

[0103] The inter-cluster transition adaptation coefficients are calibrated based on the regional dielectric fluctuation confidence interval to obtain the calibrated adaptation coefficients. Specifically, for the regional dielectric fluctuation confidence interval, the interval width (upper limit minus lower limit) and the interval mean (upper limit plus lower limit divided by 2) are calculated; the inter-cluster transition adaptation coefficients are multiplied by (interval mean divided by interval width) to obtain the calibrated adaptation coefficients.

[0104] Based on the calibrated adaptation coefficients, the first dynamic threshold and the second dynamic threshold are determined, specifically including: for two significant clusters, the attributes of the significant clusters are: the cluster corresponding to dense homogeneous materials is the lower limit cluster, and the cluster corresponding to loose fibrous materials is the upper limit cluster; calculate the mean of the structural homogeneity correlation characteristics of the core regions of the two significant clusters to obtain the mean of the lower limit cluster and the mean of the upper limit cluster respectively; multiply the mean of the lower limit cluster by the calibrated adaptation coefficient to obtain the first dynamic threshold; multiply the mean of the upper limit cluster by the calibrated adaptation coefficient to obtain the second dynamic threshold.

[0105] By calculating the interlayer penetration correction value according to the brick masonry base layer, mortar layer, etc., the interference of different wall layers on the anisotropic speckle parameters is eliminated, ensuring the accuracy of material characteristic analysis within the same layer and avoiding the obscuring of the difference between repair materials and wood structure due to interlayer dielectric differences. Then, by cluster analysis, two types of clusters are located: dense and homogeneous (repair materials) and loose and fibrous (wood structure). Combined with the inter-cluster transition adaptation coefficient and the regional dielectric interval calibration threshold, the dynamic threshold can accurately match the characteristic fluctuation of each layer of material, rather than using a uniform standard. The first and second thresholds determined in the end can clearly distinguish between the two types of materials, greatly reducing misjudgment caused by similar density.

[0106] In one embodiment, the anisotropic speckle parameters of the suspected point are compared with a first dynamic threshold and a second dynamic threshold to generate a baseline correction coefficient and a woody trait confidence level, respectively, including:

[0107] When the comprehensive value of the anisotropic speckle parameters is less than or equal to the first dynamic threshold, the internal material of the suspicious point is determined to be a dense homogeneous body. The deviation of the dielectric properties of the suspicious point from the discrimination baseline is calculated, and a baseline correction coefficient is generated. The discrimination baseline is adjusted according to the baseline correction coefficient. Specifically, when the internal material of the suspicious point is a dense homogeneous body, the absolute value of the difference between the dielectric distribution entropy value of the suspicious point and the multidimensional dielectric value of the corresponding point of the discrimination baseline is divided by the multidimensional dielectric value of the corresponding point of the discrimination baseline to obtain the dielectric property deviation. The dielectric property deviation is subtracted from 1 and multiplied by 0.6 to obtain the baseline correction coefficient. The multidimensional dielectric value of the position corresponding to the suspicious point in the discrimination baseline is multiplied by (1 + baseline correction coefficient) to obtain the updated discrimination baseline.

[0108] When the first dynamic threshold < the comprehensive value in the anisotropic speckle parameters < the second dynamic threshold, the difference between the comprehensive value in the anisotropic speckle parameters and the first and second dynamic thresholds is calculated respectively. The two differences are compared to determine whether the internal material of the suspicious point is a dense homogeneous body or a loose fibrous structure. Specifically, when the first dynamic threshold < the comprehensive value in the anisotropic speckle parameters < the second dynamic threshold, the comprehensive value of the anisotropic speckle parameters of the suspicious point is subtracted from the first dynamic threshold to obtain the first threshold difference; then the comprehensive value is subtracted from the second dynamic threshold to obtain the second threshold difference. Both differences are positive. For the first threshold difference and the second threshold difference, if the difference of the first dynamic threshold is smaller, the internal material of the suspicious point is determined to be a dense homogeneous body; if the difference of the second dynamic threshold is smaller, it is determined to be a loose fibrous structure.

[0109] When the comprehensive value of the anisotropic speckle parameters is greater than or equal to the second dynamic threshold, the internal material of the suspicious point is determined to be a loose fibrous structure. The anisotropic speckle parameters, the second dynamic threshold, and the second echo signal of the suspicious point are analyzed to generate the confidence level of woody characteristics. Specifically, this includes: when the comprehensive value of the anisotropic speckle parameters is greater than or equal to the second dynamic threshold, subtracting the second dynamic threshold from the comprehensive value of the anisotropic speckle parameters of the suspicious point, and then dividing by the second dynamic threshold to obtain the basic confidence ratio; for the interlaminar interference stripping coefficient corresponding to the suspicious point, calculating the mean of the interlaminar interference stripping coefficient at all angles; multiplying the basic confidence ratio by 0.6 and adding the mean of the interlaminar interference stripping coefficient by 0.4 to obtain the confidence level of woody characteristics.

[0110] By using the comprehensive value of anisotropic speckle parameters as the core, fine differentiation is achieved in three intervals: when the value is below the first dynamic threshold, it is clearly identified as a dense homogeneous body (such as cement, gypsum and other repair materials), and a baseline correction coefficient is generated by the dielectric property deviation to dynamically update the discrimination baseline and avoid repeated misclassification of similar materials. When the value is between the two thresholds, the difference comparison is used to further accurately classify the material. When the value is above the second dynamic threshold, the confidence level of wood properties is calculated by combining the mean value of interlayer interference peeling coefficient to further verify the wood structure properties and prevent the misclassification of highly similar repair materials as wood structures. This breaks through the limitations of single density judgment and improves the subsequent detection accuracy through dynamic correction, reducing the bias of the marking.

[0111] In one embodiment, the confidence levels of all wood properties are calculated to generate a dynamic safety threshold for the target area, including:

[0112] The confidence scores of the wood properties of all suspicious points identified as having a loose fibrous structure and the interlayer dielectric transition coefficients corresponding to each suspicious point are calculated according to different wall layers to obtain the interlayer confidence weighted value. Specifically, this includes: classifying suspicious points identified as having a loose fibrous structure into the corresponding wall layers; for each wall layer, multiplying the confidence score of the wood properties of each suspicious point within the layer by its corresponding interlayer dielectric transition coefficient and summing them to obtain the weighted confidence sum within the layer; then summing the interlayer dielectric transition coefficients of all suspicious points within the layer to obtain the weighted sum within the layer; and finally dividing the weighted confidence sum within the layer by the weighted sum within the layer to obtain the interlayer confidence weighted value for that layer.

[0113] Based on the inter-layer confidence weighted value, the distribution range of all inter-layer confidence weighted values ​​is determined, and a dynamic safety threshold is generated. Specifically, this includes: calculating the mean and standard deviation of the inter-layer confidence weighted values ​​corresponding to different wall layers; subtracting 3 from the mean and multiplying by the standard deviation to obtain the lower limit of the reasonable distribution range of all inter-layer confidence weighted values, which is the dynamic safety threshold of the target area.

[0114] In one embodiment, the spatial coordinates of all critical avoidance points are analyzed to generate and mark construction safety boundaries, including:

[0115] Based on the three-dimensional spatial coordinates of all critical avoidance points and the corresponding interlayer dielectric transition coefficients, the vertical projection deviation of the coordinates in different wall layers is calculated to obtain the interlayer projection deviation value. Specifically, this includes: assigning critical avoidance points to corresponding wall layers according to different wall layers, with each critical avoidance point in each layer containing three-dimensional coordinates (X and Y are wall surface coordinates, and Z is depth coordinate) and the corresponding interlayer dielectric transition coefficient; calculating the mean of the depth coordinates of all critical avoidance points in each layer; then, for each individual critical avoidance point in each layer, calculating the absolute value of the difference between its depth coordinate and the mean; multiplying the absolute value of the difference by the dielectric transition coefficient of that layer to obtain the corrected single-point deviation; summing the corrected single-point deviations of each layer and dividing by the total number of critical avoidance points in that layer to obtain the interlayer projection deviation value of the coordinates for that layer.

[0116] The coordinates of all critical avoidance points are corrected based on the inter-layer projection deviation of the coordinate layers, and the boundary extension redundancy is calculated in combination with the regional dielectric fluctuation confidence interval. Specifically, for each critical avoidance point, the inter-layer projection deviation of the coordinate layers of the wall layer where it is located is added to the depth coordinates corresponding to that point, while the X and Y wall coordinates remain unchanged, to obtain the corrected three-dimensional coordinates; then the width of the regional dielectric fluctuation confidence interval (upper limit minus lower limit of the interval) is multiplied by 0.5 to obtain the boundary extension redundancy.

[0117] Based on the corrected coordinates of critical avoidance points and the boundary expansion redundancy, an initial construction safety boundary is generated. Specifically, this includes: finding the minimum and maximum values ​​of the X-coordinate, Y-coordinate, and Z-coordinate for each layer of corrected critical avoidance point 3D coordinates; then adding or subtracting the boundary expansion redundancy for each coordinate extreme value (for example, subtracting the boundary expansion redundancy from the minimum value of the X-coordinate and adding the boundary expansion redundancy to the maximum value of the X-coordinate; the operations for the Y and Z coordinates are the same as for the X-coordinate). The expanded X, Y, and Z ranges for each layer constitute the initial boundary of that layer. Integrating all layer boundaries gives the overall initial construction safety boundary.

[0118] Calculate the minimum distance from all critical avoidance points within the initial construction safety boundary to the boundary to obtain the boundary coverage verification value. Specifically, this includes: calculating the distance from each critical avoidance point within the initial construction safety boundary to the boundary in the X, Y, and Z directions; taking the minimum value among the three directions for each critical avoidance point as the minimum distance from that critical avoidance point to the boundary; collecting the minimum distances of all critical avoidance points and taking the minimum value as the boundary coverage verification value.

[0119] In one embodiment, the analysis of the spatial coordinates of all critical avoidance points to generate and mark construction safety boundaries further includes:

[0120] When the boundary coverage verification value meets the preset requirements, the final construction safety boundary is generated. Specifically, the preset requirements are: boundary coverage verification value ≥ 0; when the boundary coverage verification value ≥ 0, it indicates that all critical avoidance points are within the initial construction safety boundary, and no critical avoidance points exceed the boundary; the overall initial construction safety boundary is directly used as the final construction safety boundary; where, when the boundary coverage verification value < 0, it indicates that there are critical avoidance points exceeding the initial construction safety boundary; all critical avoidance points with negative distances to the boundary are identified, and their directions and out-of-bounds values ​​are recorded; the extreme values ​​of the out-of-bounds direction corresponding to the initial boundary are extended to the out-of-bounds side by the maximum out-of-bounds value in that direction; based on the adjusted boundary, the minimum distance from all critical avoidance points to the boundary is recalculated to obtain a new boundary coverage verification value. If the boundary coverage verification value still cannot be ≥ 0, all steps after the boundary coverage verification value < 0 are repeated until the boundary coverage verification value ≥ 0; the steps are repeated a maximum of 3 times; if more than 3 times, an alarm is triggered, requiring manual intervention.

[0121] Based on the confidence level of the wood properties of the critical avoidance points, all critical avoidance points are differentiated and marked. Specifically, this includes: calculating the maximum and minimum confidence levels of the wood properties of all critical avoidance points, and using the maximum and minimum values ​​as the confidence level distribution range; dividing the confidence level distribution range into three equal confidence levels: red for high confidence level intervals, blue for medium confidence level intervals, and yellow for low confidence level intervals; determining the confidence level interval to which each critical avoidance point belongs and matching it with the corresponding color; and highlighting each critical avoidance point according to the matching color on the wall coordinate mapping diagram of the final construction safety boundary, thus completing the differentiated marking.

[0122] By classifying loose fibrous suspicious points according to wall layers and calculating interlayer confidence weights using the interlayer dielectric transition coefficient as the weight, the confidence contribution of timber points is strengthened due to the difference in transition coefficients between the repair material and the layer where the timber structure is located. This reduces the interference of repair materials and avoids misjudging the cement mortar layer as the timber tie bar of the brick base. At the same time, the interlayer projection deviation value of the coordinates is calculated based on the interlayer dielectric transition coefficient to correct the depth coordinates of key avoidance points, ensuring that the timber structure is accurately located to a specific layer. This avoids damaging the deep original wall in order to find shallow timber tenons, thus reducing the risk of mis-excavation from both the judgment and positioning dimensions.

[0123] By combining the confidence interval width of regional dielectric fluctuations to calculate the boundary extension redundancy, the extreme values ​​of the corrected critical avoidance point coordinates are expanded to prevent the omission of timber structure points due to small fluctuations in dielectric properties. Then, by calculating the minimum distance from all critical avoidance points to the boundary, the integrity of the boundary coverage is verified to ensure that no timber structure points exceed the boundary. The generated construction safety boundary is also marked with differential markings to ensure that the original structures such as Ming and Qing wooden tenons and wooden tie bars are not damaged, and to facilitate later construction positioning.

[0124] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A wood detection intelligent marking system for wall detectors, characterized in that, include: The first scanning unit is used to scan the target area according to the first scanning mode of the wall detector, acquire the first echo signal data of the target area, analyze the distribution characteristics of the preprocessed first echo signal data, and obtain the discrimination baseline. The target area is: the wall of the old house. The analysis unit is used to analyze the first echo signal data and the discrimination baseline, screen out regions whose dielectric properties deviate from the discrimination baseline, and generate a map of suspicious regions. The second scanning unit is used to scan each suspicious point in the suspicious area map based on the second scanning mode of the wall detector to obtain the second echo signal data; the coherence difference and attenuation characteristics of the preprocessed second echo signal data are calculated to obtain the anisotropic speckle parameters. The second echo signal data is multi-angle echo signal data. The comparison unit is used to perform cluster analysis on the anisotropic speckle parameters of multiple suspicious points to obtain the first dynamic threshold and the second dynamic threshold. The anisotropic speckle parameters of the suspicious point are compared with the first dynamic threshold and the second dynamic threshold respectively to generate the baseline correction coefficient and the confidence level of the woody traits. The calculation unit is used to calculate the confidence level of all wood properties and generate the dynamic safety threshold of the target area; The marking unit is used to identify suspicious points that are determined to have a loose fibrous structure and whose woody properties have a confidence level exceeding the dynamic safety threshold, and to designate them as key avoidance points; the spatial coordinates of all key avoidance points are analyzed to generate construction safety boundaries and mark them.

2. The intelligent marking system for wood detection using a wall detector according to claim 1, characterized in that, The distribution characteristics of the preprocessed first echo signal data are analyzed to obtain the discrimination baseline, including: The dielectric distribution entropy matrix of the wall material is obtained by extracting the first echo signal data after preprocessing. Based on the dielectric distribution entropy matrix of wall material, general material regions without structural anomalies are screened within the target area, and the dielectric property fluctuation range of these general material regions is calculated to obtain the confidence interval of regional dielectric fluctuation. The dielectric property variation law of the multilayer structure in the target region is analyzed to obtain the interlayer dielectric transition coefficient; The dielectric distribution entropy matrix of wall materials, the confidence interval of regional dielectric fluctuations, and the interlayer dielectric transition coefficient are integrated to generate a multi-dimensional dielectric discrimination baseline.

3. The intelligent marking system for wood detection using a wall detector according to claim 2, characterized in that, Analyzing the first echo signal data and the discrimination baseline, regions whose dielectric properties deviate from the discrimination baseline are screened out, generating a map of suspicious regions, including: Based on the preprocessed first echo signal data, the dielectric value of each scanning point is calculated point by point to exceed the confidence interval of the dielectric fluctuation in the region, and the dielectric over-limit of a single point is obtained. Based on the interlayer dielectric transition coefficient, analyze the rate of change of dielectric properties of adjacent scanning points to generate the dielectric gradient residual of adjacent points; Based on the dielectric distribution entropy matrix of the wall material, the dielectric distribution entropy of each scanning point in the local area of ​​10×10mm and the target area is calculated to generate the local-global entropy discrepancy. The composite anomaly value is obtained by fusing the single-point dielectric out-of-bounds value, the neighboring point dielectric variation residual, and the local-to-global entropy divergence. Analyze the distribution of composite anomaly quantization values ​​of all scan points, take the average of the top 5% of values ​​as the anomaly identification threshold, screen out scan points whose composite anomaly quantization values ​​exceed the anomaly identification threshold, record their three-dimensional spatial coordinates, and generate a list of suspicious point coordinates; The coordinates in the list of suspicious points are mapped according to the actual scale of the wall, and the composite anomaly quantization value of each suspicious point is used as the deviation to generate a suspicious area map.

4. The intelligent marking system for wood detection using a wall detector according to claim 2, characterized in that, The coherence difference and attenuation characteristics of the preprocessed second echo signal data are calculated to obtain anisotropic speckle parameters, including: For the preprocessed second echo signal, the signal group is divided according to the scanning angle, and the angular timing offset of each group of signals in the same propagation depth segment is calculated. The fluctuation threshold is obtained by calculating the confidence interval of regional dielectric fluctuation and the interlayer dielectric transition coefficient; Based on the angular time offset, depth segment signals with time fluctuations exceeding the fluctuation threshold are filtered out, and the interlayer interference stripping coefficient is calculated for the depth segment signals. The attenuation compensation calibration of the second echo signal intensity is performed based on the interlayer interference stripping coefficient to obtain the calibrated second echo signal set. Based on the calibrated second echo signal set, the coherence difference at different depths under the same angle and the attenuation characteristics at the same depth under different angles are analyzed to obtain the reverse coupling degree.

5. The intelligent marking system for wood detection using a wall detector according to claim 4, characterized in that, The coherence difference and attenuation characteristics of the preprocessed second echo signal data are calculated to obtain anisotropic speckle parameters, which also include: A dynamic analysis window is constructed based on the reverse coupling degree. Within the window, the degree of dispersion of the coherence-attenuation correlation curve at each scanning point is calculated to obtain the dynamic structure reference deviation value. Based on the dynamic structural reference deviation value, the directional distribution characteristics of the reverse coupling degree at different angles are analyzed to obtain the fiber orientation vector value; The dynamic structural reference deviation value and fiber orientation vector value are integrated to generate anisotropic speckle parameters.

6. The intelligent marking system for wood detection using a wall detector according to claim 5, characterized in that, Cluster analysis was performed on the anisotropic speckle parameters of multiple suspicious points to obtain the first dynamic threshold and the second dynamic threshold, including: The anisotropic speckle parameters of all suspicious points and the interlayer dielectric transition coefficients corresponding to each suspicious point are calculated to generate interlayer penetration correction values ​​for different wall layers. Based on the interlayer penetration correction value, the structural homogeneity correlation characteristics and fiber orientation correlation characteristics of each suspicious point are analyzed to generate a material property correlation matrix. Cluster analysis was performed on the correlation matrix of material properties to obtain the cluster density gradient values; Based on the cluster density gradient value, two clusters are determined, and the inter-cluster transition fit coefficient of the parameters at the boundary of the two clusters is calculated. The inter-cluster transition adaptation coefficients are calibrated based on the confidence interval of regional dielectric fluctuations to obtain the calibrated adaptation coefficients. Based on the calibrated adaptation coefficients, the first dynamic threshold and the second dynamic threshold are determined.

7. The intelligent marking system for wood detection using a wall detector according to claim 6, characterized in that, Based on the comparison of the anisotropic speckle parameters of the suspected point with the first and second dynamic thresholds, baseline correction coefficients and woody trait confidence scores are generated, including: When the comprehensive value of the anisotropic speckle parameters is less than or equal to the first dynamic threshold, the internal material of the suspicious point is determined to be a dense homogeneous body. The dielectric properties of the suspicious point are calculated to determine the deviation from the discrimination baseline, and a baseline correction coefficient is generated. When the comprehensive value of the anisotropic speckle parameters is greater than or equal to the second dynamic threshold, the internal material of the suspicious point is determined to be a loose fibrous structure. The anisotropic speckle parameters, the second dynamic threshold, and the second echo signal of the suspicious point are analyzed to generate the confidence level of the woody properties.

8. The intelligent marking system for wood detection using a wall detector according to claim 7, characterized in that, Calculate the confidence scores for all woody traits to generate a dynamic safety threshold for the target area, including: The confidence scores of the wood properties of all suspicious points identified as loose fibrous structures and the interlayer dielectric transition coefficients corresponding to each suspicious point are calculated according to different wall layers to obtain the interlayer confidence weighted value. Based on the inter-layer confidence weighted value, the distribution range of all inter-layer confidence weighted values ​​is determined, and a dynamic security threshold is generated.

9. The intelligent marking system for wood detection using a wall detector according to claim 1, characterized in that, The spatial coordinates of all critical avoidance points are analyzed to generate and mark construction safety boundaries, including: Based on the three-dimensional spatial coordinates of all key avoidance points and the corresponding interlayer dielectric transition coefficients, the vertical projection deviation of coordinates in different wall layers is calculated to obtain the interlayer projection deviation value of coordinates. The coordinates of all critical avoidance points are corrected based on the inter-coordinate projection deviation value, and the boundary extension redundancy is calculated by combining the confidence interval of regional dielectric fluctuation. Based on the corrected coordinates of the critical avoidance points and the boundary expansion redundancy, an initial construction safety boundary is generated. Calculate the minimum distance from all critical avoidance points within the initial construction safety boundary to the boundary to obtain the boundary coverage verification value.

10. The intelligent marking system for wood detection using a wall detector according to claim 9, characterized in that, The spatial coordinates of all critical avoidance points are analyzed to generate and mark construction safety boundaries, and this also includes: When the boundary coverage verification value meets the preset requirements, the final construction safety boundary is generated; Based on the confidence level of the woody characteristics of the key avoidance points, all key avoidance points are differentiated and labeled.

Images