A Smart Inspection Method for Bridge Segment Assembly Construction Quality

Abstract

This invention relates to the field of materials testing technology and proposes an intelligent detection method for the construction quality of bridge segment assembly. The method includes: acquiring ultrasonic signal sequences and transmitted signal sequences; obtaining feature values ​​of the ultrasonic signal sequences and transmitted signal sequences at the current moment based on the differences between the ultrasonic reflection amplitude and the ultrasonic signal value; obtaining ultrasonic crack correlation features corresponding to the current moment based on the differences between the feature values ​​of the ultrasonic signal sequences and transmitted signal sequences at the current moment and the feature values ​​at surrounding moments; converting the ultrasonic crack correlation features at all moments into an ultrasonic amplitude change feature image; clustering the ultrasonic amplitude change feature image to obtain clusters and identifying abnormal regions; and judging the bridge assembly construction quality based on the number of pixels in the abnormal regions. This invention avoids the drawback of poor expressiveness for bridge assembly construction quality detection in one-dimensional ultrasonic data analysis.

Description

Technical Field

[0001] This invention relates to the field of materials testing technology, specifically to an intelligent testing method for the construction quality of bridge segment assembly. Background Technology

[0002] Segmental splicing is a construction method used in large-span bridge structures, offering advantages such as fast construction speed and high reliability. Segmental splicing is divided into dry splicing and wet splicing, with wet splicing being more commonly used. Wet splicing involves connecting segments by erecting formwork and pouring concrete after the segments are installed. Dry splicing, on the other hand, involves applying a splicing adhesive to the contact surfaces during the segment splicing process. Compared to wet splicing, dry splicing offers significant advantages in terms of construction speed and convenience. However, when using dry splicing, the joint area between adjacent segments is discontinuous, allowing moisture and acidic components in the atmosphere to easily seep in and damage the joint. Increased traffic volume or the passage of heavy trucks can easily cause cracking at the joint. Under the erosion of rainwater, the cracks continue to extend, affecting the connection strength between bridge segments. When the cracks reach a certain extent, it can lead to bridge fracture.

[0003] Therefore, it is crucial to conduct timely inspections of the splicing construction quality of the assembled bridge. Rapid testing of the construction quality after splicing is essential, along with prompt detection and remediation of any cracks that appear during the splicing process, to prevent safety accidents caused by bridge defects. This paper focuses on optimizing and improving the ultrasonic non-destructive testing process during bridge splicing, addressing the issue of unclear ultrasonic characteristics of bridge splicing cracks leading to significant errors. Summary of the Invention

[0004] This invention provides an intelligent detection method for the construction quality of bridge segment assembly, to solve the problem that the ultrasonic characteristics of bridge splice cracks are not obvious. The specific technical solution adopted is as follows:

[0005] One embodiment of the present invention discloses an intelligent detection method for the construction quality of bridge segment assembly, the method comprising the following steps:

[0006] Acquire ultrasonic signal sequences and transmit signal sequences;

[0007] For an ultrasonic signal sequence, the characteristic values ​​of the ultrasonic signal sequence at the current moment are obtained based on the difference between the ultrasonic reflection amplitude at the current moment and the reflection amplitude of the surrounding adjacent small ultrasonic waves, as well as the difference between the ultrasonic reflection amplitude at the current moment and the maximum ultrasonic reflection amplitude; for a transmitted signal sequence, the characteristic values ​​of the transmitted signal sequence at the current moment are obtained.

[0008] The ultrasonic crack correlation features at the current moment are obtained based on the differences between the feature values ​​of the current moment and those of the surrounding moments in the ultrasonic signal sequence, as well as the differences between the feature values ​​of the current moment and those of the surrounding moments in the transmitted signal sequence.

[0009] The ultrasonic crack correlation features at all times are transformed into ultrasonic amplitude change feature images; clustering is performed on the ultrasonic amplitude change feature images to obtain clusters; the average gray value of each cluster is calculated; and the cluster corresponding to the maximum average gray value is taken as an abnormal region.

[0010] The quality of bridge splicing construction can be judged based on the number of pixels in abnormal areas.

[0011] Preferably, the method for acquiring the ultrasonic signal sequence and the transmitted signal sequence is as follows:

[0012] An ultrasonic probe transmits ultrasonic signal amplitudes from the transmitting end, obtaining one ultrasonic signal amplitude at each moment. All ultrasonic signal amplitudes from the initial moment to the current moment constitute the transmitted signal sequence. When the transmitted ultrasonic signal amplitudes are transmitted to the receiving end, they become ultrasonic reflection amplitudes, obtaining one ultrasonic reflection amplitude at each moment. All ultrasonic reflection amplitudes from the initial moment to the current moment constitute the ultrasonic signal sequence.

[0013] Preferably, the method for obtaining the characteristic values ​​of the ultrasonic signal sequence at the current moment based on the difference between the ultrasonic reflection amplitude at the current moment and the reflection amplitudes of neighboring ultrasonic waves, and the difference between the ultrasonic reflection amplitude at the current moment and the maximum ultrasonic reflection amplitude, is as follows:

[0014] For an ultrasonic data slice obtained at the current moment, the difference between the ultrasonic reflection amplitude at the current moment and the ultrasonic reflection amplitude at the time in the ultrasonic data slice is recorded as the first difference. The maximum ultrasonic reflection amplitude in the ultrasonic signal sequence is obtained, and the difference between the maximum ultrasonic reflection amplitude and the ultrasonic reflection amplitude at the current moment is obtained as the second difference. The feature value at the current moment is obtained by summing the first difference and multiplying it by the second difference.

[0015] Preferably, the method for obtaining an ultrasonic data slice at the current moment is as follows:

[0016] Given a preset number, an ultrasonic data slice is formed from a preset number of times before the current time. The number of ultrasonic data slices is the preset number, and the values ​​in the ultrasonic data slices are the values ​​corresponding to the corresponding sequences.

[0017] Preferably, the method for obtaining the feature value at the current time based on the product of the sum of the first difference and the second difference is as follows:

[0018] Take the current time as the i-th time.

[0019]

[0020] In the formula, s i Max represents the ultrasonic wave reflection amplitude at time i. i The maximum ultrasonic reflection amplitude in the ultrasonic signal sequence is represented by s. i,k Z represents the ultrasonic reflection amplitude at time k corresponding to the ultrasonic data slice at time i. i v represents the number of times within the ultrasonic data slice corresponding to the i-th time point. i The characteristic value at time i represents the characteristic value of the ultrasonic signal sequence, |s i,k -s i | represents the first difference, |Max i -s i | indicates the second difference.

[0021] Preferably, the method for obtaining the ultrasonic crack correlation features corresponding to the current moment based on the difference between the feature values ​​of the current moment of the ultrasonic signal sequence and the feature values ​​of the surrounding moments, and the difference between the feature values ​​of the current moment of the transmitted signal sequence and the feature values ​​of the surrounding moments, is as follows:

[0022] Take the current time as the i-th time.

[0023]

[0024] In the formula, Max0 represents the maximum amplitude of the ultrasonic signal in the transmitted signal sequence, Max1 represents the maximum amplitude of the ultrasonic reflection in the ultrasonic signal sequence, and Z... i v represents the number of times within the ultrasonic data slice corresponding to the i-th time point. i,0 v represents the characteristic value at time i in the transmitted signal sequence. i,1 v represents the characteristic value at time i in the ultrasonic signal sequence. i,k,0 v represents the feature value at time k of the ultrasonic data slice corresponding to time i in the transmitted signal sequence. i,k,1 h represents the feature value at time k of the ultrasonic data slice corresponding to time i in the ultrasonic signal sequence. i This represents the ultrasonic crack correlation feature corresponding to the i-th time moment.

[0025] Preferably, the method for converting the ultrasonic crack correlation features at all times into an ultrasonic amplitude variation feature image is as follows:

[0026] The ultrasonic crack correlation features calculated at all times are sorted in chronological order to obtain a sequence, which is denoted as the bridge splicing change feature sequence. The bridge splicing change feature sequence is converted into an ultrasonic amplitude change feature image using Gram angle field. The gray value of the ultrasonic amplitude change feature image is the ultrasonic crack correlation feature.

[0027] Preferably, the method for clustering the ultrasonic amplitude variation feature image to obtain clusters is as follows:

[0028] Given a preset cutoff radius, a cutoff circle is obtained with each pixel as the center. Confidence pixels are obtained based on the cutoff circle. Local density distances are obtained based on the distance difference between the confidence pixels and each pixel. Cluster centers are determined based on the local density distances. The number of preset clusters is given. The clustering condition is the difference in grayscale values ​​between the cluster centers and the pixels. Based on this, K-means clustering is used to obtain two clusters.

[0029] Preferably, the method for obtaining confidence pixels based on the cutoff circle is as follows:

[0030] Each pixel is designated as the center point. A confidence interval is obtained with the center point as the center. Pixels within the confidence interval 3σ are designated as trust pixels of the center point. Trust pixels within the cutoff circle of the center point are designated as confidence pixels.

[0031] Preferably, the method for obtaining the local density distance based on the confidence pixel and the distance difference between each pixel, and determining the cluster center based on the local density distance, is as follows:

[0032]

[0033] In the formula, dc represents the cutoff radius, and d p,q The distance between the p-th center point and the q-th pixel within the circle of that center point is represented by M, where M represents the maximum gray value in the ultrasonic amplitude variation feature image. p Let represent the gray value of the p-th center point, and exp() represent an exponential function with base n to the natural constant. p dt represents the number of confidence pixels centered at the p-th point. p This represents the local density distance of the p-th center point;

[0034] After obtaining the local density distance of each pixel, calculate the mean of the local density distance of all pixels, count the pixels with a local density distance greater than the mean, calculate the gray level difference between each pair of all the pixels obtained, and take the two pixels with the largest gray level difference as the two cluster centers.

[0035] The beneficial effects of this invention are as follows: This invention combines the characteristics of ultrasonic wave signal waveform changes during bridge splicing construction to construct and calculate prominent feature values ​​of ultrasonic changes in bridge splicing. This highlights and characterizes abnormal ultrasonic changes during bridge splicing quality inspection, facilitating subsequent detection and analysis of cracks and defects during bridge splicing. Simultaneously, this invention combines the attenuation changes of ultrasonic amplitude at the transmitting and receiving ends to calculate and construct corresponding ultrasonic crack correlation features for bridge splicing. Based on these features, a feature image of ultrasonic amplitude changes in bridge splicing is constructed, transforming one-dimensional ultrasonic data features into a two-dimensional space. This avoids the shortcomings of traditional one-dimensional ultrasonic data analysis, which has poor expressiveness for bridge splicing construction quality inspection. Furthermore, based on the data distribution characteristics of this feature image of ultrasonic amplitude changes in bridge splicing, appropriate cluster centers are selected to obtain abnormal change feature regions, completing the detection of bridge splicing quality. This avoids the impact of improper cluster center selection in traditional clustering algorithms, which leads to large errors in subsequent bridge splicing construction quality inspection. Attached Figure Description

[0036] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0037] Figure 1 This is a flowchart illustrating an intelligent detection method for bridge segment assembly construction quality according to an embodiment of the present invention.

[0038] Figure 2 A schematic diagram showing the setup of an ultrasonic probe. Detailed Implementation

[0039] 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.

[0040] Please see Figure 1 The diagram illustrates a flowchart of an intelligent inspection method for bridge segment assembly construction quality according to an embodiment of the present invention. The method includes the following steps:

[0041] Step S001: Use an ultrasonic probe to acquire signal sequences, obtaining the transmitted signal sequence and the ultrasonic signal sequence.

[0042] Ultrasonic waves have the advantages of strong penetrating power, ease of use, and minimal damage to the materials being tested. Therefore, appropriate ultrasonic detectors are used to inspect the construction quality of bridge splicing areas. Ultrasonic probes are installed in the bridge splicing construction area, with one probe placed at each end of the bridge. Figure 2 As shown, Figure 2 The square in the image represents the ultrasonic probe. When the ultrasonic wave emitted by the transmitter encounters a cracked area, the uneven distribution of the filling material within the crack hinders the propagation of the ultrasonic wave, resulting in a slight attenuation of the ultrasonic signal received by the receiver. This abnormal waveform change in the ultrasonic wave can be utilized to assist in the non-destructive testing of the construction quality of bridge splicing areas.

[0043] An ultrasonic probe can obtain a signal sequence at both the transmitting and receiving ends. Each data point in the signal sequence is acquired once every 1 second. To avoid the influence of measurement errors from the acquisition equipment, Gaussian filtering is used to denoise the ultrasonic probe signal sequence received at the receiving end. This aims to minimize or even eliminate the impact of random noise errors generated during the acquisition process on the accuracy of subsequent bridge splicing construction quality inspection. The resulting denoised signal sequence is denoted as the ultrasonic signal sequence.

[0044] The signal sequence obtained at the transmitting end is denoted as the transmitted signal sequence;

[0045] Thus, an ultrasonic signal sequence and a transmitted signal sequence were obtained.

[0046] Step S002: Obtain the feature values ​​of the ultrasonic signal sequence and the transmitted signal sequence at the current moment, and obtain the ultrasonic crack correlation feature corresponding to the current moment based on the difference between the feature values ​​of the two signal sequences and the feature moments of the surrounding moments.

[0047] For the acquired ultrasonic signal sequence, during the transmission of ultrasonic waves along the bridge medium, the ultrasonic data signal obtained by the receiving end is attenuated to a certain extent due to the obstruction of the abnormal crack area at the bridge construction splicing position. At this time, the ultrasonic reflection amplitude obtained by the receiving end will be weakened to a certain extent, indicating that there is a high probability of crack defects in the corresponding detection area.

[0048] In an ultrasonic signal sequence, each moment corresponds to an ultrasonic reflection amplitude. For the current moment, Z moments backward are selected to form an ultrasonic data slice. In this embodiment, the length of the ultrasonic data slice is 5. The maximum ultrasonic reflection amplitude is obtained from the ultrasonic signal sequence. The current moment of the ultrasonic signal sequence is recorded as the i-th moment. Based on the ultrasonic reflection amplitude corresponding to the i-th moment and the ultrasonic reflection amplitude in the ultrasonic data slice, the characteristic value of the i-th moment is obtained, as shown in the following formula:

[0049]

[0050] In the formula, s i Max represents the ultrasonic wave reflection amplitude at time i. i The maximum ultrasonic reflection amplitude in the ultrasonic signal sequence is represented by s. i,k Z represents the ultrasonic reflection amplitude at time k corresponding to the ultrasonic data slice at time i. i v represents the number of times within the ultrasonic data slice corresponding to the i-th time point. i This represents the characteristic value at time i, which is the characteristic value of the ultrasonic signal sequence.

[0051] Similarly, for the transmitted signal sequence, the characteristic values ​​of the transmitted signal sequence are obtained using the same method, where each data point in the transmitted signal sequence is the amplitude of the ultrasonic signal.

[0052] The above formula can be used to calculate the characteristic values ​​of the two signal sequences. The greater the difference between the ultrasonic reflection amplitude at the i-th time point and the maximum ultrasonic reflection amplitude from the initial time point to that time point, and the greater the difference between the amplitude and the data at different times in the ultrasonic data slice, the more likely that the i-th time point is to detect the characteristic region of the bridge splice crack defect.

[0053] For bridge splicing construction areas, if there are no cracks, the amplitude of the ultrasonic signal transmitted by the transmitting end and the amplitude of the ultrasonic reflection received by the receiving end should have certain similarity and correlation characteristics. The ultrasonic crack correlation characteristics at each moment are obtained based on the difference in characteristic values ​​between the ultrasonic signal sequence and the transmitted signal sequence at the same instant, as shown in the following formula:

[0054]

[0055] In the formula, Max0 represents the maximum amplitude of the ultrasonic signal in the transmitted signal sequence, Max1 represents the maximum amplitude of the ultrasonic reflection in the ultrasonic signal sequence, and Z... i v represents the number of times within the ultrasonic data slice corresponding to the i-th time point. i,0 v represents the characteristic value at time i in the transmitted signal sequence.i,1 v represents the characteristic value at time i in the ultrasonic signal sequence. i,k,0 v represents the feature value at time k of the ultrasonic data slice corresponding to time i in the transmitted signal sequence. i,k,1 h represents the feature value at time k of the ultrasonic data slice corresponding to time i in the ultrasonic signal sequence. i This represents the ultrasonic crack correlation feature corresponding to the i-th time moment.

[0056] The ultrasonic crack correlation feature h of bridge splicing can be calculated and obtained using the above formula. i The magnitude of the value indicates that when the difference between the ultrasonic data at the transmitting end and the data at the receiving end is small, it means that the amplitude of the ultrasonic data at the transmitting end and the amplitude of the ultrasonic reflection at the receiving end are relatively close, and there are no abnormal defects such as cracks. Conversely, when the calculated ultrasonic crack correlation characteristic h of the bridge splice is large, it indicates that there are no abnormal defects. i When the value is large, it is believed that the amplitude of the ultrasonic data at the transmitting and receiving ends has been significantly attenuated, and there is a high probability that cracks or abnormal defects will appear in the bridge splicing construction area to be inspected.

[0057] Thus, the ultrasonic crack correlation features corresponding to the current moment have been obtained.

[0058] Step S003: Convert the ultrasonic crack correlation features at all times into ultrasonic amplitude variation images, and obtain the heritage area based on the ultrasonic amplitude variation images.

[0059] To further obtain the internal temporal variation characteristics of the ultrasonic data features of the bridge splicing area at different times and highlight the abnormal variation characteristics of the ultrasonic data, all ultrasonic crack correlation features from the initial time to each time are used to construct the bridge splicing variation feature sequence corresponding to each time. This bridge splicing variation feature sequence is used as input and Gram angle field is used to convert it into an ultrasonic amplitude variation feature image. The specific calculation method of Gram angle field is a well-known technique and will not be described in detail here.

[0060] In the bridge splicing variation feature sequence, points with larger values ​​are more likely to represent crack defects in the corresponding bridge splicing area. In this case, the values ​​at the corresponding locations in the ultrasonic amplitude variation feature image obtained through Gram's angle field will show significant differences compared to the surrounding values. By extending the one-dimensional ultrasonic amplitude data into a two-dimensional space using Gram's angle field, the abnormal changes in crack defects during bridge splicing construction are highlighted. Therefore, further calculations are performed on the abnormal values ​​in the ultrasonic amplitude variation feature image. It is worth noting that the grayscale values ​​in the ultrasonic amplitude variation feature image are associated with ultrasonic crack features.

[0061] K-Means clustering was used to analyze the characteristic images of ultrasonic amplitude changes. However, in the K-Means clustering algorithm, improper selection of cluster centers can lead to poor clustering results, affecting the accuracy of crack anomaly detection during bridge splicing construction. Therefore, it is necessary to select appropriate cluster centers.

[0062] First, in this embodiment, the number of clusters is given as 2. For the gray value distribution of the ultrasonic amplitude change feature image, taking any pixel as the center, pixels within its confidence interval of 3σ are taken as its trust pixels. Trust pixels within the cutoff circle of the center point are recorded as trust pixels. For each pixel, it is recorded as the center point. A circle is obtained with the center point as the center, and the radius of the circle is the cutoff radius. In this embodiment, the size of the cutoff radius is 5. The Euclidean distance between the center point and the pixels inside the circle is calculated. Based on this, the local density distance of the center point is obtained, and the formula is as follows:

[0063]

[0064] In the formula, dc represents the cutoff radius, and d p,q The distance between the p-th center point and the q-th pixel within the circle of that center point is represented by M, where M represents the maximum gray value in the ultrasonic amplitude variation feature image. p Let represent the gray value of the p-th center point, and exp() represent an exponential function with base n to the natural constant. p dt represents the number of confidence pixels centered at the p-th point. p This represents the local density distance of the p-th center point.

[0065] The larger the value of the local density distance, the more similar abnormal pixels there are around the center point, and the greater the probability that the p-th center point in the ultrasonic amplitude change feature image is the cluster center.

[0066] After calculating the local density distance of all pixels, the larger the local density distance, the more likely it is to be a cluster center. In addition, the difference in gray values ​​of the cluster centers in the image should be as large as possible. Therefore, the mean of the local density distance of all pixels is calculated, and the pixels with a local density distance greater than the mean are counted and recorded as preferred pixels. The gray value difference between each pair of preferred pixels is calculated, and the two pixels corresponding to the largest gray value difference are extracted and used as the cluster center points.

[0067] After obtaining the cluster centers, the number of clusters is 2. The clustering condition is the difference in gray value between the cluster center point and the other pixels. The clusters with smaller gray value differences are divided into one group, and the cluster with the larger average gray value between the two clusters is recorded as an abnormal region.

[0068] Step S004: Determine the quality of bridge splicing construction based on the number of pixels in the abnormal area.

[0069] After identifying the abnormal region, the pixels within that region are called abnormal pixels. The proportion of abnormal pixels to the total number of pixels determines the quality anomalies in bridge splicing construction, as shown in the following formula:

[0070]

[0071] In the formula, nd is the number of abnormal pixels in the ultrasonic amplitude change feature image, N is the total number of pixels in the ultrasonic amplitude change feature image, and Q is the abnormal value of bridge splicing construction quality.

[0072] Given an anomaly threshold, in this embodiment the anomaly threshold is 0.5. When the anomaly value of the bridge splicing construction quality is greater than the anomaly threshold, it is considered that there is a serious crack defect in the bridge splicing construction quality, and further processing is required for the splicing area.

[0073] 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 within the protection scope of the present invention.

Claims

1. A method for intelligent inspection of the construction quality of bridge segment assembly, characterized in that, The method includes the following steps: Acquire ultrasonic signal sequences and transmit signal sequences; For an ultrasonic signal sequence, the characteristic values ​​of the ultrasonic signal sequence at the current moment are obtained based on the difference between the ultrasonic reflection amplitude at the current moment and the ultrasonic reflection amplitude of the surrounding adjacent ultrasonic waves, as well as the difference between the ultrasonic reflection amplitude at the current moment and the maximum ultrasonic reflection amplitude; for a transmitted signal sequence, the characteristic values ​​of the transmitted signal sequence at the current moment are obtained. The ultrasonic crack correlation features at the current moment are obtained based on the differences between the feature values ​​of the current moment and those of the surrounding moments in the ultrasonic signal sequence, as well as the differences between the feature values ​​of the current moment and those of the surrounding moments in the transmitted signal sequence. The ultrasonic crack correlation features at all times are transformed into ultrasonic amplitude change feature images; clustering is performed on the ultrasonic amplitude change feature images to obtain clusters; the average gray value of each cluster is calculated; and the cluster corresponding to the maximum average gray value is taken as an abnormal region. The quality of bridge splicing construction can be judged based on the number of pixels in abnormal areas.

2. The intelligent detection method for bridge segment assembly construction quality according to claim 1, characterized in that, The method for acquiring the ultrasonic signal sequence and the transmitted signal sequence is as follows: An ultrasonic probe transmits ultrasonic signal amplitudes from the transmitting end, obtaining one ultrasonic signal amplitude at each moment. All ultrasonic signal amplitudes from the initial moment to the current moment constitute the transmitted signal sequence. When the transmitted ultrasonic signal amplitudes are transmitted to the receiving end, they become ultrasonic reflection amplitudes, obtaining one ultrasonic reflection amplitude at each moment. All ultrasonic reflection amplitudes from the initial moment to the current moment constitute the ultrasonic signal sequence.

3. The intelligent detection method for bridge segment assembly construction quality according to claim 1, characterized in that, The method for obtaining the characteristic values ​​of the ultrasonic signal sequence at the current moment based on the differences between the ultrasonic reflection amplitude at the current moment and the ultrasonic reflection amplitudes of surrounding neighbors, as well as the differences between the ultrasonic reflection amplitude at the current moment and the maximum ultrasonic reflection amplitude, is as follows: For an ultrasonic data slice obtained at the current moment, the difference between the ultrasonic reflection amplitude at the current moment and the ultrasonic reflection amplitude at the time in the ultrasonic data slice is recorded as the first difference. The maximum ultrasonic reflection amplitude in the ultrasonic signal sequence is obtained, and the difference between the maximum ultrasonic reflection amplitude and the ultrasonic reflection amplitude at the current moment is obtained as the second difference. The feature value at the current moment is obtained by summing the first difference and multiplying it by the second difference.

4. The intelligent detection method for bridge segment assembly construction quality according to claim 3, characterized in that, The method for obtaining an ultrasonic data slice at the current moment is as follows: Given a preset number, an ultrasonic data slice is formed from a preset number of times before the current time. The number of ultrasonic data slices is the preset number, and the values ​​in the ultrasonic data slices are the values ​​corresponding to the corresponding sequences.

5. The intelligent detection method for bridge segment assembly construction quality according to claim 3, characterized in that, The method for obtaining the feature value at the current time based on the product of the sum of the first difference and the second difference is as follows: Take the current time as the i-th time. In the formula, s i Max represents the ultrasonic wave reflection amplitude at time i. i The maximum ultrasonic reflection amplitude in the ultrasonic signal sequence is represented by s. i,k Z represents the ultrasonic reflection amplitude at time k corresponding to the ultrasonic data slice at time i. i v represents the number of times within the ultrasonic data slice corresponding to the i-th time point. i The characteristic value at time i represents the characteristic value of the ultrasonic signal sequence, |s i,k -s i | represents the first difference, |Max i -s i | indicates the second difference.

6. The intelligent detection method for bridge segment assembly construction quality according to claim 1, characterized in that, The method for obtaining the ultrasonic crack correlation features corresponding to the current moment based on the difference between the feature values ​​of the current moment of the ultrasonic signal sequence and the feature values ​​of the surrounding moments, and the difference between the feature values ​​of the current moment of the transmitted signal sequence and the surrounding moments, is as follows: Take the current time as the i-th time. In the formula, Max0 represents the maximum amplitude of the ultrasonic signal in the transmitted signal sequence, Max1 represents the maximum amplitude of the ultrasonic reflection in the ultrasonic signal sequence, and Z... i v represents the number of times within the ultrasonic data slice corresponding to the i-th time point. i,0 v represents the characteristic value at time i in the transmitted signal sequence. i,1 v represents the characteristic value at time i in the ultrasonic signal sequence. i,k,0 v represents the feature value at time k of the ultrasonic data slice corresponding to time i in the transmitted signal sequence. i,k,1 h represents the feature value at time k of the ultrasonic data slice corresponding to time i in the ultrasonic signal sequence. i This represents the ultrasonic crack correlation feature corresponding to the i-th time moment.

7. The intelligent detection method for bridge segment assembly construction quality according to claim 1, characterized in that, The method for converting the ultrasonic crack correlation features at all times into an ultrasonic amplitude change feature image is as follows: The ultrasonic crack correlation features calculated at all times are sorted in chronological order to obtain a sequence, which is denoted as the bridge splicing change feature sequence. The bridge splicing change feature sequence is converted into an ultrasonic amplitude change feature image using Gram angle field. The gray value of the ultrasonic amplitude change feature image is the ultrasonic crack correlation feature.

8. The intelligent detection method for bridge segment assembly construction quality according to claim 1, characterized in that, The method for clustering ultrasonic amplitude variation feature images to obtain clusters is as follows: Given a preset cutoff radius, a cutoff circle is obtained with each pixel as the center. Confidence pixels are obtained based on the cutoff circle. Local density distances are obtained based on the distance difference between the confidence pixels and each pixel. Cluster centers are determined based on the local density distances. The number of preset clusters is given. The clustering condition is the difference in grayscale values ​​between the cluster centers and the pixels. Based on this, K-means clustering is used to obtain two clusters.

9. The intelligent detection method for bridge segment assembly construction quality according to claim 8, characterized in that, The method for obtaining confidence pixels based on the cutoff circle is as follows: Each pixel is designated as the center point. A confidence interval is obtained with the center point as the center. Pixels within the confidence interval 3σ are designated as trust pixels of the center point. Trust pixels within the cutoff circle of the center point are designated as confidence pixels.

10. The intelligent detection method for bridge segment assembly construction quality according to claim 8, characterized in that, The method for obtaining local density distances based on confidence pixels and the distance differences between each pixel, and then determining cluster centers based on these local density distances, is as follows: In the formula, dc represents the cutoff radius, and d p,q The distance between the p-th center point and the q-th pixel within the circle of that center point is represented by M, where M represents the maximum gray value in the ultrasonic amplitude variation feature image. p Let represent the gray value of the p-th center point, and exp() represent an exponential function with base n to the natural constant. p dt represents the number of confidence pixels centered at the p-th point. p This represents the local density distance of the p-th center point; After obtaining the local density distance of each pixel, calculate the mean of the local density distance of all pixels, count the pixels with a local density distance greater than the mean, calculate the gray level difference between each pair of all the pixels obtained, and take the two pixels with the largest gray level difference as the two cluster centers.

Images