Intelligent recognition and evaluation system for construction quality of bridge concrete structure

Through the on-site data acquisition subsystem and cloud computing subsystem combined with artificial intelligence identification and evaluation algorithm, the problem of fewer types and low efficiency in the appearance quality evaluation of bridge concrete is solved, and the accurate identification and rapid evaluation of multiple defects is achieved, which improves the Evaluate efficiency and quality.

CN119027802APending Publication Date: 2024-11-26CCCC HIGHWAY BRIDGES NATIONAL ENGINEERING RESEARCH CENTRE CO LTD +2
View PDF 4 Cites 0 Cited by

Patent Information

Application Number
CN202410916360.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-09
Publication Date
2024-11-26

AI Technical Summary

Technical Problem

The prior art has few types of identification, low efficiency, and inconsistent evaluation models in the evaluation of bridge concrete appearance quality, making it difficult to meet the needs of fast and accurate.

Method used

The on-site data acquisition subsystem and cloud computing subsystem are adopted, combined with mobile mobile terminals, fixed-point cameras and drones for data acquisition, and an artificial intelligence identification and evaluation algorithm based on semantic segmentation networks are used to intelligently identify concrete appearance quality, calculate and evaluate defect sizes.

Benefits of technology

It realizes accurate identification and rapid evaluation of various concrete appearance quality defects, improves evaluation efficiency and quality, and can automatically identify defects such as honeycomb lint surfaces, peeling and angle loss, hollow holes, exposed rib corrosion, cracks, and color difference of sand lines, providing Efficient and accurate concrete appearance quality assessment.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119027802A_ABST
    Figure CN119027802A_ABST
Patent Text Reader

Abstract

The invention discloses an intelligent recognition and evaluation system for construction quality of a bridge concrete structure. The intelligent recognition and evaluation system comprises a field data acquisition subsystem and a cloud computing subsystem, for different bridge components, different devices are comprehensively used for appearance image acquisition; concrete appearance defects are identified by using a semantic segmentation algorithm, defect size features are calculated, and a concrete appearance quality condition evaluation model suitable for image data is provided; an artificial intelligence identification evaluation algorithm is deployed to a cloud server, and efficient, accurate and rapid evaluation of the bridge concrete appearance quality is realized.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the technical field of concrete quality evaluation, and in particular to an intelligent identification and evaluation system for the construction quality of a bridge concrete structure. Background Art

[0002] Concrete is a common structural material in bridge engineering, and the appearance quality of concrete is directly related to the aesthetics, durability and safety of the bridge. With the improvement of the quality requirements of bridge construction, the requirements for the appearance quality of concrete are becoming more and more stringent. It is necessary to timely evaluate and monitor its appearance quality to ensure the safety and stability of the bridge. Traditional concrete appearance quality evaluation often requires a lot of manpower and time costs, and the evaluation results are easily affected by subjective factors, which is difficult to meet the rapid and accurate needs of bridge construction and maintenance. With the development of computer vision technology, concrete appearance quality detection based on image processing technology has been applied. Patent document CN112611754A uses a single gray value background plate and traditional image processing technology to extract appearance quality defects; Patent document CN111681289A uses traditional image processing technology to detect and analyze color difference problems; Patent document CN111311668A uses convolutional neural network to identify and calibrate pore problems; Patent document CN111462039A uses traditional image processing technology to extract color difference problems. However, the existing technology mainly focuses on the appearance quality inspection of plain concrete and the evaluation of two quality defects such as color difference and pores; at the same time, the data collection method and calculation processing medium limit the efficiency of appearance quality evaluation. Therefore, a method and system for rapid evaluation of concrete appearance quality is urgently needed to solve the above problems. Summary of the invention

[0003] In view of the above defects of the prior art, the present invention proposes an intelligent identification and evaluation system for the construction quality of bridge concrete structures to overcome the problems in the prior art of few types of appearance quality identification, low technical efficiency and inconsistent evaluation models.

[0004] An embodiment of the present invention discloses an intelligent identification and evaluation system for the construction quality of a bridge concrete structure, the system comprising a field data acquisition subsystem and a cloud computing subsystem; wherein the field data acquisition subsystem is used to collect appearance pictures of the concrete structure; the cloud computing subsystem intelligently evaluates the appearance quality of the concrete based on the collected appearance pictures of the concrete structure and an artificial intelligence identification and evaluation algorithm.

[0005] Furthermore, the field data acquisition subsystem includes a mobile phone terminal, a fixed-point camera subsystem and a drone acquisition subsystem.

[0006] Furthermore, the cloud computing subsystem includes a cloud server, an artificial intelligence recognition and evaluation algorithm, and a basic environment configuration for algorithm calling.

[0007] Furthermore, the artificial intelligence recognition and evaluation algorithm includes a defect recognition algorithm, a defect size calculation algorithm and an appearance quality evaluation algorithm; the concrete appearance quality intelligent evaluation includes the following steps:

[0008] S1: Intelligent identification of concrete appearance quality;

[0009] S2: Calculation of concrete appearance defect size;

[0010] S3: Assessment of concrete appearance quality.

[0011] Furthermore, in step S1, quality defects in the collected concrete structure appearance pictures are intelligently identified based on the defect recognition algorithm; the defect recognition algorithm includes the following steps:

[0012] S11: preprocessing of original concrete appearance image data;

[0013] S12: semantic segmentation and annotation of original data;

[0014] S13: Iteratively train the semantic segmentation network;

[0015] S14: Intelligent identification of concrete appearance defects.

[0016] Furthermore, in step S2, the identified appearance quality defect size information is calculated based on the defect size calculation algorithm; the appearance quality defect size information includes one or more of the cumulative defect area, the maximum area of ​​a single defect, the cumulative defect length, the maximum length of a single defect and the maximum width of the defect.

[0017] Furthermore, in step S3, the appearance quality of concrete is evaluated based on the appearance quality evaluation algorithm; the appearance quality evaluation algorithm includes the following steps:

[0018] S31: Determine the scale of different quality defects;

[0019] S32: Calculate the deduction value of each type of quality defect in the image;

[0020] S33: Calculate the score of each image;

[0021] S34: Calculate the component appearance quality score;

[0022] S35: Determine the appearance quality status grade of concrete.

[0023] Furthermore, the calculation formula for the score of each picture in step S33 is:

[0024]

[0025] Among them, S r is the overall appearance quality score, ADP k Deduct points for appearance quality, DDP k It is the deduction score for a single item of appearance quality. k and n represent the nth item among k quality defects in a picture.

[0026] Further, the step S34 is specifically as follows: according to the score of each picture, the mean of all picture scores of the component is calculated to obtain the component appearance quality score; according to the component appearance quality score, the concrete appearance quality status grade is determined: Category 1, the component appearance quality score ∈ [90, 100], the appearance quality is intact; Category 2, the component appearance quality score ∈ [80, 90), the appearance has slight defects, which have little impact on the use function; Category 3, the component appearance quality score ∈ [60, 80), the appearance has moderate defects, which have no significant impact on the use function, but repair is recommended; Category 4, the component appearance quality score ∈ [40, 60), the appearance has large defects, which affects the use function, or affects the bearing capacity, and repair and reinforcement are recommended; Category 5, the component appearance quality score ∈ [0, 40), the appearance has serious defects, which affects the use function and the bearing capacity, and it is strictly prohibited to use.

[0027] Furthermore, the cloud computing subsystem deploys the artificial intelligence recognition and evaluation algorithm to the cloud server and conducts business interaction through a network communication protocol.

[0028] In general, the above technical solutions conceived by the present invention can achieve the following beneficial effects compared with the prior art:

[0029] 1. For different bridge components, mobile terminals, fixed cameras, drones and other equipment are used to collect appearance images. The collection methods are comprehensive and diverse, which improves the quality and efficiency of concrete appearance quality assessment.

[0030] 2. The artificial intelligence algorithm based on the semantic segmentation network can automatically identify various concrete appearance quality defects such as honeycombed surface, peeling and chipping, voids and holes, exposed reinforcement corrosion, cracks, sand line color difference, etc. The recognition results are accurate and efficient; the appearance defect category and size information are extracted based on the semantic segmentation results.

[0031] 3. An artificial intelligence recognition and evaluation algorithm suitable for image data is proposed. The algorithm is deployed in the cloud to form an intelligent recognition and evaluation system for the construction quality of bridge concrete structures, which realizes automatic, rapid and accurate evaluation of the appearance quality of concrete. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 This is an overall architecture diagram of the intelligent identification and evaluation system for the construction quality of a bridge concrete structure according to an embodiment of the present invention;

[0033] Figure 2 The process of the intelligent evaluation method for concrete appearance quality according to the embodiment of the present invention is as follows: Figure 1 ;

[0034] Figure 3 A flow chart of a defect recognition algorithm according to an embodiment of the present invention;

[0035] Figure 4 It is a schematic diagram of a method for calculating length and width in a defect size calculation algorithm according to an embodiment of the present invention;

[0036] Figure 5 is a flow chart of an appearance quality assessment algorithm according to an embodiment of the present invention;

[0037] Figure 6 The process of the intelligent assessment method for concrete appearance quality according to the embodiment of the present invention is as follows: Figure 2 . DETAILED DESCRIPTION

[0038] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0039] like Figure 1 As shown, an embodiment of the present invention provides a bridge concrete structure construction quality intelligent identification and evaluation system, the system comprising: a field data acquisition subsystem and a cloud computing subsystem. The field data acquisition subsystem is used to collect concrete structure appearance pictures; the cloud computing subsystem implements intelligent evaluation of concrete appearance quality based on the collected concrete structure appearance pictures and artificial intelligence identification and evaluation algorithms.

[0040] Specifically, the field data acquisition subsystem includes equipment such as a mobile phone terminal, a fixed-point camera subsystem and a drone acquisition subsystem.

[0041] In this embodiment, for different bridge components, mobile terminals, fixed cameras, drones and other devices are used in combination to collect appearance images. For example, mobile phones can be used to collect appearance images of components such as low piers and caps at the construction site; fixed cameras can be arranged on the production line to collect appearance images of prefabricated beams and prefabricated panels; drones can be used to collect appearance images of components such as high piers and cable towers.

[0042] The data collection method adopted in the embodiment of the present invention is comprehensive and diverse. Corresponding collection equipment and methods are used for different bridge components, so that the collected concrete structure appearance pictures are of higher quality, which is conducive to the intelligent identification of defects in the pictures and the intelligent evaluation of the concrete appearance quality.

[0043] Furthermore, the cloud computing subsystem includes a cloud server, an artificial intelligence recognition and evaluation algorithm, and a basic environment configuration for algorithm calling. The cloud computing subsystem deploys the artificial intelligence recognition and evaluation algorithm to the cloud server and conducts business interaction through a network communication protocol.

[0044] Furthermore, the network communication protocol includes but is not limited to the MQTT protocol and the HTTP protocol.

[0045] The functions implemented by the cloud computing subsystem described in the embodiment of the present invention include: accepting input bridge concrete result appearance pictures, performing intelligent defect recognition, extracting defect areas, calculating defect sizes, outputting bridge concrete structure appearance quality scores, and realizing intelligent assessment of concrete appearance quality.

[0046] Specifically, the artificial intelligence recognition and evaluation algorithm includes a defect recognition algorithm, a defect size calculation algorithm and an appearance quality evaluation algorithm. Figure 2 As shown in the figure, intelligent assessment of concrete appearance quality includes the following steps:

[0047] S1: Intelligent identification of concrete appearance quality;

[0048] S2: Calculation of concrete appearance defect size;

[0049] S3: Assessment of concrete appearance quality.

[0050] Specifically, in step S1, quality defects in the collected concrete structure appearance pictures are intelligently identified based on the defect recognition algorithm. The quality defects include but are not limited to honeycomb surface, peeling corners, voids, exposed steel bars, rust, cracks, sand line color difference, etc. Specifically, Figure 3 As shown, the defect recognition algorithm includes the following steps:

[0051] S11: Preprocessing of original concrete appearance image data.

[0052] First, the original data is preprocessed by cropping, denoising and other operations.

[0053] S12: Perform semantic segmentation and annotation on the original data.

[0054] Use the labelme annotation tool to perform semantic segmentation annotation on the original data; after completing the annotation work, check the annotation accuracy.

[0055] Semantic segmentation is a research direction in computer vision. Semantic segmentation is a classification at the pixel level. Pixels belonging to the same category are classified into one category and given a color label. The semantic segmentation method can be used to understand the image at the pixel level and identify the area of ​​all pixels of a specific target in the image, that is, to extract the target area in the image. Since the appearance quality assessment of concrete requires the type and size of appearance defects, the present invention uses semantic segmentation technology as an intelligent recognition algorithm for concrete appearance defects.

[0056] S13: Iteratively train the semantic segmentation network.

[0057] The dataset is divided into training set, validation set and test set. The semantic segmentation network is iteratively trained on the training set.

[0058] S14: Intelligent identification of concrete appearance defects.

[0059] Specifically, the semantic segmentation network trained in step S13 is deployed to the cloud, the collected concrete appearance picture is input into the semantic segmentation network, and the result picture is output. Each type of appearance defect in the output result picture is set as a different color label to achieve intelligent recognition of concrete appearance defects.

[0060] Furthermore, in step S2, the relevant dimensions of the identified appearance quality defects are calculated based on the defect size calculation algorithm. Specifically, the quality defect recognition results are separated according to different color labels, and the area dimensions of surface problems such as pitting, honeycomb, and color difference are calculated; the length and width characteristics of linear defects such as cracks and exposed tendons are calculated.

[0061] Specifically, for the size calculation of surface quality defects, the polygon fitting algorithm is first used to extract the vertex coordinates A of a single defect area. 1 (x 1 ,y 1 ), A 2 (x 2 ,y 2 )……A n (x n ,y n ), and then the Shoelace Theorem formula is used to calculate the defect area.

[0062]

[0063] Where n is the number of vertices in the defect area using polygon fitting, and i is the index of the vertex; when i = n, x n+1 =x 1 ,y n+1 =y 1 .

[0064] For linear defects, edge detection and image thinning algorithms are first used to extract the edge and center lines of the problem area. Figure 4 As shown in the figure, for length feature calculation, the center line is divided into several small intervals, the distance between the start and end points of each interval is calculated, and the sum is used to get the length of the area. For width feature calculation, the normal equation at each point on the center line is calculated to get the two intersection points of the normal line and the edge line, and the distance between the intersection points is used as the width of the area.

[0065] The size of the concrete appearance quality defects in each picture is calculated, and finally the information of various quality defects in the picture is obtained, including: cumulative defect area, maximum area of ​​single defect, cumulative defect length, maximum length of single defect and maximum width of defect.

[0066] Further, in step S3, the appearance quality of the concrete is evaluated based on the appearance quality evaluation algorithm, such as Figure 5 As shown, the following steps are included:

[0067] S31: Determine the scale of different quality defects;

[0068] According to the appearance quality identification results and the dimensional feature calculation results, the scales of different quality defects are determined by comparing the control index scale and the evaluation standard comparison table. Table 1 shows an example of control indicators in a certain project. The specific control indicators can be dynamically adjusted according to the actual situation. Among them, the scale can be understood as the severity of the defect, and each degree is divided by a threshold. As shown in Table 1, the crack defects are divided into 5 types of scales according to the severity of the defects; the honeycomb surface is divided into 3 types of scales. The defect scale and thresholds at each level in the present invention are the result of continuous improvement and adjustment based on a large number of actual defect detection and evaluation results.

[0069] Table 1 Control index scale and evaluation standard comparison table

[0070]

[0071]

[0072] S32: Calculate the deduction value of each type of quality defect in the image;

[0073] After determining the scale of each type of quality defect according to step S31, further refer to Table 2 to calculate the deduction value of each type of quality defect in the image. Assume that when a crack defect is detected in an image, refer to Table 1, and calculate that the crack is in the 4th scale through indicators such as area, length, and width. Then, when calculating the deduction value of the crack defect in the image, first look at the highest scale value of the crack defect of 5, and then look at the scale value of the crack defect in the image of 4. According to Table 2, the deduction value of the crack defect in the image is 60.

[0074] Table 2 Comparison table of deduction values ​​for individual indicator grades

[0075]

[0076] S33: Calculate the score of each image;

[0077] After obtaining the deduction score for each type of quality defect according to step S32, the score of each image is calculated using the following formula.

[0078]

[0079] Among them, S r is the overall appearance quality score, ADP k Deduct points for appearance quality, DDP k It is the deduction score for a single item of appearance quality. k and n represent the nth item among k quality defects in a picture.

[0080] S34: Calculate the component appearance quality score;

[0081] After calculating the appearance quality score of each image, the average of all image scores of the component is calculated to obtain the component appearance quality score.

[0082] S35: Determine the appearance quality status grade of concrete.

[0083] After obtaining the component appearance quality score, determine the concrete appearance quality status grade by referring to Table 3 to complete the concrete appearance quality assessment task.

[0084] Table 3 Concrete appearance quality status grade comparison table

[0085]

[0086] like Figure 6 As shown, the cloud computing subsystem provided in the embodiment of the present invention deploys the artificial intelligence recognition and evaluation algorithm to the cloud server, calls the algorithm through the network end, and uses the pre-trained semantic segmentation model to identify the defects of concrete appearance quality, automatically identifies various concrete appearance quality defects such as honeycombed surface, peeling and chipping, voids, exposed reinforcement corrosion, cracks, sand line color difference, etc., and calculates the size characteristics of concrete appearance defects, and proposes a concrete appearance quality status evaluation model suitable for image data, so as to achieve efficient, accurate and rapid evaluation of concrete appearance quality.

[0087] In specific implementation, the algorithm proposed in the technical solution of the present invention can be automatically run by those skilled in the art using computer software technology. System devices that implement the algorithm, such as computer-readable storage media that store the corresponding computer program of the technical solution of the present invention and computer equipment that runs the corresponding computer program, should also be within the scope of protection of the present invention.

[0088] It will be easily understood by those skilled in the art that 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 and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. An intelligent identification and evaluation system for the construction quality of a bridge concrete structure, the system comprising a field data acquisition subsystem and a cloud computing subsystem; in, The field data acquisition subsystem is used to collect pictures of the appearance of the concrete structure; The cloud computing subsystem intelligently evaluates the appearance quality of concrete based on the collected concrete structure appearance pictures and artificial intelligence recognition and evaluation algorithms.

2. The intelligent identification and evaluation system for construction quality of bridge concrete structures according to claim 1, It is characterized in that The field data acquisition subsystem includes a mobile phone terminal, a fixed-point camera subsystem and a drone acquisition subsystem.

3. The intelligent identification and evaluation system for construction quality of bridge concrete structures according to claim 1, It is characterized in that The cloud computing subsystem includes a cloud server, an artificial intelligence recognition and evaluation algorithm, and a basic environment configuration for algorithm invocation.

4. The intelligent identification and evaluation system for construction quality of bridge concrete structures as claimed in claim 3, It is characterized in that The artificial intelligence recognition and evaluation algorithm includes a defect recognition algorithm, a defect size calculation algorithm and an appearance quality evaluation algorithm; the concrete appearance quality intelligent evaluation includes the following steps: S1: Intelligent identification of concrete appearance quality; S2: Calculation of concrete appearance defect size; S3: Assessment of concrete appearance quality.

5. The intelligent identification and evaluation system for construction quality of bridge concrete structures as claimed in claim 4, It is characterized in that In step S1, quality defects in the collected concrete structure appearance pictures are intelligently identified based on the defect identification algorithm; the defect identification algorithm includes the following steps: S11: preprocessing of original concrete appearance image data; S12: semantic segmentation and annotation of original data; S13: Iteratively train the semantic segmentation network; S14: Intelligent identification of concrete appearance defects.

6. The intelligent identification and evaluation system for construction quality of bridge concrete structures according to claim 5, It is characterized in that In step S2, the identified appearance quality defect size information is calculated based on the defect size calculation algorithm; the appearance quality defect size information includes one or more of the cumulative defect area, the maximum area of ​​a single defect, the cumulative defect length, the maximum length of a single defect and the maximum width of the defect.

7. The intelligent identification and evaluation system for construction quality of bridge concrete structures according to claim 6, It is characterized in that In step S3, the appearance quality of concrete is evaluated based on the appearance quality evaluation algorithm; the appearance quality evaluation algorithm includes the following steps: S31: Determine the scale of different quality defects; S32: Calculate the deduction value of each type of quality defect in the image; S33: Calculate the score of each image; S34: Calculate the component appearance quality score; S35: Determine the appearance quality status grade of concrete.

8. The intelligent identification and evaluation system for construction quality of bridge concrete structures according to claim 7, It is characterized in that The calculation formula for each picture score in step S33 is: Among them, S r is the overall appearance quality score, ADP k Deduct points for appearance quality, DDP k It is the deduction score for a single item of appearance quality. k and n represent the nth item among k quality defects in a picture.

9. The intelligent identification and evaluation system for construction quality of bridge concrete structures according to claim 8, It is characterized in that The step S34 is specifically to calculate the mean of all component picture scores according to the score of each picture to obtain the component appearance quality score; the step S35 is specifically to determine the concrete appearance quality status grade according to the component appearance quality score: Category 1: the appearance quality score of the component is ∈[90, 100], and the appearance quality is intact; Category 2: the component appearance quality score ∈[80,90), with slight defects in appearance and little impact on the use function; Category 3: the component appearance quality score ∈[60,80), the appearance has moderate defects, which have no significant impact on the use function, but repair is recommended; Category 4: the component appearance quality score ∈ [40, 60), the appearance has large defects, affecting the use function or bearing capacity, and it is recommended to repair or reinforce; Category 5: The appearance quality score of the component is ∈[0, 40). There are serious defects in the appearance, which affect the use function and bearing capacity, and it is strictly prohibited to use.

10. The intelligent identification and evaluation system for construction quality of bridge concrete structure according to claim 9, It is characterized in that The cloud computing subsystem deploys the artificial intelligence recognition and evaluation algorithm to the cloud server and conducts business interaction through the network communication protocol.

Citation Information

Patent Citations

  • Bare concrete appearance quality evaluation method

    CN112611754A

  • Road engineering precast beam concrete appearance quality evaluation method

    CN115565027A

  • Concrete surface defect detection and quantification method based on computer vision

    CN117372396A

  • Color difference analysis / detection method and system for fair-faced concrete

    WO2021248915A1