A method for engineering quality detection based on deep learning
Through deep learning-based methods, the lack of foundation and concrete detection in the existing technology is solved, and the comprehensiveness and reliability of construction project quality inspection is achieved, ensuring the accuracy and authenticity of the inspection results. .
Patent Information
- Application Number
- CN202411179607.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-27
- Publication Date
- 2024-11-22
- Estimated Expiration
- 2044-08-27
AI Technical Summary
The existing technology lacks effective testing of foundations and concrete in construction project quality inspection, resulting in the inability to accurately understand the quality of the project, affecting the reliability and accuracy of the inspection results.
Using deep learning-based methods, the foundation detection model and concrete detection model are trained, foundation and concrete data are obtained, settlement observation, bearing capacity testing and quality testing are carried out, foundation stability and concrete quality are judged, and early warning prompts are provided.
It realizes a comprehensive and objective analysis of construction project quality inspection, ensures the reliability and authenticity of the inspection results, avoids analysis errors caused by mistakes or equipment errors, and improves the detection effect.
Smart Images

Figure CN118710135B_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of engineering quality detection, and specifically to an engineering quality detection method based on deep learning. Background Art
[0002] As people's demand for construction continues to increase, in order to avoid the emergence of shoddy projects, the most basic foundation testing and concrete testing in the quality inspection of construction projects are also of paramount importance.
[0003] Prior art, such as the invention application patent with announcement number: CN116805314B, discloses a method for evaluating the quality of a construction project, the method comprising: obtaining a wall image of a construction project; obtaining a plurality of edge connected domains based on the wall image; calculating the change compliance of each edge connected domain; and then obtaining the degree of abnormality of each edge connected domain; obtaining a plurality of edge connected domains belonging to each pixel, calculating the correlation between each pixel and each edge connected domain, and then obtaining the degree of abnormality of each pixel; enhancing the wall image according to the degree of abnormality of each pixel to obtain an enhanced wall image, and evaluating the quality of the construction project according to the enhanced wall image, thereby accurately detecting cracks on the wall of the construction project, and thereby improving the accuracy of the quality evaluation of the construction project.
[0004] The present application obtains a foundation detection model by training a pre-trained model, obtains foundation data, and verifies whether the foundation data is valid; transmits the valid foundation data to the foundation detection model, and determines whether the stability of the foundation is abnormal; thereby solving the limitation problems existing in the feasibility analysis process of the current project quality detection development; when there is no abnormality in the stability of the foundation, obtains concrete data, verifies whether the concrete data is valid, and determines whether the concrete quality is abnormal; thereby achieving a comprehensive and objective analysis of the feasibility of project quality detection, and ensuring the reliability and authenticity of the project quality detection analysis results.
[0005] With respect to the above-mentioned scheme, the inventors of the present application have found that the above-mentioned technical scheme has at least the following technical problems: 1. Currently, the quality of a project is mainly tested by analyzing the wall surface, and there is a lack of testing of the foundation of the building when it is initially built. When there is a problem with the foundation of a building, no matter how perfect the wall surface is, it cannot be put into use, and thus it is impossible to more accurately understand the actual situation of the current project quality, and it is also impossible to guarantee the comprehensiveness and authenticity of the feasibility analysis process of the project quality detection, and thus it is impossible to guarantee the reference and accuracy of the analysis results, and it is also impossible to provide a reliable basis for the project quality detection.
[0006] 2. There is currently a lack of testing of concrete in projects. If the concrete of a building does not meet the standards initially, no problems may be revealed. However, once the building is put into use, all problems will emerge, resulting in problems such as easily chiseling open the adjacent walls during renovation and load-bearing walls being unable to bear the weight. As a result, the accuracy of the project quality inspection results cannot be guaranteed. Summary of the invention
[0007] In view of the above-mentioned technical deficiencies, the purpose of this application is to provide an engineering quality detection method based on deep learning.
[0008] In order to solve the above technical problems, the present application adopts the following technical solutions: The present application provides an engineering quality inspection method based on deep learning, including: Step 1, foundation detection model training: extracting the basic data set of the foundation from the database, and dividing the basic data set that has completed data processing into a basic data training set and a basic data verification set, and then using the basic data training set to train the pre-trained model, and then using the basic data verification set to verify the pre-trained model, thereby obtaining the foundation detection model.
[0009] Step 2: Acquisition and detection of foundation data: Observe the settlement and carry out bearing capacity test on the foundation to obtain foundation data and verify whether the foundation data is valid. Then, the valid foundation data is transmitted to the foundation detection model, the stability coefficient corresponding to the foundation is output, and it is determined whether there is any abnormality in the stability of the foundation.
[0010] Step 3. Acquisition and testing of concrete data: When there is no abnormality in the stability of the foundation and the concrete is poured, the quality of the concrete is tested to obtain the concrete data. The concrete data includes the strength, density and slump of the concrete, so as to verify whether the concrete data is valid, analyze and obtain the concrete quality assessment coefficient, and determine whether there is any abnormality in the concrete quality.
[0011] Step 4. Early warning: When the staff enters data multiple times and all fail the inspection, an early warning will be issued; when the stability of the foundation or the quality of the concrete is abnormal, an early warning will be issued.
[0012] Preferably, the data in the basic data training set is input into the pre-trained model, and the basic data verification set is input to verify the pre-trained model, and then the model is optimized based on the back propagation optimization algorithm. The model outputs the stability coefficient corresponding to the foundation, and calculates the gradient of the loss function to the parameters based on the error between the output stability coefficient and the actual value of the stability coefficient, and then updates the parameters in the opposite direction of the gradient, and gradually reduces the loss function value, thereby obtaining a foundation detection model.
[0013] Preferably, the foundation data includes settlement data and bearing capacity data of the foundation, wherein the settlement data includes foundation settlement velocity and distance, and the bearing capacity data includes deformation degree, stress value and vibration amplitude of the foundation.
[0014] Preferably, the verification of whether the foundation data is valid is carried out as follows: the staff uploads the foundation settlement speed and distance of the settlement data in the foundation data, the deformation, stress value and vibration amplitude of the foundation of the bearing capacity data to the basic data verification unit in the foundation data detection module respectively. When the basic data of the foundation pass the inspection of the basic data verification unit, the foundation data is judged to be valid; when any data in the basic data of the foundation fails to pass the inspection of the basic data verification unit, the foundation data is judged to be invalid, and the staff is prompted to re-collect and re-enter the data that failed the inspection.
[0015] Preferably, the specific inspection process for checking whether the concrete data is valid is as follows: the staff uploads the strength, density and slump of the concrete in the concrete data to the concrete data inspection unit in the foundation data detection module respectively. When all the data in the concrete data pass the inspection of the concrete data inspection unit, the concrete data is judged to be valid; when any one of the data in the concrete data fails to pass the inspection of the concrete data inspection unit, the concrete data is judged to be invalid, and the staff is prompted to re-collect and enter the data that failed the inspection.
[0016] The beneficial effects of the present application are: 1. The present application provides a method for project quality inspection based on deep learning, which obtains a foundation inspection model by training a pre-trained model, obtains foundation data, and verifies whether the foundation data is valid, transmits the valid foundation data to the foundation inspection model, and determines whether the stability of the foundation is abnormal, thereby solving the limitations existing in the feasibility analysis process of the current project quality inspection development; when the stability of the foundation is not abnormal, obtains concrete data, verifies whether the concrete data is valid, and determines whether the concrete quality is abnormal; realizes the comprehensive and objective analysis of the feasibility of project quality inspection, and ensures the reliability and authenticity of the project quality inspection analysis results.
[0017] 2. In this application, settlement observation and bearing capacity test are carried out on the foundation, wherein the bearing capacity test includes static load test and dynamic load test, so as to obtain the basic data of the foundation, and perform validity test on the basic data of the foundation to avoid errors in the analysis results due to staff mistakes or equipment errors, and to ensure the validity of the basic data of the foundation to a certain extent. Only when the valid basic data is input into the foundation detection model can the correct stability coefficient of the foundation be output, thereby completing the judgment of the stability of the foundation.
[0018] 3. In this application, when there is no abnormality in the stability of the foundation and the concrete is poured, the quality of the concrete is inspected. This ensures that there is no abnormality in the foundation before the next part of the construction and data analysis is carried out to avoid irreparable losses. In addition, data on the strength, density and slump of the concrete are obtained. After ensuring the validity of the data, analysis is carried out to determine whether there is any abnormality in the quality of the concrete based on the analysis results, which improves the detection effect of the concrete to a certain extent. BRIEF DESCRIPTION OF THE DRAWINGS
[0019] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0020] Figure 1 The figure is a flowchart of the implementation steps of the present application method. DETAILED DESCRIPTION
[0021] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.
[0022] See also Figure 1 As shown, the present application provides an engineering quality inspection method based on deep learning, comprising the following steps: Step 1, foundation detection model training: extract the basic data set of the foundation from the database, and divide the basic data set that has completed data processing into a basic data training set and a basic data verification set, and then use the basic data training set to train the pre-trained model, and then use the basic data verification set to verify the pre-trained model, thereby obtaining the foundation detection model.
[0023] In a specific example, the foundation detection model is obtained, and the specific acquisition process is as follows: the data in the basic data training set is input into the pre-trained model, and the basic data verification set is input to verify the pre-trained model, and then the model is optimized based on the back propagation optimization algorithm. The model outputs the stability coefficient corresponding to the foundation, and calculates the gradient of the loss function to the parameters based on the error between the output stability coefficient and the actual value of the stability coefficient, and then updates the parameters in the opposite direction of the gradient, and gradually reduces the loss function value, thereby obtaining the foundation detection model.
[0024] It should be noted that the basic data set includes the settlement velocity, distance, deformation, stress value and vibration amplitude of the foundation, as well as various data of the foundation combined with the corresponding stability coefficient.
[0025] Step 2: Acquisition and detection of foundation data: Observe the settlement and carry out bearing capacity test on the foundation to obtain foundation data and verify whether the foundation data is valid. Then, the valid foundation data is transmitted to the foundation detection model, the stability coefficient corresponding to the foundation is output, and it is determined whether there is any abnormality in the stability of the foundation.
[0026] In a specific example, the foundation data includes settlement data and bearing capacity data of the foundation, wherein the settlement data includes the settlement speed and distance of the foundation, and the bearing capacity data includes the deformation degree, stress value and vibration amplitude of the foundation.
[0027] In a specific example, the foundation is subjected to settlement observation and bearing capacity testing to obtain foundation data. The specific testing process is as follows: A1. When the foundation construction is completed, settlement observation equipment is installed to monitor the settlement of the foundation, and then the foundation settlement speed and distance are collected and recorded as the foundation settlement data.
[0028] A2. Bearing capacity test includes foundation static load test and foundation dynamic load test.
[0029] First, a static load test of the foundation is carried out. A preset load is applied to the foundation, and a pressure plate is installed to monitor the deformation and stress value of the foundation. Then a dynamic load test of the foundation is carried out. An impact loader device is used to apply a dynamic load to the foundation. At the same time, the dynamic response of the foundation is monitored and the vibration amplitude of the foundation is collected.
[0030] The deformation, stress value and vibration amplitude of the foundation are recorded as the bearing capacity data of the foundation.
[0031] It should be noted that the vibration amplitude of the foundation is collected by an acceleration sensor.
[0032] In a specific example, the verification of whether the foundation data is valid is carried out as follows: the staff uploads the foundation settlement speed and distance of the settlement data in the foundation data, and the deformation, stress value and vibration amplitude of the foundation of the bearing capacity data to the basic data verification unit in the foundation data detection module respectively. When the basic data of the foundation pass the inspection of the basic data verification unit, the foundation data is judged to be valid; when any data in the basic data of the foundation fails to pass the inspection of the basic data verification unit, the foundation data is judged to be invalid, and the staff is prompted to re-collect and re-enter the data that failed the inspection.
[0033] It should be noted that the basic data verification unit includes the standard range of each data in the foundation data. When the data falls within the standard range, it is valid, otherwise it is invalid.
[0034] In a specific example, the determination of whether there is an abnormality in the stability of the foundation is carried out as follows: the stability coefficient corresponding to the foundation is compared with the stability coefficient threshold of the foundation in the database. When the stability coefficient corresponding to the foundation is greater than or equal to the stability coefficient threshold of the foundation in the database, it is determined that there is an abnormality in the stability of the foundation; otherwise, it is determined that there is no abnormality in the stability of the foundation.
[0035] In this application, settlement observation and bearing capacity test are carried out on the foundation, wherein the bearing capacity test includes static load test and dynamic load test, so as to obtain the basic data of the foundation, and perform validity test on the basic data of the foundation to avoid errors in the analysis results due to staff mistakes or equipment errors, thereby ensuring the validity of the basic data of the foundation to a certain extent. Only when the valid basic data is input into the foundation detection model can the correct stability coefficient of the foundation be output, thereby completing the judgment of the foundation stability.
[0036] Step 3. Acquisition and testing of concrete data: When there is no abnormality in the stability of the foundation and the concrete is poured, the quality of the concrete is tested to obtain the concrete data. The concrete data includes the strength, density and slump of the concrete, so as to verify whether the concrete data is valid, analyze and obtain the concrete quality assessment coefficient, and determine whether there is any abnormality in the concrete quality.
[0037] In a specific example, the concrete is quality tested to obtain concrete data, and the specific testing process is as follows: B1. Concrete strength test. After the concrete is poured, concrete samples are taken from different concrete preparation equipment, and the sampled concrete is poured into a standard mold, filled and compacted in layers, and the concrete surface is smoothed, and then the concrete is cured; after the concrete is cured, a concrete sample is obtained, a universal testing machine is used to perform a crushing test on the concrete sample, and the strength of the concrete is collected.
[0038] B2. Use nuclear density meter equipment to collect the density of concrete.
[0039] B3. Slump measurement: Use a slump cone to test the slump of concrete. After turning the cone upside down and taking it out, measure the height of the concrete collapse and record it as the slump of the concrete.
[0040] The strength, density and slump of concrete are recorded as concrete data.
[0041] It should be noted that when sampling from different parts, usually at least two samples are taken from each truckload of concrete to ensure the representativeness of the sampling.
[0042] It should be noted that concrete curing means curing the concrete in a standard mold for 24 hours, then wrapping it with a wet cloth and curing it under constant temperature and humidity conditions for 28 days.
[0043] In a specific example, the specific inspection process of checking whether the concrete data is valid is as follows:
[0044] The staff uploads the strength, density and slump of the concrete in the concrete data to the concrete data verification unit in the foundation data detection module respectively. When all the data in the concrete data pass the inspection of the concrete data verification unit, the concrete data is judged to be valid; when any of the data in the concrete data fails to pass the inspection of the concrete data verification unit, the concrete data is judged to be invalid, and the staff is prompted to re-collect and re-enter the data that failed the inspection.
[0045] It should be noted that the concrete data inspection unit includes the standard range of each data in the concrete data. When the data falls within the standard range, it is valid, otherwise it is invalid.
[0046] In a specific example, the analysis results in a concrete quality assessment coefficient, and the specific analysis process is as follows: the strength, density and slump of the concrete are respectively recorded as , and , substitute into the calculation formula Concrete quality assessment coefficient ,in Expressed as a natural constant, , and They are respectively represented as the reference values of strength, density and slump of concrete in the database. and They are respectively expressed as the allowable floating value of the strength and density of the concrete in the database. , and They are respectively represented as the weight factor corresponding to the strength of concrete in the database, the weight factor corresponding to the density, and the weight factor corresponding to the slump.
[0047] It should be noted that , , ,and .
[0048] It should be noted that the weight factors corresponding to the strength, density and slump of concrete are obtained through factor analysis. First, the information of the strength, density and slump of concrete is condensed, and then the variance explanation rate after rotation is obtained, and the weight is obtained by dividing the cumulative variance explanation rate.
[0049] It should be noted that factor analysis is a well-known technology. It is a multivariate statistical analysis method that starts from studying the internal dependencies of variables and reduces some variables with intricate relationships to a few comprehensive factors; information concentration is expressed as calculating the median; the variance explanation rate is the amount of information extracted by the factor, and the variance explanation rate = characteristic root / total number of analysis items; the variance explanation rate after rotation is expressed as the variance explanation rate of the factor after maximum variance rotation.
[0050] In a specific example, the specific judgment process for judging whether the concrete quality is abnormal is as follows: comparing the concrete quality assessment coefficient with the concrete quality assessment coefficient threshold in the database; when the concrete quality assessment coefficient is less than or equal to the concrete quality assessment coefficient threshold in the database, it is judged that the concrete quality is abnormal; otherwise, it is judged that the concrete quality is not abnormal.
[0051] In this application, when there is no abnormality in the stability of the foundation and the concrete is poured, the quality of the concrete is inspected. This ensures that there is no abnormality in the foundation before the next part of the construction and data analysis is carried out to avoid irreparable losses. In addition, data on the strength, density and slump of the concrete are obtained, and analysis is performed after ensuring the validity of the data. Based on the analysis results, it is determined whether there is any abnormality in the quality of the concrete, which improves the detection effect of the concrete to a certain extent.
[0052] Step 4. Early warning: When the staff enters data multiple times and all fail the inspection, an early warning will be issued; when the stability of the foundation or the quality of the concrete is abnormal, an early warning will be issued.
[0053] The present application provides a method for project quality inspection based on deep learning. The method obtains a foundation inspection model by training a pre-trained model, obtains foundation data, and verifies whether the foundation data is valid. The valid foundation data is transmitted to the foundation inspection model, and it is determined whether the stability of the foundation is abnormal. This method solves the limitation problems existing in the feasibility analysis process of the current project quality inspection development. When the stability of the foundation is not abnormal, the concrete data is obtained, and it is verified whether the concrete data is valid, and it is determined whether the concrete quality is abnormal. This method realizes the comprehensive and objective analysis of the feasibility of project quality inspection, and ensures the reliability and authenticity of the project quality inspection analysis results.
[0054] The above contents are merely examples and explanations of the concept of the present application. The technical personnel in the relevant technical field may make various modifications or additions to the described specific embodiments or replace them in a similar manner. As long as they do not deviate from the concept of the invention or exceed the scope defined in this specification, they should all fall within the scope of protection of the present application.
Claims
1. A method for engineering quality detection based on deep learning, characterized in that: include: Step 1: Training the foundation detection model: extract the basic data set of the foundation from the database, and divide the basic data set that has completed data processing into a basic data training set and a basic data verification set, and then use the basic data training set to train the pre-trained model, and then use the basic data verification set to verify the pre-trained model, thereby obtaining the foundation detection model; The foundation detection model is obtained in this way, and the specific acquisition process is as follows: The data in the basic data training set is input into the pre-trained model, and the basic data verification set is input to verify the pre-trained model, and then the model is optimized based on the back propagation optimization algorithm. The model outputs the stability coefficient corresponding to the foundation, and calculates the gradient of the loss function to the parameter according to the error between the output stability coefficient and the actual value of the stability coefficient, and then updates the parameter according to the opposite direction of the gradient, and gradually reduces the loss function value, so as to obtain the foundation detection model; Step 2: Acquisition and detection of foundation data: Observe the settlement and test the bearing capacity of the foundation to obtain foundation data and verify whether the foundation data is valid. Then, the valid foundation data is transmitted to the foundation detection model, the stability coefficient corresponding to the foundation is output, and it is determined whether the stability of the foundation is abnormal. The foundation data includes settlement data and bearing capacity data of the foundation, wherein the settlement data includes the settlement speed and distance of the foundation, and the bearing capacity data includes the deformation degree, stress value and vibration amplitude of the foundation; The settlement observation and bearing capacity test of the foundation are carried out to obtain foundation data. The specific test process is as follows: A1. When the foundation construction is completed, install settlement observation equipment to monitor the settlement of the foundation, and then collect the settlement speed and distance of the foundation, which will be recorded as the settlement data of the foundation; A2. Bearing capacity test includes foundation static load test and foundation dynamic load test; First, a static load test of the foundation is carried out, a preset load is applied to the foundation, and a pressure plate is installed to monitor the deformation and stress value of the foundation; Then conduct a foundation dynamic load test, use an impact loader to apply a dynamic load on the foundation, monitor the dynamic response of the foundation, and collect the vibration amplitude of the foundation; The deformation, stress value and vibration amplitude of the foundation are recorded as the bearing capacity data of the foundation; The specific inspection process of checking whether the foundation data is valid is as follows: The staff uploads the foundation settlement speed and distance of the settlement data in the foundation data, the deformation degree, stress value and vibration amplitude of the foundation of the bearing capacity data to the basic data inspection unit in the foundation data detection module. When the basic data of the foundation pass the inspection of the basic data inspection unit, the foundation data is judged to be valid; When any of the basic data of the foundation fails to pass the inspection of the basic data inspection unit, the foundation data is judged to be invalid, and the staff is prompted to re-collect and enter the data that failed the inspection; The specific process of judging whether the stability of the foundation is abnormal is as follows: The stability coefficient corresponding to the foundation is compared with the stability coefficient threshold of the foundation in the database. When the stability coefficient corresponding to the foundation is greater than or equal to the stability coefficient threshold of the foundation in the database, it is determined that the stability of the foundation is abnormal; otherwise, it is determined that the stability of the foundation is not abnormal; Step 3: Acquisition and testing of concrete data: When there is no abnormality in the stability of the foundation and the concrete is poured, the quality of the concrete is tested to obtain the concrete data, which includes the strength, density and slump of the concrete, so as to verify whether the concrete data is valid, analyze and obtain the concrete quality assessment coefficient, and determine whether the concrete quality is abnormal; The concrete quality is tested to obtain concrete data. The specific testing process is as follows: B1. Concrete strength test: after the concrete pouring is completed, concrete samples are taken from different concrete preparation equipment, the sampled concrete is poured into a standard mold, filled and compacted in layers, and the concrete surface is leveled and then cured; after the concrete is cured, concrete samples are obtained, and a universal testing machine is used to perform a crushing test on the concrete samples, and the strength of the concrete is collected; B2. Use nuclear density meter equipment to collect the density of concrete; B3. Slump measurement: Use a slump cone to test the slump of concrete. After turning the cone upside down and taking it out, measure the height of the concrete collapse and record it as the slump of the concrete. The strength, density and slump of concrete are recorded as concrete data; The specific inspection process of checking whether the concrete data is valid is as follows: The staff uploads the strength, density and slump of the concrete in the concrete data to the concrete data inspection unit in the foundation data detection module. When all the data in the concrete data pass the inspection of the concrete data inspection unit, the concrete data is judged to be valid; when any of the data in the concrete data fails to pass the inspection of the concrete data inspection unit, the concrete data is judged to be invalid, and the staff is prompted to re-collect and enter the data that failed the inspection; The analysis results in a concrete quality assessment coefficient, and the specific analysis process is as follows: The strength, density and slump of concrete are respectively , and , substitute into the calculation formula Concrete quality assessment coefficient ,in is expressed as a natural constant, , and They are respectively represented as the reference values of strength, density and slump of concrete in the database. and They are respectively expressed as the allowable floating value of the strength and density of the concrete in the database. , and They are respectively represented as the weight factor corresponding to the strength of concrete in the database, the weight factor corresponding to the density, and the weight factor corresponding to the slump; Step 4: Early warning: When the staff has entered data multiple times and all have failed the inspection, an early warning will be issued; When there is an abnormality in the stability of the foundation or the quality of concrete, an early warning will be issued.
2. According to the deep learning-based engineering quality detection method of claim 1, it is characterized in that: The specific judgment process of judging whether the concrete quality is abnormal is as follows: The concrete quality assessment coefficient is compared with the concrete quality assessment coefficient threshold in the database. When the concrete quality assessment coefficient is less than or equal to the concrete quality assessment coefficient threshold in the database, it is judged that the concrete quality is abnormal; otherwise, it is judged that the concrete quality is not abnormal.
