Building material quality detection method and system

Abstract

The invention discloses a building material quality detection method and system, and relates to the technical field of quality detection.The building material quality detection method comprises the steps that a distributed sensor array is formed by deploying multifunctional intelligent nodes to collect monitoring data, and meanwhile a high-resolution image is collected in combination with an unmanned aerial vehicle; performing fusion processing on the multi-source data to generate comprehensive analysis basic data; establishing a time sequence prediction model by using a convolutional neural network and a long-short-term memory network, and predicting a material performance change trend; performing abnormal sample detection by adopting density clustering and a support vector machine algorithm; and constructing a three-dimensional digital twinborn model, visually displaying an abnormal detection result and generating a detection report. According to the invention, a complete technical system is formed, intelligentization, precision and visualization of building material quality detection are realized, and powerful support is provided for quality management decision.

Description

Technical Field

[0001] The present invention relates to the technical field of quality detection, in particular to a method and system for detecting the quality of building materials. Background Art

[0002] The quality inspection of building materials is a key link in ensuring the safety and durability of construction projects. With the continuous expansion of the scale of construction projects and the increasing complexity of structures, traditional methods of building material quality inspection mainly rely on manual on-site sampling and laboratory testing. This method has problems such as low detection efficiency, limited coverage, and poor real-time performance. In recent years, with the rapid development of emerging technologies such as the Internet of Things, artificial intelligence, and digital twins, the quality inspection of building materials has gradually evolved towards intelligence and automation. At present, the industry has begun to use various sensors to monitor material properties and introduce computer vision technology to assist in detection and analysis. However, the existing detection technology still has technical bottlenecks such as single data collection form, insufficient fusion of multi-source heterogeneous data, and limited predictive analysis capabilities, which makes it difficult to meet the needs of quality management throughout the life cycle of construction projects.

[0003] The existing building material quality detection technology mainly has the following problems: first, the traditional single detection method is difficult to fully reflect the performance status of the material, resulting in insufficient reliability and integrity of the detection results; second, there is a lack of effective prediction methods for the evolution trend of material performance, and it is impossible to timely discover potential quality risks; third, the existing abnormality detection methods generally ignore the correlation between material performance indicators, which reduces the accuracy of detection; finally, the expression of the detection results is relatively simple, lacking intuitive three-dimensional visualization, which is not conducive to the understanding and application of the detection results. In response to the above problems, the present invention proposes a building material quality detection method based on multi-source data fusion and deep learning. This method belongs to the field of construction engineering quality detection technology, and aims to achieve intelligent and precise detection of building material quality by constructing intelligent sensor networks, multi-source data fusion analysis, deep learning prediction models, and digital twin visualization. Summary of the invention

[0004] The purpose of this section is to summarize some aspects of embodiments of the present invention and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section and the specification abstract and the invention title of this application to avoid blurring the purpose of this section, the specification abstract and the invention title, and such simplifications or omissions cannot be used to limit the scope of the present invention.

[0005] In view of the above existing problems, the present invention is proposed.

[0006] Therefore, the present invention provides a method and system for detecting quality of building materials, which can solve the problems mentioned in the background technology.

[0007] In order to solve the above technical problems, the present invention provides the following technical solutions: In a first aspect, the present invention provides a method for detecting the quality of building materials, which comprises deploying multifunctional intelligent nodes at building structure nodes to form a distributed sensor array, collecting sensor monitoring data, and simultaneously using a drone to collect high-resolution images of the building; Fusing the data from the distributed sensor array with the image information collected by the drone to construct a fused data set, and performing data processing on the fused data set to generate basic data for comprehensive analysis; Based on the comprehensive analysis basic data, a time series prediction model is established using a convolutional neural network and a long short-term memory network to predict the future performance change trend of building materials and obtain trend prediction results; Based on the trend prediction results, density clustering and support vector machine algorithms are used to detect abnormal samples in the comprehensive analysis basic data set, establish an N-dimensional feature space and construct sample classification boundaries to obtain abnormal sample detection results; A three-dimensional grid model is constructed based on the geometric size data, internal structure data and surface morphology data of building materials, the abnormal sample detection results are mapped to the three-dimensional grid model, a digital twin model is established and the abnormal areas are marked, an abnormal distribution heat map and a quality inspection report are generated, and quality inspection is completed.

[0008] As a preferred solution of the building material quality detection method of the present invention, the data processing includes: The vibration signals, stress data and environmental parameters collected by the distributed sensor array are synchronized and aligned according to the timestamps to generate the sensor time series data matrix and calculate the variance of each sensor data; De-noise and filter out outliers on the sensor time series data matrix, extract the time domain and frequency domain features of various data through wavelet transform, and obtain the standardized sensor features; Perform image enhancement and feature extraction on panoramic images to obtain the texture features, crack features, and color change features of the building surface, and calculate each feature component and its variance; Based on the data heterogeneity adaptive fusion algorithm, the fusion features are calculated. The calculation formula is as follows: ; in, for The fusion feature output at each moment, For the The time series data of sensor nodes, is the corresponding sensor data variance, For the image feature components, is the corresponding image feature variance, m is the number of sensor nodes, n is the image feature dimension, and is the time-varying adaptive weight coefficient.

[0009] As a preferred solution of the building material quality detection method of the present invention, wherein: the establishment of a time series prediction model comprises: The comprehensive analysis basic data is divided into training sample sequences {X t , t=1,2,...,N}, construct a sliding time window data set; Establish a dual-branch neural network architecture, in which the convolutional neural network branch is used for spatial feature extraction and the long short-term memory network branch is used for temporal feature modeling; ; ; in, For the The feature map output of the layer, ReLU is the rectified linear unit activation function, For the The convolution kernel weights of the layer, For the The feature map input of the layer, No. The bias term of the layer, for The input gate state at the moment, σ is the sigmoid activation function, is the weight matrix input to the input gate, For input Connect to input gate , for The input vector at time instant, is the weight matrix from hidden state to input gate, From hidden state Connect to input gate , for The hidden state of the moment, is the bias term of the input gate.

[0010] As a preferred solution of the building material quality detection method of the present invention, wherein: the establishment of the time series prediction model also includes: The spatial features in each time window are extracted through a convolutional neural network, and the local structural features of building materials are obtained by multi-layer convolution and pooling operations; ; in, For the The output of the convolution kernel, is the pooling operation, C is the number of input feature channels, For the The convolution kernel weights of the channels, for Moment Channel input, For the The bias term of each channel; The local structural features are input into the long short-term memory network, and the dynamic characteristics of the performance changes of building materials are captured through temporal dependency modeling. The calculation formula is as follows: ; in, is the memory unit state, is the output gate state, For the The output of the LSTM unit, is the LSTM output layer weight matrix, for The hidden state of the moment, is the output layer bias term, is the element-wise multiplication operator; Based on the attention mechanism, the feature weights of different time steps are calculated, and the output of the long short-term memory network is adaptively weighted. The calculation formula is as follows: ; in, for Moment The attention weights of the CNN features, for Moment The attention weights of the LSTM features, For the The attention score of each CNN feature, For the The attention score of each LSTM feature, is the number of CNN features, is the number of LSTM features; The output results of the dual-branch network are integrated to generate a multi-dimensional prediction value of the future performance of building materials through a fully connected layer. The calculation formula is as follows: ; in, For the future The predicted value at time, is the fusion weight coefficient; The confidence interval of the prediction results is estimated to generate prediction results that include performance change trends and their uncertainties. The confidence interval calculation formula is as follows: ; Among them, the use of for The prediction confidence interval at time , is the standard normal distribution quantile corresponding to the confidence level, is the true value, is the predicted value.

[0011] As a preferred solution of the building material quality detection method of the present invention, the detection of abnormal samples in the comprehensive analysis basic data includes: The trend prediction results are constructed into an N-dimensional feature vector according to spatial position, time series, and performance indexes to generate an abnormality detection feature space, wherein the feature space includes physical property parameters, chemical component data, and mechanical performance indexes of building materials; The abnormality detection feature space is normalized to generate a normalized feature matrix, wherein the value range of each dimension of the normalized feature matrix is ​​in the interval [0, 1]; The density clustering algorithm is used to perform cluster analysis on the normalized feature matrix to obtain a sample density distribution map, and the sample clusters are divided based on the local density maximum. The calculation formula of the local density is as follows: ; in, is the local density, is the cutoff distance, For sample points and The Euclidean distance between them, n is the total number of samples, is the index of other sample points except sample point i, χ( ) is the truncation function; The sample clusters are classified using a support vector machine, the classification boundary between normal samples and abnormal samples is established, and the abnormal sample detection result is generated. The decision function of the classification boundary is: ; in, is the classification decision function, is the kernel function, is the Lagrange multiplier corresponding to the i-th support vector, is the bias term, is the input feature vector, m is the number of support vectors, is the category label of the support vector.

[0012] As a preferred solution of the building material quality detection method of the present invention, wherein: the construction of the three-dimensional grid model includes: Acquire geometric dimension data, internal structure data, and surface morphology data of building materials, wherein the geometric dimension data includes length, width, and height parameters, the internal structure data includes cross-sectional characteristics, pore distribution, and crack direction, and the surface morphology data includes roughness, flatness, and surface defects; A three-dimensional grid model is constructed based on the geometric dimension data, internal structure data, and surface morphology data. The three-dimensional grid model divides the space using hexahedral units.

[0013] As a preferred solution of the building material quality detection method of the present invention, the establishment of the digital twin model includes: Establishing a spatial mapping relationship between the abnormal sample detection result and the grid unit of the three-dimensional grid model to generate an abnormal position index table; Marking an abnormal area in the three-dimensional grid model according to the abnormal position index table, wherein the marked abnormal area includes marking an abnormal type identifier, an abnormal degree value and an abnormal range boundary; A digital twin model is constructed based on the three-dimensional grid model, and the digital twin model integrates physical space geometry information and anomaly detection information.

[0014] In a second aspect, the present invention provides a building material quality detection system, which includes: a data acquisition module, a data fusion module, a model building module, an anomaly detection module and a quality labeling module; The data acquisition module is used to deploy multifunctional intelligent nodes at the building structure nodes to form a distributed sensor array, collect sensor monitoring data, and simultaneously use drones to collect high-resolution images of the building; The data fusion module is used to fuse the data of the distributed sensor array with the image information collected by the drone, construct a fused data set, and perform data processing on the fused data set to generate basic data for comprehensive analysis; The model building module is used to establish a time series prediction model based on the comprehensive analysis basic data using a convolutional neural network and a long short-term memory network to predict the future performance change trend of building materials and obtain a trend prediction result; The anomaly detection module is used to detect abnormal samples in the comprehensive analysis basic data set based on the trend prediction result, establish an N-dimensional feature space and construct a sample classification boundary, and obtain an abnormal sample detection result; The quality labeling module is used to construct a three-dimensional grid model based on the geometric size data, internal structure data and surface morphology data of building materials, map the abnormal sample detection results into the three-dimensional grid model, establish a digital twin model and mark the abnormal area, and generate an abnormal distribution heat map and a quality inspection report.

[0015] In a third aspect, the present invention provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the processor executes the computer program, the steps of the building material quality detection method are implemented.

[0016] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: the computer program, when executed by a processor, implements the steps of a method for detecting quality of building materials.

[0017] Compared with the prior art, the beneficial effects of the present invention are that the present invention realizes the acquisition of multi-dimensional data of the internal state and appearance characteristics of building materials by deploying multifunctional intelligent nodes at the building structure nodes to form a distributed sensor array and combining with drone image acquisition, overcoming the limitation of insufficient information dimension of traditional single detection means; adopts data heterogeneity adaptive fusion algorithm to process multi-source data, dynamically adjusts the weight of each data source through time-varying adaptive weight coefficient, effectively solves the scale inconsistency problem in heterogeneous data fusion; constructs a dual-branch neural network architecture and introduces an attention mechanism to realize accurate prediction of the performance trend of building materials, and provides uncertainty quantification of prediction results through confidence interval estimation; finally, density clustering and support vector machine algorithms are used for anomaly detection, and the detection results are mapped to a three-dimensional digital twin model for visualization, which significantly improves the accuracy of anomaly detection and the comprehensibility of the results. The present invention forms a complete technical system from data acquisition, processing, analysis to visualization, realizes the intelligence, precision and visualization of building material quality detection, and provides strong technical support for building material quality management decision-making. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0019] Figure 1 A method flow chart of a building material quality detection method and system provided by an embodiment of the present invention; Figure 2An internal structural diagram of a computer device of a building material quality detection method and system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0020] In order to make the above-mentioned purposes, features and advantages of the present invention more understandable, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0021] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0022] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0023] Example 1, reference Figure 1-Figure 2 , which is the first embodiment of the present invention, provides a method for detecting the quality of building materials, comprising: Before describing the embodiments of the present application in detail, some related concepts are first explained for the sake of clarity.

[0024] Long Short-Term Memory (LSTM): A special recurrent neural network structure that can effectively process and predict time series data by designing input gates, forget gates, and output gates. It is particularly suitable for capturing long-term dependencies in data. In this invention, it is used for time series feature modeling and trend prediction of building material performance.

[0025] Digital twin model: refers to a digital virtual model that is built on the basis of physical space and completely corresponds to the entity. The model can reflect the state changes of the physical entity in real time. In the present invention, it is used for the three-dimensional visualization of building materials and the marking of abnormal positions.

[0026] Density clustering: A clustering analysis method based on sample density. It identifies different data clusters by calculating the local density of sample points. It is particularly suitable for discovering clusters with non-spherical distribution and processing data containing noise. The present invention uses it for the preliminary identification of abnormal samples.

[0027] Heterogeneous data fusion refers to the process of unified processing and integration of data from different sources, different features, and different structures. In this invention, it specifically refers to the technical method of fusing sensor data and image data, and realizing the effective combination of multi-source data through adaptive weights.

[0028] The present application provides a method that can effectively solve the above-mentioned problems. Next, multiple embodiments will be combined to explain in detail how to implement the building material quality detection method; Figure 1 A method flow chart of a building material quality detection method and system is shown, including: S1: Deploy multifunctional intelligent nodes at the building structure nodes to form a distributed sensor array to collect sensor monitoring data, and use drones to collect high-resolution images of the building at the same time; Furthermore, the collected sensor monitoring data includes: Multifunctional intelligent nodes with an adaptive sampling frequency adjustment mechanism are deployed at key stress points, deformation sensitive areas and material stress concentration areas of the building structure, and vibration signals, stress data and environmental parameters are collected through the multifunctional intelligent nodes; Based on the real-time monitoring data of the multifunctional intelligent node, the sampling frequency of each node is dynamically adjusted, the sampling frequency is increased in the area where the signal changes violently, and the sampling frequency is reduced in the area where the signal is stable; A wireless communication network is established between the multifunctional intelligent nodes to form a distributed sensor array, and the monitoring data collected by the distributed sensor array is transmitted to a data processing center.

[0029] Furthermore, collecting high-resolution images of buildings includes: Controlling the drone to shoot the exterior of the building from multiple angles according to a preset layered scanning strategy to collect high-resolution images of the building surface, wherein the layered scanning strategy includes a bottom-up spiral ascending path and a fixed-height horizontal circling scan; Performing geometric correction and stitching on the high-resolution images collected by the drone to generate a panoramic image of the building surface; The panoramic image is divided into image blocks corresponding to the positions of the distributed sensor array nodes, and a spatial mapping relationship between image data and sensor data is established.

[0030] S2: Fusing the data of the distributed sensor array with the image information collected by the drone to construct a fused data set, and performing data processing on the fused data set to generate basic data for comprehensive analysis; Furthermore, data processing includes: The vibration signals, stress data and environmental parameters collected by the distributed sensor array are synchronized and aligned according to the timestamps to generate the sensor time series data matrix and calculate the variance of each sensor data; De-noise and filter out outliers on the sensor time series data matrix, extract the time domain and frequency domain features of various data through wavelet transform, and obtain the standardized sensor features; Perform image enhancement and feature extraction on panoramic images to obtain the texture features, crack features, and color change features of the building surface, and calculate each feature component and its variance; Based on the data heterogeneity adaptive fusion algorithm, the fusion features are calculated. The calculation formula is as follows: ; in, for The fusion feature output at each moment, For the The time series data of sensor nodes, is the corresponding sensor data variance, For the image feature components, is the corresponding image feature variance, m is the number of sensor nodes, n is the image feature dimension, and is a time-varying adaptive weight coefficient, which is dynamically adjusted according to data quality and satisfies + =1.

[0031] Furthermore, the fusion feature F(t) is subjected to dimension reduction processing to eliminate feature redundancy and extract the main component features; The principal component characteristics are reorganized according to the time series to generate basic data for comprehensive analysis, and a basic data set for comprehensive analysis containing spatiotemporal characteristics is constructed.

[0032] S3: Based on the comprehensive analysis basic data, a time series prediction model is established using a convolutional neural network and a long short-term memory network to predict the future performance change trend of building materials and obtain trend prediction results; Further, the comprehensive analysis basic data set is divided into training sample sequences {X t , t=1,2,...,N}, construct a sliding time window data set; Establish a dual-branch neural network architecture, in which the convolutional neural network branch is used for spatial feature extraction and the long short-term memory network branch is used for temporal feature modeling; ; ; in, For the The feature map output of the layer, ReLU is the rectified linear unit activation function, For the The convolution kernel weights of the layer, For the The feature map input of the layer, No. The bias term of the layer, for The input gate state at the moment, σ is the sigmoid activation function, is the weight matrix input to the input gate, For input Connect to input gate , for The input vector at time instant, is the weight matrix from hidden state to input gate, From hidden state Connect to input gate , for The hidden state of the moment, is the bias term of the input gate; Furthermore, the spatial features in each time window are extracted through convolutional neural networks, and the local structural features of building materials are obtained by multi-layer convolution and pooling operations; ; in, For the The output of the convolution kernel, is the pooling operation, C is the number of input feature channels, For the The convolution kernel weights of the channels, for Moment Channel input, For the The bias term of each channel; The local structural features are input into the long short-term memory network, and the dynamic characteristics of the performance changes of building materials are captured through temporal dependency modeling. The calculation formula is as follows: ; in, is the memory unit state, is the output gate state, For the The output of the LSTM unit, is the LSTM output layer weight matrix, for The hidden state of the moment, is the output layer bias term, is the element-wise multiplication operator; Based on the attention mechanism, the feature weights of different time steps are calculated, and the output of the long short-term memory network is adaptively weighted. The calculation formula is as follows: ; in, for Moment The attention weights of the CNN features, for Moment The attention weights of the LSTM features, For the The attention score of each CNN feature, For the The attention score of each LSTM feature, is the number of CNN features, is the number of LSTM features; The output results of the dual-branch network are integrated to generate a multi-dimensional prediction value of the future performance of building materials through a fully connected layer. The calculation formula is as follows: ; in, For the future The predicted value at time, is the fusion weight coefficient; The confidence interval of the prediction results is estimated to generate prediction results that include performance change trends and their uncertainties. The confidence interval calculation formula is as follows: ; Among them, the use of for The prediction confidence interval at time , is the standard normal distribution quantile corresponding to the confidence level, is the true value, is the predicted value.

[0033] S4: Based on the trend prediction results, the density clustering and support vector machine algorithms are used to detect abnormal samples in the comprehensive analysis basic data set, establish an N-dimensional feature space and construct a sample classification boundary to obtain abnormal sample detection results; Further, the trend prediction results are constructed into an N-dimensional feature vector according to spatial position, time series, and performance indexes to generate an abnormality detection feature space, wherein the feature space includes physical property parameters, chemical component data, and mechanical performance indexes of building materials; The abnormality detection feature space is normalized to generate a normalized feature matrix, wherein the value range of each dimension of the normalized feature matrix is ​​in the interval [0, 1]; The density clustering algorithm is used to perform cluster analysis on the normalized feature matrix to obtain a sample density distribution map, and the sample clusters are divided based on the local density maximum. The calculation formula of the local density is as follows: ; in, is the local density, is the cutoff distance, For sample points and The Euclidean distance between them, n is the total number of samples, is the index of other sample points except sample point i, χ( ) is the truncation function; The sample clusters are classified using a support vector machine, the classification boundary between normal samples and abnormal samples is established, and the abnormal sample detection result is generated. The decision function of the classification boundary is: ; in, is the classification decision function, is the kernel function, is the Lagrange multiplier corresponding to the i-th support vector, is the bias term, is the input feature vector, m is the number of support vectors, is the category label of the support vector; if the output value of the decision function is greater than 0, it is judged as a normal sample, and if it is less than 0, it is judged as an abnormal sample. is the i-th support vector.

[0034] S5: Construct a three-dimensional grid model based on the geometric size data, internal structure data and surface morphology data of the building materials, map the abnormal sample detection results into the three-dimensional grid model, establish a digital twin model and mark the abnormal area, and generate an abnormal distribution heat map and a quality inspection report.

[0035] Furthermore, constructing a three-dimensional mesh model includes: Acquire geometric dimension data, internal structure data, and surface morphology data of building materials, wherein the geometric dimension data includes length, width, and height parameters, the internal structure data includes cross-sectional characteristics, pore distribution, and crack direction, and the surface morphology data includes roughness, flatness, and surface defects; A three-dimensional grid model is constructed based on the geometric dimension data, internal structure data, and surface morphology data. The three-dimensional grid model divides the space using hexahedral units.

[0036] Furthermore, building a digital twin model includes: Establishing a spatial mapping relationship between the abnormal sample detection result and the grid unit of the three-dimensional grid model to generate an abnormal position index table; Annotating abnormal areas in the three-dimensional grid model according to the abnormal position index table, wherein the abnormal area annotation includes an abnormal type identifier, an abnormal degree value, and an abnormal range boundary; A digital twin model is constructed based on the three-dimensional grid model, and the digital twin model integrates physical space geometry information and anomaly detection information.

[0037] Furthermore, a test report is generated, including: Performing classification and statistics on abnormal areas in the digital twin model to generate an abnormal distribution heat map; A quality inspection report is generated based on the abnormal distribution heat map, abnormal degree value and abnormal range boundary.

[0038] Embodiment 2 is the second embodiment of the present invention. This embodiment further provides a building material quality detection system, including: a data acquisition module, a data fusion module, a model building module, an anomaly detection module and a quality marking module; The data acquisition module is used to deploy multifunctional intelligent nodes at the building structure nodes to form a distributed sensor array, collect sensor monitoring data, and simultaneously use drones to collect high-resolution images of the building; The data fusion module is used to fuse the data of the distributed sensor array with the image information collected by the drone, construct a fused data set, and perform data processing on the fused data set to generate basic data for comprehensive analysis; The model building module is used to establish a time series prediction model based on the comprehensive analysis basic data using a convolutional neural network and a long short-term memory network to predict the future performance change trend of building materials and obtain trend prediction results; The anomaly detection module is used to detect abnormal samples in the comprehensive analysis basic data set based on the trend prediction results, establish an N-dimensional feature space and construct a sample classification boundary to obtain abnormal sample detection results; The quality labeling module is used to construct a three-dimensional grid model based on the geometric size data, internal structure data and surface morphology data of building materials, map the abnormal sample detection results into the three-dimensional grid model, establish a digital twin model and label the abnormal areas, and generate an abnormal distribution heat map and a quality inspection report.

[0039] This embodiment also provides a computer device, which may be a terminal, and its internal structure diagram may be as follows: Figure 2As shown. The computer device includes a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. When the computer program is executed by the processor, a method for detecting the quality of building materials is implemented. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a key, trackball or touchpad set on the computer device housing, or an external keyboard, touchpad or mouse, etc.

[0040] This embodiment also provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the following steps are implemented: deploying multifunctional intelligent nodes at building structure nodes to form a distributed sensor array, collecting sensor monitoring data, and simultaneously using drones to collect high-resolution images of buildings; Fusing the data from the distributed sensor array with the image information collected by the drone to construct a fused data set, and performing data processing on the fused data set to generate basic data for comprehensive analysis; Based on the comprehensive analysis basic data, a time series prediction model is established using a convolutional neural network and a long short-term memory network to predict the future performance change trend of building materials and obtain trend prediction results; Based on the trend prediction results, density clustering and support vector machine algorithms are used to detect abnormal samples in the comprehensive analysis basic data set, establish an N-dimensional feature space and construct sample classification boundaries to obtain abnormal sample detection results; A three-dimensional grid model is constructed based on the geometric size data, internal structure data and surface morphology data of building materials, the abnormal sample detection results are mapped to the three-dimensional grid model, a digital twin model is established and the abnormal areas are marked, an abnormal distribution heat map and a quality inspection report are generated, and quality inspection is completed.

[0041] Example 3 is the third embodiment of the present invention. It should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present invention, which should be included in the scope of the claims of the present invention.

[0042] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes. The schemes in the embodiments of the present application may be implemented in various computer languages, for example, object-oriented programming language Java and literal scripting language JavaScript, etc.

[0043] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0044] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0045] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0046] Although the preferred embodiments of the present application have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

[0047] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A method for detecting the quality of building materials, characterized in that: include: Deploy multifunctional intelligent nodes at the building structure nodes to form a distributed sensor array to collect sensor monitoring data, and use drones to collect high-resolution images of the building at the same time; Fusing the data from the distributed sensor array with the image information collected by the drone to construct a fused data set, and performing data processing on the fused data set to generate basic data for comprehensive analysis; Based on the comprehensive analysis basic data, a time series prediction model is established using a convolutional neural network and a long short-term memory network to predict the future performance change trend of building materials and obtain trend prediction results; Based on the trend prediction results, density clustering and support vector machine algorithms are used to detect abnormal samples in the comprehensive analysis basic data set, establish an N-dimensional feature space and construct sample classification boundaries to obtain abnormal sample detection results; A three-dimensional grid model is constructed based on the geometric size data, internal structure data and surface morphology data of building materials, the abnormal sample detection results are mapped to the three-dimensional grid model, a digital twin model is established and the abnormal areas are marked, an abnormal distribution heat map and a quality inspection report are generated, and quality inspection is completed.

2. The method for detecting the quality of building materials according to claim 1, characterized in that: The data processing includes: The vibration signals, stress data and environmental parameters collected by the distributed sensor array are synchronized and aligned according to the timestamps to generate the sensor time series data matrix and calculate the variance of each sensor data; De-noise and filter out outliers on the sensor time series data matrix, extract the time domain and frequency domain features of various data through wavelet transform, and obtain the standardized sensor features; Perform image enhancement and feature extraction on panoramic images to obtain the texture features, crack features, and color change features of the building surface, and calculate each feature component and its variance; Based on the data heterogeneity adaptive fusion algorithm, the fusion features are calculated. The calculation formula is as follows: ； in, for The fusion feature output at each moment, For the The time series data of sensor nodes, is the corresponding sensor data variance, For the image feature components, is the corresponding image feature variance, m is the number of sensor nodes, n is the image feature dimension, and is the time-varying adaptive weight coefficient.

3. The method for detecting the quality of building materials according to claim 2, characterized in that: The time series prediction model is established including: The comprehensive analysis basic data is divided into training sample sequences {X t , t=1,2,...,N}, construct a sliding time window data set; Establish a dual-branch neural network architecture, in which the convolutional neural network branch is used for spatial feature extraction and the long short-term memory network branch is used for temporal feature modeling; ； ； in, For the The feature map output of the layer, ReLU is the rectified linear unit activation function, For the The convolution kernel weights of the layer, For the The feature map input of the layer, No. The bias term of the layer, for The input gate state at the moment, σ is the sigmoid activation function, is the weight matrix input to the input gate, For input Connect to input gate , for The input vector at time instant, is the weight matrix from hidden state to input gate, From hidden state Connect to input gate , for The hidden state of the moment, is the bias term of the input gate.

4. The method for detecting the quality of building materials according to claim 3, characterized in that: The said establishing a time series prediction model also includes: The spatial features in each time window are extracted through a convolutional neural network, and the local structural features of building materials are obtained by multi-layer convolution and pooling operations; ； in, For the The output of the convolution kernel, is the pooling operation, C is the number of input feature channels, For the The convolution kernel weights of the channels, for Moment Channel input, For the The bias term of each channel; The local structural features are input into the long short-term memory network, and the dynamic characteristics of the performance changes of building materials are captured through temporal dependency modeling. The calculation formula is as follows: ； in, is the memory unit state, is the output gate state, For the The output of the LSTM unit, is the LSTM output layer weight matrix, for The hidden state of the moment, is the output layer bias term, is the element-wise multiplication operator; Based on the attention mechanism, the feature weights of different time steps are calculated, and the output of the long short-term memory network is adaptively weighted. The calculation formula is as follows: ； in, for Moment The attention weights of the CNN features, for Moment The attention weights of the LSTM features, For the The attention score of each CNN feature, For the The attention score of each LSTM feature, is the number of CNN features, is the number of LSTM features; The output results of the dual-branch network are integrated to generate a multi-dimensional prediction value of the future performance of building materials through a fully connected layer. The calculation formula is as follows: ； in, For the future The predicted value at time, is the fusion weight coefficient; The confidence interval of the prediction results is estimated to generate prediction results that include performance change trends and their uncertainties. The confidence interval calculation formula is as follows: ； Among them, the use of for The prediction confidence interval at time , is the standard normal distribution quantile corresponding to the confidence level, is the true value, is the predicted value.

5. The method for detecting the quality of building materials according to claim 4, characterized in that: The detecting of abnormal samples in the comprehensive analysis basic data includes: The trend prediction results are constructed into an N-dimensional feature vector according to spatial position, time series, and performance indexes to generate an abnormality detection feature space, wherein the feature space includes physical property parameters, chemical component data, and mechanical performance indexes of building materials; The abnormality detection feature space is normalized to generate a normalized feature matrix, wherein the value range of each dimension of the normalized feature matrix is ​​in the interval [0, 1]; The density clustering algorithm is used to perform cluster analysis on the normalized feature matrix to obtain a sample density distribution map, and the sample clusters are divided based on the local density maximum. The calculation formula of the local density is as follows: ； in, is the local density, is the cutoff distance, For sample points and The Euclidean distance between them, n is the total number of samples, is the index of other sample points except sample point i, χ( ) is the truncation function; The sample clusters are classified using a support vector machine, the classification boundary between normal samples and abnormal samples is established, and the abnormal sample detection result is generated. The decision function of the classification boundary is: ； in, is the classification decision function, is the kernel function, is the Lagrange multiplier corresponding to the i-th support vector, is the bias term, is the input feature vector, m is the number of support vectors, is the category label of the support vector.

6. The method for testing the quality of building materials according to claim 5, characterized in that: The constructing of the three-dimensional grid model comprises: Acquire geometric dimension data, internal structure data, and surface morphology data of building materials, wherein the geometric dimension data includes length, width, and height parameters, the internal structure data includes cross-sectional characteristics, pore distribution, and crack direction, and the surface morphology data includes roughness, flatness, and surface defects; A three-dimensional grid model is constructed based on the geometric dimension data, internal structure data, and surface morphology data. The three-dimensional grid model divides the space using hexahedral units.

7. The method for testing the quality of building materials according to claim 6, characterized in that: The establishment of the digital twin model includes: Establishing a spatial mapping relationship between the abnormal sample detection result and the grid unit of the three-dimensional grid model to generate an abnormal position index table; Marking an abnormal area in the three-dimensional grid model according to the abnormal position index table, wherein the marked abnormal area includes marking an abnormal type identifier, an abnormal degree value and an abnormal range boundary; A digital twin model is constructed based on the three-dimensional grid model, and the digital twin model integrates physical space geometry information and anomaly detection information.

8. A building material quality detection system, based on the building material quality detection method according to any one of claims 1 to 7, characterized in that: It includes data acquisition module, data fusion module, model building module, anomaly detection module and quality labeling module; The data acquisition module is used to deploy multifunctional intelligent nodes at the building structure nodes to form a distributed sensor array, collect sensor monitoring data, and simultaneously use drones to collect high-resolution images of the building; The data fusion module is used to fuse the data of the distributed sensor array with the image information collected by the drone, construct a fused data set, and perform data processing on the fused data set to generate basic data for comprehensive analysis; The model building module is used to establish a time series prediction model based on the comprehensive analysis basic data using a convolutional neural network and a long short-term memory network to predict the future performance change trend of building materials and obtain trend prediction results; The anomaly detection module is used to detect abnormal samples in the comprehensive analysis basic data set based on the trend prediction result, establish an N-dimensional feature space and construct a sample classification boundary, and obtain an abnormal sample detection result; The quality labeling module is used to construct a three-dimensional grid model based on the geometric size data, internal structure data and surface morphology data of building materials, map the abnormal sample detection results into the three-dimensional grid model, establish a digital twin model and label the abnormal area, and generate an abnormal distribution heat map and a quality inspection report.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the building material quality detection method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the building material quality detection method according to any one of claims 1 to 7 are implemented.

Images