Intelligent mechanical property detection method for reinforcing wood structure column of building

By installing strain sensors on the surface of the building wooden structure column and analyzing the timing correlation characteristics of the strain variables using deep learning technology, the problems of strong detection complexity and subjectivity in the existing technology are solved, and the intelligence of building wooden structure columns is realized Mechanical performance detection improves detection accuracy and safety.

CN119043886APending Publication Date: 2024-11-29NANYANG NORMAL UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411065776.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-05
Publication Date
2024-11-29

AI Technical Summary

Technical Problem

In the prior art, the mechanical properties detection of building wooden structure columns has problems such as strong subjectivity and complex operation, and it is difficult to effectively monitor and evaluate the structural properties of the columns, resulting in safety hazards.

Method used

By installing strain sensors on the surface of the column of the building wooden structure, the strain value is monitored in real time, and using deep learning artificial intelligence technology to analyze the timing correlation characteristics between the strain value, we can judge whether reinforcement is needed.

Benefits of technology

Intelligent detection of the mechanical properties of the cylindrical body of building wooden structures is realized, which reduces subjectivity, improves the accuracy and efficiency of detection, and can promptly identify structural problems and take reinforcement measures, improving safety.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119043886A_ABST
    Figure CN119043886A_ABST
Patent Text Reader

Abstract

The invention discloses a mechanical property intelligent detection method for building wood structure column reinforcement. The method comprises the following steps: firstly, acquiring dependent variables of a plurality of preset time points in a preset time period acquired by a plurality of strain sensors mounted on the surface of a building wood structure main body, and then, carrying out data preprocessing on the dependent variables of the plurality of preset time points in the preset time period acquired by the strain sensors to obtain a sequence of dependent variable time sequence images; the method comprises the following steps: firstly, acquiring a dependent variable time sequence image, then extracting dependent variable time sequence correlation characteristics in the sequence of the dependent variable time sequence image to obtain an autocorrelation significant dependent variable time sequence correlation characteristic graph, and finally, determining whether the building wood structure main body needs to be reinforced or not based on the autocorrelation significant dependent variable time sequence correlation characteristic graph. In this way, the time sequence correlation characteristics between the dependent variables can be analyzed based on the artificial intelligence technology of deep learning, and whether the column needs to be reinforced or not is judged.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of intelligent detection, and more specifically, to an intelligent detection method for mechanical properties of reinforced building wood structure columns. Background Art

[0002] Due to factors such as natural aging, environmental erosion, and human damage, the structural performance of wooden columns in buildings will gradually decline, and may even cause deformation, cracks, or other mechanical performance problems, posing safety hazards. Therefore, regular inspection and reinforcement of wooden columns in buildings is an important measure to protect and repair such buildings. Currently, the commonly used inspection method is mainly based on manual observation, but this inspection method has problems such as strong subjectivity and complex operation.

[0003] Therefore, an optimized mechanical properties testing scheme for strengthening building wood structure columns is expected. Summary of the invention

[0004] In order to solve the above technical problems, this application is proposed. The embodiment of this application provides a mechanical property intelligent detection method for the reinforcement of building wood structure columns, which can install strain sensors on the surface of the main body of the building wood structure to monitor the strain of the column in real time, and analyze the time series correlation characteristics between the strains through artificial intelligence technology based on deep learning, so as to judge whether the column needs to be reinforced.

[0005] According to one aspect of the present application, a method for intelligently detecting mechanical properties of building wood structure columns for reinforcement is provided, which comprises: Acquire strain amounts at a plurality of predetermined time points within a predetermined time period collected by a plurality of strain sensors installed on the surface of the main body of the building wood structure; Performing data preprocessing on the strain quantities at a plurality of predetermined time points within a predetermined time period collected by each strain sensor to obtain a sequence of time-series images of the strain quantities; Extracting strain variable time series correlation features in the sequence of strain variable time series images to obtain an autocorrelation significant strain variable time series correlation feature map; and Based on the time series correlation characteristic diagram of the autocorrelation significant strain variable, it is determined whether the main body of the building wooden structure needs to be reinforced.

[0006] Compared with the prior art, the intelligent detection method for mechanical properties of building wood structure columns provided by the present application first obtains the strains at multiple predetermined time points within a predetermined time period collected by multiple strain sensors installed on the surface of the main body of the building wood structure, then performs data preprocessing on the strains at multiple predetermined time points within a predetermined time period collected by each strain sensor to obtain a sequence of strain time series images, then extracts the strain time series correlation features in the sequence of strain time series images to obtain an autocorrelated significant strain time series correlation feature map, and finally, based on the autocorrelated significant strain time series correlation feature map, determines whether the main body of the building wood structure needs to be reinforced. In this way, the time series correlation features between strains can be analyzed based on deep learning artificial intelligence technology to determine whether the column needs to be reinforced. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following briefly introduces the drawings required for use in the description of the embodiments. The following drawings are not intentionally scaled to the actual size, and the focus is on illustrating the main purpose of the present application.

[0008] Figure 1 The present invention is a flow chart of a method for intelligently detecting mechanical properties of building wood structure columns according to an embodiment of the present application.

[0009] Figure 2 It is a schematic diagram of the architecture of the intelligent detection method of mechanical properties for reinforcing building wood structure columns according to an embodiment of the present application.

[0010] Figure 3 This is a flowchart of sub-step S120 of the intelligent detection method of mechanical properties for reinforcing building wood structure columns according to an embodiment of the present application.

[0011] Figure 4 This is a flowchart of sub-step S130 of the intelligent detection method of mechanical properties for reinforcing building wood structure columns according to an embodiment of the present application.

[0012] Figure 5 It is a block diagram of a mechanical property intelligent detection system for reinforcing building wood structure columns according to an embodiment of the present application.

[0013] Figure 6 This is a diagram of an application scenario of the intelligent detection method of mechanical properties for reinforcing building wood structure columns according to an embodiment of the present application. DETAILED DESCRIPTION

[0014] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, not all embodiments. Based on the embodiments of the present application, all other embodiments obtained by ordinary technicians in this field without creative work also fall within the scope of protection of the present application.

[0015] As shown in this application and claims, unless the context clearly indicates an exception, the words "a", "an", "an" and / or "the" do not refer to the singular and may also include the plural. Generally speaking, the terms "include" and "comprise" only indicate the inclusion of the steps and elements that have been clearly identified, and these steps and elements do not constitute an exclusive list. The method or device may also include other steps or elements.

[0016] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, any number of different modules can be used and run on the user terminal and / or server. The modules are only illustrative, and different aspects of the system and method can use different modules.

[0017] Flowcharts are used in the present application to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed accurately in order. On the contrary, various steps may be processed in reverse order or simultaneously as required. Meanwhile, other operations may also be added to these processes, or a certain step or several steps of operations may be removed from these processes.

[0018] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0019] In response to the above technical problems, the technical concept of the present application is to install strain sensors on the surface of the main body of the building's wooden structure to monitor the strain of the column in real time, and to analyze the time series correlation characteristics between the strains through artificial intelligence technology based on deep learning, so as to determine whether the column needs to be reinforced.

[0020] Based on this, Figure 1 The present invention is a flow chart of a method for intelligently detecting mechanical properties of building wood structure columns according to an embodiment of the present application. Figure 2 FIG. 1 is a schematic diagram of the structure of a mechanical property intelligent detection method for strengthening a building wood structure column according to an embodiment of the present application. Figure 1 and Figure 2As shown, according to the embodiment of the present application, the intelligent detection method of mechanical properties for strengthening the columns of building wood structures includes the following steps: S110, obtaining strain quantities at multiple predetermined time points within a predetermined time period collected by multiple strain sensors installed on the surface of the main body of the building wood structure; S120, performing data preprocessing on the strain quantities at multiple predetermined time points within the predetermined time period collected by each strain sensor to obtain a sequence of strain quantity time series images; S130, extracting the strain quantity time series correlation features in the sequence of strain quantity time series images to obtain an autocorrelation significant strain quantity time series correlation feature map; and, S140, determining whether the main body of the building wood structure needs to be reinforced based on the autocorrelation significant strain quantity time series correlation feature map.

[0021] It should be understood that the purpose of step S110 is to collect strain data within a predetermined time period by installing multiple strain sensors on the surface of the main body of the building wood structure. The strain sensors can measure the strain of the wood structure column at different time points, and these data are used for analysis and processing in subsequent steps. The purpose of step S120 is to pre-process the strain data collected from different strain sensors to obtain a time-series image sequence of the strain variable. The data pre-processing may include operations such as denoising, filtering, and correction to ensure the quality and accuracy of the data used in subsequent analysis. The purpose of step S130 is to extract the time-series correlation features of the strain variable from the time-series image sequence of the strain variable. By analyzing the correlation between the strain variables, the time-series correlation features of the strain variable with significant autocorrelation can be identified. These features may include peak values, periodicity, change trends, etc., which are used for decision-making and judgment in subsequent steps. The purpose of step S140 is to determine whether the main body of the building's wooden structure needs to be reinforced based on the time-series correlation characteristic diagram of the strain variable with significant autocorrelation. By analyzing the information in the characteristic diagram, the mechanical properties of the main body of the wooden structure can be evaluated and whether reinforcement is needed. Based on the anomalies or trends in the characteristic diagram, potential structural problems or damages can be identified, and appropriate reinforcement measures can be taken.

[0022] Specifically, in the technical solution of the present application, the strain quantities at multiple predetermined time points within a predetermined time period collected by multiple strain sensors installed on the surface of the main body of the building wood structure are first obtained. Then, the strain quantities at multiple predetermined time points within a predetermined time period collected by each strain sensor are subjected to data preprocessing to obtain a sequence of strain quantity time series images. That is, the strain quantities at multiple predetermined time points within a predetermined time period collected by each discretely distributed strain sensor are converted into image data by means of data preprocessing, so as to represent the information of the time dimension in the spatial dimension, and to convert the one-dimensional single time series relationship representation into the time series association representation between spatial neighborhoods.

[0023] In a specific example of this application, Figure 3As shown, the encoding process of performing data preprocessing on the strain variables at multiple predetermined time points within a predetermined time period collected by each strain sensor to obtain a sequence of strain variable time series images includes: S121, arranging the strain variables at multiple predetermined time points within a predetermined time period collected by each strain sensor into strain variable time series input vectors according to the time dimension to obtain a sequence of strain variable time series input vectors; and, S122, passing the sequence of strain variable time series input vectors through a vector-image converter to obtain a sequence of strain variable time series images.

[0024] It should be understood that the purpose of step S121 is to arrange the strain data collected from each strain sensor according to the time dimension to form a strain time series input vector, each strain time series input vector contains the strain measurement value at different time points, which is used to construct a strain time series image. The purpose of step S122 is to use a vector-image converter to convert the sequence of strain time series input vectors into a sequence of strain time series images. The vector-image converter is a technology that converts vector data into image data. It can convert the strain time series input vector into a corresponding time series image. These time series images can more intuitively represent the change of the strain over time, providing a basis for the feature extraction and analysis of the subsequent steps. In general, these two steps organize and convert the strain data collected by each strain sensor to obtain a strain time series image sequence. These time series images will be used for feature extraction and analysis in the subsequent steps to evaluate the mechanical properties of the building wood structure and determine whether reinforcement is required.

[0025] Wherein, in step S122, the sequence of strain variable time series input vectors is respectively passed through a vector-image converter to obtain a sequence of strain variable time series images, including: vector segmentation of the sequence of strain variable time series input vectors is respectively performed to obtain a sequence of strain variable time series input sub-vectors; the sequence of strain variable time series input sub-vectors is respectively arranged into a sequence of strain variable time series input matrices; and, normalization processing is performed on the sequence of strain variable time series input matrices to obtain a sequence of strain variable time series images; wherein the range of values ​​at each position in the sequence of strain variable time series images is 0-255.

[0026] It should be understood that for two-dimensional strain variable time series images, the network model based on deep learning has a natural advantage, that is, it can capture the neighborhood correlation feature pattern in the spatial dimension, thereby capturing and representing the changing trend of the mechanical properties of the wooden structure column over time. Therefore, in the technical solution of the present application, it is expected to extract the strain variable time series correlation features in the sequence of strain variable time series images to obtain an autocorrelated significant strain variable time series correlation feature map. Here, the sequence of strain variable time series images contains the stress distribution and changes of the main body of the building's wooden structure at different time points and different spatial positions. Extracting the strain variable time series correlation features therefrom can reflect the mechanical properties and reinforcement requirements of the main body of the building's wooden structure.

[0027] In a specific example of this application, Figure 4 As shown, the encoding process of extracting the strain variable time series correlation features in the sequence of the strain variable time series images to obtain the autocorrelation significant strain variable time series correlation feature map includes: S131, using a deep learning network model to extract features from the sequence of the strain variable time series images to obtain the strain variable time series correlation feature map; and S132, passing the strain variable time series correlation feature map through a feature map autocorrelation significant module to obtain the autocorrelation significant strain variable time series correlation feature map. Among them, the feature map autocorrelation significant module can construct the similarity correlation relationship between each pixel in the feature map, thereby aggregating the complete feature distribution of whether the main body of the building wooden structure needs to be reinforced.

[0028] It should be understood that the purpose of step S131 is to use the deep learning network model to extract features from the time series image of the corresponding variable. The deep learning model can learn the key features in the image, such as texture, shape, edge, etc., so as to convert the time series image of the strain variable into a higher-level feature representation, that is, the time series correlation feature map of the strain variable, which can capture the time series correlation information between the strain variables. The purpose of step S132 is to process the time series correlation feature map of the corresponding variable using the feature map autocorrelation significant module to obtain the autocorrelation significant strain variable time series correlation feature map. The autocorrelation significant module can highlight the area with a strong correlation with the time series of the strain variable by calculating the autocorrelation between different positions in the feature map. This helps to extract the important features in the time series image of the strain variable, which is further used to determine whether the main body of the building wood structure needs to be reinforced. In general, these two steps combine the deep learning network model and the autocorrelation significant module to convert the original time series image of the strain variable into the autocorrelation significant strain variable time series correlation feature map to extract the key time series correlation features and use them to determine whether the main body of the building wood structure needs to be reinforced.

[0029] Among them, in step S131, the sequence of strain variable time series images is feature extracted by using a deep learning network model to obtain a strain variable time series correlation feature map, including: the sequence of strain variable time series images is passed through a strain variable time series correlation feature extractor based on a three-dimensional convolutional neural network model to obtain the strain variable time series correlation feature map. It is worth mentioning that a three-dimensional convolutional neural network (3D Convolutional Neural Network) is a deep learning model used to process three-dimensional data with a time dimension. Unlike a traditional two-dimensional convolutional neural network, a three-dimensional convolutional neural network can simultaneously consider spatial and temporal features. In step S131, a strain variable time series correlation feature extractor based on a three-dimensional convolutional neural network model is used to extract features from a sequence of corresponding variable time series images to obtain a strain variable time series correlation feature map. The three-dimensional convolutional neural network model here can learn the time series correlation features in the strain variable time series image, including patterns and structures in space and time. By using a three-dimensional convolutional neural network model, the time series correlation information in the strain variable time series image sequence can be captured. These features are very important for subsequent analysis and decision-making, such as determining the deformation of the building wood structure, detecting abnormal behavior, predicting future change trends, etc. In summary, the three-dimensional convolutional neural network model is used in step S131 to extract the temporal correlation features of the strain variable time series image sequence, thereby providing more accurate and comprehensive information for subsequent analysis and decision-making.

[0030] Among them, in step S131, the strain variable time series correlation feature map is passed through the feature map autocorrelation significance module to obtain the autocorrelation significant strain variable time series correlation feature map, including: passing the strain variable time series correlation feature map through the first convolution layer of the feature autocorrelation association enhancement module to obtain a first feature map; passing the first feature map through the second convolution layer of the feature autocorrelation association enhancement module to obtain a second feature map; expanding each feature matrix of the second feature map along the channel dimension into feature vectors to obtain a sequence of first feature vectors; calculating the cosine similarity between any two first feature vectors in the sequence of the first feature vectors to obtain a cosine similarity feature map; passing the cosine similarity feature map through the Softmax function Perform normalization processing to obtain a normalized cosine similarity feature map; multiply the normalized cosine similarity feature map and the cosine similarity feature map by position point to obtain a similarity mapping optimization feature map; pass the similarity mapping optimization feature map through the first deconvolution layer of the feature autocorrelation association enhancement module to obtain a first deconvolution feature map; calculate the element-by-element sum of the first deconvolution feature map and the first feature map to obtain a first fused feature map; pass the first fused feature map through the second deconvolution layer of the feature autocorrelation association enhancement module to obtain a second deconvolution feature map; and calculate the element-by-element sum of the second deconvolution feature map and the strain variable time series association feature map to obtain the autocorrelation significant strain variable time series association feature map.

[0031] Then, the time series correlation characteristic diagram of the autocorrelation significant strain variable is passed through a classifier to obtain a classification result, and the classification result is used to indicate whether the main body of the building wooden structure needs to be reinforced.

[0032] Accordingly, based on the time-series correlation feature graph of the autocorrelation significant strain variable, determining whether the main body of the wooden structure of the building needs to be reinforced includes: passing the time-series correlation feature graph of the autocorrelation significant strain variable through a classifier to obtain a classification result, and the classification result is used to indicate whether the main body of the wooden structure of the building needs to be reinforced. In a specific example of the present application, passing the time-series correlation feature graph of the autocorrelation significant strain variable through a classifier to obtain a classification result, and the classification result is used to indicate whether the main body of the wooden structure of the building needs to be reinforced, including: expanding the time-series correlation feature graph of the autocorrelation significant strain variable into a corrected classification feature vector according to a row vector or a column vector; using the fully connected layer of the classifier to fully connect encode the corrected classification feature vector to obtain an encoded classification feature vector; and inputting the encoded classification feature vector into the Softmax classification function of the classifier to obtain the classification result.

[0033] That is, in the technical solution of the present application, the labels of the classifier include that the main body of the building's wooden structure needs to be reinforced (first label), and that the main body of the building's wooden structure does not need to be reinforced (second label), wherein the classifier determines which classification label the time series correlation feature map of the autocorrelation significant dependent variable belongs to by means of a soft maximum function. It is worth noting that the first label p1 and the second label p2 here do not contain artificially set concepts. In fact, during the training process, the computer model does not have the concept of "whether the main body of the building's wooden structure needs to be reinforced". It only has two classification labels and the probability of the output feature under these two classification labels, that is, the sum of p1 and p2 is one. Therefore, the classification result of whether the main body of the building's wooden structure needs to be reinforced is actually converted into a binary class probability distribution that conforms to natural laws through the classification label. In essence, the physical meaning of the natural probability distribution of the label is used, rather than the linguistic text meaning of "whether the main body of the building's wooden structure needs to be reinforced".

[0034] It should be understood that the role of the classifier is to use the given categories and known training data to learn classification rules and classifiers, and then classify (or predict) unknown data. Logistic regression, SVM, etc. are often used to solve binary classification problems. For multi-class classification problems, logistic regression or SVM can also be used, but multiple binary classifications are required to form a multi-classification, but this is prone to errors and inefficient. Commonly used multi-classification methods include the Softmax classification function.

[0035] It is worth mentioning that the fully connected layer, also known as the densely connected layer or fully connected layer, is a common layer type in deep learning neural networks. In the fully connected layer, each neuron is connected to all neurons in the previous layer to form a fully connected graph structure. The function of the fully connected layer is to combine and transform the feature information of the previous layer to generate a new feature representation. Each neuron has a set of weights and biases in the fully connected layer for linearly combining input features. Then, the linear combination result is nonlinearly transformed by the activation function to introduce nonlinear relationships and complex feature representation capabilities. In step S142, the fully connected layer is used to fully connect encode the corrected classification feature vector to obtain an encoded classification feature vector. The fully connected layer linearly combines the corrected feature vector with the weight of each neuron and performs a nonlinear transformation by the activation function. This process can map the feature vector to a higher-dimensional feature space and extract richer and more abstract feature representations. After being processed by the fully connected layer, the encoded classification feature vector can contain more information about whether the main body of the building's wooden structure needs to be reinforced. Finally, the encoded classification feature vector is input into the Softmax classification function of the classifier, which can map the feature vector to the probability distribution of different categories to obtain the final classification result. The fully connected layer plays the role of feature conversion and feature extraction in this process, helping the classifier to make accurate classification decisions.

[0036] In a specific example of the present application, the intelligent detection method for mechanical properties of building wood structure columns reinforcement also includes: a training step for training the vector-to-image converter, the strain variable time series correlation feature extractor based on the three-dimensional convolutional neural network model, the feature map autocorrelation significance module and the classifier; wherein the training step includes: obtaining training data, the training data includes: training strain variables at multiple predetermined time points within a predetermined time period collected by multiple strain sensors; arranging the training strain variables at multiple predetermined time points within a predetermined time period collected by each strain sensor into training strain variable time series input vectors according to the time dimension to obtain a sequence of training strain variable time series input vectors; passing the sequence of training strain variable time series input vectors through the vector-to-image converter to obtain a sequence of training strain variable time series images; A sequence of time-series images is passed through a strain variable temporal correlation feature extractor based on a three-dimensional convolutional neural network model to obtain a training strain variable temporal correlation feature map; the training strain variable temporal correlation feature map is passed through a feature map autocorrelation significance module to obtain a training autocorrelation significant strain variable temporal correlation feature map; the training autocorrelation significant strain variable temporal correlation feature map is passed through a classifier to obtain a classification loss function value; based on the training autocorrelation significant strain variable temporal correlation feature map, an autocorrelation significant strain variable temporal correlation loss function value is calculated; a weighted sum of the classification loss function value and the autocorrelation significant strain variable temporal correlation loss function value is used as the final loss function value and the vector-image converter, the strain variable temporal correlation feature extractor based on a three-dimensional convolutional neural network model, the feature map autocorrelation significance module and the classifier are trained through the propagation direction of gradient descent.

[0037] Here, after the sequence of the training strain variable time series images passes through the strain variable inter-temporal correlation feature extractor based on the three-dimensional convolutional neural network model, each feature matrix of the training strain variable time series correlation feature map obtained expresses the local intra-local time domain-inter-local time domain strain variable time series correlation characteristics of the local time domain determined after the global time domain through vector-image conversion, and the various feature matrices follow the cross-sample correlation based on the three-dimensional convolutional neural network model. Then, after the training strain variable time series correlation feature map passes through the feature map autocorrelation significance module, the channel dimension of the training strain variable time series correlation feature map can be autocorrelated based on the local intra-local time domain-inter-local time domain temporal correlation feature expression of each feature matrix, so that the channel correlation of the training autocorrelation significant strain variable time series correlation feature map is constrained by the temporal correlation. In this way, the training autocorrelation significant response variable time series correlation feature map has complex time series correlation distribution properties across time domain scales in both the spatial distribution dimension within the feature matrix and the channel distribution dimension between feature matrices. This makes the training autocorrelation significant response variable time series correlation feature map have a relatively obvious problem of insufficient understanding of the complex feature correlation distribution expression classification regression, affecting the accuracy of the classification results obtained through the classifier.

[0038] Based on this, the present application further introduces a loss function for promoting classification understanding based on the complex feature expression of the training autocorrelation significant strain variable time series correlation feature map, that is, based on the training autocorrelation significant strain variable time series correlation feature map, the autocorrelation significant strain variable time series correlation loss function value is calculated. Wherein, based on the training autocorrelation significant strain variable time series correlation feature map, the autocorrelation significant strain variable time series correlation loss function value is calculated, including: first, by calculating the training autocorrelation significant strain variable time series correlation feature vector obtained by expanding the training autocorrelation significant strain variable time series correlation feature map, for example, recorded as The mean and variance between the eigenvalues ​​at any two positions of get the mean weight matrix and the variance weight matrix , and then the time series correlation feature vector of the training autocorrelation significant response variable Respectively with the mean weight matrix and the variance weight matrix After performing query multiplication, the autocorrelation matrix of the two eigenvectors is calculated. norm, and subtract the training autocorrelation significant dependent variable time series correlation eigenvector The autocorrelation matrix of The product of the norm and the weight as a hyperparameter is used to obtain the value of the autocorrelation significant response variable time series association loss function.

[0039] For example, expressed as in, represents the training autocorrelation significant dependent variable time series correlation feature vector obtained after the training autocorrelation significant dependent variable time series correlation feature map is expanded, is the mean weight matrix, is the variance weight matrix, represents the transpose of a vector, represents the matrix multiplication operation, is the weight as a hyperparameter, Represents the matrix norm, represents the time series correlation loss function value of the autocorrelation significant dependent variable, and is as described above: in, The first eigenvalues, The first eigenvalues, represents the mean weight matrix The eigenvalues ​​of the positions, represents the variance weight matrix The eigenvalue of the location.

[0040] The autocorrelation significant strain variable time-series association loss function value determines the overall distributed structural bottleneck of the training autocorrelation significant strain variable time-series association feature graph through multi-dimensional interaction of the detail group of the training autocorrelation significant strain variable time-series association feature graph, and simulates the correlation between the structural details of the training autocorrelation significant strain variable time-series association feature graph and the macro-structured feature representation behavior through low-rank structured reasoning of the training autocorrelation significant strain variable time-series association feature graph. In this way, by training the classifier with the autocorrelation significant strain variable time-series association loss function value, the structural detail dependency relative to the macro-complexity can be modeled in the classification scenario, thereby promoting the understanding of the macro-complex feature expression based on the structural details of the training autocorrelation significant strain variable time-series association feature graph, and improving the accuracy of the classification results.

[0041] In summary, the intelligent detection method of mechanical properties for strengthening building wood structure columns based on the embodiment of the present application is explained, which can analyze the time series correlation characteristics between strain variables based on deep learning artificial intelligence technology to determine whether the column needs to be reinforced.

[0042] Figure 5 FIG. 1 is a block diagram of a mechanical property intelligent detection system 100 for strengthening a building wood structure column according to an embodiment of the present application. Figure 5As shown, according to the embodiment of the present application, the intelligent mechanical property detection system 100 for strengthening the building wood structure column includes: a data acquisition module 110, which is used to obtain the strain quantity at multiple predetermined time points within a predetermined time period collected by multiple strain sensors installed on the surface of the building wood structure body; a data preprocessing module 120, which is used to perform data preprocessing on the strain quantity at multiple predetermined time points within the predetermined time period collected by each strain sensor to obtain a sequence of strain quantity time series images; a feature extraction module 130, which is used to extract the strain quantity time series correlation features in the sequence of strain quantity time series images to obtain an autocorrelation significant strain quantity time series correlation feature map; and a building wood structure body analysis module 140, which is used to determine whether the building wood structure body needs to be reinforced based on the autocorrelation significant strain quantity time series correlation feature map.

[0043] In one example, in the above-mentioned intelligent mechanical property detection system 100 for strengthening building wood structure columns, the data preprocessing module 120 includes: a vectorization unit, which is used to arrange the strain variables at multiple predetermined time points within a predetermined time period collected by each strain sensor into strain variable time series input vectors according to the time dimension to obtain a sequence of strain variable time series input vectors; and a conversion unit, which is used to pass the sequence of strain variable time series input vectors through a vector-to-image converter to obtain a sequence of strain variable time series images.

[0044] Here, those skilled in the art can understand that the specific functions and operations of each module in the above-mentioned intelligent mechanical property detection system 100 for strengthening building wood structure columns have been referred to above. Figures 1 to 5 The intelligent detection method of mechanical properties for strengthening building wooden structure columns has been introduced in detail, and therefore, its repeated description will be omitted.

[0045] As described above, the intelligent detection system 100 for mechanical properties of building wood structure columns reinforcement according to the embodiment of the present application can be implemented in various wireless terminals, such as a server having an intelligent detection algorithm for mechanical properties of building wood structure columns reinforcement. In one example, the intelligent detection system 100 for mechanical properties of building wood structure columns reinforcement according to the embodiment of the present application can be integrated into the wireless terminal as a software module and / or a hardware module. For example, the intelligent detection system 100 for mechanical properties of building wood structure columns reinforcement can be a software module in the operating system of the wireless terminal, or can be an application developed for the wireless terminal; of course, the intelligent detection system 100 for mechanical properties of building wood structure columns reinforcement can also be one of the many hardware modules of the wireless terminal.

[0046] Alternatively, in another example, the intelligent detection system 100 for mechanical properties of building wooden structure columns and the wireless terminal may be separate devices, and the intelligent detection system 100 for mechanical properties of building wooden structure columns may be connected to the wireless terminal via a wired and / or wireless network and transmit interactive information in accordance with an agreed data format.

[0047] Figure 6 FIG. 1 is an application scenario diagram of the mechanical properties intelligent detection method for strengthening building wood structure columns according to an embodiment of the present application. Figure 6 As shown, in this application scenario, first, the strain values ​​(for example, Figure 6 D), then, the strain values ​​collected by each strain sensor at a plurality of predetermined time points within a predetermined time period are input into a server deployed with a mechanical property intelligent detection algorithm for strengthening building wood structure columns (for example, Figure 6 S), wherein the server is able to use the intelligent detection algorithm for mechanical properties of building wood structure columns to process the strain quantities at multiple predetermined time points within a predetermined time period collected by each strain sensor to obtain a classification result indicating whether the main body of the building wood structure needs to be reinforced.

[0048] In addition, it will be appreciated by those skilled in the art that various aspects of the present application may be illustrated and described by a number of patentable categories or situations, including any new and useful process, machine, product or combination of substances, or any new and useful improvements thereto. Accordingly, various aspects of the present application may be performed entirely by hardware, entirely by software (including firmware, resident software, microcode, etc.), or by a combination of hardware and software. The above hardware or software may all be referred to as "data blocks", "modules", "engines", "units", "components" or "systems". In addition, various aspects of the present application may be represented as a computer product located in one or more computer-readable media, which includes computer-readable program code.

[0049] Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the invention belongs. It should also be understood that terms such as those defined in common dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant technology and should not be interpreted in an idealized or extremely formal sense, unless explicitly defined as such herein.

[0050] The above is an explanation of the present invention and should not be considered as a limitation thereof. Although several exemplary embodiments of the present invention have been described, it will be readily appreciated by those skilled in the art that many modifications may be made to the exemplary embodiments without departing from the novel teachings and advantages of the present invention. Therefore, all such modifications are intended to be included within the scope of the present invention as defined in the claims. It should be understood that the above is an explanation of the present invention and should not be considered as being limited to the specific embodiments disclosed, and modifications to the disclosed embodiments and other embodiments are intended to be included within the scope of the appended claims. The present invention is defined by the claims and their equivalents.

Claims

1. A mechanical properties intelligent detection method for strengthening building wood structure columns, characterized in that: include: Acquire strain amounts at a plurality of predetermined time points within a predetermined time period collected by a plurality of strain sensors installed on the surface of the main body of the building wood structure; Performing data preprocessing on the strain quantities at a plurality of predetermined time points within a predetermined time period collected by each strain sensor to obtain a sequence of time-series images of the strain quantities; Extracting strain variable time series correlation features in the sequence of strain variable time series images to obtain an autocorrelation significant strain variable time series correlation feature map; as well as Based on the time series correlation characteristic diagram of the autocorrelation significant strain variable, it is determined whether the main body of the building wooden structure needs to be reinforced.

2. The intelligent detection method for mechanical properties of building wood structure column reinforcement according to claim 1 is characterized in that: The method of preprocessing the strain quantities at a plurality of predetermined time points within a predetermined time period collected by each strain sensor to obtain a sequence of time-series images of the strain quantities comprises: Arranging the strain quantities at a plurality of predetermined time points within a predetermined time period collected by each strain sensor into strain quantity time series input vectors according to the time dimension to obtain a sequence of strain quantity time series input vectors; and The sequence of the strain variable time series input vectors is respectively passed through a vector-to-image converter to obtain a sequence of the strain variable time series images.

3. The intelligent detection method for mechanical properties of building wood structure column reinforcement according to claim 2 is characterized in that: The sequence of strain variable time series input vectors is respectively passed through a vector-image converter to obtain a sequence of strain variable time series images, including: Performing vector segmentation on the sequence of strain variable time series input vectors to obtain a sequence of strain variable time series input sub-vectors; Arranging the sequences of the dependent variable time series input sub-vectors into sequences of dependent variable time series input matrices respectively; and Normalizing the sequence of the strain variable time series input matrices respectively to obtain the sequence of the strain variable time series images; The value of each position in the sequence of the strain variable time series image ranges from 0 to 255.

4. The intelligent detection method for mechanical properties of building wood structure column reinforcement according to claim 3 is characterized in that: Extracting the strain variable time series correlation features in the sequence of the strain variable time series images to obtain an autocorrelated significant strain variable time series correlation feature map, including: Using a deep learning network model to extract features from the sequence of strain variable time series images to obtain a strain variable time series correlation feature map; and The strain variable time series correlation feature map is passed through a feature map autocorrelation significant module to obtain the autocorrelation significant strain variable time series correlation feature map.

5. The intelligent detection method for mechanical properties of building wood structure column reinforcement according to claim 4 is characterized in that: Using a deep learning network model to extract features from the sequence of strain variable time series images to obtain a strain variable time series correlation feature map, including: The sequence of the strain variable time series images is passed through a strain variable time series correlation feature extractor based on a three-dimensional convolutional neural network model to obtain the strain variable time series correlation feature map.

6. The intelligent detection method for mechanical properties of building wood structure column reinforcement according to claim 5 is characterized in that: The strain variable time series correlation feature map is passed through a feature map autocorrelation significant module to obtain the autocorrelation significant strain variable time series correlation feature map, including: Passing the strain variable time series correlation feature map through the first convolution layer of the feature autocorrelation correlation enhancement module to obtain a first feature map; Passing the first feature map through the second convolutional layer of the feature autocorrelation association enhancement module to obtain a second feature map; Expanding each feature matrix of the second feature map along the channel dimension into feature vectors to obtain a sequence of first feature vectors; Calculating the cosine similarity between any two first feature vectors in the sequence of the first feature vectors to obtain a cosine similarity feature graph; The cosine similarity feature map is normalized by a Softmax function to obtain a normalized cosine similarity feature map; Multiplying the normalized cosine similarity feature map and the cosine similarity feature map by position points to obtain a similarity mapping optimization feature map; Passing the similarity map optimization feature map through the first deconvolution layer of the feature autocorrelation association enhancement module to obtain a first deconvolution feature map; Calculating an element-by-element sum of the first deconvolution feature map and the first feature map to obtain a first fused feature map; Passing the first fused feature map through the second deconvolution layer of the feature autocorrelation association enhancement module to obtain a second deconvolution feature map; and The element-by-element sum of the second deconvolution feature map and the strain variable time series correlation feature map is calculated to obtain the autocorrelation significant strain variable time series correlation feature map.

7. The intelligent detection method for mechanical properties of building wood structure column reinforcement according to claim 6 is characterized in that: Determining whether the main body of the building wood structure needs to be reinforced based on the time series correlation characteristic diagram of the autocorrelation significant strain variable includes: The time series correlation characteristic diagram of the autocorrelation significant strain variable is passed through a classifier to obtain a classification result, and the classification result is used to indicate whether the main body of the building wooden structure needs to be reinforced.

8. The intelligent detection method for mechanical properties of building wood structure column reinforcement according to claim 7 is characterized in that: Also includes: A training step for training the vector-to-image converter, the temporal correlation feature extractor between dependent variables based on the three-dimensional convolutional neural network model, the feature map autocorrelation significant module and the classifier; Wherein, the training step includes: Acquiring training data, the training data comprising: training strain quantities at a plurality of predetermined time points within a predetermined time period collected by a plurality of strain sensors; Arranging the training strain variables at a plurality of predetermined time points within a predetermined time period collected by each strain sensor into training strain variable time series input vectors according to the time dimension to obtain a sequence of training strain variable time series input vectors; Passing the sequence of the training strain variable time series input vectors through a vector-to-image converter to obtain a sequence of training strain variable time series images; The sequence of the training strain variable time series images is passed through a strain variable time series correlation feature extractor based on a three-dimensional convolutional neural network model to obtain a training strain variable time series correlation feature map; The training strain variable time series correlation feature map is passed through a feature map autocorrelation significance module to obtain a training autocorrelation significant strain variable time series correlation feature map; Passing the training autocorrelation significant strain variable time series correlation feature map through a classifier to obtain a classification loss function value; Based on the training autocorrelation significant strain variable time series correlation feature map, calculating the autocorrelation significant strain variable time series correlation loss function value; The weighted sum of the classification loss function value and the autocorrelation significant strain time series correlation loss function value is used as the final loss function value and the vector-to-image converter, the strain time series correlation feature extractor based on the three-dimensional convolutional neural network model, the feature map autocorrelation significant module and the classifier are trained through the propagation direction of gradient descent.