Method and system for identifying the quality of a surface of a building material

Abstract

The application discloses a building material surface quality identification method and system, and relates to the field of material identification, and comprises the following steps: step 1: according to the type of target building material and its quality requirement in a preset construction scene, a plurality of surface feature collection indexes are defined in advance; step 2: based on the constructed collection index set, the surface data of the current building material sample is collected, the surface data is preprocessed and feature extraction is performed, and the current surface feature value corresponding to each collection index is output; step 3: each current surface feature value obtained is compared with the standard feature value range corresponding to the pre-stored qualified building material of the same type; the application can output the deterioration coefficient in the future preset construction scene, realize the quantitative prediction of the long-term performance degradation risk of the material, and allow the construction party and the manager to predict the risk in advance and make the material replacement or reinforcement scheme in advance.

Description

Technical Field

[0001] This invention relates to the field of material identification technology, specifically to a method and system for identifying the surface quality of building materials. Background Technology

[0002] The performance evolution and surface quality degradation of building materials during long-term service have become key factors affecting structural safety, functionality, and life-cycle costs. Accurate identification and assessment of the surface condition of materials such as concrete, steel, and decorative panels during construction, acceptance, and operation and maintenance are crucial for ensuring project quality. Traditionally, this process relies heavily on manual visual inspection or simple physical measurements, which presents challenges in terms of efficiency and traceability. Meanwhile, the rapid development of the Internet of Things (IoT), sensor technology, and artificial intelligence has provided technological possibilities for building a digital and intelligent engineering quality management system.

[0003] Traditional methods rely heavily on the experience and subjective judgment of inspectors, resulting in large fluctuations in assessment results. They are difficult to conduct accurate horizontal comparisons and historical tracing, and have limitations in the depth and efficiency of information processing. Existing methods focus on the immediate discovery and recording of apparent defects, but fail to effectively integrate multi-dimensional feature data. Furthermore, they lack the ability to deeply mine massive amounts of inspection data to reveal the evolutionary patterns of defects. Most existing technical systems can only reflect the current state of materials and cannot scientifically predict and assess the performance degradation trend under specific harsh environments in the future based on current data and historical patterns, which often leads to delayed maintenance decisions. Summary of the Invention

[0004] (a) Technical problems to be solved

[0005] In view of the above-mentioned shortcomings of the existing technology, the present invention provides a method and system for identifying the surface quality of building materials, which can effectively solve the problems of the existing technology.

[0006] (II) Technical Solution

[0007] To achieve the above objectives, the present invention provides the following technical solution:

[0008] This invention discloses a method for identifying the surface quality of building materials, comprising the following steps:

[0009] Step 1: Based on the type of the target building material and its quality requirements under the preset construction scenario, predefine several surface feature acquisition indicators;

[0010] Step 2: Based on the constructed set of acquisition indicators, collect surface data of the current building material sample through an image sensor, laser scanner or hyperspectral imaging device, preprocess the surface data and extract features, and output the current surface feature value corresponding to each acquisition indicator;

[0011] Step 3: Compare each current surface feature value obtained with the standard feature value range corresponding to the pre-stored qualified building materials of the same type. Based on the comparison deviation, calculate the anomaly coefficient of each surface feature relative to the standard state through a preset normalization algorithm. The anomaly coefficient is used to quantitatively characterize the degree of deviation of the feature item.

[0012] Step 4: Integrate the anomaly coefficients of all calculated surface features to form an anomaly feature vector. Associate and store the anomaly feature vector with the current building material identification information, collection time, and environmental parameters to construct an anomaly feature dataset.

[0013] Step 5: Construct a deterioration prediction model. The model uses historical abnormal feature datasets and corresponding subsequent actual deterioration records as training samples, and the current abnormal feature dataset as model input. The deterioration prediction model evaluates and outputs the comprehensive deterioration coefficient of the current building material under a specific preset construction scenario in the future. The deterioration coefficient is used to predict the risk level of material performance degradation.

[0014] Step 6: Based on the deterioration coefficient and the preset risk threshold rules, automatically generate a quality assessment report that includes the current quality status of materials and predictions of future deterioration risks. Submit the report to the construction management end through the communication network to guide the use of materials, replacement, or adjustment of construction plans.

[0015] Furthermore, the set of collected indicators in step 1 includes at least one or more of the following: surface texture complexity, color uniformity, pore distribution density, microcrack length, reflectance spectral characteristics, and local deformation curvature.

[0016] Furthermore, the normalization algorithm in step 3 operates as follows:

[0017] The current surface feature value is compared with the pre-stored corresponding qualified standard feature value range to determine the direction and magnitude of the deviation of the current value from the midpoint of the standard range;

[0018] For a feature value that falls within the standard range, its anomaly coefficient is obtained by calculating the distance of the value from the midpoint of the range and dividing it by the width of the entire standard range. This makes the coefficient a dimensionless value between zero and one, used to characterize its relative deviation from the normal range.

[0019] For feature values ​​that fall outside the standard range, the anomaly coefficient is calculated based on the ratio above, and a penalty factor proportional to the distance beyond the boundary is introduced, so that the final calculated anomaly coefficient is greater than one, thereby quantitatively characterizing the severity of the out-of-standard anomaly. Through this method, surface feature measurements of different dimensions and magnitudes are uniformly converted into comparable anomaly coefficients.

[0020] Furthermore, the environmental parameters in step 4 include the ambient temperature, relative humidity, and light intensity at the time of data collection; the storage format of the abnormal feature dataset is a time series database, used to record the changes in abnormal feature vectors of the same building material at different time points to support trend analysis.

[0021] Furthermore, the construction scenarios preset in step 5 include one or more combinations of long-term high humidity environment, periodic freeze-thaw cycle, strong acid and alkali corrosion environment, or continuous high load stress state.

[0022] Furthermore, the process of constructing the deterioration prediction model in step 5 includes:

[0023] Obtain a historical anomaly feature dataset and corresponding actual deterioration data records; the historical anomaly feature dataset is obtained by processing historical building material samples using the methods in steps 1 to 4, and each data record contains an anomaly feature vector, material identifier, collection time, environmental parameters, and subsequent preset construction scenario identifier; the actual deterioration record is a quantitative deterioration degree label obtained by actual measurement or expert evaluation of the corresponding building material sample after experiencing the subsequent construction scenario.

[0024] A deep learning network is constructed as a deterioration prediction model. This network includes at least one long short-term memory network for capturing the temporal dependencies of anomalous features, and a fully connected branch for fusing environmental parameters and scene labels. The historical anomalous feature dataset and the corresponding actual deterioration records are used as training samples. With the goal of minimizing the error between the predicted deterioration degree and the actual deterioration label, the parameters of the deep learning network are iteratively trained using a time series backpropagation algorithm and a gradient descent optimizer until the model converges.

[0025] Furthermore, the deterioration prediction model organizes the generated current abnormal feature dataset into time series model input data according to the collection time sequence. The organized time series data, current environmental parameters, and target future construction scenario identifiers are input together into the trained and converged deterioration prediction model. The model extracts the evolution pattern of abnormal features through its long short-term memory network, and fuses environmental and scenario information through a fully connected layer. Finally, the output layer calculates and outputs a comprehensive deterioration coefficient that represents future risk.

[0026] Furthermore, the preset logic of the risk threshold rule in step 6 is as follows: based on the numerical range of the comprehensive deterioration coefficient, at least three consecutive risk threshold intervals are pre-divided, corresponding to low risk, medium risk and high risk levels respectively; for low risk level, the recommended measures are periodic re-inspection and routine monitoring; for medium risk level, the recommended measures include shortening the detection cycle, applying temporary protective coating or carrying out local reinforcement; for high risk level, the recommended measures are to immediately replace materials, adjust the structural design or change the construction process parameters.

[0027] Furthermore, the preprocessing in step 2 includes denoising, enhancing, scaling, and image stitching of the collected raw surface data; the feature extraction uses the gray-level co-occurrence matrix algorithm to extract texture features, the color histogram statistical method to extract color distribution features, and edge detection and morphological algorithms to extract crack and pore geometric features.

[0028] This invention also discloses a surface quality identification system for building materials, comprising:

[0029] The indicator definition module is used to predefine several surface feature collection indicators based on the type of target building material and its quality requirements under preset construction scenarios, and form a collection indicator set.

[0030] The feature extraction module collects surface data of the current building material sample based on the collection index set, preprocesses and extracts features from the surface data, and outputs the current surface feature value corresponding to each collection index.

[0031] The anomaly calculation module is used to compare each current surface feature value with the pre-stored standard feature value range corresponding to the same type of qualified building materials. Based on the comparison deviation, the anomaly coefficient of each surface feature is calculated by a preset normalization algorithm.

[0032] The dataset construction module is used to integrate the anomaly coefficients of all surface features into anomaly feature vectors, and associate and store the anomaly feature vectors with the current building material identification information, collection time and environmental parameters to construct anomaly feature datasets;

[0033] The deterioration prediction module is used to build a deterioration prediction model using historical abnormal feature datasets and corresponding subsequent actual deterioration records as training samples. The current abnormal feature dataset is input into the model to evaluate and output the comprehensive deterioration coefficient of the current building materials under a specific preset construction scenario in the future.

[0034] The report generation module is used to automatically generate a quality assessment report containing the current quality status of materials and predictions of future deterioration risks based on the deterioration coefficient and preset risk threshold rules, and submit it to the construction management end through the communication network.

[0035] The indicator definition module and the feature extraction module are interconnected via a wireless network. The feature extraction module and the anomaly calculation module are interconnected via a wireless network. The anomaly calculation module and the dataset construction module are interconnected via a wireless network. The deterioration prediction module, the dataset construction module, and the report generation module are interconnected via a wireless network.

[0036] (III) Beneficial Effects

[0037] Compared with the known prior art, the technical solution provided by this invention has the following beneficial effects:

[0038] 1. By predefining quantifiable acquisition indicators and using sensors and algorithms to automatically extract surface feature values, and by calculating anomaly coefficients and constructing structured datasets, the originally vague qualitative descriptions are transformed into calculable digital indicators, which improves detection efficiency and consistency, provides an objective data basis for quality assessment, and enables accurate comparison and analysis of the quality of materials from different batches and at different times.

[0039] 2. By introducing a deterioration prediction model based on deep learning, this model can learn the complex mapping relationship between surface anomaly features in historical data and subsequent actual performance deterioration. By inputting the current material's anomaly feature dataset, the model can output its deterioration coefficient under the future preset construction scenario, realizing the quantitative prediction of the material's long-term performance degradation risk. This allows construction parties and managers to anticipate risks in advance and formulate material replacement or reinforcement plans in advance. Attached Figure Description

[0040] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0041] Figure 1 This is a flowchart illustrating the method for identifying the surface quality of building materials in this invention.

[0042] Figure 2 This is a schematic diagram of the framework of the building material surface quality identification system of the present invention.

[0043] The labels in the diagram represent: 1. Indicator definition module; 2. Feature extraction module; 3. Anomaly calculation module; 4. Dataset construction module; 5. Deterioration prediction module; 6. Report generation module. Detailed Implementation

[0044] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0045] The present invention will be further described below with reference to embodiments.

[0046] The method for identifying the surface quality of building materials in this embodiment, such as Figure 1 As shown, it includes the following steps:

[0047] Step 1: Based on the type of the target building material and its quality requirements under the preset construction scenario, predefine several surface feature acquisition indicators; the acquisition indicator set includes at least one or more of the following: surface texture complexity, color uniformity, pore distribution density, microcrack length, reflectance spectral characteristics, and local deformation curvature.

[0048] Step 2: Based on the constructed set of acquisition indicators, surface data of the current building material sample is collected using an image sensor, laser scanner, or hyperspectral imaging device. The surface data is preprocessed and features are extracted, and the current surface feature value corresponding to each acquisition indicator is output. The preprocessing includes denoising, enhancement, scale normalization, and image stitching of the collected raw surface data. The feature extraction uses the gray-level co-occurrence matrix algorithm to extract texture features, the color histogram statistical method to extract color distribution features, and edge detection and morphological algorithms to extract the geometric features of cracks and pores.

[0049] Step 3: Compare each current surface feature value with the pre-stored standard feature value range corresponding to the same type of qualified building materials. Based on the comparison deviation, calculate the anomaly coefficient of each surface feature relative to the standard state using a preset normalization algorithm. The anomaly coefficient is used to quantify the degree of deviation of the feature item. The normalization algorithm's operating logic is as follows:

[0050] The current surface feature value is compared with the pre-stored corresponding qualified standard feature value range to determine the direction and magnitude of the deviation of the current value from the midpoint of the standard range;

[0051] For a feature value that falls within the standard range, its anomaly coefficient is obtained by calculating the distance of the value from the midpoint of the range and dividing it by the width of the entire standard range. This makes the coefficient a dimensionless value between zero and one, used to characterize its relative deviation from the normal range.

[0052] For feature values ​​that fall outside the standard range, the anomaly coefficient is calculated based on the ratio above, and a penalty factor proportional to the distance beyond the boundary is introduced, so that the final calculated anomaly coefficient is greater than one, thereby quantitatively characterizing the severity of the out-of-standard anomaly. Through this method, surface feature measurements of different dimensions and magnitudes are uniformly converted into comparable anomaly coefficients.

[0053] Step 4: Integrate the anomaly coefficients of all calculated surface features to form an anomaly feature vector. Associate and store the anomaly feature vector with the current building material identification information, collection time, and environmental parameters to construct an anomaly feature dataset. Environmental parameters include ambient temperature, relative humidity, and light intensity at the time of collection. The anomaly feature dataset is stored in a time series database format to record the changes in the anomaly feature vectors of the same building material at different time points to support trend analysis.

[0054] Step 5: Construct a deterioration prediction model. The model uses historical anomaly feature datasets and corresponding subsequent actual deterioration records as training samples, and the current anomaly feature dataset as input. The deterioration prediction model evaluates and outputs the comprehensive deterioration coefficient of the current building materials under specific preset construction scenarios in the future. The deterioration coefficient is used to predict the risk level of material performance degradation. The preset construction scenarios include one or more combinations of long-term high humidity environment, periodic freeze-thaw cycle, strong acid and alkali corrosion environment, or continuous high load stress state. The construction process of the deterioration prediction model includes:

[0055] Obtain historical anomaly feature datasets and corresponding actual deterioration data records; the historical anomaly feature datasets are obtained by processing historical building material samples using the methods in steps 1 to 4, and each data record contains an anomaly feature vector, material identifier, collection time, environmental parameters, and subsequent preset construction scenario identifier; the actual deterioration records are the quantitative deterioration degree labels obtained by actual measurement or expert evaluation of the corresponding building material samples after experiencing subsequent construction scenarios.

[0056] A deep learning network is constructed as a deterioration prediction model. This network includes at least one long short-term memory network for capturing the temporal dependencies of anomalous features, and a fully connected branch for fusing environmental parameters and scene labels. Using historical anomalous feature datasets and corresponding actual deterioration records as training samples, and aiming to minimize the error between the predicted deterioration degree and the actual deterioration label, the parameters of the deep learning network are iteratively trained using a time series backpropagation algorithm and a gradient descent optimizer until the model converges.

[0057] The deterioration prediction model organizes the generated current abnormal feature dataset into time series data according to the collection time sequence. The organized time series data, current environmental parameters, and target future construction scenario identifiers are input into the trained and converged deterioration prediction model. The model extracts the evolution pattern of abnormal features through its long short-term memory network, and integrates environmental and scenario information through a fully connected layer. Finally, the output layer calculates and outputs a comprehensive deterioration coefficient that represents future risk.

[0058] Step 6: Based on the deterioration coefficient and the preset risk threshold rules, automatically generate a quality assessment report that includes the current quality status of materials and predictions of future deterioration risks. Submit the report to the construction management end through the communication network to guide the use of materials, replacement, or adjustment of construction plans.

[0059] The default logic of the risk threshold rule is as follows: based on the numerical range of the comprehensive deterioration coefficient, at least three consecutive risk threshold intervals are pre-divided, corresponding to low risk, medium risk and high risk levels respectively; for low risk level, the recommended measures are periodic re-inspection and routine monitoring; for medium risk level, the recommended measures include shortening the inspection cycle, applying temporary protective coatings or carrying out local reinforcement; for high risk level, the recommended measures are to immediately replace materials, adjust the structural design or change the construction process parameters.

[0060] Compared with existing technologies, this embodiment achieves objective identification of material surface defects by constructing multi-dimensional acquisition indicators and quantifying anomaly coefficients. It introduces a deterioration prediction model based on deep learning, upgrading static quality inspection to dynamic performance degradation risk assessment. This effectively overcomes the shortcomings of existing technologies, such as strong subjectivity and inability to predict material performance changes in specific future environments. It provides data-driven decision-making basis for material selection and scheme optimization before construction, and improves the predictability and initiative of quality control.

[0061] Example 2

[0062] At other levels, this embodiment also provides another optimization mechanism based on Embodiment 1, specifically a building material surface quality identification system, such as... Figure 2 As shown, it includes:

[0063] The indicator definition module 1 is used to predefine several surface feature collection indicators based on the type of target building material and its quality requirements under a preset construction scenario, and form a collection indicator set.

[0064] Feature extraction module 2 collects surface data of the current building material sample based on the collection index set, preprocesses and extracts features from the surface data, and outputs the current surface feature value corresponding to each collection index.

[0065] Anomaly calculation module 3 is used to compare each current surface feature value with the pre-stored standard feature value range corresponding to the same type of qualified building materials. Based on the comparison deviation, the anomaly coefficient of each surface feature is calculated by a preset normalization algorithm.

[0066] The dataset construction module 4 is used to integrate the anomaly coefficients of all surface features into anomaly feature vectors, and to associate and store the anomaly feature vectors with the current building material identification information, collection time and environmental parameters to construct anomaly feature datasets;

[0067] The deterioration prediction module 5 is used to construct a deterioration prediction model based on historical abnormal feature datasets and corresponding subsequent actual deterioration records as training samples. The current abnormal feature dataset is input into the model to evaluate and output the comprehensive deterioration coefficient of the current building materials under a specific preset construction scenario in the future.

[0068] The report generation module 6 is used to automatically generate a quality assessment report containing the current status of material quality and predictions of future deterioration risks based on the deterioration coefficient and preset risk threshold rules, and submit it to the construction management end through the communication network.

[0069] The indicator definition module 1 and the feature extraction module 2 are connected via a wireless network. The feature extraction module 2 and the anomaly calculation module 3 are connected via a wireless network. The anomaly calculation module 3 and the dataset construction module 4 are connected via a wireless network. The deterioration prediction module 5 is connected via a wireless network to the dataset construction module 4 and the report generation module 6.

[0070] In summary, this invention achieves a comprehensive profile of the material surface condition through multi-dimensional feature acquisition and fusion analysis, avoiding the one-sidedness of single-index evaluation. Through the deterioration prediction model, it can uncover the deterioration patterns hidden in historical data, enabling prediction of the material's future performance. It has a high degree of automation, and the generated quantitative construction report is directly connected to the management end, improving the efficiency of quality assessment and the response speed of decision-making, and reducing the reliance on human subjective experience.

[0071] This invention can not only effectively avoid construction safety risks and engineering rework costs caused by potential material defects, but also provide scientific suggestions for material maintenance and replacement cycles, thereby achieving cost optimization throughout the entire life cycle.

[0072] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for identifying the surface quality of building materials, characterized in that, Includes the following steps: Step 1: Based on the type of the target building material and its quality requirements under the preset construction scenario, predefine several surface feature acquisition indicators; Step 2: Based on the constructed collection index set, collect surface data of the current building material sample, preprocess and extract features from the surface data, and output the current surface feature value corresponding to each collection index; Step 3: Compare each current surface feature value obtained with the standard feature value range corresponding to the pre-stored qualified building materials of the same type. Based on the comparison deviation, calculate the anomaly coefficient of each surface feature relative to the standard state using a preset normalization algorithm. Step 4: Integrate the anomaly coefficients of all calculated surface features to form an anomaly feature vector. Associate and store the anomaly feature vector with the current building material identification information, collection time, and environmental parameters to construct an anomaly feature dataset. Step 5: Construct a deterioration prediction model. The model uses historical abnormal feature datasets and corresponding subsequent actual deterioration records as training samples, and the current abnormal feature dataset as model input. The deterioration prediction model evaluates and outputs the comprehensive deterioration coefficient of the current building materials under a specific preset construction scenario in the future. Step 6: Based on the deterioration coefficient and the preset risk threshold rules, automatically generate a quality assessment report and submit the report to the construction management end through the communication network.

2. The method for identifying the surface quality of building materials according to claim 1, characterized in that, The set of collected indicators in step 1 includes at least one or more of the following: surface texture complexity, color uniformity, pore distribution density, microcrack length, reflectance spectral characteristics, and local deformation curvature.

3. The method for identifying the surface quality of building materials according to claim 1, characterized in that, The normalization algorithm in step 3 operates as follows: The current surface feature value is compared with the pre-stored corresponding qualified standard feature value range to determine the direction and magnitude of the deviation of the current value from the midpoint of the standard range; For a feature value that falls within the standard range, its anomaly coefficient is obtained by calculating the distance of the value from the midpoint of the range and dividing it by the width of the entire standard range; For feature values ​​that fall outside the standard range, the anomaly coefficient is calculated based on the ratio above, but a penalty factor proportional to the distance beyond the boundary is introduced, so that the final calculated anomaly coefficient is greater than one.

4. The method for identifying the surface quality of building materials according to claim 1, characterized in that, The environmental parameters in step 4 include the ambient temperature, relative humidity, and light intensity at the time of data collection; the storage format of the abnormal feature dataset is a time series database, used to record the changes in abnormal feature vectors of the same building material at different time points.

5. The method for identifying the surface quality of building materials according to claim 1, characterized in that, The construction scenarios preset in step 5 include one or more combinations of long-term high humidity environment, periodic freeze-thaw cycle, strong acid and alkali corrosion environment, or continuous high load stress state.

6. The method for identifying the surface quality of building materials according to claim 1, characterized in that, The process of constructing the deterioration prediction model in step 5 includes: Acquire historical anomaly feature datasets and corresponding actual deterioration data records. Each data record contains an anomaly feature vector, material identifier, collection time, environmental parameters, and subsequent preset construction scenario identifier. The actual deterioration record is a quantitative deterioration degree label obtained by actual measurement or expert evaluation of the corresponding building material sample after experiencing the subsequent construction scenario. A deep learning network is constructed as a deterioration prediction model. This network includes at least one long short-term memory network for capturing the temporal dependencies of anomalous features, and a fully connected branch for fusing environmental parameters and scene labels. The historical anomalous feature dataset and the corresponding actual deterioration records are used as training samples. With the goal of minimizing the error between the predicted deterioration degree and the actual deterioration label, the parameters of the deep learning network are iteratively trained using a time series backpropagation algorithm and a gradient descent optimizer until the model converges.

7. The method for identifying the surface quality of building materials according to claim 6, characterized in that, The deterioration prediction model organizes the generated current abnormal feature dataset into time series model input data according to the collection time sequence. The organized time series data, current environmental parameters, and target future construction scene identifiers are input together into the trained and converged deterioration prediction model. The model extracts the evolution pattern of abnormal features through its long short-term memory network, and fuses environmental and scene information through a fully connected layer. Finally, the comprehensive deterioration coefficient is calculated and output through the output layer.

8. The method for identifying the surface quality of building materials according to claim 1, characterized in that, The preset logic of the risk threshold rule in step 6 is as follows: based on the numerical range of the comprehensive deterioration coefficient, at least three consecutive risk threshold intervals are pre-divided, corresponding to low risk, medium risk and high risk levels respectively. For low-risk levels, the recommended measures are periodic retesting and routine monitoring; For medium-risk levels, recommended measures include shortening the testing cycle, applying temporary protective coatings, or carrying out local reinforcement; for high-risk levels, recommended measures include immediately replacing materials, adjusting structural design, or changing construction process parameters.

9. The method for identifying the surface quality of building materials according to claim 1, characterized in that, The preprocessing in step 2 includes denoising, enhancing, scaling, and stitching images of the collected raw surface data; the feature extraction uses the gray-level co-occurrence matrix algorithm to extract texture features, the color histogram statistical method to extract color distribution features, and edge detection and morphological algorithms to extract crack and pore geometric features.

10. A building material surface quality identification system, wherein the system is an implementation system based on the building material surface quality identification method according to any one of claims 1-9, characterized in that, include: The indicator definition module (1) is used to predefine several surface feature collection indicators based on the type of target building material and its quality requirements under the preset construction scenario, and form a collection indicator set; Feature extraction module (2) collects surface data of the current building material sample based on the collection index set, preprocesses and extracts features from the surface data, and outputs the current surface feature value corresponding to each collection index; The anomaly calculation module (3) is used to compare each current surface feature value with the standard feature value range corresponding to the pre-stored qualified building materials of the same type, and calculate the anomaly coefficient of each surface feature based on the comparison deviation by using a preset normalization algorithm. The dataset construction module (4) is used to integrate the abnormal coefficients of all surface features into an abnormal feature vector, and to associate and store the abnormal feature vector with the current building material identification information, collection time and environmental parameters to construct an abnormal feature dataset; The deterioration prediction module (5) is used to construct a deterioration prediction model based on historical abnormal feature datasets and corresponding subsequent actual deterioration records as training samples. The current abnormal feature dataset is input into the model to evaluate and output the comprehensive deterioration coefficient of the current building materials under a specific preset construction scenario in the future. The report generation module (6) is used to automatically generate a quality assessment report containing the current status of material quality and prediction of future deterioration risks based on the deterioration coefficient and preset risk threshold rules, and submit it to the construction management end through the communication network. The indicator definition module (1) and the feature extraction module (2) are interconnected via a wireless network. The feature extraction module (2) and the anomaly calculation module (3) are interconnected via a wireless network. The anomaly calculation module (3) and the dataset construction module (4) are interconnected via a wireless network. The deterioration prediction module (5) and the dataset construction module (4) and the report generation module (6) are interconnected via a wireless network.

Images