A construction progress assessment method and system based on multi-UAV visual collaboration

Abstract

The present invention relates to the field of construction progress assessment, and discloses a multi-UAV visual collaborative construction progress assessment method and system, the method comprising the following steps: obtaining an original visual data set; extracting key features of a construction area; correcting importance scores and construction difficulty scores; calculating a comprehensive assessment coefficient for each construction area; sorting the construction areas according to the comprehensive assessment coefficients, and adjusting the construction plan and resource allocation; the system comprises a visual data acquisition module, a data processing module, a score acquisition module, a comprehensive assessment coefficient acquisition module, and a construction adjustment module. Through in-depth analysis and processing of image data, the importance and construction difficulty of each construction area can be more accurately assessed, thereby providing strong support for construction management and decision-making, significantly improving the efficiency and quality of construction progress assessment, and providing guarantees for the successful completion of construction projects.

Description

Technical Field

[0001] The present invention relates to the field of construction progress assessment, and in particular to a multi-UAV visual collaborative construction progress assessment method and system. Background Art

[0002] In the field of modern construction, project progress assessment is a key link to ensure that construction tasks are completed on time. Traditional progress assessment methods rely on manual on-site inspections and reports, which are not only time-consuming and labor-intensive, but also easily affected by subjective judgments, and accuracy and efficiency are often difficult to guarantee. With the advancement of technology, drone technology has been introduced into construction progress monitoring due to its flexibility and low cost. They can obtain high-resolution images of construction sites from the air, providing a new perspective for construction progress assessment.

[0003] Although the application of single drones in construction monitoring has achieved certain results, due to the complexity and vastness of the construction site, it is often difficult for a single drone to fully cover the entire construction site or complete large-scale data collection in a short period of time. Therefore, the concept of multi-drone system came into being. Through collaborative operation, the multi-drone system can not only significantly improve the efficiency of data collection, but also obtain more comprehensive and three-dimensional construction site information by shooting at different angles and heights.

[0004] For example, Chinese patent 201910637196.8 discloses a construction progress assessment method based on multi-UAV visual collaboration, which inputs the current 3D model information into the actual BIM model, and estimates the total project completion time of the BIM plan model based on the fuzzy algorithm using the actual BIM model; the real-time construction progress index is obtained by comparing the actual working hours of the actual BIM model with the total project completion time of the BIM plan model. However, the above method still has the following shortcomings: it does not fully consider the importance of the area during the construction process and the impact of the construction difficulty, resource allocation may become arbitrary, resulting in delayed construction progress in key areas, and lacks an assessment of the construction difficulty, that is, it is impossible to accurately identify areas where quality problems may occur, and thus it is impossible to take preventive measures in advance.

[0005] Currently, no effective solution has been proposed for the problems in the related technologies. Summary of the invention

[0006] In view of this, the purpose of the embodiments of the present application is to provide a construction progress assessment method and system based on multi-UAV visual collaboration to overcome the above-mentioned technical problems existing in the existing related technologies.

[0007] To this end, the specific technical solution adopted by the present invention is as follows:

[0008] According to one aspect of the present invention, a construction progress assessment method based on multi-UAV visual collaboration is provided, and the construction progress assessment method based on multi-UAV visual collaboration includes the following steps:

[0009] S1. Based on the collaborative flight of multiple UAVs, we capture images and video data of the construction site and obtain the original visual dataset.

[0010] S2. Preprocess the original visual data set to obtain a processed visual data set, and extract key features of the construction area to obtain a feature vector set.

[0011] S3. Evaluate the importance and construction difficulty of each construction area based on the feature vector set to obtain an importance score and a construction difficulty score, calculate the correlation coefficient between each construction area, and correct the importance score and the construction difficulty score.

[0012] S4. Construct a characteristic information matrix based on the corrected importance score and construction difficulty score, analyze the characteristic information matrix, and calculate the comprehensive evaluation coefficient of each construction area.

[0013] S5. Sort the construction areas according to the comprehensive evaluation coefficients and adjust the construction plan and resource allocation.

[0014] Optionally, the importance and construction difficulty of each construction area are evaluated based on the feature vector set to obtain an importance score and a construction difficulty score, and the correlation coefficient between each construction area is calculated. Correcting the importance score and the construction difficulty score includes the following steps:

[0015] S31, analyzing the characteristic vectors, evaluating the importance of each construction area, and obtaining a preliminary importance score;

[0016] S32, identifying and correcting a scoring error caused by an image quality problem, and obtaining a corrected importance score;

[0017] S33, analyzing the construction difficulty based on the importance score to obtain a construction difficulty score;

[0018] S34. Correct the importance score and the construction difficulty score by calculating the correlation coefficient between the construction areas.

[0019] Optionally, analyzing the feature vectors, evaluating the importance of each construction area, and obtaining a preliminary importance score includes the following steps:

[0020] S311, training a classification model based on a support vector machine through feature vectors collected historically, substituting the feature vectors acquired in real time into the trained classification model, outputting classification results, and identifying different construction elements;

[0021] S312. Based on the type and quantity of construction elements, a preliminary importance score is assigned to each construction area using a weighted scoring method;

[0022] S313. Normalize the preliminary importance scores.

[0023] Optionally, identifying and correcting a score error caused by an image quality problem to obtain a corrected importance score comprises the following steps:

[0024] S321, performing time series processing on the preliminary importance score data to obtain symbolized importance score time series data;

[0025] S322. Using a similarity measurement method, compare the symbolized importance score time series data to identify and remove deviation values, missing values, and outliers in the data;

[0026] S323, randomly selecting features and corresponding thresholds from the importance score time series data set, and dividing the data set into two subsets according to the selected features and thresholds, and constructing nodes of an isolated tree;

[0027] S324, repeating the partitioning process until the data points are isolated or a predetermined tree depth is reached, thereby obtaining an isolation forest model;

[0028] S325, using the isolation forest model to detect each importance score data point, calculating the path length of each importance score data point in each tree, and calculating the average path length of the importance score data points, to obtain an anomaly index;

[0029] S326, determining a threshold of anomaly index to distinguish normal data points from outliers;

[0030] S327, marking outliers whose abnormality index is higher than the threshold;

[0031] S328. Correct the importance scores of the points marked as outliers to obtain corrected importance score data.

[0032] Optionally, performing time series processing on the preliminary importance score data to obtain symbolized importance score time series data includes the following steps:

[0033] S3211. Convert the preliminary importance score data into time series data, and process the time series data using a data standardization formula so that the time series data has zero mean and unit variance;

[0034] S3212, reducing the dimension of the time series data, and calculating the mean of each segment;

[0035] S3213. Map the mean of each segment to a predefined symbol set to obtain symbolized importance score time series data.

[0036] Optionally, analyzing the construction difficulty based on the importance score to obtain the construction difficulty score includes the following steps:

[0037] S331, extracting features related to construction difficulty using an image processing algorithm, and outputting a feature vector related to construction difficulty;

[0038] S332. Based on a supervised learning algorithm, a construction difficulty scoring model is trained using features including the corrected importance score and construction difficulty related features to obtain a trained construction difficulty scoring model;

[0039] S333, applying the construction difficulty scoring model to generate a construction difficulty score for each construction area;

[0040] Among them, the construction difficulty scoring model is applied to generate a construction difficulty score for each construction area. The calculation formula is: ;

[0041] Where, T 1 , T 2 , T 3 and T 4 They represent the corrected importance score value, space restriction coded value, working height coded value and material accessibility coded value respectively;

[0042] w 1 , w 2 , w 3 and 4 Respectively represent the T obtained through training 1 , T 2 , T 3 and T 4 The weight of

[0043] b Represents the bias term, which is used to provide the output value of the baseline. A Indicates the construction difficulty score.

[0044] Optionally, correcting the importance score and the construction difficulty score by calculating the correlation coefficient between the construction areas includes the following steps:

[0045] S341. For each pair of construction areas, calculate the correlation coefficient between the importance score and the construction difficulty score, and obtain a correlation coefficient matrix;

[0046] S342, using the correlation coefficient matrix as input data to perform principal component analysis, and according to the principal component analysis results, selecting the principal component that explains the data variation exceeding a preset threshold to reconstruct the importance score and construction difficulty score data;

[0047] S343. Use the selected principal components and inversely transform back to the original feature space to obtain the corrected importance score and construction difficulty score.

[0048] Optionally, constructing a characteristic information matrix based on the corrected importance score and construction difficulty score, analyzing the characteristic information matrix, and calculating the comprehensive evaluation coefficient of each construction area includes the following steps:

[0049] S41. Integrate the existing and corrected importance scores and construction difficulty scores with the existing construction task completion data, resource utilization efficiency data, safety performance data, environmental suitability data, time schedule compliance data, and economic data;

[0050] S42, recursively eliminating the integrated data, and constructing a feature information matrix for the eliminated data;

[0051] S43. Train the random forest model through the feature data of the feature information matrix to obtain a comprehensive evaluation coefficient prediction model, and input the feature data of the new feature information matrix into the comprehensive evaluation coefficient prediction model to output the comprehensive evaluation coefficient of each construction area.

[0052] Optionally, training a random forest model with the characteristic data of the characteristic information matrix to obtain a comprehensive evaluation coefficient prediction model, and inputting the characteristic data of the new characteristic information matrix into the comprehensive evaluation coefficient prediction model, and outputting the comprehensive evaluation coefficient of each construction area includes the following steps:

[0053] S431, converting the characteristic data of the characteristic information matrix into numerical values ​​and performing standardization processing;

[0054] S432, dividing the feature data into a training set and a test set;

[0055] S433, using the training set to train the random forest model, and adjusting the parameters of the random forest model, and finally obtaining a comprehensive evaluation coefficient prediction model;

[0056] S434, using the test set to evaluate the prediction performance of the comprehensive evaluation coefficient prediction model;

[0057] S435. Input the characteristic data of the new characteristic information matrix into the comprehensive evaluation coefficient prediction model to predict the comprehensive evaluation coefficient of each construction area.

[0058] According to another aspect of the present invention, a multi-UAV visual collaborative construction progress assessment system is provided, which includes a visual data acquisition module, a data processing module, a score acquisition module, a comprehensive evaluation coefficient acquisition module and a construction adjustment module.

[0059] Among them, the visual data acquisition module is used to capture images and video data of the construction site based on the collaborative flight of multiple UAVs to obtain the original visual data set.

[0060] The data processing module is used to preprocess the original visual data set to obtain a processed visual data set, and extract key features of the building construction area to obtain a feature vector set.

[0061] The scoring acquisition module is used to evaluate the importance and construction difficulty of each construction area based on the feature vector set, obtain the importance score and construction difficulty score, calculate the correlation coefficient between each construction area, and correct the importance score and construction difficulty score.

[0062] The comprehensive evaluation coefficient acquisition module is used to construct a characteristic information matrix based on the corrected importance score and construction difficulty score, analyze the characteristic information matrix, and calculate the comprehensive evaluation coefficient of each construction area.

[0063] The construction adjustment module is used to sort the building construction areas according to the comprehensive evaluation coefficient and adjust the construction plan and resource allocation.

[0064] Embodiments of the present invention include the following benefits:

[0065] (1) The present invention provides a method and system for construction progress assessment based on multi-UAV visual collaboration, which uses multiple UAVs to collaborate to capture images and video data of a construction site from different angles and positions, thereby improving the coverage and efficiency of data collection. The visual data collected by multiple UAVs is converted into quantitative indicators that are of practical significance for construction progress assessment. Through in-depth analysis and processing of image data, the importance and construction difficulty of each construction area can be more accurately assessed, thereby providing strong support for construction management and decision-making, significantly improving the efficiency and quality of construction progress assessment, and providing guarantees for the successful completion of construction projects.

[0066] (2) By obtaining a preliminary importance score, the present invention provides a quantitative basis for subsequent construction planning and resource allocation, which helps to determine which areas should be given priority for construction, thereby improving the efficiency of the overall project. The importance score is corrected based on the isolation forest model, which reduces misjudgments caused by poor image quality and improves the robustness of the evaluation method. The construction difficulty score can help project managers understand the specific challenges of construction in each area, so as to better plan construction methods and allocate resources. The score corrected by the correlation coefficient more comprehensively considers the mutual influence between regions and improves the accuracy of the score.

[0067] (3) The present invention integrates the corrected importance score and construction difficulty score with other key construction indicators, including construction task completion, resource utilization efficiency, safety performance, environmental suitability, time schedule compliance and economic data, and can comprehensively reflect the actual construction progress of the construction area through integration. By recursively eliminating the integrated data, the features with small prediction contributions are eliminated to construct a feature information matrix; by obtaining the comprehensive evaluation coefficient of each construction area, a quantitative evaluation tool is provided for project managers, which helps project managers make more informed decisions, such as optimizing construction plans, adjusting resource allocation and taking risk mitigation measures.

[0068] (4) The present invention adds importance scores and construction difficulty scores to the comprehensive evaluation, making the construction progress evaluation more comprehensive, taking into account not only the completion rate, but also the quality and risk factors of the construction. Through comprehensive evaluation, resources such as manpower, materials and equipment can be allocated more effectively, ensuring that resources are allocated to the areas where they are most needed, thereby improving overall work efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0070] Figure 1 is a flow chart of a construction progress assessment method based on multi-UAV visual collaboration according to an embodiment of the present invention;

[0071] Figure 2 It is a principle block diagram of a multi-UAV visual collaborative construction progress assessment system according to an embodiment of the present invention.

[0072] In the figure:

[0073] 1. Visual data acquisition module; 2. Data processing module; 3. Score acquisition module; 4. Comprehensive evaluation coefficient acquisition module; 5. Construction adjustment module. DETAILED DESCRIPTION

[0074] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0075] To further illustrate each embodiment, the present invention provides drawings, which are part of the disclosure of the present invention and are mainly used to illustrate the embodiments and can be used in conjunction with the relevant descriptions in the specification to explain the operating principles of the embodiments. With reference to these contents, ordinary technicians in the field should be able to understand other possible implementations and advantages of the present invention. The components in the figures are not drawn to scale, and similar component symbols are generally used to represent similar components.

[0076] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which this application belongs. The terms used herein are only for the purpose of describing the embodiments of this application and are not intended to limit this application.

[0077] According to an embodiment of the present invention, a construction progress assessment method and system based on multi-UAV visual collaboration are provided.

[0078] The present invention is further described with reference to the accompanying drawings and specific embodiments. Figure 1 As shown, according to one embodiment of the present invention, a construction progress assessment method based on multi-UAV visual collaboration is provided, and the construction progress assessment method based on multi-UAV visual collaboration includes the following steps:

[0079] S1. Based on the coordinated flight of multiple drones, capture the image and video data of the construction site to obtain the original visual data set. Specifically:

[0080] Determine the flight path and range of the drone, including the take-off point, flight altitude, aerial photography route, etc. Set flight parameters: Set the flight parameters of the drone, such as flight speed, flight altitude, shooting angle, exposure, etc. Start the drone, let it take off and perform aerial photography tasks along the predetermined path. During the flight, the drone uses the onboard camera to capture images and video data of the construction site.

[0081] S2. Preprocess the original visual data set to obtain a processed visual data set, and use image processing or machine learning algorithms to extract key features of the construction area to obtain a feature vector set. Specifically:

[0082] The original visual data is cleaned and denoised, the image is segmented into different areas or objects, and image processing techniques (such as edge detection, color features, texture features, etc.) are used to extract feature information related to the construction area from the image.

[0083] The video data is processed to extract images of key frames or key time periods, and then features are extracted from these images.

[0084] The selected features are converted into feature vectors for use as input to machine learning algorithms. The feature vectors are analyzed and processed using machine learning algorithms (e.g., clustering, classification, regression, etc.) to identify and classify key features of the construction area. The processed and analyzed feature vector set is saved or output as a representation of the key features of the construction area.

[0085] S3. Evaluate the importance and construction difficulty of each construction area based on the feature vector set to obtain an importance score and a construction difficulty score, calculate the correlation coefficient between each construction area, and correct the importance score and the construction difficulty score.

[0086] In a further embodiment, the importance and construction difficulty of each construction area are evaluated based on the feature vector set to obtain an importance score and a construction difficulty score, and the correlation coefficient between each construction area is calculated. Correcting the importance score and the construction difficulty score includes the following steps:

[0087] S31. Analyze the feature vectors, evaluate the importance of each construction area, and obtain a preliminary importance score.

[0088] In a further embodiment, analyzing the feature vectors, evaluating the importance of each construction area, and obtaining a preliminary importance score comprises the following steps:

[0089] S311. Train the classification model based on support vector machine through the feature vectors collected historically, and substitute the feature vectors acquired in real time into the trained classification model to output the classification results and identify different construction elements; support vector machine is a supervised learning algorithm that can find an optimal hyperplane in the feature space to maximize the boundaries between different categories. After the training is completed, the feature vectors acquired in real time are input into the model for classification, and the classification results are output, so as to identify different construction elements in the construction site.

[0090] S312. Based on the category and quantity of construction elements, a preliminary importance score is assigned to each construction area using a weighted scoring method; the weighted scoring method is a scoring mechanism that takes the importance of different factors into account. A weight is assigned to each construction element, and then the importance score of each construction area is calculated based on its category and quantity.

[0091] S313. Normalize the preliminary importance scores. The preliminary importance scores are normalized so that the scores are standardized to a uniform range, which helps to compare the importance between different construction areas.

[0092] S32, identifying and correcting a scoring error caused by an image quality problem, and obtaining a corrected importance score;

[0093] In a further embodiment, identifying and correcting the score error caused by the image quality problem to obtain the corrected importance score comprises the following steps:

[0094] S321, performing time series processing on the preliminary importance score data to obtain symbolized importance score time series data; that is, converting the preliminary importance score data into a time series format, which means that each score is associated with a time point.

[0095] S322. Use similarity measurement methods (such as Euclidean distance, Pearson correlation coefficient, etc.) to compare the symbolized importance score time series data, identify and eliminate deviation values, missing values ​​and outliers in the data. These values ​​are caused by image quality problems (such as blur, occlusion, lighting changes, etc.) and affect the accuracy of the score.

[0096] S323, randomly select features (features are time points of scores or other related attributes) and corresponding thresholds from the importance score time series data set, and divide the data set into two subsets according to the selected features and thresholds, and construct nodes of the isolated tree. For example, select the score as a feature, and select the corresponding threshold, and divide the data set into two subsets according to the selected features and thresholds.

[0097] S324, repeating the partitioning process until the data point is isolated or reaches a predetermined tree depth, thereby obtaining an isolation forest model, i.e., multiple isolated trees;

[0098] It should be noted that Isolation Forest is a tree-based anomaly detection algorithm. It randomly selects a feature and a split value to divide the data, constructs the nodes of the isolation tree, and then repeats this process until the data point is isolated or the predetermined tree depth is reached. The isolation forest model constructed by multiple operations can be used for subsequent anomaly detection.

[0099] S325. Use the isolation forest model to detect each importance score data point, calculate the path length of each importance score data point in each tree, and calculate the average path length of the importance score data points to obtain an anomaly index.

[0100] S326, determining a threshold of anomaly index, which is used to distinguish normal data points from outliers; short path length usually means anomaly. Determine a threshold of anomaly index through statistical analysis, which is used to distinguish normal data points from outliers.

[0101] S327, marking outliers whose abnormality index is higher than a threshold, which are scoring outliers caused by image quality problems;

[0102] S328. The importance scores of the points marked as outliers are corrected by using an interpolation method, an averaging method, or a score estimation method based on adjacent normal data points to obtain corrected importance score data, thereby reducing errors caused by image quality issues.

[0103] In a further embodiment, performing time series processing on the preliminary importance score data to obtain symbolized importance score time series data includes the following steps:

[0104] S3211. Convert the preliminary importance score data into time series data, and process the time series data using a data standardization formula so that the time series data has zero mean and unit variance;

[0105] The data normalization formula is: ;

[0106] In the formula, C i Represents the time series data after normalization. i elements;

[0107] x i Represents the time series data i elements;

[0108] μ Represents the average value of time series data;

[0109] δ Represents the standard deviation of the time series data.

[0110] In calculation C i When each data point x iThe mean of the original time series is subtracted from the data. The mean of the standardized data set is zero. Because the mean is subtracted from each data point, the mean of the entire data set is moved to zero. Dividing by the standard deviation, we get C i yes x i The degree of deviation is in the units of standard deviation. Therefore, the calculated normalized data set has unit variance.

[0111] S3212, reduce the dimension of the time series data and calculate the mean of each segment; that is, divide the standardized time series into multiple segments according to the set interval, and calculate the mean of each segment. This helps to reduce the dimension and complexity of the data, while retaining the average value of the data points in each segment, which helps to further analyze and process the data.

[0112] S3213, mapping the mean of each segment to a predefined symbol set to obtain symbolized importance score time series data. The predefined symbol set refers to a set of symbols or labels defined in advance to represent different numerical ranges or trends in the time series data. This symbol set is usually set based on domain knowledge, experience or a specific algorithm, and its purpose is to convert continuous numerical data into discrete symbols or labels.

[0113] S33, analyzing the construction difficulty based on the importance score to obtain a construction difficulty score;

[0114] In a further embodiment, analyzing the construction difficulty based on the importance score to obtain the construction difficulty score includes the following steps:

[0115] S331. Use image processing algorithms (such as edge detection, object recognition, etc.) to extract features related to construction difficulty, such as space restrictions, working height, material accessibility, etc., and output feature vectors related to construction difficulty.

[0116] S332: Based on a supervised learning algorithm, a construction difficulty scoring model is trained using the corrected importance score and construction difficulty related features to obtain a trained construction difficulty scoring model. Specifically:

[0117] Supervised learning algorithms include decision trees, support vector machines, neural networks, etc. The selected supervised learning model is used for training so that the model can learn the relationship between the data and generate a suitable construction difficulty score prediction model. The trained model is tuned to optimize the model's hyperparameters or structure to improve the model's generalization ability and prediction performance. The trained model is evaluated using methods such as validation sets or cross-validation to evaluate the model's performance and check for problems such as overfitting or underfitting.

[0118] S333. Apply the construction difficulty scoring model to generate a construction difficulty score for each construction area, taking into account the importance of the area and other difficulty-related characteristics.

[0119] Among them, the construction difficulty scoring model is applied to generate a construction difficulty score for each construction area. The calculation formula is: ;

[0120] Where, T 1 , T 2 , T 3 and T 4 They represent the corrected importance score value, space restriction coded value, working height coded value and material accessibility coded value respectively;

[0121] Spatial constraints: refers to the spatial conditions of the construction area, including the size, shape, restrictive factors (such as existing obstacles, adjacent buildings, etc.) of the available space, and the layout of the work area.

[0122] Working height: refers to the height at which work is performed on the construction site, which may involve operations above ground level, the erection of scaffolding or the use of lifting equipment.

[0123] Accessibility of materials: refers to the ease of obtaining and using materials at various locations on the construction site. This involves the storage location of materials, supply chain management, material transportation, etc.

[0124] w 1 , w 2 , w 3 and 4 Respectively represent the T obtained through training 1 , T 2 , T 3 and T 4 The weight of

[0125] b Represents a bias term, which provides a baseline output value. Even if all feature values ​​are zero, the model will output a non-zero score, which is the value of the bias term.

[0126] S34. Correct the importance score and the construction difficulty score by calculating the correlation coefficient between the construction areas.

[0127] In a further embodiment, by calculating the correlation coefficients between the construction areas, correcting the importance score and the construction difficulty score comprises the following steps:

[0128] S341. For each pair of construction areas, the correlation coefficient between the importance score and the construction difficulty score is calculated, and a correlation coefficient matrix is ​​obtained to show the score correlation between all construction areas; the correlation coefficient is a statistic used to measure the degree of correlation between two variables. It indicates the strength and direction of the linear correlation between the two variables.

[0129] S342. Perform principal component analysis using the correlation coefficient matrix as input data, and based on the results of the principal component analysis, select principal components that explain data variation exceeding a preset threshold to reconstruct the importance score and construction difficulty score data, and ignore those components that explain less variation. This can reduce multicollinearity between variables because principal components are orthogonal (i.e., linearly independent).

[0130] S343. Use the selected principal components and transform back to the original feature space to obtain the corrected importance scores and construction difficulty scores. These corrected scores will retain the most important variation information in the data while reducing the impact of other minor variations.

[0131] S4. Construct a characteristic information matrix based on the corrected importance score and construction difficulty score, analyze the characteristic information matrix, and calculate the comprehensive evaluation coefficient of each construction area.

[0132] In a further embodiment, constructing a characteristic information matrix based on the corrected importance score and construction difficulty score, and analyzing the characteristic information matrix to calculate the comprehensive evaluation coefficient of each building construction area includes the following steps:

[0133] S41. Integrate the existing and corrected importance scores and construction difficulty scores with the existing construction task completion data, resource utilization efficiency data, safety performance data, environmental suitability data, time schedule compliance data and economic data.

[0134] S42. Recursively eliminate the integrated data and construct a feature information matrix for the eliminated data. Recursive elimination means improving the generalization ability and interpretability of the model by continuously eliminating features that have little impact on the model performance.

[0135] S43. Train the random forest model through the feature data of the feature information matrix to obtain a comprehensive evaluation coefficient prediction model, and input the feature data of the new feature information matrix into the comprehensive evaluation coefficient prediction model to output the comprehensive evaluation coefficient of each construction area.

[0136] In a further embodiment, the random forest model is trained by the feature data of the feature information matrix to obtain a comprehensive evaluation coefficient prediction model, and the feature data of the new feature information matrix is ​​input into the comprehensive evaluation coefficient prediction model, and the comprehensive evaluation coefficient of each construction area is output, which includes the following steps:

[0137] S431, converting the characteristic data of the characteristic information matrix into numerical values ​​and performing standardization processing;

[0138] S432, dividing the feature data into a training set and a test set;

[0139] S433, training a random forest model using the training set, and adjusting random forest model parameters, such as the number of trees, the maximum depth of the trees, etc., to optimize model performance, and finally obtain a comprehensive evaluation coefficient prediction model;

[0140] S434. Use the test set to evaluate the prediction performance of the comprehensive evaluation coefficient prediction model. Common evaluation indicators include accuracy, recall, F1 score and AUC value, and perform cross-validation to ensure the stability and generalization ability of the model.

[0141] S435. Input the characteristic data of the new characteristic information matrix into the comprehensive evaluation coefficient prediction model to predict the comprehensive evaluation coefficient of each construction area.

[0142] S5. Sort the construction areas according to the comprehensive evaluation coefficients and adjust the construction plan and resource allocation. Specifically:

[0143] Priority adjustment: Allocate an earlier construction start time to areas with a high comprehensive assessment coefficient. This means that the project timeline needs to be reviewed and adjusted based on the importance and difficulty of the construction area. At the same time, for important or difficult construction areas, increase the frequency of progress monitoring to cope with potential delays.

[0144] Resource reallocation: Allocate more resources (such as funds, manpower, equipment, etc.) to areas with high comprehensive evaluation coefficients.

[0145] like Figure 2 As shown, according to another embodiment of the present invention, a multi-UAV visual collaborative construction progress assessment system is provided, and the multi-UAV visual collaborative construction progress assessment system includes a visual data acquisition module 1, a data processing module 2, a score acquisition module 3, a comprehensive evaluation coefficient acquisition module 4 and a construction adjustment module 5.

[0146] Among them, the visual data acquisition module 1 is used to capture images and video data of the construction site based on the collaborative flight of multiple UAVs to obtain the original visual data set.

[0147] The data processing module 2 is used to preprocess the original visual data set to obtain a processed visual data set, and extract key features of the construction area to obtain a feature vector set.

[0148] The score acquisition module 3 is used to evaluate the importance and construction difficulty of each construction area based on the feature vector set, obtain the importance score and construction difficulty score, calculate the correlation coefficient between each construction area, and correct the importance score and construction difficulty score.

[0149] The comprehensive evaluation coefficient acquisition module 4 is used to construct a characteristic information matrix based on the corrected importance score and construction difficulty score, analyze the characteristic information matrix, and calculate the comprehensive evaluation coefficient of each construction area.

[0150] The construction adjustment module 5 is used to sort the construction areas according to the comprehensive evaluation coefficients and adjust the construction plan and resource allocation.

[0151] In summary, the present invention provides a multi-UAV visual collaborative construction progress assessment method and system, which captures images and video data of a construction site from different angles and positions through multi-UAV collaboration, thereby improving the coverage and efficiency of data collection. The visual data collected by multiple UAVs is converted into quantitative indicators that are of practical significance for construction progress assessment. Through in-depth analysis and processing of image data, the importance and construction difficulty of each construction area can be more accurately assessed, thereby providing strong support for construction management and decision-making, significantly improving the efficiency and quality of construction progress assessment, and providing guarantees for the successful completion of construction projects. By obtaining a preliminary importance score, the present invention provides a quantitative judgment basis for subsequent construction planning and resource allocation, which helps to determine which areas should be prioritized for construction, thereby improving the efficiency of the overall project. And based on the isolation forest model, the importance score is corrected, which reduces the misjudgment caused by poor image quality and improves the robustness of the assessment method. The construction difficulty score can help project managers understand the specific challenges of construction in each area, so as to better plan construction methods and allocate resources. The score corrected by the correlation coefficient more comprehensively considers the mutual influence between regions and improves the accuracy of the score. The present invention integrates the corrected importance score and construction difficulty score with other key construction indicators, including the completion of construction tasks, resource utilization efficiency, safety performance, environmental suitability, time schedule compliance and economic data, and can comprehensively reflect the actual situation of the construction progress in the construction area through integration. By recursively eliminating the integrated data, the features with small prediction contributions are eliminated, thereby constructing a feature information matrix; through the comprehensive evaluation coefficient of each construction area obtained, a quantitative evaluation tool is provided for project managers, which helps project managers make more wise decisions, such as optimizing construction plans, adjusting resource allocation and taking risk mitigation measures. The present invention adds the importance score and construction difficulty score to the comprehensive evaluation, making the construction progress evaluation more comprehensive, considering not only the completion rate, but also the quality and risk factors of the construction. Through the comprehensive evaluation, resources such as manpower, materials and equipment can be allocated more effectively to ensure that resources are allocated to the areas where they are most needed, thereby improving overall work efficiency.

[0152] Those skilled in the art will appreciate that all or some of the steps in the methods disclosed above, and the functional charging modules / units in the systems and devices may be implemented as software, firmware, hardware, or appropriate combinations thereof.

[0153] The terms "first", "second", "third", "fourth", etc. (if any) in the specification of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0154] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A construction progress assessment method based on multi-UAV visual collaboration, characterized in that: The multi-UAV visual collaborative construction progress assessment method includes the following steps: S1. Based on the coordinated flight of multiple UAVs, images and video data of the construction site are captured to obtain the original visual data set; S2, preprocessing the original visual data set to obtain a processed visual data set, and extracting key features of the building construction area to obtain a feature vector set; S3. Evaluate the importance and construction difficulty of each construction area based on the feature vector set to obtain an importance score and a construction difficulty score, calculate the correlation coefficient between each construction area, and correct the importance score and the construction difficulty score; S4. construct a characteristic information matrix based on the corrected importance score and construction difficulty score, analyze the characteristic information matrix, and calculate the comprehensive evaluation coefficient of each construction area; S5. Sort the construction areas according to the comprehensive evaluation coefficients and adjust the construction plan and resource allocation; The importance and construction difficulty of each construction area are evaluated based on the feature vector set to obtain an importance score and a construction difficulty score, and the correlation coefficient between each construction area is calculated. The importance score and the construction difficulty score are corrected, and the following steps are included: S31, analyzing the characteristic vectors, evaluating the importance of each construction area, and obtaining a preliminary importance score; S32, identifying and correcting a scoring error caused by an image quality problem, and obtaining a corrected importance score; S33, analyzing the construction difficulty based on the importance score to obtain a construction difficulty score; S34, correcting the importance score and the construction difficulty score by calculating the correlation coefficient between the construction areas; The analysis of the construction difficulty based on the importance score to obtain the construction difficulty score includes the following steps: S331, extracting features related to construction difficulty using an image processing algorithm, and outputting a feature vector related to construction difficulty; S332. Based on a supervised learning algorithm, a construction difficulty scoring model is trained using features including the corrected importance score and construction difficulty related features to obtain a trained construction difficulty scoring model; S333, applying the construction difficulty scoring model to generate a construction difficulty score for each construction area; Among them, the construction difficulty scoring model is applied to generate a construction difficulty score for each construction area. The calculation formula is: ； Where, T1, T2, T3 and T4 represent the corrected importance score value, space restriction coded value, working height coded value and material accessibility coded value respectively; w1, w2, w3 and w4 represent the weights of T1, T2, T3 and T4 learned after training respectively; b Represents the bias term, which is used to provide the output value of the baseline; A Indicates the construction difficulty score; The calculation of the correlation coefficients between the construction areas and the correction of the importance scores and the construction difficulty scores comprises the following steps: S341. For each pair of construction areas, calculate the correlation coefficient between the importance score and the construction difficulty score, and obtain a correlation coefficient matrix; S342, using the correlation coefficient matrix as input data to perform principal component analysis, and according to the principal component analysis results, selecting the principal component that explains the data variation exceeding a preset threshold to reconstruct the importance score and construction difficulty score data; S343. Use the selected principal components and inversely transform back to the original feature space to obtain the corrected importance score and construction difficulty score.

2. The construction progress assessment method based on multi-UAV visual collaboration according to claim 1 is characterized in that: The analysis of the feature vectors, evaluation of the importance of each construction area, and obtaining a preliminary importance score includes the following steps: S311, training a classification model based on a support vector machine through feature vectors collected historically, substituting the feature vectors acquired in real time into the trained classification model, outputting classification results, and identifying different construction elements; S312. Based on the type and quantity of construction elements, a preliminary importance score is assigned to each construction area using a weighted scoring method; S313. Normalize the preliminary importance scores.

3. The construction progress assessment method based on multi-UAV visual collaboration according to claim 2 is characterized in that: The step of identifying and correcting the scoring error caused by the image quality problem to obtain a corrected importance score comprises the following steps: S321, performing time series processing on the preliminary importance score data to obtain symbolized importance score time series data; S322. Using a similarity measurement method, compare the symbolized importance score time series data to identify and remove deviation values, missing values, and outliers in the data; S323, randomly selecting features and corresponding thresholds from the importance score time series data set, and dividing the data set into two subsets according to the selected features and thresholds, and constructing nodes of an isolated tree; S324, repeating the partitioning process until the data points are isolated or a predetermined tree depth is reached, thereby obtaining an isolation forest model; S325, using the isolation forest model to detect each importance score data point, calculating the path length of each importance score data point in each tree, and calculating the average path length of the importance score data points, to obtain an anomaly index; S326, determining a threshold of anomaly index to distinguish normal data points from outliers; S327, marking outliers whose abnormality index is higher than the threshold; S328. Correct the importance scores of the points marked as outliers to obtain corrected importance score data.

4. The construction progress assessment method based on multi-UAV visual collaboration according to claim 3 is characterized in that: The step of performing time series processing on the preliminary importance score data to obtain symbolized importance score time series data comprises the following steps: S3211. Convert the preliminary importance score data into time series data, and process the time series data using a data standardization formula so that the time series data has zero mean and unit variance; S3212, reducing the dimension of the time series data, and calculating the mean of each segment; S3213. Map the mean of each segment to a predefined symbol set to obtain symbolized importance score time series data.

5. The construction progress assessment method based on multi-UAV visual collaboration according to claim 1 is characterized in that: The process of constructing a characteristic information matrix based on the corrected importance score and construction difficulty score, analyzing the characteristic information matrix, and calculating the comprehensive evaluation coefficient of each construction area includes the following steps: S41. Integrate the existing and corrected importance scores and construction difficulty scores with the existing construction task completion data, resource utilization efficiency data, safety performance data, environmental suitability data, time schedule compliance data, and economic data; S42, recursively eliminating the integrated data, and constructing a feature information matrix for the eliminated data; S43. Train the random forest model through the feature data of the feature information matrix to obtain a comprehensive evaluation coefficient prediction model, and input the feature data of the new feature information matrix into the comprehensive evaluation coefficient prediction model to output the comprehensive evaluation coefficient of each construction area.

6. The construction progress assessment method based on multi-UAV visual collaboration according to claim 5 is characterized in that: The method of training the random forest model with the feature data of the feature information matrix to obtain the comprehensive evaluation coefficient prediction model, and inputting the feature data of the new feature information matrix into the comprehensive evaluation coefficient prediction model to output the comprehensive evaluation coefficient of each construction area comprises the following steps: S431, converting the characteristic data of the characteristic information matrix into numerical values ​​and performing standardization processing; S432, dividing the feature data into a training set and a test set; S433, using the training set to train the random forest model, and adjusting the parameters of the random forest model, and finally obtaining a comprehensive evaluation coefficient prediction model; S434, using the test set to evaluate the prediction performance of the comprehensive evaluation coefficient prediction model; S435. Input the characteristic data of the new characteristic information matrix into the comprehensive evaluation coefficient prediction model to predict the comprehensive evaluation coefficient of each construction area.

7. A multi-UAV visual collaborative construction progress assessment system, used to implement the multi-UAV visual collaborative construction progress assessment method according to any one of claims 1 to 6, characterized in that: The multi-UAV visual collaborative construction progress assessment system includes a visual data acquisition module, a data processing module, a score acquisition module, a comprehensive assessment coefficient acquisition module and a construction adjustment module; The visual data acquisition module is used to capture images and video data of the construction site based on the coordinated flight of multiple drones to obtain the original visual data set; The data processing module is used to preprocess the original visual data set to obtain a processed visual data set, and extract key features of the building construction area to obtain a feature vector set; The score acquisition module is used to evaluate the importance and construction difficulty of each construction area based on the feature vector set, obtain the importance score and construction difficulty score, calculate the correlation coefficient between each construction area, and correct the importance score and construction difficulty score; The comprehensive evaluation coefficient acquisition module is used to construct a characteristic information matrix based on the corrected importance score and construction difficulty score, and analyze the characteristic information matrix to calculate the comprehensive evaluation coefficient of each construction area; The construction adjustment module is used to sort the construction areas according to the comprehensive evaluation coefficients and adjust the construction plan and resource allocation.

Images