Building measurement system and method based on unmanned aerial vehicle surveying and mapping

Abstract

The invention relates to the technical field of three-dimensional modeling, in particular to a building measurement system and method based on unmanned aerial vehicle surveying and mapping, and the system comprises an environment sensing and monitoring module which collects real-time illumination intensity, air temperature and humidity data, records environment changes, carries out the analysis according to the difference between the real-time data and a standard threshold value, dynamically adjusts the parameters of an unmanned aerial vehicle camera, and carries out the measurement of the unmanned aerial vehicle. Comprising exposure and sensitivity, matching the current environment, and recording environment acquisition parameters. According to the method, illumination, air temperature and humidity are collected, equipment parameters are dynamically adjusted, multi-scene adaptability is improved, a filtering strategy is set according to environment variables, abnormal data are removed, interpolation is carried out on a point cloud fracture area based on space continuity, data coherence is improved, a building structure is recognized through feature library comparison, and structural damage is recognized in combination with an image. And on the basis of form and trend prediction change, a process covers acquisition and analysis, closed-loop linkage is constructed, and environmental adaptability, data integrity and recognition precision are enhanced.

Description

Technical Field

[0001] The present invention relates to the technical field of three-dimensional modeling, and in particular to a building measurement system and method based on unmanned aerial vehicle (UAV) surveying and mapping. Background Art

[0002] The technical field of three-dimensional modeling involves the process of reconstructing and expressing three-dimensional spatial structures using multi-source data. The core content of this technical field is to achieve digital modeling of the geometric features of objects or environments in real scenes through multi-modal inputs such as visual images, laser scanning, and surveying data. Three-dimensional modeling is widely used in multiple fields such as building information modeling, digital twin, geographic information system, virtual reality, and intelligent manufacturing. In the field of building surveying and mapping, three-dimensional modeling technology is often combined with sub-fields such as remote sensing, surveying engineering, and photogrammetry, mainly involving key steps such as scene acquisition, data preprocessing, point cloud reconstruction, mesh generation, and texture mapping, with the aim of forming an accurate and visual three-dimensional expression model. This field has high requirements for the accuracy of data acquisition, the efficiency of modeling algorithms, and the reconstruction accuracy. In recent years, with the development of computing power and image analysis methods, three-dimensional modeling has gradually expanded from static models to dynamic modeling and real-time update directions.

[0003] Among them, a building measurement system based on UAV surveying and mapping refers to using a UAV equipped with an image acquisition device to perform multi-angle aerial photography in building construction or operation and maintenance scenarios, combining image acquisition with coordinate positioning information to form an original data set for modeling. This patent theme addresses technical issues such as complex building shapes and limitations of traditional measurement methods. It completes image data acquisition through low-altitude multi-angle flight, and then based on the image overlap degree and flight path control strategy, realizes high-density acquisition of building external space information. Subsequently, an image stitching method based on feature point matching is used to fuse the acquired images to generate a continuous-view image sequence for constructing a three-dimensional point cloud. On the basis of the point cloud, further combined with building geometric rules and image gray-scale texture analysis methods, building contour extraction and patch mesh generation are completed, thus completing the spatial modeling process of the building structure. This system integrates UAV navigation control, visual positioning, image stitching, and modeling processes to form a complete three-dimensional modeling path for building measurement applications.

[0004] The prior art does not configure a feedback mechanism for real-time environmental perception, lacks the ability to adaptively adjust to different lighting and climate conditions, resulting in quality fluctuations of image data in strong contrast or variable light environments. The filtering process does not adjust processing parameters according to changes in error sources, and is prone to generating residual errors. The missing area of the point cloud relies on external supplementation, and there is no systematic vacancy repair mechanism. Building structure recognition is mainly based on geometric feature rules and does not have a cross-type recognition logic. The structural state judgment focuses on the current performance and lacks trend inference based on morphological evolution. In practical applications, the above limitations reduce the integrity of modeling and the accuracy of damage recognition. Summary of the Invention

[0005] The object of the present invention is to solve the disadvantages existing in the prior art, and to propose a building measurement system and method based on UAV mapping.

[0006] To achieve the above object, the present invention adopts the following technical solutions: A building measurement system based on UAV mapping includes: The environmental perception and monitoring module collects real-time light intensity, temperature, and humidity data, records environmental changes, analyzes according to the difference between the real-time data and the standard threshold, dynamically adjusts the parameters of the UAV camera, including exposure and sensitivity, matches the current environment, and records the environmental acquisition parameters; The adaptive filtering module uses the environmental acquisition parameters to analyze the impact of the current environment on the data quality, matches the filtering window size and type, filters the collected point cloud data, removes the errors introduced by environmental variables, and outputs optimized point cloud data; The data interpolation and optimization module automatically fills in the missing areas according to the gaps and discontinuities in the optimized point cloud data by using spatial analysis technology, and generates a complete point cloud data set through continuity evaluation; The building feature recognition module processes the complete point cloud data set, analyzes the building elements, scores the matching degree of each element by comparing multiple elements with a preset building feature library, sorts and marks them by category, and generates a building feature classification result; The damage prediction and analysis module combines the image data collected by the UAV camera and the building feature classification result, uses visual analysis to detect structural damage, identifies and classifies the damage type, predicts future changes based on the damage structural features, and generates a structural damage prediction result.

[0007] The environmental acquisition parameters include exposure and sensitivity, the optimized point cloud data includes filtering window size, filtering type, and error removal, the complete point cloud data set is specifically the filling of missing areas and data continuity, the building feature classification result includes element matching score and category marking, and the structural damage prediction result specifically refers to damage type identification, damage classification, and future change prediction.

[0008] As a further solution of the present invention, the step of recording the environmental acquisition parameters is specifically: Collect real-time light intensity data and environmental temperature, obtain a light intensity matrix and a temperature matrix through preliminary processing of sensor data, and generate an environmental data set; Analyze the light intensity matrix in the environmental data set, calculate the difference between it and the light standard threshold, and adjust the exposure parameter of the UAV camera according to the size of the difference value to obtain a preliminary exposure setting; Based on the preliminary exposure settings and the current environmental temperature matrix described above, use the formula: ; Adjust the exposure parameters of the drone camera to the optimal value to obtain the exposure parameters; Among them, represents the exposure parameter, represents the light intensity, represents the temperature, represents the adjustment coefficient of the ambient light, ensuring that the exposure adjustment is proportional to the light intensity, represents the smoothing coefficient of the exposure adjustment due to environmental changes, reducing the impact of small fluctuations on the exposure, represents the absolute value of the difference between the light intensity and the temperature, used to capture the sensitivity to environmental changes; According to the exposure parameters and the humidity information in the current environmental dataset, adjust the sensitivity of the camera to match the current environmental conditions and record the environmental acquisition parameters.

[0009] As a further solution of the present invention, the steps for obtaining the optimized point cloud data are specifically as follows: Analyze the environmental acquisition parameters, including the measurement of real-time light, temperature, and humidity, extract parameters from the sensor data to generate an environmental parameter matrix, and obtain the environmental parameter configuration; Based on the environmental parameter configuration, select the filter window size and type that match the current environmental conditions, determine the matching of the filter through experimental data, prepare the configuration for point cloud data filtering, and obtain the filter configuration; Apply the filter configuration to process the original point cloud data, using the formula: ; Perform filtering processing, reduce the error introduced by environmental variables through the weighted average method, and obtain the optimized point cloud data; Among them, represents the optimized point cloud data, represents the i-th data point in the original point cloud dataset, is the average value obtained by the filter configuration according to the environmental parameters, represents the environmental sensitivity adjustment coefficient, used to adjust the relationship between data sensitivity and environmental variables, represents the absolute value of the difference between the original data and the filtered average value, used to calculate the deviation of the point cloud data, represents the total number of data points, the number of data points in the original point cloud data, used to calculate the average value.

[0010] As a further solution of the present invention, the steps for obtaining the complete point cloud dataset are specifically as follows: Identify the vacant areas in the optimized point cloud data, mark the boundaries and positions of the areas through spatial mapping technology, and generate a missing area index table; Utilize the missing area index table to select the spatial analysis technology that best matches the current environment, perform the filling operation, and evenly fill each missing area to obtain the preliminarily filled point cloud data; Perform continuity and consistency evaluation on the preliminarily filled point cloud data, using the formula: ; Evaluate the quality of the filling by comprehensively considering the position differences between each point and its neighboring points, and generate a complete point cloud data set; Wherein, represents the complete point cloud data set, represents the i-th point in the filled point cloud, is the spatial average position of adjacent points, is the total number of evaluation points, is the maximum difference between the point and the average position, measures the gap between the filled point and the surrounding average position, is the maximum value of the point differences.

[0011] As a further solution of the present invention, the steps for obtaining the building feature classification result are specifically as follows: Extract the building elements in the complete point cloud data set, perform spatial positioning and identification of structural characteristics, and extract the shape and material characteristics of each element through visual processing algorithms to generate an element feature index table; Use the element feature index table for data analysis, compare it with a preset building feature library, calculate the matching degree using statistical analysis methods, assign a preliminary matching score to each element, and generate a matching score result; Perform comprehensive analysis on the matching score result, calculate the score, using the formula: ; Determine the classification result of each element based on the comprehensive score and adjustment parameters, and generate a building feature classification result; Wherein, represents the classification score, refers to the score of a single element, reflecting its matching degree with the model in the feature library, is the number of elements, is the highest score, used for normalizing the score, is the weight parameter, emphasizing the key of certain features, is the offset, adjusting the score threshold to ensure the fairness and matching of the score.

[0012] As a further solution of the present invention, the steps for obtaining the structural damage prediction result are specifically as follows: Analyze the image data collected by the UAV camera, extract the building structure features in the images, identify the structural elements through image processing algorithms, create feature vectors for each element, and generate a building element feature library; Integrate the building element feature library and the building feature classification result, use a deep learning model to compare and analyze the element features, classify different structural damage types, and output the damage type classification result; Conduct a quantitative analysis on the damage type classification result, calculate the probability of future structural changes, and adopt the formula: ; Estimate the severity of the structural damage within a future time period and generate a structural damage prediction result; Wherein, represents the future change probability, Assign weights according to the criticality of the damage type, represents the damage depth, represents the time since the damage was discovered, Adjust the baseline prediction value, Adjust the ratio of the damage depth to the time to provide a sensitive adjustment to the damage deterioration rate, represents the total number of damage types.

[0013] A building measurement method based on UAV mapping, which is executed based on the above-mentioned building measurement system based on UAV mapping, and includes the following steps: S1: Dynamically adjust the exposure and sensitivity parameters of the UAV camera based on the current light intensity, temperature, and humidity data, adjust the filter window size and type, and apply them to the preliminary filtering process of the point cloud data to remove the influence of environmental factors and obtain environmentally adapted point cloud data; S2: Conduct a spatial analysis on the environmentally adapted point cloud data, identify the vacancies and discontinuities in the data, automatically fill the vacant areas, and supplement the point cloud data set through continuity evaluation to generate a complete point cloud data set; S3: Based on the complete point cloud data set, analyze the building elements, compare multiple elements in the preset building feature library, score the matching degree of each element, classify and mark the building features, and generate a building feature classification result; S4: Combine the image data collected by the UAV camera and the building feature classification result, detect the structural damage of the building through visual analysis, identify and classify the damage types, and predict future structural changes to obtain a structural damage prediction result.

[0014] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In the present invention, by collecting environmental light, temperature and humidity and analyzing the deviation from the standard value, the parameters of the image acquisition device are directly adjusted to improve the multi-scene adaptation ability. The filtering strategy is dynamically set according to environmental variables to effectively eliminate abnormal data points. The broken area in the point cloud is automatically interpolated based on spatial continuity to improve data coherence. The building elements are compared with a preset feature library to classify and identify the structural types. Combining image information to identify the structural damage category, and based on the damage form and feature trend, time series prediction is carried out. The processing flow covers multiple links from data collection to result analysis, forming a closed-loop linkage, enhancing environmental adaptability, data integrity and structural recognition accuracy in building measurement. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 is the system flow chart of the present invention; Figure 2 is the step flow chart for the present invention to record environmental acquisition parameters; Figure 3 is the step flow chart for the present invention to optimize the acquisition of point cloud data; Figure 4 is the step flow chart for the present invention to obtain a complete point cloud data set; Figure 5 is the step flow chart for the present invention to obtain the building feature classification result; Figure 6 is the step flow chart for the present invention to obtain the structural damage prediction result. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0016] In order to make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0017] In the description of the present invention, it should be understood that the orientation or positional relationship indicated by the terms "length", "width", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc. is based on the orientation or positional relationship shown in the accompanying drawings, and is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as limiting the present invention. In addition, in the description of the present invention, the meaning of "a plurality" is two or more, unless otherwise specifically defined.

[0018] Embodiment 1: Please refer to Figure 1 , the present invention provides a technical solution: A building measurement system based on unmanned aerial vehicle surveying and mapping includes: The environmental perception monitoring module collects real-time light intensity, temperature, and humidity data, records environmental changes, analyzes the differences between the real-time data and the standard thresholds, dynamically adjusts the parameters of the UAV camera, including exposure and sensitivity, to match the current environment, and records the environmental acquisition parameters; The adaptive filtering module uses the environmental acquisition parameters to analyze the impact of the current environment on the data quality, matches the filtering window size and type, filters the collected point cloud data, removes the errors introduced by environmental variables, and outputs optimized point cloud data; The data interpolation and optimization module automatically fills in the missing areas according to the gaps and discontinuities in the optimized point cloud data by using spatial analysis techniques, and generates a complete point cloud data set through continuity evaluation; The building feature recognition module processes the complete point cloud data set, analyzes the building elements, scores the matching degree of each element by comparing multiple elements with the preset building feature library, sorts and marks them by category, and generates the building feature classification result; The damage prediction and analysis module combines the image data collected by the UAV camera and the building feature classification result, uses visual analysis to detect structural damage, identifies and classifies the damage types, predicts future changes based on the damaged structural features, and generates the structural damage prediction result.

[0019] The environmental acquisition parameters include exposure and sensitivity. The optimized point cloud data includes the filtering window size, filtering type, and error removal. The complete point cloud data set specifically refers to the filling of missing areas and data continuity. The building feature classification result includes the element matching score and category marking. The structural damage prediction result specifically refers to the identification of damage types, damage classification, and prediction of future changes.

[0020] Please refer to Figure 2 , and the steps of recording the environmental acquisition parameters are specifically as follows: Collect real-time light intensity data and environmental temperature, obtain the light intensity matrix and temperature matrix through preliminary processing of sensor data, and generate an environmental data set; Through the deployment of multiple sensor nodes at different geographical locations, each node includes a light intensity sensor and a temperature sensor. The nodes send the data to the central processing unit, which is responsible for collecting the data of all nodes and performing preliminary data cleaning to remove outliers or damaged data points. After data cleaning, the data fusion technology is used to integrate the data from different sensors to form a multi-dimensional environmental data set. This data set not only includes the original light and temperature data, but also undergoes spatial and temporal interpolation to fill in the blank areas that appear during the data collection process. This data processing technology ensures the integrity and availability of the data, providing support for subsequent data analysis and decision-making.

[0021] Analyze the light intensity matrix in the environmental dataset, calculate the difference between it and the light standard threshold, and adjust the exposure parameters of the UAV camera according to the magnitude of the difference value to obtain the preliminary exposure settings. By statistically analyzing the environmental dataset, first calculate the difference between the light intensity of each data point and the pre-set standard threshold. The difference value is obtained through arithmetic difference. Then, based on the statistical distribution of the difference value, determine the benchmark for exposure parameter adjustment. The adjustment strategy is set according to the degree to which the light intensity is lower or higher than the standard threshold. For the case where the light intensity is lower than the threshold, increase the exposure time of the camera to allow more light to enter the lens. Conversely, reduce the exposure time to reduce the amount of light entering the lens. Through this method, ensure that the UAV camera can obtain the optimal image quality under different lighting conditions and obtain the preliminary exposure settings.

[0022] Integrate the preliminary exposure settings and the current environmental temperature matrix, and use the formula: ; Adjust the exposure parameters of the UAV camera to the optimal and obtain the exposure parameters. Among them, represents the exposure parameter, represents the light intensity, represents the temperature, represents the adjustment coefficient of environmental light, ensuring that the exposure adjustment is proportional to the light intensity, represents the smoothing coefficient of exposure adjustment due to environmental changes, reducing the impact of small fluctuations on exposure, represents the absolute value of the difference between the light intensity and the temperature, used to capture the sensitivity of environmental changes; Formula: ; The advantage of the formula is that it allows the exposure settings of the camera to be dynamically adjusted to match the cyclically changing environmental light and temperature conditions. By combining the changes in light intensity and temperature, the exposure amount of the camera can be more accurately controlled, thus maintaining the consistency of image quality in various environments.

[0023] Detailed explanation of the formula and the derivation process of formula calculation: First, obtain the light intensity and environmental temperature through the data collected by the environmental sensor, and then calculate the absolute difference between them . is the environmental light adjustment coefficient, determined through experiments or according to the response curve of the camera, and is a smoothing coefficient used to reduce the exposure adjustment fluctuations caused by environmental changes, obtained based on previous data analysis. If the light intensity is 500 lux and the temperature is 25 °C, set and , the calculation process is as follows: ; ; ; The results show that according to the current environmental conditions, the exposure parameter of the camera should be set to 2.08 to ensure image quality. This value directly affects the exposure time of the camera, thereby ensuring the consistency of image quality under different lighting and temperature conditions.

[0024] According to the exposure parameter and the humidity information in the current environmental dataset, adjust the sensitivity of the camera to match the current environmental conditions and record the environmental acquisition parameters.

[0025] Collect the humidity information of the current environment through an environmental sensor, and then combine it with the previously calculated exposure parameter to adjust the sensitivity of the camera. The adjustment process refers to the influence of humidity on light propagation and reflection. In a high-humidity environment, the light scattering increases, and the sensitivity needs to be increased to capture more light. Conversely, in a low-humidity environment, the sensitivity is reduced to avoid overexposure. The adjusted sensitivity setting is determined through experiments and previous records under similar conditions to ensure that the camera can obtain appropriate exposure effects in different environments. At the same time, record all adjusted parameters and environmental data for future analysis and comparison. In this way, real-time monitoring of environmental parameters and dynamic adjustment of camera settings are carried out to ensure that the drone can capture the optimal images under various environmental conditions and match the current environmental conditions.

[0026] Please refer to Figure 3 , the specific steps for optimizing the acquisition of point cloud data are as follows: Analyze the environmental acquisition parameters, including the measurement of real-time illumination, temperature, and humidity, extract parameters from the sensor data, generate an environmental parameter matrix, and obtain the environmental parameter configuration; During the process of analyzing the environmental acquisition parameters, first continuously obtain illumination, temperature, and humidity data according to the real-time monitoring system. Each data point is automatically recorded every five minutes, and the real-time update and accuracy of the data are ensured through environmental monitoring sensors. Then, use data preprocessing algorithms, including noise removal and data smoothing, to optimize the data quality, thereby ensuring the reliability of subsequent analysis. The analysis results will be used to adjust and optimize the environmental parameter configuration to ensure that the system settings can reflect environmental changes in real time. Through this series of steps, the obtained environmental parameter configuration can be directly used in the subsequent steps, ensuring the consistency and real-time nature of data processing.

[0027] Based on the environmental parameter configuration, select the filtering window size and type that match the current environmental conditions, determine the matching of the filter through experimental data, prepare the configuration for point cloud data filtering, and obtain the filter configuration; In the decision-making process of selecting the filtering window size and type according to the analysis results of the environmental parameter configuration, first match the environmental parameters with the pre-stored environmental configurations in the database, search for the optimal filtering settings under the nearest conditions. This matching process includes comparing how the performance of multiple filters responds to the corresponding environmental parameters, using machine learning algorithms for pattern recognition to identify the filter type most suitable for the current environmental conditions, and then conducting experimental verification to ensure that the selected filter has the best effect under similar environmental conditions. Through this process, ensure that the optimal noise suppression effect can be achieved for each data processing, and the obtained filter configuration will directly affect the next data processing effect.

[0028] Apply the filter configuration to process the original point cloud data, using the formula: ; Perform filtering processing, reduce the error introduced by environmental variables through the weighted average method, and obtain the optimized point cloud data; Among them, represents the optimized point cloud data, represents the i-th data point in the original point cloud dataset, is the average value obtained by the filter configuration according to the environmental parameters, represents the environmental sensitivity adjustment coefficient, which is used to adjust the relationship between data sensitivity and environmental variables, represents the absolute value of the difference between the original data and the filtered average value, which is used to calculate the deviation of the point cloud data, represents the total number of data points, the number of data points in the original point cloud data, which is used to calculate the average value.

[0029] Formula: ; The benefit of the formula is that by introducing the environmental sensitivity adjustment coefficient , the response intensity of each data point to environmental changes can be flexibly adjusted, enhancing the matching and accuracy of the model under different environmental conditions.

[0030] Detailed explanation of the formula and the derivation process of formula calculation: Refer to an actual scenario, where the original point cloud dataset , the average environmental impact , the environmental sensitivity adjustment coefficient , substitute into the formula to get: ; ; The results show that through the above calculation process, the adjusted point cloud data is obtained. The value is 1.35, which indicates that the deviation of the point cloud data after considering the environmental impact is effectively reduced, making the data processing results more stable and accurate, thereby improving the data quality and reducing the errors introduced by environmental variables.

[0031] Please refer to Figure 4 , and the specific steps for obtaining the complete point cloud dataset are as follows: Identify the vacant areas in the optimized point cloud data, mark the boundaries and positions of the areas through spatial mapping technology, and generate a missing area index table. Through spatial mapping technology, mark the boundaries and positions of the missing areas. This process relies on geographic information systems and can analyze the original point cloud data to identify vacant and discontinuous areas, converting these areas into accurate digital representations. Each missing area will be assigned a unique identifier, and these identifiers are stored in the index table for subsequent data processing and filling processes. The purpose of doing this is to ensure that each point that needs to be corrected can be located during the filling process, thereby ensuring the integrity and accuracy of the data. This index table provides the necessary input information for the subsequent steps, enabling the filling process to target the vacant areas.

[0032] Using the missing area index table, select the spatial analysis technology that best matches the current environment, perform the filling operation, and evenly fill each missing area to obtain the preliminarily filled point cloud data. This algorithm can analyze the spatial continuity in the point cloud data and recommend the most suitable filling method for the current data characteristics. Each selected filling technology will configure corresponding parameters for the missing type. For large, continuously missing areas, an interpolation-based method will be selected, while for scattered point missing, statistical reconstruction technology will be used. Through the application of these technologies, the preliminarily filled point cloud data can effectively repair the vacancies while maintaining the original data characteristics, thus providing a good basis for data continuity assessment.

[0033] For the preliminarily filled point cloud data, perform continuity and consistency assessments using the formula: ; Generate the complete point cloud dataset by comprehensively evaluating the filling quality based on the position differences between each point and its neighboring points. Among them, represents the complete point cloud dataset, represents the i-th point in the filled point cloud, is the spatial average position of the adjacent points, is the total number of evaluation points, is the maximum difference between the point and the average position, Measure the gap between the filled points and the surrounding average positions. It is the maximum value of the point differences.

[0034] Formula: ; The benefit of the formula is that by quantifying the position differences between each point and its neighboring points, it evaluates the filling quality, ensures continuity and consistency, and at the same time improves the precision and efficiency of data processing.

[0035] Detailed explanation of the formula and the derivation process of formula calculation: A simplified point cloud data set is set up, where , the point cloud data is as follows: , the average position , the maximum difference (the distance from point 13 to point 20). Calculate the absolute value of the difference between each point and the average position to obtain ; Substitute into the formula for calculation: ; The results show that the deviation between the filled data set and the average position in the ideal state is within an acceptable range, indicating that the filling operation maintains the consistency and continuity of the data.

[0036] Please refer to Figure 5 , the specific steps for obtaining the building feature classification results are as follows: Extract the building elements in the complete point cloud data set, perform spatial positioning and identification of structural characteristics, extract the shape and material characteristics of each element through visual processing algorithms, and generate an element feature index table; By analyzing the building elements in the complete point cloud data set, perform spatial positioning and structural characteristic identification, use visual processing algorithms to extract the shape and material characteristics of each element from the data set. This process involves image recognition technology and machine learning algorithms. Each building element is located by three-dimensional space coordinates, and its structural strength and material composition are analyzed. The identification of these characteristics is based on batch experimental data and previous training models to optimize the parameter settings of the algorithms. Through this method, determine the physical and chemical properties of the building elements, as well as their positions and functions in the building, and generate an element feature index table, providing the necessary basic data for subsequent comparison and classification.

[0037] Use the element feature index table for data analysis, compare with the preset building feature library, adopt statistical analysis methods to calculate the matching degree, assign a preliminary matching score to each element, and generate a matching score result; After obtaining the element feature indices, these indices are used for data analysis. The features of each building element are compared with a preset building feature library. The statistical analysis method is used in this process, including multivariate analysis and pattern recognition techniques. The fitness of each element to the features in the library is evaluated by calculating the Euclidean distance between the element feature vector and the multi-vectors in the feature library. The smaller the distance, the higher the fitness. In addition, the supervised learning technique in machine learning is also adopted to improve the accuracy of matching. Through this method, not only can the matching score of each element be obtained, but also the elements can be sorted according to the score, providing a basis for building classification and marking.

[0038] Conduct a comprehensive analysis of the matching score results, calculate the score, using the formula: ; Determine the classification result of each element based on the comprehensive score and adjustment parameters, generating the building feature classification result; Among them, represents the classification score, refers to the score of a single element, reflecting its matching degree with the model in the feature library, is the number of elements, is the highest score, used for normalizing the score, is the weight parameter, emphasizing the key of certain features, is the bias, adjusting the score threshold to ensure the fairness and matching of the score.

[0039] Formula: ; The advantage of the formula is that by introducing the weight factor and the bias term , the characteristics of the scoring system for matching different building elements can be flexibly adjusted, enhancing the applicability and flexibility of the model. At the same time, using the method of weighted average can more fairly handle the influence of multiple elements, ensuring the fairness and accuracy of the score result.

[0040] Detailed explanation of the formula and the derivation process of the formula calculation: There is a building element data set set, including 3 elements, and the matching scores are respectively , , ; The weight factor , the bias term . The highest score . The formula calculation is as follows: ; ; ; The results show that, on average, the classification scores of these three building elements are 5.444, which reflects the overall matching degree of the elements with the feature library according to the set weights and biases. This score helps to analyze how the elements are effectively classified according to the preset criteria.

[0041] Please refer to Figure 6 , and the steps to obtain the structural damage prediction results are specifically as follows: Analyze the image data collected by the drone camera, extract the building structure features in the images, identify the structural elements through image processing algorithms, create feature vectors for each element, and generate a building element feature library; Extract the building structure features from the images captured by the drone camera through image processing techniques, including using methods such as edge detection, morphological operations, and texture analysis to identify and label the building elements. Creating feature vectors for each element includes calculating its geometric, optical, and texture attributes. This feature vector will provide a data basis for the next step of model training and feature comparison. The focus of this stage is to ensure the accuracy and usability of the data through image processing algorithms, so as to provide support for the automatic classification and recognition of building elements. Through these image analysis steps, a building element feature library is generated, and the feature library can effectively support the subsequent structural damage identification and classification work.

[0042] Integrate the building element feature library and the building feature classification results, use a deep learning model to compare and analyze the element features, classify different types of structural damage, and output the damage type classification results; Combine the building element feature library and the existing building feature classification results, adopt a deep learning model for feature analysis and matching, including using algorithms such as convolutional neural networks or support vector machines to compare and classify different types of structural damage. The model identifies the characteristic patterns of various damages through training, including cracks, spalling, or corrosion. The output damage type classification results provide a scientific basis for damage assessment. This process involves the analysis and identification of the features of each damage type, ensuring the accuracy and practicality of the classification results.

[0043] Conduct a quantitative analysis of the damage type classification results, calculate the probability of future structural changes, using the formula: ; Estimate the severity of the structural damage in the future time period and generate the structural damage prediction results; Among them, represents the future change probability, Assign weights according to the criticality of the damage type, represents the damage depth, represents the time since the damage was discovered, Adjust the baseline prediction value, Adjust the ratio of damage depth to time to provide a sensitive adjustment to the rate of damage deterioration. Represents the total number of damage types.

[0044] Formula: ; The benefit of the formula is that by weighing the damage depth and time, it can dynamically evaluate the deterioration speed of structural damage and provide a more accurate prediction of future changes.

[0045] Detailed explanation of the formula and the derivation process of formula calculation: There are three types of damage set, with weights being 0.5, 0.3, 0.2 respectively, and the damage depths being 2, 3, 1 respectively, and the self-damage excavation times being 12, 15, 10 years respectively, and the baseline deviation , calculated as follows: 1. For each type, calculate ; ; ; ; 2. Apply the weights and the baseline deviation: ; The results show that under the given damage type and its corresponding depth and time conditions, the prediction index of future structural changes is , indicating that under the current parameter settings, the rate of damage deterioration is at a medium to low level, thus providing a basis for maintenance decisions.

[0046] A building survey method based on UAV mapping. A building survey method based on UAV mapping is executed based on the above-mentioned building survey system based on UAV mapping, and includes the following steps: S1: Based on the current light intensity, temperature, and humidity data, dynamically adjust the exposure and sensitivity parameters of the UAV camera, adjust the size and type of the filtering window, and apply it to the preliminary filtering process of the point cloud data to remove the influence of environmental factors and obtain environment-adapted point cloud data; S2: Conduct a spatial analysis of the environment-adapted point cloud data, identify the gaps and discontinuities in the data, automatically fill the gap areas, and supplement the point cloud data set through continuity evaluation to generate a complete point cloud data set; S3: Based on the complete point cloud data set, analyze the building elements, compare multiple elements in the preset building feature library, score the matching degree of each element, and conduct classification and marking of the building features to generate a building feature classification result; S4: Combine the image data collected by the drone camera and the building feature classification results, detect the structural damage of the building through visual analysis, identify and classify the damage types, predict future structural changes, and obtain the structural damage prediction results.

[0047] The above are only the preferred embodiments of the present invention, and do not limit the present invention in other forms. Any person skilled in the art may use the technical content disclosed above to make changes or modifications into equivalent embodiments with equivalent changes and apply them to other fields. However, as long as it does not depart from the technical solution content of the present invention, any simple modification, equivalent change, and modification made to the above embodiments based on the technical essence of the present invention still fall within the protection scope of the technical solution of the present invention.

Claims

1. A building measurement system based on drone mapping, characterized in that: The system comprises: The environmental perception monitoring module collects real-time light intensity, temperature, and humidity data, records environmental changes, analyzes the difference between real-time data and standard thresholds, dynamically adjusts drone camera parameters, including exposure and sensitivity, to match the current environment, and records environmental acquisition parameters; The adaptive filtering module uses the environmental acquisition parameters to analyze the impact of the current environment on the data quality, matches the filter window size and type, performs filtering on the collected point cloud data, removes the errors introduced by the environmental variables, and outputs optimized point cloud data; The data interpolation optimization module uses spatial analysis technology to automatically fill in the missing areas according to the vacancies and discontinuities in the optimized point cloud data, and generates a complete point cloud data set through continuity evaluation; The building feature recognition module processes the complete point cloud data set, analyzes the building elements, scores the matching degree of each element by comparing multiple elements with a preset building feature library, organizes and marks them by category, and generates a building feature classification result; The damage prediction analysis module combines the image data collected by the drone camera and the classification results of the building features, uses visual analysis to detect structural damage, identifies and classifies damage types, predicts future changes based on the damaged structural characteristics, and generates structural damage prediction results.

2. The building measurement system based on drone mapping according to claim 1 is characterized in that: The environmental acquisition parameters include exposure and sensitivity, the optimized point cloud data includes filter window size, filter type, and error removal, the complete point cloud data set specifically includes missing area filling and data continuity, the building feature classification results include element matching scores and category tags, and the structural damage prediction results specifically refer to damage type identification, damage classification, and future change prediction.

3. The building measurement system based on drone mapping according to claim 2 is characterized in that: The steps of recording the environmental acquisition parameters are specifically as follows: Collect real-time light intensity data and ambient temperature, obtain light intensity matrix and temperature matrix through preliminary processing of sensor data, and generate environmental data set; Analyze the light intensity matrix in the environmental data set, calculate the difference between the light intensity matrix and the light standard threshold, adjust the exposure parameters of the drone camera according to the difference value, and obtain a preliminary exposure setting; Combining the preliminary exposure settings and the current ambient temperature matrix, the formula is used: ； Adjust the exposure parameters of the drone camera to the optimal level and obtain the exposure parameters; in, represents the exposure parameters, represents the light intensity, represents temperature, Represents the adjustment factor of the ambient light, ensuring that the exposure adjustment is proportional to the light intensity. Represents the smoothing coefficient of the exposure adjustment due to environmental changes, reducing the impact of small fluctuations on exposure. Represents the absolute value of light intensity and temperature difference, used to capture the sensitivity of environmental changes; According to the exposure parameters and the humidity information in the current environmental data set, the sensitivity of the camera is adjusted to match the current environmental conditions, and the environmental acquisition parameters are recorded.

4. The building measurement system based on drone mapping according to claim 3 is characterized in that: The steps for obtaining the optimized point cloud data are specifically as follows: Analyze the environmental acquisition parameters, including the measurement of real-time light, temperature, and humidity, extract parameters from sensor data, generate an environmental parameter matrix, and obtain environmental parameter configuration; Based on the environmental parameter configuration, a filter window size and type matching the current environmental conditions are selected, the matching of the filter is determined through experimental data, a configuration is prepared for point cloud data filtering, and a filter configuration is obtained; Apply the filter configuration to process the raw point cloud data using the formula: ； Perform filtering processing and reduce the error introduced by environmental variables through weighted average method to obtain optimized point cloud data; in, Represents the optimized point cloud data, represents the i-th data point in the original point cloud dataset, The filter configuration is the average value obtained according to the environmental parameters. Represents the environmental sensitivity adjustment coefficient, which is used to adjust the relationship between data sensitivity and environmental variables. Represents the absolute value of the difference between the original data and the filtered average value, which is used to calculate the deviation of the point cloud data. Represents the total number of data points, the number of data points in the original point cloud data, used to calculate the average value.

5. The building measurement system based on drone mapping according to claim 4 is characterized in that: The steps for obtaining the complete point cloud data set are specifically as follows: Identify the missing areas in the optimized point cloud data, mark the boundaries and positions of the areas by using a spatial mapping technique, and generate a missing area index table; Using the missing area index table, selecting the spatial analysis technology that best matches the current environment, performing a filling operation, filling each missing area evenly, and obtaining preliminary filled point cloud data; The continuity and consistency of the initially filled point cloud data are evaluated using the formula: ； The quality of filling is evaluated by comprehensively analyzing the position difference between each point and its neighboring points to generate a complete point cloud dataset; in, Represents the complete point cloud dataset, represents the i-th point in the filled point cloud, is the spatial average position of adjacent points, is the total number of evaluation points, is the maximum difference between the point and the average position, Measures the distance between the filled point and the surrounding average position, is the maximum value of the point difference.

6. The building measurement system based on drone mapping according to claim 5 is characterized in that: The steps for obtaining the building feature classification results are specifically as follows: Extracting architectural elements from the complete point cloud data set, performing spatial positioning and structural characteristics recognition, extracting the shape and material characteristics of each element through a visual processing algorithm, and generating an element feature index table; Using the element feature index table to perform data analysis, comparing with a preset building feature library, using a statistical analysis method to calculate the degree of matching, assigning a preliminary matching score to each element, and generating a matching score result; The matching score results are comprehensively analyzed and the score is calculated using the formula: ； The classification results of each element are determined by comprehensive scoring and adjustment parameters to generate the classification results of building characteristics; in, Represents the classification score, Refers to the score of a single element, reflecting its matching degree with the model in the feature library. is the number of elements, is the highest score, used for normalized score, is the weight parameter, emphasizing the key of certain features, To offset bias, adjust the scoring threshold to ensure fairness and matching of the scoring.

7. The building measurement system based on drone mapping according to claim 6 is characterized in that: The steps for obtaining the structural damage prediction results are specifically as follows: Analyze the image data collected by the drone camera, extract the building structure features in the image, identify the structural elements through image processing algorithms, create feature vectors for each element, and generate a building element feature library; Integrate the building element feature library and the building feature classification results, use a deep learning model to compare and analyze element features, classify differentiated structural damage types, and output damage type classification results; The damage type classification results are quantitatively analyzed to calculate the probability of future structural changes using the formula: ； Estimate the severity of structural damage in the future time period and generate structural damage prediction results; in, represents the probability of future changes, Assign weights based on the criticality of the damage type, represents the damage depth, represents the time since the injury was discovered, Adjusting the baseline predictions, Adjusts the ratio of damage depth to time, providing sensitive adjustment of the damage progression rate, Represents the total number of damage types.

8. A building measurement method based on drone mapping, characterized in that: The construction measurement system based on drone mapping according to any one of claims 1 to 7 comprises the following steps: Based on the current light intensity, temperature, and humidity data, the exposure and sensitivity parameters of the drone camera are dynamically adjusted, and the filter window size and type are adjusted. This is applied to the preliminary filtering of the point cloud data to remove the influence of environmental factors and obtain environmentally adapted point cloud data. Performing spatial analysis on the environmental adaptation point cloud data, identifying gaps and discontinuities in the data, automatically filling in the gaps, supplementing the point cloud data set through continuity assessment, and generating a complete point cloud data set; Based on the complete point cloud data set, the architectural elements are analyzed, multiple elements in the preset architectural feature library are compared, the matching degree of each element is scored, the architectural features are classified and marked, and the architectural feature classification results are generated; Combining the image data collected by the drone camera and the classification results of the building features, the structural damage of the building is detected through visual analysis, the damage type is identified and classified, and future structural changes are predicted to obtain structural damage prediction results.

Images