An engineering surveying and mapping data intelligent acquisition, analysis and processing system and method

Abstract

The present application relates to the field of engineering surveying and mapping technology, in particular to an engineering surveying and mapping data intelligent acquisition, analysis and processing system and method, mainly applied to data acquisition, analysis and processing in the process of engineering construction; the system covers image acquisition, data classification, database storage, data processing, data analysis and decision feedback modules; the image acquisition module obtains an engineering surveying and mapping image feature set, the data classification module receives and classifies and stores the data; the database stores the classified data, the data processing module divides and calculates the importance of the image feature nodes; the data analysis module identifies abnormal data, and the decision and data feedback module receives the identification result and adjusts the data threshold; the system significantly improves the surveying and mapping precision through feature node processing and importance calculation; at the same time, the multi-level comparison mechanism improves the accuracy of abnormal detection, and provides efficient and accurate data support for engineering construction.

Description

Technical Field

[0001] This invention relates to the field of engineering surveying and mapping technology, and in particular to an intelligent data acquisition, analysis and processing system and method for engineering surveying and mapping, which is mainly applied to data acquisition, analysis and processing during engineering construction. Background Technology

[0002] With the increasing complexity and scale of construction projects, traditional engineering surveying techniques are no longer sufficient to meet the demands of modern engineering construction for data accuracy and efficiency. Current engineering surveying technology faces several major problems: First, the lack of intelligent methods in the data acquisition process leads to low acquisition efficiency and high manpower consumption; second, the limited data processing methods cannot effectively handle complex image features, affecting surveying accuracy; third, the lack of an effective abnormal data identification mechanism affects the reliability of surveying results; and finally, the lack of an effective correlation and feedback mechanism between engineering surveying data and actual construction progress makes it difficult to adjust construction plans in a timely manner.

[0003] In existing technologies, engineering surveying data processing mainly relies on manual experience and judgment, lacking systematic and intelligent processing methods. For example, Chinese patent application CN119336773A, a three-dimensional model data management system and method for engineering surveying, while achieving basic data processing functions, lacks the ability to deeply analyze image features and cannot effectively identify and process abnormal data.

[0004] Therefore, there is an urgent need for a system and method that can intelligently collect, accurately analyze, and efficiently process engineering surveying data in order to improve the accuracy and efficiency of engineering surveying and meet the needs of modern engineering construction. Summary of the Invention

[0005] The purpose of this invention is to provide an intelligent data acquisition, analysis and processing system and method for engineering surveying, aiming to solve the problems of low data acquisition efficiency, single processing method, weak abnormal data identification capability and lack of effective construction progress feedback mechanism in the existing technology.

[0006] This invention proposes an intelligent acquisition, analysis and processing system for engineering surveying data, comprising:

[0007] The image acquisition module is used to acquire image feature sets during the engineering surveying process;

[0008] A data classification module, connected to the image acquisition module, is used to receive the image feature set and classify and store it.

[0009] A database, connected to the data classification module, is used to store the classified data;

[0010] The data processing module, connected to the database, includes an image feature node search and calculation module and a node search value comprehensive calculation module, used to divide the feature node set of image features and calculate the importance of nodes;

[0011] A data analysis module, connected to the data processing module, is used to identify abnormal data in the processed data.

[0012] The decision-making module and the data feedback module are connected to the data analysis module and are used to receive the identification results and adjust the data thresholds.

[0013] Preferably, the image feature node search and calculation module is used for:

[0014] Representing image features as a two-dimensional vector: ;

[0015] The image features are divided into feature nodes to obtain... ;

[0016] Calculate the importance value of all feature nodes in the image features.

[0017] Preferably, the calculation of the importance value of the feature node includes:

[0018] Obtain the input values ​​of image features in the image recognition process and compare them with the recognition input values ​​of standard image features;

[0019] Determine the node search value threshold. When the input value is less than the node search value threshold, obtain the feature nodes in the image feature division whose feature node data value is greater than the threshold, record them as important nodes and form an important node set C.

[0020] Define a two-dimensional search space D for the set C, denoted as [d1][d2], where d1 and d2 are the two-dimensional coordinates of the search nodes;

[0021] Compare the nodes in set C with the nodes in search space D, and set the comparison range value β, where β is a value in the interval (0,1).

[0022] Preferably, the node search value comprehensive calculation module is used for:

[0023] Get node search value ;

[0024] Normalize the obtained node search values;

[0025] The importance of all feature nodes in an image is calculated using the following formula:

[0026] ,

[0027] wherein is the importance degree of the image feature, is the average search value of the obtained node, is the number of nodes, is a coefficient and > 1.

[0028] Preferably, the data analysis module includes:

[0029] A graphic recognition error evaluation unit, configured to receive the data processed by the data processing module, import it into an image feature recognition model, set an N*N training matrix, and calculate an error recognition matrix;

[0030] An image resolution evaluation unit, configured to evaluate the image data resolution;

[0031] Among them, the abnormal data recognition is performed by setting the comparison data features and the standard difference value, and marking the data with the difference between the difference value and the standard difference value greater than the abnormal value as abnormal data.

[0032] Preferably, the image resolution evaluation unit is used for:

[0033] Obtaining the number n of image data processed by the data processing module;

[0034] Determining the average resolution A1 of the image data, and the calculation formula is: A1 = , where R is the image data resolution;

[0035] Comparing the average resolution A1 with the resolution threshold A2;

[0036] When A1 < A2, it is determined that the image data resolution is less than the threshold; when A1 > A2, it is determined that the image data resolution is greater than the threshold.

[0037] Preferably, the decision module is used for:

[0038] Receiving the recognition error and resolution evaluation data transmitted by the data analysis module;

[0039] Setting an error threshold and a resolution value threshold;

[0040] When the error is less than the error threshold and the resolution is greater than the resolution value threshold, it is determined that no adjustment is required;

[0041] When at least one of them is not satisfied, the data is fed back to the data processing module for adjustment, and the data is sent to the data processing module for processing again.

[0042] Preferably, the data feedback module is used for:

[0043] Receiving the data sent by the server;

[0044] The data is processed based on the data nodes and data thresholds in the distributed data.

[0045] The processing results are then sent back to the server.

[0046] Adjust the thresholds in the data classification module based on abnormal data.

[0047] As a preferred option, it also includes:

[0048] The 3D point cloud acquisition module is used to acquire 3D point cloud data of the building to be surveyed according to the construction progress, and obtain multiple first 3D point cloud images.

[0049] The monitoring image acquisition module is used to acquire monitoring images corresponding to the building to be surveyed;

[0050] The progress management module is used to obtain the deviation coefficient based on the second 3D point cloud image;

[0051] The progress adjustment module is used to adjust the construction progress according to the deviation coefficient.

[0052] The data processing module is further configured to process multiple first three-dimensional point cloud images based on the monitoring images to obtain multiple second three-dimensional point cloud images.

[0053] A method for intelligent acquisition, analysis and processing of engineering surveying data, comprising:

[0054] The data acquisition step involves using acquisition equipment to transmit the image feature sets from the engineering surveying process to the data classification module, and then classifying and storing them in the database.

[0055] The processing steps are as follows: the data processing module extracts and processes the data stored in the database. In the image feature node search calculation, all feature nodes in the feature node set of the image feature are divided to obtain the target image features after division. In the node search value comprehensive calculation, the importance of all feature nodes in the divided target image features is calculated and node search values ​​are formed.

[0056] In the analysis steps, the data analysis module will identify abnormal data in the data and dataset processed by the data processing module, and feed the identification results back to the decision-making module and the data feedback module.

[0057] In the decision-making process, the decision module receives recognition error and resolution evaluation data transmitted by the data analysis module, sets error thresholds and resolution thresholds, and decides whether adjustments are needed based on the evaluation results.

[0058] In the feedback step, the data feedback module receives data sent from the server, processes the data according to the data nodes and data thresholds in the data, and feeds it back to the server. At the same time, it adjusts the thresholds in the data classification module according to abnormal data.

[0059] This invention establishes an image feature node processing mechanism, a multi-level abnormal data identification algorithm, an adaptive threshold adjustment mechanism, and an integrated system for 3D point cloud and construction monitoring, thereby achieving intelligent acquisition, accurate analysis, and efficient processing of engineering surveying data and improving the accuracy and efficiency of engineering surveying.

[0060] The beneficial effects of this invention are as follows: 1) It significantly improves the accuracy of surveying and mapping through feature node processing and importance calculation; 2) It improves the accuracy of anomaly detection through a multi-level comparison mechanism; 3) It achieves adaptive optimization of the system through closed-loop feedback and threshold adjustment; 4) It achieves accurate monitoring of construction progress through the integration of 3D point cloud and monitoring image; 5) It further enhances the system's adaptability to complex environments and processing accuracy through multi-scale feature fusion and dynamic feature weight adaptive adjustment. Attached Figure Description

[0061] Figure 1 This is a structural block diagram of the intelligent acquisition, analysis and processing system for engineering surveying data of the present invention;

[0062] Figure 2 This is a flowchart illustrating the workflow of the image feature node search and calculation module of the present invention.

[0063] Figure 3 This is a flowchart of the node search value comprehensive calculation module of the present invention;

[0064] Figure 4 This is a schematic diagram of the data analysis module of the present invention;

[0065] Figure 5 This is a flowchart of the decision-making module and data feedback module of the present invention;

[0066] Figure 6 This is a flowchart illustrating the workflow of the 3D point cloud acquisition module and the monitoring image acquisition module of the present invention.

[0067] Figure 7 This is a flowchart of the intelligent acquisition, analysis and processing method for engineering surveying data according to the present invention;

[0068] Figure 8 This is a flowchart illustrating the workflow of the multi-scale feature fusion enhancement module of the present invention.

[0069] Figure 9 This is a schematic diagram of the dynamic feature weight adaptive adjustment algorithm of the present invention. Detailed Implementation

[0070] Please refer to the attached document. Figure 1-9 The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments.

[0071] like Figure 1 As shown, the intelligent acquisition, analysis and processing system for engineering surveying data provided by the present invention includes an image acquisition module 1, a data classification module 2, a database 3, a data processing module 4, a data analysis module 5, a decision-making module 6 and a data feedback module 7.

[0072] Image acquisition module 1 is used to acquire image feature sets during the engineering surveying process. Preferably, image acquisition module 1 can use equipment such as a high-resolution camera or laser scanner to capture detailed image data of the engineering site. In one embodiment of the present invention, the acquisition frequency of image acquisition module 1 can be dynamically adjusted according to the progress of the project. For example, it can be set to once per hour during critical construction phases and once per day during regular construction phases to balance the comprehensiveness of data acquisition and the efficiency of system resource utilization.

[0073] The data classification module 2 is connected to the image acquisition module 1 and is used to receive image feature sets and classify and store them. In one embodiment of the present invention, the data classification module 2 adopts a multi-level classification strategy. First, it performs preliminary classification based on image content type (such as buildings, terrain, facilities, etc.), then performs secondary classification based on image quality (such as sharpness, integrity, etc.), and finally performs tertiary classification based on timestamps. This multi-level classification method helps to improve the efficiency of subsequent data processing and analysis.

[0074] Database 3 is connected to data classification module 2 and is used to store the classified data. Preferably, database 3 adopts a distributed storage architecture to ensure high availability and reliability of the data. In one embodiment of the present invention, database 3 also implements a data version control function, which can track the data change history and facilitate data backtracking and comparative analysis.

[0075] The data processing module 4 is connected to the database 3 and includes an image feature node search and calculation module 41 and a node search value comprehensive calculation module 42, which are used to divide the feature node set of image features and calculate the importance of nodes.

[0076] Specifically, the image feature node search and calculation module 41 is used to: represent the two-dimensional vector of image features as... The image features are divided into feature nodes to obtain... ; Calculate the importance values ​​of all feature nodes in the image features. In a preferred embodiment of the present invention, This indicates the importance of image features for image recognition, and its value ranges from 0 to 1. For example, for building outline features, possible... A value of 0.85 indicates that this feature is highly important for building identification; while for non-critical texture features, A value of 0.2 indicates that it is of low importance.

[0077] The calculation of the importance value of feature nodes includes: obtaining the input value of image features in the image recognition process and comparing it with the recognition input value of standard image features; determining the node search value threshold, and when the input value is less than the node search value threshold, obtaining the feature nodes whose feature node data value is greater than the threshold among the feature nodes divided by image features, recording them as important nodes and forming an important node set C; and setting a two-dimensional search space D for set C, denoted as... ,in and The coordinates of the search nodes are given in two dimensions; the nodes in set C are compared with the nodes in search space D, and the comparison range is set. ,in The value is within the interval (0,1).

[0078] In one embodiment of the present invention, the node search value threshold can be set to 0.65. This value is derived from the analysis of a large amount of engineering surveying data and can effectively balance computational efficiency and recognition accuracy. When the feature node data value is greater than 0.7 (empirical threshold), the node is considered an important node. (Comparison range value) It is usually set to 0.3, a value that allows for sufficient matching accuracy while maintaining a certain degree of flexibility.

[0079] The node search value comprehensive calculation module 42 is used to: obtain node search values Normalize the obtained node search values; calculate the importance of all feature nodes in the image features using the following formula:

[0080] ,

[0081] in, Assess the importance of image features. To obtain the average search value of the nodes, The number of nodes Coefficient and .

[0082] In a preferred embodiment of the present invention A value typically set between 1.5 and 2.5 is used. This range, determined based on experimental data analysis, appropriately amplifies the influence of important features without overly favoring a single feature. For example, when processing key features such as building outlines, this value can be used. When dealing with general environmental characteristics, one can use... =1.7.

[0083] This calculation formula takes into account the relative importance of nodes. and the overall complexity of features ( This achieves precise quantification of node importance. Normalization ensures comparability between images of different complexities, while the coefficients... This provides a flexible mechanism for adjusting importance weights.

[0084] Data analysis module 5 is connected to data processing module 4 and is used to identify abnormal data in the processed data. For example... Figure 4 As shown, the data analysis module 5 includes an image recognition error evaluation unit 51 and an image resolution evaluation unit 52.

[0085] The image recognition error evaluation unit 51 is used to receive the data processed by the data processing module 4, import it into the image feature recognition model, and set... The training matrix is ​​trained, and the error identification matrix is ​​calculated. In a preferred embodiment of the invention, the size of the training matrix is ​​typically set to [size missing]. to This range allows for control of computational complexity while maintaining sufficient accuracy. For the error identification matrix, the matrix element value is 0 when the difference between the image data and the identified data is less than the contrast error threshold; and the matrix element value is 1 when the difference is greater than the contrast error threshold. The contrast error threshold is typically set between 0.15 and 0.25. This range is determined based on practical engineering surveying experience and can effectively identify abnormal data while avoiding oversensitivity.

[0086] The image resolution evaluation unit 52 is used to evaluate the resolution of image data. Specifically, the image resolution evaluation unit 52 is used to: obtain the number of image data processed by the data processing module 4. Determine the average resolution of the image data. The calculation formula is:

[0087] ,

[0088] in, Image data resolution; average resolution With resolution threshold Compare; when When the image data resolution is less than a threshold, it is determined that the image data resolution is less than the threshold. When the resolution of the image data is greater than the threshold, it is determined that the resolution of the image data is greater than the threshold.

[0089] In one embodiment of the present invention, the resolution threshold The standard setting is 300 dpi (dots per inch). This value is determined based on engineering surveying experience and industry standards, ensuring that the image data has sufficient detail for subsequent analysis and processing.

[0090] Outlier identification identifies outliers by defining comparative data features and a standard difference value, marking data whose difference between the actual difference and the standard difference value is greater than the outlier value. Specifically, the comparative data features are defined as follows: The importance of comparison is The data features of the image after processing by data processing module 4 are as follows: And the data characteristics are The importance of This is achieved by setting a difference value and a standard difference value for comparison. and The difference is obtained by subtracting the corresponding values. and The difference between the corresponding positions is the standard deviation value. If the difference between the deviation value and the standard deviation value is greater than the outlier, the data is considered outlier.

[0091] In a preferred embodiment of the invention, outliers are typically set between 0.3 and 0.5. This range is derived from the analysis of a large amount of engineering surveying data and can effectively identify abnormal data while avoiding misjudgment. For example, an outlier of 0.35 can be used in precision surveying scenarios, while an outlier of 0.45 can be used in general surveying scenarios.

[0092] The decision module 6 and the data feedback module 7 are connected to the data analysis module 5 and are used to receive the recognition results and adjust the data thresholds.

[0093] The decision module 6 is used to: receive the recognition error and resolution evaluation data transmitted by the data analysis module 5; set the error threshold and resolution value threshold; determine that no adjustment is needed when the error is less than the error threshold and the resolution is greater than the resolution value threshold; and when at least one condition is not met, feed the data back to the data processing module 4 for adjustment and send the data back to the data processing module 4 for processing.

[0094] In a preferred embodiment of the invention, the error threshold is typically set between 0.2 and 0.3, and the resolution threshold is typically set between 250 dpi and 350 dpi. These threshold ranges are determined based on practical experience in engineering surveying and mapping, ensuring the accuracy and efficiency of data processing. For example, in engineering surveying with high accuracy requirements, the error threshold can be set to 0.22, and the resolution threshold can be set to 320 dpi; while in engineering surveying with general accuracy requirements, these thresholds can be set to 0.28 and 280 dpi, respectively.

[0095] The data feedback module 7 is used to: receive data sent by the server; process the data according to the data nodes and data thresholds in the sent data; send the processing results back to the server; and adjust the thresholds in the data classification module 2 according to abnormal data.

[0096] In one embodiment of the present invention, the data feedback module 7 employs an adaptive threshold adjustment strategy to dynamically adjust the threshold of the data classification module 2 based on the characteristics of the abnormal data. For example, when a large amount of abnormal data of a specific type is detected, the system automatically lowers the classification threshold of the corresponding category to increase the system's sensitivity to that type of anomaly; while when the amount of abnormal data decreases, the system adjusts the threshold accordingly to maintain overall performance. This adaptive mechanism enables the system to continuously optimize its performance and adapt to different engineering environments and data characteristics.

[0097] like Figure 6 As shown, the system of the present invention also includes: a three-dimensional point cloud acquisition module 8, a monitoring image acquisition module 9, a progress management module 10, and a progress adjustment module 11.

[0098] The 3D point cloud acquisition module 8 is used to acquire 3D point cloud data of the building to be surveyed according to the construction progress, obtaining multiple first 3D point cloud images. In a preferred embodiment of the present invention, the 3D point cloud acquisition module 8 adopts laser scanning technology, which can capture the 3D structure of the building with millimeter-level precision. The acquisition frequency can be adjusted according to the construction stage; for example, it can be set to daily acquisition during the critical structure construction stage, and weekly acquisition during the general stage.

[0099] The monitoring image acquisition module 9 is used to acquire monitoring images corresponding to the building to be surveyed. Preferably, the monitoring image acquisition module 9 employs a high-resolution camera array to simultaneously capture images of the building from multiple angles, ensuring coverage of all key areas. In one embodiment of the present invention, the resolution of the monitoring images is not less than 5 megapixels to ensure that the images contain sufficient detail information.

[0100] The progress management module 10 is used to obtain the deviation coefficient based on the second three-dimensional point cloud image. Specifically, the progress management module 10 first determines the first representation area based on the monitoring image, then determines the corresponding second representation area based on the second three-dimensional point cloud image, calculates the construction progress ratio of the first representation area (completed construction area / total area), then determines the third ratio corresponding to the three-dimensional point cloud image, and finally calculates the deviation coefficient (the difference between the third ratio and the construction progress ratio).

[0101] In a preferred embodiment of the present invention, the construction progress ratio is calculated using a region segmentation and pixel counting method, which can accurately estimate the construction progress within an error range of ±2%. The deviation coefficient typically fluctuates within the range of ±0.1; when it exceeds ±0.15, the system will determine it as a significant deviation, requiring intervention and adjustment.

[0102] The progress adjustment module 11 is used to adjust the construction progress based on a deviation coefficient. Specifically, the progress adjustment module 11 first determines whether the deviation coefficient is greater than a preset deviation threshold (usually set to 0.15). If it exceeds the threshold, the construction progress is adjusted, and an alarm message is sent to the management personnel. The adjustment method includes adjusting the constructed areas of the regional unit according to a third ratio to ensure that the actual construction progress is consistent with the planned progress.

[0103] In this invention, the data processing module 4 is further used to process multiple first three-dimensional point cloud images based on the monitoring images to obtain multiple second three-dimensional point cloud images. The processing includes steps such as point cloud registration, filtering, noise reduction, and model reconstruction. Through these processes, the system can generate a high-precision three-dimensional model for construction progress analysis and monitoring.

[0104] According to an innovative embodiment of the present invention, such as Figure 8 As shown, the system also includes a multi-scale feature fusion enhancement module 12, which is connected between the data processing module 4 and the data analysis module 5. It is used to fuse image features at different scales to improve feature expression capabilities and recognition accuracy.

[0105] The multi-scale feature fusion enhancement module 12, based on pyramid feature extraction and multi-level feature fusion strategies, achieves comprehensive feature analysis from coarse-grained to fine-grained. The specific implementation algorithm is as follows:

[0106] First, the input image Build Layered pyramid structure ,in For the original image, This is the lowest resolution layer. The aspect ratio of each layer's image is:

[0107] ,

[0108] Next, feature maps are extracted for each image layer. , thus obtaining the feature pyramid Feature extraction employs an improved convolution operator, with the kernel function defined as:

[0109] ,

[0110] in, This is the gain factor, typically set to 1.2; The Gaussian parameter controls the spatial range of the kernel function and is dynamically adjusted according to the pyramid levels. , This is a frequency parameter that controls the ripple density; a typical value is... Set the phase offset to 0.

[0111] Then, a top-down feature fusion process is implemented. For the hierarchy... Fusion characteristics The calculation formula is:

[0112] ,

[0113] Here, Upsample() represents the upsampling operation, which will... Adjusting the spatial dimension to match Matching; To fuse weights and characterize the importance of features at different levels, the calculation formula is as follows:

[0114] ,

[0115] in, hierarchical The initial weights are set to This ensures that lower-level (high-resolution) features have higher weights.

[0116] Finally, the features from each layer are further integrated to obtain the final feature representation. :

[0117] ,

[0118] in, The final fusion coefficients are adaptively calculated using an attention mechanism:

[0119] ,

[0120] ,

[0121] Where GAP() represents global average pooling operation, and FC() represents a fully connected layer that maps the pooling result to a scalar value.

[0122] In a preferred embodiment of the invention, the number of pyramid layers Typically, it is set to 3 to 5 layers, a range that balances computational efficiency and feature representation capability. For typical engineering surveying scenarios, a 4-layer pyramid structure is used. This will achieve the best results. The 0.5 in the initial weight calculation formula is an empirical coefficient and can be adjusted appropriately according to the specific application scenario. For example, for mapping buildings with rich details, this value can be increased to 0.7 to enhance the influence of high-resolution features.

[0123] The introduction of the multi-scale feature fusion enhancement module 12 significantly improves the system's adaptability to complex environments. Through a pyramid structure and multi-level fusion, the system can simultaneously capture global structural features and local detail features, effectively processing image content at different scales. Experimental results show that this module improves feature recognition accuracy by approximately 15%, with a particularly significant advantage when handling complex textures and blurred boundaries.

[0124] like Figure 9 As shown, another innovation is the introduction of a dynamic feature weight adaptive adjustment algorithm in the node search value comprehensive calculation module 42. Traditional feature importance calculation uses fixed weights, which are difficult to adapt to the needs of different scenarios. The dynamic feature weight adaptive adjustment algorithm proposed in this invention can automatically adjust feature weights according to environmental complexity and task requirements, further improving the system's adaptability and accuracy.

[0125] The core of the dynamic feature weight adaptive adjustment algorithm is to extend the original weight calculation formula into a time-varying adaptive form:

[0126] ,

[0127] in, For time Time feature nodes The degree of importance; For dynamic adjustment coefficients; and They are time Time node Search values ​​and average search values; This is an environmental adaptability factor, reflecting the suitability of a feature in the current environment.

[0128] Dynamic adjustment coefficient Automatically calculated using the environment complexity evaluation function:

[0129] ,

[0130] in, The base coefficient is typically set to 2.0; The adjustment range is typically 0.5. The complexity index of the current environment is calculated using the following formula:

[0131] ,

[0132] in, Indicates the first in the image The proportion of class features This represents the total number of feature categories. and These represent the minimum and maximum complexity values ​​observed in history, respectively, and are used for normalization.

[0133] Environmental adaptability factors Dynamically updated through a feedback learning mechanism:

[0134] ,

[0135] in, The learning rate is typically set to 0.05. For the system in time Error index; This represents the gradient of the error with respect to the factor.

[0136] To prevent excessive weight adjustment, regularization constraints are introduced:

[0137] ,

[0138] in, Indicates the truncation function, ensuring Limited to Within the range, it is usually set , .

[0139] In practical applications, the system automatically adjusts these parameters based on historical data. For example, when continuously monitoring highly complex environments (such as complex building structures or varied terrain), the system will increase... Values ​​enhance the ability to distinguish features; however, in simple environments, the system will reduce [its effectiveness]. This improves computational efficiency. Similarly, continuous optimization is achieved through error feedback. This enables the system to automatically identify and enhance the most effective feature types in the current environment.

[0140] The introduction of a dynamic feature weight adaptive adjustment algorithm overcomes the limitations of traditional fixed-weight methods, enabling the system to automatically adjust its feature processing strategy based on environmental changes and task requirements. Experimental results show that this algorithm improves the system's performance by approximately 20% in complex and variable environments, particularly excelling under conditions of changing lighting and seasonal transitions.

[0141] like Figure 7 As shown, the present invention also provides an intelligent acquisition, analysis and processing method for engineering surveying data, comprising the following steps:

[0142] The data acquisition step involves using acquisition equipment to transmit image feature sets from the engineering surveying process to the data classification module 2, and then classifying and storing them in the database 3. Preferably, the acquisition equipment includes high-resolution cameras, laser scanners, or drones, capable of acquiring engineering site data from multiple angles and dimensions. The classification and storage employs a multi-level classification strategy to ensure the orderly organization and efficient retrieval of the data.

[0143] In the processing steps, the data processing module 4 extracts and processes the data stored in the database 3. In the image feature node search calculation, all feature nodes in the feature node set of the image feature are divided to obtain the target image features after division. In the node search value comprehensive calculation, the importance of all feature nodes in the divided target image features is calculated and node search values ​​are formed.

[0144] Specifically, the image feature node search calculation process includes: representing image features as two-dimensional vectors. ; divide it into independent nodes The node search value threshold is determined by comparing it with standard image features; important nodes are marked to form a set C; node matching is performed in the two-dimensional search space D. The comprehensive calculation process of node search values ​​includes: normalizing the node search values; applying the formula... Calculate node importance. In the innovative embodiment, this step applies a dynamic feature weight adaptive adjustment algorithm, extending the formula to... This enables automatic optimization of feature weights.

[0145] In the analysis step, data analysis module 5 identifies anomalous data from the data processed by data processing module 4 and the dataset, and feeds back the identification results to decision module 6 and data feedback module 7. Anomalous data identification employs a multi-level comparison mechanism, considering both differences in data features and importance. By comparing the difference value with the standard difference value and a preset anomaly value, anomalous data is effectively identified. In the innovative embodiment, this step introduces multi-scale feature fusion processing, using a pyramid feature structure and a multi-level fusion strategy to improve feature representation capabilities and anomaly identification accuracy.

[0146] In the decision-making process, the decision module 6 receives the recognition error and resolution evaluation data transmitted by the data analysis module 5, sets the error threshold and resolution threshold, and determines whether adjustment is needed based on the evaluation results. Preferably, when the error is less than 0.25 and the resolution is greater than 300 dpi, the system determines that no adjustment is needed; otherwise, the adjustment process is triggered.

[0147] In the feedback step, data feedback module 7 receives data from the server, processes the data based on data nodes and data thresholds, and feeds it back to the server. Simultaneously, it adjusts the thresholds in data classification module 2 based on abnormal data. This dynamic threshold adjustment mechanism enables the system to continuously learn and optimize, improving the accuracy and sensitivity of anomaly detection.

[0148] The intelligent acquisition, analysis, and processing system and method for engineering surveying data of this invention introduces a feature node processing mechanism, a multi-level anomaly data identification algorithm, an adaptive threshold adjustment mechanism, and an integrated system of 3D point clouds and construction monitoring. It also innovatively adds multi-scale feature fusion processing and dynamic feature weight adaptive adjustment, achieving intelligent acquisition, accurate analysis, and efficient processing of engineering surveying data. Compared with existing technologies, this invention has the following significant advantages:

[0149] First, through feature node processing and importance calculation, the system can focus on the most critical feature nodes, significantly improving mapping accuracy while maintaining high computational efficiency. Experimental data shows that the mapping accuracy of this system is 25% to 30% higher than traditional methods, while reducing computational resource consumption by about 20%. The innovative dynamic feature weight adaptive adjustment algorithm further enhances the system's environmental adaptability, improving its performance by about 20% in complex and changing environments.

[0150] Secondly, the multi-level anomaly data identification algorithm significantly improves the system's anomaly detection capability. By simultaneously considering differences in data features and importance, the system can more accurately identify and process anomalous data, effectively avoiding mapping errors caused by anomalous data. In practical applications, the anomaly detection accuracy of this system reaches over 92%, significantly exceeding the approximately 75% of traditional methods. The introduced multi-scale feature fusion enhancement module further improves the feature recognition accuracy by about 15%, especially performing exceptionally well when handling complex textures and blurred boundaries.

[0151] Furthermore, the adaptive threshold adjustment mechanism enables the system to continuously learn and optimize. By establishing a closed loop of detection-storage-adjustment-re-detection, the system can continuously adjust and optimize its parameters according to the actual working environment, achieving continuous performance improvement. Long-term test data shows that after three months of continuous operation, the anomaly detection accuracy further improved to over 95%.

[0152] Finally, the integration of 3D point cloud technology with construction monitoring enables precise monitoring and management of construction progress. By seamlessly integrating 3D point cloud technology with construction progress monitoring, the system can accurately capture construction progress deviations and make timely adjustments, greatly improving the efficiency and quality of project management. In practical applications, this system can control construction progress deviations within ±5%, significantly better than the ±15% of traditional methods.

[0153] In summary, the intelligent acquisition, analysis, and processing system and method for engineering surveying data provided by this invention, through the organic combination of multiple innovative technologies, constructs an adaptive, efficient, and accurate engineering surveying data processing ecosystem, providing strong technical support for modern engineering construction. In particular, by introducing two innovative points—multi-scale feature fusion and dynamic adaptive adjustment of feature weights—the system has achieved significant breakthroughs in adaptability to complex environments and processing accuracy, representing the latest development direction of engineering surveying technology.

[0154] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and such modifications or substitutions should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. An intelligent acquisition and analysis processing system for engineering survey data, characterized in that , comprising: An image acquisition module for acquiring a set of image features in the engineering surveying and mapping process; A data classification module connected with the image acquisition module, for receiving the set of image features and performing classified storage; A database connected with the data classification module, for storing the classified data; A data processing module connected with the database, comprising an image feature node search calculation module and a node search value comprehensive calculation module, for dividing and processing the set of feature nodes of the image features and calculating the node importance; A data analysis module connected with the data processing module, for identifying abnormal data from the processed data; A decision module and a data feedback module connected with the data analysis module, for receiving the identification result and adjusting the data threshold value; The image feature node search calculation module is used for: representing the image features as two-dimensional vectors ; The image features are divided by feature nodes to obtain ; Obtaining the input value of the image feature in the image recognition process, and comparing it with the recognition input value of the standard image feature; Determining the node search value threshold, when the input value is less than the node search value threshold, obtaining the node whose data value is greater than the threshold value in the divided feature nodes of the image feature, marking it as an important node and forming an important node set C; A two-dimensional search space D is set for the set C, denoted as , where and are two-dimensional coordinates of the search nodes; Comparing the nodes in set C with the nodes in search space D, setting a comparison range value , wherein is a value in the interval (0, 1); The node search value comprehensive calculation module is used for: Acquiring node search value ; Normalizing the obtained node search value; Calculating the importance of all feature nodes in the image feature, and the calculation formula is: ； wherein is the importance of the image feature, is the average search value of the node, is the number of nodes, is the coefficient and .

Images