Earthwork measurement and engineering progress ai inspection and identification system based on visual analysis

Abstract

The application discloses a visual analysis-based earthwork measurement and engineering progress AI inspection and identification system, relates to the field of power monitoring, and comprises a data management module, which is used for acquiring engineering project information, setting a BIM linkage model and basic task information; a data distribution module, which is used for task distribution on the basic task information and acquisition of monitoring task information; a data acquisition module, which is used for acquisition of data acquisition results and synchronous linkage; a progress analysis module, which is used for measurement analysis and progress evaluation according to the synchronous linkage results and acquisition of joint progress data; an inspection feedback module, which is used for deviation analysis according to the joint progress data obtained at each position in the BIM linkage model and acquisition of inspection feedback data; and an inspection identification module, which is used for acquisition of corresponding feedback monitoring data according to the inspection feedback data, multi-dimensional evaluation of the obtained feedback monitoring data based on corresponding deviation analysis results; and the application improves the accuracy in the process of earthwork measurement and engineering progress identification to a certain extent.

Description

Technical Field

[0001] This invention relates to the field of physical fitness monitoring technology, and in particular to an AI-based inspection and recognition system for earthwork measurement and engineering progress based on visual analysis. Background Technology

[0002] Earthwork surveying and project progress inspection are core components of engineering construction, directly impacting project cost accounting, schedule control, and quality and safety. Traditional inspection methods primarily rely on manual on-site inspections, followed by progress verification through document reports. This approach suffers from drawbacks such as low inspection frequency, limited coverage, data lag, non-standard recording, and inability to monitor on-site dynamics in real time. In recent years, the development of drones, computer vision, and AI technologies has provided technical support to address the shortcomings of traditional earthwork surveying and project progress inspection.

[0003] A search revealed Chinese patent CN115077489A, which discloses a method for calculating earthwork volume using oblique photography from a drone. The method includes the following steps: Step 1: Collecting and processing weather information to obtain information indicating whether collection is permitted or prohibited; Step 2: When permitted collection information is obtained, a drone is used for surveying to acquire surrounding images and preliminary collection location images. When prohibited collection information is obtained, the drone is controlled only after weather conditions are met; Step 3: Processing the preliminary collection location images to obtain calculation type information, and then processing the surrounding images and collection location images to obtain collection warning information and preliminary analysis information; Step 4: Processing the preliminary analysis information, and after confirming the safety of the collection location, collecting collection location information in different ways according to the size of the collection area to obtain real-time collection location parameter information. This invention can calculate earthwork volume information with smaller errors and is therefore more worthy of widespread use.

[0004] Compared with existing technologies, the Chinese patent with patent number CN115077489A can set different collection methods according to the size of the collection area at the collection location, thereby improving the accuracy of earthwork volume measurement and calculation.

[0005] However, in actual use, the above method relies on waiting for suitable weather conditions before controlling the drone and processing the initial location images based on the control results to obtain real-time location parameter information. In this process, the progress planning data during the project implementation is not considered, and dynamic adjustments cannot be made according to the project implementation process. Earthwork measurement and project progress inspection and identification are not based on the results of dynamic adjustments, which to some extent affects the accuracy of earthwork measurement and project progress identification. Summary of the Invention

[0006] The purpose of this invention is to address the shortcomings of insufficient accuracy in existing technologies by proposing an AI-based inspection and recognition system for earthwork measurement and project progress based on visual analysis.

[0007] To achieve the above objectives, the present invention adopts the following technical solution: A vision-based AI-powered inspection and recognition system for earthwork measurement and project progress includes: The data management module is used to acquire project information, set up BIM linkage models based on project information, and set up basic task information through BIM linkage models. The data allocation module is used to acquire multi-source task monitoring terminals, allocate the basic task information set in the BIM linkage model according to the multi-source task monitoring terminals, and acquire the corresponding monitoring task information. The data acquisition module is used by each multi-source task monitoring terminal to collect data according to the corresponding monitoring task information, obtain the corresponding data acquisition results, perform spatiotemporal registration processing on each of the acquired data acquisition results, and perform synchronous linkage based on the spatiotemporal registration processing results. The progress analysis module is used to measure and analyze the synchronous linkage results corresponding to each data collection result in sequence, perform multi-dimensional progress evaluation based on the measurement and analysis results, and obtain joint progress data. The inspection feedback module is used to perform deviation analysis based on the joint progress data obtained at various locations within the BIM linkage model, perform progress progression analysis on the obtained deviation analysis results, and obtain inspection feedback data. The inspection feedback data is the monitoring task information at the corresponding location after dynamic adjustment. The inspection and identification module is used to obtain corresponding feedback monitoring data based on the dynamically adjusted monitoring task information, and to conduct multi-dimensional evaluation of the obtained feedback monitoring data based on the corresponding deviation analysis results.

[0008] The above technical solution further includes: the data management module includes: The data entry unit is used to acquire project information, which includes project identification information, survey and exploration information, project planning information, and construction management information. The data management unit is used to set up a BIM linkage model based on the obtained project information. The BIM linkage model includes a geometric structure sub-model, an earthwork exploration sub-model, and a project planning sub-model. The geometric structure sub-model contains the spatial structure information corresponding to the project identification information. The earthwork exploration sub-model contains the earthwork planning information set at various locations within the geometric structure sub-model based on the corresponding project information. The project planning sub-model contains the project implementation information set at various locations within the geometric structure sub-model based on the corresponding project information. The task identification unit is used to identify the earthwork exploration sub-model and project planning sub-model corresponding to each location in the geometric structure sub-model of the BIM linkage model, and obtain the basic task information corresponding to each location.

[0009] Furthermore, the data allocation module includes: The monitoring and management unit is used to acquire information about multi-source task monitoring terminal equipment, including basic equipment information, equipment monitoring information, and equipment operation information. The task allocation unit is used to time-mark the basic task information at each location within the BIM linkage model according to the project planning information, set the task monitoring cycle for each basic task information according to the time-marking results, allocate tasks according to the basic task information corresponding to each task monitoring cycle and the information of each multi-source task monitoring terminal device, and obtain the corresponding monitoring task information according to the task allocation results. The monitoring task information includes the multi-source task monitoring terminal corresponding to each basic task information and the corresponding operation configuration information of the multi-source task monitoring terminal.

[0010] Furthermore, the data acquisition module includes: The task monitoring unit is used by each multi-source task monitoring terminal to collect data according to the monitoring task information and obtain corresponding types of data collection results. The data collection results include three types: fixed visual acquisition data, laser visual acquisition data, and aerial visual acquisition data. The monitoring and processing unit is used to acquire the corresponding data acquisition results obtained by each multi-source task monitoring terminal, set the corresponding timestamp, select a reference coordinate system based on the same timestamp for each type of data acquisition result, map the acquired data acquisition results to the reference coordinate system, synchronize and link the data acquisition results in each reference coordinate system according to the order of the timestamps, acquire the synchronization and linkage results, and upload them to the corresponding location in the BIM linkage model.

[0011] Furthermore, the progress analysis module includes: The progress analysis unit is used to perform visual analysis on the synchronous linkage results at corresponding locations within the BIM linkage model. It sequentially acquires the 3D point cloud images in the reference coordinate system corresponding to each time stamp, and performs traversal analysis on the obtained 3D point cloud images according to the changes of each pixel in each time stamp to generate a 3D point cloud change image. The multi-dimensional evaluation unit is used to acquire project planning information and construction control information at corresponding locations within the BIM linkage model. Based on the project planning information and construction control information, a planning control link is set up. The planning control link contains multiple planning control nodes. Each planning control node contains reference point cloud change images, accuracy planning data, and progress planning data corresponding to the project implementation process. The obtained three-dimensional point cloud change images are sequentially compared with each planning control node to perform multi-dimensional progress evaluation, obtain the multi-dimensional progress evaluation results of each planning control node, and map them into the planning control link to obtain joint progress data.

[0012] Furthermore, the inspection feedback module includes: The deviation analysis unit is used to acquire joint progress data obtained at various locations within the BIM linkage model, and to perform deviation analysis on the joint progress data at various locations within the BIM linkage model in sequence. The deviation analysis includes node deviation analysis and link deviation analysis. The node deviation analysis is the multi-dimensional completion measurement deviation corresponding to the planning and control node. The multi-dimensional completion measurement deviation includes time deviation, quantity deviation, and quality deviation. The link deviation analysis integrates the multi-dimensional completion measurement deviation results of various types at various planning and control nodes within the corresponding planning and control link to obtain the link deviation analysis results within the planning and control link. The node deviation analysis results and the link deviation analysis results are marked as the deviation analysis results at the corresponding locations within the BIM linkage model. The progress feedback unit is used to perform progress analysis based on the obtained deviation analysis results and obtain inspection feedback data.

[0013] Furthermore, the process by which the progress feedback unit acquires inspection feedback data includes: The joint progress data and deviation analysis results within and between various locations in the BIM linkage model are associated and combined based on project planning information and construction control information. An associated comparison dataset is constructed based on the association combination results. The associated comparison dataset is analyzed and processed based on deep learning algorithms to construct a project progress association prediction model. The project progress association prediction model is used to obtain predicted deviation data at other locations based on the joint progress data and deviation analysis results obtained at the current location. The predicted deviation data obtained at the corresponding locations are weighted and calculated to obtain comprehensive predicted deviation data. The comprehensive prediction deviation data obtained at each location is compared with the basic task information at the corresponding location in the BIM linkage model to evaluate the deviation. Based on the deviation evaluation results, inspection feedback data is generated, which is the monitoring task information at the corresponding location after dynamic adjustment.

[0014] Furthermore, the inspection and identification module includes: The feedback monitoring unit is used to acquire dynamically adjusted monitoring task information and corresponding feedback monitoring data, and to map the acquired feedback monitoring data to the corresponding location in the BIM linkage model. The inspection and identification unit is used to acquire joint progress data at corresponding locations within the BIM linkage model. Based on the deviation analysis results of each planning and control link and planning and control node within the joint progress data, the accuracy of the visual analysis process is adjusted. Based on the accuracy adjustment results, the obtained feedback monitoring data is sent to the progress analysis module for multi-dimensional evaluation to obtain the joint progress data corresponding to the feedback monitoring data.

[0015] The present invention has the following beneficial effects: 1. In this invention, the entire process of engineering project information is jointly coordinated and managed through the established BIM linkage model. Corresponding basic task information is set according to the relevant engineering project information. The obtained basic task information is assigned to tasks through multi-source task monitoring terminals to obtain monitoring task information. Data is collected based on the obtained monitoring task information. The obtained data collection results are processed for spatiotemporal registration. Synchronous linkage is performed based on the spatiotemporal registration results, which can improve the accuracy of the visual analysis process to a certain extent.

[0016] 2. In this invention, the obtained data collection results are measured and analyzed sequentially according to the synchronous linkage results. During the measurement and analysis process, a planning and control link is set according to the process information corresponding to the corresponding project information, and multiple planning and control nodes are set within the planning and control link. The data collection results are evaluated in multiple dimensions according to each planning and control node, which refines the evaluation process of project progress and improves the accuracy of project progress inspection and identification.

[0017] 3. In this invention, the data acquisition and data analysis processes of relevant procedures at relevant locations within the BIM linkage model are dynamically adjusted based on the multi-dimensional progress assessment results and the corresponding comprehensive prediction deviation data. Based on the dynamic adjustment results, the accuracy and processing efficiency of earthwork measurement and project progress inspection and identification processes can be flexibly improved. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of the structure of the AI-based inspection and recognition system for earthwork measurement and project progress based on visual analysis proposed in this invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] Example 1, as Figure 1 As shown, the AI-based inspection and recognition system for earthwork measurement and project progress based on visual analysis proposed in this invention includes, in this embodiment: The data management module is used to acquire project information, set up BIM linkage models based on project information, and set up basic task information through BIM linkage models. The data allocation module is used to acquire multi-source task monitoring terminals, allocate the basic task information set in the BIM linkage model according to the multi-source task monitoring terminals, and acquire the corresponding monitoring task information. The data acquisition module is used by each multi-source task monitoring terminal to collect data according to the corresponding monitoring task information, obtain the corresponding data acquisition results, perform spatiotemporal registration processing on each of the acquired data acquisition results, and perform synchronous linkage based on the spatiotemporal registration processing results. The progress analysis module is used to measure and analyze the synchronous linkage results corresponding to each data collection result in sequence, perform multi-dimensional progress evaluation based on the measurement and analysis results, and obtain joint progress data. The inspection feedback module is used to perform deviation analysis based on the joint progress data obtained at various locations within the BIM linkage model, perform progress progression analysis on the obtained deviation analysis results, and obtain inspection feedback data. The inspection feedback data is the monitoring task information at the corresponding location after dynamic adjustment. The inspection and identification module is used to obtain corresponding feedback monitoring data based on the dynamically adjusted monitoring task information, and to conduct multi-dimensional evaluation of the obtained feedback monitoring data based on the corresponding deviation analysis results.

[0021] In the specific implementation process, the data management module is used to acquire project information, set up a BIM linkage model based on the project information, and set up basic task information through the BIM linkage model. The process includes: The data entry unit is used to acquire project information, which includes project identification information, survey and exploration information, project planning information, and construction management information, wherein: Project identification information includes basic project information, project construction address, geographical coordinate range, etc. Exploration and survey information includes geological exploration information, digital elevation models, etc. Project planning information includes construction planning information, schedule information, and preliminary calculation baseline data; Construction control information includes constraint information for different work processes within the corresponding project planning information; The data management unit is used to set up a BIM linkage model based on the obtained project information. The BIM linkage model includes a geometric structure sub-model, an earthwork exploration sub-model, and a project planning sub-model. The geometric structure sub-model contains the spatial structure information corresponding to the project identification information. The earthwork exploration sub-model contains the earthwork planning information set at various locations within the geometric structure sub-model based on the corresponding project information. The project planning sub-model contains the project implementation information set at various locations within the geometric structure sub-model based on the corresponding project information. It should be further explained that in the process of setting up the BIM linkage model, the geometric structure sub-model is set up first. Based on the geometric structure sub-model, the earthwork exploration sub-model and the project planning sub-model are constructed according to the corresponding exploration and detection information, project planning information and construction control information in the project information. In this process, the geometric structure sub-model is spatially coded, the project information is spatially decomposed according to the spatial area code, and the corresponding spatial area code is set according to the spatial decomposition result. The spatial area code includes four levels: work area, zone, sub-zone and unit, which are used to describe the area range within the spatial structure. The exploration and detection information, project planning information, and construction control information at the locations corresponding to the spatial regions within the geometric structure sub-model are respectively encoded as earthwork procedures. The earthwork procedure codes are divided into four levels: geological type, foundation pit area, procedure information, and constraint conditions. Based on the earthwork planning information represented by the earthwork procedure codes, the corresponding earthwork exploration sub-model is obtained. The exploration and survey information, project planning information, and construction control information at the locations corresponding to the spatial regions within the geometric structure sub-model are respectively encoded into planning procedures. The planning procedure code includes project implementation information at the corresponding spatial region code location. The planning procedure code includes sub-projects, sub-items, procedure information, and constraints to obtain the corresponding project planning sub-model. The geometric structure sub-model, earthwork exploration sub-model and project planning sub-model are superimposed and integrated according to their corresponding location information to construct a BIM linkage model. The BIM linkage model is used to describe the earthwork planning information and project implementation information of the engineering project at the corresponding spatial location. The task identification unit is used to identify the earthwork exploration sub-model and project planning sub-model corresponding to each location in the geometric structure sub-model of the BIM linkage model, and obtain the basic task information corresponding to each location. It should be further explained that, in the specific implementation process, the basic task information corresponding to each location is the process task information and the corresponding progress planning information of the earthwork process code and planning process code corresponding to the corresponding spatial area code.

[0022] In the specific implementation process, the data allocation module is used to acquire multi-source task monitoring terminals, allocate the basic task information set in the BIM linkage model according to the multi-source task monitoring terminals, and the process of acquiring the corresponding monitoring task information includes: The monitoring and management unit is used to acquire information about multi-source task monitoring terminal devices. This information includes basic device information, device monitoring information, and device operation information, wherein: The basic equipment information includes the corresponding equipment identification information and equipment operating parameter information; The equipment monitoring information corresponds to the monitoring type, such as drone monitoring terminals, lidar terminals, and fixed camera monitoring terminals; The equipment operation information refers to the current operating status information of the corresponding type of multi-source task monitoring terminal equipment; The task allocation unit is used to time-mark the basic task information at each location within the BIM linkage model according to the project planning information, set the task monitoring cycle for each basic task information according to the time-marking results, allocate tasks according to the basic task information corresponding to each task monitoring cycle and the information of each multi-source task monitoring terminal device, and obtain the corresponding monitoring task information according to the task allocation results. The monitoring task information includes the multi-source task monitoring terminal corresponding to each basic task information and the operation configuration information corresponding to the multi-source task monitoring terminal. It should be further explained that, in the specific implementation process, when acquiring basic task information at each location within the BIM linkage model and setting the task monitoring cycle corresponding to each basic task information, extension coefficients are set on both sides according to the progress plan. The set extension coefficients are extended based on the upper and lower limits of the corresponding progress plan to obtain the task monitoring cycle corresponding to the basic task information. Obtain the basic task information and multi-source task monitoring terminal equipment information corresponding to each basic task information in the current BIM linkage model. Match the equipment monitoring information, equipment operation information and equipment basic information corresponding to the multi-source task monitoring terminal equipment information in sequence. Obtain the multi-source task monitoring terminal corresponding to the equipment basic information that matches the equipment operation information. Assign the corresponding basic task information to the corresponding multi-source task monitoring terminal. The basic task information obtained by the multi-source task monitoring terminal is used to set monitoring task information according to the spatial area code, earthwork process code and planning process code in the BIM linkage model to which it belongs. The monitoring task information includes the monitoring spatial range and monitoring time range of the corresponding multi-source task monitoring terminal, as well as the operation configuration information corresponding to the monitoring range. The operation configuration information includes acquisition frequency data, acquisition resolution data, etc.

[0023] In the specific implementation process, the data acquisition module is used by each multi-source task monitoring terminal to collect data according to the corresponding monitoring task information, obtain the corresponding data acquisition results, perform spatiotemporal registration processing on each of the obtained data acquisition results, and perform synchronous linkage based on the spatiotemporal registration processing results. The process includes: The task monitoring unit is used by each multi-source task monitoring terminal to collect data according to the monitoring task information and obtain corresponding types of data collection results. The data collection results include three types: fixed visual acquisition data, laser visual acquisition data, and aerial visual acquisition data. The monitoring and processing unit is used to acquire the corresponding type of data acquisition results obtained by each multi-source task monitoring terminal. The data acquisition results are fixed visual acquisition data, laser visual acquisition data and aerial visual acquisition data obtained by the corresponding multi-source task monitoring terminal at each corresponding position in the BIM linkage model, and a corresponding timestamp is set. The timestamp is the time stamp information of the corresponding fixed visual acquisition data, laser visual acquisition data and aerial visual acquisition data. The data acquisition results of each type are selected from the reference coordinate system based on the same timestamp on the multi-source task monitoring terminal. The reference coordinate system is an independent coordinate system corresponding to the earthwork process code and the planning process code at the corresponding spatial area code. The obtained data acquisition results are mapped to the reference coordinate system in sequence, with priority given to mapping the fixed visual acquisition data to the reference coordinate system. Multiple coordinate anchor points are set. The coordinate anchor points are used to mark the fixed pixels at the corresponding positions within the corresponding spatial area code. The laser visual acquisition data and aerial visual acquisition data are proportionally rotated and adjusted according to the multiple coordinate anchor points. Based on the adjustment results, the corresponding laser visual acquisition data and aerial visual acquisition data are mapped to the reference coordinate system to obtain the visual acquisition image at the corresponding spatial area code. It should be further explained that during the mapping process, the obtained fixed visual acquisition data, laser visual acquisition data and aerial visual acquisition data are pre-processed in advance. The image preprocessing process includes color equalization processing and noise reduction processing, etc. Based on the timestamp order, the visual acquisition images corresponding to each mapping result in the reference coordinate system are synchronized and linked. That is, the corresponding visual acquisition images are sorted and processed in sequence, and the synchronization and linkage result is obtained based on the sorting result. The synchronization and linkage result is the change of each visual acquisition image in the reference coordinate system with the timestamp order, and is uploaded to the corresponding position in the BIM linkage model.

[0024] In the specific implementation process, the progress analysis module is used to sequentially measure and analyze the synchronous linkage results corresponding to each data acquisition result, and perform multi-dimensional progress evaluation based on the measurement and analysis results. The process of obtaining joint progress data includes: The progress analysis unit is used to perform visual analysis on the synchronous linkage results at corresponding locations within the BIM linkage model. Based on the edge processing algorithm, it extracts the boundary range of the corresponding spatial region encoding location within each visual acquisition image. Based on the extracted boundary range, it obtains the 3D point cloud image in the reference coordinate system corresponding to the corresponding timestamp. The obtained 3D point cloud image is then traversed and analyzed according to the changes of each pixel within each timestamp to generate a 3D point cloud change image. It should be further explained that, in the specific implementation process, during the generation of the 3D point cloud transformation image: Convert the 3D point cloud image corresponding to the corresponding timestamp into a grid of the same resolution, set the corresponding resolution grid precision, and calculate the elevation difference for each grid of the 3D point cloud images at adjacent timestamps. Based on the obtained elevation difference Perform change type marking, if If <0, then earthwork excavation begins; if If >0, then backfill the earthwork; if If ≈0, then there is no change; set the corresponding elevation difference threshold HY according to the resolution grid accuracy. If the corresponding grid is marked as a changing grid, the changing regions at each adjacent timestamp are extracted according to the changes of the changing grid. The extraction results of the changing regions are combined with the timestamps for temporal change rendering. The change type and intensity are rendered at the corresponding changing grid mark position, where the intensity is the size of the corresponding changing region. A three-dimensional point cloud change image is generated based on the temporal change rendering results. The multi-dimensional evaluation unit is used to acquire project planning information and construction control information at corresponding locations within the BIM linkage model. Based on the project planning information and construction control information, a planning control link is set up. The planning control link contains multiple planning control nodes. Each planning control node contains reference point cloud change images, accuracy planning data, and progress planning data corresponding to the project implementation process. The acquired three-dimensional point cloud change images are sequentially compared with each planning control node to perform multi-dimensional progress evaluation, obtain the multi-dimensional progress evaluation results of each planning control node, and map them into the planning control link to obtain joint progress data. It should be further explained that, in the specific implementation process, the process of obtaining joint progress data includes: The planning and control link consists of the steps involved in the corresponding type of process set at the corresponding spatial area code based on the earthwork process code and the planning process code. Each step information is set as a corresponding planning and control node. The planning and control node contains the reference point cloud change image, accuracy planning data, and schedule planning data corresponding to the step information, wherein: The reference point cloud change image represents the change information at the corresponding spatial region coding position, the starting position, the key planning progress position, and the ending position within the corresponding step information. The schedule planning data includes the planned completion time and planned amount of work to be completed at the corresponding spatial area code location within the corresponding step information. The accuracy planning data is set for the constraints at the corresponding spatial area code location within the corresponding step information, corresponding to the earthwork process code and the planning process code; Multidimensional progress assessment is performed on the reference point cloud change images, accuracy planning data, and progress planning data corresponding to each planning and control node, and the corresponding 3D point cloud change images between the corresponding timestamps. This multidimensional progress assessment compares the reference point cloud change images, accuracy planning data, and progress planning data with the corresponding 3D point cloud change images at the starting position, key planning progress points, and ending position, sequentially. Each change region is integrated to obtain the corresponding comparison analysis results. These results are the sum of deviations between each change region in the 3D point cloud change image and the corresponding reference point cloud change image, accuracy planning data, and progress planning data. The obtained comparison analysis results are marked as the multidimensional progress assessment results for each planning and control node. The multidimensional progress assessment results between each planning and control node are then balanced and scheduled. The balanced scheduling results are visualized within the planning and control link, and a multidimensional comprehensive progress assessment result within the planning and control link is obtained. The obtained multidimensional progress assessment results and visualization results are then used to set up joint progress data.

[0025] In the specific implementation process, the inspection feedback module is used to perform deviation analysis based on the joint progress data obtained at various locations within the BIM linkage model, perform progress progression analysis on the obtained deviation analysis results, and obtain inspection feedback data. The process of obtaining inspection feedback data as the monitoring task information at the corresponding location after dynamic adjustment includes: The deviation analysis unit is used to acquire joint progress data obtained at various locations within the BIM linkage model, and to perform deviation analysis on the joint progress data at various locations within the BIM linkage model in sequence. The deviation analysis includes node deviation analysis and link deviation analysis. The node deviation analysis is the multi-dimensional completion measurement deviation corresponding to the planning and control node. The multi-dimensional completion measurement deviation includes time deviation, quantity deviation, and quality deviation. The link deviation analysis integrates the multi-dimensional completion measurement deviation results of various types at various planning and control nodes within the corresponding planning and control link to obtain the link deviation analysis results within the planning and control link. The node deviation analysis results and the link deviation analysis results are marked as the deviation analysis results at the corresponding locations within the BIM linkage model. It should be further explained that in the process of obtaining the deviation analysis results: The deviation data includes three types: time deviation data, quantity deviation data, and quality deviation data. Each type of deviation data is assigned a corresponding weight coefficient at the location of its respective planning and control node. The corresponding weight coefficient is weighted and calculated with the corresponding deviation data to obtain the node deviation analysis results at each planning and control node. Furthermore, the process of obtaining the node deviation analysis results includes: The deviation data corresponding to time deviation data, quantity deviation data, and quality deviation data are set into statistical unit intervals respectively. Normalization mapping is performed based on the position information of the statistical unit intervals to which the time deviation data, quantity deviation data, and quality deviation data belong. The normalization mapping results corresponding to the time deviation data, quantity deviation data, and quality deviation data are obtained respectively. The weighted calculation is performed based on the normalization mapping results and the set weight coefficients. The node deviation analysis results at each planning and control node are obtained based on the weighted calculation results. The obtained node deviation analysis results are mapped to the planning and control link respectively. The deviation data between each planning and control node is balanced and scheduled. The balanced scheduling results are visualized and marked in the planning and control link, and the link deviation analysis results in the planning and control link are obtained. The progress feedback unit is used to perform progress analysis based on the obtained deviation analysis results and obtain inspection feedback data. The process includes: The joint progress data and deviation analysis results within and between different locations in the BIM linkage model are correlated and combined based on project planning information and construction control information. A correlation comparison dataset is then constructed based on the correlation and combination results, including: The spatial region codes and corresponding earthwork process codes and planning process codes for each location are obtained. Process combinations within a single location, cross-process combinations within a single location, same-process combinations between multiple locations, and cross-process combinations between multiple locations are performed. Feature extraction is performed based on the joint progress data and deviation analysis results involved in the associated combination results. Based on the feature extraction results, spatial correlation analysis, process correlation analysis, and resource correlation analysis are performed on each combination. The resource correlation analysis results are weighted to obtain the corresponding correlation quantification results. The results of each associated combination are filtered to obtain an associated comparison dataset. An associated comparison dataset is constructed based on the associated combination results. A project schedule correlation prediction model is constructed by analyzing and processing the correlation comparison dataset based on deep learning algorithms. The process includes: The joint progress data, deviation analysis results, and corresponding correlation quantification results corresponding to the corresponding processes at the corresponding positions in the key comparison dataset are labeled as the input layer; a feature extraction layer is set up, which obtains the corresponding spatial correlation features based on the graph convolutional network and the corresponding temporal correlation features based on the long short-term memory network; the joint progress data and deviation analysis results corresponding to the corresponding processes at other positions are labeled as the output layer; The smoothing loss function is selected first for training and analysis to build a project progress correlation prediction model. The model parameters corresponding to the project progress correlation prediction model are obtained. The obtained project progress correlation prediction model is validated according to the corresponding validation set based on the correlation comparison dataset. The corresponding project progress correlation prediction model is output according to the validation results. The project progress correlation prediction model is used to obtain prediction deviation data at other locations based on the joint progress data and deviation analysis results obtained at the current location, and to perform weighted calculation on the prediction deviation data obtained at the corresponding locations to obtain comprehensive prediction deviation data. The comprehensive prediction deviation data obtained at each location is compared with the basic task information at the corresponding location in the BIM linkage model to evaluate the deviation. Based on the deviation evaluation results, the operation configuration information of the corresponding multi-source task monitoring terminal in the relevant monitoring task information is dynamically adjusted. If the deviation evaluation is larger, the acquisition frequency data and acquisition resolution data are higher. Based on the adjustment results of the operation configuration information, inspection feedback data is generated. The inspection feedback data is the dynamically adjusted monitoring task information at the corresponding location.

[0026] In the specific implementation process, the inspection and identification module is used to obtain corresponding feedback monitoring data based on the dynamically adjusted monitoring task information, and the process of performing multi-dimensional evaluation on the obtained feedback monitoring data based on the corresponding deviation analysis results includes: The feedback monitoring unit is used to acquire dynamically adjusted monitoring task information and corresponding feedback monitoring data, and to map the acquired feedback monitoring data to the corresponding location in the BIM linkage model. The inspection and identification unit is used to acquire joint progress data at corresponding locations within the BIM linkage model. Based on the deviation analysis results of each planning and control link and planning and control node within the joint progress data, the accuracy of the visual analysis process is adjusted. According to the deviation analysis results, the accuracy of each 3D point cloud image within the progress analysis unit is adjusted to the corresponding resolution grid. A linear relationship is set between the error analysis results and the resolution grid accuracy; that is, the larger the deviation analysis result, the higher the resolution grid accuracy. Accuracy adjustments are made according to this linear relationship to obtain the accuracy adjustment results. Based on the accuracy adjustment results, the obtained feedback monitoring data is sent to the progress analysis module for multi-dimensional evaluation to obtain the joint progress data corresponding to the feedback monitoring data.

[0027] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A visual analysis-based AI-powered inspection and recognition system for earthwork measurement and project progress, characterized in that: include: The data management module is used to acquire project information, set up a BIM linkage model based on the project information, and set up basic task information through the BIM linkage model. The data allocation module is used to acquire multi-source task monitoring terminals, allocate the basic task information set in the BIM linkage model according to the multi-source task monitoring terminals, and acquire the corresponding monitoring task information. The data acquisition module is used by each multi-source task monitoring terminal to collect data according to the corresponding monitoring task information, obtain the corresponding data acquisition results, perform spatiotemporal registration processing on each of the acquired data acquisition results, and perform synchronous linkage based on the spatiotemporal registration processing results. The progress analysis module is used to measure and analyze the synchronous linkage results corresponding to each data collection result in sequence, perform multi-dimensional progress evaluation based on the measurement and analysis results, and obtain joint progress data. The inspection feedback module is used to perform deviation analysis based on the joint progress data obtained at various locations within the BIM linkage model, perform progress progression analysis on the obtained deviation analysis results, and obtain inspection feedback data. The inspection feedback data is the monitoring task information at the corresponding location after dynamic adjustment. The inspection and identification module is used to obtain corresponding feedback monitoring data based on the dynamically adjusted monitoring task information, and to conduct multi-dimensional evaluation of the obtained feedback monitoring data based on the corresponding deviation analysis results.

2. The AI-based inspection and recognition system for earthwork measurement and project progress based on visual analysis as described in claim 1, characterized in that, The data management module includes: The data entry unit is used to acquire project information, which includes project identification information, survey and exploration information, project planning information, and construction management and control information. The data management unit is used to set up a BIM linkage model based on the obtained project information. The BIM linkage model includes a geometric structure sub-model, an earthwork exploration sub-model, and a project planning sub-model. The geometric structure sub-model contains the spatial structure information corresponding to the project identification information. The earthwork exploration sub-model contains the earthwork planning information set at various locations within the geometric structure sub-model based on the corresponding project information. The project planning sub-model contains the project implementation information set at various locations within the geometric structure sub-model based on the corresponding project information. The task identification unit is used to identify the earthwork exploration sub-model and project planning sub-model corresponding to each location in the geometric structure sub-model of the BIM linkage model, and obtain the basic task information corresponding to each location.

3. The AI-based inspection and recognition system for earthwork measurement and project progress based on visual analysis as described in claim 2, characterized in that, The data allocation module includes: The monitoring and management unit is used to acquire information about multi-source task monitoring terminal equipment, including basic equipment information, equipment monitoring information, and equipment operation information. The task allocation unit is used to time-mark the basic task information at each location within the BIM linkage model according to the project planning information, set the task monitoring cycle for each basic task information according to the time-marking results, allocate tasks according to the basic task information corresponding to each task monitoring cycle and the information of each multi-source task monitoring terminal device, and obtain the corresponding monitoring task information according to the task allocation results. The monitoring task information includes the multi-source task monitoring terminal corresponding to each basic task information and the corresponding operation configuration information of the multi-source task monitoring terminal.

4. The AI-based inspection and recognition system for earthwork measurement and project progress based on visual analysis according to claim 3, characterized in that, The data acquisition module includes: The task monitoring unit is used by each multi-source task monitoring terminal to collect data according to the monitoring task information and obtain corresponding types of data collection results. The data collection results include three types: fixed visual acquisition data, laser visual acquisition data, and aerial visual acquisition data. The monitoring and processing unit is used to acquire the corresponding data acquisition results obtained by each multi-source task monitoring terminal, set the corresponding timestamp, select a reference coordinate system based on the same timestamp for each type of data acquisition result, map the acquired data acquisition results to the reference coordinate system, synchronize and link the data acquisition results in each reference coordinate system according to the order of the timestamps, acquire the synchronization and linkage results, and upload them to the corresponding location in the BIM linkage model.

5. The AI-based inspection and recognition system for earthwork measurement and project progress based on visual analysis according to claim 4, characterized in that, The progress analysis module includes: The progress analysis unit is used to perform visual analysis on the synchronous linkage results at corresponding locations within the BIM linkage model. It sequentially acquires the 3D point cloud images in the reference coordinate system corresponding to each time stamp, and performs traversal analysis on the obtained 3D point cloud images according to the changes of each pixel in each time stamp to generate a 3D point cloud change image. The multi-dimensional evaluation unit is used to acquire project planning information and construction control information at corresponding locations within the BIM linkage model. Based on the project planning information and construction control information, a planning control link is set up. The planning control link contains multiple planning control nodes. Each planning control node contains reference point cloud change images, accuracy planning data, and progress planning data corresponding to the project implementation process. The obtained three-dimensional point cloud change images are sequentially compared with each planning control node to perform multi-dimensional progress evaluation, obtain the multi-dimensional progress evaluation results of each planning control node, and map them into the planning control link to obtain joint progress data.

6. The AI-based inspection and recognition system for earthwork measurement and project progress based on visual analysis according to claim 5, characterized in that, The inspection feedback module includes: The deviation analysis unit is used to acquire joint progress data obtained at various locations within the BIM linkage model, and to perform deviation analysis on the joint progress data at various locations within the BIM linkage model in sequence. The deviation analysis includes node deviation analysis and link deviation analysis. The node deviation analysis is the multi-dimensional completion measurement deviation corresponding to the planning and control node. The multi-dimensional completion measurement deviation includes time deviation, quantity deviation, and quality deviation. The link deviation analysis integrates the multi-dimensional completion measurement deviation results of various types at various planning and control nodes within the corresponding planning and control link to obtain the link deviation analysis results within the planning and control link. The node deviation analysis results and the link deviation analysis results are marked as the deviation analysis results at the corresponding locations within the BIM linkage model. The progress feedback unit is used to perform progress analysis based on the obtained deviation analysis results and obtain inspection feedback data.

7. The AI-based inspection and recognition system for earthwork measurement and project progress based on visual analysis as described in claim 6, characterized in that, The process by which the progress feedback unit acquires inspection feedback data includes: The joint progress data and deviation analysis results within and between various locations in the BIM linkage model are associated and combined based on project planning information and construction control information. An associated comparison dataset is constructed based on the association combination results. The associated comparison dataset is analyzed and processed based on deep learning algorithms to construct a project progress association prediction model. The project progress association prediction model is used to obtain predicted deviation data at other locations based on the joint progress data and deviation analysis results obtained at the current location. The predicted deviation data obtained at the corresponding locations are weighted and calculated to obtain comprehensive predicted deviation data. The comprehensive prediction deviation data obtained at each location is compared with the basic task information at the corresponding location in the BIM linkage model to evaluate the deviation. Based on the deviation evaluation results, inspection feedback data is generated, which is the monitoring task information at the corresponding location after dynamic adjustment.

8. The AI-based inspection and recognition system for earthwork measurement and project progress based on visual analysis according to claim 7, characterized in that, The inspection and identification module includes: The feedback monitoring unit is used to acquire dynamically adjusted monitoring task information and corresponding feedback monitoring data, and to map the acquired feedback monitoring data to the corresponding location in the BIM linkage model. The inspection and identification unit is used to acquire joint progress data at corresponding locations within the BIM linkage model. Based on the deviation analysis results of each planning and control link and planning and control node within the joint progress data, the accuracy of the visual analysis process is adjusted. Based on the accuracy adjustment results, the obtained feedback monitoring data is sent to the progress analysis module for multi-dimensional evaluation to obtain the joint progress data corresponding to the feedback monitoring data.

Images