A Method and System for Dynamically Annotating Project Cost Data Based on BIM Technology

Abstract

The present invention provides a method and system for dynamically annotating engineering cost data based on BIM technology. By combining a standardized deep learning network and a customized deep learning network, it efficiently extracts cost component elements from target BIM engineering cost dynamic data and generates a confidence sequence associated with a preset cost mapping knowledge set, thereby accurately identifying the target cost component elements. Further, the standardized deep learning network and the annotation network are used to carefully annotate the target cost component elements, generating a confidence sequence associated with another preset cost mapping knowledge set to accurately obtain the target cost description label of the target cost component elements. This method not only significantly improves the accuracy and efficiency of dynamically annotating engineering cost data, but also realizes in-depth mining and refined management of BIM engineering cost data, contributing to accurate estimation of engineering costs, cost control, and project management.

Description

Technical Field

[0001] The present invention relates to the technical field of deep learning. Specifically, it relates to a method and system for dynamically annotating project cost data based on BIM technology. Background Art

[0002] In the field of construction engineering, the dynamic data management of project cost plays a crucial role in project cost control, budget planning, and decision-making support. With the rapid development of Building Information Modeling (BIM) technology, BIM has become an important tool in project cost management, providing rich and visual building data, which strongly supports the accurate calculation of project cost. However, in the process of processing BIM project cost data, how to efficiently and accurately annotate the cost component elements in dynamic data has always been an urgent problem to be solved.

[0003] Traditional methods for annotating project cost data often rely on manual operations, which are not only time-consuming and laborious but also easily affected by human factors, making it difficult to guarantee the accuracy and consistency of annotation results. In recent years, although some automated annotation methods based on machine learning or deep learning technologies have been proposed, these methods usually rely on a single deep learning network, and their annotation effects and generalization abilities are still limited for complex and variable BIM project cost dynamic data.

[0004] In addition, existing automated annotation methods often lack detailed differentiation and accurate description of different cost component elements, resulting in the annotation results being difficult to meet the refined requirements of project cost management in practical applications. Therefore, how to develop a method that can efficiently and accurately annotate the cost component elements in project cost dynamic data in combination with the characteristics of BIM technology has become an urgent technical problem to be solved. Summary of the Invention

[0005] In view of the above-mentioned problems, in combination with the first aspect of the present invention, embodiments of the present invention provide a method for dynamically annotating project cost data based on BIM technology, and the method includes:

[0006] Obtain the target BIM project cost dynamic data to be annotated, extract cost component elements from the target BIM project cost dynamic data through a standardized deep learning network, generate a first confidence sequence associated with a first preset cost mapping knowledge set, extract cost component elements from the target BIM project cost dynamic data through a customized deep learning network, and generate a second confidence sequence associated with the first preset cost mapping knowledge set, where the first preset cost mapping knowledge set includes a plurality of reference cost component elements;

[0007] Based on the first confidence sequence and the second confidence sequence, obtain the target cost component elements in the target BIM project cost dynamic data;

[0008] Based on the target BIM project cost dynamic data and the target cost component elements, use the standardized deep learning network to label the target cost component elements to generate a third confidence sequence associated with the second preset cost mapping knowledge set, and use the annotation network to label the target cost component elements to generate a fourth confidence sequence associated with the second preset cost mapping knowledge set. The second preset cost mapping knowledge set includes multiple reference cost description labels;

[0009] Based on the third confidence sequence and the fourth confidence sequence, obtain the target cost description label of the target cost component element;

[0010] Based on the target cost component element and the target cost description label of the target cost component element, generate the annotation result of the target BIM project cost dynamic data.

[0011] In another aspect, an embodiment of the present invention further provides a system for annotating project cost dynamic data based on BIM technology, including a processor and a machine-readable storage medium. The machine-readable storage medium is connected to the processor. The machine-readable storage medium is used to store programs, instructions or codes, and the processor is used to execute the programs, instructions or codes in the machine-readable storage medium to implement the above method.

[0012] Based on the above aspects, the embodiments of the present application combine a standardized deep learning network and a customized deep learning network to efficiently extract cost component elements from target BIM project cost dynamic data, and generate a confidence sequence associated with a preset cost mapping knowledge set, so as to accurately identify the target cost component elements. Further, use the standardized deep learning network and the annotation network to carefully label the target cost component elements to generate a confidence sequence associated with another preset cost mapping knowledge set, so as to accurately obtain the target cost description label of the target cost component element. This method not only significantly improves the accuracy and efficiency of annotating project cost dynamic data, but also realizes the in-depth mining and refined management of BIM project cost data, which is helpful for accurate estimation of project cost, cost control and project management. Description of the Drawings

[0013] Figure 1 is a schematic execution flowchart of a method for annotating project cost dynamic data based on BIM technology provided by an embodiment of the present invention.

[0014] Figure 2It is a schematic diagram of the hardware architecture of the project cost dynamic data annotation system based on BIM technology provided by the embodiments of the present invention. Specific embodiments

[0015] The present invention will be specifically described below with reference to the accompanying drawings of the specification. Figure 1 It is a schematic flowchart of the project cost dynamic data annotation method based on BIM technology provided by an embodiment of the present invention. The project cost dynamic data annotation method based on BIM technology will be introduced in detail below.

[0016] Step S110: Obtain the target BIM project cost dynamic data to be annotated, extract cost component elements from the target BIM project cost dynamic data through a standardized deep learning network to generate a first confidence sequence associated with a first preset cost mapping knowledge set, and extract cost component elements from the target BIM project cost dynamic data through a customized deep learning network to generate a second confidence sequence associated with the first preset cost mapping knowledge set. The first preset cost mapping knowledge set includes a plurality of reference cost component elements.

[0017] In this embodiment, in a construction project, there is target BIM project cost dynamic data to be annotated. Suppose this construction project is a construction project of a comprehensive commercial building, which includes various building structure parts, such as foundation works, main structure, decoration, water supply and drainage system, electrical system, etc. Each part contains numerous cost component elements.

[0018] First, obtain the target BIM project cost dynamic data of this commercial building project. These target BIM project cost dynamic data can come from the cost estimates, budgets, and actual cost records of each stage of the project and are constantly updated as the project progresses. For example, the cost data of the foundation works part may include the costs of different types of foundation piles, the amount and price of concrete used, the cost of earth excavation, etc. These target BIM project cost dynamic data are dynamically changing and may change due to factors such as material price fluctuations and engineering quantity changes as the project progresses.

[0019] Next, a standardized deep learning network is used to extract cost component elements from the target BIM project cost dynamic data. The standardized deep learning network is constructed and trained based on the general cost knowledge of multiple projects. For example, this network has learned the patterns of basic cost component elements in a large number of different types of construction projects. When processing the target BIM project cost dynamic data of a commercial building project, possible cost component elements can be identified, and a first confidence sequence associated with the first preset cost mapping knowledge set is generated for each identified cost component element. The first preset cost mapping knowledge set contains multiple reference cost component elements, such as various types of basic components (such as precast piles, cast-in-place piles, etc.), main structure components (such as beams, columns, slabs, etc.). Suppose the standardized deep learning network identifies the cost component element of a cast-in-place pile in the foundation project, and a confidence value of this cast-in-place pile cost component element in the first preset cost mapping knowledge set can be given, forming an element in the first confidence sequence. If this confidence value is relatively high, it indicates that the network is relatively certain about this recognition result.

[0020] At the same time, a customized deep learning network also extracts cost component elements from the target BIM project cost dynamic data. The customized deep learning network is generated by learning network parameters according to the feature distance between the first estimated cost component elements and the actual cost component elements of each first template BIM project cost dynamic data. For this commercial building project, the customized deep learning network may be specifically customized and trained for commercial building projects of similar scale and function. It also analyzes the target BIM project cost dynamic data of the commercial building. For example, when analyzing the foundation project part, because it is customized for commercial buildings, it may be able to more accurately identify cost component elements suitable for the characteristics of commercial buildings. For example, there may be special anti-floating design component elements in the foundation project of large commercial buildings. The customized deep learning network can better identify these elements and generate a second confidence sequence associated with the first preset cost mapping knowledge set. For example, for the special anti-floating component element in the foundation project, the customized deep learning network gives a confidence value in the first preset cost mapping knowledge set, forming an element in the second confidence sequence.

[0021] Step S120: Obtain the target cost component elements in the target BIM project cost dynamic data according to the first confidence sequence and the second confidence sequence.

[0022] Continuing with the above commercial building project as an example, the first confidence sequence and the second confidence sequence contain information about the possibility degrees of each reference cost component element in the target BIM project cost dynamic data.

[0023] Suppose the confidence level of the cast-in-place pile cost component element in the basic engineering part in the first confidence level sequence is 0.8, while the confidence level of the cast-in-place pile cost component element in the second confidence level sequence generated by the customized deep learning network is 0.9. At the same time, in the first confidence level sequence, for another component element in the basic engineering - the isolated foundation, the confidence level is 0.3, and in the second confidence level sequence, the confidence level of the isolated foundation is 0.2.

[0024] According to the feature distance between the first confidence level sequence and the second confidence level sequence, determine the first extraction loss of the second confidence level sequence. For example, if the differences between the corresponding elements in the two sequences are large, such as the confidence level difference of the cast-in-place pile is 0.1 and the confidence level difference of the isolated foundation is 0.1, comprehensively consider these differences to calculate the first extraction loss. Then, according to the deviation value between the second confidence level sequence and the first extraction loss, determine the first corrected confidence level of the first confidence level sequence. If the deviation value of the second confidence level sequence as a whole from the first extraction loss indicates that it is more reliable, then the confidence level of the same cost component element in the first confidence level sequence as in the second confidence level sequence may be increased. For example, correct the confidence level of the cast-in-place pile in the first confidence level sequence to 0.85.

[0025] Next, according to the deviation value between the corrected first confidence level sequence and other relevant confidence level sequences (such as the confidence level sequences related to the processing results of other networks that may be involved later), determine the second target confidence level sequence. For example, assume that there are also confidence level sequences of the extraction results of the cost component elements in the target BIM project cost dynamic data by other networks. After comprehensively considering these factors, obtain the final second target confidence level sequence. According to this second target confidence level sequence, obtain the target cost component elements in the target BIM project cost dynamic data. For example, finally determine that the cast-in-place pile and the special anti-floating component elements are the target cost component elements because these target cost component elements have a relatively high confidence level in the final confidence level sequence, while the isolated foundation is excluded due to its low confidence level.

[0026] Step S130, according to the target BIM project cost dynamic data and the target cost component elements, label the target cost component elements through the standardized deep learning network to generate a third confidence level sequence associated with the second preset cost mapping knowledge set, and label the target cost component elements through the labeling network to generate a fourth confidence level sequence associated with the second preset cost mapping knowledge set, where the second preset cost mapping knowledge set includes multiple reference cost description tags.

[0027] In the commercial building project, the target cost component elements have been determined, such as the cast-in-place pile and the special anti-floating component elements.

[0028] Based on the target BIM project cost dynamic data and these target cost component elements, the target cost component elements are labeled through a standardized deep learning network. The standardized deep learning network is generated by learning network parameters based on the feature distance between the third annotation results of each third-template BIM project cost dynamic data and the template annotation data. This network will generate a third confidence sequence associated with the second preset cost mapping knowledge set for each target cost component element. The second preset cost mapping knowledge set includes multiple reference cost description labels, such as "cast-in-place pile - high-strength concrete cast-in-place pile (diameter 1 meter, depth 30 meters), C30 concrete, cost per meter is XXX yuan" and other similar description labels. Suppose in the third confidence sequence given by the standardized deep learning network for the target cost component element of the cast-in-place pile, the confidence for the description label "high-strength concrete cast-in-place pile (diameter 1 meter, depth 30 meters), C30 concrete, cost per meter is XXX yuan" is 0.7.

[0029] Meanwhile, the target cost component elements are labeled through a labeling network. The labeling network is generated by learning network parameters based on the feature distance between the second estimated cost description label and the actual cost description label of each second-template BIM project cost dynamic data. The labeling network will also generate a fourth confidence sequence associated with the second preset cost mapping knowledge set for the target cost component elements. For example, in the fourth confidence sequence given by the labeling network for the target cost component element of the cast-in-place pile, the confidence for the description label "high-strength concrete cast-in-place pile (diameter 1 meter, depth 30 meters), C30 concrete, cost per meter is XXX yuan" is 0.8.

[0030] Step S140, based on the third confidence sequence and the fourth confidence sequence, obtain the target cost description label of the target cost component element.

[0031] Still taking the target cost component element of the cast-in-place pile in the commercial building project as an example, the target cost description label of the target cost component element is obtained based on the third confidence sequence generated by the standardized deep learning network and the fourth confidence sequence generated by the labeling network.

[0032] Suppose the confidence level of the description label "High-strength concrete cast-in-place pile (diameter 1 meter, depth 30 meters), C30 concrete, cost per meter is XXX yuan" in the third confidence level sequence is 0.7, and the confidence level of this description label in the fourth confidence level sequence is 0.8. At the same time, there may be other networks (such as the confidence level sequence corresponding to the results of the customized deep learning network in annotation mentioned in step S120, assumed here to be the sixth confidence level sequence) that affect the confidence level of this description label. For example, the confidence level of this description label in the sixth confidence level sequence is 0.75. Considering the information of these confidence level sequences comprehensively, it is determined that the target cost description label corresponding to the target cost component element of the cast-in-place pile is "High-strength concrete cast-in-place pile (diameter 1 meter, depth 30 meters), C30 concrete, cost per meter is XXX yuan", because the comprehensive confidence level of this description label in multiple confidence level sequences is the highest.

[0033] Step S150, generate the annotation result of the target BIM project cost dynamic data according to the target cost component element and the target cost description label of the target cost component element.

[0034] In the commercial building project, the target cost component elements (such as cast-in-place piles and special anti-floating component elements) and their respective target cost description labels (such as "High-strength concrete cast-in-place pile (diameter 1 meter, depth 30 meters), C30 concrete, cost per meter is XXX yuan" for cast-in-place piles, etc.) have been determined.

[0035] Generate the annotation result of the target BIM project cost dynamic data according to these target cost component elements and the target cost description labels of the target cost component elements. This annotation result will include all target cost component elements and their corresponding target cost description labels. For example, the annotation result will clearly list the cost component element information of the cast-in-place pile and its detailed cost description label, the cost component element information of the special anti-floating component element and its corresponding cost description label, etc. This annotation result can be used in multiple aspects such as project cost analysis, cost control, and budget review. For example, in terms of cost control, by comparing the actual cost with the cost per meter and other information in the cost description label in the annotation result, cost deviations can be detected in a timely manner and measures can be taken for adjustment. At the same time, during budget review, the cost component elements and their cost description labels in the annotation result can be used to check whether the budget preparation is reasonable, whether there are any omissions or overestimated costs, etc.

[0036] Based on the above steps, in the embodiments of the present application, by combining a standardized deep learning network and a customized deep learning network, efficient extraction of cost component elements from the target BIM project cost dynamic data is performed, and a confidence sequence associated with a preset cost mapping knowledge set is generated, thereby accurately identifying the target cost component elements. Further, the standardized deep learning network and the annotation network are used to carefully annotate the target cost component elements, generating a confidence sequence associated with another preset cost mapping knowledge set to accurately obtain the target cost description label of the target cost component elements. This method not only significantly improves the annotation accuracy and efficiency of project cost dynamic data, but also realizes in-depth mining and refined management of BIM project cost data, which helps with accurate estimation of project costs, cost control, and project management.

[0037] In a possible implementation manner, the customized deep learning network is generated by learning network parameters according to the feature distance between the first estimated cost component elements and the actual cost component elements of each first template BIM project cost dynamic data. The annotation network is generated by learning network parameters according to the feature distance between the second estimated cost description labels and the actual cost description labels of each second template BIM project cost dynamic data.

[0038] The standardized deep learning network is generated by learning network parameters according to the feature distance between the third annotation result of each third template BIM project cost dynamic data and the template annotation data. The third annotation result includes the third estimated cost component elements and the third cost description labels, and the template annotation data includes the actual cost component elements and the actual cost description labels of the third template BIM project cost dynamic data.

[0039] In a possible implementation manner, the method further includes:

[0040] Step A110: According to the target BIM project cost dynamic data, use the annotation network to extract cost component elements from the target BIM project cost dynamic data, generating a fifth confidence sequence associated with a first preset cost mapping knowledge set.

[0041] Step A120: According to the target BIM project cost dynamic data and the target cost component elements, use the customized deep learning network to annotate the target cost component elements, generating a sixth confidence sequence associated with a second preset cost mapping knowledge set.

[0042] Among them, step S120 includes:

[0043] Step S121: Determine the first extraction loss of the second confidence sequence based on the feature distance between the first confidence sequence and the second confidence sequence, and determine the second extraction loss of the fifth confidence sequence based on the feature distance between the first confidence sequence and the fifth confidence sequence.

[0044] Step S122: Determine the first corrected confidence of the first confidence sequence based on the deviation value between the second confidence sequence and the first extraction loss, and correct the first confidence sequence according to the first corrected confidence.

[0045] Step S123: Determine the second target confidence sequence based on the deviation value between the corrected first confidence sequence and the second extraction loss, and obtain the target cost component element according to the second target confidence sequence.

[0046] Step S140 includes: Obtain the target cost description label corresponding to the target cost component element according to the third confidence sequence, the fourth confidence sequence, and the sixth confidence sequence.

[0047] In this embodiment, first, the customized deep learning network is generated by learning network parameters based on the feature distance between the first estimated cost component element and the actual cost component element of each first template BIM project cost dynamic data. Taking multiple previously completed large commercial building projects as the source of the first template BIM project cost dynamic data, in these projects, for each first template BIM project cost dynamic data, the cost component element is extracted through the customized deep learning network with initialized parameters to obtain the first estimated cost component element. For example, in a completed large commercial shopping center project, the foundation engineering part actually includes various cost component elements, such as large-diameter cast-in-place piles, raft foundations, and diaphragm walls, etc., which are actual cost component elements. The first estimated cost component elements initially identified by the customized deep learning network may have some deviations. For example, some deep large-diameter cast-in-place piles may be misidentified as small-diameter cast-in-place piles, or the thickness of the raft foundation may be estimated incorrectly. By calculating the feature distance between these first estimated cost component elements and the actual cost component elements, such as calculating the differences between the two in terms of structural features, material usage, and cost features related to construction techniques, the network parameters of the customized deep learning network are adjusted and learned according to these feature distances, so that the network can more accurately identify cost component elements when processing similar commercial building projects.

[0048] The annotation network is generated by learning network parameters based on the characteristic distance between the second estimated cost description tags and the actual cost description tags of each second-template BIM project cost dynamic data. For example, for a series of commercial building projects as the second-template BIM project cost dynamic data, in a commercial office building project, for the beam component elements, the actual cost description tag is "rectangular-section reinforced concrete beam, C30 concrete, beam height 1 meter, width 0.5 meter, cost per meter is XXX yuan", while the second estimated cost description tag initially obtained through the annotation network may be "reinforced concrete beam, C30 concrete, size of the beam not accurately described, large deviation in cost estimation per meter". By calculating the characteristic distance between such second estimated cost description tags and actual cost description tags in terms of description accuracy, cost value accuracy, etc., the network parameters of the annotation network are adjusted, thereby improving the annotation accuracy of the annotation network in subsequent projects.

[0049] The standardized deep learning network is generated by learning network parameters based on the characteristic distance between the third annotation results and the template annotation data of each third-template BIM project cost dynamic data. Among many different types of building projects as the third-template BIM project cost dynamic data, for a large shopping mall renovation project, there may be differences between the third estimated cost component elements and the third cost description tags in the third annotation results and the actual cost component elements and the actual cost description tags in the template annotation data. For example, for the exterior renovation part of the shopping mall, some special decorative components may be missing in the estimated cost component elements, and there are also deviations between the cost values in the cost description tags and the actual ones. By calculating the characteristic distance of these differences, such as considering from aspects such as the integrity of component elements and the accuracy of cost values, the parameters of the standardized deep learning network are optimized so that it can operate more precisely when dealing with the target commercial building project.

[0050] In the target commercial building project, based on the target BIM project cost dynamic data, the annotation network extracts cost component elements from the target BIM project cost dynamic data to generate a fifth confidence sequence associated with the first preset cost mapping knowledge set. For example, in the electrical system part of the target commercial building, the annotation network will identify various cost component elements such as distribution boxes and cable trays and generate a fifth confidence sequence. For the distribution box, the annotation network gives a confidence value in the first preset cost mapping knowledge set according to its own learning results and the characteristics of the electrical system cost data of the target commercial building. This value reflects the certainty of the annotation network in identifying the cost component element of the distribution box. The same applies to other cost component elements, thus forming a fifth confidence sequence.

[0051] Based on the target BIM project cost dynamic data and the target cost component elements, the target cost component elements are labeled through a customized deep learning network to generate a sixth confidence sequence associated with the second preset cost mapping knowledge set. For example, for the target cost component elements determined in the target commercial building, such as large-diameter cast-in-place piles in the foundation project, the customized deep learning network will label them and generate a sixth confidence sequence according to the relevant information in the target BIM project cost dynamic data, such as the depth, diameter, and concrete strength of the cast-in-place piles. For a possible label like "large-diameter cast-in-place pile, C30 concrete, 30 meters deep, with a cost of XXX yuan per meter", the customized deep learning network will give a confidence value in the second preset cost mapping knowledge set. The labeling of other target cost component elements is the same, thus forming a sixth confidence sequence.

[0052] In the process of obtaining the target cost component elements in the target BIM project cost dynamic data based on the first confidence sequence and the second confidence sequence, take the structural part of the target commercial building as an example. Based on the feature distance between the first confidence sequence and the second confidence sequence, determine the first extraction loss of the second confidence sequence. Suppose the confidence in a certain structural component in the first confidence sequence is 0.8 and in the second confidence sequence is 0.7. By comparing the differences between the two in each cost component element, such as the confidence differences corresponding to relevant features such as the type and scale of the structural component, calculate the first extraction loss. At the same time, based on the feature distance between the first confidence sequence and the fifth confidence sequence, determine the second extraction loss of the fifth confidence sequence. For example, there is a difference in the confidence of a certain component in the first confidence sequence and the fifth confidence sequence. Consider this difference situation for all components to determine the second extraction loss. Based on the deviation value between the second confidence sequence and the first extraction loss, determine the first corrected confidence of the first confidence sequence. If the deviation value between the second confidence sequence and the first extraction loss shows that the second confidence sequence has unique advantages in some component elements, then adjust the confidence of the corresponding component elements in the first confidence sequence to obtain the first corrected confidence. For example, for a certain key structural component, the first corrected confidence will be appropriately increased or decreased according to the deviation value. Correct the first confidence sequence according to the first corrected confidence, and then based on the deviation value between the corrected first confidence sequence and the second extraction loss, determine the second target confidence sequence. Comprehensively consider the relationship between the corrected first confidence sequence and the second extraction loss in each component element to obtain the second target confidence sequence. Finally, based on the second target confidence sequence, obtain the target cost component elements. For example, among the structural components of the commercial building, those with a higher confidence in the second target confidence sequence are determined as the target cost component elements.

[0053] When obtaining the target cost description label of the target cost component element based on the third confidence sequence and the fourth confidence sequence, take the decoration part of the target commercial building as an example. Based on the third confidence sequence, the fourth confidence sequence, and the sixth confidence sequence, obtain the target cost description label corresponding to the target cost component element. Suppose the target cost component element is a marble floor. The confidence of the description label "natural marble floor, 2 cm thick, with a cost of XXX yuan per square meter" in the third confidence sequence is 0.7, the confidence of this description label in the fourth confidence sequence is 0.8, and in the sixth confidence sequence is 0.75. Considering the confidence of this description label and other possible description labels in these three confidence sequences comprehensively, finally determine that the target cost description label of the marble floor is "natural marble floor, 2 cm thick, with a cost of XXX yuan per square meter" because the overall confidence of this description label is the highest after comprehensively considering the three confidence sequences.

[0054] In a possible implementation manner, the method further includes: obtaining auxiliary guiding data of the project cost dynamic data annotation request.

[0055] Wherein, the auxiliary guiding data includes component extraction guiding data and component annotation guiding data. The component extraction guiding data is used to guide the extraction of cost component elements of the target cost knowledge domain in the target BIM project cost dynamic data, and the component annotation guiding data is used to guide the annotation of the extracted cost component elements.

[0056] Wherein, the first confidence sequence is obtained by extracting cost component elements from the target BIM project cost dynamic data through a standardized deep learning network based on the target BIM project cost dynamic data and the component extraction guiding data. The second confidence sequence is obtained by extracting cost component elements from the target BIM project cost dynamic data through a customized deep learning network based on the target BIM project cost dynamic data and the component extraction guiding data.

[0057] The third confidence sequence is obtained by annotating the target cost component element through a standardized deep learning network based on the target BIM project cost dynamic data, the target cost component element, and the component annotation guiding data. The fourth confidence sequence is obtained by annotating the target cost component element through an annotation network based on the target BIM project cost dynamic data, the target cost component element, and the component annotation guiding data.

[0058] In a possible implementation manner, the step of obtaining the auxiliary guiding data of the project cost dynamic data annotation request includes:

[0059] Step B110: Obtain target data from the source data node corresponding to the engineering cost dynamic data annotation request. The target data includes the type of engineering project involved, the usage scenario of the cost data, and the ultimate application purpose of the annotation data.

[0060] Step B120: Determine the target data source according to the target data. The target data source includes existing engineering cost databases, project document libraries, industry standard specification libraries, and relevant expert experience knowledge bases.

[0061] Step B130: Establish a data screening rule according to the annotation target of the engineering cost dynamic data annotation request, and extract the initial engineering cost-related data set from the target data source based on the data screening rule.

[0062] Step B140: Conduct feature analysis on the initial engineering cost-related data set to generate a data feature analysis result. The data feature analysis result includes the structural features, semantic features, time series features, and numerical features of the initial engineering cost-related data set, as well as the representation methods and association relationships of cost component elements in different data sources, the semantic changes and consistency requirements of cost description tags in different contexts, and the differences in description tags of the same cost component in different engineering projects. The time series features include the dynamic change features of the cost with the project progress and the impact of the dynamic change features on the extraction and annotation of cost component elements.

[0063] Step B150: Determine the type of auxiliary guiding data for the engineering cost dynamic data annotation request according to the data feature analysis result. The types include: rule-based guiding data, example-based guiding data, and association-based guiding data.

[0064] Step B160: Construct the structural framework of the guiding data according to the type of the auxiliary guiding data. Among them, the rule-based guiding data is classified and organized according to the cost knowledge domain, the example-based guiding data is stored hierarchically according to the component type and project type, and the association-based guiding data is constructed in the form of a relationship graph.

[0065] Step B170: Determine the hierarchical structure of the auxiliary guiding data to establish the dependency relationship and call order between different levels of data.

[0066] Step B180: Extract the corresponding target data content from the initial engineering cost-related data set according to the determined guiding data type and the hierarchical structure of the auxiliary guiding data, and integrate the extracted different types of target data content according to the constructed structural framework to construct the auxiliary guiding data.

[0067] In this embodiment, first, obtaining the auxiliary guiding data of the request for dynamic data annotation of project cost is an important prerequisite for accurate annotation of the dynamic data of project cost. The auxiliary guiding data includes component extraction guiding data and component annotation guiding data, which play a key guiding role in the extraction and annotation of cost component elements of the target BIM project cost dynamic data. Among them, the component extraction guiding data is used to guide the extraction of cost component elements of the target cost knowledge domain in the target BIM project cost dynamic data, and the component annotation guiding data is used to guide the annotation of the extracted cost component elements. For example, in a commercial building project, for the target BIM project cost dynamic data, which covers cost information from the foundation project to the decoration, and the target cost knowledge domain may be a specific structural part or a specific system (such as the electrical system). The component extraction guiding data can help accurately extract relevant cost component elements from these complex data, and the component annotation guiding data can ensure the accurate annotation of the extracted cost component elements.

[0068] The first confidence sequence is obtained by extracting cost component elements from the target BIM project cost dynamic data through a standardized deep learning network based on the target BIM project cost dynamic data and the component extraction guiding data. Taking the electrical system of a commercial building as an example, the target BIM project cost dynamic data contains cost-related information such as various electrical equipment and line laying. The component extraction guiding data may contain characteristic information about cost component elements of the electrical system, such as typical characteristics of cost component elements such as different types of distribution boxes and cable specifications. The standardized deep learning network analyzes and processes the target BIM project cost dynamic data based on this information. When identifying the cost component element of the distribution box, according to the relevant data in the target BIM project cost dynamic data (such as equipment model, power, etc.) combined with the characteristic information in the component extraction guiding data, the confidence of the distribution box in the entire cost component element system is calculated, thus forming an element in the first confidence sequence. The same applies to other cost component elements, and finally a complete first confidence sequence is obtained.

[0069] The second confidence sequence is obtained by extracting cost component elements from the target BIM project cost dynamic data through a customized deep learning network based on the target BIM project cost dynamic data and component extraction guidance data. The customized deep learning network also extracts cost component elements based on the target BIM project cost dynamic data and component extraction guidance data in a commercial building project. For example, for the special structure part of a commercial building, there may be some customized structural components. The customized deep learning network uses the special structure component feature information (such as relevant information on cost component elements such as special shapes and special materials) in the component extraction guidance data and the relevant data (such as structural dimensions and material usage) in the target BIM project cost dynamic data to identify these special structural components and give confidence levels, forming the second confidence sequence.

[0070] The third confidence sequence is obtained by annotating the target cost component elements through a standardized deep learning network based on the target BIM project cost dynamic data, the target cost component elements, and component annotation guidance data. In a commercial building project, after the target cost component elements (such as cast-in-place piles in foundation engineering) have been determined, the target BIM project cost dynamic data contains various relevant data on the cast-in-place piles (such as depth, diameter, concrete strength, etc.), and the component annotation guidance data may contain information such as the specifications and requirements for the cost annotation of cast-in-place piles. The standardized deep learning network annotates the cast-in-place piles based on this information. For example, for an annotation like "cast-in-place pile - high-strength cast-in-place pile (1-meter diameter, 30-meter depth), C30 concrete, with a cost of XXX yuan per meter", the confidence level of this annotation in the entire annotation system is calculated according to the specific data in the target BIM project cost dynamic data and the annotation specifications in the component annotation guidance data, thus forming an element in the third confidence sequence. The annotation of other target cost component elements follows the same principle, and finally the third confidence sequence is obtained.

[0071] The fourth confidence sequence is obtained by annotating target cost components based on target BIM project cost dynamic data, target cost component elements, and component annotation guidance data through an annotation network. In a commercial building project, the annotation network annotates based on target BIM project cost dynamic data, target cost component elements, and component annotation guidance data. For example, for target cost component elements in the decoration part of a commercial building (such as marble floors), the target BIM project cost dynamic data includes information such as the area and material grade of the marble floor, and the component annotation guidance data includes relevant specifications and requirements for marble floor cost annotation (such as price ranges for different grades of marble and cost adjustments corresponding to laying processes). The annotation network annotates the marble floor based on this information. For an annotation like "natural marble floor, 2 cm thick, with a cost of XXX yuan per square meter", the confidence of this annotation in the entire annotation system is calculated according to the specific data in the target BIM project cost dynamic data and the annotation specifications in the component annotation guidance data, thus forming an element in the fourth confidence sequence. The annotation of other target cost component elements follows the same principle, and finally the fourth confidence sequence is obtained.

[0072] In the specific steps of obtaining auxiliary guidance data for a project cost dynamic data annotation request, target data is obtained from the source data node corresponding to the project cost dynamic data annotation request. These target data include the type of engineering project involved, the usage scenario of cost data, and the ultimate application purpose of the annotation data. In a commercial building project, the type of engineering project is a large commercial building, and the usage scenario of cost data may include project budget preparation, cost control, settlement review, etc. The ultimate application purpose of the annotation data may be to accurately evaluate the project cost and ensure that the project cost is within a reasonable range, etc.

[0073] Based on these target data, target data sources are determined. The target data sources include existing project cost databases, project document libraries, industry standard specification libraries, and relevant expert experience knowledge bases. In a commercial building project, the existing project cost database may contain cost data of previous similar commercial building projects, the project document library contains documents such as design documents and construction plans of the commercial building project, which may involve cost-related information, the industry standard specification library contains standards and specifications for cost calculation and cost component definition in the construction industry, and the relevant expert experience knowledge base contains expert experience knowledge in commercial building cost management.

[0074] According to the annotation target of the project cost dynamic data annotation request, establish data screening rules, and extract the initial project cost-related data set from the target data source based on these data screening rules. For example, if the annotation target is to accurately annotate the cost component elements of the structural part of a commercial building, the data screening rules may be to screen out the data related to the cost of the structural part, screen out the cost data of the structural parts of previous commercial buildings from the project cost database, screen out the cost information in the structural design-related documents from the project document library, screen out the structural cost calculation specifications from the industry standard specification library, and screen out the expert experience knowledge on the structural cost management of commercial buildings from the expert experience knowledge base, so as to form an initial project cost-related data set.

[0075] Conduct feature analysis on the initial project cost-related data set to generate the data feature analysis results. In the commercial building project, the data feature analysis results include the structural features, semantic features, time series features, and numerical features of the initial project cost-related data set, as well as the representation methods and correlation relationships of cost component elements in different data sources, the semantic changes and consistency requirements of cost description tags in different contexts, and the description tag differences of the same cost component in different engineering projects. In terms of structural features, for example, the data structure in the project cost database may be stored according to different building parts and cost classifications, and the cost information in the project document library may be a structure associated with specific design structures and construction processes. In terms of semantic features, there may be differences in the semantic expressions of the same cost component element in different data sources. For example, in the project cost database, professional cost terms may be used to describe beam components, while in the project document library, it may be described from a design perspective. The time series features include the dynamic change features of costs with the project progress and the impact of the dynamic change features on the extraction and annotation of cost component elements. For example, in the commercial building project, as the project progresses from the foundation project to the main structure and then to the decoration and finishing, the importance and value of cost component elements at different stages will change. In the foundation project stage, the cost component element of the foundation pile is the key object of attention, and its dynamic cost changes (such as the increase in pile length due to changes in geological conditions) will affect the extraction and annotation of cost component elements.

[0076] Based on the results of data feature analysis, determine the types of auxiliary guiding data for the dynamic data annotation request of project cost. The types include rule-based guiding data, example-based guiding data, and association-based guiding data. In a commercial building project, the rule-based guiding data may be the standard specifications for cost calculation and the definition of cost component elements, such as the cost calculation rules for structural components; the example-based guiding data may be the specific cost component elements and annotation examples in previous similar commercial building projects, such as the cost component elements and annotation instances of the foundation piles of a completed commercial building; the association-based guiding data may be the association relationships between different cost component elements (such as the impact of the association relationship between beams and columns on the structure's force-bearing on the cost) constructed in the form of a relationship map.

[0077] According to the types of auxiliary guiding data, construct the structural framework of the guiding data. The rule-based guiding data is classified and organized according to the cost knowledge domain. For example, the rules related to structural cost, the rules related to electrical cost, etc. are classified separately. The example-based guiding data is stored hierarchically according to the component type and project type. For example, the cost component element examples of the foundation project and the cost component element examples of the main structure are stored hierarchically according to different project types (such as commercial buildings, commercial office buildings, etc.). The association-based guiding data is constructed in the form of a relationship map. For example, the force-bearing association relationships between structural components such as beams, columns, and slabs and the impact of such association relationships on the cost are constructed into a relationship map.

[0078] Determine the hierarchical structure of the auxiliary guiding data to establish the dependency relationships and call sequences between different levels of data. For example, in a commercial building project, the rule-based guiding data may be at a higher level, providing the basic rule basis for the extraction and annotation of cost component elements; the example-based guiding data is at an intermediate level, providing example references for specific operations; the association-based guiding data is at a deeper level, used to handle the complex association relationships between cost component elements.

[0079] According to the determined guiding data types and the hierarchical structure of the auxiliary guiding data, extract the corresponding target data content from the initial project cost-related data set, and integrate the extracted target data content of different types according to the constructed structural framework to construct the auxiliary guiding data. For example, extract the rule-based guiding data content such as the structural cost calculation rules from the initial project cost-related data set and classify and organize them according to the cost knowledge domain; extract the example-based guiding data content such as the cost component element examples and annotation examples of the foundation piles in different commercial building projects and store them hierarchically according to the component type and project type; extract the association-based guiding data content such as the association relationships between structural components such as beams and columns and construct them into a relationship map. Finally, integrate these different types of target data content to construct a complete auxiliary guiding data, which is used to provide accurate guidance in the process of extracting and annotating cost component elements in commercial building projects.

[0080] In a possible implementation, step S120 may further include:

[0081] Determine a first extraction loss of the second confidence sequence according to a feature distance between the first confidence sequence and the second confidence sequence.

[0082] Determine a first corrected confidence for the first confidence sequence according to a deviation value between the second confidence sequence and the first extraction loss.

[0083] Correct the first confidence sequence according to the first corrected confidence to generate a first target confidence sequence.

[0084] Obtain the target cost component element according to the first target confidence sequence.

[0085] In this embodiment, in the cost analysis of a commercial building, the first confidence sequence and the second confidence sequence contain confidence information about various cost component elements. For example, in the structural part of a commercial building, the first confidence sequence is obtained by a standardized deep learning network based on target BIM project cost dynamic data and component extraction guidance data. There may be a confidence value for a beam component, and a corresponding confidence value for a column component, etc.; the second confidence sequence is obtained by a customized deep learning network based on the same target BIM project cost dynamic data and component extraction guidance data, and also contains confidence values of components such as beams and columns. The calculation of the feature distance is based on factors such as the confidence difference of these components in the two sequences and the attributes of the components themselves. For example, for a beam component, the confidence in the first confidence sequence is 0.8, and the confidence in the second confidence sequence is 0.7. At the same time, consider the influence of relevant attributes such as the size and material of the beam on this difference to comprehensively calculate the feature distance, and then determine the first extraction loss of the second confidence sequence. This first extraction loss reflects the deviation degree of the second confidence sequence in terms of cost component element extraction relative to the first confidence sequence.

[0086] Next, based on the deviation value between the second confidence sequence and the first extraction loss, determine the first corrected confidence for the first confidence sequence. In a commercial building project, there is a certain relationship between the second confidence sequence and the first extraction loss. For example, if the deviation value between the confidence of some cost component elements in the second confidence sequence and the first extraction loss is large, it indicates that the performance of these component elements in the second confidence sequence is inconsistent with the overall extraction loss situation. Taking the column component as an example, if the confidence of the column component in the second confidence sequence is 0.6, and the deviation value related to the column component calculated according to the first extraction loss is large, this may mean that there are special circumstances in the identification of the column component by the customized deep learning network. Adjust the confidence of the column component in the first confidence sequence according to this deviation value, so as to determine the first corrected confidence for the first confidence sequence. The same applies to other cost component elements. By comprehensively considering the deviation value situations of all component elements, the confidence values in the first confidence sequence are adjusted accordingly.

[0087] Then, correct the first confidence sequence according to the first corrected confidence to generate the first target confidence sequence. In a commercial building project, replace each confidence value in the first confidence sequence with the first corrected confidence to obtain the first target confidence sequence. For example, the original confidence of the beam component in the first confidence sequence is 0.8, and after correction, it becomes 0.85 according to the first corrected confidence; the original confidence of the column component is 0.7, and after correction, it becomes 0.72, etc. This first target confidence sequence combines the information of the first confidence sequence and the second confidence sequence. Through the analysis and processing of the feature distance, the first extraction loss, the deviation value, etc. between the two, a more optimized confidence sequence is obtained.

[0088] Finally, based on the first target confidence sequence, obtain the target cost component elements. In the cost management of a commercial building, the confidence values in the first target confidence sequence reflect the possibility degree of each cost component element in the target BIM project cost dynamic data. For example, in the electrical system of a commercial building, for cost component elements such as distribution boxes and cable trays, screen according to the confidence values in the first target confidence sequence. If the confidence value of the distribution box in the first target confidence sequence is higher than the set threshold, then determine the distribution box as the target cost component element; if the confidence value of the cable tray is lower than the threshold, then exclude it. In this way, screen the cost component elements of each part (such as the structural part, the decoration part, the electrical system, etc.) of the commercial building, and finally determine the target cost component elements in the target BIM project cost dynamic data. These target cost component elements will be used as the basis for subsequent annotation and other operations, which is of great significance for accurately evaluating the project cost of the commercial building.

[0089] In a possible implementation, the step of determining the first confidence sequence and the second confidence sequence includes:

[0090] Step C110, based on the target BIM engineering cost dynamic data, cost component elements are extracted from the target BIM engineering cost dynamic data through a standardized deep learning network to generate a first basic confidence sequence associated with a first preset cost mapping knowledge set, and cost component elements are extracted from the target BIM engineering cost dynamic data through a customized deep learning network to generate a second basic confidence sequence associated with the first preset cost mapping knowledge set.

[0091] The first basic confidence sequence includes the first basic confidence of each reference cost component element, and the second basic confidence includes the second basic confidence of each reference cost component element.

[0092] Step C120: extracting, from the reference cost component elements, target reference component elements whose first basic confidence and second basic confidence are both greater than or equal to a set confidence, according to the first basic confidence sequence and the second basic confidence sequence.

[0093] Step C130: remove the confidences corresponding to the non-target reference component elements in the first basic confidence sequence and the second basic confidence sequence, respectively, to generate the first confidence sequence and the second confidence sequence.

[0094] In this embodiment, taking the foundation project of a commercial building as an example, the target BIM project cost dynamic data includes cost information in all aspects of the foundation project, such as cost data of different types of foundation piles (cast-in-place piles, precast piles, etc.), cost information of earth excavation and backfilling, and cost of the foundation waterproofing project. Based on its existing learning models and algorithms, the standardized deep learning network analyzes and processes this data, and matches and calculates the confidence level for the reference cost component elements in the first preset cost mapping knowledge set. The first preset cost mapping knowledge set contains numerous cost component elements that may appear in the foundation project, such as various types of pile foundations, different grades of concrete, and various types of steel bars. For the reference cost component element of the cast-in-place pile, the standardized deep learning network calculates a confidence value according to the relevant data in the target BIM project cost dynamic data (such as information on the diameter, depth, and concrete strength of the cast-in-place pile), and this value is the first foundation confidence level of the cast-in-place pile in the first foundation confidence sequence. Similarly, the customized deep learning network extracts cost component elements for the same target BIM project cost dynamic data. For the reference cost component element of the cast-in-place pile, according to its own customized learning results and algorithms, it calculates the second foundation confidence level in the second foundation confidence sequence. Other reference cost component elements, such as precast piles and foundation waterproofing membranes, also obtain their respective first foundation confidence levels and second foundation confidence levels under the processing of the two networks, thus forming a complete first foundation confidence sequence and second foundation confidence sequence.

[0095] Then, based on the first foundation confidence sequence and the second foundation confidence sequence, target reference component elements with the first foundation confidence level and the second foundation confidence level both greater than or equal to the set confidence level are extracted from each reference cost component element. In the commercial building project, the set confidence level is a threshold determined according to project requirements and past experience, for example, set to 0.6. In the foundation project part, for the reference cost component element of the cast-in-place pile, if its first foundation confidence level in the first foundation confidence sequence is 0.7 and its second foundation confidence level in the second foundation confidence sequence is 0.8, both greater than the set confidence level of 0.6, then the cast-in-place pile is determined as the target reference component element. For other reference cost component elements, such as the first foundation confidence level of the precast pile is 0.5 and the second foundation confidence level is 0.65, since the first foundation confidence level is less than the set confidence level, the precast pile cannot be determined as the target reference component element. Through such comparison and screening, the target reference component elements that meet the conditions are determined from numerous reference cost component elements.

[0096] Finally, the confidence levels corresponding to the non-target reference component elements in the first basic confidence level sequence and the second basic confidence level sequence are respectively removed to generate the first confidence level sequence and the second confidence level sequence. In the basic project of a commercial building, for the first basic confidence level sequence, after determining the target reference component element (such as bored pile), the confidence levels corresponding to other non-target reference component elements (such as precast pile, etc.) are removed from the first basic confidence level sequence. For example, if the first basic confidence level sequence originally includes the confidence levels of reference cost component elements such as bored pile, precast pile, and foundation waterproofing membrane, after removing the confidence levels of non-target reference component elements such as precast pile, the confidence levels of target reference component elements such as bored pile that remain form the first confidence level sequence. Similarly, for the second basic confidence level sequence, after removing the confidence levels corresponding to the non-target reference component elements, the second confidence level sequence is obtained. The first confidence level sequence and the second confidence level sequence obtained in this way are more focused on the target reference component elements, providing a more targeted data basis for accurately obtaining the target cost component elements in the target BIM project cost dynamic data, and helping to improve the accuracy and efficiency of the entire project cost analysis and management.

[0097] In a possible implementation manner, the target BIM project cost dynamic data includes multiple target cost component elements, and the annotation result includes multiple target cost component elements and target cost description labels of the target cost component elements.

[0098] The annotation result is generated by enabling an x-round deep learning process. Each round of the deep learning process is a round of deep learning component extraction process or a round of deep learning component annotation process. The deep learning result of each round of the deep learning process is a target cost component element or a target cost description label.

[0099] Among them, the specific steps of the a-th round of deep learning process include:

[0100] Step D110, determine the output category of the result to be output according to the deep learning result of the (a - 1)-th round of deep learning process. The output category is a component extraction category or a component annotation category. The output category corresponding to the first round of deep learning process is a component extraction category.

[0101] Step D120, according to the target BIM project cost dynamic data and the deep learning results of the previous (a - 1) rounds of deep learning process, obtain a standardized confidence level sequence through a standardized deep learning network, obtain a customized confidence level sequence through a customized deep learning network, and obtain an annotated confidence level sequence through an annotation network.

[0102] Among them, the standardized confidence sequence is the first confidence sequence or the third confidence sequence, the customized confidence sequence is the second confidence sequence or the sixth confidence sequence, and the labeled confidence sequence is the fourth confidence sequence or the fifth confidence sequence.

[0103] Step D130: Generate the deep learning result of the a-th round of deep learning process according to the standardized confidence sequence, the customized confidence sequence, and the labeled confidence sequence.

[0104] In this embodiment, taking the electrical system of a commercial building as an example, the target cost component elements may include numerous elements such as distribution boxes, cable trays, wires and cables, etc. Each round of deep learning process is either a deep learning component extraction process or a deep learning component labeling process, and the deep learning result of each round is a target cost component element or a target cost description label.

[0105] For the a-th round of deep learning process, first determine the output category of the result to be output according to the deep learning result of the (a - 1)-th round of deep learning process. If it is the first round of deep learning process, its corresponding output category is the component extraction category. This is because in the initial stage, it is necessary to first identify the target cost component elements from the target BIM project cost dynamic data. For example, in the first round of deep learning process, for the target BIM project cost dynamic data of the electrical system of a commercial building, the first thing to do is to find out cost component elements such as distribution boxes and cable trays. This process is based on the algorithm of the deep learning network and the cost information related to the electrical system in the target BIM project cost dynamic data, such as equipment model, quantity, specification, etc.

[0106] Next, obtain the standardized confidence sequence through the standardized deep learning network, the customized confidence sequence through the customized deep learning network, and the labeled confidence sequence through the labeling network according to the target BIM project cost dynamic data and the deep learning results of the previous (a - 1) rounds of deep learning process. Assume that in the second round of deep learning process (at this time a = 2), the distribution box, a target cost component element, has been identified in the first round. For the standardized deep learning network, according to the cost information of the electrical system in the target BIM project cost dynamic data (such as the cost corresponding to different functional modules of the distribution box, the price difference of different brands, etc.) and the deep learning result of the first round (it has been determined that there is a distribution box component), analyze the cost description labels related to the distribution box, calculate the confidence corresponding to each possible cost description label, and thus obtain the standardized confidence sequence (here it may be the third confidence sequence because it is in the round related to component labeling). For example, for the possible cost description label "Distribution box - 10 circuits, brand A, cost per unit is XXX yuan", the standardized deep learning network calculates a confidence value based on the data.

[0107] Similarly, the customized deep learning network will also perform analysis and processing based on the same target BIM project cost dynamic data and the deep learning results of the first round. Since the customized deep learning network is generated by learning network parameters according to the feature distance between the first estimated cost component elements and the actual cost component elements of each first-template BIM project cost dynamic data, in the scenario of the electrical system of a commercial building, the cost description labels of the distribution boxes can be analyzed based on the learning experience of similar electrical systems in previous commercial buildings to obtain a customized confidence sequence (which may be the sixth confidence sequence here). For example, for the label "Distribution box - 10 circuits, brand A, cost per unit is XXX yuan", the customized deep learning network gives a confidence value based on its own learning results.

[0108] The annotation network will also operate based on the target BIM project cost dynamic data and the deep learning results of the first round. The annotation network is generated by learning network parameters according to the feature distance between the second estimated cost description labels and the actual cost description labels of each second-template BIM project cost dynamic data. It will analyze the cost description labels of the distribution boxes to obtain an annotation confidence sequence (which may be the fourth confidence sequence here). For example, for the label "Distribution box - 10 circuits, brand A, cost per unit is XXX yuan", the annotation network gives a confidence value based on its own learning and calculation.

[0109] Finally, based on the standardized confidence sequence, the customized confidence sequence, and the annotation confidence sequence, the deep learning result of the a-th round of the deep learning process is generated. In the above-mentioned second round of the deep learning process, the confidence levels of the label "Distribution box - 10 circuits, brand A, with a cost of XXX yuan per unit" and other possible labels in the three confidence sequences are comprehensively considered. If the confidence level of this label in the standardized confidence sequence is 0.7, in the customized confidence sequence is 0.8, and in the annotation confidence sequence is 0.75, a specific algorithm (such as weighted average or other comprehensive evaluation algorithms) is used to determine whether this label is the final target cost description label or to further adjust the deep learning operations in subsequent rounds based on this result. If this label is finally determined as the target cost description label, then this is the deep learning result of the second round of the deep learning process; if further analysis of other possible labels or operations such as the extraction of components in the next round is required, corresponding adjustments are made based on this result to provide a basis for the subsequent deep learning process until a complete annotation result is generated, including all target cost component elements and their target cost description labels in the electrical system of the commercial building. This process is continuously repeated in each round of the deep learning process. The operation category of the next round is determined based on the result of the previous round. The corresponding confidence sequences are obtained through different deep learning networks, and then the deep learning results of each round are generated based on these confidence sequences, gradually constructing a complete annotation result.

[0110] In a possible implementation manner, the training steps of the customized deep learning network include:

[0111] Step S101, obtain a first sample learning data sequence, where the first sample learning data sequence includes a plurality of first template BIM project cost dynamic data carrying template annotation data, and the template annotation data of each first template BIM project cost dynamic data includes the actual cost component elements of the first template BIM project cost dynamic data.

[0112] In this embodiment, a series of completed large commercial building projects are taken as an example. These projects serve as the source of the first template BIM project cost dynamic data. Each project has its unique project cost composition. For example, in the foundation engineering part of one commercial building project, the actual cost component elements in the template annotation data include large-diameter cast-in-place piles, raft foundations, diaphragm walls, etc., and also include the relevant cost information of these component elements in this project, such as the depth, diameter, concrete strength grade of the cast-in-place piles and the corresponding cost values. The data of these projects together constitute the first sample learning data sequence, providing basic data for the training of the customized deep learning network.

[0113] Step S102: Based on each piece of first-template BIM project cost dynamic data, use a customized deep learning network with initialized parameters to extract cost component elements from the each piece of first-template BIM project cost dynamic data, and generate first estimated cost component elements for each piece of first-template BIM project cost dynamic data.

[0114] For example, when processing the data of one commercial building project, the customized deep learning network analyzes the project's project cost data according to its initialized parameters and algorithms. When analyzing the foundation engineering part, since the network parameters may be inaccurate initially, some features of large-diameter cast-in-place piles may be misidentified, and thus the generated first estimated cost component elements may have deviations. For example, the diameter of the large-diameter cast-in-place pile may be estimated incorrectly, or part of the concrete consumption of the raft foundation may be miscomputed to other components, resulting in less accurate first estimated cost component elements. This situation is common in the initial stage of training because the network is still in the learning process.

[0115] Step S103: Based on the each piece of first-template BIM project cost dynamic data and the first estimated cost component elements, use a customized deep learning network with initialized parameters to label the first estimated cost component elements, generate first estimated cost description labels for the first estimated cost component elements, and generate annotation results for the each piece of first-template BIM project cost dynamic data based on the first estimated cost component elements and the first estimated cost description labels of the each piece of first-template BIM project cost dynamic data.

[0116] Taking the large-diameter cast-in-place pile in the foundation engineering as an example, according to the previously extracted first estimated cost component elements (even if there are some errors), the customized deep learning network attempts to label it. If the estimated diameter of the cast-in-place pile in the first estimated cost component elements is incorrect, then when labeling its cost description label, an inaccurate first estimated cost description label may be generated based on the incorrect diameter information, such as "large-diameter cast-in-place pile (incorrect diameter), C30 concrete, cost per meter is XXX yuan (incorrect value)". Then, based on the first estimated cost component elements and the first estimated cost description labels of each piece of first-template BIM project cost dynamic data, the annotation results for each piece of first-template BIM project cost dynamic data are generated. This annotation result contains inaccurate cost component elements and cost description label information, but it is an intermediate result in the network training process.

[0117] Step S104: Determine the first network learning error based on the feature distance between the first estimated cost component elements and the actual cost component elements of each piece of first-template BIM project cost dynamic data, and train the customized deep learning network with initialized parameters according to the first network learning error.

[0118] For example, for the component element of a large-diameter cast-in-place pile, compare the incorrect diameter, concrete consumption, etc. in the first estimated cost component element with the correct values in the actual cost component element, calculate the differences in terms of structural characteristics, material consumption, etc., so as to obtain the characteristic distance. These characteristic distances comprehensively reflect the first network learning error. According to this first network learning error, train the customized deep learning network with initialized parameters. During the training process, the network will adjust its internal parameters according to the error, such as adjusting the neuron connection weights related to the identification and annotation of the cost component element of the cast-in-place pile, etc., to reduce the error. With the processing and learning of more first-template BIM project cost dynamic data, the customized deep learning network gradually improves the accuracy of extracting and annotating cost component elements.

[0119] The training steps of the annotation network include:

[0120] Step S201, obtain a second sample learning data sequence, where the second sample learning data sequence includes multiple second-template BIM project cost dynamic data carrying template annotation data, and the template annotation data of each second-template BIM project cost dynamic data includes the actual cost description labels of each actual cost component element of this second-template BIM project cost dynamic data.

[0121] Taking multiple completed commercial building projects as an example, the template annotation data of each second-template BIM project cost dynamic data includes the actual cost description labels of each actual cost component element of this project. For example, in a commercial office building project, for the beam component element, its actual cost description label is "rectangular-section reinforced concrete beam, C30 concrete, beam height 1 meter, width 0.5 meter, cost per meter is XXX yuan", and there are corresponding accurate actual cost description labels for the column component element as well. The data of these different commercial building projects together constitute the second sample learning data sequence, providing a data basis for the training of the annotation network.

[0122] Step S202, based on each second-template BIM project cost dynamic data, use the annotation network with initialized parameters to extract cost component elements from each second-template BIM project cost dynamic data, and generate second estimated cost component elements for each second-template BIM project cost dynamic data.

[0123] For example, when processing the data of a commercial building project, due to the initialized parameters, the annotation network may have deviations when identifying cost component elements. For example, in the structural part of the building, some beams with special shapes may be misidentified as ordinary beams, resulting in inaccurate second estimated cost component elements for each second-template BIM project cost dynamic data.

[0124] Step S203: Based on each piece of second-template BIM project cost dynamic data and the second estimated cost component elements, use the annotation network with initialized parameters to annotate the second estimated cost component elements, generate the second estimated cost description tags for the second estimated cost component elements, and generate the annotation results for each piece of second-template BIM project cost dynamic data based on the second estimated cost component elements and the second estimated cost description tags of each piece of second-template BIM project cost dynamic data.

[0125] For example, for a special-shaped beam that was previously misidentified as a common beam, inaccurate second estimated cost description tags may be generated during annotation, such as "common beam (actually a special-shaped beam), C30 concrete, cost per meter is YYY yuan (incorrect value)". Then, based on the second estimated cost component elements and the second estimated cost description tags of each piece of second-template BIM project cost dynamic data, the annotation results for each piece of second-template BIM project cost dynamic data are generated, and this result contains the error information in the extraction and annotation process.

[0126] Step S204: Determine the second network learning error based on the feature distance between the second estimated cost description tags and the actual cost description tags of each piece of second-template BIM project cost dynamic data, and train the annotation network with initialized parameters based on the second network learning error.

[0127] For example, for beam component elements, compare the beam type, cost value, etc. in the inaccurate second estimated cost description tags with the correct information in the actual cost description tags, calculate the feature distances in terms of description accuracy, cost value accuracy, etc., and these feature distances are combined to determine the second network learning error. Based on this second network learning error, the annotation network with initialized parameters is trained. During the training process, the annotation network will adjust its internal parameters, such as adjusting the algorithm parameters related to the annotation of beam component elements, etc., to improve the accuracy of annotating cost component elements. With the learning of more second-template BIM project cost dynamic data, the annotation network gradually improves its annotation ability.

[0128] The training steps of the standardized deep learning network include:

[0129] Step S301: Obtain a third sample learning data sequence, where the third sample learning data sequence includes multiple third-template BIM project cost dynamic data carrying template annotation data, and the template annotation data of each third-template BIM project cost dynamic data includes the actual cost component elements of the third-template BIM project cost dynamic data and the actual cost description tags of each actual cost component element.

[0130] Taking various types of completed construction projects (including commercial buildings, residential buildings, etc.) as examples, the template annotation data of the BIM project cost dynamic data for each third template includes the actual cost component elements of the project and the actual cost description labels of each actual cost component element. For example, in a large shopping mall renovation project, the actual cost component elements include the exterior facade decoration components, interior decoration components, etc., and their actual cost description labels detail information such as the type, specification, material of the components, and the corresponding cost values. The data of these different projects constitute the third sample learning data sequence.

[0131] Step S302: According to each third-template BIM project cost dynamic data, use the standardized deep learning network with initialized parameters to extract cost component elements from each third-template BIM project cost dynamic data, and generate the third estimated cost component elements of each third-template BIM project cost dynamic data.

[0132] Step S303: According to each third-template BIM project cost dynamic data and the third estimated cost component elements, use the standardized deep learning network with initialized parameters to label the third estimated cost component elements, generate the third estimated cost description labels of the third estimated cost component elements, and generate the annotation results of each third-template BIM project cost dynamic data according to the third estimated cost component elements and the third estimated cost description labels of each third-template BIM project cost dynamic data.

[0133] For example, when processing the data of a shopping mall renovation project, the standardized deep learning network may miss some small decoration components or misestimate the size of some components due to inaccurate initialized parameters when extracting the exterior facade decoration components, thus obtaining third estimated cost component elements with some errors. Then, according to each third-template BIM project cost dynamic data and the third estimated cost component elements, use the standardized deep learning network with initialized parameters to label the third estimated cost component elements and generate the third estimated cost description labels. For example, for the exterior facade decoration components with misestimated sizes, inaccurate third estimated cost description labels will be generated during the labeling. Then, generate the annotation results of each third-template BIM project cost dynamic data according to the third estimated cost component elements and the third estimated cost description labels of each third-template BIM project cost dynamic data, and this result contains inaccurate information in the extraction and labeling processes.

[0134] Step S304: Determine the third network learning error according to the feature distance between the third annotation results of each third-template BIM project cost dynamic data and the template annotation data, and train the standardized deep learning network with initialized parameters according to the third network learning error.

[0135] For example, for the exterior decoration components, compare the incorrect component elements and inaccurate cost description labels in the third annotation result with the correct actual cost component elements and actual cost description labels in the template annotation data, calculate the feature distances in terms of component integrity, cost value accuracy, etc., and these feature distances are combined to determine the third network learning error. According to this third network learning error, train the standardized deep learning network with initialized parameters. During the training process, the standardized deep learning network will adjust various internal parameters, such as neuron connection weights, coefficients in the algorithm, etc., to improve the accuracy of cost component element extraction and annotation. With the learning of more third-template BIM project cost dynamic data, the performance of the standardized deep learning network is continuously improved, and it can more accurately handle the cost component element extraction and annotation tasks in the target BIM project cost dynamic data.

[0136] Figure 2 FIG. shows the hardware structure diagram of a BIM technology-based project cost dynamic data annotation system 100 for implementing the above-mentioned BIM technology-based project cost dynamic data annotation method, as Figure 2 shown, the BIM technology-based project cost dynamic data annotation system 100 may include a processor 110, a machine-readable storage medium 120, a bus 130, and a communication unit 140.

[0137] The machine-readable storage medium 120 can store data and / or instructions. In some embodiments, the machine-readable storage medium 120 can store data obtained from an external terminal. In some embodiments, the machine-readable storage medium 120 can store the data and / or instructions that the BIM technology-based project cost dynamic data annotation system 100 uses to execute or complete the exemplary methods described in the present invention.

[0138] In a specific implementation process, one or more processors 110 execute the computer-executable instructions stored in the machine-readable storage medium 120, so that the processor 110 can execute the BIM technology-based project cost dynamic data annotation method in the above method embodiments. The processor 110, the machine-readable storage medium 120, and the communication unit 140 are connected through the bus 130, and the processor 110 can be used to control the transceiver actions of the communication unit 140.

[0139] For the specific implementation process of the processor 110, reference can be made to the various method embodiments executed by the above-mentioned BIM technology-based project cost dynamic data annotation system 100. The implementation principles and technical effects are similar, and will not be elaborated here in this embodiment.

[0140] In addition, an embodiment of the present invention further provides a readable storage medium, in which computer-executable instructions are preset. When a processor executes the computer-executable instructions, the above method for dynamically annotating project cost data based on BIM technology is implemented.

[0141] It should be noted that, in order to simplify the description of the present invention disclosure and thus help the understanding of one or more embodiments of the invention, in the foregoing description of the embodiments of the present invention, sometimes multiple features are merged into one embodiment, drawing or description thereof.

Claims

1. A method for dynamically annotating project cost data based on BIM technology, characterized in that, The method includes: Obtain the target BIM project cost dynamic data to be labeled, extract cost component elements from the target BIM project cost dynamic data through a standardized deep learning network, generate a first confidence sequence associated with a first preset cost mapping knowledge set, extract cost component elements from the target BIM project cost dynamic data through a customized deep learning network, generate a second confidence sequence associated with the first preset cost mapping knowledge set, and the first preset cost mapping knowledge set includes multiple reference cost component elements; Based on the first confidence sequence and the second confidence sequence, obtain the target cost component elements in the target BIM project cost dynamic data; Based on the target BIM project cost dynamic data and the target cost component elements, label the target cost component elements through the standardized deep learning network, generate a third confidence sequence associated with a second preset cost mapping knowledge set, label the target cost component elements through a labeling network, generate a fourth confidence sequence associated with the second preset cost mapping knowledge set, and the second preset cost mapping knowledge set includes multiple reference cost description labels; Based on the third confidence sequence and the fourth confidence sequence, obtain the target cost description labels of the target cost component elements; Based on the target cost component elements and the target cost description labels of the target cost component elements, generate the labeling result of the target BIM project cost dynamic data; The step of obtaining the target cost component elements in the target BIM project cost dynamic data based on the first confidence sequence and the second confidence sequence includes: Determine the first extraction loss of the second confidence sequence based on the feature distance between the first confidence sequence and the second confidence sequence; Determine the first corrected confidence for the first confidence sequence based on the deviation value between the second confidence sequence and the first extraction loss; Correct the first confidence sequence based on the first corrected confidence to generate a first target confidence sequence; Based on the first target confidence sequence, obtain the target cost component elements.

2. The method for dynamically annotating project cost data based on BIM technology according to claim 1, wherein The customized deep learning network is generated by learning network parameters based on the feature distance between the first estimated cost component elements and the actual cost component elements of each first template BIM project cost dynamic data; the labeling network is generated by learning network parameters based on the feature distance between the second estimated cost description labels and the actual cost description labels of each second template BIM project cost dynamic data; The standardized deep learning network is generated by learning network parameters based on the feature distance between the third labeling result of each third template BIM project cost dynamic data and the template labeling data, the third labeling result includes the third estimated cost component elements and the third cost description labels, and the template labeling data includes the actual cost component elements and the actual cost description labels of the third template BIM project cost dynamic data.

3. The method for dynamically annotating project cost data based on BIM technology according to claim 2, wherein The method further includes: Based on the target BIM project cost dynamic data, extract cost component elements from the target BIM project cost dynamic data through a labeled network, and generate a fifth confidence sequence associated with the first preset cost mapping knowledge set; Based on the target BIM project cost dynamic data and the target cost component elements, label the target cost component elements through the customized deep learning network, and generate a sixth confidence sequence associated with the second preset cost mapping knowledge set; Among them, obtaining the target cost component elements in the target BIM project cost dynamic data based on the first confidence sequence and the second confidence sequence includes: Determine the first extraction loss of the second confidence sequence according to the feature distance between the first confidence sequence and the second confidence sequence, and determine the second extraction loss of the fifth confidence sequence according to the feature distance between the first confidence sequence and the fifth confidence sequence; Determine the first corrected confidence of the first confidence sequence according to the deviation value between the second confidence sequence and the first extraction loss, and correct the first confidence sequence according to the first corrected confidence; Determine the second target confidence sequence according to the deviation value between the corrected first confidence sequence and the second extraction loss, and obtain the target cost component elements according to the second target confidence sequence; Among them, obtaining the target cost description label of the target cost component element according to the third confidence sequence and the fourth confidence sequence includes: Obtain the target cost description label corresponding to the target cost component element according to the third confidence sequence, the fourth confidence sequence and the sixth confidence sequence.

4. The method for dynamically annotating project cost data based on BIM technology according to claim 1, wherein, The method further includes: Obtain auxiliary guiding data for the project cost dynamic data annotation request; Among them, the auxiliary guiding data includes component extraction guiding data and component annotation guiding data. The component extraction guiding data is used to guide the extraction of cost component elements in the target cost knowledge domain of the target BIM project cost dynamic data, and the component annotation guiding data is used to guide the annotation of the extracted cost component elements; Among them, the first confidence sequence is obtained by extracting cost component elements from the target BIM project cost dynamic data through a standardized deep learning network based on the target BIM project cost dynamic data and the component extraction guiding data; the second confidence sequence is obtained by extracting cost component elements from the target BIM project cost dynamic data through a customized deep learning network based on the target BIM project cost dynamic data and the component extraction guiding data; The third confidence sequence is obtained by labeling the target cost component elements through a standardized deep learning network based on the target BIM project cost dynamic data, the target cost component elements and the component annotation guiding data; the fourth confidence sequence is obtained by labeling the target cost component elements through a labeled network based on the target BIM project cost dynamic data, the target cost component elements and the component annotation guiding data.

5. The method for dynamically annotating project cost data based on BIM technology according to claim 4, wherein The steps of obtaining the auxiliary guiding data for the dynamic data annotation request of project cost include: Obtain target data from the source data node corresponding to the dynamic data annotation request of project cost, where the target data includes the type of engineering project involved, the usage scenario of cost data, and the ultimate application purpose of the annotation data; Determine the target data source according to the target data, where the target data source includes the existing project cost database, project document library, industry standard specification library, and relevant expert experience knowledge library; Establish a data screening rule according to the annotation target of the dynamic data annotation request of project cost, and extract the initial project cost-related data set from the target data source based on the data screening rule; Conduct feature analysis on the initial project cost-related data set to generate a data feature analysis result, where the data feature analysis result includes the structural features, semantic features, time series features, and numerical features of the initial project cost-related data set, as well as the representation methods and association relationships of cost component elements in different data sources, the semantic changes and consistency requirements of cost description tags in different contexts, and the description tag differences of the same cost component in different engineering projects. The time series features include the dynamic change features of cost with the project progress, and the impact of the dynamic change features on the extraction and annotation of cost component elements; Determine the type of the auxiliary guiding data for the dynamic data annotation request of project cost according to the data feature analysis result, where the type includes: rule-based guiding data, example-based guiding data, and association-based guiding data; Construct the structural framework of the guiding data according to the type of the auxiliary guiding data. Among them, the rule-based guiding data is classified and organized according to the cost knowledge domain, the example-based guiding data is stored hierarchically according to the component type and project type, and the association-based guiding data is constructed in the form of a relationship graph; Determine the hierarchical structure of the auxiliary guiding data to establish the dependency relationship and call order between different levels of data; Extract the corresponding target data content from the initial project cost-related data set according to the determined guiding data type and the hierarchical structure of the auxiliary guiding data, and integrate the extracted target data content of different types according to the constructed structural framework to construct the auxiliary guiding data.

6. The method for dynamically annotating project cost data based on BIM technology according to any one of claims 1-5, characterized in that, The steps of determining the first confidence sequence and the second confidence sequence include: Based on the target BIM project cost dynamic data, extract cost component elements from the target BIM project cost dynamic data through a standardized deep learning network to generate a first basic confidence sequence associated with a first preset cost mapping knowledge set, and extract cost component elements from the target BIM project cost dynamic data through a customized deep learning network to generate a second basic confidence sequence associated with the first preset cost mapping knowledge set; Among them, the first basic confidence sequence includes the first basic confidence of each reference cost component element, and the second basic confidence includes the second basic confidence of each reference cost component element; According to the first basic confidence sequence and the second basic confidence sequence, extract the target reference component elements from each of the reference cost component elements, where the first basic confidence and the second basic confidence are both greater than or equal to a set confidence level. Respectively eliminate the confidence levels corresponding to the non-target reference component elements in the first basic confidence sequence and the second basic confidence sequence to generate the first confidence sequence and the second confidence sequence.

7. The method for dynamically annotating engineering cost data based on BIM technology according to claim 3, characterized in that, The target BIM project cost dynamic data includes multiple target cost component elements, and the annotation result includes multiple target cost component elements and target cost description labels of the target cost component elements. The annotation result is generated by enabling x rounds of deep learning processes. Each round of deep learning process is either a round of deep learning component extraction process or a round of deep learning component annotation process. The deep learning result of each round of deep learning process is a target cost component element or a target cost description label. Among them, the specific steps of the a-th round of deep learning process include: Determine the output category of the result to be output according to the deep learning result of the (a - 1)-th round of deep learning process. The output category is either a component extraction category or a component annotation category. The output category corresponding to the first round of deep learning process is the component extraction category. According to the target BIM project cost dynamic data and the deep learning results of the previous (a - 1) rounds of deep learning processes, obtain a standardized confidence sequence through a standardized deep learning network, obtain a customized confidence sequence through a customized deep learning network, and obtain a labeled confidence sequence through a labeling network. Among them, the standardized confidence sequence is the first confidence sequence or the third confidence sequence, the customized confidence sequence is the second confidence sequence or the sixth confidence sequence, and the labeled confidence sequence is the fourth confidence sequence or the fifth confidence sequence. Generate the deep learning result of the a-th round of deep learning process according to the standardized confidence sequence, the customized confidence sequence, and the labeled confidence sequence.

8. The method for dynamically annotating project cost data based on BIM technology according to any one of claims 3-5, characterized in that The training steps of the customized deep learning network include: Obtain a first sample learning data sequence, which includes multiple first template BIM project cost dynamic data carrying template annotation data. The template annotation data of each first template BIM project cost dynamic data includes the actual cost component elements of the first template BIM project cost dynamic data. According to each first template BIM project cost dynamic data, extract cost component elements from each first template BIM project cost dynamic data through a customized deep learning network with initialized parameters to generate the first estimated cost component elements of each first template BIM project cost dynamic data. Based on each of the first template BIM project cost dynamic data and the first estimated cost component elements, the first estimated cost component elements are labeled through a customized deep learning network with initialized parameters to generate the first estimated cost description labels of the first estimated cost component elements. Based on the first estimated cost component elements and the first estimated cost description labels of each of the first template BIM project cost dynamic data, the annotation results of each of the first template BIM project cost dynamic data are generated; Based on the feature distances between the first estimated cost component elements and the actual cost component elements of each of the first template BIM project cost dynamic data, the first network learning error is determined. Based on the first network learning error, the customized deep learning network with initialized parameters is trained; The training steps of the annotation network include: Obtain a second sample learning data sequence, where the second sample learning data sequence includes multiple second template BIM project cost dynamic data carrying template annotation data. The template annotation data of each second template BIM project cost dynamic data includes the actual cost description labels of the actual cost component elements of the second template BIM project cost dynamic data; Based on each of the second template BIM project cost dynamic data, the cost component elements of each of the second template BIM project cost dynamic data are extracted through an annotation network with initialized parameters to generate the second estimated cost component elements of each of the second template BIM project cost dynamic data; Based on each of the second template BIM project cost dynamic data and the second estimated cost component elements, the second estimated cost component elements are labeled through an annotation network with initialized parameters to generate the second estimated cost description labels of the second estimated cost component elements. Based on the second estimated cost component elements and the second estimated cost description labels of each of the second template BIM project cost dynamic data, the annotation results of each of the second template BIM project cost dynamic data are generated; Based on the feature distances between the second estimated cost description labels and the actual cost description labels of each of the second template BIM project cost dynamic data, the second network learning error is determined. Based on the second network learning error, the annotation network with initialized parameters is trained; The training steps of the standardized deep learning network include: Obtain a third sample learning data sequence, where the third sample learning data sequence includes multiple third template BIM project cost dynamic data carrying template annotation data. The template annotation data of each third template BIM project cost dynamic data includes the actual cost component elements of the third template BIM project cost dynamic data and the actual cost description labels of the actual cost component elements; Based on each of the third template BIM project cost dynamic data, the cost component elements of each of the third template BIM project cost dynamic data are extracted through a standardized deep learning network with initialized parameters to generate the third estimated cost component elements of each of the third template BIM project cost dynamic data; According to each of the third template BIM project cost dynamic data and the third estimated cost component elements, the third estimated cost component elements are labeled through a standardized deep learning network with initialized parameters to generate third estimated cost description labels for the third estimated cost component elements. According to the third estimated cost component elements and the third estimated cost description labels of each of the third template BIM project cost dynamic data, the annotation results of each of the third template BIM project cost dynamic data are generated; According to the feature distances between the third annotation results of each of the third template BIM project cost dynamic data and the template annotation data, the third network learning error is determined, and the standardized deep learning network with initialized parameters is trained according to the third network learning error.

9. A dynamic data annotation system for project cost based on BIM technology, characterized in that, The BIM technology-based project cost dynamic data annotation system includes a processor and a memory. The memory is connected to the processor. The memory is used to store programs, instructions or codes, and the processor is used to execute the programs, instructions or codes in the memory to implement the BIM technology-based project cost dynamic data annotation method described in any one of claims 1-8 above.

Images