AI-driven green building construction drawing review and evaluation method and system

Abstract

The present application provides an AI-driven green building construction drawing review and evaluation method and system, which relates to the field of construction drawing review technology, including: obtaining digital information of target green building construction drawings; obtaining predetermined review and evaluation dimensions, and extracting a first evaluation dimension; matching a first related indicator set in multi-dimensional indicators, and matching a first indicator parameter set in multi-indicator data; performing transfer learning on the initial AI review and evaluation model; evaluating and analyzing the first indicator parameter set through the target AI review and evaluation model to obtain a first evaluation result; and forming a target review and evaluation result. The present application can solve the technical problem in the prior art that the accuracy of construction drawing review and evaluation is insufficient due to the lack of dynamic adjustment capabilities of evaluation standards and technical indicators for specific green building projects. Through transfer learning, the specific dimensional requirements of specific projects are adapted, thereby improving the accuracy of green building construction drawing review and evaluation.

Description

Technical Field

[0001] The present application relates to the technical field of construction drawing review, and in particular to an AI-driven green building construction drawing review and evaluation method and system. Background Art

[0002] In the construction process of green building projects, construction drawing review is one of the key links. Its purpose is to ensure that the construction drawings comply with green building standards and relevant technical specifications, and promote the realization of building energy conservation, water conservation, material conservation, environmental protection and other goals. At present, the green building construction drawing review and evaluation methods are mainly based on the design indicators and technical requirements of green buildings, and usually follow the green building evaluation standards, including technical requirements for building energy conservation, water conservation, material conservation, environmental protection, indoor environmental quality, etc. Although the existing green building construction drawing review and evaluation methods have formed a certain system and can effectively evaluate most green building projects, there are still some technical problems. The existing green building evaluation standards and technical indicators are mostly universal and difficult to flexibly adapt to the specific needs of different projects. In particular, with the continuous innovation of green building technology and the diversification of architectural design, static evaluation methods often cannot fully consider the specific conditions of the project, such as the climate characteristics of the building location, design innovation, and the application of green technology. Therefore, the accuracy and adaptability of the construction drawing review and evaluation are limited, and it is difficult to meet the personalized and diversified green building design requirements.

[0003] In summary, there are technical problems in the existing technology that the accuracy of green building construction drawing review and evaluation is insufficient due to the lack of dynamic adjustment capabilities of evaluation standards and technical indicators for specific green building projects. Summary of the invention

[0004] The purpose of this application is to provide an AI-driven green building construction drawing review and evaluation method and system to solve the technical problem in the prior art that the green building construction drawing review and evaluation lacks accuracy due to the lack of dynamic adjustment capabilities of evaluation standards and technical indicators for specific green building projects.

[0005] In view of the above problems, this application provides an AI-driven green building construction drawing review and evaluation method and system.

[0006] In a first aspect, the present application provides an AI-driven green building construction drawing review and evaluation method, which is implemented by an AI-driven green building construction drawing review and evaluation system, wherein the AI-driven green building construction drawing review and evaluation method includes: obtaining target digital information of a target green building construction drawing, wherein the target digital information includes multi-indicator data corresponding to multi-dimensional indicators; obtaining a predetermined review and evaluation dimension, and extracting a first evaluation dimension from the predetermined review and evaluation dimension; matching a first related indicator set corresponding to the first evaluation dimension in the multi-dimensional indicator, and matching a first indicator parameter set of the first related indicator set in the multi-indicator data; performing transfer learning on an initial AI review and evaluation model according to the multi-indicator data to obtain a target AI review and evaluation model; evaluating and analyzing the first indicator parameter set through the target AI review and evaluation model to obtain a first evaluation result of the target green building construction drawing in the first evaluation dimension; and forming a target review and evaluation result of the target green building construction drawing based on the first evaluation result.

[0007] In the second aspect, the present application also provides an AI-driven green building construction drawing review and evaluation system for executing the AI-driven green building construction drawing review and evaluation method as described in the first aspect, wherein the AI-driven green building construction drawing review and evaluation system includes: a target information acquisition module, the target information acquisition module is used to acquire target digital information of the target green building construction drawing, the target digital information includes multi-indicator data corresponding to the multi-dimensional indicators; an evaluation dimension acquisition module, the evaluation dimension acquisition module is used to acquire predetermined audit and evaluation dimensions, and extract the first evaluation dimension from the predetermined audit and evaluation dimensions; an indicator matching module, the indicator matching module is used to match the target digital information in the multi-dimensional indicators. A first relevant indicator set corresponding to the first evaluation dimension, and matching the first indicator parameter set of the first relevant indicator set in the multi-indicator data; a transfer learning module, the transfer learning module is used to perform transfer learning on the initial AI audit and evaluation model according to the multi-indicator data, and obtain a target AI audit and evaluation model; an evaluation and analysis module, the evaluation and analysis module is used to evaluate and analyze the first indicator parameter set through the target AI audit and evaluation model, and obtain a first evaluation result of the target green building construction drawing in the first evaluation dimension; an evaluation result assembly module, the evaluation result assembly module is used to assemble a target audit and evaluation result of the target green building construction drawing based on the first evaluation result.

[0008] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0009] By obtaining the target digital information of the target green building construction drawing, the target digital information includes multi-indicator data corresponding to the multi-dimensional indicators; obtaining the predetermined audit and evaluation dimension, and extracting the first evaluation dimension in the predetermined audit and evaluation dimension; matching the first relevant indicator set corresponding to the first evaluation dimension in the multi-dimensional indicators, and matching the first indicator parameter set of the first relevant indicator set in the multi-indicator data; performing transfer learning on the initial AI audit and evaluation model according to the multi-indicator data to obtain the target AI audit and evaluation model; evaluating and analyzing the first indicator parameter set through the target AI audit and evaluation model to obtain the first evaluation result of the target green building construction drawing in the first evaluation dimension; and forming the target audit and evaluation result of the target green building construction drawing based on the first evaluation result. That is to say, by collecting the multi-dimensional indicator data of the construction drawing, selecting the evaluation dimension, finding the relevant indicator set from the multi-dimensional indicators, and extracting the corresponding indicator parameter set from the multi-indicator data, performing transfer learning on the initial AI audit and evaluation model according to the multi-indicator data, adjusting the model according to the digital information of the target green building project and the specific evaluation dimension, so as to adapt it to the special needs of different projects, realize targeted evaluation, avoid the limitations of fixed standards, and improve the accuracy of the audit and evaluation of green building construction drawings.

[0010] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented according to the contents of the specification, and in order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are specifically cited below. It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become easy to understand through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In order to more clearly illustrate the technical solutions in the present application or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only exemplary, and for ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0012] Figure 1 A flowchart of the AI-driven green building construction drawing review and evaluation method for this application;

[0013] Figure 2 This is a structural diagram of the AI-driven green building construction drawing review and evaluation system for this application.

[0014] Explanation of the accompanying drawings: target information acquisition module 11, evaluation dimension acquisition module 12, indicator matching module 13, transfer learning module 14, evaluation analysis module 15, evaluation result assembly module 16. DETAILED DESCRIPTION

[0015] This application solves the technical problem in the prior art that the accuracy of green building construction drawing review and evaluation is insufficient due to the lack of dynamic adjustment capabilities of evaluation standards and technical indicators for specific green building projects, by providing an AI-driven green building construction drawing review and evaluation method and system. By collecting multi-dimensional indicator data of construction drawings, selecting evaluation dimensions, finding relevant indicator sets from multi-dimensional indicators, and extracting corresponding indicator parameter sets from multi-indicator data, the initial AI review and evaluation model is transferred and learned based on the multi-indicator data, and the model is adjusted according to the digital information of the target green building project and the specific evaluation dimensions to adapt it to the special needs of different projects, so as to achieve targeted evaluation, avoid the limitations of fixed standards, and improve the accuracy of green building construction drawing review and evaluation.

[0016] Below, the technical solutions in the present application will be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments of the present application. It should be understood that the present application is not limited to the example embodiments described herein. Based on the embodiments of the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present application. It should also be noted that, for the convenience of description, only the parts related to the present application are shown in the accompanying drawings, rather than all of them.

[0017] For example, please refer to the attached Figure 1 The present application provides an AI-driven green building construction drawing review and evaluation method, wherein the AI-driven green building construction drawing review and evaluation method is applied to an AI-driven green building construction drawing review and evaluation system, and the AI-driven green building construction drawing review and evaluation method specifically includes the following steps:

[0018] S100: Acquire target digital information of a target green building construction drawing, wherein the target digital information includes multi-indicator data corresponding to multi-dimensional indicators.

[0019] Specifically, the target digital information of the target green building construction drawing is obtained, that is, all relevant information in the green building construction drawing is converted into a digital format. The relevant data of the green building construction drawing is collected, including multi-dimensional information such as building shape and orientation, enclosure structure performance, window-to-wall ratio, shading and insulation measures, site design, greening and vegetation, wind environment design, indoor air quality, acoustic environment design, building material selection, construction material management, and barrier-free design. For example, the south-facing window-to-wall ratio is 0.42, the east-facing window-to-wall ratio is 0.38, and the greening rate is 28%. Multi-dimensional indicators refer to multiple technical and environmental indicators in different fields or dimensions that are concerned in green building assessments. Multi-indicator data refers to data sets related to various designs and performances of green buildings, which contain digital data collected from multiple dimensions, that is, specific values ​​or parameters corresponding to indicators in each dimension.

[0020] The multi-dimensional information of the target green building construction drawing is usually extracted from the construction drawings and design documents of the green building project. It is digitized through tools such as BIM (Building Information Modeling) and CAD (Computer-Aided Design) software to convert the original paper or graphic data into operable digital information, which is convenient for subsequent data analysis and evaluation. After the data conversion is completed, further data cleaning, format conversion and feature extraction are required. Multi-index data refers to the data obtained after operations such as data cleaning, format conversion, feature extraction, etc., to remove duplicate or invalid data, ensure the accuracy of the final data, and convert the data into a unified standard format. All green building design data is converted into digital information in a standard format through digital tools and classified according to each green building evaluation dimension, including specific data of multiple dimensions, such as the building's envelope performance (thermal performance, seismic resistance, etc.), window-to-wall ratio, building energy-saving performance, green material selection, etc. The multi-dimensional information of the target green building construction drawing is converted through digital tools to ensure the accuracy, operability and standardization of the construction drawing design data, and to comprehensively and accurately present all dimensions of the building design, which not only improves the availability of data, but also reduces the evaluation risk caused by data errors.

[0021] S200: Obtain predetermined audit evaluation dimensions, and extract a first evaluation dimension from the predetermined audit evaluation dimensions.

[0022] Further, the present application S200 includes:

[0023] Establish an energy-saving audit and evaluation dimension set; establish a water-saving audit and evaluation dimension set; establish an environmental protection audit and evaluation dimension set; establish a comfort audit and evaluation dimension set; based on the energy-saving audit and evaluation dimension set, the water-saving audit and evaluation dimension set, the environmental protection audit and evaluation dimension set and the comfort audit and evaluation dimension set, establish the predetermined audit and evaluation dimensions; wherein the energy-saving audit and evaluation dimension set includes energy-saving control item evaluation indicators, energy-saving conventional item evaluation indicators and energy-saving scoring item evaluation indicators, the water-saving audit and evaluation dimension set includes water-saving control item evaluation indicators, water-saving conventional item evaluation indicators and water-saving scoring item evaluation indicators, the environmental protection audit and evaluation dimension set includes environmental protection control item evaluation indicators, environmental protection conventional item evaluation indicators and environmental protection scoring item evaluation indicators, and the comfort audit and evaluation dimension set includes comfort control item evaluation indicators, comfort conventional item evaluation indicators and comfort scoring item evaluation indicators.

[0024] Specifically, identify all dimensional indicators related to energy conservation and form an energy conservation review and evaluation dimensional set, including energy conservation control item evaluation indicators, energy conservation conventional item evaluation indicators and energy conservation scoring item evaluation indicators. Energy conservation control item evaluation indicators refer to design elements that directly control building energy efficiency (such as window sealing and thermal conductivity of enclosure structure); energy conservation conventional item evaluation indicators refer to conventional energy conservation measures in building design (such as natural ventilation design and shading measures); energy conservation scoring item evaluation indicators refer to the quantitative evaluation of energy conservation design results (such as the overall energy efficiency score of the building).

[0025] Identify the main design elements that affect the use of building water resources, and form a water-saving review and evaluation dimension set to evaluate whether the building effectively saves water resources, including water-saving control item evaluation indicators, water-saving conventional item evaluation indicators and water-saving scoring item evaluation indicators. Water-saving control item evaluation indicators refer to design measures that directly affect the use of building water resources (such as water pipe layout, water recycling system, etc.); water-saving conventional item evaluation indicators refer to common water-saving equipment or technologies (such as water-saving faucets, low-flow flushing equipment, etc.); water-saving scoring item evaluation indicators are quantitative evaluations based on the effectiveness of water-saving design (such as building water consumption, water resource recycling rate, etc.).

[0026] Identify dimensions related to building environmental protection, such as building material selection, waste management, energy recycling, etc., and form an environmental audit and evaluation dimension set, including environmental control item evaluation indicators, environmental conventional item evaluation indicators, and environmental scoring item evaluation indicators. Environmental control item evaluation indicators refer to specific control items of environmental design and technology (such as environmental certification of building materials, waste treatment systems, etc.); environmental conventional item evaluation indicators refer to conventional environmental protection measures (such as low pollution emission systems, energy-saving and environmental protection facilities, etc.); environmental scoring item evaluation indicators include scoring or quantitative evaluation of the effectiveness of environmental protection measures.

[0027] Identify the design elements that affect the comfort of building residents and users, and form a comfort audit assessment dimension set to ensure that the building provides a suitable living and working environment in design, such as temperature and humidity control, air quality, noise control, etc. The comfort audit assessment dimension set includes comfort control item assessment indicators, comfort general item assessment indicators and comfort score item assessment indicators. Among them, the comfort control item assessment indicators refer to the design and technology that directly affects the comfort of the building (such as ventilation system, sound insulation wall design, indoor temperature control); the comfort general item assessment indicators include common comfort design elements (such as lighting, air purification system, etc.); the comfort score item assessment indicators refer to the quantitative evaluation of the comfort design effect (such as indoor temperature and humidity, noise control effect score, etc.).

[0028] The evaluation results of the four dimensions of energy saving, water saving, environmental protection and comfort are integrated to establish a comprehensive predetermined review and evaluation dimension, including all relevant control items, conventional items and scoring items, to comprehensively evaluate the greenness and sustainability of building design. The predetermined review and evaluation dimension refers to a multi-dimensional evaluation framework set in advance based on the green building evaluation goals, which integrates the energy saving, water saving, environmental protection, comfort and other aspects of green buildings as the standard and basis for evaluating green building construction drawings. Any specific dimension is extracted from the predetermined review and evaluation dimension as the first evaluation dimension for in-depth analysis.

[0029] By forming a review and evaluation dimension set of four dimensions, namely energy saving, water saving, environmental protection and comfort, we can comprehensively evaluate various indicators of green building design, which can not only effectively ensure that the design of the building meets the standards in terms of energy saving, water saving, environmental protection and comfort, but also provide data support and improvement direction for building design through the application of quantitative evaluation items. By extracting the first evaluation dimension, we can conduct an in-depth analysis of the core goals of the project and improve the accuracy and pertinence of the evaluation results.

[0030] S300: Matching a first related indicator set corresponding to the first evaluation dimension in the multi-dimensional indicators, and matching a first indicator parameter set of the first related indicator set in the multi-indicator data.

[0031] Specifically, among the multi-dimensional indicators, a group of related indicators corresponding to the first evaluation dimension is selected, which is called the first related indicator set. For example, if the first evaluation dimension is energy saving, identify and extract all indicators related to energy saving, such as energy consumption, thermal insulation performance, etc. According to the first related indicator set, the corresponding actual parameter data is matched from the multi-indicator data, that is, the first indicator parameter set. By matching the first evaluation dimension related indicator set from the multi-dimensional indicators, the key indicators related to the specific dimension can be quickly screened out, reducing the complexity of data processing.

[0032] S400: Perform transfer learning on the initial AI audit and evaluation model according to the multi-indicator data to obtain a target AI audit and evaluation model.

[0033] Specifically, an audit and evaluation task of a green building construction drawing of the same type as the target green building construction drawing is obtained as a source audit and evaluation task. A source data set is formed based on the multi-indicator data and the audit and evaluation results of the source audit and evaluation task, and an initial AI audit and evaluation model is constructed in combination with predetermined audit and evaluation rules. The first evaluation indicator and the second evaluation indicator in the first evaluation dimension are extracted, and the relevant indicator sets of the two evaluation indicators are obtained to obtain a first general indicator set, which includes all relevant data of the two evaluation indicators of the first evaluation dimension. Similarly, evaluation dimensions other than the first evaluation dimension are randomly extracted from the predetermined audit and evaluation dimensions as the second evaluation dimension, and a corresponding second general indicator set is also obtained, which includes all relevant data of the two evaluation indicators of the second evaluation dimension.

[0034] According to the first general indicator set and the second general indicator set, the corresponding first general indicator data set and the second general indicator data set are matched in the multi-indicator data. According to the initial AI audit evaluation model, parallel transfer learning is performed on the first general indicator data set and the second general indicator data set, and multiple related tasks are learned simultaneously in the source domain and the target domain, which helps the model learn more general representations and avoid negative transfer. Use the feature extraction layer that has been trained in the initial model to extract the common features of the target data. Map the characteristics of the target data to the output layer of the model, and improve adaptability through retraining. Design a transfer learning loss function based on the target task requirements, and use the validation set to evaluate the model after transfer learning to ensure that its performance meets the requirements. Through transfer learning, the target AI audit evaluation model can adapt to the target project data more accurately and quickly achieve better performance. Transfer learning can help the model achieve higher prediction accuracy in new fields.

[0035] S500: Evaluate and analyze the first indicator parameter set through the target AI audit and evaluation model to obtain a first evaluation result of the target green building construction drawing in the first evaluation dimension.

[0036] Specifically, the first indicator parameter set is a data set related to the first evaluation dimension, which contains all necessary parameters for evaluating the performance of the target green building. The first indicator parameter set is preprocessed, including data cleaning, normalization or standardization. The first indicator parameter set is evaluated and analyzed according to the target AI audit and evaluation model, and the preprocessed first indicator parameter set is input into the target AI audit and evaluation model to ensure that the format of the data is consistent with the format used during model training. The target AI audit and evaluation model is calculated based on the input first indicator parameter set, and the target AI audit and evaluation model obtains the result of the first evaluation dimension through the learned feature mapping and weight calculation. Through the evaluation and analysis, the first evaluation result of the target green building construction drawing in the first evaluation dimension is obtained, that is, the performance evaluation result of the target green building construction drawing in the first evaluation dimension, including a score, grade or specific evaluation report. The first evaluation result obtained by evaluating and analyzing the first indicator parameter set through the target AI audit and evaluation model can provide a quantitative evaluation of the performance of the target green building construction drawing in a specific evaluation dimension, and corresponding optimization and adjustment are performed according to the evaluation results, thereby improving the efficiency and accuracy of the evaluation.

[0037] S600: Based on the first evaluation result, a target review and evaluation result of the target green building construction drawing is established.

[0038] Specifically, according to a similar process of obtaining the first evaluation result, the above steps are also performed for other evaluation dimensions in the predetermined review and evaluation dimensions to obtain the evaluation results corresponding to each evaluation dimension, that is, multiple evaluation results. The evaluation results of the first evaluation dimension and other dimensions are combined to form a comprehensive evaluation result, that is, the target review and evaluation result of the target green building construction drawing, including a comprehensive evaluation report of all relevant evaluation dimension results, providing a comprehensive evaluation of the overall performance of the target green building construction drawing. By integrating the evaluation results of multiple dimensions, the one-sidedness that may be caused by single-dimensional analysis is avoided, and not only a quantitative evaluation of the performance of the building in each evaluation dimension is provided, but also a comprehensive evaluation of the overall performance is provided.

[0039] Further, the present application S400 includes:

[0040] Acquire source digital information of a source audit and evaluation task, wherein the source audit and evaluation task refers to an audit and evaluation task of a green building construction drawing of the same type as the target green building construction drawing; construct a source data set based on the preprocessed source multi-indicator data of the source digital information and the source audit and evaluation results; acquire predetermined audit and evaluation rules, and construct the initial AI audit and evaluation model in combination with the predetermined audit and evaluation rules and the source data set.

[0041] Specifically, the audit and evaluation tasks of green building construction drawings of the same type as the target green building construction drawings completed in previous projects are obtained, and the relevant digital information is extracted from the completed source audit and evaluation tasks to obtain the source digital information. The source digital information refers to the digital data of the construction drawings involved in the source audit and evaluation tasks, covering multi-dimensional indicators and corresponding specific parameters.

[0042] Data preprocessing is performed on the source digitized information, including data cleaning, data standardization, format conversion, and feature extraction, to ensure data quality and consistency. After preprocessing the source digitized information, source multi-indicator data and source audit evaluation results are obtained. Source multi-indicator data refers to specific indicator data extracted and preprocessed from the source digitized information, and source audit evaluation results refer to the evaluation conclusions or scoring results for these indicators in the source task. The source multi-indicator data is combined with the source audit evaluation results to form a source data set, which is used as training data to build the initial AI audit evaluation model.

[0043] Obtain the predetermined review and evaluation rules, i.e., the specific requirements of the predefined construction drawing review and evaluation tasks, including the standardized content of control items, routine items, and scoring items, which define the standards for how to extract features from input data, how to process these features, and how to evaluate building performance. Use the predetermined review and evaluation rules as feature constraints and combine them with the source data set to train the initial AI review and evaluation model. The initial AI review and evaluation model can predict the evaluation results of different construction drawings by learning the relationship between rule constraints and historical evaluation data. Extract relevant features from the source data set according to the predetermined review and evaluation rules, including selecting indicators directly related to green building performance, creating new derived features, or using feature selection techniques to identify the most important features. Select a suitable machine learning algorithm to build the model, such as decision trees, random forests, support vector machines, neural networks, etc.

[0044] Use the training set data to train the selected model, and the model continuously learns how to predict the audit and evaluation results based on the input feature data. Use the test set data to verify the performance of the model. Evaluate the effect of the model based on indicators such as the model's accuracy and mean square error. The accuracy refers to the proportion of the number of positive and negative samples correctly predicted by the model to the total number of samples, and the mean square error refers to the average of the square of the error between the predicted value and the true value. If the performance does not meet the requirements, adjust the parameters of the model or select a different algorithm to retrain. Once the model has been verified and meets the performance requirements, it is deployed in the actual application environment for the audit and evaluation of new green building construction drawings. By using the historical data of the audit and evaluation tasks, fully tapping the experience of previous tasks, combined with the predetermined audit rules, accurately capture the standard requirements of green building evaluation, ensure the consistency of the evaluation results with industry specifications, accurately evaluate the advantages and disadvantages of the construction drawing design, and reduce the deviation and time consumption of manual audits.

[0045] Furthermore, the present application also includes the following steps:

[0046] Extract the first evaluation indicator and the second evaluation indicator in the first evaluation dimension in sequence; obtain the union of the first related indicator set corresponding to the first evaluation indicator and the second related indicator set of the second evaluation indicator, recorded as the first general indicator set; obtain the second general indicator set of the second evaluation dimension, the second evaluation dimension refers to any dimension in the predetermined audit evaluation dimension that is different from the first evaluation dimension; match the first general indicator data set corresponding to the first general indicator set and the second general indicator data set corresponding to the second general indicator set in the multi-indicator data in sequence; perform parallel transfer learning on the first general indicator data set and the second general indicator data set based on the initial AI audit evaluation model to obtain the target AI audit evaluation model.

[0047] Specifically, a first evaluation indicator and a second evaluation indicator are extracted from the first evaluation dimension in sequence, and the two evaluation indicators are adjacent and are specific evaluation indicators directly related to the first evaluation dimension. A set of related indicators for each evaluation indicator is identified, namely, a first related indicator subset corresponding to the first evaluation indicator and a second related indicator subset corresponding to the second evaluation indicator. According to the union of the first related indicator subset and the second related indicator subset, a first general indicator set is formed, which includes all relevant data of the two evaluation indicators.

[0048] From the predetermined audit evaluation dimensions, any evaluation dimension other than the first evaluation dimension is extracted as the second evaluation dimension. As in the previous steps, a second general indicator set of the second evaluation dimension is obtained, which includes all relevant data of the two evaluation indicators in the second evaluation dimension. According to the first general indicator set and the second general indicator set, the corresponding values ​​are matched in the multi-indicator data to obtain the first general indicator data set and the second general indicator data set.

[0049] Migrate the initial AI audit evaluation model to the target task, and perform parallel transfer learning on the first general indicator dataset and the second general indicator dataset of the target domain. In the process of transfer learning, first freeze some general layers, keep the core layer of the source task in the initial model unchanged, freeze the underlying weights related to the source task in the initial model (such as convolutional layer, Transformer layer), and retain the general features extracted. Only train the weights of the task-specific part, and transfer the training of the top-level network specific to the target task. Add two independent task-specific output layers, one for the first evaluation dimension and the other for the second evaluation dimension, and each task-specific layer outputs a separate prediction result.

[0050] Define the loss function for each task (such as mean squared error MSE or cross entropy CE), and combine the losses of multiple tasks in a weighted form. Use an optimizer (such as Adam or SGD) to update the weights of the model. In the training cycle, the first general indicator dataset and the second general indicator dataset are updated alternately to ensure that the weight of each task is optimized. Through parallel transfer learning, the target AI audit evaluation model can consider two dimensions of data at the same time, learn more general feature representations, and adapt to the target domain.

[0051] Transfer learning refers to applying the model trained for the source task (initial AI audit evaluation model) to the target task, and improving the model's adaptability to new tasks by learning the target domain data in parallel. Parallel transfer learning refers to migrating the training model on two or more data sets (such as the first general indicator data set and the second general indicator data set) at the same time, so that the model can adapt to audit tasks of different dimensions, which helps the model learn more general representations and avoid negative transfer. Negative transfer refers to the fact that during the transfer learning process, the model introduces irrelevant characteristics of the source domain into the target domain task, resulting in a decrease in model performance. This problem can be avoided through multi-task parallel learning. By extracting relevant data that meets the evaluation dimension from multi-indicator data, ensuring the accuracy and completeness of the input data, and improving the training effect of the model, parallel transfer learning enables the model to learn the characteristics of multi-dimensional audit tasks at the same time, avoiding the one-sidedness of single-dimensional training.

[0052] Furthermore, the present application also includes the following steps:

[0053] Introduce a knowledge distillation dimensionality reduction function to perform distillation analysis on the source multi-indicator data to obtain source multi-indicator dimensionality reduction data; perform machine learning on the dimensionality reduction source data set constructed based on the source multi-indicator dimensionality reduction data and the source audit and evaluation results to obtain an audit and evaluation prediction model; integrate the audit and evaluation prediction model into the initial AI audit and evaluation model.

[0054] Specifically, the knowledge distillation dimensionality reduction function is used to process the source multi-index data to reduce the data dimension while retaining the key information of the original data as much as possible. The goal of the dimensionality reduction task is to map high-dimensional data to a low-dimensional space, extract key features, eliminate redundant information, and retain the key information and complex relationships of the original data as much as possible. Knowledge distillation is a model compression method. Its core idea is to guide a smaller and simpler model (student model) to learn through a complex model with higher performance (teacher model), so that the student model can achieve higher efficiency while retaining the performance of the teacher model.

[0055] In the dimensionality reduction task, knowledge distillation maps high-dimensional data to low-dimensional data space through a distillation dimensionality reduction function, extracts key features, and retains important information related to the original data. For example, the source multi-indicator data set contains 5,000 records, each with 100 features, and the distillation temperature is set to 5 (moderate smoothness) to ensure moderate smoothness. Finally, the dimensionality reduction process simplifies the original 100-dimensional data into 10 key features, and the divergence between the reduced data and the original data is reduced to 0.02, and the performance of the reduced model in the prediction task remains at 95% accuracy. Source multi-indicator data refers to multi-indicator data obtained from source audit and evaluation tasks, such as relevant indicators of green building construction drawings (window-to-wall ratio, energy consumption level, air quality, etc.); source multi-indicator dimensionality reduction data refers to source multi-indicator data processed by the knowledge distillation dimensionality reduction function, which has a lower dimension but retains key features and information structure.

[0056] The specific implementation steps of knowledge distillation are to utilize the collaboration between the teacher model and the student model. The teacher model is a model that has been trained in a complex way and has high performance. Its role is to learn complex feature relationships from high-dimensional data, predict the input high-dimensional data, and generate a probability distribution for each sample, called a soft label. The student model is a model with a simpler structure and fewer parameters. By learning the soft labels and original data of the teacher model, it can achieve the performance of the teacher model and reduce the data dimension. The temperature parameter is to smooth the probability distribution generated by the teacher model so that the student model can better learn the details of the teacher model. The teacher model is trained with high-dimensional data (such as multi-dimensional indicators of green building construction drawings) to enable it to capture complex feature relationships. The loss function consists of two parts, including distillation loss and reconstruction loss. The distillation loss is used to measure the similarity between the output of the student model and the soft label of the teacher model, and the reconstruction loss is used to measure the difference between the data after dimensionality reduction and the original data. The student model trained by distillation maps high-dimensional data to low-dimensional data space to obtain reduced dimensionality data, that is, source multi-indicator dimensionality reduction data.

[0057] Combine the reduced source multi-indicator dimensionality reduction data with the source audit assessment results to form a reduced source data set. Combine the source multi-indicator dimensionality reduction data (input features) and the source audit assessment results (labels) to generate a reduced source data set. According to the reduced source data set, select a suitable machine learning model (such as random forest, gradient boosting tree). For classification problems, select random forest, support vector machine, neural network, etc.; for regression problems, select linear regression, decision tree regression, etc. Set the parameters of the model, such as the number of layers and nodes of the neural network, the number of trees of the random forest, etc.

[0058] Divide the reduced-dimensionality source data set into a training set and a validation set. The training set is used to train the model, and the validation set is used to evaluate the performance of the model. Use the training set data to train the selected machine learning model. During the training process, the model will learn the relationship between the input features (the source multi-indicator data after dimensionality reduction) and the target variable (the source audit evaluation results). Use the test set to verify the performance of the model. The evaluation indicators may include accuracy, mean square error, etc. Accuracy refers to the proportion of positive and negative samples correctly predicted by the model to the total number of samples. The mean square error refers to the average of the squared error between the predicted value and the true value. According to the evaluation results of the validation set, adjust the parameters of the model to improve the performance of the model. Deploy the trained audit evaluation prediction model to the actual application environment to predict the audit evaluation results of new green building construction drawings.

[0059] Embed the audit and evaluation prediction model into the initial AI audit and evaluation model to achieve functional integration and performance optimization. Achieve efficient collaboration between the prediction model and the initial AI audit and evaluation model to improve overall performance. The output layer of the initial model combines the output of the prediction model to generate a more comprehensive evaluation result. Ensure the stability of the dimensionality reduction prediction model in the initial AI model, standardize the output results of different modules to make them comparable in the same evaluation dimension, and achieve joint optimization of different tasks by sharing the feature extraction part of the initial model. Reduce the dimensionality of the source multi-index data through the knowledge distillation dimensionality reduction function, reduce the dimensional complexity of model training, and improve the training speed. By integrating the dimensionality reduction prediction model into the initial AI audit and evaluation model, the flexibility and scalability of the model are enhanced, and it can be applied to more tasks.

[0060] Furthermore, the present application also includes the following steps:

[0061] The expression of knowledge distillation dimensionality reduction function is as follows: ;in, Characterize the distillation dimension reduction function, Characterize the source multi-index dimensionality reduction data Multiple indicator data with the source The divergence between Characterize the temperature coefficient, Characterizing the source multi-indicator data In The first dimension indicator indicator data.

[0062] Specifically, the knowledge distillation dimensionality reduction function is used to convert high-dimensional data into low-dimensional data in machine learning while retaining the key information in the data. The expression of the knowledge distillation dimensionality reduction function is: Among them, L kl Characterize the distillation dimension reduction function, Characterize the source multi-index dimension reduction data P T With the source multi-index data P S The closer the two are, the closer the divergence is to 0. The divergence is an indicator of the difference between two probability distributions, which measures the average amount of information required to transfer from one distribution to another. T It is the distribution data after dimensionality reduction, which is obtained by adjusting the temperature parameter t to the original data P S The result obtained after smoothing the information; S is the original distribution data, which represents the dimensional features and their weight distribution in the source multi-index data; t represents the temperature coefficient, which is used to control the smoothness of the output distribution. When t is higher, the output distribution is smoother (weakening the difference between large values), and when t is lower, the distribution is sharper (emphasizing the features with larger weights); i represents the source multi-index data P S The i-th indicator data among the n dimensional indicators in . Perform exponential operations on the features after dimensionality reduction to highlight significant features; All features are normalized to ensure that the output distribution sum is 1. The knowledge distillation dimensionality reduction function retains the key information of the original data, so that the data after dimensionality reduction can still meet the audit and evaluation requirements.

[0063] Furthermore, the present application also includes the following steps:

[0064] Rendering the target audit and evaluation results to the target green building construction drawing to obtain a target audit and evaluation visual graph; obtaining a target audit report of the target audit and evaluation visual graph, wherein the target audit report includes any audit-failed indicators; generating any rectification suggestions for the any audit-failed indicators in combination with the predetermined audit and evaluation rules; and rectifying the target green building construction drawing according to the any rectification suggestions.

[0065] Specifically, the target audit assessment results are associated with the target green building construction drawings, and visualization technology is used to highlight the target audit assessment results in the construction drawings in the form of colors, symbols or annotations on the corresponding building components. Use the rendering function of the building information modeling (BIM) software to generate a visual map containing the assessment results, so that you can intuitively see which parts have not passed the audit. Automatically summarize all failed indicators in the target audit assessment visual map, including the current value, qualified threshold and gap of each indicator. According to any audit failed indicators, a target audit report is formed, which lists the failed indicators and the corresponding assessment results in detail, and lists the analysis details and impact evaluation of each failed indicator in the form of a table or paragraph.

[0066] According to the failed indicators and the predetermined audit and evaluation rules, the applicable rectification plan is automatically matched, and specific rectification suggestions are proposed for the indicators that have not passed the audit in the report, and any rectification suggestions are generated. Specifically, for each failed indicator listed in the report, its reasons and potential impacts are analyzed; according to the predetermined audit and evaluation rules, specific rectification suggestions are formulated for each failed indicator, including specific rectification measures, possible resources, expected effects and implementation schedules. For example, the failed indicator shows that the energy saving dimension score is 60 points (window-to-wall ratio and enclosure structure are not up to standard); the comfort dimension score is 65 points (air quality is not up to standard). Based on this, the optimization suggestions given are: when the window-to-wall ratio is greater than 0.45, the external windows of the living space should be equipped with movable external shading devices to reduce the solar radiation heat entering the room in summer; the enclosure structure uses rock wool sandwich panels to reduce the wall heat transfer coefficient to 1.4W / (m²·K). Rock wool sandwich panels are a material with excellent thermal insulation properties that can effectively reduce the energy consumption of buildings; in terms of air purification, a fresh air system is added, and the air exchange volume per hour is ≥30m³ / person. The adjusted digital construction drawings passed the review and finally scored 90 points, meeting the green building standards.

[0067] According to any rectification suggestions for any audited indicators that failed, make necessary modifications and adjustments to the target green building construction drawings. Update the rectification measures to the target green building construction drawings to ensure that all relevant parties can see the latest design changes. After the rectification is completed, rerun the AI ​​audit and evaluation model to verify the effectiveness of the rectification measures. By combining predetermined audit and evaluation rules and automated tools, ensure that the modified construction drawings meet the green building design specifications.

[0068] In summary, the AI-driven green building construction drawing review and evaluation method provided in this application has the following technical effects:

[0069] By obtaining the target digital information of the target green building construction drawing, the target digital information includes multi-indicator data corresponding to the multi-dimensional indicators; obtaining the predetermined audit and evaluation dimension, and extracting the first evaluation dimension in the predetermined audit and evaluation dimension; matching the first relevant indicator set corresponding to the first evaluation dimension in the multi-dimensional indicators, and matching the first indicator parameter set of the first relevant indicator set in the multi-indicator data; performing transfer learning on the initial AI audit and evaluation model according to the multi-indicator data to obtain the target AI audit and evaluation model; evaluating and analyzing the first indicator parameter set through the target AI audit and evaluation model to obtain the first evaluation result of the target green building construction drawing in the first evaluation dimension; and forming the target audit and evaluation result of the target green building construction drawing based on the first evaluation result. That is to say, by collecting the multi-dimensional indicator data of the construction drawing, selecting the evaluation dimension, finding the relevant indicator set from the multi-dimensional indicators, and extracting the corresponding indicator parameter set from the multi-indicator data, performing transfer learning on the initial AI audit and evaluation model according to the multi-indicator data, adjusting the model according to the digital information of the target green building project and the specific evaluation dimension, so as to adapt it to the special needs of different projects, realize targeted evaluation, avoid the limitations of fixed standards, and improve the accuracy of the audit and evaluation of green building construction drawings.

[0070] Embodiment 2, based on the same inventive concept as the AI-driven green building construction drawing review and evaluation method in the aforementioned embodiment 1, this application also provides an AI-driven green building construction drawing review and evaluation system, please refer to the attached Figure 2 The AI-driven green building construction drawing review and evaluation system includes:

[0071] A target information acquisition module 11, the target information acquisition module 11 is used to acquire the target digital information of the target green building construction drawing, and the target digital information includes multi-indicator data corresponding to the multi-dimensional indicators; an evaluation dimension acquisition module 12, the evaluation dimension acquisition module 12 is used to acquire a predetermined audit evaluation dimension, and extract a first evaluation dimension in the predetermined audit evaluation dimension; an indicator matching module 13, the indicator matching module 13 is used to match a first related indicator set corresponding to the first evaluation dimension in the multi-dimensional indicators, and match a first indicator parameter set of the first related indicator set in the multi-indicator data; a transfer learning module 14, the transfer learning module 14 is used to perform transfer learning on the initial AI audit evaluation model according to the multi-indicator data to obtain a target AI audit evaluation model; an evaluation analysis module 15, the evaluation analysis module 15 is used to evaluate and analyze the first indicator parameter set through the target AI audit evaluation model to obtain a first evaluation result of the target green building construction drawing in the first evaluation dimension; an evaluation result assembly module 16, the evaluation result assembly module 16 is used to assemble a target audit evaluation result of the target green building construction drawing based on the first evaluation result.

[0072] Furthermore, the evaluation dimension acquisition module 12 in the AI-driven green building construction drawing review and evaluation system is also used for:

[0073] Establish an energy-saving audit and evaluation dimension set; establish a water-saving audit and evaluation dimension set; establish an environmental protection audit and evaluation dimension set; establish a comfort audit and evaluation dimension set; based on the energy-saving audit and evaluation dimension set, the water-saving audit and evaluation dimension set, the environmental protection audit and evaluation dimension set and the comfort audit and evaluation dimension set, establish the predetermined audit and evaluation dimensions; wherein the energy-saving audit and evaluation dimension set includes energy-saving control item evaluation indicators, energy-saving conventional item evaluation indicators and energy-saving scoring item evaluation indicators, the water-saving audit and evaluation dimension set includes water-saving control item evaluation indicators, water-saving conventional item evaluation indicators and water-saving scoring item evaluation indicators, the environmental protection audit and evaluation dimension set includes environmental protection control item evaluation indicators, environmental protection conventional item evaluation indicators and environmental protection scoring item evaluation indicators, and the comfort audit and evaluation dimension set includes comfort control item evaluation indicators, comfort conventional item evaluation indicators and comfort scoring item evaluation indicators.

[0074] Furthermore, the transfer learning module 14 in the AI-driven green building construction drawing review and evaluation system is also used for:

[0075] Acquire source digital information of a source audit and evaluation task, wherein the source audit and evaluation task refers to an audit and evaluation task of a green building construction drawing of the same type as the target green building construction drawing; construct a source data set based on the preprocessed source multi-indicator data of the source digital information and the source audit and evaluation results; acquire predetermined audit and evaluation rules, and construct the initial AI audit and evaluation model in combination with the predetermined audit and evaluation rules and the source data set.

[0076] Furthermore, the transfer learning module 14 in the AI-driven green building construction drawing review and evaluation system is also used for:

[0077] Extract the first evaluation indicator and the second evaluation indicator in the first evaluation dimension in sequence; obtain the union of the first related indicator set corresponding to the first evaluation indicator and the second related indicator set of the second evaluation indicator, recorded as the first general indicator set; obtain the second general indicator set of the second evaluation dimension, the second evaluation dimension refers to any dimension in the predetermined audit evaluation dimension that is different from the first evaluation dimension; match the first general indicator data set corresponding to the first general indicator set and the second general indicator data set corresponding to the second general indicator set in the multi-indicator data in sequence; perform parallel transfer learning on the first general indicator data set and the second general indicator data set based on the initial AI audit evaluation model to obtain the target AI audit evaluation model.

[0078] Furthermore, the transfer learning module 14 in the AI-driven green building construction drawing review and evaluation system is also used for:

[0079] Introduce a knowledge distillation dimensionality reduction function to perform distillation analysis on the source multi-indicator data to obtain source multi-indicator dimensionality reduction data; perform machine learning on the dimensionality reduction source data set constructed based on the source multi-indicator dimensionality reduction data and the source audit and evaluation results to obtain an audit and evaluation prediction model; integrate the audit and evaluation prediction model into the initial AI audit and evaluation model.

[0080] Furthermore, the transfer learning module 14 in the AI-driven green building construction drawing review and evaluation system is also used for:

[0081] The expression of knowledge distillation dimensionality reduction function is as follows: ;in, Characterize the distillation dimension reduction function, Characterize the source multi-index dimensionality reduction data Multiple indicator data with the source The divergence between Characterize the temperature coefficient, Characterizing the source multi-indicator data In The first dimension indicator indicator data.

[0082] Furthermore, the AI-driven green building construction drawing review and evaluation system also includes an evaluation and rectification module, which is also used to:

[0083] Rendering the target audit and evaluation results to the target green building construction drawing to obtain a target audit and evaluation visual graph; obtaining a target audit report of the target audit and evaluation visual graph, wherein the target audit report includes any audit-failed indicators; generating any rectification suggestions for the any audit-failed indicators in combination with the predetermined audit and evaluation rules; and rectifying the target green building construction drawing according to the any rectification suggestions.

[0084] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. Figure 1 The AI-driven green building construction drawing review and evaluation method and specific examples in Example 1 are also applicable to the AI-driven green building construction drawing review and evaluation system in this embodiment. Through the above detailed description of the AI-driven green building construction drawing review and evaluation method, those skilled in the art can clearly understand the AI-driven green building construction drawing review and evaluation system in this embodiment, so for the sake of brevity of the specification, it will not be described in detail here. For the system disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part description.

[0085] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

[0086] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application belong to the scope of the present application and its equivalent technology, the present application is also intended to include these modifications and variations.

Claims

1. AI-driven green building construction drawing review and evaluation method, characterized by: include: Acquire target digital information of a target green building construction drawing, wherein the target digital information includes multi-indicator data corresponding to multi-dimensional indicators; Obtaining predetermined audit evaluation dimensions, and extracting a first evaluation dimension from the predetermined audit evaluation dimensions; Matching a first related indicator set corresponding to the first evaluation dimension in the multi-dimensional indicators, and matching a first indicator parameter set of the first related indicator set in the multi-indicator data; Perform transfer learning on the initial AI audit evaluation model according to the multi-indicator data to obtain a target AI audit evaluation model; The first indicator parameter set is evaluated and analyzed by the target AI review and evaluation model to obtain a first evaluation result of the target green building construction drawing in the first evaluation dimension; Based on the first evaluation result, forming a target review and evaluation result of the target green building construction drawing; Among them, the initial AI audit evaluation model is transferred and learned according to the multi-indicator data to obtain the target AI audit evaluation model, including: Acquire source digital information of a source audit and evaluation task, wherein the source audit and evaluation task refers to an audit and evaluation task of a green building construction drawing of the same type as the target green building construction drawing; Building a source data set based on the preprocessed source multi-indicator data of the source digital information and the source audit and evaluation results; Obtaining predetermined audit evaluation rules, and building the initial AI audit evaluation model by combining the predetermined audit evaluation rules with the source data set; Among them, after obtaining the predetermined audit evaluation rules and building the initial AI audit evaluation model in combination with the predetermined audit evaluation rules and the source data set, it also includes: Introducing a knowledge distillation dimensionality reduction function to perform distillation analysis on the source multi-index data to obtain source multi-index dimensionality reduction data; Performing machine learning on the reduced-dimensional source data set formed based on the source multi-index reduced-dimensional data and the source audit and evaluation results to obtain an audit and evaluation prediction model; Integrating the audit evaluation prediction model into the initial AI audit evaluation model; The expression of the knowledge distillation dimensionality reduction function is as follows: ； in, Characterize the distillation dimensionality reduction function, Characterize the source multi-index dimensionality reduction data Multiple indicator data with the source The divergence between Characterize the temperature coefficient, Characterize the source multi-indicator data In The first dimension indicator indicator data.

2. The AI-driven green building construction drawing review and evaluation method according to claim 1, characterized in that: Obtaining predetermined audit assessment dimensions, and extracting a first assessment dimension from the predetermined audit assessment dimensions, including: Establish a set of energy-saving audit and assessment dimensions; Establish a set of water conservation audit and assessment dimensions; Establishing a set of environmental audit assessment dimensions; Build a set of comfort audit assessment dimensions; Based on the energy-saving audit and evaluation dimension set, the water-saving audit and evaluation dimension set, the environmental audit and evaluation dimension set, and the comfort audit and evaluation dimension set, the predetermined audit and evaluation dimension is formed; Among them, the energy-saving audit evaluation dimension set includes energy-saving control item evaluation indicators, energy-saving conventional item evaluation indicators and energy-saving scoring item evaluation indicators, the water-saving audit evaluation dimension set includes water-saving control item evaluation indicators, water-saving conventional item evaluation indicators and water-saving scoring item evaluation indicators, the environmental protection audit evaluation dimension set includes environmental protection control item evaluation indicators, environmental protection conventional item evaluation indicators and environmental protection scoring item evaluation indicators, and the comfort audit evaluation dimension set includes comfort control item evaluation indicators, comfort conventional item evaluation indicators and comfort scoring item evaluation indicators.

3. The AI-driven green building construction drawing review and evaluation method according to claim 1, characterized in that: After obtaining the predetermined audit evaluation rules and building the initial AI audit evaluation model in combination with the predetermined audit evaluation rules and the source data set, it includes: Sequentially extracting a first evaluation indicator and a second evaluation indicator in the first evaluation dimension; Obtaining a union of a first relevant indicator set corresponding to the first evaluation indicator and a second relevant indicator set corresponding to the second evaluation indicator, recorded as a first general indicator set; Acquire a second general indicator set of a second evaluation dimension, where the second evaluation dimension refers to any dimension in the predetermined audit evaluation dimension that is different from the first evaluation dimension; Sequentially matching in the multi-indicator data to obtain a first general indicator data set corresponding to the first general indicator set and a second general indicator data set corresponding to the second general indicator set; Based on the initial AI audit and evaluation model, parallel transfer learning is performed on the first general indicator data set and the second general indicator data set to obtain the target AI audit and evaluation model.

4. The AI-driven green building construction drawing review and evaluation method according to claim 1, characterized in that: Also includes: Rendering the target audit and evaluation results to the target green building construction drawing to obtain a target audit and evaluation visual diagram; Obtaining a target audit report of the target audit evaluation visual graph, wherein the target audit report includes any audit failure indicators; Generate any rectification suggestions for any audit-failed indicators in combination with the predetermined audit evaluation rules; Correct the target green building construction drawings according to any correction suggestions.

5. AI-driven green building construction drawing review and evaluation system, characterized by: The steps for implementing the AI-driven green building construction drawing review and evaluation method according to any one of claims 1 to 4, wherein the AI-driven green building construction drawing review and evaluation system comprises: A target information acquisition module, wherein the target information acquisition module is used to acquire target digital information of a target green building construction drawing, wherein the target digital information includes multi-indicator data corresponding to multi-dimensional indicators; An evaluation dimension acquisition module, the evaluation dimension acquisition module is used to acquire predetermined audit evaluation dimensions and extract a first evaluation dimension from the predetermined audit evaluation dimensions; An indicator matching module, the indicator matching module is used to match a first related indicator set corresponding to the first evaluation dimension in the multi-dimensional indicators, and match a first indicator parameter set of the first related indicator set in the multi-indicator data; A transfer learning module, wherein the transfer learning module is used to perform transfer learning on the initial AI audit evaluation model according to the multi-indicator data to obtain a target AI audit evaluation model; An evaluation and analysis module, the evaluation and analysis module is used to evaluate and analyze the first indicator parameter set through the target AI review and evaluation model to obtain a first evaluation result of the target green building construction drawing in the first evaluation dimension; An evaluation result building module is used to build a target review evaluation result of the target green building construction drawing based on the first evaluation result.

Images