CN115759037A - Intelligent auditing frame and auditing method for building construction scheme - Google Patents

Abstract

The invention relates to an intelligent auditing framework and an auditing method for a building construction scheme. The integrity analysis of the construction scheme to be audited is realized by analyzing the text keywords, the calculation formula symbols and the rich text information in the construction scheme to be audited and comparing the analyzed text keywords, the calculation formula symbols and the rich text information with the character information in the sensitive word bank and the knowledge map, and an excellent construction scheme is provided as a reference case to help technicians develop further modification and perfection of the construction scheme to be audited.

Description

Intelligent audit framework and audit method of building construction plan

technical field

The invention belongs to the field of building construction, and in particular relates to an intelligent auditing framework and an auditing method of a building construction plan.

Background technique

The construction plan is the implementation plan of the engineering project, and it is a large number of document materials written daily by the technical personnel of the construction company. These include organizational plans, personnel composition plans, technical plans, safety plans, material supply plans, calculation books, etc. Since the building construction plan directly guides the building construction, its completeness and accuracy are related to the safety and quality of the engineering project.

Usually, after the construction plan is compiled by the on-site technicians, it is submitted to the technical management line personnel for proofreading and review. The audit work mainly includes checking the integrity of the implementation plan based on key information, checking the accuracy of calculation formulas based on key data information, and comparing key elements of the plan to improve the content of the plan, etc. In the traditional review process, the workload of technicians is heavy, repetitive, and inefficient. How to further improve the quality of work and shorten the working hours is a problem worthy of study.

Contents of the invention

The purpose of the present invention is to provide an intelligent review framework and review method for building construction plans. The framework uses artificial intelligence to replace traditional technicians to pre-check construction plans, improve the review efficiency and quality of construction plans, and reduce the workload of technicians.

The present invention is realized through the following technical solutions:

An intelligent review framework for building construction schemes, characterized in that it includes keyword analysis based on natural language processing technology, information matching based on deep semantic similarity models, algorithm-based computational symbol recognition, and rich text analysis;

The keyword analysis based on natural language processing technology is used to analyze and segment the construction plan to be reviewed through natural language processing technology, and obtain all keywords in the construction plan to be reviewed;

The algorithm-based calculation symbol identification is used to accurately identify the calculation formula symbols in the construction plan to be reviewed through the algorithm;

The rich text analysis is used to extract all the rich text information required in various documents and generate a directory structure;

The information matching based on the deep semantic similarity model is used to establish sensitive thesaurus and knowledge graphs based on current building-related laws and regulations, standards and specifications, and other excellent construction schemes, through the presentation layer and matching based on the deep semantic similarity model Layer, LSTM, match the information in the sensitive lexicon and knowledge map with the text of the construction plan to be reviewed, complete the leak detection and filling of the construction plan to be reviewed, and list the missing items of the construction plan to be reviewed in the form of a list.

In an embodiment of the present invention, the keyword analysis based on natural language processing technology includes: using the core semantic analysis technology of natural language processing technology to carry out lexical analysis, sentence analysis, pragmatic analysis and linguistic analysis of the text of the construction plan to be reviewed. Environmental analysis, realize the keyword analysis of the construction plan text to be reviewed, and prepare for keyword matching.

In an embodiment of the present invention, the algorithm-based recognition of calculation symbols includes: using an image correction algorithm, a text detection algorithm, a text recognition algorithm, and a semantic correction algorithm to identify the calculation formula symbols in the construction plan to be reviewed, so as to be consistent with the standard Match the sensitive words in the specification, and then realize the missing and filling of calculation formula symbols

In an embodiment of the present invention, the rich text parsing includes: performing rich text parsing on each element in the text of the construction plan to be reviewed through Node parsing, table parsing, graph parsing, and document structure detection, and forming a directory structure so as to be compatible with knowledge The similar excellent schemes in the map are matched, and the matched similar excellent schemes can be used as reference cases for writing construction schemes to be reviewed.

An intelligent auditing method for building construction schemes realized based on the above-mentioned intelligent auditing framework for building construction schemes, comprising the steps of:

The first step is keyword analysis based on natural language processing technology, algorithm-based computational symbol recognition and rich text analysis, including:

Using the keyword analysis based on natural language processing technology, the construction plan to be reviewed is analyzed and word-segmented by natural language processing technology, and all keywords in the construction plan to be reviewed are obtained;

Using the algorithm-based calculation symbol recognition, the algorithm is used to accurately identify the calculation formula symbols in the construction plan to be reviewed;

Using the rich text analysis, extract all the rich text information required in various documents, and generate a directory structure;

The second step is information matching based on the deep semantic similarity model, including:

Establish a sensitive vocabulary by collecting relevant laws and regulations, standards and specifications, and excellent construction plans in the construction field;

Form a domain knowledge map by collecting domain expert knowledge and Internet domain resources;

Use the representation layer, matching layer and LSTM of the deep semantic similarity model to match sensitive thesaurus and knowledge graphs with the text information of the construction plan to be reviewed;

The third step is to output the result file, including:

Through the matching of the sensitive lexicon, the leaked items of the construction plan to be reviewed are output in the form of a list, which is convenient for technicians to check and supplement;

Through knowledge map matching, it provides references for technical personnel to modify and improve construction plans to be reviewed in the form of recommended content of excellent construction plans.

Compared with the prior art, the present invention has the following beneficial effects:

1. Quickly capture the key information in the construction plan to be reviewed by analyzing the key words of the construction plan to be reviewed, calculation formula symbols and rich text information, and avoid the blindness of the construction plan review;

2. Establish a sensitive lexicon and knowledge map through artificial intelligence, realize the rapid matching of laws, regulations, standards and specifications, and excellent construction plans with construction plans to be reviewed, improve the speed of completeness inspection of plans to be reviewed, and provide technical personnel with reference to excellent construction plans plan;

3. Through the review framework, 90% of the program review workload can be reduced, and the work efficiency and quality of technical personnel can be improved.

Description of drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings that need to be used in the description of the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention. For those skilled in the art, other drawings can also be obtained based on these drawings without creative effort.

Fig. 1 is a diagram of the audit steps of the intelligent audit framework of the building construction scheme provided by the embodiment of the present invention.

Detailed ways

Name explanation:

NLP: Natural Language Processing (NLP, Natural Language Processing) is a discipline that studies the language problems of human-computer interaction. According to the difficulty of technical implementation, such systems can be divided into three types: simple matching, fuzzy matching and paragraph understanding.

DSSM: Deep Structured Semantic Models, based on the deep semantic similarity model, the principle of DSSM is to use DNN to express Query and Title in the search engine as low-latitude semantic vectors, and calculate the distance between the two semantic vectors through the cosine distance, Finally, a semantic similarity model is trained. This model can be used not only to predict the semantic similarity of two sentences, but also to obtain the low-dimensional semantic vector representation of a sentence. DSSM can be divided into three layers from bottom to top: input layer, presentation layer, and matching layer.

The specific embodiments of the present invention will be further described below in conjunction with the accompanying drawings. It should be noted here that the descriptions of these embodiments are used to help understand the present invention, but are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below may be combined with each other as long as they do not constitute a conflict with each other.

The invention provides an intelligent review framework for building construction schemes. After receiving the construction scheme, the framework operates in parallel through multiple processing components such as element extraction, element audit, and expert advice, and finally outputs the preprocessed result text. The framework uses artificial intelligence to preprocess the construction plan, which reduces a lot of repetitive labor for technicians, further improves the efficiency and quality of construction plan review, and reduces labor costs.

To achieve the above object, the technical solution adopted in the present invention is:

An intelligent review framework for building construction schemes, including keyword analysis based on natural language processing technology (NLP), information matching based on deep semantic similarity model (DSSM), algorithm-based computational symbol recognition and rich text analysis. The keyword analysis based on natural language processing technology (NLP) is characterized in that: the core semantic analysis technology of natural language processing technology is adopted to carry out lexical analysis, sentence analysis, pragmatic analysis and context analysis of the construction plan text to be reviewed, and realize The keyword analysis of the text of the construction plan to be reviewed prepares for keyword matching; the information matching feature based on the deep semantic similarity model (DSSM) is: the establishment of relevant laws and regulations based on current construction, standard specifications, and other excellent construction plans Sensitive thesaurus and knowledge map, through the presentation layer, matching layer, and LSTM in DSSM, match the information in the sensitive thesaurus and knowledge map with the text of the construction plan to be reviewed, complete the leak detection and filling of the construction plan to be reviewed, and List the missing items of the construction plan to be reviewed in the form of a list; the algorithm-based calculation symbol recognition is characterized in that: through the image correction algorithm, text detection algorithm, text recognition algorithm and semantic correction algorithm, the calculation formula symbols in the construction plan to be reviewed are processed. Recognize, so as to match with the sensitive words in the standard specification, and then realize the missing filling of calculation formula symbols; the rich text analysis is characterized in that: through Node analysis, table analysis, chart analysis and document structure detection, the text of the construction plan to be reviewed Contents, titles, paragraphs, headers, footers, tables, pictures and other elements in the content are analyzed in rich text, and a directory structure is formed to match with similar excellent solutions in the knowledge graph. Reference case for program writing.

The working principle of an intelligent review framework for a building construction plan in the present invention is: analyze the text keywords, calculation formula symbols, and rich text information in the construction plan to be reviewed, and compare it with the text information in the sensitive lexicon and knowledge map , realize the integrity analysis of the construction plan to be reviewed, and provide excellent construction plans as reference cases to help technicians further modify and improve the construction plan to be reviewed.

Referring to Fig. 1, the audit steps based on the intelligent audit framework of the building construction scheme of the present invention include:

(1) The first step of the intelligent review framework includes keyword analysis based on natural language processing technology (NLP), algorithm-based calculation symbol recognition and rich text analysis:

1) Keyword analysis based on natural language processing technology (NLP), which aims to analyze and segment the construction plan to be reviewed through natural language processing technology (NLP), and obtain all keywords in the construction plan to be reviewed. Among them, the core of natural language processing technology is semantic analysis technology, including lexical analysis, syntactic analysis, pragmatic analysis and context analysis. Analysis of the meaning contained in the paragraph;

2) Algorithm-based calculation symbol recognition aims to accurately identify the calculation formula symbols in the construction plan to be reviewed through the algorithm. Among them, the algorithm mainly includes image correction algorithm, text detection algorithm, text recognition algorithm and semantic correction algorithm, so as to realize the transformation of symbol pictures into computer character sequences and prepare for the next step of matching;

3) Rich text analysis, which aims to extract all the rich text information required in various documents such as PDF and WORD, and generate a directory structure. Among them, the rich text analysis mainly uses NODE analysis, table analysis, chart analysis and document structure detection, and forms a directory structure after analyzing all elements including directories, titles, paragraphs, headers, footers, tables, and pictures.

(2) The second step of the intelligent review framework is information matching based on the deep semantic similarity model (DSSM):

1) Establish a sensitive lexicon by collecting relevant laws and regulations, standards and specifications, and excellent construction plans in the construction field;

2) Form a domain knowledge map by collecting domain expert knowledge and Internet domain resources;

3) Use the representation layer and matching layer of the deep semantic similarity model DSSM and the DSSM model of LSTM to match the sensitive lexicon and knowledge graph with the text information of the construction plan to be reviewed.

(3) The third step of the intelligent review framework is to output the result file:

1) Output the leakage items of the construction plan to be reviewed in the form of a list through the matching of sensitive thesaurus, which is convenient for technicians to check and supplement;

2) Through knowledge map matching, in the form of recommended content of excellent construction plans, it provides references for technicians to modify and improve construction plans to be reviewed.

Compared with the prior art, the present invention has the following advantages:

1. Quickly capture the key information in the construction plan to be reviewed by analyzing the key words of the construction plan to be reviewed, calculation formula symbols and rich text information, and avoid the blindness of the construction plan review;

2. Establish a sensitive lexicon and knowledge map through artificial intelligence, realize the rapid matching of laws, regulations, standards and specifications, and excellent construction plans with construction plans to be reviewed, improve the speed of completeness inspection of plans to be reviewed, and provide technical personnel with reference to excellent construction plans plan;

3. Through the review framework, 90% of the program review workload can be reduced, and the work efficiency and quality of technical personnel can be improved.

The embodiments of the present invention have been described in detail above with reference to the accompanying drawings, but the present invention is not limited to the described embodiments. For those skilled in the art, without departing from the principle and spirit of the present invention, various changes, modifications, substitutions and modifications to these embodiments still fall within the protection scope of the present invention.

From the foregoing it will be seen that the invention is well adapted to carry out all of the above objects and objects, as well as other advantages which are obvious and inherent in the structure. It should be understood that certain features and subcombinations are useful and can be used without reference to other features and subcombinations. This is within the scope of the present invention.

Claims (5)

1. An intelligent auditing framework for a building construction scheme is characterized by comprising keyword analysis based on a natural language processing technology, information matching based on a deep semantic similarity model, calculation symbol identification based on an algorithm and rich text analysis;

the keyword analysis based on the natural language processing technology is used for analyzing and segmenting the construction scheme to be audited through the natural language processing technology to obtain all keywords in the construction scheme to be audited;

the calculation symbol identification based on the algorithm is used for accurately identifying the calculation formula symbol in the construction scheme to be audited through the algorithm;

the rich text analysis is used for extracting all rich text information required in various documents and generating a directory structure;

the information matching based on the deep semantic similarity model is used for establishing a sensitive word bank and a knowledge map based on the relevant laws and regulations, standard specifications and other excellent construction schemes of the existing building, matching the information in the sensitive word bank and the knowledge map with the text of the construction scheme to be checked through a presentation layer, a matching layer and an LSTM based on the deep semantic similarity model, completing the missing and filling of the construction scheme to be checked, and listing the missing and writing items of the construction scheme to be checked in a list form.

2. The intelligent building construction plan review framework of claim 1, wherein the keyword parsing based on natural language processing techniques comprises: and performing lexical analysis, statement analysis, pragmatic analysis and context analysis on the construction scheme text to be audited by adopting a core semantic analysis technology of a natural language processing technology, realizing keyword analysis on the construction scheme text to be audited, and preparing for keyword matching.

3. The building construction scheme intelligent audit framework of claim 1 wherein the algorithm based computational symbol identification includes: and identifying the calculation formula symbols in the construction scheme to be checked through an image correction algorithm, a character detection algorithm, a character identification algorithm and a semantic correction algorithm so as to be matched with sensitive words in the standard specification, and further realizing the omission and filling of the calculation formula symbols.

4. The building construction plan intelligent audit framework of claim 1, wherein the rich text parsing comprises: through Node analysis, table analysis, chart analysis and document structure detection, rich text analysis is carried out on all elements in the construction scheme text to be audited, a directory structure is formed so as to be matched with similar excellent schemes in a knowledge graph, and the similar excellent schemes obtained through matching can be used as writing reference cases of the construction scheme to be audited.

5. An intelligent auditing method for construction scheme based on the intelligent auditing framework for construction scheme of any one of claims 1 to 4, comprising the steps of:

the method comprises the following steps of firstly, keyword analysis based on natural language processing technology, calculation symbol identification based on algorithm and rich text analysis, wherein the method comprises the following steps:

analyzing and segmenting the construction scheme to be audited through the natural language processing technology by adopting the keyword analysis based on the natural language processing technology to obtain all keywords in the construction scheme to be audited;

adopting the calculation symbol identification based on the algorithm to accurately identify the calculation formula symbol in the construction scheme to be audited through the algorithm;

adopting the rich text analysis to extract all the rich text information required in various documents and generating a directory structure;

and secondly, matching information based on the deep semantic similarity model, including:

establishing a sensitive word bank by acquiring relevant laws, regulations, standard specifications and excellent construction schemes in the building field;

forming a domain knowledge map by collecting domain expert knowledge and Internet domain resources;

matching the sensitive word bank, the knowledge graph and the text information of the construction scheme to be checked by adopting an expression layer, a matching layer and an LSTM of a depth semantic similarity model;

thirdly, outputting a result file, comprising:

through sensitive word bank matching, leakage items of the construction scheme to be checked are output in a list mode, and technicians can conveniently check and supplement the leakage items;

and through knowledge map matching, in an excellent construction scheme content recommendation form, reference is provided for technical personnel to modify and perfect the construction scheme to be checked.

Images