Intelligent graph auditing method and device based on knowledge graph

Abstract

The invention provides an intelligent graph auditing method and device based on a knowledge graph, and the method comprises the steps: carrying out the analysis of a to-be-audited engineering drawing in the field of rail transit, obtaining to-be-audited information in the engineering drawing, and enabling the to-be-audited information to comprise a graph, a character, and a corresponding relation between the graph and the character; based on the rail transit knowledge graph and the to-be-audited information, performing drawing audit on the engineering drawing to obtain a drawing audit result of the engineering drawing; the rail transit knowledge graph is constructed based on expert drawing checking experience in the field of rail transit and engineering execution standards and / or engineering drawing specifications, the defects that in a traditional scheme, manual drawing checking is long in period and low in efficiency, and a universal drawing checking method is not suitable for drawings in the field of rail transit and is poor in drawing checking effect are overcome, and the drawing checking efficiency is improved. The accuracy and effectiveness of drawing checking are guaranteed, the drawing checking efficiency is greatly improved, and a guarantee is provided for the safety and quality of engineering construction in the field of rail transit.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to an intelligent image review method and device based on a knowledge graph. Background Art

[0002] In the rapidly developing rail transit industry, the review of engineering drawings is a key link in ensuring project safety and quality, and its importance is self-evident.

[0003] At present, most engineering drawings in the field of rail transit are reviewed manually, that is, review experts review the engineering drawings based on their professional knowledge and rich experience. Although this review method can accurately identify problems and hidden dangers in the drawings, it has problems such as high labor costs, long review cycle, and low efficiency.

[0004] However, with the continuous expansion of the scale of rail transit construction, the traditional review model can no longer meet the growing engineering needs. Therefore, how to improve the efficiency of review while ensuring its effectiveness has become an urgent problem to be solved. Summary of the invention

[0005] The present invention provides an intelligent drawing review method and device based on knowledge graph, which is used to solve the problems of high cost, long cycle and low efficiency of manual drawing review in the prior art, realize intelligent review of engineering drawings in the field of rail transit, and improve the efficiency and accuracy of drawing review.

[0006] The present invention provides an intelligent image review method based on knowledge graph, comprising: Obtain engineering drawings to be reviewed in the field of rail transit; Parsing the engineering drawings to obtain the information to be reviewed in the engineering drawings, wherein the information to be reviewed includes graphics, text, and a corresponding relationship between the graphics and the text; Based on the rail transit knowledge graph and the information to be reviewed, the engineering drawings are reviewed to obtain the review results of the engineering drawings; The rail transit knowledge graph is constructed based on the expert drawing review experience in the rail transit field, as well as engineering execution standards and / or engineering drawing specifications.

[0007] According to an intelligent drawing review method based on knowledge graph provided by the present invention, the engineering drawing is parsed to obtain the information to be reviewed in the engineering drawing, including: Parsing the engineering calculation sheet and / or design drawing in the engineering drawing to obtain the drawing type of the engineering drawing and the information to be reviewed in the engineering drawing; The step of performing drawing review on the engineering drawings based on the rail transit knowledge graph and the information to be reviewed to obtain drawing review results of the engineering drawings includes: Based on the drawing type, a target knowledge graph is determined from the rail transit knowledge graph, and based on the target knowledge graph and the information to be reviewed, a drawing review is performed on the engineering drawings to obtain a drawing review result of the engineering drawings.

[0008] According to an intelligent drawing review method based on a knowledge graph provided by the present invention, based on the target knowledge graph and the information to be reviewed, the engineering drawing is reviewed to obtain the drawing review result of the engineering drawing, including: Based on the information to be reviewed, determining an entity triple from the target knowledge graph, wherein the entity triple includes multiple target entities and entity relationships between the target entities; Based on the entity triples and the information to be reviewed, the engineering drawings are reviewed to obtain the review results of the engineering drawings.

[0009] According to an intelligent drawing review method based on a knowledge graph provided by the present invention, based on the rail transit knowledge graph and the information to be reviewed, the engineering drawings are reviewed to obtain the drawing review results of the engineering drawings, and then the method further includes: In the case where the drawing review result indicates that the engineering drawing is non-compliant, determining the non-compliant part of the engineering drawing, and the items and contents of the engineering execution standards and / or engineering drawing specifications in the field of rail transit that the non-compliant part violates; The non-compliance is displayed, as well as the items and contents in the engineering execution standards and / or engineering drawing specifications in the rail transit field that are violated by the non-compliance.

[0010] According to an intelligent map review method based on a knowledge graph provided by the present invention, the rail transit knowledge graph is determined based on the following steps: Perform entity extraction on the expert drawing review experience, the engineering execution standard and / or the engineering drawing specification to obtain a plurality of entities; the entities include graphic entities and text entities; Extracting relationships based on each entity and the engineering execution standard and / or the engineering drawing specification to obtain entity relationships between the entities; Based on the entities and the entity relationships between the entities, a rail transit knowledge graph is constructed.

[0011] According to an intelligent drawing review method based on knowledge graph provided by the present invention, the expert drawing review experience, the engineering execution standard and / or the engineering drawing specification are subjected to entity extraction to obtain multiple entities, including: Extract the key points of the expert review experience to obtain the key points of the expert review; Based on the expert review key points, the engineering implementation standard and / or the engineering drawing specification are screened to obtain the target implementation standard and / or the target drawing specification; Perform entity extraction on the expert review key points, the target execution standard and / or the target mapping specification to obtain multiple entities; The extracting of relationships based on each entity and the engineering execution standard and / or the engineering drawing specification to obtain entity relationships between the entities includes: Relationship extraction is performed based on the entities, the target execution standard and / or the target mapping specification to obtain entity relationships between the entities.

[0012] The present invention also provides an intelligent image review device based on knowledge graph, comprising: An acquisition unit, used to acquire engineering drawings to be reviewed in the field of rail transit; A parsing unit, used to parse the engineering drawings to obtain the information to be reviewed in the engineering drawings, wherein the information to be reviewed includes graphics, text, and a corresponding relationship between the graphics and the text; An audit unit, configured to perform a drawing audit on the engineering drawing based on the rail transit knowledge graph and the information to be audited, and obtain a drawing audit result of the engineering drawing; The rail transit knowledge graph is constructed based on the expert drawing review experience in the rail transit field, as well as engineering execution standards and / or engineering drawing specifications.

[0013] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, an intelligent image review method based on a knowledge graph as described above is implemented.

[0014] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-described intelligent image review methods based on knowledge graphs.

[0015] The present invention also provides a computer program product, including a computer program, which, when executed by a processor, implements any of the above-described intelligent image review methods based on knowledge graphs.

[0016] The intelligent drawing review method and device based on knowledge graph provided by the present invention reviews the information to be reviewed in the engineering drawings through the rail transit knowledge graph constructed on the basis of the expert review experience in the rail transit field, engineering execution standards and / or engineering drawing specifications, and obtains the drawing review results, thereby realizing efficient and automated review of engineering drawings in the rail transit field, overcoming the defects of long manual review cycle and low efficiency in traditional solutions, as well as the general review method being unsuitable for drawings in the rail transit field and poor review effect, which not only ensures the accuracy and effectiveness of the review, but also greatly improves the review efficiency; in addition, it can also effectively assist review experts in conducting highly complex drawing reviews, saving a lot of manpower and material costs, and providing guarantees for the safety and quality of engineering construction in the rail transit field. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the present invention or the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0018] Figure 1 It is a flowchart of the intelligent image review method based on knowledge graph provided by the present invention; Figure 2 It is a schematic diagram of the rail transit knowledge graph construction process provided by the present invention; Figure 3 It is the overall framework diagram of the intelligent image review method based on knowledge graph provided by the present invention; Figure 4 It is a structural schematic diagram of an intelligent image review device based on knowledge graph provided by the present invention; Figure 5 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0019] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0020] In the rapidly developing rail transit industry, the review of engineering drawings is a key link in ensuring engineering safety and quality, and its importance is self-evident. The traditional manual review method is highly dependent on the professional knowledge and experience of review experts. Although it can accurately identify problems in drawings and potential safety hazards, it faces challenges such as high labor costs, long review cycles, and low efficiency. With the continuous expansion of the scale of rail transit construction, the traditional manual review method can no longer meet the growing engineering needs, and there is an urgent need to introduce more efficient and intelligent drawing review technology.

[0021] At present, although intelligent drawing review technology has shown great potential in many fields, its application in the rail transit industry is still in the initial exploration stage. The main obstacle is that the engineering drawings in this industry are highly professional and complex, the number of samples is relatively limited, and the drawing standards and specifications are diverse, which leads to many difficulties in directly applying general intelligent drawing review technology.

[0022] In this regard, the present invention provides an intelligent drawing review method based on a knowledge graph, which aims to construct a rail transit knowledge graph through the expert review experience in the rail transit field, as well as engineering execution standards and / or engineering drawing specifications, and use the rail transit knowledge graph to perform automated and intelligent review of engineering drawings, thereby overcoming the current defects of long drawing review cycle, low efficiency, and high review cost in the rail transit field, improving the review efficiency and accuracy, reducing the review cost, and providing guarantees for the safety and quality of engineering construction in the rail transit field.

[0023] Figure 1 is a flow chart of the intelligent image review method based on knowledge graph provided by the present invention, such as Figure 1 As shown, the method includes: Step 110, obtaining engineering drawings to be reviewed in the field of rail transit; Step 120, parsing the engineering drawings to obtain the information to be reviewed in the engineering drawings, where the information to be reviewed includes graphics, text, and the corresponding relationship between the graphics and text; Step 130, based on the rail transit knowledge graph and the information to be reviewed, the engineering drawings are reviewed to obtain the review results of the engineering drawings; Among them, the rail transit knowledge graph is constructed based on the expert review experience in the rail transit field, as well as engineering execution standards and / or engineering drawing specifications.

[0024] Specifically, considering that most of the current engineering drawings in the field of rail transit are reviewed manually, although this method can more accurately identify the problems in the drawings and the potential safety hazards, it is costly, time-consuming, and inefficient. In addition, considering that the current common intelligent drawing review technology is difficult to apply in the field of rail transit due to the high complexity and professionalism of drawings in the field of rail transit, as well as the scarcity of samples, non-standard drawings and many other problems, the review effect is not good.

[0025] Based on this, in an embodiment of the present invention, it is proposed that a rail transit knowledge graph can be pre-constructed based on the expert review experience in the rail transit field, as well as engineering execution standards and / or engineering drawing specifications. Based on this rail transit knowledge graph, the engineering drawings in the rail transit field are reviewed, so as to utilize the rich information in the knowledge graph to automatically review the contents in the engineering drawings, thereby realizing intelligent review of engineering drawings in the rail transit field, which can improve the review efficiency while ensuring the accuracy and effectiveness of the review.

[0026] It is understandable that in the actual rail transit drawing review process, before the review, the first thing to be determined is the drawings to be reviewed, that is, the engineering drawings to be reviewed in the rail transit field. The engineering drawings can be obtained from the relevant documents of the project engineering in the rail transit field. They can be design drawings of the corresponding project engineering, such as plane structure drawings in CAD (Computer Aided Design) format, three-dimensional structure drawings drawn by Solidwordks, etc., or calculation books in document format in the project engineering, or other documents related to the project engineering, such as files representing specific data of the design object of the project engineering (such as materials used, service life, environment, location, load-bearing, etc.), which are not limited to the drawings themselves.

[0027] After determining the engineering drawings to be reviewed, in an embodiment of the present invention, the engineering drawings can be parsed to obtain the information that needs to be reviewed and confirmed, thereby obtaining the information to be reviewed in the engineering drawings. Specifically, after the project is launched, different software may be used in the process of making engineering drawings, or the drawings may be made in different ways. The formats of the resulting drawings are different. In order to facilitate subsequent processing and accurately capture the information that needs to be reviewed, the engineering drawings to be reviewed can be pre-processed here to make them more standardized and unified, so as to meet the requirements for input drawings during the review. Specifically, the design drawings in the engineering drawings can be converted / unified in format (such as converting the design drawings into Drawing format), the calculation books and other documents can be standardized in terms of terminology and content, etc., so as to obtain the pre-processed engineering drawings.

[0028] Furthermore, the preprocessed engineering drawings can be parsed to obtain the information to be reviewed, which can include the text and graphics in the engineering drawings, as well as the correspondence between the text and the graphics. For example, if a graphic symbol in the engineering drawing is accompanied by a text description, then the graphic symbol and the text description, as well as the correspondence between the two, are all information that needs to be paid attention to during the drawing review process, and are all information to be reviewed.

[0029] Specifically, the design drawings and calculation books in the engineering drawings can be parsed to find out the specific text and graphics in the design drawings that need to be reviewed during the review process, as well as the correspondence between text and images, and the data in the calculation books and other project documents that need to be focused on during the review process, such as the environment, geology, materials used, etc., so as to obtain the information to be reviewed. The analysis of engineering drawings here can be achieved through computer software, algorithms, etc.

[0030] After obtaining the information to be reviewed in the engineering drawings, in an embodiment of the present invention, a pre-constructed rail transit knowledge graph can be applied to review the information to be reviewed to obtain the drawing review results, thereby realizing automatic review of engineering drawings in the rail transit field, avoiding the time-consuming and labor-intensive manual review in traditional solutions, as well as the long cycle and low efficiency.

[0031] Specifically, here, on the basis of the parsed information to be reviewed, the rail transit knowledge graph can be used to review the engineering drawings, that is, according to the implementation standards, specifications, drawing requirements, etc. in the rail transit field represented by the rail transit knowledge graph, the graphics, text, and graphic-text correspondence information extracted from the engineering drawings are reviewed to determine whether they comply with the standards, specifications, etc. in the rail transit field reflected by the rail transit knowledge graph, thereby obtaining the review results of the engineering drawings, that is, the drawing review results.

[0032] It should be noted that before reviewing drawings with the help of the rail transit knowledge graph, it is necessary to pre-construct the rail transit knowledge graph. In the embodiment of the present invention, in order to ensure the accuracy of the review and the reliability and practicality of the review results, a large number of implementation standards, drawing specifications, etc. in the field of rail transit are selected to be collected in advance, and the review experts in this field have obtained many years of review experience, and the text is organized. According to the obtained expert review experience, as well as the engineering implementation standards and / or engineering drawing specifications, the rail transit knowledge graph is constructed. The rail transit knowledge graph constructed in this way can represent many specifications, restriction and constraint items, drawing review points, etc. in the field of rail transit, and includes many key points, difficulties, and error-prone points involved in the review process. Therefore, based on this, the review of engineering drawings can not only greatly improve the efficiency of review, but also effectively ensure the comprehensiveness and accuracy of the review.

[0033] In an embodiment of the present invention, a rail transit knowledge graph is constructed based on the expert review experience in the rail transit field, as well as engineering execution standards and / or engineering drawing specifications. This can effectively integrate professional knowledge, design specifications, etc. in the rail transit field, such as engineering execution standards, engineering drawing specifications, expert review experience and other multi-source heterogeneous information, to form a rich rail transit field knowledge base, and use structured semantic relationships to express information to obtain a rail transit knowledge graph. Drawing review based on this can achieve a deep understanding and semantic analysis of the information to be reviewed in the engineering drawings, so that problems and hidden dangers in the drawings can be quickly and accurately identified, thereby ensuring the safety and quality of rail transit engineering construction.

[0034] The intelligent review method based on knowledge graph provided by the present invention reviews the information to be reviewed in the engineering drawings through the rail transit knowledge graph constructed on the basis of the expert review experience in the rail transit field, engineering execution standards and / or engineering drawing specifications, and obtains the drawing review results of the engineering drawings, thereby realizing efficient and automated review of engineering drawings in the rail transit field, overcoming the defects of long manual review cycle and low efficiency in traditional solutions, and the general review method being unsuitable for drawings in the rail transit field and poor review effect, which not only ensures the accuracy and effectiveness of the review, but also greatly improves the review efficiency; in addition, it can also effectively assist review experts in conducting highly complex drawing reviews, saving a lot of manpower and material costs, and providing guarantees for the safety and quality of engineering construction in the rail transit field.

[0035] Based on the above embodiment, step 120 and step 130 include: Parse the engineering calculation sheet and / or design drawing in the engineering drawing to obtain the drawing type of the engineering drawing and the information to be reviewed in the engineering drawing; Based on the drawing type, the target knowledge graph is determined from the rail transit knowledge graph, and based on the target knowledge graph and the information to be reviewed, the engineering drawings are reviewed to obtain the drawing review results of the engineering drawings.

[0036] Specifically, in step 120 and step 130, the engineering drawings are parsed to obtain the information to be reviewed in the engineering drawings, and the engineering drawings are reviewed according to the rail transit knowledge graph and the information to be reviewed to obtain the process of drawing review results, which may specifically include: First, the engineering drawings can be parsed to extract the information that needs to be paid attention to in the drawing review. That is, the design drawings and / or engineering calculation books in the engineering drawings of the project can be parsed through computer software. Specifically, the design drawings in the Drawing format after pre-processing and / or the calculation books in the document format of the project can be parsed to identify the type of design drawings, such as plan views, section views, or general drawings, etc., and to extract the graphics and text that need to be paid attention to in the drawing review process, as well as the correspondence between images and text, so as to obtain the drawing type and information to be reviewed of the engineering drawings.

[0037] Then, according to this drawing type, the target knowledge graph can be determined from the rail transit knowledge graph. The target knowledge graph here is a type of rail transit knowledge graph that corresponds to the drawing type of the engineering drawing. Since there are many types of engineering drawings in the rail transit field, the types of engineering drawings produced in different engineering projects may be different. What's more, multiple types of drawings may be produced in one engineering project for the convenience of engineers to review. For different types of drawings, the production requirements and corresponding specifications are different. Therefore, in order to achieve more accurate drawing review, in an embodiment of the present invention, when constructing the rail transit knowledge graph, knowledge graphs can be constructed in advance for various drawing types to obtain rail transit field knowledge graphs corresponding to each type. In this way, in the actual drawing review process, the corresponding rail transit knowledge graph can be selected according to the drawing type, which is called the target knowledge graph. Based on the target knowledge graph, the information to be reviewed is reviewed to obtain the drawing review result.

[0038] Here, specifically, based on the information to be reviewed, the engineering drawings are reviewed using the rich information in the field of rail transit contained in the target knowledge graph. Specifically, the text, image and corresponding relationship extracted from the engineering drawings meet the corresponding relationship in the target knowledge graph. In other words, the graphics and text are queried in the target knowledge graph, and it is confirmed whether they meet the corresponding relationship, whether this relationship matches the directly extracted corresponding relationship, and then the graphics and text in the engineering drawings meet the specific requirements of the implementation standards and drawing specifications in the field of rail transit, so as to obtain the drawing review results of the engineering drawings. The drawing review results here can be an overall result, such as passing the review or failing the review, and can also include the review results of the specific content in the engineering drawings, for example, which part of the graphics and text is wrong and does not meet the provisions of a certain item in the engineering implementation standards and / or engineering drawing specifications in the field of rail transit, and the content of the item is XXXX.

[0039] Based on the above embodiment, based on the target knowledge graph and the information to be reviewed, the engineering drawings are reviewed to obtain the review results of the engineering drawings, including: Based on the information to be reviewed, determine entity triples from the target knowledge graph, where the entity triples contain multiple target entities and entity relationships between the target entities; Based on the entity triples and the information to be reviewed, the engineering drawings are reviewed to obtain the review results of the engineering drawings.

[0040] Specifically, the process of conducting a drawing review on the engineering drawings according to the target knowledge graph and the information to be reviewed to obtain the drawing review results of the engineering drawings may specifically include: First, we can find the corresponding entity triples from the target knowledge graph based on the correspondence between the graphics, text, and images and text in the information to be reviewed. That is, we can find the entities corresponding to the graphics and text in the information to be reviewed from the target knowledge graph, which are called target entities, and extract the triples containing this target entity, so that we can get the entity triples corresponding to the information to be reviewed. In addition to the target entity, the entity triples also contain the entity relationships between the target entities, which can be understood as the correspondence between graphics and text.

[0041] Then, the information to be reviewed can be reviewed according to each entity triplet, that is, based on the entity triples found in the target knowledge graph for each group of graphics and texts in the information to be reviewed, the group of graphics, texts and corresponding relationships can be reviewed and judged to determine whether the target entity and entity relationship match the graphics, texts and corresponding relationships, that is, whether the graphics and texts in the information to be reviewed satisfy the relationship represented by the entity triples in the target knowledge graph, thereby obtaining the drawing review result.

[0042] Based on the above embodiment, based on the rail transit knowledge graph and the information to be reviewed, the engineering drawings are reviewed to obtain the review results of the engineering drawings, and then the following steps are further included: If the drawing review results indicate that the engineering drawings are non-compliant, determine the non-compliant parts of the engineering drawings, as well as the items and contents of the engineering execution standards and / or engineering drawing specifications in the field of rail transit that are violated by the non-compliant parts; Display the non-compliances and the items and contents of the engineering execution standards and / or engineering drawing specifications in the field of rail transit that the non-compliances violate.

[0043] Specifically, in step 130, the engineering drawings are reviewed according to the rail transit knowledge graph and the information to be reviewed. After obtaining the review results, if the review results indicate that the engineering drawings are wrong and there are non-compliant areas, the non-compliant areas in the engineering drawings can be marked to facilitate subsequent drafting personnel to review and modify. For example, the non-compliant areas in the engineering drawings found by the review can be marked by underlining, background color, annotation boxes, etc., and the specific items and item contents corresponding to the non-compliant areas can be determined, that is, the specific items and contents in the engineering execution standards and / or engineering drawing specifications in the field of rail transit violated by each non-compliant area in the engineering drawings can be determined. After that, the marked content and the violated items can be displayed, that is, the non-compliant areas in the engineering drawings and the items and contents in the engineering execution standards and / or engineering drawing specifications in the field of rail transit violated by the non-compliant areas can be displayed, so that relevant personnel can clearly and intuitively find the errors, so as to make corrections in time and avoid them in the subsequent process.

[0044] It should be noted that the determined non-compliances, as well as the specific items violated and their contents can be attached to the engineering drawings. After intelligent review through the rail transit knowledge graph, the engineering drawings are output, which may include the overall review results, the review results of each part, the non-compliances, and the specific items violated by the non-compliances and their contents.

[0045] Based on the above embodiment, the rail transit knowledge graph is determined based on the following steps: Entity extraction is performed on the expert review experience, engineering execution standards and / or engineering drawing specifications to obtain multiple entities; the entities include graphic entities and text entities; Relationship extraction is performed based on each entity, as well as engineering execution standards and / or engineering drawing specifications to obtain entity relationships between each entity; A rail transit knowledge graph is constructed based on each entity and the entity relationships between entities.

[0046] Specifically, the process of determining the rail transit knowledge graph may include the following steps: Figure 2 is a schematic diagram of the rail transit knowledge graph construction process provided by the present invention, such as Figure 2 As shown, when constructing the rail transit knowledge graph, we first need to collect the data required for the construction of the rail transit knowledge graph. There are two main parts: one is the industry experience of the review experts, that is, the expert review experience, which can be converted into text through written records and other methods; the other is the implementation standards and / or drawing specifications of engineering construction in the rail transit field, that is, engineering implementation standards and / or engineering drawing specifications.

[0047] Next, entity extraction can be performed on the acquired expert review experience, as well as engineering implementation standards and / or engineering drawing specifications to extract graphic entities and text entities therefrom, that is, to extract expert review experience in the field of rail transit, as well as graphic entities and text entities in engineering implementation standards and / or engineering drawing specifications. The text entities may be graphic names, calculation word names, etc.

[0048] Afterwards, the relationship between the engineering execution standards and / or engineering drawing specifications can be extracted based on the extracted entities to obtain the entity relationship between the entities. That is, the relationship between the entities can be extracted from the specific expression of the items in the engineering execution standards and / or engineering drawing specifications to obtain the entity relationship.

[0049] Then, based on the entities extracted from the entities and the entity relationships between the entities extracted from the relationships, a rail transit knowledge graph can be constructed to represent the standards and specifications in the rail transit field reflected by the entities and entity relationships through a structured semantic network. Here, the constructed rail transit knowledge graph can have multiple types, which can be specifically divided according to the drawing types of the project engineering, including the general drawing knowledge graph, the section drawing knowledge graph, the plan view knowledge graph, etc. Different types of knowledge graphs may contain the same entity, which can be associated in different types of knowledge graphs. Therefore, when applying the rail transit knowledge graph for drawing review, multiple types of knowledge graphs can be combined to review the drawings based on the entity triples in each knowledge graph and the entity triples associated across the knowledge graphs, making the drawing review process more comprehensive and accurate.

[0050] Based on the above embodiment, entity extraction is performed on the expert review experience, and the engineering execution standards and / or engineering drawing specifications to obtain multiple entities; relationship extraction is performed based on each entity, and the engineering execution standards and / or engineering drawing specifications to obtain entity relationships between each entity, including: Extract the key points of expert review experience and obtain the key points of expert review; Based on the key points of expert review, the key points of engineering implementation standards and / or engineering drawing specifications are screened to obtain the target implementation standards and / or target drawing specifications; Entity extraction is performed on the key points of expert drawing review, target execution standards and / or target drawing specifications to obtain multiple entities; Relationship extraction is performed based on each entity, as well as a target execution standard and / or a target mapping specification to obtain entity relationships between entities.

[0051] Specifically, the above-mentioned entity extraction of expert drawing review experience, engineering execution standards and / or engineering drawing specifications to obtain multiple entities, and the process of extracting relationships based on each entity and engineering execution standards and / or engineering drawing specifications to obtain entity relationships between each entity specifically includes the following steps: After obtaining the expert review experience, considering that the expert review experience is generally of primary and secondary importance, in the embodiment of the present invention, the expert review experience can be screened first to select the key points and main points that need to be paid attention to in the review, thereby obtaining the expert review main points. That is, the key points of the expert review experience can be extracted first to extract the content that needs to be paid attention to in the review process, thereby obtaining the expert review main points.

[0052] After that, the engineering implementation standards and / or engineering drawing specifications can be screened according to the key points of the expert review to select the items that the review experience focuses on, that is, all the contents of the engineering implementation standards and / or engineering drawing specifications are screened to select the important items corresponding to the key points of the expert review, so as to obtain the screened engineering implementation standards and / or engineering drawing specifications, that is, the target implementation standards and / or target drawing specifications. Then, entity extraction can be performed on the key points of the expert review, as well as the target implementation standards and / or target drawing specifications to extract graphic entities and text entities therefrom, and relationship extraction can be performed on the target implementation standards and / or target drawing specifications based on the extracted graphic entities and text entities, so as to extract entity relationships from the specific expressions of each item in the target implementation standards and / or target drawing specifications.

[0053] Figure 3 This is the overall framework diagram of the intelligent image review method based on knowledge graph provided by the present invention, such as Figure 3 As shown, the method includes: First, obtain the engineering drawings to be reviewed in the rail transit field.

[0054] Next, the engineering drawings are parsed to obtain the information to be reviewed in the engineering drawings, which includes graphics, text, and the corresponding relationship between the graphics and text.

[0055] Here, the engineering drawings are parsed to obtain the information to be reviewed in the engineering drawings, including: parsing the engineering calculation sheets and / or design drawings in the engineering drawings to obtain the drawing types of the engineering drawings and the information to be reviewed in the engineering drawings.

[0056] Subsequently, based on the rail transit knowledge graph and the information to be reviewed, the engineering drawings are reviewed to obtain the review results of the engineering drawings.

[0057] Here, based on the rail transit knowledge graph and the information to be reviewed, a drawing review is performed on the engineering drawings to obtain the drawing review results of the engineering drawings, including: based on the drawing type, determining the target knowledge graph from the rail transit knowledge graph, and based on the target knowledge graph and the information to be reviewed, a drawing review is performed on the engineering drawings to obtain the drawing review results of the engineering drawings.

[0058] Among them, based on the target knowledge graph and the information to be reviewed, the engineering drawings are reviewed to obtain the drawing review results of the engineering drawings, including: based on the information to be reviewed, determining entity triples from the target knowledge graph, the entity triples containing multiple target entities and the entity relationships between the target entities; based on the entity triples and the information to be reviewed, the engineering drawings are reviewed to obtain the drawing review results of the engineering drawings.

[0059] Furthermore, based on the rail transit knowledge graph and the information to be reviewed, a drawing review is performed on the engineering drawings to obtain the drawing review results of the engineering drawings, which then includes: when the drawing review results indicate that the engineering drawings are non-compliant, determining the non-compliance of the engineering drawings, and the items and contents in the engineering implementation standards and / or engineering drawing specifications in the rail transit field that are violated by the non-compliant items; displaying the non-compliance, and the items and contents in the engineering implementation standards and / or engineering drawing specifications in the rail transit field that are violated by the non-compliant items.

[0060] Among them, the rail transit knowledge graph is determined based on the following steps: entity extraction is performed on expert review experience, as well as engineering execution standards and / or engineering drawing specifications to obtain multiple entities; entities include graphic entities and text entities; relationship extraction is performed based on each entity, as well as engineering execution standards and / or engineering drawing specifications to obtain entity relationships between entities; based on each entity and the entity relationships between entities, a rail transit knowledge graph is constructed.

[0061] Entity extraction is performed on the expert review experience, as well as the engineering implementation standards and / or engineering drawing specifications to obtain multiple entities; relationship extraction is performed based on each entity, as well as the engineering implementation standards and / or engineering drawing specifications to obtain entity relationships between each entity, including: key points are extracted from the expert review experience to obtain expert review key points; based on the expert review key points, key points of the engineering implementation standards and / or engineering drawing specifications are screened to obtain target implementation standards and / or target drawing specifications; entity extraction is performed on the expert review key points, as well as the target implementation standards and / or target drawing specifications to obtain multiple entities; relationship extraction is performed based on each entity, as well as the target implementation standards and / or target drawing specifications to obtain entity relationships between each entity.

[0062] The intelligent drawing review method based on knowledge graph provided by the present invention reviews the information to be reviewed in the engineering drawings through the rail transit knowledge graph constructed on the basis of the expert review experience in the rail transit field, engineering execution standards and / or engineering drawing specifications, and obtains the drawing review results of the engineering drawings, thereby realizing efficient and automated review of engineering drawings in the rail transit field, overcoming the defects of long manual review cycle and low efficiency in traditional solutions, and the general review method being unsuitable for drawings in the rail transit field and poor review effect, not only ensuring the accuracy and effectiveness of the review, but also greatly improving the review efficiency; in addition, it can also effectively assist review experts in conducting highly complex drawing reviews, saving a lot of manpower and material costs, and providing guarantees for the safety and quality of engineering construction in the rail transit field.

[0063] The following is a description of the knowledge graph-based intelligent image review device provided by the present invention. The knowledge graph-based intelligent image review device described below and the knowledge graph-based intelligent image review method described above can refer to each other.

[0064] Figure 4 is a schematic diagram of the structure of the intelligent image review device based on the knowledge graph provided by the present invention, such as Figure 4 As shown, the device comprises: An acquisition unit 410 is used to acquire engineering drawings to be reviewed in the field of rail transit; The parsing unit 420 is used to parse the engineering drawing to obtain the information to be reviewed in the engineering drawing, where the information to be reviewed includes graphics, text, and the corresponding relationship between the graphics and the text; The review unit 430 is used to review the engineering drawings based on the rail transit knowledge graph and the information to be reviewed, and obtain the review results of the engineering drawings; The rail transit knowledge graph is constructed based on the expert drawing review experience in the rail transit field, as well as engineering execution standards and / or engineering drawing specifications.

[0065] The intelligent drawing review device based on knowledge graph provided by the present invention reviews the information to be reviewed in the engineering drawings through the rail transit knowledge graph constructed on the basis of the expert review experience in the rail transit field, and the engineering execution standards and / or engineering drawing specifications, and obtains the drawing review results of the engineering drawings, thereby realizing efficient and automated review of engineering drawings in the rail transit field, overcoming the defects of the traditional scheme of long manual review cycle and low efficiency, and the general review method being unsuitable for drawings in the rail transit field and poor review effect, not only ensuring the accuracy and effectiveness of the review, but also greatly improving the review efficiency; in addition, it can also effectively assist the review experts in conducting highly complex drawing reviews, saving a lot of manpower and material costs, and providing guarantee for the safety and quality of engineering construction in the rail transit field.

[0066] Based on the above embodiment, the parsing unit 420 is used for: Parsing the engineering calculation sheet and / or design drawing in the engineering drawing to obtain the drawing type of the engineering drawing and the information to be reviewed in the engineering drawing; The audit unit 430 is used for: Based on the drawing type, a target knowledge graph is determined from the rail transit knowledge graph, and based on the target knowledge graph and the information to be reviewed, a drawing review is performed on the engineering drawings to obtain a drawing review result of the engineering drawings.

[0067] Based on the above embodiment, the review unit 430 is used for: Based on the information to be reviewed, determining an entity triple from the target knowledge graph, wherein the entity triple includes multiple target entities and entity relationships between the target entities; Based on the entity triples and the information to be reviewed, the engineering drawings are reviewed to obtain the review results of the engineering drawings.

[0068] Based on the above embodiment, the device further includes a display unit, which is used to: In the case where the drawing review result indicates that the engineering drawing is non-compliant, determining the non-compliant part of the engineering drawing, and the items and contents of the engineering execution standards and / or engineering drawing specifications in the field of rail transit that the non-compliant part violates; Display the non-compliance, as well as the items and contents of the engineering execution standards and / or engineering drawing specifications in the field of rail transit that the non-compliance violates

[0069] Based on the above embodiment, the device further includes a knowledge graph determination unit, which is used to: Perform entity extraction on the expert drawing review experience, the engineering execution standard and / or the engineering drawing specification to obtain a plurality of entities; the entities include graphic entities and text entities; Extracting relationships based on each entity and the engineering execution standard and / or the engineering drawing specification to obtain entity relationships between the entities; Based on the entities and the entity relationships between the entities, a rail transit knowledge graph is constructed.

[0070] Based on the above embodiment, the knowledge graph determination unit is used to: Extract the key points of the expert review experience to obtain the key points of the expert review; Based on the expert review key points, the engineering implementation standard and / or the engineering drawing specification are screened to obtain the target implementation standard and / or the target drawing specification; Perform entity extraction on the expert review key points, the target execution standard and / or the target mapping specification to obtain multiple entities; Relationship extraction is performed based on the entities, the target execution standard and / or the target mapping specification to obtain entity relationships between the entities.

[0071] Figure 5 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 5 As shown, the electronic device may include: a processor 510, a communication interface 520, a memory 530 and a communication bus 540, wherein the processor 510, the communication interface 520 and the memory 530 communicate with each other through the communication bus 540. The processor 510 may call the logic instructions in the memory 530 to execute the intelligent drawing review method based on the knowledge graph, the method comprising: obtaining engineering drawings to be reviewed in the field of rail transit; parsing the engineering drawings to obtain the information to be reviewed in the engineering drawings, the information to be reviewed including graphics, text, and the corresponding relationship between the graphics and the text; based on the rail transit knowledge graph and the information to be reviewed, the engineering drawings are reviewed to obtain the drawing review results of the engineering drawings; the rail transit knowledge graph is constructed based on the expert review experience in the field of rail transit, as well as the engineering execution standards and / or engineering drawing specifications.

[0072] In addition, the logic instructions in the above-mentioned memory 530 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0073] On the other hand, the present invention also provides a computer program product, which includes a computer program stored on a non-transitory computer-readable storage medium, and the computer program includes program instructions. When the program instructions are executed by a computer, the computer can execute the intelligent drawing review method based on the knowledge graph provided by the above-mentioned methods, and the method includes: obtaining engineering drawings to be reviewed in the field of rail transit; parsing the engineering drawings to obtain information to be reviewed in the engineering drawings, and the information to be reviewed includes graphics, text, and the correspondence between the graphics and the text; based on the rail transit knowledge graph and the information to be reviewed, performing a drawing review on the engineering drawings to obtain the drawing review results of the engineering drawings; the rail transit knowledge graph is constructed based on the expert review experience in the rail transit field, as well as engineering execution standards and / or engineering drawing specifications.

[0074] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the intelligent drawing review method based on the knowledge graph provided by the above-mentioned methods, the method comprising: obtaining engineering drawings to be reviewed in the field of rail transit; parsing the engineering drawings to obtain information to be reviewed in the engineering drawings, the information to be reviewed including graphics, text, and the correspondence between the graphics and the text; based on the rail transit knowledge graph and the information to be reviewed, performing a drawing review on the engineering drawings to obtain the drawing review results of the engineering drawings; the rail transit knowledge graph is constructed based on the expert review experience in the rail transit field, as well as engineering execution standards and / or engineering drawing specifications.

[0075] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0076] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0077] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An intelligent image review method based on knowledge graph, characterized in that: include: Obtain engineering drawings to be reviewed in the field of rail transit; Parsing the engineering drawings to obtain the information to be reviewed in the engineering drawings, wherein the information to be reviewed includes graphics, text, and a corresponding relationship between the graphics and the text; Based on the rail transit knowledge graph and the information to be reviewed, the engineering drawings are reviewed to obtain the review results of the engineering drawings; The rail transit knowledge graph is constructed based on the expert drawing review experience in the rail transit field, as well as engineering execution standards and / or engineering drawing specifications.

2. The intelligent image review method based on knowledge graph according to claim 1 is characterized in that: The step of parsing the engineering drawings to obtain the information to be reviewed in the engineering drawings includes: Parsing the engineering calculation sheet and / or design drawing in the engineering drawing to obtain the drawing type of the engineering drawing and the information to be reviewed in the engineering drawing; The step of performing drawing review on the engineering drawings based on the rail transit knowledge graph and the information to be reviewed to obtain drawing review results of the engineering drawings includes: Based on the drawing type, a target knowledge graph is determined from the rail transit knowledge graph, and based on the target knowledge graph and the information to be reviewed, a drawing review is performed on the engineering drawings to obtain a drawing review result of the engineering drawings.

3. The intelligent image review method based on knowledge graph according to claim 2 is characterized in that: The step of performing a drawing review on the engineering drawing based on the target knowledge graph and the information to be reviewed to obtain a drawing review result of the engineering drawing includes: Based on the information to be reviewed, determining an entity triple from the target knowledge graph, wherein the entity triple includes multiple target entities and entity relationships between the target entities; Based on the entity triples and the information to be reviewed, the engineering drawings are reviewed to obtain the review results of the engineering drawings.

4. The intelligent drawing review method based on knowledge graph according to any one of claims 1 to 3, characterized in that: Based on the rail transit knowledge graph and the information to be reviewed, the engineering drawings are reviewed to obtain the review results of the engineering drawings, and then the following steps are further included: In the case where the drawing review result indicates that the engineering drawing is non-compliant, determining the non-compliant part of the engineering drawing, and the items and contents of the engineering execution standards and / or engineering drawing specifications in the field of rail transit that the non-compliant part violates; The non-compliance is displayed, as well as the items and contents in the engineering execution standards and / or engineering drawing specifications in the rail transit field that are violated by the non-compliance.

5. The intelligent drawing review method based on knowledge graph according to any one of claims 1 to 3, characterized in that: The rail transit knowledge graph is determined based on the following steps: Perform entity extraction on the expert drawing review experience, the engineering execution standard and / or the engineering drawing specification to obtain a plurality of entities; the entities include graphic entities and text entities; Extracting relationships based on each entity and the engineering execution standard and / or the engineering drawing specification to obtain entity relationships between the entities; Based on the entities and the entity relationships between the entities, a rail transit knowledge graph is constructed.

6. The intelligent image review method based on knowledge graph according to claim 5 is characterized in that: The entity extraction of the expert drawing review experience, the engineering execution standard and / or the engineering drawing specification is performed to obtain multiple entities, including: Extract the key points of the expert review experience to obtain the key points of the expert review; Based on the expert review key points, the engineering implementation standard and / or the engineering drawing specification are screened to obtain the target implementation standard and / or the target drawing specification; Perform entity extraction on the expert review key points, the target execution standard and / or the target mapping specification to obtain multiple entities; The extracting of relationships based on each entity and the engineering execution standard and / or the engineering drawing specification to obtain entity relationships between the entities includes: Relationship extraction is performed based on the entities, the target execution standard and / or the target mapping specification to obtain entity relationships between the entities.

7. An intelligent image review device based on knowledge graph, characterized in that: include: An acquisition unit, used to acquire engineering drawings to be reviewed in the field of rail transit; A parsing unit, used to parse the engineering drawings to obtain the information to be reviewed in the engineering drawings, wherein the information to be reviewed includes graphics, text, and a corresponding relationship between the graphics and the text; An audit unit, configured to perform a drawing audit on the engineering drawing based on the rail transit knowledge graph and the information to be audited, and obtain a drawing audit result of the engineering drawing; The rail transit knowledge graph is constructed based on the expert drawing review experience in the rail transit field, as well as engineering execution standards and / or engineering drawing specifications.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the intelligent image review method based on the knowledge graph is implemented as described in any one of claims 1 to 6.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the intelligent image review method based on the knowledge graph is implemented as described in any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the intelligent image review method based on the knowledge graph is implemented as described in any one of claims 1 to 6.

Images