Engineering drawing recognition and automated inspection system based on deep learning
Through the deep learning-based engineering drawing identification and automation inspection system, the inaccurate inspection conditions caused by incomplete analysis of engineering drawing content in the existing technology are solved, the accuracy of identification and inspection is improved, and the correct implementation of engineering tasks and projects are ensured. Management efficiency.
Patent Information
- Application Number
- CN202411103151.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-13
- Publication Date
- 2024-11-22
- Estimated Expiration
- 2044-08-13
AI Technical Summary
The prior art has incomplete content analysis in engineering drawing identification and inspection, resulting in inaccurate inspection conditions, which reduces the accuracy of identification and inspection.
Adopt deep learning-based engineering drawing identification and automation inspection system, including drawing database, complete condition limiting module, drawing integrity inspection module, calibration condition limiting module, drawing association module, drawing verification module and adjustment module, and engineering drawings The content is analyzed in detail, the element integrity, special label integrity and correlation level are determined, the identification process parameters are adjusted, and the inspection is ensured.
It improves the accuracy of the analysis of the content of the engineering drawings, ensures the accuracy of the inspection conditions, reduces potential errors and accidents, improves the correct implementation of engineering tasks, and enhances the management efficiency and design consistency of the engineering project.
Smart Images

Figure CN118629066B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of drawing recognition and inspection, and in particular to an engineering drawing recognition and automated inspection system based on deep learning. Background Art
[0002] The engineering drawing recognition and automated inspection system is a system that uses computer vision and image processing technology to realize automated recognition and inspection of engineering drawings. Traditional automated recognition and inspection technologies for engineering drawings include using image processing and analysis technology to preprocess and extract features from engineering drawings, recognize and classify symbols in engineering drawings, and then automatically inspect the dimensions and geometric requirements in engineering drawings. However, traditional technologies cannot accurately analyze engineering drawings to determine the corresponding inspection conditions, thereby reducing the efficiency of engineering drawing recognition and inspection.
[0003] Chinese patent application publication number: CN111401231A discloses a method and device for identifying and reviewing engineering drawings, which converts imported initial format drawings into standard format drawings; parses standard format drawings to determine the graphic object set corresponding to the standard format drawings; uses a preset graphic filtering algorithm to filter the graphic object set according to the drawing type of the standard format drawings to determine the graphic object set to be reviewed; uses a preset graphic completion algorithm to verify the connection relationship of all graphic objects to be reviewed in the graphic object set to be reviewed according to the drawing type, and obtains automatic modification information; modifies the graphic object set to be reviewed according to the automatic modification information to obtain a standardized graphic object set; uses a pre-acquired building specification library to verify and review the standardized graphic object set to obtain the review result, so that the client can obtain and display the review result. The use of this technical solution can save time for reviewing drawings and improve review efficiency and review accuracy.
[0004] It can be seen that the existing technology has the problem of incomplete analysis of the content of engineering drawings, resulting in inaccurate inspection conditions for matching engineering drawings and low accuracy of engineering drawing recognition and inspection. Summary of the invention
[0005] To this end, the present invention provides an engineering drawing recognition and automated inspection system based on deep learning, so as to overcome the problem in the prior art that incomplete analysis of the contents of engineering drawings results in inaccurate inspection conditions for matching engineering drawings, resulting in low accuracy in engineering drawing recognition and inspection.
[0006] To achieve the above objectives, the present invention provides an engineering drawing recognition and automated inspection system based on deep learning, comprising:
[0007] Drawing database, used to acquire and store engineering drawings;
[0008] a complete condition qualification module, which is connected to the drawing database and includes a plurality of drawing completeness conditions and thresholds corresponding to the drawing completeness conditions, and is used to perform image / data recognition on engineering drawings, determine drawing information according to identification information of a single engineering drawing, and determine the completeness conditions corresponding to the drawings according to the drawing information, wherein the drawing completeness conditions include element completeness conditions and special annotation completeness conditions;
[0009] A drawing integrity checking module, which is used to determine whether the drawing is complete according to the integrity condition corresponding to the drawing;
[0010] A verification condition definition module, which is connected to the drawing integrity verification module, includes a number of drawing verification conditions, and determines corresponding verification conditions according to drawing information, wherein the drawing verification conditions include standard part / standard part verification conditions, associated adaptation verification conditions and single drawing adaptation verification conditions;
[0011] A drawing association module, which is connected to the verification condition definition module, is used to determine the associated drawings of the engineering drawings according to the drawing information, determine the association level of the associated drawings, and determine the consistency content of the engineering drawings according to the association level;
[0012] A drawing verification module, which is connected to the drawing association module, is used to determine whether to perform drawing association according to the integrity condition determination result of the drawing to be verified so as to use the consistency content to supplement the drawing content, and to determine the verification result of the drawing to be verified based on the integrity condition determination result;
[0013] An adjustment module, which is connected to the drawing verification module and is used to determine the adjustment amount of the drawing recognition process parameters according to the proportion of the accurate number of drawing verification results;
[0014] The drawing recognition process parameters include the number of preset element types and the number of preset special annotations.
[0015] Furthermore, the complete condition limitation module determines the special annotation integrity condition corresponding to the engineering drawing according to the presence of warning words in the content of the engineering drawing, and the complete condition limitation module determines the element integrity condition corresponding to the engineering drawing according to the absence of warning words in the content of the engineering drawing.
[0016] Furthermore, the drawing integrity verification module determines the integrity of the elements of the engineering drawing based on a comparison result of the number of element types in the engineering drawing and the number of preset element types, under the condition of determining the integrity of the corresponding elements of the engineering drawing. The drawing integrity verification module determines the integrity of the elements of the engineering drawing based on whether there is unknown data in the engineering drawing.
[0017] Furthermore, the drawing integrity verification module determines whether the special markings of the engineering drawing are complete based on the element integrity of the engineering drawing and the comparison result between the number of special markings and a preset number of special markings, under the condition of determining the integrity of the special markings corresponding to the engineering drawing.
[0018] Furthermore, the preset number of element types is determined according to an average value of element types included in a number of engineering drawings of the same type, and the preset number of special markings is determined according to the types of warning words in the drawing description and the number of positions corresponding to the warning words included in the engineering drawings.
[0019] Further, the drawing association module determines the association level based on whether other engineering drawings contain the same project identifier or prefix as the engineering drawing and other engineering drawing numbers appearing in the engineering drawing, under the condition that the engineering drawing is complete, wherein:
[0020] The drawing association module determines that the engineering drawing has no associated drawings based on the absence of other engineering drawings containing the same project identifier or prefix as the engineering drawing and the absence of other engineering drawing numbers appearing in the engineering drawing;
[0021] The drawing association module determines that the association level between the other engineering drawings and the engineering drawing is a first association level according to the existence of other engineering drawings containing the same project identifier or prefix as the engineering drawing;
[0022] The drawing association module determines that the association level between the other engineering drawings and the engineering drawings is a second association level according to the numbers of the other engineering drawings that appear in the engineering drawings.
[0023] Furthermore, the drawing verification module determines the verification conditions for standard parts / specification parts corresponding to the engineering drawing and the single drawing adaptation verification conditions under the condition that it is determined that there are no associated drawings for the engineering drawing. The drawing verification module determines the verification conditions for standard parts / specification parts corresponding to the engineering drawing, the single drawing adaptation verification conditions and the associated adaptation verification conditions under the condition that it is determined that there are associated drawings for the engineering drawing.
[0024] Furthermore, the drawing verification module uses the consistency content to supplement the drawing content when determining that the engineering drawing is incomplete.
[0025] Furthermore, the adjustment module determines the adjustment amount of the drawing recognition process parameters according to the proportion of the drawing inspection results, wherein the drawing recognition process parameters include the number of preset element types and the number of preset special annotations.
[0026] Furthermore, the adjustment amount of the preset number of element types is negatively correlated with the accurate proportion of the drawing inspection results, and the adjustment amount of the preset number of special annotations is negatively correlated with the accurate proportion of the drawing inspection results.
[0027] Compared with the prior art, the beneficial effect of the present invention is that the present invention determines the integrity conditions corresponding to the engineering drawings by whether the engineering drawings contain warning words involving safety issues, dangerous operations or requiring special attention. If there are warning words in the engineering drawings, it means that there are operating instructions requiring special attention in the engineering drawings. At this time, it is necessary not only to ensure that the elements of the engineering drawings are complete, but also to ensure that the special markings are complete. The above method improves the accuracy of the analysis of the content of the engineering drawings and thus improves the accuracy of the inspection conditions for matching the engineering drawings, ensuring the correct execution of engineering tasks and reducing potential errors or accidents.
[0028] Furthermore, the present invention determines the integrity of the elements of the engineering drawing according to the number of element types in the engineering drawing and whether there is any unknown data. If the number of element types in the engineering drawing is greater than or equal to a preset number of types and there is no unknown data, the engineering drawing elements are determined to be complete. The above method improves the accuracy of the analysis of the content of the engineering drawings and thus improves the accuracy of the inspection conditions for matching the engineering drawings, thereby ensuring the correct execution of engineering tasks and reducing potential errors or accidents.
[0029] Furthermore, when determining the completeness of special markings on engineering drawings, the present invention first determines that the elements of the engineering drawings are complete, and then determines the completeness of the special markings on the engineering drawings based on a comparison result between the number of special markings on the engineering drawings and a preset number of special markings, wherein the preset number of special markings is the sum of the number of positions corresponding to warning words in the engineering drawings and the number of warning words that do not exist in the corresponding positions. The above method avoids the phenomenon of subsequent construction errors caused by the lack of warning words at the corresponding positions where warning words exist in the engineering drawings. By comprehensively analyzing the contents of the engineering drawings, the accuracy of the inspection conditions for matching the engineering drawings is improved, thereby ensuring the correct execution of engineering tasks and reducing potential errors or accidents.
[0030] Furthermore, the present invention determines the association level based on other engineering drawings containing the same project identifier or prefix as the engineering drawing and other engineering drawing numbers appearing in the engineering drawing, so that the relationship between the engineering drawings can be established, which is convenient for managing and tracking the drawings of the entire engineering project, which helps to ensure the consistency, coordination and integrity between the drawings. When a design change occurs, the affected drawings and associated drawings can be quickly determined, which helps to carry out effective change management and ensure that all relevant drawings are updated in a timely manner to reflect the content of the design change. Understanding the association level between engineering drawings can help determine which drawings are references or references to other drawings, which is very important for ensuring correct drawing references and references, and avoiding conflicts and errors between drawings. The drawing integrity inspection module can help ensure that the engineering drawings meet quality standards and audit requirements. Potential problems or inconsistencies can be discovered by checking the association level. Associating engineering drawing numbers enables the source and evolution of drawings to be traced, which is of great significance for quality traceability, problem solving and project history. Determining the association level of engineering drawings helps to improve the management efficiency of engineering projects, reduce errors, ensure design consistency, and provide better support and guarantee for subsequent work.
[0031] Furthermore, the present invention can obtain missing parts of the content by associating other complete drawings or related information, making the engineering drawings more complete and reducing missing or inaccurate information. The consistent content is usually reviewed or verified, and using it to supplement drawings can ensure the consistency between drawings and avoid contradictions or inconsistencies. Using the associated consistent content can avoid repeatedly entering the same information on different drawings, reduce the occurrence of errors, and improve work efficiency. Complete and consistent drawings help improve design quality and reduce problems in subsequent construction and operation.
[0032] Furthermore, the present invention can evaluate the accuracy of the current drawing recognition process by comparing the accurate proportion of drawing inspection results with the preset proportion, and make corresponding adjustments as needed. If the accurate proportion of drawing inspection results is lower than the preset proportion, it means that there are errors or inaccuracies in drawing recognition, and the parameters need to be adjusted to improve the recognition accuracy. The adjustment amount of the drawing recognition process parameters is determined based on the comparison result of the accurate proportion of drawing inspection results with the preset proportion, which helps to ensure the accuracy and reliability of drawing recognition, adapt to changing needs, and improve production efficiency and project quality. BRIEF DESCRIPTION OF THE DRAWINGS
[0033] Figure 1 It is a structural schematic diagram of an engineering drawing recognition and automated inspection system based on deep learning according to an embodiment of the present invention;
[0034] Figure 2This is a workflow diagram of a complete condition limitation module in an engineering drawing recognition and automated inspection system based on deep learning according to an embodiment of the present invention;
[0035] Figure 3 This is a workflow diagram of the adjustment module in the engineering drawing recognition and automated inspection system based on deep learning in an embodiment of the present invention. DETAILED DESCRIPTION
[0036] In order to make the objects and advantages of the present invention more clearly understood, the present invention is further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.
[0037] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are only used to explain the technical principles of the present invention and are not intended to limit the protection scope of the present invention.
[0038] It should be noted that, in the description of the present invention, terms such as "up", "down", "left", "right", "inside" and "outside" indicating directions or positional relationships are based on the directions or positional relationships shown in the drawings. This is merely for the convenience of description and does not indicate or imply that the device or element must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it cannot be understood as a limitation on the present invention.
[0039] In addition, it should be noted that in the description of the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For those skilled in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.
[0040] See also Figure 1-Figure 3 As shown, Figure 1 It is a structural schematic diagram of an engineering drawing recognition and automated inspection system based on deep learning according to an embodiment of the present invention; Figure 2 This is a workflow diagram of a complete condition limitation module in an engineering drawing recognition and automated inspection system based on deep learning according to an embodiment of the present invention; Figure 3 This is a workflow diagram of the adjustment module in the engineering drawing recognition and automated inspection system based on deep learning in an embodiment of the present invention.
[0041] The engineering drawing recognition and automated inspection system based on deep learning in the embodiment of the present invention includes:
[0042] Drawing database, used to acquire and store engineering drawings;
[0043] a complete condition qualification module, which is connected to the drawing database and includes a plurality of drawing completeness conditions and thresholds corresponding to the drawing completeness conditions, and is used to perform image / data recognition on engineering drawings, determine drawing information according to identification information of a single engineering drawing, and determine the completeness conditions corresponding to the drawings according to the drawing information, wherein the drawing completeness conditions include element completeness conditions and special annotation completeness conditions;
[0044] A drawing integrity checking module, which is used to determine whether the drawing is complete according to the integrity condition corresponding to the drawing;
[0045] A verification condition definition module, which is connected to the drawing integrity verification module, includes a number of drawing verification conditions, and determines corresponding verification conditions according to drawing information, wherein the drawing verification conditions include standard part / standard part verification conditions, associated adaptation verification conditions and single drawing adaptation verification conditions;
[0046] A drawing association module, which is connected to the verification condition definition module, is used to determine the associated drawings of the engineering drawings according to the drawing information, determine the association level of the associated drawings, and determine the consistency content of the engineering drawings according to the association level;
[0047] A drawing verification module, which is connected to the drawing association module, is used to determine whether to perform drawing association according to the integrity condition determination result of the drawing to be verified so as to use the consistency content to supplement the drawing content, and to determine the verification result of the drawing to be verified based on the integrity condition determination result;
[0048] An adjustment module, which is connected to the drawing verification module and is used to determine the adjustment amount of the drawing recognition process parameters according to the proportion of the accurate number of drawing verification results;
[0049] Specifically, the complete condition limiting module determines the drawing completeness condition corresponding to the engineering drawing according to whether there is a warning word in the engineering drawing content;
[0050] If there is no warning word in the engineering drawing content, the complete condition limiting module determines the completeness condition of the element corresponding to the engineering drawing;
[0051] If there are warning words in the engineering drawing content, the complete condition limiting module determines that the engineering drawing corresponds to a special marking completeness condition.
[0052] The warning words in the embodiments of the present invention include but are not limited to "caution, special instructions, and text that is different in color from the graphics and general instructions in the engineering drawings." For example, if the graphics in the engineering drawings are black and most of the instructions in the drawings are black, then the text that is different in color from black is a warning word.
[0053] The present invention determines the integrity condition corresponding to the engineering drawing by whether the engineering drawing contains warning words involving safety issues, dangerous operations or warning words requiring special attention. If there are warning words in the engineering drawing, it means that there are operating instructions requiring special attention in the engineering drawing. At this time, it is necessary not only to ensure that the elements of the engineering drawing are complete, but also to ensure that the special markings are complete. The above method improves the accuracy of the analysis of the content of the engineering drawing and thus improves the accuracy of the inspection conditions for matching the engineering drawings, ensures the correct execution of engineering tasks, and reduces potential errors or accidents.
[0054] Specifically, the complete condition limiting module determines the integrity of the corresponding elements of the engineering drawing, and the drawing integrity checking module determines whether the elements of the engineering drawing are complete according to the comparison result between the number of element types N of the engineering drawing and the preset number of element types N0;
[0055] If N≥N0, the drawing integrity verification module performs a secondary determination on the integrity of the engineering drawing elements;
[0056] If N<N0, the drawing integrity inspection module determines that the engineering drawing element is incomplete;
[0057] The value of the preset number of element types N0 is the average value of the element types contained in several engineering drawings of the same type, but the above value is not limited thereto, and those skilled in the art may also adjust the value according to actual needs.
[0058] Specifically, under the condition that the drawing integrity verification module performs a secondary determination on the integrity of the engineering drawing elements, the drawing integrity verification module determines whether the engineering drawing elements are complete according to whether there is unknown data in the engineering drawing;
[0059] If the engineering drawing contains unknown data, the drawing integrity verification module determines that the engineering drawing element is incomplete;
[0060] If the engineering drawing does not contain unknown data, the drawing integrity verification module determines that the elements of the engineering drawing are complete.
[0061] The unknown data in the embodiment of the present invention includes but is not limited to "dimensional values, component values and process parameters".
[0062] The present invention determines the integrity of the elements of the engineering drawing according to the number of element types in the engineering drawing and whether there is any unknown data. If the number of element types in the engineering drawing is greater than or equal to a preset number of types and there is no unknown data, the engineering drawing elements are determined to be complete. The above method improves the accuracy of the analysis of the content of the engineering drawing and thus improves the accuracy of the inspection conditions for matching the engineering drawings, ensures the correct execution of engineering tasks, and reduces potential errors or accidents.
[0063] Specifically, the complete condition limiting module determines the completeness of the special annotation corresponding to the engineering drawing, and the drawing completeness checking module determines whether the special annotation of the engineering drawing is complete according to the element completeness of the engineering drawing;
[0064] If the elements of the engineering drawing are complete, the drawing integrity verification module performs a secondary determination on the completeness of the special markings on the engineering drawing;
[0065] If the elements of the engineering drawing are complete, the drawing integrity verification module determines that the special markings of the engineering drawing are incomplete.
[0066] Specifically, under the condition that the drawing integrity inspection module performs a secondary determination on whether the special markings of the engineering drawing are complete, the drawing integrity inspection module determines whether the special markings of the engineering drawing are complete according to the comparison result between the special marking quantity M of the engineering drawing and the preset special marking quantity M0;
[0067] If M≥M0, the drawing integrity inspection module determines that the special markings on the engineering drawing are complete;
[0068] If M<M0, the drawing integrity inspection module determines that the special marking of the engineering drawing is incomplete.
[0069] The preset number of special markings in the embodiment of the present invention is the sum of the number of positions corresponding to the warning words in the engineering drawings and the number of warning words without corresponding positions. For example, when the warning words in the engineering drawings include a power supply voltage requirement of 200V-220V and a temperature range of components of 20℃-30℃, and the number of components included in the engineering drawings is 4, then the preset number of special markings is 5.
[0070] When determining the completeness of special markings on engineering drawings, the present invention first determines that the elements of the engineering drawings are complete, and then determines the completeness of the special markings on the engineering drawings based on a comparison result between the number of special markings on the engineering drawings and a preset number of special markings, wherein the preset number of special markings is the sum of the number of positions corresponding to warning words in the engineering drawings and the number of warning words that do not exist in the corresponding positions. The above method avoids the phenomenon of subsequent construction errors caused by the lack of warning words at the corresponding positions where warning words exist in the engineering drawings, improves the accuracy of the inspection conditions for matching the engineering drawings through a comprehensive analysis of the contents of the engineering drawings, ensures the correct execution of engineering tasks, and reduces potential errors or accidents.
[0071] Specifically, the drawing integrity inspection module determines the association level based on whether other engineering drawing numbers appear in the engineering drawing under the condition that the engineering drawing is complete;
[0072] If other engineering drawing numbers appear in the engineering drawing, the drawing association module determines that the association level between the other engineering drawings and the engineering drawing is the second association level;
[0073] If no other engineering drawing numbers appear in the engineering drawing, the drawing association module performs a secondary determination on the association level.
[0074] Specifically, the drawing association module determines the association level based on whether the other engineering drawings contain the same project identifier or prefix as the engineering drawing under the condition that the drawing association module performs a secondary determination on the determined association level;
[0075] If the other engineering drawings do not contain the same project identifier or prefix as the engineering drawing, the drawing association module determines that there is no associated drawing with the engineering drawing;
[0076] If the other engineering drawings contain the same project identifier or prefix as the engineering drawing, the drawing association module determines that the association level between the other engineering drawings and the engineering drawing is a first association level.
[0077] The same project identifier or prefix described in the embodiment of the present invention includes but is not limited to "AMP001 and AMP002, PRJ-2021-001 and PRJ-2021-002, ABC Electronics and BBC Electronics".
[0078] Specifically, when the drawing association module determines that there are no associated drawings for the engineering drawing, the drawing verification module determines the verification conditions for the standard parts / specification parts corresponding to the engineering drawing and the single drawing adaptation verification conditions. When the drawing association module determines that there are associated drawings for the engineering drawing, the drawing verification module determines the verification conditions for the standard parts / specification parts corresponding to the engineering drawing, the single drawing adaptation verification conditions and the associated adaptation verification conditions.
[0079] The standard parts / standard parts verification conditions in the embodiment of the present invention include but are not limited to "size and specifications: verify whether the size, aperture, pitch, etc. of the standard parts / standard parts meet the requirements of the engineering drawings, material and performance: verify whether the material of the standard parts / standard parts meets the requirements of the engineering drawings, and whether its mechanical properties, physical properties, etc. meet the use requirements, surface quality: verify whether the surface quality of the standard parts / standard parts, such as roughness, defects, etc., meet the requirements of the drawings, standard compliance: verify whether the standard parts / standard parts meet the corresponding standards and specifications, such as national standards and industry standards"; the single image adaptation verification conditions include but are not limited to "verify whether the clarity, contrast, color, etc. of the image meet the requirements to ensure that the image can clearly convey information, size and proportion: verify the image Whether the size and proportion are consistent with the design requirements to ensure that the image can be displayed correctly in actual applications, content accuracy: verify whether the content of the image accurately reflects the actual situation to avoid erroneous or misleading information"; the association adaptation verification conditions include but are not limited to "verifying the consistency between related data to ensure the accuracy and completeness of related data, logical relationship: verify whether the logical relationship between related data is correct, such as parent-child relationship, subordinate relationship, etc., data range: check whether the value range of related data is reasonable and whether it complies with business rules, update synchronization: confirm whether the update of related data can be carried out synchronously to avoid data inconsistency, reference integrity: verify whether there are reference integrity issues in related data, such as whether foreign key constraints are correct.
[0080] The present invention determines the association level according to other engineering drawings containing the same project identifier or prefix as the engineering drawing and other engineering drawing numbers appearing in the engineering drawing, and can establish the relationship between the engineering drawings, so as to facilitate the management and tracking of the drawings of the entire engineering project, which helps to ensure the consistency, coordination and integrity between the drawings. When a design change occurs, the affected drawings and associated drawings can be quickly determined, which helps to carry out effective change management and ensure that all relevant drawings are updated in a timely manner to reflect the content of the design change. Understanding the association level between engineering drawings can help determine which drawings are references or references to other drawings, which is very important for ensuring correct drawing references and references, and avoiding conflicts and errors between drawings. The drawing integrity inspection module can help ensure that the engineering drawings meet quality standards and audit requirements. Potential problems or inconsistencies can be discovered by checking the association level. The associated engineering drawing numbers enable the source and evolution of the drawings to be traced, which is of great significance for quality traceability, problem solving and project history records. Determining the association level of engineering drawings helps to improve the management efficiency of engineering projects, reduce errors, ensure design consistency, and provide better support and guarantee for subsequent work.
[0081] Specifically, under the condition that the drawing verification module determines that the engineering drawing is incomplete, the drawing verification module performs drawing association to use the consistent content to supplement the drawing content.
[0082] In an embodiment of the present invention, using consistency content to supplement drawing content includes supplementing associated drawing content that does not exist in the drawing to the drawing. For example, if drawing A is associated with drawing B and drawing C, and the association level between drawing A and drawing B is the first association level, and the association level between drawing A and drawing C is the second association level, the size of the circuit board does not exist in drawing A, and both drawings B and drawings C have the size of the circuit board, then the size of the circuit board is supplemented to the drawing content. For example, if drawing A is associated with drawing B and drawing C, and the association level between drawing A and drawing B is the first association level, and the association level between drawing A and drawing C is the second association level, the size of the circuit board does not exist in drawing A, the size of the circuit board in drawing B is 50 cm, and the size of the circuit board in drawing C is 60 cm, then the size of the circuit board of 60 cm is supplemented to the drawing content.
[0083] The present invention can obtain missing parts of the content by associating other complete drawings or related information, making the engineering drawings more complete and reducing missing or inaccurate information. The consistent content is usually reviewed or verified, and using it to supplement drawings can ensure the consistency between drawings and avoid contradictions or inconsistencies. Using the associated consistent content can avoid repeatedly entering the same information on different drawings, reduce the occurrence of errors, and improve work efficiency. Complete and consistent drawings help improve design quality and reduce problems in subsequent construction and operation.
[0084] Specifically, the adjustment module determines the adjustment amount of the drawing recognition process parameter by comparing the accurate proportion P of the drawing inspection result with the preset proportion P0;
[0085] If P≥P0, the adjustment module determines that there is no need to adjust the drawing recognition process parameters;
[0086] If P<P0, the adjustment module determines to adjust the drawing recognition process parameters by the adjustment coefficient;
[0087] Among them, the value range of the preset proportion is set to 0.7-9, and the value of the preset proportion is preferably 0.85, but the above value is not limited to this, and technical personnel in this field can also adjust the value according to actual needs.
[0088] The drawing recognition process parameters in the embodiment of the present invention include a preset number of element types and a preset number of special annotations.
[0089] Specifically, the adjustment coefficient is calculated and set by the following formula:
[0090] ;
[0091] The adjusted number of preset element types N0' is set to N0'=K×N0.
[0092] The proportion of accurate drawing inspection results in the embodiment of the present invention is the ratio of the number of accurate drawing inspection results to the total number of inspected drawings.
[0093] The present invention can evaluate the accuracy of the current drawing recognition process by comparing the accurate proportion of drawing inspection results with the preset proportion, and make corresponding adjustments as needed. If the accurate proportion of drawing inspection results is lower than the preset proportion, it means that there is an error or inaccuracy in drawing recognition, and the parameters need to be adjusted to improve the recognition accuracy. The adjustment amount of the drawing recognition process parameters is determined according to the comparison result of the accurate proportion of drawing inspection results with the preset proportion, which helps to ensure the accuracy and reliability of drawing recognition, adapt to changing needs, and improve production efficiency and project quality.
[0094] So far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present invention.
[0095] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.
Claims
1. A deep learning-based engineering drawing recognition and automated inspection system, characterized in that: include: Drawing database, used to acquire and store engineering drawings; a complete condition qualification module, which is connected to the drawing database and includes a plurality of drawing completeness conditions and thresholds corresponding to the drawing completeness conditions, and is used to perform image / data recognition on engineering drawings, determine drawing information according to identification information of a single engineering drawing, and determine the completeness conditions corresponding to the drawings according to the drawing information, wherein the drawing completeness conditions include element completeness conditions and special annotation completeness conditions; A drawing integrity checking module, which is used to determine whether the drawing is complete according to the integrity condition corresponding to the drawing; A verification condition definition module, which is connected to the drawing integrity verification module, includes a number of drawing verification conditions, and determines corresponding verification conditions according to drawing information, wherein the drawing verification conditions include standard part / standard part verification conditions, associated adaptation verification conditions and single drawing adaptation verification conditions; A drawing association module, which is connected to the verification condition definition module, is used to determine the associated drawings of the engineering drawings according to the drawing information, determine the association level of the associated drawings, and determine the consistency content of the engineering drawings according to the association level; A drawing verification module, which is connected to the drawing association module, is used to determine whether to perform drawing association according to the integrity condition determination result of the drawing to be verified so as to use the consistency content to supplement the drawing content, and to determine the verification result of the drawing to be verified based on the integrity condition determination result; An adjustment module, which is connected to the drawing verification module and is used to determine the adjustment amount of the drawing recognition process parameters according to the proportion of the accurate number of drawing verification results; The drawing recognition process parameters include the number of preset element types and the number of preset special annotations.
2. The engineering drawing recognition and automated inspection system based on deep learning according to claim 1, characterized in that: The complete condition limitation module determines the special annotation integrity condition corresponding to the engineering drawing according to the presence of warning words in the engineering drawing content, and the complete condition limitation module determines the element integrity condition corresponding to the engineering drawing according to the absence of warning words in the engineering drawing content.
3. The engineering drawing recognition and automated inspection system based on deep learning according to claim 2, characterized in that: The drawing integrity verification module determines the integrity of the elements of the engineering drawing based on a comparison result of the number of element types in the engineering drawing and the number of preset element types, under the condition of determining the integrity of the corresponding elements of the engineering drawing. The drawing integrity verification module determines the integrity of the elements of the engineering drawing based on whether there is unknown data in the engineering drawing.
4. The engineering drawing recognition and automated inspection system based on deep learning according to claim 2, characterized in that: The drawing integrity verification module determines the integrity of the special annotations corresponding to the engineering drawing based on the element integrity of the engineering drawing and the comparison result of the special annotation quantity with the preset special annotation quantity, under the condition of determining the integrity of the special annotations corresponding to the engineering drawing.
5. The deep learning-based engineering drawing recognition and automated inspection system according to claim 3 or 4, characterized in that: The preset number of element types is determined according to the average value of element types included in a number of engineering drawings of the same type, and the preset number of special markings is determined according to the types of warning words in the drawing description and the number of positions corresponding to the warning words included in the engineering drawings.
6. The deep learning-based engineering drawing recognition and automated inspection system according to claim 5, characterized in that: The drawing association module determines the association level based on the fact that the engineering drawing is complete and that other engineering drawings contain the same project identifier or prefix as the engineering drawing and other engineering drawing numbers appearing in the engineering drawing, wherein: The drawing association module determines that the engineering drawing has no associated drawings based on the absence of other engineering drawings containing the same project identifier or prefix as the engineering drawing and the absence of other engineering drawing numbers appearing in the engineering drawing; The drawing association module determines that the association level between the other engineering drawings and the engineering drawing is a first association level according to the existence of other engineering drawings containing the same project identifier or prefix as the engineering drawing; The drawing association module determines that the association level between the other engineering drawings and the engineering drawings is a second association level according to the numbers of the other engineering drawings that appear in the engineering drawings.
7. The deep learning-based engineering drawing recognition and automated inspection system according to claim 6, characterized in that: The drawing verification module determines the verification conditions of the standard parts / standard parts corresponding to the engineering drawing and the single drawing adaptation verification conditions under the condition that it is determined that there are no associated drawings for the engineering drawing. The drawing verification module determines the verification conditions of the standard parts / standard parts corresponding to the engineering drawing, the single drawing adaptation verification conditions and the associated adaptation verification conditions under the condition that it is determined that there are associated drawings for the engineering drawing.
8. The deep learning-based engineering drawing recognition and automated inspection system according to claim 7, characterized in that: The drawing verification module uses the consistency content to supplement the drawing content when determining that the engineering drawing is incomplete.
9. The deep learning-based engineering drawing recognition and automated inspection system according to claim 8, characterized in that: The adjustment module determines the adjustment amount of the drawing recognition process parameters according to the proportion of the drawing verification results, wherein the drawing recognition process parameters include the number of preset element types and the number of preset special annotations.
10. The deep learning-based engineering drawing recognition and automated inspection system according to claim 9, characterized in that: The adjustment amount of the preset number of element types is negatively correlated with the accurate proportion of the drawing verification results, and the adjustment amount of the preset number of special annotations is negatively correlated with the accurate proportion of the drawing verification results.
