Methods, apparatus, electronic devices and storage media for cross-sectional view recognition

Abstract

This invention discloses a method, apparatus, device, and storage medium for cross-sectional view recognition. The method includes: determining a top-view cross-sectional view of the pile cap representing a top-down perspective from acquired cross-sectional views of the pile cap; inputting the top-view cross-sectional view of the pile cap into a section type judgment model to obtain the target section type of the top-view cross-sectional view of the pile cap; extracting rebar vector data representing rebar entities from the cross-sectional views of the pile cap from different perspectives; inputting the rebar vector data and the target section type corresponding to the rebar vector data into a rebar type judgment model to obtain the target rebar type of the rebar vector data; and determining the target rebar parameters corresponding to the rebar vector data based at least on the dimensional parameter information of the rebar vector data and the target rebar type. The technical solution provided by this invention can improve the efficiency of component recognition in cross-sectional views of pile caps to a certain extent.

Description

Technical Field

[0001] This invention relates to the field of image processing, and more specifically to a method, apparatus, electronic device, and storage medium for cross-sectional image recognition. Background Technology

[0002] A pile cap is a structure where, when a building uses a pile foundation, the tops of the piles are connected by a reinforced concrete platform or slab to form a monolithic foundation to bear the loads on top. Before constructing a pile cap, the pile cap components in the pile cap drawings need to be identified. In existing technologies, identifying reinforcement component information based on CAD or other vector sectional drawings requires manual reading of the reinforcement style and dimensional parameters from the sectional drawing, followed by matching the corresponding reinforcement component information. However, this manual method of identifying components in the pile cap sectional drawing is extremely time-consuming. Summary of the Invention

[0003] In view of this, embodiments of the present invention provide a method, apparatus, electronic device and storage medium for cross-sectional view recognition, which can improve the efficiency of component recognition in cross-sectional views of pile caps to a certain extent.

[0004] This invention provides a method for cross-sectional view recognition, the method comprising: determining a top-view cross-sectional view of the pile cap representing a top-down perspective from acquired cross-sectional views of the pile cap; inputting the top-view cross-sectional view of the pile cap into a section type judgment model to obtain a target section type of the top-view cross-sectional view of the pile cap; extracting rebar vector data representing rebar entities from the cross-sectional views of the pile cap from different perspectives; inputting the rebar vector data and the target section type corresponding to the rebar vector data into a rebar type judgment model to obtain a target rebar type of the rebar vector data; determining target rebar parameters corresponding to the rebar vector data based at least on the dimensional parameter information of the rebar vector data and the target rebar type; wherein the dimensional parameter information represents the rebar dimensional parameter information of the cross-sectional views of the pile cap from different perspectives; and the target rebar parameters represent the rebar component information corresponding to the cross-sectional views of the pile cap from different perspectives.

[0005] In another aspect, the present invention provides a cross-sectional view recognition device, comprising: a top-view pile cap cross-sectional view determination unit, used to determine a top-view pile cap cross-sectional view representing a top-view perspective from pile cap cross-sectional views acquired from different perspectives; a target section type determination unit, used to input the top-view pile cap cross-sectional view into a section type judgment model to obtain the target section type of the top-view pile cap cross-sectional view; a reinforcement vector data extraction unit, used to extract reinforcement vector data representing reinforcement entities from the pile cap cross-sectional views from different perspectives; a target reinforcement type determination unit, used to input the reinforcement vector data and the target section type corresponding to the reinforcement vector data into a reinforcement type judgment model to obtain the target reinforcement type of the reinforcement vector data; and a target reinforcement parameter determination unit, used to determine the target reinforcement parameters corresponding to the reinforcement vector data based at least on the dimensional parameter information of the reinforcement vector data and the target reinforcement type; the dimensional parameter information represents the reinforcement dimensional parameter information of the pile cap cross-sectional views from different perspectives; and the target reinforcement parameters represent the reinforcement component information corresponding to the pile cap cross-sectional views from different perspectives.

[0006] In another aspect, the present invention provides an electronic device, the electronic device comprising a processor and a memory, the memory being used to store a computer program, which, when executed by the processor, implements the above-described cross-sectional view recognition method.

[0007] In another aspect, the present invention provides a computer-readable storage medium for storing a computer program that, when executed by a processor, implements the above-described cross-sectional view recognition method.

[0008] By identifying the top-view pile cap section from the user-selected pile cap section drawing, and then calculating the similarity between this top-view pile cap section drawing and the preset pile cap template data, the target section type of the top-view pile cap section drawing is obtained. Then, based on the similarity between the reinforcement template data and the reinforcement vector data corresponding to the target section type, the target reinforcement type of the reinforcement vector data is determined. Finally, based on the extracted reinforcement size parameter information and the target reinforcement type, the target reinforcement parameters corresponding to the reinforcement vector data are determined. This can improve the efficiency of component identification in the pile cap section drawing to a certain extent. Attached Figure Description

[0009] The features and advantages of the invention will be more clearly understood by referring to the accompanying drawings, which are schematic and should not be construed as limiting the invention in any way. In the drawings:

[0010] Figure 1 A schematic diagram of the steps of a cross-sectional view recognition method according to one embodiment of the present invention is shown;

[0011] Figure 2A schematic diagram of a cross-sectional view recognition device according to one embodiment of the present invention is shown;

[0012] Figure 3 A schematic diagram of the structure of an electronic device according to one embodiment of the present invention is shown. Detailed Implementation

[0013] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0014] CAD vector graphics, also known as vector diagrams, record the position and color information of points, lines, and surfaces. CAD vector graphics can be used to describe various relevant information and data in building engineering. A pile cap refers to a structure where, when a building uses a pile foundation, the tops of the piles are connected by a reinforced concrete platform or slab to form a monolithic foundation to bear the loads on top. Before constructing a pile cap, it is necessary to identify the pile cap components in the pile cap drawings. In existing technologies, the process of identifying reinforcement component information based on CAD vector cross-sectional drawings requires manual reading of the reinforcement style and dimensional parameters on the cross-sectional drawing, followed by matching to obtain the corresponding reinforcement component information. However, this manual method of identifying components in the pile cap cross-sectional drawing is time-consuming. With the development of information technology, more and more civil engineering cost estimators are starting to use 3D modeling software. To address the difficulty and time-consuming nature of manually modeling pile caps using cross-sectional drawings, a high-precision AI cross-sectional drawing recognition algorithm is urgently needed, which can automatically identify and establish pile cap components based on the characteristics of CAD pile cap drawings.

[0015] Please see Figure 1 One embodiment of this application provides a cross-sectional view recognition method, which may include the following steps.

[0016] S110: Among the pile cap cross-sectional views obtained from different perspectives, determine the top-view pile cap cross-sectional view that represents the top-view perspective.

[0017] In this embodiment, the pile cap is a structural component that makes up the pile foundation. It connects the pile group into a whole at the top of the piles, supports the loads acting on the pile foundation, and transfers them to the piles and the ground. The planar shape of the pile cap depends on the arrangement of the piles and is usually rectangular or strip-shaped. Pile caps can include three-pile caps, four-pile caps, etc. The type of pile cap can be clearly determined by judging the bottom or top view of the pile cap. Therefore, the pile cap sectional view representing the top view can be determined first from different pile cap sectional views. The different pile cap sectional views can be the front view, side view (left view or right view), bottom view, or top view of the pile cap. For the pile cap, its upper and lower structures are consistent, so its bottom or top view structures are the same. Therefore, the top view pile cap sectional view representing the top view can also represent the bottom view pile cap sectional view.

[0018] S120: Input the top-view pile cap cross-section diagram into the cross-section type judgment model to obtain the target cross-section type of the top-view pile cap cross-section diagram.

[0019] In this embodiment, the cross-sectional type of the pile cap can be determined by judging the top-view pile cap cross-sectional diagram, such as a rectangular pile cap, a three-pile pile cap, a stepped four-pile pile cap, or a stepped five-pile pile cap. The cross-sectional type judgment model is used to calculate the similarity between the top-view pile cap cross-sectional diagram and each preset pile cap cross-sectional diagram template, and the cross-sectional type of the preset pile cap cross-sectional diagram template corresponding to the maximum similarity is taken as the model's output. For example, if five preset pile cap cross-sectional diagram templates (A, B, C, D, and E) are stored in the database, and the similarity between the top-view pile cap cross-sectional diagram and these five templates is calculated, the output similarity results are 0.75, 0.7, 0.85, 0.8, and 0.61 respectively. The cross-sectional type of the preset pile cap template cross-sectional diagram with a similarity of 0.85 is then taken as the target cross-sectional type of the top-view pile cap cross-sectional diagram. Of course, the preset pile cap section drawing template can include multiple section drawings with the same section type but different vector characteristics. This can avoid interference caused by annotation, leader lines and other issues to a certain extent when using the section type to determine the model for calculation.

[0020] S130: Extract the reinforcement vector data representing the reinforcement entity from the pile cap cross-sectional view from different perspectives.

[0021] In this embodiment, the cross-sectional view of the pile cap includes reinforcement data. The reinforcement data is labeled with its length using leader lines and may also include a corresponding name. During the identification process, it is necessary to identify the reinforcement vector data representing the reinforcement entity. Then, each reinforcement vector data is identified to obtain the component information of the pile cap.

[0022] S140: Input the rebar vector data and the target section type corresponding to the rebar vector data into the rebar type judgment model to obtain the target rebar type of the rebar vector data.

[0023] In this embodiment, after obtaining the target section type of the pile cap, it is also necessary to determine the type of rebar in the rebar vector data included in the pile cap cross-section drawing. Rebar patterns may include alternating upturns, staggered upturns, and no upturns at all. The rebar type determination model is used to calculate the similarity between the rebar vector data and the rebar template data, and assigns the rebar template data corresponding to the maximum similarity. Specifically, for example, the rebar vector data and the target section type can be input into the rebar type determination model. The model can first determine the rebar templates corresponding to several rebar types encompassed by the target section type, and then calculate the similarity with these rebar templates. Alternatively, similarity calculations can be performed with all rebar templates first, and then the rebar types and templates corresponding to the similarity from high to low can be sorted to obtain a rebar type sequence. This sequence is then iterated until a rebar type corresponding to the target section type is matched. For example, there are 17 types of rebar, A1, A2, A3, A4, ..., A17. Among them, the rebar types that match the target section type are A1, A3, A5, A8, A12, and A16. Then, the rebar type judgment model can calculate the similarity between the rebar vector data and the rebar templates corresponding to the six rebar types A1, A3, A5, A8, A12, and A16 to obtain the target rebar type of the rebar vector data.

[0024] S150: Based on the dimensional parameter information and the target rebar type, at least the rebar vector data is used to determine the target rebar parameters corresponding to the rebar vector data; the dimensional parameter information represents the rebar dimensional parameter information of pile cap cross-sectional views from different perspectives; the target rebar parameters represent the rebar component information corresponding to pile cap cross-sectional views from different perspectives.

[0025] In this embodiment, after determining the type of reinforcing steel, it is also necessary to obtain the parameter information of the reinforcing steel. The reinforcing steel members in the pile cap cross-sectional view can be identified based on the dimensions and structural relationships of the reinforcing steel to obtain the target reinforcing steel parameters. The reinforcing steel parameter information includes dimensional parameters (such as length information) and positional information (such as top reinforcing steel, bottom reinforcing steel, side reinforcing steel, inclined reinforcing steel, etc.). Specifically, for example, firstly, the dimensional parameters of each reinforcing steel entity in the pile cap cross-sectional view are extracted. Then, based on the dimensional parameters of the reinforcing steel entities, the positional relationships between different reinforcing steel vector data, and the target reinforcing steel type, the reinforcing steel parameter information corresponding to the reinforcing steel vector data is determined. Here, the reinforcing steel parameter information is the result of identifying the components in the pile cap cross-sectional view.

[0026] By constructing a hierarchical model for identifying components in sectional drawings, the process involves first identifying the cross-sectional type of the pile cap, then determining the type of reinforcement based on the cross-sectional type and reinforcement vector data, and finally identifying the reinforcement parameter information in the CAD drawings based on the reinforcement type and dimensional parameter information. This process yields the component information in the sectional drawings of the pile cap, thereby improving the efficiency of component identification in the sectional drawings of the pile cap to a certain extent.

[0027] In one implementation, determining a top-view pile cap cross-section representing a top-down perspective from the acquired pile cap cross-section images from different viewpoints may include: performing cluster analysis on the elements in the pile cap cross-section images from different viewpoints according to density to obtain clustered elements and scattered elements; the clustered elements are used to represent elements with a density greater than or equal to a preset threshold; the scattered elements are used to represent elements with a density less than the preset threshold; extracting name elements representing the name of the pile cap cross-section from the elements in the pile cap cross-section images; matching the clustered elements and the name elements to obtain the pile cap cross-section name corresponding to the clustered elements; determining the pile cap cross-section name to which the scattered elements belong based on the distance between the scattered elements and the clustered elements' respective cluster categories; determining the top-down pile cap cross-section representing a top-down perspective based on the pile cap cross-section name and target elements; the target elements include the clustered elements corresponding to the pile cap cross-section name and the scattered elements corresponding to the pile cap cross-section name.

[0028] In this embodiment, since users cannot directly determine which perspective corresponds to different viewpoints of the pile cap sectional view, and since the user cannot directly determine the affiliation of each element in the pile cap sectional view during the selection process, it is first necessary to combine, match, and identify the relevant elements of the three views of the pile cap. The elements are points, lines, surfaces, labels, leader lines, etc., in the pile cap sectional view. The clustered elements are used to characterize elements whose number of elements assigned to the same category and the area ratio of that category are greater than or equal to a preset threshold when dividing the sectional view into different categories. The scattered elements are used to characterize elements whose number of elements assigned to the same category and the area ratio of that category are less than a preset threshold when dividing the sectional view into different categories. The distance between the scattered elements and the clustered elements can be the distance between the center of the scattered elements and the center point of the clustered element's corresponding cluster range, or it can be the shortest distance from the center of the scattered elements to the plane of the clustered element's corresponding cluster range. Specifically, for example, a density clustering segmentation algorithm can first be used to calculate the density of each graphic element from the center point of the three views, dividing the graphic elements into clustered graphic elements and scattered graphic elements. Then, the graphic elements representing the names of the three views are identified to obtain named graphic elements. Next, the clustered graphic elements and named graphic elements are matched to obtain the pile cap section view names corresponding to each clustered graphic element. Then, scattered CAD graphic elements can be processed, and their affiliation can be determined by optimizing distance and name metrics. Finally, pile cap section views from different perspectives can be determined based on the names and clustered graphic elements. This yields the top-view pile cap section view representing the top-down perspective. By using an unsupervised approach, the vector map segmentation problem is transformed into a CAD graphic element calculation problem. By calculating the density of CAD graphic elements in a certain area, high-density section view areas are separated, which improves the accuracy of locating and segmenting section views from different perspectives to a certain extent.

[0029] In one embodiment, inputting the top-view pile cap sectional view into a section type determination model to obtain the target section type of the top-view pile cap sectional view may include: extracting multiple target section style features of the top-view pile cap sectional view; calculating the feature similarity between the target section style features and the section style features of multiple preset pile cap templates; obtaining the similarity between the top-view pile cap sectional view and each preset pile cap template based on the sum of the feature similarity between the multiple target section style features of the top-view pile cap sectional view and the feature similarity between each preset pile cap template; and taking the section type of the preset pile cap template data corresponding to the maximum similarity as the target section type of the top-view pile cap.

[0030] In this embodiment, CAD drawings are generally vector graphics, from which a large amount of vector data information can be extracted. Features of these vector features are then mined and extracted, such as the position of vector points, line lengths, surface shapes, the relationships between them, and text categories, to achieve effective classification and recognition of cross-sectional styles. The target cross-sectional style features in different feature dimensions may include: the ratio of the longest to the shortest side of the largest enclosed region, the number of sides in the largest enclosed region, whether the largest enclosed region is rectangular, the number of dimension annotations, the special instruction "main reinforcement" for the three-pile cap, whether it contains the phrase "three sides," and whether it is a quadrilateral, etc. Then, the features of the pile cap cross-section drawing in these feature dimensions and the features of the preset pile cap template data in these feature dimensions are extracted and their similarity is calculated, as shown in Table 1.

[0031] Table 1. Feature similarity of pile cap cross-sectional view and pre-set pile cap template data across different feature dimensions.

[0032] Feature 1 Feature 2 Feature 3 Feature 4 Feature 5 Feature 6 Feature 7 Template 1 0.7 0.48 0.91 0.85 0.69 0.75 0.82 Template 2 0.71 0.84 … … … … … Template 3 0.74 0.75 … … … … … Template 4 0.68 0.72 … … … … … Template 7 0.75 0.68 … … … … …

[0033] Then, the feature similarities in each cross-sectional template are accumulated to obtain the similarity between the top view pile cap cross-sectional view and each preset pile cap template. The cross-sectional view type of the pile cap template data corresponding to the maximum similarity is then used as the target section type of the top view pile cap cross-sectional view.

[0034] In one embodiment, the rebar vector data representing the rebar entities in the pile cap cross-sectional views from different perspectives are extracted; combined data of several rebar entities in the pile cap cross-sectional views from different perspectives are obtained; the combined data includes rebar annotation information, leader lines, and rebar entities; and the rebar vector data representing the rebar entities is extracted from the combined data of each rebar entity.

[0035] In this embodiment, since the description of the steel reinforcement entity in the pile cap sectional drawing includes not only the steel reinforcement vector data, but also the annotation information and leader lines of the steel reinforcement entity, the determination of the structural type of the pile cap needs to be based on the steel reinforcement vector data. Therefore, it is first necessary to divide the steel reinforcement entities in the pile cap sectional drawing to obtain steel reinforcement entity combination data, including the annotation information, leader lines and steel reinforcement vector data of the steel reinforcement entities. Then, the steel reinforcement vector data is extracted from each steel reinforcement entity combination data.

[0036] In one implementation, inputting the rebar vector data and the target section type corresponding to the rebar vector data into a rebar type determination model to obtain the target rebar type of the rebar vector data may include: determining a set of target rebar types corresponding to the pile cap section based on the target section type; the set of target rebar types includes the rebar types encompassed by the target section type; calculating the matching degree between the rebar vector data and the preset rebar template data corresponding to the rebar types in the target rebar type set; the preset rebar template characterizes the rebar template corresponding to the rebar types in the target rebar type set; and determining the rebar type of the preset rebar template data corresponding to the maximum matching degree as the target rebar type of the rebar vector data.

[0037] In this embodiment, during the process of determining the target rebar type, since there is a corresponding relationship between the target cross-section type and the target rebar type, the rebar type can be initially screened by the target cross-section type. Then, the rebar template and rebar vector data corresponding to the screened rebar type are similar to each other, thereby reducing the matching time of the target rebar type to a certain extent.

[0038] In this embodiment, the target rebar type set includes rebar types corresponding to the target cross-section type. Specifically, for example, if there are 20 rebar types in total, and the target cross-section type includes 5 rebar types, then a similarity calculation can be directly performed with the rebar templates corresponding to these 5 rebar types to obtain the matching degree between the rebar vector data and the preset rebar templates corresponding to the rebar types in the target rebar type set. Then, the maximum matching degree among these 5 rebar types is taken as the target rebar type.

[0039] In one embodiment, calculating the matching degree between the rebar vector data and the preset rebar template data corresponding to the rebar types in the target rebar type set may include: extracting multiple target rebar style features from the rebar vector data; calculating the feature matching degree between the target rebar style features and the rebar style features of the rebar template data corresponding to the rebar types in the target rebar type set; and obtaining the matching degree between the rebar vector data and each rebar template corresponding to the rebar types in the target rebar type set based on the sum of the feature matching degrees between the multiple target rebar style features of the rebar vector data and the rebar template data corresponding to the rebar types in the target rebar type set.

[0040] In this embodiment, the matching degree between the rebar vector data and the preset rebar template data corresponding to the rebar type in the target rebar type set can be calculated in the same way as calculating the similarity between the top view of the pile cap section and the preset pile cap template data. In the feature design of the rebar type, the design can be based on principles such as extensive analysis of drawings, incorporating business features, vector features, and geometric features.

[0041] The characteristics of the rebar style can include: the presence of Y-shaped tie bars, the presence of upper longitudinal rebar, the presence of lower longitudinal rebar, the presence of upper transverse rebar, the presence of lower transverse rebar, the presence of side transverse rebar, the presence of stirrups, the aspect ratio of the upper and lower rebars, the presence of side longitudinal rebar, the presence of multiple cross-sectional views, the width of the bottom transverse rebar, whether it is a stirrup form for specific longitudinal rebars, whether the edge bars are distributed vertically, whether all lines of the edge bars are contained within the baseline, the number of entities in different stirrup groups, and the number of edge lines with a length greater than one text height within the stirrup range. Then, the features of the rebar vector data in these dimensions are extracted and their matching degree is calculated with the features of the preset rebar templates corresponding to the rebar types in the target rebar type set. Then, the features in each cross-sectional template are accumulated to obtain the matching degree between the rebar vector data and the preset rebar templates corresponding to each rebar type in the target rebar type set. Finally, the rebar type of the rebar template corresponding to the maximum matching degree is taken as the target rebar type of the rebar vector data.

[0042] In one embodiment, determining the target rebar parameters corresponding to the rebar vector data based at least on the dimensional parameter information of the rebar vector data and the target rebar type may include: extracting the dimensional parameter information of the pile cap cross-sectional views from different perspectives; determining the target dimensional parameter information of the rebar vector data based on the dimensional parameter information and the positional relationship between the graphic elements included in the pile cap cross-sectional views; and determining the target rebar parameters corresponding to the rebar vector data based at least on the target dimensional parameter information and the target rebar type.

[0043] In this embodiment, the target rebar type is used to characterize the overall structural information of the rebar, and it is also necessary to identify local areas in the pile cap cross-section. Therefore, the rebar parameter information can be obtained by matching the identification results based on the target rebar type and the rebar's dimensional parameter information. Of course, information such as business relationships and vector data features of the pile cap cross-section can also be fused to make the identification results of the rebar parameter information more accurate.

[0044] In one embodiment, the cross-sectional view recognition method may further include: receiving user modifications to the target reinforcement parameters to obtain corrected reinforcement parameters; and updating the first similarity calculation model and / or the second similarity calculation model based on the corrected reinforcement parameters.

[0045] In some cases, the results generated by the cross-section type judgment model and / or the rebar type judgment model differ from the actual results and do not meet practical needs. Therefore, real-time feedback of user correction data on the recognition results can be received and fed back to the online model for incremental training to update the similarity parameter matrix, thereby improving the model's recognition accuracy to a certain extent. An incremental support vector machine model is introduced as the "mother model" in the cloud; a distance-based cosine similarity model is placed on the local PC as the "child model." Incremental online learning of the mother model is achieved, and the feature recommendation matrix parameters on the PC are updated to achieve optimal recommendation of results. The incremental model uses a Bayesian optimization algorithm for incremental optimization, enabling rapid training to obtain the updated similarity parameter matrix after new user data is added, and also allowing rapid learning of modified parameters during user modeling modifications. Optimizing the use of user modeling data and learning CAD drawing features provides a new approach to parameter learning and updating for vector graphics recognition models, making it more suitable for large-scale engineering applications. In this embodiment, the model can be updated and trained after a single user modification, or after a preset number of user modifications.

[0046] Please see Figure 2 One embodiment of this application also provides a cross-sectional view recognition device, which may include: a top-view pile cap cross-sectional view determination unit, a target section type determination unit, a reinforcement vector data extraction unit, a target reinforcement type determination unit, and a target reinforcement parameter determination unit.

[0047] The top-view pile cap section determination unit is used to determine the top-view pile cap section that represents the top-view perspective from the pile cap section drawings obtained from different angles.

[0048] The target section type determination unit is used to input the top-view pile cap sectional view into the section type judgment model to obtain the target section type of the top-view pile cap sectional view.

[0049] The reinforcement vector data extraction unit is used to extract the reinforcement vector data representing the reinforcement entity from the pile cap cross-sectional view from different perspectives.

[0050] The target rebar type determination unit is used to input the rebar vector data and the target section type corresponding to the rebar vector data into the rebar type judgment model to obtain the target rebar type of the rebar vector data.

[0051] The target reinforcement parameter determination unit is used to determine the target reinforcement parameters corresponding to the reinforcement vector data based at least on the dimensional parameter information of the reinforcement vector data and the target reinforcement type; the dimensional parameter information represents the reinforcement dimensional parameter information of the pile cap section view from different perspectives; the target reinforcement parameters represent the reinforcement member information corresponding to the pile cap section view from different perspectives.

[0052] In one embodiment, the top-view pile cap sectional view determination unit may include: a primitive classification unit, used to perform cluster analysis on primitives in pile cap sectional views from different perspectives according to density, to obtain clustered primitives and scattered primitives; the clustered primitives are used to characterize primitives with a density greater than or equal to a preset threshold; the scattered primitives are used to characterize primitives with a density less than the preset threshold; a name primitive extraction unit, used to extract name primitives characterizing the name of the pile cap sectional view from the primitives in the pile cap sectional view; and a name matching unit, used to match the clustered primitives and... The name primitives are matched to obtain the pile cap section view name corresponding to the cluster primitive; the scattered primitive category determination unit is used to determine the pile cap section view name to which the scattered primitive belongs based on the distance between the scattered primitive and the cluster primitive's respective cluster category; the top-view pile cap section view determination unit is used to determine the top-view pile cap section view representing the top-view perspective based on the pile cap section view name and the target primitive; the target primitive includes the cluster primitive corresponding to the pile cap section view name and the scattered primitive corresponding to the pile cap section view name.

[0053] In one embodiment, the target section type determination unit may include: a section style extraction unit, used to extract target section style features of the top-view pile cap cross-section in different feature dimensions; a section style feature similarity calculation unit, used to calculate the feature similarity between each target section style feature and the section style features of a plurality of preset pile cap templates; a section style similarity calculation unit, used to obtain the similarity between the top-view pile cap cross-section and each preset pile cap template based on the sum of the feature similarities of the plurality of target section style features of the top-view pile cap cross-section and the feature similarity of each preset pile cap template data; and a target section type determination unit, used to take the section type of the preset pile cap template data corresponding to the maximum similarity as the target section type of the top-view pile cap.

[0054] In one embodiment, the rebar vector data extraction unit may include: a rebar entity combination data acquisition unit, used to acquire combination data of several rebar entities in a pile cap cross-sectional view from different perspectives; the combination data includes rebar annotation information, leader lines, and rebar entities; and a rebar vector data extraction unit, used to extract rebar vector data representing the rebar entities from the combination data of each rebar entity.

[0055] In one embodiment, the target rebar type determination unit may include: a target rebar type set generation unit, used to determine the target rebar type set corresponding to the pile cap cross-section based on the target cross-section type; the target rebar type set includes the rebar types encompassed by the target cross-section type; a rebar matching degree calculation unit, used to calculate the matching degree between the rebar vector data and the preset rebar template data; the preset rebar template represents the rebar template corresponding to the rebar type in the target rebar type set; and a target rebar type determination unit, used to determine the rebar type of the preset rebar template data corresponding to the maximum matching degree as the target rebar type of the rebar vector data.

[0056] In one embodiment, the rebar matching degree calculation unit may include: a target rebar style feature extraction unit, used to extract multiple target rebar style features from the rebar vector data; a rebar feature matching degree calculation unit, used to calculate the feature matching degree between the target rebar style features and the rebar style features of the rebar template data corresponding to the rebar types in the target rebar type set; and a rebar matching degree calculation unit, used to obtain the matching degree between the rebar vector data and each rebar template corresponding to the rebar types in the target rebar type set based on the sum of the feature matching degrees of the multiple target rebar style features of the rebar vector data and the rebar template data corresponding to the rebar types in the target rebar type set.

[0057] In one embodiment, the target reinforcement parameter determination unit may include: a size parameter information extraction unit, used to extract size parameter information from the pile cap cross-sectional views from different perspectives; a target size parameter information determination unit, used to determine the target size parameter information of the reinforcement vector data based on the size parameter information and the positional relationship between the graphic elements included in the pile cap cross-sectional views; and a target reinforcement parameter determination unit, used to determine the target reinforcement parameters corresponding to the reinforcement vector data based at least on the target size parameter information and the target reinforcement type.

[0058] In one embodiment, the cross-sectional view recognition device may further include: a rebar parameter correction receiving unit, used to receive user modifications to the target rebar parameters to obtain corrected rebar parameters; and a model updating unit, used to update the cross-section type judgment model and / or the rebar type judgment model based on the corrected rebar parameters.

[0059] The specific functions and effects of the cross-sectional view recognition device can be explained by referring to other embodiments in this specification, and will not be repeated here. Each module in the cross-sectional view recognition device can be implemented entirely or partially through software, hardware, or a combination thereof. Each module can be embedded in or independent of the processor in a computer device in hardware form, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0060] Please see Figure 3 One embodiment of this application also provides an electronic device, which includes a processor and a memory. The memory is used to store a computer program, which, when executed by the processor, implements the above-described cross-sectional view recognition method.

[0061] The processor can be a central processing unit (CPU). It can also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or combinations thereof.

[0062] Memory, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs, non-transitory computer-executable programs, and modules, such as the program instructions / modules corresponding to the methods in the embodiments of this invention. The processor executes various functional applications and data processing by running the non-transitory software programs, instructions, and modules stored in the memory, thereby implementing the methods described in the above embodiments.

[0063] The memory may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function; the data storage area may store data created by the processor, etc. Furthermore, the memory may include high-speed random access memory and non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, the memory may optionally include memory remotely located relative to the processor, which can be connected to the processor via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0064] One embodiment of this application also provides a computer-readable storage medium for storing a computer program that, when executed by a processor, implements the above-described cross-sectional view recognition method.

[0065] Those skilled in the art will understand that implementing all or part of the processes in the methods described in this specification can be accomplished by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments described. Any references to memory, storage, databases, or other media used in the embodiments provided in this specification can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, or optical storage, etc. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0066] It should be understood that each block of a flowchart and / or block diagram, and combinations of blocks in a flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing device, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0067] This specification describes various embodiments in a progressive manner. Different embodiments focus on describing the parts that differ from other embodiments. Those skilled in the art, upon reading this specification, will realize that the various embodiments and the technical features disclosed in these embodiments can be combined in numerous ways. For the sake of brevity, not all possible combinations of the technical features in the described embodiments are described. However, any combination of these technical features that does not contradict each other should be considered within the scope of this specification.

[0068] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0069] The various embodiments described in this specification emphasize the parts that differ from other embodiments, and these embodiments can be explained by comparison with each other. Any combination of the various embodiments described in this specification, based on general technical knowledge, is covered within the scope of this specification.

[0070] The above description is merely an embodiment of this invention and is not intended to limit the scope of protection of the claims. Various modifications and variations are possible for those skilled in the art. Any modifications, equivalent substitutions, or improvements made within the spirit and principle of this invention should be included within the scope of the claims.

Claims

1. A profile identification method characterized by, The method includes: From the pile cap cross-sectional views obtained from different perspectives, determine the top-view pile cap cross-sectional view that represents the top-view perspective. Input the top-view pile cap cross-section diagram into the cross-section type determination model to obtain the target cross-section type of the top-view pile cap cross-section diagram; Extract the reinforcement vector data representing the reinforcement entity from the pile cap cross-sectional views from different perspectives; Input the rebar vector data and the target section type corresponding to the rebar vector data into the rebar type determination model to obtain the target rebar type of the rebar vector data; Based at least on the dimensional parameter information of the rebar vector data and the target rebar type, the target rebar parameters corresponding to the rebar vector data are determined; the dimensional parameter information represents the rebar dimensional parameter information of pile cap cross-sectional views from different perspectives; the target rebar parameters represent the rebar component information corresponding to pile cap cross-sectional views from different perspectives. The top-view sectional view of the pile cap is input into the section type determination model to obtain the target section type of the top-view sectional view of the pile cap, including: Extract multiple target section style features from the top-view pile cap cross-section; Calculate the feature similarity between each target section style feature and the pre-set section style features of multiple pile cap templates; Based on the sum of the feature similarity of the multiple target section style features of the top view pile cap sectional view and the feature similarity of each preset pile cap template data, the similarity between the top view pile cap sectional view and each preset pile cap template is obtained. The cross-sectional type of the preset pile cap template data corresponding to the maximum similarity is taken as the target cross-sectional type of the top-view pile cap.

2. The method of claim 1, wherein, From the pile cap cross-sectional views obtained from different perspectives, determine the top-view pile cap cross-sectional view representing the top-view perspective, including: Cluster analysis is performed on the elements in the pile cap cross-sectional view from different perspectives according to density to obtain clustered elements and scattered elements; the clustered elements are used to represent elements with a density greater than or equal to a preset threshold; the scattered elements are used to represent elements with a density less than the preset threshold. Extract the name element representing the name of the pile cap section view from the elements in the pile cap section view. The clustering primitives and the name primitives are matched to obtain the names of the pile cap cross-sections corresponding to the clustering primitives; Based on the distance between the scattered primitives and the clustered primitives, the name of the pile cap section view to which the scattered primitives belong is determined; Based on the name of the pile cap section view and the target elements, a top-view pile cap section view representing the top-view perspective is determined; the target elements include clustered elements corresponding to the name of the pile cap section view and scattered elements corresponding to the name of the pile cap section view.

3. The method of claim 1, wherein, Extract the reinforcement vector data representing the reinforcement entity from the pile cap cross-sectional views from different perspectives; including: Obtain combined data of several steel reinforcement entities from different perspectives of the pile cap cross-section; the combined data includes steel reinforcement annotation information, leader lines, and steel reinforcement entities; Extract the rebar vector data that characterizes the rebar entity from the combined data of each rebar entity.

4. The method of claim 1, wherein, Inputting the rebar vector data and the corresponding target section type into the rebar type determination model yields the target rebar type of the rebar vector data, including: Based on the target section type, determine the target reinforcement type set corresponding to the pile cap section view; the target reinforcement type set includes the reinforcement types encompassed by the target section type; Calculate the matching degree between the rebar vector data and the preset rebar template data; the preset rebar template represents the rebar template corresponding to the rebar type in the target rebar type set; The type of rebar in the preset rebar template data corresponding to the maximum matching degree is determined as the target rebar type in the rebar vector data.

5. The method of claim 4, wherein, Calculating the matching degree between the rebar vector data and the preset rebar template data corresponding to the rebar types in the target rebar type set includes: Extract multiple target rebar pattern features from the rebar vector data; Calculate the feature matching degree between the target rebar style features and the rebar style features of the rebar template data corresponding to the rebar types in the target rebar type set; The matching degree between the rebar vector data and each rebar template corresponding to the rebar type in the target rebar type set is obtained by summing the feature matching degrees of multiple target rebar style features of the rebar vector data and the feature matching degrees of the rebar template data corresponding to the rebar type in the target rebar type set.

6. The method of claim 1, wherein, Determining the target rebar parameters corresponding to the rebar vector data based at least on the dimensional parameter information of the rebar vector data and the target rebar type includes: Extract the dimensional parameters of the pile cap cross-sectional views from different perspectives; Based on the dimensional parameter information and the positional relationship between the elements included in the pile cap cross-section, the target dimensional parameter information of the reinforcement vector data is determined; The target rebar parameters corresponding to the rebar vector data are determined based at least on the target size parameter information and the target rebar type.

7. The method of claim 1, wherein, The method further includes: Receive user modifications to the target reinforcement parameters and obtain corrected reinforcement parameters; The cross-section type determination model and / or the rebar type determination model are updated based on the corrected rebar parameters.

8. A profile identification device, characterized by The cross-sectional view recognition device includes: The top-view pile cap section determination unit is used to determine the top-view pile cap section that represents the top-view perspective from the pile cap section drawings obtained from different angles. The target section type determination unit is used to input the top-view pile cap sectional view into the section type judgment model to obtain the target section type of the top-view pile cap sectional view; The reinforcement vector data extraction unit is used to extract the reinforcement vector data representing the reinforcement entity from the pile cap cross-sectional view from different perspectives; The target rebar type determination unit is used to input the rebar vector data and the target section type corresponding to the rebar vector data into the rebar type judgment model to obtain the target rebar type of the rebar vector data; The target reinforcement parameter determination unit is used to determine the target reinforcement parameters corresponding to the reinforcement vector data based at least on the dimensional parameter information of the reinforcement vector data and the target reinforcement type; the dimensional parameter information represents the reinforcement dimensional parameter information of pile cap cross-sectional views from different perspectives; the target reinforcement parameters represent the reinforcement member information corresponding to pile cap cross-sectional views from different perspectives; The target cross-section type determination unit includes: The cross-section style extraction unit is used to extract the target cross-section style features of the top-view pile cap cross-section in different feature dimensions; The cross-section style feature similarity calculation unit is used to calculate the feature similarity between each target cross-section style feature and the cross-section style features of multiple preset pile cap templates; The cross-section style similarity calculation unit is used to obtain the similarity between the top view of the pile cap cross-section and each preset pile cap template based on the sum of the feature similarity of multiple target cross-section style features of the top view of the pile cap cross-section and the feature similarity of each preset pile cap template data. The target section type determination unit is used to take the section type of the preset pile cap template data corresponding to the maximum similarity as the target section type of the top-view pile cap.

9. An electronic device, comprising: The electronic device includes a processor and a memory, the memory being used to store a computer program that, when executed by the processor, implements the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium is used to store a computer program that, when executed by a processor, implements the method as described in any one of claims 1 to 7.

Images