Railway signal drawing analysis method, system and device

Abstract

The invention belongs to the technical field of engineering drawing management, and discloses a railway signal drawing analysis method, system and device, and the method comprises the steps: carrying out the preprocessing of all signal drawings of a target station, carrying out the drawing frame recognition and forward mapping of obtained first data, and obtaining a first data frame; performing table recognition on the obtained second data to obtain table structure data; identifying characters in the second data by utilizing a plurality of character identification technologies, correspondingly obtaining a plurality of original data, then performing fusion operation to obtain table character data, and performing reverse mapping after merging the table character data with the table structure data; extracting terminal relation data according to the obtained third data to construct an equipment relation tree; after the name of the target element is received, node terminal relation data of the target element is retrieved in the equipment relation tree, a matched target graph is searched in the second data and then is drawn with the node terminal relation data, an equipment relation graph is obtained, efficient query is achieved, and all associated drawings of the signal equipment are visually displayed in a graphic mode.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of engineering drawing management, in particular to a railway signal drawing analysis method, system and device. BACKGROUND

[0002] The railway signal drawing comprehensively reflects the installation position and connection relationship of various signal devices in the signal system, and the daily maintenance and fault handling of the railway electric department must use the signal drawing for construction guidance and technical reference. After the railway station is opened for operation, the signal drawing is handed over to the railway operation management department as a device file, which is archived and managed in the form of a CAD file. The CAD (Computer-Aided Design) file is a general term for all data files generated during the design process.

[0003] The railway signal drawing has the characteristics of dense graphical elements, numerous drawings, and complex mutual relationships between drawings. A signal device is scattered in different drawings from the cabinet position, cable path, connection box, and all connection relationships. When the device is daily maintained or handled in an emergency, the user needs to open multiple drawings at the same time, and then find the cable path relationship, device connection relationship, and terminal relationship one by one in each DWG file. Then, switching between multiple pages for browsing not only is inefficient but also is prone to errors. Moreover, although it is still possible to manually query associated graphics one by one and then manually draw the signal drawing, there are problems such as large workload and low drawing efficiency. The DWG (Drawing) is a proprietary file format created by AutoCAD.

[0004] Therefore, the prior art still needs to be improved and improved. SUMMARY

[0005] The present application provides a railway signal drawing analysis method, system and device, which aims to solve the problem that the signal drawing cannot be intuitively displayed and the associated graphics of the signal device cannot be quickly browsed in the prior art.

[0006] In a first aspect, an embodiment of the present application provides a railway signal drawing analysis method, comprising: Preprocessing all signal drawings of a target station to obtain first data, sequentially performing frame recognition and forward mapping on the first data to obtain second data, and then performing table recognition on the second data to obtain table structure data; Using multiple text recognition technologies to recognize the text in the second data, respectively, to obtain multiple original data after fusion operation, to obtain table text data, and then performing reverse mapping on the table structure data and the table text data after merging, to obtain third data; constructing a device relationship tree according to terminal relationship data extracted from the third data; receiving a target element name input by a user and searching for node terminal relationship data of the target element in the device relationship tree to find a matching target graph in the second data and drawing a device relationship graph with the node terminal relationship data.

[0007] In some embodiments, the first data is obtained by preprocessing all signal diagram sheets of a target station, the second data is obtained by sequentially performing frame recognition and forward mapping on the first data, and the table structure data is obtained by performing table recognition on the second data, including: comparing the resolution of each signal picture with a plurality of preset standard resolutions, scaling the current signal picture to the preset standard resolution with the smallest difference, sequentially performing binarization and line enhancement on the obtained scaled signal picture to obtain the first data; performing frame recognition on the first data using a target recognition algorithm and converting the data into an image coordinate system using a preset coordinate mapping relationship to obtain the second data; performing table recognition on the second data using the target recognition algorithm to obtain the table structure data.

[0008] In some embodiments, a plurality of character recognition technologies are used to recognize characters in the second data respectively, a plurality of original data are obtained by corresponding recognition, and a fusion operation is performed to obtain table character data, and the table structure data and the table character data are merged and reverse mapped, including: recognizing characters in the second data using an optical character recognition method to obtain first character information, and recognizing the characters using a target learning algorithm to obtain second character information; determining whether the second character information contains the first character information, if not, writing the first character information into the second character information, if not all, calculating the area size of the character area of the first character information and the second character information respectively, and retaining the character with larger area to obtain the table character data; merging the table structure data and the table character data using icon coordinates in the table structure data to obtain merged table data; converting the merged table data into data in a world coordinate system using the preset coordinate mapping relationship to obtain the third data.

[0009] In some embodiments, the device relationship tree is constructed according to terminal relationship data extracted from the third data, including: Extracting a target field in the third data to obtain the terminal relationship data, so as to construct a device name and a terminal connected thereto to obtain the device relationship tree.

[0010] In some embodiments, the target element name input by the user is received and searched in the device relationship tree to obtain node terminal relationship data of the target element, and after a matching target graph is found in the second data, the device relationship graph is drawn according to the node terminal relationship data, including: The target element name input by the user is received and searched in the device relationship tree to obtain node terminal relationship data of the target element, and a frame corresponding to the target graph is separated from the frame according to the graph area coordinates in the node terminal relationship data. The target graph is constructed and drawn to obtain the device relationship graph for online browsing by the user.

[0011] In some embodiments, the pre-processing of all signal graph papers of a target station to obtain the first data further includes: The signal graph paper in a standard format is structurally parsed to identify a layer structure, a view configuration, a layout space and a model space, conversion parameters are determined according to a preset rule, and a signal picture in a lossless compression image format is obtained by vector rendering according to the conversion parameters by using a rendering engine.

[0012] In some embodiments, the frame recognition of the first data by using the target recognition algorithm further includes: The frame size in the second data is compared with a preset frame threshold value. If the target graph is smaller than a preset minimum frame threshold value or the coincidence rate between a plurality of target graphs is greater than a preset maximum frame threshold value, the frame recognition of the first data is performed again by using the target recognition algorithm, and the repeatedly recognized frame is deleted to obtain corrected second data. The second data includes a frame name and a frame coordinate.

[0013] In some embodiments, the conversion of the second data into data in an image coordinate system by using a preset coordinate mapping relationship further includes: A preset coordinate mapping relationship between an image coordinate system and a world coordinate system is established to perform forward mapping or reverse mapping according to the preset coordinate mapping relationship.

[0014] In a second aspect, the embodiments of the present application provide a railway signal graph paper analysis system, including: ​The table structure acquisition module is configured to preprocess all signal drawing sheets of a target station to obtain first data, sequentially perform frame recognition and forward mapping on the first data to obtain second data, and perform table recognition on the second data to obtain table structure data. The table text acquisition module is configured to recognize texts in the second data by using multiple text recognition techniques, fuse multiple original data obtained by the recognition to obtain table text data. The third data acquisition module is configured to perform reverse mapping on the table structure data and the table text data after merging to obtain third data. The device relationship tree construction module is configured to construct a device relationship tree according to terminal relationship data extracted from the third data. The device relationship diagram drawing module is configured to receive a target element name input by a user, search the device relationship tree to obtain node terminal relationship data of the target element, search the second data for a matching target graph, and draw a device relationship diagram according to the node terminal relationship data.

[0015] In a third aspect, an apparatus is provided, which includes a memory, a processor, a display, and a computer program stored in the memory and executable on the processor, and the computer program, when executed by the processor, implements the steps of the railway signal drawing sheet analysis method.

[0016] Compared with the prior art, the railway signal drawing sheet analysis method and system and apparatus provided by the present application preprocess all signal drawing sheets of a target station, sequentially perform frame recognition and forward mapping on obtained first data, perform table recognition on obtained second data to obtain table structure data, recognize texts in the second data by using multiple text recognition techniques, fuse multiple original data obtained by the recognition to obtain table text data, perform reverse mapping on the table structure data and the table text data after merging to obtain third data, construct a device relationship tree according to terminal relationship data extracted from the third data, search the device relationship tree to obtain node terminal relationship data of a target element after receiving a target element name, search the second data for a matching target graph, and draw a device relationship diagram according to the node terminal relationship data, thereby realizing efficient query and intuitively displaying all associated drawing sheets of signal devices in a graphical manner. BRIEF DESCRIPTION OF DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiments. It should be understood that the following drawings only show some of the embodiments of the present application, and therefore should not be regarded as limiting the scope, and all other embodiments obtained by those skilled in the art without making creative efforts based on the drawings are within the scope of the present application.

[0018] Figure 1 A flow chart of the railway signal drawing analysis method provided by the present application; Figure 2 A flow chart of the table structure data acquisition in the railway signal drawing analysis method provided by the present application; Figure 3 A flow chart of the second data correction in the railway signal drawing analysis method provided by the present application; Figure 4 A flow chart of the third data acquisition in the railway signal drawing analysis method provided by the present application; Figure 5 A flow chart of the arrival and departure chart of the combined cabinet in the terminal relationship data provided by the present application; Figure 6 A flow chart of the equipment relationship chart drawing in the railway signal drawing analysis method provided by the present application; Figure 7 A schematic diagram of the equipment relationship chart provided by the present application; Figure 8 A structural schematic diagram of the railway signal drawing analysis system provided by the present application.

[0019] Reference signs: 10-table structure acquisition module; 20-table text acquisition module; 30-third data acquisition module; 40-equipment relationship tree construction module; 50-equipment relationship chart drawing module. DETAILED DESCRIPTION

[0020] The technical solutions of the embodiments of the present application will be described clearly and completely in the following with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments.

[0021] The components of the embodiments of the present application generally described and shown in the drawings herein can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the claimed present application, but only represents selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without making creative efforts are within the scope of the present application.

[0022] Hereinafter, the terms "include", "have", and their conjugates, used in the various embodiments of the present application, merely indicate the presence of the features, numbers, steps, operations, elements, components, or combinations thereof, and do not preclude the presence or addition of one or more other features, numbers, steps, operations, elements, components, or combinations thereof.

[0023] Unless defined otherwise, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art to which various embodiments of the present application belong. Such terms, as commonly used in the art, are to be interpreted in the same way as they are interpreted in the context of the relevant art and will not be interpreted in an idealized or overly formal sense unless clearly defined in various embodiments of the present application.

[0024] The present application provides a railway signal drawing analysis method and system and device. The railway signal drawing analysis method pre-processes all signal drawings of a target station, sequentially performs frame recognition and forward mapping on the obtained first data, and performs table recognition on the obtained second data to obtain table structure data; uses multiple character recognition technologies to recognize characters in the second data, obtains multiple original data, and performs fusion operation to obtain table character data; and performs reverse mapping on the merged table structure data and table character data to obtain third data; constructs a device relationship tree according to terminal relationship data extracted from the third data; receives a target element name, retrieves node terminal relationship data of the target element in the device relationship tree, finds a matching target graph in the second data, and draws a device relationship graph, which realizes the conversion of paper signal drawings into online signal drawings for users to browse and search the device relationship graph, effectively improves the drawing browsing efficiency, and improves the accuracy and efficiency of signal device searching.

[0025] The railway signal drawing analysis method design scheme is described below through some specific embodiments.

[0026] Please refer to Figure 1 The present application provides a railway signal drawing analysis method, which includes steps S100-400: S100, pre-processes all signal drawings of a target station to obtain first data, sequentially performs frame recognition and forward mapping on the first data to obtain second data, and then performs table recognition on the second data to obtain table structure data.

[0027] Wherein, in the application, the target station is taken as an example of a train station (which can also be a high-speed rail station), and can also be a subway station, etc. The signal drawing is a railway electric power engineering completion drawing, mainly in the form of DWG file and management, including cable route drawing, cable wiring drawing, combination cabinet / rack side wiring drawing, distribution panel wiring drawing, lightning protection cabinet wiring drawing, interface cabinet / rack wiring drawing, acquisition wiring drawing, power supply wiring drawing, combination internal wiring drawing, track cabinet wiring drawing and zero layer drawing, etc. The first data is image data. The second data is frame data, including frame name and frame coordinates. The table text data includes table text content and table text coordinates.

[0028] Exemplarily, first, all signal drawings of the target station are imported into the CAD software in units of stations, and all signal drawings of the target station are preprocessed, such as resolution scaling, binarization operation and line enhancement operation, to obtain the first data.

[0029] Then, the frame recognition algorithm (such as yolo11 target recognition algorithm, which is used to realize high-precision real-time recognition of target objects in images) is used to recognize the frame of the first data, and the frame recognition result is output, including the frame name and the frame xy coordinates.

[0030] After that, the first data after recognizing the frame is subjected to coordinate conversion: first, forward mapping is performed, i.e. mapping and converting the picture xy coordinates (two-dimensional Cartesian coordinate system) to cad coordinates (coordinate system in CAD software), intercepting all graphics in the coordinate range, saving to a new dwg file, obtaining the second data, and saving to the frame file library.

[0031] Further, the yolo11 target recognition algorithm is also used to recognize the table of the second data to output the table structure data, including the table name and the table xy coordinates, to obtain the table structure data.

[0032] It can be understood that in the application, the frame and icon in the signal drawing are recognized by using the target recognition algorithm to recognize the frame and icon in the signal drawing, so as to form an online file and provide an opportunity for online browsing.

[0033] In one implementation method, all signal drawings of the target station are preprocessed to obtain the first data, which further includes: The standard format signal drawing is structurally parsed to recognize the layer structure, view configuration, layout space and model space, and the conversion parameters are determined according to the preset rules, and the rendering engine is used to perform vector rendering according to the conversion parameters to obtain the lossless compressed image format signal picture.

[0034] Wherein, the conversion parameters include: output image resolution, image size, background color setting and whether to retain layer visibility setting.

[0035] Exemplarily, in another embodiment in the present application, before the signal drawing is preprocessed, the signal drawing is first converted in format, and the signal drawing of the Dwg file is converted into a picture in png format, that is, the view range of the standard format CAD drawing (i.e. the signal drawing) is read and the effective view area is extracted, so as to parse the structure of the signal drawing, and the layer structure, view configuration, layout space and model space are recognized.

[0036] Then, conversion parameters such as the size of the scale factor, color, line type and target image resolution (DPI) or output image size, and optional parameters such as whether to draw the figure centrally, whether to fill the background, etc. are set. And the rendering engine is used to perform vector rendering according to the conversion parameters, and the entity objects (such as lines, text, block references, polylines, etc.) in the CAD layer are rendered to the bitmap cache according to the attributes such as color and line type, and then saved as a signal picture in PNG format (lossless compression image format).

[0037] In one embodiment, please refer to Figure 2 , step S100, pre-processing all signal drawings of the target station to obtain first data, sequentially performing frame recognition and forward mapping on the first data to obtain second data, and then performing table recognition on the second data to obtain table structure data, including: S110, comparing the resolution of each signal picture with a plurality of preset standard resolutions, and scaling the current signal picture to the preset standard resolution with the smallest difference, and sequentially performing binarization operation and line enhancement on the obtained scaled signal picture to obtain first data; S120, performing frame recognition on the first data by using a target recognition algorithm, and converting the data into image coordinate system by using a preset coordinate mapping relationship to obtain second data; S130, performing table recognition on the second data by using a target recognition algorithm to obtain table structure data.

[0038] Exemplarily, all signal drawings on the target station are first preprocessed: Because the drawing scale of Dwg drawings is different, when Dwg is converted into png picture according to cad coordinate size, it is easy to appear super large or super small resolution, which leads to low accuracy of target recognition algorithm, so it is necessary to scale the resolution of all signal drawings, that is, the resolution of each signal picture is compared with a plurality of preset standard resolutions (for example, [640*640, 1280*1280, 2000*2000, 5000*5000] and the like), and the current signal picture is scaled to the preset standard resolution closest to it, that is, according to the resolution size of the input signal picture, the closest resolution is matched for scaling and saving, and the scaled signal picture is obtained.

[0039] Then, the scaled signal picture is subjected to a binarization operation, the pixel value in the image is converted into 0 or 1, so that the picture is divided into black and white parts. After that, the signal picture subjected to the binarization operation is subjected to line enhancement processing, that is, the line is thickened after being recognized, and is redrawn to the input picture, and the first data is obtained.

[0040] Secondly, the first data is subjected to frame recognition by using a target recognition algorithm. Firstly, the first data is normalized to a fixed size (such as 640*640 pixels). Secondly, feature extraction and multi-scale fusion are performed: deep semantic features of the image are extracted by using an improved backbone network (such as CSPDarknet11 and a Transformer integrated module and the like), and a multi-scale feature pyramid structure such as FPN++ / PANet is introduced to fuse features of different levels to enhance the small target detection capability, and three scale feature maps (such as 80*80, 40*40, 20*20) are outputted.

[0041] Next, in the third step, the detection head outputs original prediction information, including: boundary box coordinates, target confidence and class probability. Among them, each boundary box is represented by four coordinate values: (x_center, y_center, width, height) (that is, the x coordinate and y coordinate of the center point of the boundary box on the image, the width and height of the boundary box itself), which is usually normalized to the image size; the target confidence represents the probability value (0~1) that the region contains a target object; the class probability is the probability distribution of all classes outputted for each boundary box.

[0042] Fourth step, post-processing and final frame generation: first, filter the confidence and category threshold, that is, filter out the bounding box with confidence lower than the set threshold (such as 0.25), and select the target category name corresponding to the maximum category probability for the remaining bounding box. And carry out non-maximum suppression, using improved NMS strategy (such as Soft-NMS, DIoU-NMS), eliminate overlapping detection boxes, in order to improve the detection accuracy in dense scene. Then carry out frame coordinate denormalization and integer conversion, convert the normalized (x_center, y_center, width, height) to the actual pixel coordinates of the image, scale according to the original image size, output the left top corner and right bottom corner coordinates of the frame. Then encapsulate the recognition result into a standard structure or JSON format, etc., to obtain the second data.

[0043] Further, the second data is converted into image coordinate system data by using a preset coordinate mapping relationship, that is, the picture xy coordinate is mapped and converted to cad coordinate, and all graphics in the coordinate range are intercepted and saved to a new dwg file, to obtain the second data.

[0044] Among them, the image xy coordinate system, that is, the two-dimensional Cartesian coordinate with pixels as the unit, the origin is located at the left top corner of the image (0, 0), the x-axis increases to the right, and the y-axis increases downward. The cad coordinate system is usually the world coordinate system (WCS) or user-defined coordinate system (UCS) in engineering drawings, expressed in physical units such as millimeters and meters.

[0045] It can be understood that in the present application, the feature extraction structure and dynamic non-maximum suppression mechanism are used by using the target recognition algorithm, so as to improve the recognition ability and reasoning efficiency of the frame.

[0046] In one implementation method, referring to Figure 3 , the target recognition algorithm is used to identify the frame of the first data, and then further comprising: S121, comparing the frame size in the second data with the preset frame threshold; S122, if the target graphic is smaller than the preset minimum frame threshold, or the coincidence rate between multiple target graphics is greater than the preset maximum frame threshold, the target recognition algorithm is used again to identify the frame of the first data, and the repeatedly identified frame is deleted, to obtain the corrected second data.

[0047] Exemplarily, after identifying the frame of the first data, the second data can also be corrected: There are two cases when comparing the frame size in the second data with the preset frame threshold. First, compare each frame size with the preset minimum frame threshold (for example, 25% of the average value of all recognition frame areas in the frame recognition result). Second, select two frames from all frames each time for overlap rate comparison.

[0048] Therefore, if the target graph is smaller than the preset minimum threshold, it means that the target is too small, which may be a recognition error. Or if the overlap rate is greater than the preset maximum frame threshold, it means that the overlap rate is too high, which may appear multiple times. Therefore, in both cases, the frame needs to be recognized again, and the redundant results are deleted, that is, the frames that are repeatedly recognized in the two recognitions are deleted to obtain the corrected second data.

[0049] In one implementation method, the second data is converted into data in the image coordinate system using the preset coordinate mapping relationship, which further includes: Establishing a preset coordinate mapping relationship between the image coordinate system and the world coordinate system to perform forward mapping or reverse mapping according to the preset coordinate mapping relationship.

[0050] Exemplarily, the process of obtaining the preset coordinate mapping relationship is as follows: First, a preset coordinate mapping relationship between the xy coordinate system (i.e., the image coordinate system) and the cad coordinate system (i.e., the world coordinate system) is established.

[0051] Then, the position information of the detected text and graphic elements in the image (such as the position of the text recognized by OCR and the endpoints of the lines on the drawing) is converted into real coordinate values in the CAD drawing, so as to perform subsequent size calculation, component positioning, and three-dimensional modeling operations. Therefore, the specific process of forward mapping is as follows: first, use OCR or image recognition algorithm to extract the text, lines, endpoints, and other elements on the image to obtain the pixel coordinates (x_img, y_img) of these elements in the image; and manually or automatically select several control points with known coordinates, such as several standard marking points on the drawing; and then use these control points to establish a transformation function T: (x_img, y_img) → (x_cad, y_cad).

[0052] In the picture represented by the xy coordinate, the upper left corner is the coordinate origin, the right x positive direction, and the downward y positive direction, denoted as x, y. In the cad coordinate, the lower left corner is the coordinate origin, the right x positive direction, and the upward y positive direction, denoted as X, Y. Therefore, the preset coordinate mapping relationship (transformation function) between them is: X = X0+ scale*x (1); Y = Y0+ scale * ( h - y) (2); Wherein X0 is the minimum value of X coordinate in CAD drawing view. Y0 is the minimum value of Y coordinate in CAD drawing view. Scale is the scale ratio in Png scaling processing, and h is the picture height input to the recognition algorithm.

[0053] Finally, the new point (x_img, y_img) in any image is mapped to the coordinate (x_cad, y_cad) in the CAD space through the transformation function.

[0054] S200, the text in the second data is recognized by using a plurality of text recognition technologies, and a plurality of original data are obtained after fusion operation, table text data are obtained, and the table structure data and the table text data are merged and reversely mapped to obtain third data.

[0055] In the present application, the plurality of text recognition technologies are taken as examples of OCR recognition algorithm and CAD file analysis algorithm (analysis algorithm provided by CAD software).

[0056] Exemplarily, after obtaining the second data, the target recognition algorithm, the OCR recognition algorithm and the CAD file analysis algorithm are used to fuse and analyze the drawing content. Considering that the OCR recognition technology cannot achieve 100% recognition, there is a certain recognition omission, and the recognition omission may occur. At the same time, part of the text in the CAD drawing is drawn by lines, and the CAD analysis cannot obtain such text, so the CAD file analysis and the OCR technology are fused to realize 100% recognition of the drawing text content.

[0057] Then, the text in the second data is recognized and analyzed by using the OCR recognition algorithm and the CAD file analysis tool respectively, and two kinds of original data are obtained, and then the two kinds of original data are fused: the two kinds of original data with high similarity or only in the OCR recognition algorithm are taken as table text data.

[0058] Then, after the table structure data and the table text data are merged, the transformation function in the forward mapping is also used for reverse mapping, and finally the third data is obtained.

[0059] It can be understood that, by using a plurality of text recognition technologies to recognize and fuse the recognition results, the accuracy and quality of recognition can be improved, the error rate can be reduced, and the accuracy of generating online signal drawings can be improved.

[0060] In one embodiment, please refer to Figure 4 , the text in the second data is recognized by using a plurality of text recognition technologies, and a plurality of original data are obtained after fusion operation, table text data are obtained, and the table structure data and the table text data are merged and reversely mapped, including: S201, using an optical character recognition method to recognize characters in the second data to obtain first text information, and using a target learning algorithm to recognize the characters to obtain second text information; S202: Determine whether the second text information contains the first text information. If not, write the first text information into the second text information. If not, calculate the area of ​​the text regions of the first text information and the second text information, respectively, and retain the text with the larger area to obtain table text data. S203, merging the table structure data and the table text data using the icon coordinates in the table structure data to obtain merged table data; S204: Convert the merged table data into data in the world coordinate system using a preset coordinate mapping relationship to obtain third data.

[0061] The original data includes text and various coordinate data, and the text includes first text information and second text information. The third data is chart data.

[0062] Exemplarily, the specific process of obtaining the third data is as follows: After obtaining the table structure data, multiple text recognition technologies are used to recognize the characters in the second data respectively, that is, the OCR recognition algorithm (i.e., optical character recognition method) is called to recognize all text contents in the frame image (i.e., the second data), and the first text information is output, including the first text content and its corresponding text area frame coordinates (x1, y1, x2, y2); and the CAD file parsing tool is called to parse all text contents in the frame image, and the second text information is output, including the second text content and its corresponding character area frame coordinates (X1, Y1, X2, Y2).

[0063] Then, the original data identified twice above (i.e. (x1, y1, x2, y2) and (X1, Y1, X2, Y2)) are merged and filtered: The first text information identified by the OCR algorithm is traversed in sequence, and matched one by one in the second text information parsed by CAD: if the same text in the first text information is not found in the CAD parsing result (or the same text does not exceed the minimum threshold), that is, they are not similar, the first text information is written into the second text information as table text data; if it is found (or the same text exceeds the maximum threshold), the coordinates of the text area boxes of the first text information and the second text information are calculated respectively, the area size of each text area is calculated, and the text with the larger area is taken and written into the table text data.

[0064] Secondly, the third data is generated by merging the table structure data and the text content corresponding to the table text data according to the coordinates, and the third data is reversely mapped, and the cad coordinate points are converted into image coordinate points by using the transformation function in the forward mapping, and the points are marked on the image to obtain the third data.

[0065] S300, the terminal relationship data extracted from the third data is used to construct the device relationship tree.

[0066] Exemplarily, after obtaining the third data, the terminal relationship data including the terminal coming direction, the terminal going direction, the file source, the graphic area coordinates and the device name is extracted according to the drawing rules of different types of drawings, and the terminal relationship database is established, and the terminal relationship data is added to the terminal relationship database, wherein the coming and going direction chart of the combined cabinet in the terminal relationship data is as shown in Figure 5 .

[0067] Then, the device relationship tree is constructed according to the device name and the terminal coming and going direction relationship by traversing the terminal relationship database.

[0068] In one embodiment, the device relationship tree is constructed according to the terminal relationship data extracted from the third data, including: The target field in the third data is extracted to obtain the terminal relationship data, so as to construct the device name and the terminal coming and going direction relationship connected thereto, and obtain the device relationship tree; The target field includes the terminal coming direction, the terminal going direction, the file source, the graphic area coordinates and the device name.

[0069] Exemplarily, after obtaining the third data, the terminal coming direction, the terminal going direction, the file source, the graphic area coordinates and the device name and other target fields are extracted by analyzing the combined cabinet, the number of layers, the terminal board number and the terminal number in different drawings, so as to obtain the terminal relationship data. Then, the device relationship tree with the device name as the main body and the terminal coming and going direction relationship connected thereto as the branch is constructed according to the terminal relationship data.

[0070] S400, the node terminal relationship data of the target element is obtained by receiving the target element name input by the user and searching in the device relationship tree, so as to find the matching target graphic in the second data, and draw the device relationship diagram according to the node terminal relationship data.

[0071] Optionally, after the device relationship tree is constructed, the user can input the name of the target element such as the device, the cable name or the terminal, the tree traversal algorithm is used to search the target element name in the device relationship tree, and the target element is found by traversing the nodes of the device relationship tree.

[0072] Then, according to the searching of the frame matching the target component in the second data, and according to the terminal relationship data of the target component, the device relationship diagram of the target component is drawn for the user to check.

[0073] It can be understood that, in the present application, the node terminal relationship data of the target component is obtained by searching the device relationship tree, to match the frame in the second data, and according to the frame and the node terminal relationship data, the device relationship diagram is drawn for the user to check online, which improves the convenience and simplicity of online checking.

[0074] In an implementation method, referring to Figure 6 , the step S400 receives the target component name input by the user and searches in the device relationship tree to obtain the node terminal relationship data of the target component, to search for the matching target frame in the second data, and draw the device relationship diagram according to the node terminal relationship data, including: S410, receiving the target component name input by the user and searching in the device relationship tree to obtain the node terminal relationship data of the target component, and matching the frame corresponding to each node in the node terminal relationship data in the second data; S420, separating the target frame corresponding to the frame region from the frame according to the frame region coordinates in the node terminal relationship data; S430, constructing and drawing the target frame to obtain the device relationship diagram for the user to check online.

[0075] Wherein, the target component name includes: input device, cable and terminal name. Wherein, the input device can be a cabinet or a rack; the cabinet is an interface cabinet, a combination cabinet, a power supply cabinet or a track cabinet, etc. device, used to realize different wiring.

[0076] Exemplarily, the specific construction process of the device relationship diagram is as follows: After obtaining the device relationship tree, the node terminal relationship data of the target component can be searched in the device relationship tree according to the target component name input by the user, and the frame corresponding to each node in the node terminal relationship data is matched in the second data. Wherein, the terminal relationship data schematic diagram is shown in the following table (1).

[0077]

[0078] Table (1) Wherein, "210-1" in table (1) is the cabinet serial number, "RC-ODJ{12,14},16" is the combination name, and 1-2 and 01-04 are terminal directions. Other table contents are terminal numbers (such as 211-1005-14).

[0079] Then, according to the figure area coordinates in the node terminal relationship data, the target figure corresponding to the figure area is cropped and extracted from the frame, and the target figure is constructed and drawn to obtain a DWG file of the device relationship diagram for the user to browse online. The schematic diagram of the device relationship diagram is shown in Figure 7 . In the diagram, Figure 7 the middle annulus D2 represents a device, the serial numbers , and are frame files, the serial numbers and are partial figures in the cabinet wiring diagram corresponding to the frame files, and the serial numbers with the same numbers are in one-to-one correspondence, that is, the serial number corresponds to the serial number .

[0080] Then, the user can browse all the figures associated with the device in a single figure interface. Meanwhile, the user can also browse the frame file in which the local figure corresponding to the node of the device relationship tree is located by clicking the node figure, and can also automatically switch to the original drawing to browse all the figures. The user can also select a specified terminal to browse the cable to and from the terminal.

[0081] The new drawing (i.e., the device relationship diagram) contains all the cable and terminal relationship figures of the signal device from the indoor to the outdoor, and the user can master the signal device connection wiring, cable direction, and terminal to and from relationship based on the original drawing in a single browsing interface, thereby greatly improving the drawing browsing efficiency.

[0082] Referring to Figure 8 , the embodiment of the present application provides a railway signal drawing analysis system, which comprises: a table structure acquisition module 10 configured to pre-process all signal drawings of a target station to obtain first data, perform frame recognition and forward mapping on the first data to obtain second data, and perform table recognition on the second data to obtain table structure data.

[0083] a table text acquisition module 20 configured to identify the text in the second data by using a plurality of text recognition technologies, perform fusion operation on the plurality of original data obtained by the identification to obtain table text data.

[0084] a third data acquisition module 30 configured to perform reverse mapping on the table structure data and the table text data after merging to obtain third data.

[0085] a device relationship tree construction module 40 configured to construct a device relationship tree according to terminal relationship data extracted from the third data.

[0086] The device relationship drawing module 50 is configured to receive a target element name input by a user and search in a device relationship tree to obtain node terminal relationship data of the target element, and after finding a matching target graph in the second data, draw a device relationship graph based on the node terminal relationship data.

[0087] Exemplarily, the present application takes railway electric power engineering as-completion drawings (i.e. signal drawings) as input data, first traverses each DWG file in sequence, uses a frame recognition algorithm to segment all frames in the file, crops the frame graph to generate a new file, and saves it to a frame file library. Then traverse the frame file library in sequence, obtain the file name, type and text content by parsing the CAD file, use a table recognition algorithm (also yolo11 target recognition algorithm in the present application) to recognize the CAD file table, and then coordinate fuse the text content and table graph, so as to realize the parsing and conversion of the CAD file graph content to text content, and save it to a third database. Finally, according to the drawing rules of different types, traverse all CAD file recognition contents in sequence, establish a relationship array containing terminal direction, terminal destination, file source, graph area coordinates and device name, and then traverse all relationship arrays to establish a device relationship tree, so as to realize intelligent analysis of signal drawings.

[0088] Then, the user can input a device or terminal or cable name through the application system, automatically search the device relationship tree and frame file library, according to the direction from each indoor cabinet to outdoor equipment box, crop and extract target graphs from the original drawings according to the graph area coordinates, and finally redraw to generate a device relationship drawing file based on the original drawings. Moreover, the user can browse the local graph corresponding to the node by clicking the node graph of the device relationship tree, and also can automatically switch to the original drawing to browse all graphs.

[0089] It can be understood that the user imports all signal drawings of the station yard, the system automatically processes and constructs a device relationship tree based on the drawings, so as to generate signal drawing connection terminal direction data, signal cable route data and signal device connection wiring data, etc. The user inputs a device name or cable name or terminal name in the interactive interface, the system automatically analyzes, crops and redraws to generate a device relationship drawing based on the original drawing, which can realize real-time browsing of local drawings and one-key switching to the original drawing for browsing, thereby improving the drawing browsing efficiency. Moreover, using the signal drawing connection terminal direction data, signal cable route data and signal device connection wiring data, the functions of terminal checking, drawing checking and automatic drawing of cable route graph can be realized.

[0090] The present application also provides a device, which exemplarily includes a processor, a display, and a memory, wherein the memory stores a computer program, and the processor runs the computer program to enable the device to execute the above-mentioned railway signal drawing parsing method or the functions of each module in the above-mentioned railway signal drawing parsing system.

[0091] The processor can be an integrated circuit chip with signal processing capabilities. The processor can be a general-purpose processor, including at least one of a central processing unit (CPU), a graphics processing unit (GPU), a network processor (NP), a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. The general-purpose processor can be a microprocessor or any conventional processor, etc., and can implement or execute the various methods, steps, and logic block diagrams disclosed in the embodiments of this application.

[0092] The memory may be, but is not limited to, random access memory (RAM), read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), and electrically erasable programmable read-only memory (EEPROM). The memory is used to store computer programs, and the processor can execute the computer programs accordingly after receiving an execution instruction.

[0093] In several embodiments provided in the present application, it should be understood that the disclosed apparatus and method can also be implemented by other means. The apparatus embodiments described above are only illustrative, for example, the flowcharts and structural diagrams in the drawings show the possible implementation architecture, function and operation of the apparatus, method and computer program product according to the embodiments of the present application. In this regard, each block in the flowchart or block diagram can represent a module, a program segment or a part of code, which contains one or more executable instructions for implementing the specified logic function. It should also be noted that in alternative implementation, the functions noted in the block can also occur in different order from that noted in the drawings. For example, two consecutive blocks can actually be executed substantially in parallel, and sometimes they can be executed in reverse order, depending on the functions involved. It should also be noted that each block in the structural diagram and / or flowchart, and the combination of blocks in the structural diagram and / or flowchart, can be implemented by a dedicated hardware-based system for performing the specified function or action, or can be implemented by a combination of dedicated hardware and computer instructions.

[0094] In addition, each functional module or unit in the embodiments of the present application can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.

[0095] If the functions are implemented in the form of software function modules and sold or used as independent products, they can be stored in a computer readable storage medium. Based on this understanding, the technical solutions of the present application can be embodied in the form of software products, and the computer software products are stored in a storage medium, including a number of instructions for causing a computer device (which can be a smart phone, a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the embodiments of the present application.

[0096] The above description is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto, and any person skilled in the art can easily think of changes or replacements within the technical scope disclosed in the present application, which should be covered within the protection scope of the present application.

Claims

1. A method for analyzing railway signal drawings, characterized in that: include: Preprocessing all signal drawings of a target station to obtain first data, performing frame recognition and forward mapping on the first data in sequence to obtain second data, and then performing table recognition on the second data to obtain table structure data; Recognize the characters in the second data using multiple character recognition technologies, perform a fusion operation on the obtained multiple original data to obtain table text data, and perform reverse mapping on the table structure data and the table text data to obtain third data; constructing a device relationship tree based on the terminal relationship data extracted from the third data; Receive the target component name input by the user and search in the device relationship tree to obtain the node-terminal relationship data of the target component, and then find a matching target graphic in the second data and draw a device relationship diagram with the node-terminal relationship data.

2. The railway signal drawing analysis method according to claim 1, characterized in that: The method of preprocessing all signal drawings of the target station to obtain first data, sequentially performing frame recognition and forward mapping on the first data to obtain second data, and then performing table recognition on the second data to obtain table structure data includes: Comparing the resolution of each signal image with a plurality of preset standard resolutions, scaling the current signal image to a preset standard resolution having the smallest difference therebetween, and sequentially performing binarization and line enhancement on the scaled signal image to obtain the first data; Performing frame recognition on the first data using a target recognition algorithm, and converting the data into data in an image coordinate system using a preset coordinate mapping relationship to obtain the second data; The target recognition algorithm is used to perform table recognition on the second data to obtain the table structure data.

3. The railway signal drawing analysis method according to claim 2, characterized in that: The method comprises: using a plurality of character recognition technologies to respectively recognize characters in the second data, performing a fusion operation on the obtained plurality of original data to obtain table character data, and merging the table structure data and the table character data and performing reverse mapping thereon, including: Recognizing characters in the second data using an optical character recognition method to obtain first text information, and recognizing the characters using a target learning algorithm to obtain second text information; determining whether the second text information contains the first text information; if not, writing the first text information into the second text information; and if not, calculating the area of ​​the text regions of the first text information and the second text information respectively, retaining the text with the larger area, to obtain the table text data; Merging the table structure data and the table text data using the icon coordinates in the table structure data to obtain merged table data; The merged table data is converted into data in a world coordinate system using the preset coordinate mapping relationship to obtain the third data.

4. The railway signal drawing analysis method according to claim 1, characterized in that: The step of constructing a device relationship tree based on the terminal relationship data obtained by extracting the third data includes: The target field in the third data is extracted to obtain the terminal relationship data, so as to construct the relationship between the device name and the terminal connected thereto, and obtain the device relationship tree.

5. The railway signal drawing analysis method according to claim 4, characterized in that: The receiving of the target component name input by the user and searching the device relationship tree to obtain the node-terminal relationship data of the target component, searching the second data for a matching target graphic, and then drawing the device relationship diagram with the node-terminal relationship data, including: receiving a target component name input by a user and searching the device relationship tree to obtain node-terminal relationship data of the target component, and matching the second data with a frame corresponding to each node in the node-terminal relationship data; Separating the target graphic corresponding to the graphic area from the frame according to the graphic area coordinates in the node-terminal relationship data; The target graph is constructed and drawn to obtain the device relationship diagram for online reference by the user.

6. The railway signal drawing analysis method according to claim 2, characterized in that: The pre-processing of all signal drawings of the target station to obtain the first data further includes: The signal drawing in the standard format is structurally analyzed to identify the layer structure, view configuration, layout space and model space, and conversion parameters are determined according to preset rules. The rendering engine is used to perform vector rendering according to the conversion parameters to obtain a signal image in a lossless compressed image format.

7. The railway signal drawing analysis method according to claim 2, characterized in that: The method further comprises: performing frame recognition on the first data by using a target recognition algorithm; comparing the frame size in the second data with a preset frame threshold; If the target pattern is smaller than a preset minimum frame threshold, or the overlap rate between multiple target patterns is greater than a preset maximum frame threshold, the target recognition algorithm is used again to perform frame recognition on the first data, and the repeatedly recognized frames are deleted to obtain corrected second data; The second data includes: a frame name and frame coordinates.

8. The railway signal drawing analysis method according to claim 2, characterized in that: The method of converting the data into the image coordinate system by using the preset coordinate mapping relationship also includes: A preset coordinate mapping relationship between the image coordinate system and the world coordinate system is established to perform forward mapping or reverse mapping according to the preset coordinate mapping relationship.

9. A railway signal drawing analysis system, characterized in that: include: a table structure acquisition module, configured to pre-process all signal drawings of a target station to obtain first data, perform frame recognition and forward mapping on the first data in sequence to obtain second data, and then perform table recognition on the second data to obtain table structure data; a table text acquisition module, configured to respectively recognize characters in the second data using a plurality of character recognition technologies, and perform a fusion operation on the obtained plurality of original data to obtain table text data; A third data acquisition module is used to merge the table structure data and the table text data and then perform reverse mapping to obtain third data; a device relationship tree construction module, configured to construct a device relationship tree based on the terminal relationship data extracted from the third data; The device relationship diagram drawing module is used to receive the target component name input by the user and search in the device relationship tree to obtain the node-terminal relationship data of the target component, so as to find the matching target graphic in the second data and then draw the device relationship diagram with the node-terminal relationship data.

10. A device, characterized in that: include: A memory, a processor, a display, and a computer program stored in the memory and executable on the processor, wherein when the computer program is executed by the processor, the steps of the method for parsing a railway signal drawing according to any one of claims 1 to 8 are implemented.

Images