Aircraft matching material change identification method and system, storage medium and computer

Abstract

The invention provides an aircraft matching material change identification method and system, a storage medium and a computer, and the identification method comprises the steps: receiving an assembly instruction in an aircraft production process, and carrying out the analysis to generate a first instruction matching list; obtaining a second instruction matching list corresponding to the instruction version number of the first instruction matching list on the assembly production line, and comparing the first instruction matching list with the second instruction matching list; obtaining change items in the instruction matching list according to a comparison result, marking the change items, and performing change type division and material state division on the materials marked with the change items; and according to the divided change type and the material state, the graph number, the version and the demand quantity of the material corresponding to the original instruction matching list are refreshed so as to complete updating of the airplane matching material information. The aircraft matching material change identification method provided by the invention is high in efficiency and high in accuracy.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a method, system, storage medium and computer for identifying changes in supporting materials for an aircraft. Background Art

[0002] There are tens of thousands of parts for large parts assembly of domestic large aircraft fuselages, and the process instructions are frequently changed, which has a great impact on the warehouse and the actual assembly parts matching work. At present, many data in the new and old versions of the instruction matching list are recorded in the form of tables, and the tables recording the data are displayed to the staff through image projection. The staff compares the differences based on the displayed table images to obtain the changes.

[0003] Due to the complex table structure and text arrangement, it is difficult to identify the changes between the previous and subsequent versions of the instruction supporting lists through the existing text detection and comparison recognition methods, and the accuracy is low. Although the accuracy is improved to a certain extent through manual active recognition, the efficiency of manual recognition is low. Summary of the invention

[0004] Based on this, the purpose of the present invention is to provide a method, system, storage medium and computer for identifying changes in aircraft supporting materials to solve the technical problems existing in the prior art.

[0005] The present invention provides a method for identifying changes in aircraft supporting materials, comprising:

[0006] Receiving assembly instructions in the aircraft production process, parsing the assembly instructions, generating a first instruction matching list and using it as a material assembly basis for the warehouse;

[0007] Obtaining an instruction version number of the first instruction matching list, obtaining a second instruction matching list corresponding to the instruction version number on the assembly production line according to the instruction version number, and comparing the first instruction matching list with the second instruction matching list;

[0008] According to the comparison result, the changed items in the supporting list of the first instruction are obtained and marked, and the marked changed items are classified into change types and material states;

[0009] The drawing number, edition and required quantity of the material corresponding to the second instruction supporting list are refreshed according to the classified change type and material status to complete the update of the aircraft supporting material information.

[0010] Optionally, the step of comparing the first instruction supporting list with the second instruction supporting list includes:

[0011] Acquire table images containing tables in the first instruction supporting list and the second instruction supporting list, and pre-process the two table images respectively to improve the quality of the table images;

[0012] The preprocessed table image is input into a multi-layer convolutional neural network to identify the cells of the preprocessed table image and the center position of each cell, thereby obtaining the table structure in the table image and the position information of each cell in the table in the table image;

[0013] The preprocessed table image is input into the backbone network for feature extraction, the extracted features are used to predict the text area corner points and segmentation images, the predicted corner points are matched with the segmentation map, and the text detection area is obtained;

[0014] Scaling the image of the text detection area in equal proportion, extracting features from the scaled image, identifying the post-verification probability matrix of the characters through a circular convolutional network, obtaining the maximum post-verification probability in the post-verification probability matrix to obtain the character category and then decoding, and identifying the characters in the text detection area;

[0015] Combination matching is performed according to the table structure, the position of each cell in the table image and the recognized text, and the combination matching result is adjusted to a table document format, and the table information corresponding to the first instruction supporting list and the second instruction supporting list is obtained and then compared.

[0016] Optionally, the step of preprocessing the two table images respectively to improve the quality of the table images includes:

[0017] Performing grayscale processing on the table image to obtain a grayscale image, and performing binarization processing on the grayscale image to obtain a binarized contour of the table image;

[0018] Setting an image threshold to filter out pixel points in the binary contour that are lower than the image threshold, and performing secondary filtering on the table image at the position of the table image according to the pixel points that are not filtered out;

[0019] Obtaining a correction straight line and an edge straight line in the table image after the secondary filtering, performing comparative calculation based on the correction straight line and the edge straight line, obtaining the position and tilt angle of the image edge straight line and performing image correction;

[0020] Line compensation is performed on the table lines in the filtered and corrected table image, and the line compensation at least includes: connecting and extending the missing lines, cutting off the lines beyond the periphery of the table, and expanding and thickening the lines whose line width is less than a width threshold.

[0021] Optionally, the step of inputting the preprocessed table image into the backbone network for feature extraction, predicting the text area corner points and the segmentation image for the extracted features, matching the predicted corner points with the segmentation image, and obtaining the text detection area includes:

[0022] Adding feature pyramid and deconvolution to the single target multi-stage detection network to construct a feature extraction model, and using the preprocessed table image as input of the feature extraction model to extract a feature map;

[0023] Predicting corner boxes based on the extracted feature map, taking the center point of the predicted corner box as the anchor point of the feature map, sampling the feature map using the prior box, and calculating the intersection and union ratio of the prior box and the annotation information on the table image based on the corner detection network to obtain the corner point position of the text area;

[0024] The midpoint of the border of the text area is calculated according to the position of the corner point of the text area, and the position-sensitive segmentation map of each cell is obtained according to the midpoint of the border of the text area and the position of the corner point of the text area;

[0025] The corner points of the text area are filtered based on a preset exclusion rule to exclude corner points that are not in the same text area, and the prediction box composed of the remaining corner points and the position-sensitive segmentation map are average-pooled in the region of interest to obtain the target text detection area.

[0026] Optionally, the step of obtaining and comparing table information corresponding to the first instruction matching list and the second instruction matching list comprises:

[0027] Obtain a first table text corresponding to the first instruction supporting list and a second table text corresponding to the second instruction supporting list;

[0028] Using the first table text and the second table text as inputs of an Apache POI document processor to obtain a set of horizontal titles and vertical titles corresponding to the two table texts;

[0029] Vectorize the horizontal and vertical titles in the two table texts, take the upper left corner of the table text as the starting point, and perform relevance scoring according to the distance of the horizontal and vertical titles from the starting point to adjust the title weights of the horizontal and vertical titles;

[0030] The texts of the two tables are compared based on the horizontal and vertical titles after weight adjustment to obtain the comparison results of the horizontal and vertical titles and the corresponding contents in each title.

[0031] The present invention also provides an aircraft supporting material change identification system, comprising:

[0032] A parsing module, used for receiving assembly instructions in the aircraft production process, parsing the assembly instructions, generating a first instruction matching list and serving as a material assembly basis for the warehouse;

[0033] a comparison module, configured to obtain an instruction version number of the first instruction matching list, obtain a second instruction matching list corresponding to the instruction version number on the assembly production line according to the instruction version number, and compare the first instruction matching list with the second instruction matching list;

[0034] A classification module, used for obtaining and marking the changed items in the supporting list of the first instruction according to the comparison result, and classifying the marked changed item materials by change type and material status;

[0035] The updating module is used to update the drawing number, edition and required quantity of the material corresponding to the second instruction supporting list according to the divided change type and material status, so as to complete the update of the aircraft supporting material information.

[0036] Optionally, the comparison module specifically includes:

[0037] a preprocessing unit, used for acquiring table images containing tables in the first instruction supporting list and the second instruction supporting list, and preprocessing the two table images respectively to improve the quality of the table images;

[0038] A first recognition unit is used to input the preprocessed table image into a multi-layer convolutional neural network, identify the cells of the preprocessed table image and the center position of each cell, and then obtain the table structure in the table image and the position information of each cell in the table in the table image;

[0039] An extraction unit is used to input the preprocessed table image into the backbone network for feature extraction, predict the text area corner points and segmentation image for the extracted features, match the predicted corner points with the segmentation map, and obtain the text detection area;

[0040] A second recognition unit is used to scale the image of the text detection area, extract features from the scaled image, recognize the post-verification probability matrix of the characters through a circular convolutional network, obtain the maximum post-verification probability in the post-verification probability matrix to obtain the character category and then decode it to recognize the characters in the text detection area;

[0041] The adjustment unit is used to perform combination matching according to the table structure, the position of each cell in the table image and the recognized text, and adjust the combination matching result to a table document format, and then compare the table information corresponding to the first instruction supporting list and the second instruction supporting list.

[0042] Optionally, the preprocessing unit is specifically used for:

[0043] Performing grayscale processing on the table image to obtain a grayscale image, and performing binarization processing on the grayscale image to obtain a binarized contour of the table image;

[0044] Setting an image threshold to filter out pixel points in the binary contour that are lower than the image threshold, and performing secondary filtering on the table image at the position of the table image according to the pixel points that are not filtered out;

[0045] Obtaining a correction straight line and an edge straight line in the table image after the secondary filtering, performing comparative calculation based on the correction straight line and the edge straight line, obtaining the position and tilt angle of the image edge straight line and performing image correction;

[0046] Line compensation is performed on the table lines in the filtered and corrected table image, and the line compensation at least includes: connecting and extending the missing lines, cutting off the lines beyond the periphery of the table, and expanding and thickening the lines whose line width is less than a width threshold.

[0047] The present invention also provides a storage medium on which a computer program is stored, and when the program is executed by a processor, the above-mentioned aircraft supporting material change identification method is implemented.

[0048] The present invention also proposes a computer, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned aircraft supporting material change identification method when executing the computer program.

[0049] Compared with the prior art, the present invention has the following beneficial effects: the aircraft supporting material change identification method provided by the present application first receives assembly instructions in the aircraft production process, and generates a corresponding first instruction supporting list, compares the first instruction supporting list with the second instruction supporting list existing in the assembly production line, obtains and marks the change items in the instruction supporting list according to the comparison results, and classifies the materials with the marked change items into change types and material statuses; refreshes the drawing number, edition and required quantity of the materials corresponding to the original instruction supporting list according to the classified change types and material statuses, so as to update the aircraft supporting material information; automatically refreshes the original instruction supporting material list data based on real-time reception and comparison of change data, does not need to rely on manual identification and judgment, greatly improves the efficiency of actual warehouse material matching operations, has high data accuracy, and can effectively connect upstream and downstream data, facilitates product quality traceability, and is suitable for large-scale promotion.

[0050] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 This is a flow chart of the aircraft supporting material change identification method in Embodiment 1 of the present invention;

[0052] Figure 2 This is a structural block diagram of a computer in Embodiment 4 of the present invention.

[0053] The following specific implementation manner will further illustrate the present invention in conjunction with the above-mentioned drawings. DETAILED DESCRIPTION

[0054] In order to facilitate the understanding of the present invention, the present invention will be described more fully below with reference to the relevant drawings. Several embodiments of the present invention are given in the drawings. However, the present invention can be implemented in many different forms and is not limited to the embodiments described herein. On the contrary, the purpose of providing these embodiments is to make the disclosure of the present invention more thorough and comprehensive.

[0055] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art of the present invention. The terms used herein in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The term "and / or" used herein includes any and all combinations of one or more related listed items.

[0056] Embodiment 1

[0057] See also Figure 1 , which shows a method for identifying changes in supporting materials for an aircraft in a first embodiment of the present invention. The method for identifying changes in supporting materials for an aircraft specifically includes steps S10 to S40:

[0058] S10, receiving assembly instructions in the aircraft production process, parsing the assembly instructions, generating a first instruction matching list and using it as a material assembly basis for the warehouse;

[0059] Optionally, in this embodiment, the assembly production line can automatically receive production assembly instructions issued after confirmation, and the received information includes the instruction number, the version number of the instruction, the manufacturing serial number, the kit number of the instruction, the material details of the supporting list, etc.; the issued assembly instructions are parsed, and the corresponding instruction supporting list is generated and used as the basis for material assembly in the warehouse.

[0060] S20, obtaining an instruction version number of the first instruction matching list, obtaining a second instruction matching list corresponding to the instruction version number on the assembly production line according to the instruction version number, and comparing the first instruction matching list with the second instruction matching list;

[0061] Optionally, on the assembly line, due to the large number of parts to be assembled, the process instructions are frequently changed; before the latest instruction list is generated, the last changed instruction list has been saved on the assembly line; therefore, in order to find the saved instruction list in the system, it is necessary to parse the instruction number, manufacturing serial number, instruction kit number, instruction version number and other information based on the received production assembly instructions, and obtain the instruction list corresponding to the current instruction version number on the assembly line based on the instruction version number; then automatically compare the two instruction lists to find the location where the instruction list has changed.

[0062] Optionally, in this embodiment, the step of comparing the first instruction matching list with the second instruction matching list includes:

[0063] Acquire table images containing tables in the first instruction supporting list and the second instruction supporting list, and pre-process the two table images respectively to improve the quality of the table images;

[0064] The preprocessed table image is input into a multi-layer convolutional neural network to identify the cells of the preprocessed table image and the center position of each cell, thereby obtaining the table structure in the table image and the position information of each cell in the table in the table image;

[0065] The preprocessed table image is input into the backbone network for feature extraction, the extracted features are used to predict the text area corner points and segmentation images, the predicted corner points are matched with the segmentation map, and the text detection area is obtained;

[0066] Scaling the image of the text detection area in equal proportion, extracting features from the scaled image, identifying the post-verification probability matrix of the characters through a circular convolutional network, obtaining the maximum post-verification probability in the post-verification probability matrix to obtain the character category and then decoding, and identifying the characters in the text detection area;

[0067] Combination matching is performed according to the table structure, the position of each cell in the table image and the recognized text, and the combination matching result is adjusted to a table document format, and the table information corresponding to the first instruction supporting list and the second instruction supporting list is obtained and then compared.

[0068] Optionally, in this embodiment, since the instruction support list often records data in the form of a table, for the changes between the previous and subsequent versions of the instruction support list, how to accurately identify the table and accurately locate the table text is crucial to timely identifying the aircraft supporting material change information.

[0069] The step of preprocessing the two table images respectively to improve the quality of the table images comprises:

[0070] Performing grayscale processing on the table image to obtain a grayscale image, and performing binarization processing on the grayscale image to obtain a binarized contour of the table image;

[0071] Setting an image threshold to filter out pixel points in the binary contour that are lower than the image threshold, and performing secondary filtering on the table image at the position of the table image according to the pixel points that are not filtered out;

[0072] Obtaining a correction straight line and an edge straight line in the table image after the secondary filtering, performing comparative calculation based on the correction straight line and the edge straight line, obtaining the position and tilt angle of the image edge straight line and performing image correction;

[0073] Line compensation is performed on the table lines in the filtered and corrected table image, and the line compensation at least includes: connecting and extending the missing lines, cutting off the lines beyond the periphery of the table, and expanding and thickening the lines whose line width is less than a width threshold.

[0074] The steps of inputting the preprocessed table image into the backbone network for feature extraction, predicting the corner points of the text area and the segmentation image for the extracted features, matching the predicted corner points with the segmentation image, and obtaining the text detection area include:

[0075] Adding feature pyramid and deconvolution to the single target multi-stage detection network to construct a feature extraction model, and using the preprocessed table image as input of the feature extraction model to extract a feature map;

[0076] Predicting corner boxes based on the extracted feature map, taking the center point of the predicted corner box as the anchor point of the feature map, sampling the feature map using the prior box, and calculating the intersection and union ratio of the prior box and the annotation information on the table image based on the corner detection network to obtain the corner point position of the text area;

[0077] The midpoint of the border of the text area is calculated according to the position of the corner point of the text area, and the position-sensitive segmentation map of each cell is obtained according to the midpoint of the border of the text area and the position of the corner point of the text area;

[0078] The corner points of the text area are filtered based on a preset exclusion rule to exclude corner points that are not in the same text area, and the prediction box composed of the remaining corner points and the position-sensitive segmentation map are average-pooled in the region of interest to obtain the target text detection area.

[0079] Table recognition and text recognition are a method for processing document type images. Before recognition, image preprocessing is required to provide good conditions for accurately obtaining table data in the later stage. Through grayscale and binarization processing and then pixel value filtering, the interference caused by background color and acquisition light in the image can be effectively reduced. By comparing the through straight line with the edge straight line in the image, the position and inclination angle of the edge straight line of the image are obtained and the image is corrected, the influence of table tilt can be reduced. However, filtering and correction may cause partial loss of table frame lines. In order to avoid the loss of table lines caused by acquisition, the preprocessing process also includes line compensation for the table lines in the filtered and corrected table image. After the preprocessing process, the image features are more obvious, which is conducive to the subsequent recognition of tables and content.

[0080] The input of the multi-layer convolutional neural network is a preprocessed image, and the output is a table and the position of each cell in the table in the image; the corner points of the text area are predicted by the corner point detection network, and the text content is segmented and predicted by the region segmentation network. In this embodiment, when performing segmentation map prediction, the technology of position-sensitive segmentation map is added to divide the text segmentation area into sub-areas. When predicting the text box, the large-grained prediction results are used to assist in filtering prediction; finally, the predicted corner points and the segmentation image are matched, and the prediction results are output. The combination of corner points and segmentation images has a better detection effect on text areas of different lengths and widths in the image.

[0081] Further, the step of obtaining the table information corresponding to the first instruction matching list and the second instruction matching list and then comparing them includes:

[0082] Obtain a first table text corresponding to the first instruction supporting list and a second table text corresponding to the second instruction supporting list;

[0083] Using the first table text and the second table text as inputs of an Apache POI document processor to obtain a set of horizontal titles and vertical titles corresponding to the two table texts;

[0084] Vectorize the horizontal and vertical titles in the two table texts, take the upper left corner of the table text as the starting point, and perform relevance scoring according to the distance of the horizontal and vertical titles from the starting point to adjust the title weights of the horizontal and vertical titles;

[0085] The texts of the two tables are compared based on the horizontal and vertical titles after weight adjustment to obtain the comparison results of the horizontal and vertical titles and the corresponding contents in each title.

[0086] In this embodiment, Apache POI is used as a document processor to effectively extract horizontal and vertical titles at special positions in two table texts, and can further extract multi-level titles in horizontal or vertical titles, and store the multi-level titles in a collection container; the titles are vectorized to convert the text matching problem into a vector calculation problem in a vector space, and the similarity of the vectors is calculated through a model, which greatly simplifies the complexity of the similarity calculation; further, for the special structure of the table data, the upper left corner of the table text is taken as the starting point. It is generally believed that the closer the title is to the starting point, the higher the weight, and the greater the impact on the assembly process after the change. Therefore, in this embodiment, the relevance score is performed according to the distance of the horizontal and vertical titles from the starting point to adjust the title weights of the horizontal and vertical titles; and then the different change items in the new and old versions of the instruction matching list are distinguished.

[0087] S30, obtaining and marking the changed items in the first instruction supporting list according to the comparison result, and classifying the marked changed item materials by change type and material status;

[0088] S40, refreshing the drawing number, edition and required quantity of the material corresponding to the second instruction supporting list according to the classified change type and material status, so as to complete the update of the aircraft supporting material information.

[0089] Optionally, after comparing the first instruction supporting list with the second instruction supporting list, the change results can be differentiated into change types according to the required quantity, including but not limited to adding supporting materials, canceling supporting materials, increasing the number of supporting materials, reducing the number of supporting materials, etc.; obtain the supporting status of the materials in the original instruction supporting list details, and determine whether the materials have been supported or shipped out through the supporting and outbound records of the material list. Through the change type and material status of the material change item in the supporting list, when the changeable conditions are met, the drawing number, edition and required quantity of the corresponding material in the original instruction supporting material list are automatically refreshed to update the material information to the latest information.

[0090] In summary, the aircraft supporting material change identification method provided by the present application first receives the assembly instructions in the aircraft production process, and generates a corresponding first instruction supporting list, compares the first instruction supporting list with the second instruction supporting list existing in the assembly production line, and mainly stores data in the form of a table. The table structure and text arrangement are complex, which makes manual recognition difficult. After preprocessing the acquired table image, a multi-layer convolutional neural network is used to identify the table structure in the preprocessed table image and the position information of each cell in the table in the table image, and a text detection area is obtained based on the backbone network. Finally, a circular convolutional network is used to identify the text in the text detection area, and a combination match is performed according to the table structure, cell position and recognized text, and the text is adjusted to a table document format for comparison, thereby improving the efficiency of table data comparison; according to the real-time reception and comparison of the change data, the original instruction supporting material list data is automatically refreshed, without relying on human judgment, and the efficiency of the actual warehouse material matching operation is greatly improved. The data accuracy is high, and the upstream and downstream data can be effectively connected in series, which is convenient for product quality traceability and suitable for large-scale promotion.

[0091] Embodiment 2

[0092] This embodiment provides an aircraft supporting material change identification system, including:

[0093] A parsing module, used for receiving assembly instructions in the aircraft production process, parsing the assembly instructions, generating a first instruction matching list and serving as a material assembly basis for the warehouse;

[0094] a comparison module, configured to obtain an instruction version number of the first instruction matching list, obtain a second instruction matching list corresponding to the instruction version number on the assembly production line according to the instruction version number, and compare the first instruction matching list with the second instruction matching list;

[0095] A classification module, used for obtaining and marking the changed items in the supporting list of the first instruction according to the comparison result, and classifying the marked changed item materials by change type and material status;

[0096] The updating module is used to update the drawing number, edition and required quantity of the material corresponding to the second instruction supporting list according to the divided change type and material status, so as to complete the update of the aircraft supporting material information.

[0097] The comparison module specifically includes:

[0098] a preprocessing unit, used for acquiring table images containing tables in the first instruction supporting list and the second instruction supporting list, and preprocessing the two table images respectively to improve the quality of the table images;

[0099] A first recognition unit is used to input the preprocessed table image into a multi-layer convolutional neural network, identify the cells of the preprocessed table image and the center position of each cell, and then obtain the table structure in the table image and the position information of each cell in the table in the table image;

[0100] An extraction unit is used to input the preprocessed table image into the backbone network for feature extraction, predict the text area corner points and segmentation image for the extracted features, match the predicted corner points with the segmentation map, and obtain the text detection area;

[0101] A second recognition unit is used to scale the image of the text detection area, extract features from the scaled image, recognize the post-verification probability matrix of the characters through a circular convolutional network, obtain the maximum post-verification probability in the post-verification probability matrix to obtain the character category and then decode it to recognize the characters in the text detection area;

[0102] The adjustment unit is used to perform combination matching according to the table structure, the position of each cell in the table image and the recognized text, and adjust the combination matching result to a table document format, and then compare the table information corresponding to the first instruction supporting list and the second instruction supporting list.

[0103] The pre-processing unit is specifically used for:

[0104] Performing grayscale processing on the table image to obtain a grayscale image, and performing binarization processing on the grayscale image to obtain a binarized contour of the table image;

[0105] Setting an image threshold to filter out pixel points in the binary contour that are lower than the image threshold, and performing secondary filtering on the table image at the position of the table image according to the pixel points that are not filtered out;

[0106] Obtaining a correction straight line and an edge straight line in the table image after the secondary filtering, performing comparative calculation based on the correction straight line and the edge straight line, obtaining the position and tilt angle of the image edge straight line and performing image correction;

[0107] Line compensation is performed on the table lines in the filtered and corrected table image, and the line compensation at least includes: connecting and extending the missing lines, cutting off the lines beyond the periphery of the table, and expanding and thickening the lines whose line width is less than a width threshold.

[0108] The extraction unit is specifically used for:

[0109] Adding feature pyramid and deconvolution to the single target multi-stage detection network to construct a feature extraction model, and using the preprocessed table image as input of the feature extraction model to extract a feature map;

[0110] Predicting corner boxes based on the extracted feature map, taking the center point of the predicted corner box as the anchor point of the feature map, sampling the feature map using the prior box, and calculating the intersection and union ratio of the prior box and the annotation information on the table image based on the corner detection network to obtain the corner point position of the text area;

[0111] The midpoint of the border of the text area is calculated according to the position of the corner point of the text area, and the position-sensitive segmentation map of each cell is obtained according to the midpoint of the border of the text area and the position of the corner point of the text area;

[0112] The corner points of the text area are filtered based on a preset exclusion rule to exclude corner points that are not in the same text area, and the prediction box composed of the remaining corner points and the position-sensitive segmentation map are average-pooled in the region of interest to obtain the target text detection area.

[0113] The adjustment unit is specifically used for:

[0114] Obtain a first table text corresponding to the first instruction supporting list and a second table text corresponding to the second instruction supporting list;

[0115] Using the first table text and the second table text as inputs of an Apache POI document processor to obtain a set of horizontal titles and vertical titles corresponding to the two table texts;

[0116] Vectorize the horizontal and vertical titles in the two table texts, take the upper left corner of the table text as the starting point, and perform relevance scoring according to the distance of the horizontal and vertical titles from the starting point to adjust the title weights of the horizontal and vertical titles;

[0117] The texts of the two tables are compared based on the horizontal and vertical titles after weight adjustment to obtain the comparison results of the horizontal and vertical titles and the corresponding contents in each title.

[0118] Embodiment 3

[0119] This embodiment provides a storage medium on which a computer program is stored. When the program is executed by a processor, the above-mentioned aircraft supporting material change identification method is implemented.

[0120] Embodiment 4

[0121] The present invention also provides a computer, see Figure 2 , shown is a computer in an embodiment of the present invention, including a memory 10, a processor 20, and a computer program 30 stored in the memory 10 and executable on the processor 20. When the processor 20 executes the computer program 30, the above-mentioned aircraft supporting material change identification method is implemented.

[0122] The memory 10 includes at least one type of storage medium, including flash memory, hard disk, multimedia card, card-type memory (e.g., SD or DX memory, etc.), magnetic memory, disk, optical disk, etc. In some embodiments, the memory 10 may be an internal storage unit of a computer, such as a hard disk of the computer. In other embodiments, the memory 10 may also be an external storage device, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card, etc. Further, the memory 10 may also include both an internal storage unit of a computer and an external storage device. The memory 10 may be used not only to store application software and various types of data installed in the computer, but also to temporarily store data that has been output or is to be output.

[0123] Among them, in some embodiments, the processor 20 can be an electronic control unit (Electronic Control Unit, abbreviated as ECU, also known as a vehicle computer), a central processing unit (Central Processing Unit, CPU), a controller, a microcontroller, a microprocessor or other data processing chip, used to run the program code stored in the memory 10 or process data, such as executing access restriction programs, etc.

[0124] It should be pointed out that Figure 2 The structure shown does not constitute a limitation on the computer. In other embodiments, the computer may include fewer or more components than shown in the figure, or combine certain components, or arrange the components differently.

[0125] Those skilled in the art will appreciate that the logic and / or steps represented in the flowchart or otherwise described herein, for example, may be considered as an ordered list of executable instructions for implementing logical functions, and may be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For purposes of this specification, a "computer-readable medium" may be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0126] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.

[0127] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, a plurality of steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or a combination thereof: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0128] The technical features of the above-described embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0129] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the attached claims.

Claims

1. A method for identifying changes in aircraft supporting materials, characterized in that: include: Receiving assembly instructions in the aircraft production process, parsing the assembly instructions, generating a first instruction matching list and using it as a material assembly basis for the warehouse; Obtaining an instruction version number of the first instruction matching list, obtaining a second instruction matching list corresponding to the instruction version number on the assembly production line according to the instruction version number, and comparing the first instruction matching list with the second instruction matching list; According to the comparison result, the changed items in the supporting list of the first instruction are obtained and marked, and the marked changed items are classified into change types and material states; The drawing number, edition and required quantity of the material corresponding to the second instruction supporting list are refreshed according to the classified change type and material status to complete the update of the aircraft supporting material information.

2. The method for identifying changes in aircraft supporting materials according to claim 1, characterized in that: The step of comparing the first instruction supporting list with the second instruction supporting list comprises: Acquire table images containing tables in the first instruction supporting list and the second instruction supporting list, and pre-process the two table images respectively to improve the quality of the table images; The preprocessed table image is input into a multi-layer convolutional neural network to identify the cells of the preprocessed table image and the center position of each cell, thereby obtaining the table structure in the table image and the position information of each cell in the table in the table image; The preprocessed table image is input into the backbone network for feature extraction, the extracted features are used to predict the text area corner points and segmentation images, the predicted corner points are matched with the segmentation map, and the text detection area is obtained; Scaling the image of the text detection area in equal proportion, extracting features from the scaled image, identifying the post-verification probability matrix of the characters through a circular convolutional network, obtaining the maximum post-verification probability in the post-verification probability matrix to obtain the character category and then decoding, and identifying the characters in the text detection area; Combination matching is performed according to the table structure, the position of each cell in the table image and the recognized text, and the combination matching result is adjusted to a table document format, and the table information corresponding to the first instruction supporting list and the second instruction supporting list is obtained and then compared.

3. The method for identifying changes in aircraft supporting materials according to claim 2, characterized in that: The step of preprocessing the two table images respectively to improve the quality of the table images comprises: Performing grayscale processing on the table image to obtain a grayscale image, and performing binarization processing on the grayscale image to obtain a binarized contour of the table image; Setting an image threshold to filter out pixel points in the binary contour that are lower than the image threshold, and performing secondary filtering on the table image at the position of the table image according to the pixel points that are not filtered out; Obtaining a correction straight line and an edge straight line in the table image after the secondary filtering, performing comparative calculation based on the correction straight line and the edge straight line, obtaining the position and tilt angle of the image edge straight line and performing image correction; Line compensation is performed on the table lines in the filtered and corrected table image, and the line compensation at least includes: connecting and extending the missing lines, cutting off the lines beyond the periphery of the table, and expanding and thickening the lines whose line width is less than a width threshold.

4. The method for identifying changes in aircraft supporting materials according to claim 2, characterized in that: The steps of inputting the preprocessed table image into the backbone network for feature extraction, predicting the corner points of the text area and the segmentation image for the extracted features, matching the predicted corner points with the segmentation image, and obtaining the text detection area include: Adding feature pyramid and deconvolution to the single target multi-stage detection network to construct a feature extraction model, and using the preprocessed table image as input of the feature extraction model to extract a feature map; Predicting corner boxes based on the extracted feature map, taking the center point of the predicted corner box as the anchor point of the feature map, sampling the feature map using the prior box, and calculating the intersection and union ratio of the prior box and the annotation information on the table image based on the corner detection network to obtain the corner point position of the text area; The midpoint of the border of the text area is calculated according to the position of the corner point of the text area, and the position-sensitive segmentation map of each cell is obtained according to the midpoint of the border of the text area and the position of the corner point of the text area; The corner points of the text area are filtered based on a preset exclusion rule to exclude corner points that are not in the same text area, and the prediction box composed of the remaining corner points and the position-sensitive segmentation map are average-pooled in the region of interest to obtain the target text detection area.

5. The method for identifying changes in aircraft supporting materials according to claim 2, characterized in that: The step of obtaining the table information corresponding to the first instruction matching list and the second instruction matching list and then comparing them comprises: Obtain a first table text corresponding to the first instruction supporting list and a second table text corresponding to the second instruction supporting list; Using the first table text and the second table text as inputs of an Apache POI document processor to obtain a set of horizontal titles and vertical titles corresponding to the two table texts; Vectorize the horizontal and vertical titles in the two table texts, take the upper left corner of the table text as the starting point, and perform relevance scoring according to the distance of the horizontal and vertical titles from the starting point to adjust the title weights of the horizontal and vertical titles; The texts of the two tables are compared based on the weight-adjusted horizontal and vertical titles to obtain comparison results of the horizontal titles, vertical titles and the corresponding contents of each title.

6. An aircraft supporting material change identification system, characterized in that: include; A parsing module, used for receiving assembly instructions in the aircraft production process, parsing the assembly instructions, generating a first instruction matching list and serving as a material assembly basis for the warehouse; a comparison module, configured to obtain an instruction version number of the first instruction matching list, obtain a second instruction matching list corresponding to the instruction version number on the assembly production line according to the instruction version number, and compare the first instruction matching list with the second instruction matching list; A classification module, used for obtaining and marking the changed items in the supporting list of the first instruction according to the comparison result, and classifying the marked changed item materials by change type and material status; The updating module is used to update the drawing number, edition and required quantity of the material corresponding to the second instruction supporting list according to the divided change type and material status, so as to complete the update of the aircraft supporting material information.

7. The aircraft supporting material change identification system according to claim 6, characterized in that: The comparison module specifically includes: a preprocessing unit, used for acquiring table images containing tables in the first instruction supporting list and the second instruction supporting list, and preprocessing the two table images respectively to improve the quality of the table images; A first recognition unit is used to input the preprocessed table image into a multi-layer convolutional neural network, identify the cells of the preprocessed table image and the center position of each cell, and then obtain the table structure in the table image and the position information of each cell in the table in the table image; An extraction unit is used to input the preprocessed table image into the backbone network for feature extraction, predict the text area corner points and segmentation image for the extracted features, match the predicted corner points with the segmentation map, and obtain the text detection area; A second recognition unit is used to scale the image of the text detection area, extract features from the scaled image, recognize the post-verification probability matrix of the characters through a circular convolutional network, obtain the maximum post-verification probability in the post-verification probability matrix to obtain the character category and then decode it to recognize the characters in the text detection area; The adjustment unit is used to perform combination matching according to the table structure, the position of each cell in the table image and the recognized text, and adjust the combination matching result to a table document format, and then compare the table information corresponding to the first instruction supporting list and the second instruction supporting list.

8. The aircraft supporting material change identification system according to claim 7, characterized in that: The pre-processing unit is specifically used for: Performing grayscale processing on the table image to obtain a grayscale image, and performing binarization processing on the grayscale image to obtain a binarized contour of the table image; Setting an image threshold to filter out pixel points in the binary contour that are lower than the image threshold, and performing secondary filtering on the table image at the position of the table image according to the pixel points that are not filtered out; Obtaining a correction straight line and an edge straight line in the table image after the secondary filtering, performing comparative calculation based on the correction straight line and the edge straight line, obtaining the position and tilt angle of the image edge straight line and performing image correction; Line compensation is performed on the table lines in the filtered and corrected table image, and the line compensation at least includes: connecting and extending the missing lines, cutting off the lines beyond the periphery of the table, and expanding and thickening the lines whose line width is less than a width threshold.

9. A storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method for identifying changes in aircraft supporting materials as described in any one of claims 1 to 5 is implemented.

10. A computer comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the method for identifying changes in aircraft supporting materials as described in any one of claims 1 to 5 is implemented.

Images