Drawing recognition method and drawing recognition device

Abstract

This invention provides a drawing recognition method and a drawing recognition device that can accurately extract symbols and annotations from drawings and associate them with each other to create data. [Solution] A drawing recognition method for extracting symbols and annotations contained in a drawing file, comprising: a drawing file formation step of forming a symbol extraction drawing file S2 from an original drawing file S1; a symbol extraction step of extracting symbols from the symbol extraction drawing file S2; an annotation search area setting step of setting an annotation search area SA centered on each of the multiple symbols extracted in the symbol extraction step; and an annotation extraction step of extracting annotations corresponding to each symbol from the figures detected in the annotation search area SA centered on each of the symbols extracted in the annotation search area setting step.

Description

Technical Field

[0001] The present invention relates to a drawing recognition method and a drawing recognition apparatus. More specifically, the present invention relates to a drawing recognition method and a drawing recognition apparatus for associating and extracting characters and the like that supplement specifications of facilities, equipment, etc. represented by various symbols and symbols described near the symbols from various drawings.

Background Art

[0002] Currently, in the creation of various drawings, the creation of drawings using CAD software has become common. When drawings created using CAD software are provided, they may be provided not in CAD data but in data converted to PDF format or paper-based drawings (hereinafter sometimes simply referred to as paper-based drawings, etc.), and there may be a need to obtain various information described in these drawings. For example, in the case of architectural design drawings, there are many pieces of information such as facility symbols (hereinafter sometimes simply referred to as symbols), characters and figures related to the symbols (hereinafter sometimes simply referred to as side notes), wiring, etc., and it may be necessary to extract the symbols and side notes. However, in architectural design drawings, since symbols, side notes, and wiring, etc. are described in a state where they overlap or symbols are very close to each other, it is difficult to accurately extract the symbols and side notes, and technologies for accurately extracting symbols and side notes from paper-based drawings, etc. have been developed (see Patent Documents 1 to 5).

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Patent Document 2

Patent Document 3

Patent Document 4

Patent Document 5

[0004] The technologies described in Patent Documents 1 to 5 extract figures and text from drawings using template matching. However, in drawings containing wiring diagrams, symbols, etc., it is difficult to perform appropriate template matching, and it is difficult to extract symbols and other elements with high accuracy. In particular, when symbols and annotations are written in association, it is difficult to extract symbols and annotations with high accuracy while maintaining their relationship.

[0005] In view of the above circumstances, the present invention aims to provide a drawing recognition method and a drawing recognition device that can accurately extract symbols and annotations from drawings and associate them with each other to create data. [Means for solving the problem]

[0006] <Method for recognizing drawings> The drawing recognition method of the first invention is a drawing recognition method for extracting symbols and annotations contained in a drawing file, characterized by performing the following steps: a drawing file formation step of forming a symbol extraction drawing file for use in symbol extraction from the drawing file; a symbol extraction step of extracting symbols from the symbol extraction drawing file formed in the drawing file formation step; an annotation search area setting step of setting annotation search areas in the symbol extraction drawing file for each of the multiple symbols extracted in the symbol extraction step, with each annotation search area centered on the symbol; and an annotation extraction step of extracting annotations corresponding to each symbol from figures detected in the annotation search areas set in the annotation search area setting step. The drawing recognition method of the second invention is characterized in that, in the annotation extraction step of the first invention, the size of the annotation template registered in the annotation template list is changed while matching the annotation template with the figure detected in the annotation search area, and the annotations included in the annotation search area are extracted. The drawing recognition method of the third invention is characterized in that, in the annotation search area setting step, in the annotation search area set for a plurality of representative symbols extracted in the symbol extraction step, a matching process is performed between the annotation template and the figures detected in the annotation search area while changing the size of the annotation template to determine the extraction template size which is the size of the annotation template that yields the highest similarity, and in the annotation extraction step, annotations included in the annotation search area are extracted using the annotation template of the extraction template size determined in the annotation search area setting step. The drawing recognition method of the fourth invention is characterized in that, in the annotation extraction step of the first invention, a plurality of figures included in the annotation search area are extracted as annotation candidates, and the annotation candidate that is closest in distance to the symbol is associated with the symbol. The drawing recognition method of the fifth invention is characterized in that, in the annotation extraction step of the first invention, a plurality of figures included in the annotation search area are extracted as annotation candidates, and for the extracted plurality of annotation candidates, an annotation candidate to be associated with the symbol that set the annotation search area is determined based on a management list that stores the relationship between symbols and annotations. <Drawing Recognition Device> The sixth invention is a drawing recognition device for extracting symbols and annotations contained in a drawing file, the drawing recognition device comprising: a drawing file forming unit that forms a symbol extraction drawing file for use in symbol extraction from the drawing file; a symbol extraction unit that extracts symbols from the symbol extraction drawing file formed by the drawing file forming unit; an annotation search area setting unit that sets an annotation search area centered on each symbol in the symbol extraction drawing file for each of the multiple symbols extracted by the symbol extraction unit; and an annotation extraction unit that extracts annotations corresponding to each symbol from figures detected in the annotation search area set by the annotation search area setting unit. The drawing recognition device of the seventh invention is characterized in that, in the first invention, the annotation extraction unit performs a matching process between the annotation template and the figure detected in the annotation search area while changing the size of the annotation template registered in the annotation template list, and extracts the annotations included in the annotation search area. The drawing recognition device of the eighth invention is characterized in that, in the seventh invention, the annotation search area setting unit has a function to determine the extraction template size, which is the size of the annotation template that yields the highest similarity, by performing a matching process between the annotation template and the figures detected in the annotation search area while changing the size of the annotation template in the annotation search area set for a plurality of representative symbols extracted by the symbol extraction unit, and the annotation extraction unit has a function to extract annotations included in the annotation search area using the annotation template of the extraction template size determined in the annotation search area setting unit. The drawing recognition device of the ninth invention is characterized in that, in the sixth invention, the annotation extraction unit has a function of extracting a plurality of figures included in the annotation search area as annotation candidates, and a function of associating the annotation candidate that is closest in distance to the symbol from among the plurality of annotation candidates with the symbol. The drawing recognition device of the 10th invention is characterized in that, in the 6th invention, the annotation extraction unit has a function to extract a plurality of figures included in the annotation search area as annotation candidates, and has a function to determine the annotation candidates to be associated with the symbol that sets the annotation search area from which the annotation candidates were extracted, based on a management list that stores the relationship between symbols and annotations. [Effects of the Invention]

[0007] <Method for recognizing drawings> According to the first to third inventions, the extraction accuracy of marginal notes can be increased, and related symbols and marginal notes can be appropriately linked and stored. According to the fourth invention, a symbol can be easily linked to a related annotation. According to the fifth invention, the accuracy of linking a symbol with a highly related annotation can be improved. <Drawing Recognition Device> According to the sixth to eighth inventions, the extraction accuracy of marginal notes can be increased, and related symbols and marginal notes can be appropriately linked and stored. According to the ninth invention, a symbol can be easily linked to a related annotation. According to the tenth invention, the accuracy of linking a symbol with a highly related annotation can be increased. [Brief explanation of the drawing]

[0008] [Figure 1] This is a flowchart of the drawing recognition method in this embodiment. [Figure 2] This is a schematic block diagram of the drawing recognition device 1 of this embodiment. [Figure 3] (A) is a diagram showing an example of the original drawing file S1, and (B) is a diagram showing an example of the symbol extraction drawing file S2. [Figure 4] (A) is a diagram showing an example of a state in which a side-reading search area SA has been set for each symbol included in the symbol extraction drawing file S2, and (B) is a diagram showing an example of a state in which a side-reading search area SA has been set for a selected symbol included in the symbol extraction drawing file S2. [Figure 5] (A) is a diagram showing an example of an extracted symbol file S4, (B) is a diagram showing an example of a side note template list TP, and (C) is a diagram showing an example of a management list CR. [Modes for carrying out the invention]

[0009] The drawing recognition method of this embodiment is a method for extracting symbols and annotations from drawing files such as construction drawings, and is characterized by its ability to extract annotations with high accuracy while maintaining the relationship between symbols and annotations.

[0010] The drawing file from which symbols and the like are extracted by the drawing recognition method of the present embodiment is not particularly limited. Examples of drawing files include drawing files obtained by scanning and digitizing drawings printed on paper (drawing files in which images are composed of pixel data), and drawing files created by converting drawings drawn in CAD (drawing files in which symbols and the like in the data have individual information). For example, as the former type of file, raster format drawing files such as raster PDF files and PNG files can be mentioned, and as the latter type of file, vector format drawing files such as vector PDF files and SVG files can be mentioned. The drawing file from which symbols and the like are extracted by the drawing recognition method of the present embodiment may be either a raster format drawing file or a vector format drawing file described above. Hereinafter, the drawing file from which symbols and the like are extracted will be simply referred to as the original drawing file.

[0011] In the following description, the case where the drawing from which symbols and the like are extracted is a construction drawing will be described as a representative, but the drawing from which symbols and the like are extracted by the drawing recognition method of the present embodiment is not particularly limited. Also, in the following description, since the drawing from which symbols and the like are extracted is a construction drawing, "symbol" means a symbol that graphically represents electrical and mechanical parts and equipment used in building facilities. Also, "side note" means characters and symbols that supplement the specifications of the electrical and mechanical parts and equipment represented by the above-mentioned symbols and are described adjacent to the symbols.

[0012] <The drawing recognition device 1 of the present embodiment> First, the drawing recognition method of the present embodiment and the drawing recognition device 1 will be described. As shown in FIG. 2, the drawing recognition device 1 may have a display function of displaying a drawing file or the like formed by each part described later on the display DS, and an input function of inputting necessary information or instructing necessary operations using an input device IS such as a keyboard and a mouse.

[0013] <Data Input Section 2> As shown in Figure 2, the device 1 for implementing the drawing recognition method of this embodiment (hereinafter simply referred to as the drawing recognition device 1) has a data input unit 2 that creates a source drawing file S1 (see Figure 3(A)). This data input unit 2 has the function of inputting the source drawing file that will be used to create the source drawing file S1. For example, if the source drawing file is a vector format drawing file such as a vector PDF file or a raster format drawing file such as a raster PDF file, it has the function of receiving drawing data from a storage medium such as a USB or receiving it via a line such as the Internet (data reception function). If the source drawing file S1 is a drawing file, the drawing file received by the data input unit 2 can be used as the source drawing file S1 as is. Furthermore, the data input unit 2 has a function to create the original drawing file S1 if the drawing that will be used as the source for creating the original drawing file S1 is a drawing printed on paper, etc., by digitizing the drawing through scanning or other processes and saving it as a raster format drawing file such as a raster PDF file. The original drawing file S1 received or created by the data input unit 2 is transmitted to and stored in the storage unit 10.

[0014] <Drawing file formation unit 3> As shown in Figure 2, the drawing recognition device 1 has a drawing file formation unit 3. The drawing file formation unit 3 has the function of forming a symbol extraction drawing file S2 from the original drawing file S1 received or created by the data input unit 2. The drawing file formation unit 3 also has the function of transmitting the created symbol extraction drawing file S2 to the storage unit 10, and the symbol extraction drawing file S2 transmitted to the storage unit 10 is stored in the storage unit 10.

[0015] The method by which the drawing file formation unit 3 forms the symbol extraction drawing file S2 from the original drawing file S1 is not particularly limited. For example, the drawing file formation unit 3 can form the symbol extraction drawing file S2, which is a PNG file, from the original drawing file S1 by performing background removal processing (such as erasing thin lines exceeding a threshold or filled areas using image processing, or background processing by manually specifying feature areas).

[0016] <Symbol Extraction Unit 4> As shown in Figure 2, the drawing recognition device 1 has a symbol extraction unit 4. The symbol extraction unit 4 has the function of extracting symbols from the symbol extraction drawing file S2 created by the drawing file formation unit 3. The method by which the symbol extraction unit 4 extracts symbols is not particularly limited, but symbols can be extracted by known methods such as pattern matching based on the extracted symbol file S4 (see Figure 5(A)) stored in the storage unit 10. Pattern matching can be performed using AI (artificial intelligence) or the like, but the means of performing pattern matching is not particularly limited. The symbol extraction unit 4 also has the function of forming an extracted symbol file S4 for each extracted symbol, which associates symbol data obtained by cutting out an image of the area containing the extracted symbol with location information of the area (i.e., information regarding the coordinates of the area on the symbol extraction drawing file S2). Each formed extracted symbol file S4 is transmitted to the storage unit 10 and stored.

[0017] Furthermore, when symbol extraction is performed using pattern matching with AI (artificial intelligence), the confidence level of the symbol extraction (displayed numerically: closer to 1 indicates higher confidence) is also included in the extracted symbol file S4. Alternatively, the extracted symbol file S4 may be an image of the area containing the extracted symbols, and the file name of the extracted symbol file S4 may include the location information and confidence level, thereby associating the area image with the location information and confidence level. This method is desirable because it allows the extracted symbol file S4 group to be sorted in descending order of confidence level (or in descending order) on the display DS, enabling efficient deletion of falsely detected extracted symbol file S4. For example, in the case of an image of a rectangular area, the file name of the extracted symbol file S4 can be set to 0.9224@17560×4782_17634×4854_0.png. In the aforementioned file name, "0.9224" represents the confidence level, and "17560×4782_17634×4854" represents the X and Y axis coordinates (position information) of the corners located diagonally opposite each other in the rectangular area of ​​the image.

[0018] <Annotation Search Area Setting Section 5> As shown in Figure 2, the drawing recognition device 1 has a note search area setting unit 5. The note search area setting unit 5 has the function of setting an area (note search area SA) for searching for notes associated with a symbol in the symbol extraction drawing file S2. Specifically, the note search area setting unit 5 has the function of setting a certain area centered on each symbol as the note search area SA in the symbol extraction drawing file S2 for all symbols extracted by the symbol extraction unit 4. Specifically, it has the function of setting a certain area centered on each symbol as the note search area SA for each symbol in the symbol extraction drawing file S2 based on the information contained in each extracted symbol file S4 (see Figure 4). When setting the note search area SA for each symbol, the note search area setting unit 5 sets the range in the symbol extraction drawing file S2 as the note search area SA, which is assumed to include both each symbol and the notes associated with that symbol. For example, based on the size of each extracted symbol, the reference position may be the center of each symbol, and the area calculated by applying a certain multiplier to the top, bottom, left, and right relative to the size of each symbol may be set as the annotation search area SA. Alternatively, if a management list CR (see Figure 5(C)) containing pre-registered annotations corresponding to each symbol (i.e., those used for each symbol) is stored in the storage unit 10 (or may be stored in the annotation search area setting unit 5), the annotation search area SA may be set as a certain range calculated based on the number of characters registered in the management list CR (number of characters in the annotation string), etc., using the center of each symbol as a reference. This annotation search area SA information includes the coordinates of the annotation search area SA (for example, if a rectangular area is set as the annotation search area SA, the coordinate information of the corners). Once this annotation search area SA information is calculated, it is transmitted from the annotation search area setting unit 5 to the storage unit 10 and stored in the storage unit 10 in association with the extracted symbol information of each symbol for which the annotation search area SA has been set. The annotation search area SA information may also include information on the corresponding extracted symbol.

[0019] Figure 5(C) shows a list of multiple annotations corresponding to a single symbol, and the management list CR stores such lists of annotations in association with each symbol.

[0020] <sidenote extraction part 7> As shown in Figure 2, the drawing recognition device 1 has a side note extraction unit 7. The side note extraction unit 7 has the function of extracting side notes from the side note search area SA formed by the side note search area setting unit 5, and the function of linking the extracted side notes with symbols. Specifically, the side note extraction unit 7 performs pattern matching using the side note template list TP for all symbols file S4, for the figures contained within the side note search area SA set by the side note search area setting unit 5, and extracts figures with high similarity to the side note template as side note candidates. If there is only one extracted side note candidate, that side note candidate is linked to the symbol for which the side note search area SA is set, and is transmitted to the storage unit 10 as symbol data for storage.

[0021] Furthermore, the annotation extraction unit 7 has a function to create symbol data by associating the annotation candidate that is considered to have the highest relevance to the symbol with the symbol when there are multiple annotation candidates within the annotation search area SA. The method by which the annotation extraction unit 7 selects the annotation candidate that is considered to have a high relevance is not particularly limited, but for example, it may associate the annotation candidate that is closest in distance from the symbol (for example, the distance from the center of the symbol to the center of the annotation candidate) with the symbol.

[0022] <Extraction and linking of symbols and annotations> The process of extracting symbols and annotations using the aforementioned drawing recognition device 1 (i.e., the drawing recognition method of this embodiment) will now be explained (see Figures 1 and 2).

[0023] <Drawing file formation process> As shown in Figures 1 and 2, the drawing file formation unit 3 receives the original drawing file S1, which has been formed or received by the data input unit 2 and stored in the storage unit 10. The drawing file formation unit 3 forms a symbol extraction drawing file S2 from the original drawing file S1 (see Figure 3(B)). The formed symbol extraction drawing file S2 is transmitted to the storage unit 10 for storage, but the symbol extraction drawing file S2 may also be transmitted directly from the drawing file formation unit 3 to the symbol extraction unit 4 or the annotation extraction unit 7.

[0024] <Symbol Extraction Process> As shown in Figures 1 and 2, when the symbol extraction unit 4 receives the symbol extraction drawing file S2 from the storage unit 10 (or when the symbol extraction drawing file S2 is transmitted to the symbol extraction unit 4 from the drawing file formation unit 3), the symbol extraction unit 4 extracts symbols from the symbol extraction drawing file S2 based on the symbol extraction drawing file S2 using AI-based pattern matching or known pattern matching methods, and forms a symbol file S4. Specifically, when symbols are extracted from the symbol extraction drawing file S2, an extracted symbol file S4 is formed which includes symbol data extracted from the image of the area containing the extracted symbols, location information of the area, confidence level, etc. This extracted symbol file S4 is transmitted to the storage unit 10 and stored.

[0025] The extracted symbol file S4 may also be sent directly to the annotation search area setting unit 5. Furthermore, the extracted symbol files S4 may be sent to the storage unit 10 or the reference search area setting unit 5 once the extraction of all symbols is complete, or each time a symbol is extracted and an extracted symbol file S4 is formed, an extracted symbol file S4 may be sent to the storage unit 10 or the reference search area setting unit 5. Alternatively, once a certain number of extracted symbol files S4 have been formed, a certain number of extracted symbol files S4 may be sent to the storage unit 10 or the reference search area setting unit 5 all at once.

[0026] <Steps for setting up the search area> As shown in Figures 1 and 2, when the Observation Search Area Setting Unit 5 receives the extracted symbol file S4 from the storage unit 10 (or when the extracted symbol file S4 is transmitted from the symbol extraction unit 4 to the Observation Search Area Setting Unit 5), the Observation Search Area Setting Unit 5 sets an Observation Search Area SA in the extracted symbol file S4 for each symbol. In other words, the Observation Search Area Setting Unit 5 sets a certain area centered on the symbol as the Observation Search Area SA in the extracted symbol file S4 based on the information contained in the extracted symbol file S4 (see Figure 4). The information of the set Observation Search Area SA is then transmitted from the Observation Search Area Setting Unit 5 to the storage unit 10 and stored there.

[0027] The information for the observation search area SA may also be transmitted directly from the observation search area setting unit 5 to the observation extraction unit 7. Furthermore, the information of the behest search area SA may be transmitted to the storage unit 10 or the behest extraction unit 7 once the formation of information for the behest search area SA for all symbols has been completed, or the information of the behest search area SA may be transmitted to the storage unit 10 or the behest extraction unit 7 each time information for the behest search area SA is formed for each symbol. Alternatively, once information for the behest search area SA has been formed for a certain number of symbols, the information for a certain number of behest search area SAs may be transmitted to the storage unit 10 or the behest extraction unit 7 all at once.

[0028] <sidenote extraction process> As shown in Figures 1 and 2, when the information of the annotation search area SA is obtained from the storage unit 10 (or when the information of the annotation search area SA is transmitted from the annotation search area setting unit 5), the annotation extraction unit 7 obtains the annotation template list TP stored in the storage unit 10. The annotation extraction unit 7 may have already obtained the annotation template list TP stored in the storage unit 10.

[0029] When the annotation template list TP is obtained, the annotation extraction unit 7 matches the figures included in the annotation search area SA with the annotation template list TP for all symbols and extracts annotation candidates. If there is only one annotation candidate and the similarity (confidence level in AI matching) is above a certain level, or if multiple annotation candidates are extracted but only one has a similarity above a certain level, the unit determines that the annotation candidate is an annotation of the symbol for which the annotation search area SA is set. The unit then links the annotation candidate determined to be an annotation of a symbol with the symbol for which the annotation search area SA is set and transmits it to the storage unit 10 as symbol data for storage.

[0030] Furthermore, if multiple occult candidates with a similarity level above a certain threshold are extracted, the occult extraction unit 7 determines that the occult candidate considered to have the highest relevance to the symbol is the occult candidate of the symbol, associates this occult candidate with the symbol for which the occult search area SA is set, and transmits it to the storage unit 10 for storage as symbol data. For example, the occult candidate with the closest (shortest) distance to the symbol is determined to be the occult candidate considered to have the highest relevance to the symbol, associates this occult candidate with the symbol for which the occult search area SA is set, and transmits it to the storage unit 10 for storage as symbol data.

[0031] By extracting symbols and annotations using the method described above, the accuracy of symbol and annotation extraction can be increased, and related symbols and annotations can be appropriately linked and stored.

[0032] <About the side note template list TP> The font, color, size, etc., of the marginal notes stored in the marginal note template list TP (see Figure 5(B)) used for matching marginal notes are not particularly limited. The characters that make up the marginal note may be registered as a whole, or each character that makes up the marginal note (for example, a single alphanumeric character) may be registered. Multiple marginal notes that are the same in composition but expressed in multiple fonts may be included in the marginal note template list TP.

[0033] In addition, a management list CR may be provided separately from the annotation template list TP, which stores the number of characters used for annotations in symbols (see Figures 2 and 5(C)), and this management list CR may be stored in the storage unit 10. Furthermore, the annotation extraction unit 7 may be provided with a function (verification function) to determine whether the annotations selected by matching using the management list CR are appropriate. By providing such a management list CR and verification function, it is possible to prevent the annotation extraction unit 7 from selecting unsuitable figures or other elements as annotations for symbols.

[0034] For example, when multiple symbols are drawn adjacent to each other, the side notes of adjacent symbols (hereinafter referred to as false side notes) may be included in the side note search area SA. In this case, depending on how the side notes are written in the drawing, the false side notes may have a higher matching reliability than the side notes that are actually associated with the symbol, or the distance between the false side notes and the symbol may be shorter than the distance between the actual side notes and the symbol. This situation is particularly likely to occur when only a single alphanumeric character is used as the side note template. Therefore, if a side note is selected by matching the figure included in the side note search area SA with the side note template, and then verified by a verification function to determine whether that side note is suitable for that symbol, it is possible to prevent the side note extraction unit 7 from selecting a side note that is not suitable for the symbol.

[0035] <Regarding the annotation extraction section 7> The annotation extraction unit 7 may have a function to change the size of the annotation template when matching the figures included in the annotation search area SA with the annotation template list TP. If such a function is provided, it becomes possible to appropriately extract annotations even if the sizes of the figures and the annotation template list TP are different. In other words, even if the scale of the original drawing in the original drawing file S1 is different from the scale of the stored annotation template list TP, the matching accuracy between the figures and the annotation template list TP can be improved, and annotations can be appropriately extracted.

[0036] <When using the extraction template size> If the annotation extraction unit 7 has the function of matching figures included in the annotation search area SA with the annotation template list TP while changing the size of the annotation template, it may perform matching while changing the size of the annotation template for all figures included in the annotation search area SA, or it may perform matching using the extraction template size. In other words, the annotation search area setting unit 5 may be provided with a function to determine the extraction template size (i.e., the annotation search area setting process may be provided with a process to determine the extraction template size), and the annotation extraction unit 7 may perform matching using the extraction template size set by the annotation search area setting unit 5. As shown in Figure 4(A), the annotation search area setting unit 5 sets the annotation search area SA for all symbols, and then selects several representative symbols from the symbols extracted by the symbol extraction unit 4. Then, as shown in Figure 4(B), it determines the extraction template size using the figures detected in the annotation search area SA of the selected symbols (annotation search area SA(1), annotation search area SA(2)). Then, when the extraction template size determined by the annotation search area setting unit 5 is transmitted to the annotation extraction unit 7 (or when the extraction template size is temporarily stored in the storage unit 10 and then transmitted from the storage unit 10 to the annotation extraction unit 7), the annotation extraction unit 7 uses the annotation template adjusted to the determined extraction template size to perform pattern matching between the figures contained in the annotation search area SA and the annotation template for all annotation search areas SA. This makes it possible to appropriately extract annotations from the figures present in the annotation search area SA and makes it easier to prevent the erroneous extraction of figures other than annotations as annotations. As a method for determining the extraction template size, for example, when matching figures detected in the annotation search area SA set by the annotation search area setting unit 5 with the size of the annotation template while changing the size of the annotation template for several representative symbols, the size of the annotation template that yields the highest similarity can be set as the extraction template size.

[0037] The function for setting the extraction template size may be provided in the annotation extraction unit 7 instead of the annotation search area setting unit 5. In other words, the annotation extraction process may include a step for determining the extraction template size. In this case, the annotation extraction unit 7 selects several representative symbols from the symbols extracted by the symbol extraction unit 4. Then, for the selected symbols, the extraction template size can be determined using the shapes detected in the annotation search area SA set in the annotation search area setting unit 5.

[0038] Furthermore, the method for selecting a representative symbol to be used when determining the template size for extraction is not particularly limited. For example, the annotation search area setting unit 5 and the annotation extraction unit 7 may automatically select symbols in order of highest confidence level when symbols are extracted in the symbol extraction unit 4. Alternatively, the operator may select a representative symbol by checking the images of the extracted symbols (i.e., the images contained in the extracted symbol file S4 and the images in the extraction drawing file S2) and their confidence levels. [Industrial applicability]

[0039] The drawing recognition method of the present invention is suitable for extracting symbols and characters from drawings that contain a large number of symbols, characters, and lines, such as construction drawings. [Explanation of Symbols]

[0040] 1. Drawing Recognition Device 2. Data Input Section 3. Drawing file formation unit 4. Symbol Extraction Unit 5. Side note search area setting section 7 Paragraph extraction part 10 Storage section S1 Original drawing file S2 Symbol Extraction Drawing File S4 Extracted Symbol File SA (Side Area) TP (Postscript) Template List CR Management List

Claims

1. A drawing recognition method for extracting symbols and annotations contained in a drawing file, A drawing file formation step of forming a drawing file for symbol extraction to be used for symbol extraction from the aforementioned drawing file, A symbol extraction step for extracting symbols from the symbol extraction drawing file formed in the drawing file formation step, A step of setting an adjacent search area for each of the multiple symbols extracted in the symbol extraction step, setting an adjacent search area centered on each symbol in the symbol extraction drawing file, The process involves extracting annotations corresponding to each symbol from the figures detected in the annotation search area set in the annotation search area setting step. A drawing recognition method characterized by the following:

2. In the aforementioned annotation extraction process, The process involves changing the size of the side note templates registered in the side note template list, matching the side note templates with the shapes detected in the side note search area, and extracting the side notes contained in the side note search area. The drawing recognition method according to claim 1, characterized by the feature.

3. In the aforementioned note search area setting step and / or note extraction step, In the symbol extraction process, a matching process is performed between the side-view template and the figures detected in the side-view search area, which is set for the multiple representative symbols extracted in the symbol extraction process, while changing the size of the side-view template, to determine the extraction template size, which is the size of the side-view template that yields the highest similarity. In the aforementioned annotation extraction process, Using the determined annotation template of the extraction template size, the annotations contained in the annotation search area are extracted. The drawing recognition method according to claim 2, characterized in that it is a drawing recognition method.

4. In the aforementioned annotation extraction process, Multiple figures included in the aforementioned search area are extracted as candidates for annotations, Among the multiple potential annotation candidates, the one closest in distance to the symbol is associated with the symbol. The drawing recognition method according to claim 1, characterized by the feature.

5. In the aforementioned annotation extraction process, Multiple figures included in the aforementioned search area are extracted as candidates for annotations, For the multiple extracted candidate annotations, the system determines which candidate annotations to associate with the symbol that defined the annotation search area from which the candidate annotations were extracted, based on a management list that stores the relationship between symbols and annotations. The drawing recognition method according to claim 1, characterized by the feature.

6. A drawing recognition device for extracting symbols and annotations contained in a drawing file, The drawing recognition device is A drawing file forming unit that forms a drawing file for symbol extraction to be used for symbol extraction from the aforementioned drawing file, A symbol extraction unit extracts symbols from the symbol extraction drawing file formed by the drawing file forming unit, A side-view search area setting unit sets a side-view search area centered on each of the multiple symbols extracted by the symbol extraction unit in the symbol extraction drawing file, The system includes a marking extraction unit that extracts markings corresponding to each symbol from figures detected in the marking search area set by the marking search area setting unit. A drawing recognition device characterized by the following features.

7. The aforementioned annotation extraction unit is, The process involves changing the size of the side note templates registered in the side note template list, matching the side note templates with the shapes detected in the side note search area, and extracting the side notes contained in the side note search area. The drawing recognition device according to claim 6.

8. The annotation extraction unit and / or the annotation search area setting unit are: The symbol extraction unit has a function to determine the extraction template size, which is the size of the side note template that yields the highest similarity, by performing a matching process between the side note template and the figures detected in the side note search area, while changing the size of the side note template in the side note search area set for the multiple representative symbols extracted in the symbol extraction unit. The aforementioned annotation extraction unit is, It has the function of extracting annotations contained in the annotation search area using the annotation template of the determined extraction template size. The drawing recognition method according to claim 7, characterized in that it is a drawing recognition method.

9. The aforementioned annotation extraction unit is, A function to extract multiple figures included in the aforementioned search area as candidates for annotations, It has a function to associate the symbol with the symbol that is closest to the multiple candidate for accompanying text. The drawing recognition device according to claim 6.

10. The aforementioned annotation extraction unit is, It has a function to extract multiple figures included in the aforementioned annotation search area as annotation candidates, The system has a function to determine which of the extracted candidate annotations should be associated with the symbol that defined the annotation search area, based on a management list that stores the relationship between the symbol and the annotation. The drawing recognition device according to claim 6.

Images