Drawing reader

Abstract

To provide a drawing reader which reduces the number of character recognition processes and is capable of converting necessary information into text even when a drawing is in a similar format and is difficult to define.SOLUTION: A drawing reader 1 disclosed herein is configured to acquire a target image obtained by capturing an image of a reading target drawing or form, determine a reading range of the target image through pattern matching processing with a pre-stored image, and cut out and output the reading range to an external character recognition device 2.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present disclosure relates to a drawing reading device. [Background technology]

[0002] Patent Document 1 discloses a technology for performing OCR (Optical Character Recognition) on non-standard forms via a cloud service. In this technology, when it is not possible to use a reference form associated with form information used to read a target form to read a first form image of the target form, which is the form to be processed, form information to be used to read the target form is generated based on the positions of tables, lines, and characters included in the first form image, and this generated form information is stored in association with the first form image. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent Publication No. 2021-111044 Summary of the Invention [Problem to be solved by the invention]

[0004] However, in the above-mentioned Patent Document 1, if the character recognition function is provided as a cloud service, a fee is charged for each character recognition, which is costly. In addition, if tables, lines, characters, etc. have irregular shapes, it is not possible to define the reading position, and it is not possible to define templates such as reference forms.

[0005] The present disclosure has been made in consideration of the above, and aims to provide a drawing reading device that can reduce the number of times character recognition is performed and convert the necessary information into text even for drawings that are difficult to define in a similar format. [Means for solving the problem]

[0006] In order to solve the above-mentioned problems and achieve the objectives, the drawing reading device of the present disclosure is a drawing reading device equipped with a processor, which acquires a target image of the drawing or document to be read, determines a reading range for the target image by pattern matching processing with a pre-stored image, and extracts and outputs the reading range to an external character recognition device. [Effects of the Invention]

[0007] According to the present disclosure, it is possible to reduce the number of times character recognition is performed and to convert necessary information into text even for drawings that are difficult to define in a similar format. [Brief explanation of the drawings]

[0008] [Figure 1] FIG. 1 is a block diagram showing the functional configuration of a drawing reading apparatus according to an embodiment. [Figure 2] FIG. 2 is a flowchart showing an outline of a process executed by the drawing reading apparatus according to an embodiment. [Figure 3] FIG. 3 is a diagram illustrating an outline of a process executed by a drawing reading apparatus according to an embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0009] Hereinafter, a drawing reading device according to an embodiment of the present disclosure will be described with reference to the drawings. Note that the present disclosure is not limited to the following embodiment. In the following description, the same parts are denoted by the same reference numerals.

[0010] [Functional configuration of the drawing reading device] Fig. 1 is a block diagram showing the functional configuration of a drawing reading device according to one embodiment. Drawing reading device 1 shown in Fig. 1 outputs captured images of figures and forms to be read via network N100 to character recognition device 2 of an external cloud service that can recognize characters, figures, ruled lines, etc. using OCR. Drawing reading device 1 includes input unit 10, output unit 20, storage unit 30, and processing unit 40.

[0011] The input unit 10 accepts input of various operations or various information, and outputs signals or various information corresponding to the accepted operations to the processing unit 40. The input unit 10 is configured using input devices such as a mouse, keyboard, and touch panel, a reading reader capable of acquiring various information from a storage medium such as a memory card, a scanner and a camera capable of capturing images of figures and forms to be read, and the like.

[0012] The output unit 20 outputs various information and images output from the processing unit 40. The output unit 20 is configured using a liquid crystal display or an organic electroluminescent display (EL display), a speaker, a printer, and the like.

[0013] The storage unit 30 stores various information and programs related to the drawing reading device 1. The storage unit 30 is configured using an SSD (Solid State Drive), an HDD (Hard Disk Drive), a volatile memory, a non-volatile memory, etc. The storage unit 30 has a determination rule storage unit 31, a cut-out rule storage unit 32, a parts list storage unit 33, and a reading result storage unit 34. The determination rule storage unit 31 stores determination rules for determining the reading range by pattern matching processing for a target image obtained by capturing a drawing or form to be read. The cut-out rule storage unit 32 stores cut-out rules for the reading range. The parts list storage unit 33 stores a parts list that associates drawing numbers, product numbers, and product names. The reading result storage unit 34 stores the results of character recognition and reading by the character recognition device 2.

[0014] The processing unit 40 is configured using a memory and a processor having hardware such as a CPU (Central Processing Unit). The processing unit 40 reads a program stored in the storage unit 30 into a working area of ​​the memory, executes the program, and controls each component through the execution of the program by the processor. In this way, the processing unit 40 realizes a functional module that meets a predetermined purpose through cooperation between hardware and software. Specifically, the processing unit 40 includes an acquisition unit 41, a determination unit 42, an extraction unit 43, an association unit 44, and a storage control unit 45. In one embodiment, the processing unit 40 functions as a processor.

[0015] The acquisition unit 41 acquires a target image obtained by capturing an image of a drawing or a document to be read via the input unit 10 or the network N100.

[0016] The determination unit 42 determines the reading range of the target image P1 according to the determination rule stored in the determination rule storage unit 31 by pattern matching the target image P1 with a pre-stored image.

[0017] The cutout unit 43 cuts out the read range determined by the determination unit 42 in accordance with the cutout rules stored in the cutout rule storage unit 32 .

[0018] The associating unit 44 associates the product number and product name extracted by the extracting unit 43 and outputs them to the character recognition device 2.

[0019] The memory control unit 45 compares the parts list stored in the parts list memory unit 33 with the processing results of the OCR processing input from the character recognition device 2, matches the product names, stores the matched processing results in the reading result memory unit 34, and outputs a list of inconsistencies to the output unit 20.

[0020] [Processing by drawing reader] Next, a description will be given of the processing executed by the drawing reading apparatus 1. Fig. 2 is a flowchart showing an outline of the processing executed by the drawing reading apparatus 1. Fig. 3 is a diagram showing a schematic outline of the processing executed by the drawing reading apparatus 1.

[0021] 2, first, the acquisition unit 41 executes an image input process to acquire a target image obtained by capturing an image of a drawing or a document to be read via the input unit 10 or the network N100 (step S1). Specifically, as shown in (a) of FIG. 3, the acquisition unit 41 acquires a target image P1.

[0022] Next, the determination unit 42 executes a read range determination process for the target image P1 in accordance with the determination rules stored in the determination rule storage unit 31, in which the read range is determined by pattern matching with pre-stored images for the target image P1 (step S2). Here, the determination rules include a product number frame determination rule and a product name frame determination rule.

[0023] The product number frame determination rule uses multiple images for pattern matching processing with different granularity, each with a rectangular size of 1300px ±15% width and 180px ±15% height. The product number frame determination rule calculates the similarity at a size 0.8 to 1.2 times larger than the target image P1, and determines that a match occurs when the similarity is 0.8 or greater. Here, multiple images for pattern matching processing with different granularity are, for example, multiple images with different shapes, granularity, and fonts for the "product number" text.

[0024] The product name frame determination rule also uses multiple images for pattern matching processing with different granularities, each with a rectangular size of 1300px ±15% and 240px ±15%. The product name frame determination rule calculates similarity using a size 0.8 to 1.2 times the size of the target image P1, and determines a match if the similarity is 0.8 or greater. Here, multiple images for pattern matching processing with different granularities are, for example, multiple images with different shapes, granularities, and fonts for the "product name" text. In this way, the determination unit 42 performs pattern matching processing on the target image P1 for each predetermined rectangular size, calculates similarity, and determines a match if the similarity is 0.8 or greater, in accordance with the product number frame determination rule and product name frame determination rule stored in the determination rule storage unit 31. In this case, as shown in (b) of Figure 3, the judgment unit 42 judges the rectangular size that matches the product number for the target image P1 as a product number frame W1, the rectangular size that matches the product name as a product name frame W2, and the rectangular size that matches other than the product number and product name as an other frame W3.

[0025] Thereafter, the cutout unit 43 performs a read range cutout process to cut out the read range determined by the determination unit 42 in accordance with the cutout rules stored in the cutout rule storage unit 32 (step S3). There are two cutout rules. The first rule is to cut out the target image P1 in the rectangular size determined by the determination unit 42. The second rule is to delete the area above the matched coordinates in the image for pattern matching processing. For example, as shown in FIG. 3(c), the cutout unit 43 cuts out each of the product number box W1, product name box W2, and other box W3 of the target image P1 determined by the determination unit 42 in accordance with the first rule to generate cutout images K1 and K2 (FIG. 3(b) → FIG. 3(c)). Then, as shown in (d) of Figure 3, the cut-out unit 43, in accordance with the second rule, deletes characters, figures, etc. in areas W11 and W21 above the matched coordinates in the image for pattern matching processing for each of the cut-out images K1 and K2 ((c) of Figure 3 → (d) of Figure 3).

[0026] Next, the associating unit 44 executes a product number / product name associating process in which the product number and product name cut out by the cutout unit 43 are associated with each other and output to the character recognition device 2 (step S4). Specifically, the associating unit 44 associates the cutout image of the product number cut out by the cutout unit 43 with the cutout image of the product name and outputs them to the character recognition device 2. For example, as shown in FIG. 3(e), the associating unit 44 associates the cutout image K1 of the product number and the cutout image K2 of the product name cut out by the cutout unit 43 with each other and outputs them to the character recognition device 2 (FIG. 3(d) → FIG. 3(e)).

[0027] Thereafter, the storage control unit 45 executes a matching storage process in which it compares the parts list stored in the parts list storage unit 33 with the processing result of the OCR processing input from the character recognition device 2, and stores the matching result of the item names in the reading result storage unit 34 (step S5). In this case, the storage control unit 45 may output to the output unit 20 a list of inconsistencies between the parts list stored in the parts list storage unit 33 and the processing result of the OCR processing input from the character recognition device 2. After step S5, the drawing reading device 1 ends this process.

[0028] According to the embodiment described above, the acquisition unit 41 acquires a target image P1 obtained by capturing an image of a drawing or form to be read, and the determination unit 42 determines the read range of the target image P1 by pattern matching it with a pre-stored image. Then, the processing unit 40 extracts the read range and outputs it to the character recognition device 2. This reduces the number of times character recognition is performed, and makes it possible to convert necessary information into text even for drawings that are difficult to define in a similar format.

[0029] In one embodiment, the determination unit 42 determines the read range of the target image P1 by pattern matching with a pre-stored image in accordance with the determination rule stored in the determination rule storage unit 31. However, this is not limiting. For example, the determination unit 42 may appropriately change the rectangular size and shape of the read range for the character string surrounding the surface shape symbol. Specifically, the determination rule may be set to a rectangular size and shape smaller than the rectangular sizes of the product number and product name. In this case, the determination unit 42 further determines the read range determined by the pattern matching with a pre-stored image by pattern matching with a pre-stored image. In this case, the cutout unit 43 cuts out a rectangular size detected within the image area of ​​the pattern matching process to generate a cutout image. Then, the association unit 44 associates the cutout image cut out by the cutout unit 43 with the drawing number and outputs the result to the character recognition device 2.

[0030] In one embodiment, the present invention can be applied to the character strings of designated certification marks, character strings of warning labels, product name consistency checks on labels, various characteristic marks, and instruction contents. For example, by setting similar judgment rules, pre-stored images, and cut-out rules, the drawing reading device 1 can appropriately cut out the reading range, thereby accurately converting the character strings around the symbols into text.

[0031] In the explanation of the flowcharts in this specification, the order of processing between steps is clearly indicated using expressions such as "first," "then," and "continue," but the order of processing required to implement this embodiment is not uniquely determined by these expressions. In other words, the order of processing in the flowcharts described in this specification can be changed within a consistent range.

[0032] Further advantages and modifications will readily occur to those skilled in the art. The invention in its broader aspects is not limited to the specific details and representative embodiments shown and described above. Accordingly, various modifications may be made without departing from the spirit or scope of the general inventive concept as defined by the appended claims and their equivalents.

[0033] Although some of the embodiments of the present application have been described in detail above with reference to the drawings, these are merely examples, and the present invention can be implemented in other forms that have undergone various modifications and improvements based on the knowledge of those skilled in the art, including the aspects described in the disclosure of the present invention. [Explanation of symbols]

[0034] 1. Drawing reader 2 Character recognition device 31 Judgment rule memory unit 32 Extraction rule memory 33 Parts list storage section 34 Reading result storage unit 40 Processing section 41 Acquisition Department 42 Judgment section 43 Cutout 44 Mapping section 45 Memory control unit

Claims

[Claim 1] A drawing reader including a processor, The processor: Acquire an image of the drawing or document to be read, determining a plurality of read ranges for the target image by pattern matching a plurality of images having different shapes, granularity, and fonts of pre-stored product number text and a plurality of images having different shapes, granularity, and fonts of pre-stored product name text; extracting each of the plurality of read ranges to generate a plurality of extracted images of the product number or product name; For each of the plurality of cut-out images of the product number or product name, characters and graphics in the area above the coordinates that match in the image for pattern matching processing are deleted; an extracted image of the product number obtained by deleting characters and graphics in the area above the coordinates that match in the image for pattern matching processing, and an extracted image of the product name obtained by deleting characters and graphics in the area above the coordinates that match in the image for pattern matching processing, and outputting the extracted images to an external character recognition device in association with each other; comparing a pre-stored parts list with the processing result input from the character recognition device, and storing the matching result of matching the product names; Drawing reading device.

Images