JPS63278177A - Drawing recognizing system - Google Patents

Abstract

PURPOSE:To convert a large quality of the drawing information of a drawing sheet to code data and to efficiently automatically register by optically inputting the drawing information and automatically recognizing graphic information and character information in the drawing information. CONSTITUTION:The drawing sheet is inputted by a drawing reader 2 and temporarily held on a working memory 4 in the form of image data for one sheet unit. A drawing recognition device 5 refers to a drawing recognizing dictionary memory 6 to recognize a vector, a symbol and extract an area corresponding to a character. At such a time, an optimum recognition program is selected to the image data and a graphic element is recognized according to the algorithm. The program makes access to the memory 6 to recognize the image data and temporarily store on the memory 4. At such a time, the reader 2 and the device 5 can successively input the drawing sheet since a parallel processing can be executed. Further, a graphic segment device 7 segments the image data corresponding to the character area and a character recognition device 8 refers to a character recognizing dictionary memory 9 to recognize the character and convent to a prescribed character code.

Description

[Detailed description of the invention] [Field of application in industry] This invention is a drawing that automates the processing of various drawings such as power distribution line diagrams, topographic maps used for gas and water piping network diagrams, and equipment management maps. It concerns recognition systems.

[Conventional technology]

FIG. 4 is a block diagram showing an example of a figure recognition system using a conventional CAD system disclosed in, for example, Japanese Unexamined Patent Publication No. 59-125468. The information input device is used to input topographical information and equipment location information. Reference numeral 12 denotes standard pattern information, which is a pre-registered standard pattern related to figures used during manual digitizing. 13 is the drawing information file C1 drawing management information 13a4 drawing information 1
It consists of 3b. 14, the attribute information is input using an attribute information input device such as a key fob. 15
is an attribute information file that stores attribute information input from the attribute information input device 14. Reference numeral 16 denotes a graphic correspondence device that associates each drawing information I (I:1 to N) of the drawing information 13b registered in the drawing information file 13 with attribute information registered in the attribute information file 15. A drawing correction device 17 reads the drawing information 13b while referring to the attribute information registered in the attribute information file 15, and corrects the graphic information.

Next, the operation will be explained.

First, when an operator instructs the operator to select the drawing position big and standard pattern information 12 on the drawing information input device 11, for example, a digitizer, the drawing information file 13 is accessed, and the drawing information file 13 is accessed from the code information of the vector set. The configured drawing information 1 to N are stored as drawing information 13b.

One drawing is separated into drawing management information 13a and drawing information 13b of the data itself and is registered in the drawing information file 13. In addition to the above vector code, the drawing information 13b includes symbol codes in accordance with the contents of operator instructions. , is registered in the drawing information file 13 to character code 1.

On the other hand, the attribute information according to the drawing content is inputted by the operator by operating the attribute information input device 14, or by inputting information created all at once.
Register for 5.

The graphic correspondence device 16 synthesizes the graphic information 13b and the attributes while checking the attribute information file 15 and returns it to drawing information. The drawing correction device 17 is composed of, for example, a display device and a tufflet device, and according to an operator's instructions, drawing information 1 registered in the drawing information file 13 is displayed.
3b and the attribute information registered in the attribute information file 15 are read out at any time, the contents thereof are changed or corrected, and the attributes are re-registered or newly registered.

[The problem that the invention attempts to solve]

Conventional figure recognition systems are configured as shown below, so all of the hundreds of thousands of existing drawings must be input directly by an operator using the direct tracing method. It is necessary to register and check all the coordinates and character codes of vectors, symbols, etc. for each figure one by one, and there is a problem that drawing information cannot be registered unless a huge number of inputs are processed.

This invention was made in order to solve the above-mentioned problems, and it is possible to input drawing information by optically inputting drawing information and automatically recognizing one character information of figure information in the drawing information. The purpose of this invention is to obtain a figure recognition system that can automatically register information along with its location information.

[Means for solving problems]

The figure recognition system according to the present invention includes a drawing reading means for optically reading an existing drawing sheet, and read image data from line segment vector information, predetermined symbol information, and nine character information area while referring to figure recognition dictionary information. a graphic information recognition means for decomposing and recognizing graphic element information into graphic element information; a graphic information storage means for sequentially storing the recognized graphic element information along with position information; and image data in the recognized character information area. and character information recognition means for analyzing the text while referring to a predetermined character recognition dictionary and converting it into a predetermined character code.

[Effect]

In the figure recognition system according to the present invention, when an existing drawing sheet is read by the drawing reading means, the figure information recognition means reads the read image data by referring to the figure recognition dictionary information and adding line segment vector information, predetermined symbol information 9, etc. It is decomposed into graphic element information consisting of a character information area, recognized, and stored in a graphic information storage means. The character information recognition means analyzes the image data in the recognized character information area while referring to a predetermined character recognition dictionary and converts it into a predetermined character code.

〔Example〕 An embodiment of the present invention will be described below with reference to the drawings.

In FIG. 1, reference numeral 1 denotes a drawing reading recognition device, which is roughly divided into a drawing reading device 2 and a drawing recognition device 3.
The drawing reading device 2 optically reads a drawing sheet 10, which will be described later.

The drawing recognition device 3 includes a working memory 4. which temporarily stores image data read by the drawing reading device 2. A figure recognition device 5 which decomposes and recognizes the image data stored in the working memory 4 into figure element information consisting of line segment vector information, predetermined symbol information and 9 character information areas while referring to the figure recognition dictionary memory 6. A figure cutting device 7 that cuts out image data corresponding to the character information area recognized by the figure recognition device 5. The figure cutting device 7 is composed of a character recognition device 8 that analyzes image data corresponding to the cut out character information area while referring to a character recognition dictionary memory 9 and converts it into a predetermined character code, and a drawing recognition device 3. Drawing information consisting of recognized graphical information and character information is transferred to the drawing information file 13 and automatically registered.

Next, the operation shown in FIG. 1 will be explained with reference to FIGS. 2 and 3.

FIG. 2 is a schematic diagram illustrating the graphic information recognition processing operation, and in the figure, 4a represents image data output from the drawing reading device 2 and is stored in the working memory 4. 5a is a recognition program processed by the figure recognition device 5, and an optimum recognition program is selected based on the image data 4a. 5b is table data, which is composed of a line segment vector information section 5c, a symbol information section 5d, a character information area section 50, etc., and the line segment solid information sections 1 to 5C are based on entry numbers, addresses, line types, line widths, etc. configured.

The symbol information section 5d is an entry number.

It consists of address, type, angle, etc. The character information area section 5e is composed of an entry number, a circumscribed rectangular address, and the like. 10 is a drawing sheet, line diagrams, symbols,
It is made up of characters.

Note that the arrows in the figure indicate the processing route.

FIG. 3 is a schematic diagram illustrating the character information recognition processing operation. In the figure, 7a and 7b are cutout image data, and the figure cutter 7 refers to the entry number and circumscribed rectangle address of the character information area section 5e. This corresponds to image data cut out from the image data in the working memory 4. 8a shows a state where the character code full is registered together with the character code corresponding to the cutout image data 7a and 7b or the circumscribed rectangle address. In addition,
Arrows in the figure indicate processing routes.

The drawing sheet 10 is input by the drawing reading device 2,
Working memory in the form of image data for each image 4
”Secondly, it is temporarily held. The figure recognition device 5 performs vector recognition, symbol recognition, and extraction of an area corresponding to one character while referring to the drawing recognition dictionary memory 6.

This poem, as shown in Figure 2, image data 4a
The recognition program 5a that is most suitable for the description content and description method is selected for each, and the graphic element recognition process is executed according to the selected algorithm.

Then, the selected recognition program 5a or figure recognition dictionary memory 6 is accessed to recognize the image data 4a in accordance with the format of the chiffle data 5b and temporarily store it in the working memory 4.

In particular, the line segment vector information section 5C contains entry numbers, A1 to
Address, line type, line width, etc. are registered for each continuous line segment. The symbol information section 5d is an entry number.

Address, type, angle, etc. are registered. In the character information area section 5C, entry numbers, circumscribed rectangular addresses, etc. are registered.

Further, at this time, since the drawing reading device 2 and the figure recognition device 5 can perform parallel processing, it is possible to input a plurality of drawing sheets 10 in succession. Furthermore, for the character information area section 5e, the above-mentioned circumscribed rectangle area 1 is extracted from the image data 4a stored in the working memory 4''2.
~ While referring to the responses, only the image data corresponding to the character area is cut out and extracted by the figure cutting device 7, and normalized as cut out image data 7a and 7b. Next, as shown in FIG. 3, using the cutout image data 7a, 7b as a unit amount, the characters in the cutout image data 7a, 7b are recognized by referring to the character recognition device 8 or the character recognition dictionary memory 9, and a predetermined character is Convert to code. Note that, depending on the cutout size of the character, it is possible to refer to a character recognition dictionary of the optimum size in the character recognition dictionary memory 9.

These results are edited and registered in the drawing information file 13, or for the drawing information, a reference point is set in the drawing management information 13a, and the drawing information (table data 5b) is stored in the drawing information 13b in the order of entries. .

On the other hand, the attribute information for the drawing is input to the attribute information input device 14.
, and is stored in the attribute information file 15. In addition, the drawing information 13b stored in the drawing information file 13
The attribute information or figure centering stored in the attribute information file 15 is read out by the device 16 and can be checked and synthesized at any time.

In addition, the drawing information 13 registered in the drawing information file 13
If it is necessary to modify the drawing correction device 17, an instruction is given from the tufflet, keyboard, etc. (not shown) of the drawing correction device 17, and the instruction is displayed on the display, so that the contents of the figure and attributes can be changed and corrected.

In the above embodiment, the figure recognition device 5 has explained the case where the figure element information is recognized by breaking it down into three types of vectors, symbols, and characters, or the recognition level is raised to a higher format and the figure element information is recognized by, for example, vectors. If configured to recognize closed areas, houses in the drawing sheet 10 can be recognized. Furthermore, by configuring the system to recognize closed regions of solid one-dimensional shapes that include characters, it becomes possible to run a graphic processing program that corresponds to the meaning of the input closed shape, such as automatically forming data on a representative house.

In addition, if it is attribute information or existing information, the drawing information 13b registered in the drawing information file 13 can be used by connecting online to a computer that can process this attribute information or using magnetic tape. I'll come.

〔Effect of the invention〕

According to the present invention, an existing drawing sheet is inputted by a drawing reading means, and the inputted image data is read into line segment vector information, while referring to figure recognition dictionary information.
The predetermined symbol information is decomposed into graphic element information consisting of five character information areas, recognized, and stored in the graphic information storage means, and the character information recognition means stores the image data in the recognized character information area in a predetermined character recognition dictionary. The system is configured to analyze and convert the drawing information into predetermined character codes while referring to 1.It is possible to convert the drawing information of a large number of drawing sheets into reproducible code data and register it efficiently and automatically. It has the advantage of being able to immediately respond to requests for corrections, changes, and output of drawn drawing information.

[Brief explanation of drawings]

FIG. 1 is a block diagram explaining the configuration of a graphic recognition system showing an embodiment of the present invention, FIG. 2 is a schematic diagram explaining graphic information recognition processing operation, and FIG. 3 is a schematic diagram explaining character information recognition processing operation. The schematic diagram in FIG. 4 is a block diagram showing an example of a figure recognition system using a conventional CAD system. 1/, 1 is a drawing reading recognition device, 2 is a drawing reading bag layer (drawing reading means), 3 is a drawing recognition device, 4 is a working memory, 5 is a figure recognition device (figure recognition means),
6 is a figure recognition dictionary memory, 7 is a figure cutting device, 8 is a character recognition device (character information recognition means), 9 is a character recognition dictionary memory, 13 is a drawing information file (graphic information storage means),
13a is drawing management information, 13b is drawing information, 14 is an attribute information input device, 15 is an attribute information file, 16 is a figure centering device, and 17 is a drawing correction device. In addition, in the figures, the same reference numerals indicate the same or corresponding parts. Procedural amendment (voluntary)

Claims (2)

[Claims] (1) A drawing reading means that optically reads an existing drawing sheet, and the image data read by this drawing reading means is composed of line segment vector information, predetermined symbol information, and character information area while referring to figure recognition dictionary information. a graphical information recognition means for decomposing and recognizing graphical element information;
A graphic information storage means for sequentially storing the graphic element information on the drawing sheet recognized by the graphic information recognition means together with position information; A graphic recognition system comprising character information recognition means for analyzing character information while referring to a character recognition dictionary and converting it into a predetermined character code. (2) The graphic recognition system according to claim (1), wherein the character information recognition means stores the recognized character code in the graphic information storage means in dependence on the position information of the character information area. .