JPH0785262B2 - Drawing input device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial field of application] In the present invention, when detecting and separating an object in an image, even if the type of object or the constraint condition is changed, only the rule is changed.
The present invention relates to a drawing input device capable of detecting and separating an object.

[Conventional technology]

FIG. 10 is a block diagram of a conventional drawing input device in, for example, the 31st IPSJ National Convention Proceedings Collection 6N-1 “Input of map information and facility information using drawing reading” in the latter half of 1985.

In the figure, (1) is an image input means for inputting an image to obtain binary digital image data, (2) is a first storage means for storing the input binary image data, and (3) is an image Image processing means for image-processing data to extract line information and area information, (4) second storage means for storing line information and area information, and (14) symbol or line in image Element detection / separation means for detecting / separating elements, (10) fourth storage means for storing the detected / separated elements, (12) object recognition means for recognizing target elements, and (13) for recognized objects. It is a fifth storage means for storing code information.

In the conventional drawing input device configured as described above, for example, a drawing as shown in FIG. 2 is input by the image input means (1), and binary digital image data having a thickness is stored in the first storage. It is stored in the means (2). Image processing means (3)
Expands and contracts image data and performs logical operations between images to extract small white closed regions and black filled regions, and thin image data with a thickness to draw lines. The xy coordinates (point sequence) of the center of the and the connection information of the line are extracted. Area information and line information are stored in the second storage means (4). The element detecting / separating means (14) takes out the object from the area information and the line information, and stores it in the fourth storage means (10) for each object taken out. When the target is a symbol, the target recognition means (12) recognizes the type of the symbol, and when the target is a line, it is line-coded and stores the code information in the fifth storage means (13).

[Problems to be solved by the invention]

In the conventional drawing input device as described above, the type of the symbol to be detected / separated is changed, the target of the line figure to be separated is changed, and the constraint condition for writing the line figure is different for each drawing. Therefore, there is a problem in that the element detection / separation means must be changed and created for each object each time the type of symbol is changed or the constraint condition of the line figure is changed.

The present invention has been made in order to solve such a problem, and the type of the symbol to be detected and separated in the drawing is changed, or the constraint condition of writing the line figure is changed. Among the required feature amounts and determination methods, it has a rule for calculating each feature amount and a rule for determining, and the purpose is to be able to change / add / delete rules according to changes in the target.

[Means for Solving Problems] A drawing input device according to the present invention includes a region for detecting and recognizing the specific recognition target, a rule for calculating a characteristic amount of a line, the characteristic amount and the characteristic amount. A table in which the other feature amount related to the amount is associated, and a feature extraction rule storage unit that stores a table in which the rule for calculating the feature amount based on this table and the feature amount are stored. Based on the tables stored in the feature extraction rule storage means, a rule for calculating the feature amount of the element extracted by the image processing means is selected from the feature extraction rule storage means, and is selected. Feature extraction control means for calculating the feature amount of the element based on the rule, a rule for determining whether the feature amount of the area or line is the feature amount of the recognition target, and the recognition target To detect the specific recognition target based on the judgment rule storage unit that stores a correspondence table with the judgment rule necessary for detecting and recognizing the recognition target, and the correspondence table stored in the judgment rule storage unit. The determination rule for detecting the specific recognition target by selecting the determination rule from the determination rule storage unit and determining the feature amount of each element calculated by the feature extraction control unit based on the selected determination rule. And a control means.

[Operation] According to the present invention, the feature extraction control means selects the rules of the feature amounts of the areas and lines extracted as elements from the image information based on the tables stored in the feature extraction rule storage means. It is determined whether the feature amount calculated by the determination rule control unit is the feature amount of the recognition target by the rule selected from the determination rule storage unit based on the correspondence table between the recognition target and the determination rule. As a result, the recognition target is detected. Therefore, by simply changing the rules stored in the feature extraction rule storage means or the above two tables, the symbols and lines with different rules used for calculating the feature amount are described. Different types of drawings can be easily input.

[Embodiment] FIG. 1 is an overall configuration diagram showing an embodiment of the present invention, in which (1) to (4), (10) and (12) to (13) are the same as the above conventional means. belongs to.

(5) is a feature extraction rule storage means for storing rules for calculating feature amounts such as regions and lines, and (6) is a feature extraction rule storage means (5 for storing rules for calculating necessary feature amounts according to objects. ), A feature extraction control means for selecting and importing the feature quantity, (7) a third storage means for storing information on the calculated feature quantity, and (8) a feature quantity for which the feature quantity is a recognition target. Determination rule storage means for storing a rule for determining whether or not, based on the rule stored in the determination rule storage means (8) based on the feature amount calculated by the feature extraction control means (6) Judgment rule control means for detecting a recognition target by judgment, (10) fourth storage means for storing information of the judged element, (11) feature extraction rule (5) and judgment rule (8) Rules for operators to change It is a correction means.

Next, the operation of the above embodiment will be described with reference to FIGS. For example, a drawing in which equipment information (equipment symbols and equipment lines) are superimposed on a map as shown in FIG. 2 is input as a digital image having a thickness by the image inputting means (1), and 2 The image data of the value is stored in the first storage means (2). The image processing means (3) inputs the image data, thins a line from the image data as shown in FIG. 3 (a), and x at the center of the line as shown in FIG. 3 (b). -Take out the y coordinate sequence as point sequence data,
This point sequence data is stored in the second storage means (4) together with the connection information of the lines at the branch points and the intersections. Further, the image processing means (3) uses image contraction / expansion of the image data or inter-image calculation to calculate a black filled area (hereinafter referred to as a black area) as shown in FIGS. A small white closed area (hereinafter referred to as a white area) is taken out and stored in the second storage means (4). In the feature extraction rule storage means (5), the feature quantity of the region, for example, the area S of the region, the perimeter l, the vertical m of the circumscribing rectangle, the horizontal n, the center of gravity M, and the vicinity are present as shown in FIG. Rules for calculating the distance d between the regions, the feature amount of the line, for example, the length L of the line as shown in FIG.
There are rules for determining the inclination θ of a line, closed-loop property α, the size of a group of connected lines, and the like. The feature extraction control means (6) is a feature quantity necessary for detecting and separating the symbol or line to be recognized stored in the feature quantity extraction rule storage means (5) and other features related to this feature quantity. Using a table in which the amounts are associated with each other and a table in which the rules for calculating the feature amounts based on this table and the feature amounts are associated with each other, the rules according to the need are stored in the feature amount extraction rule storage means ( Select from 5), take in, and calculate the feature amount. For example, for the symbol shown in FIG. 7, the area of the black area, the circumscribed rectangle, the area of the white area within a certain distance between the black area and the circumscribed rectangle, the size of the area including the black area and the white area, etc. The calculation rule is selected. Which other feature amount is required to calculate a certain feature amount fa when calling each rule? The feature extraction means (6) stores the correspondence table stored in the feature extraction rule storage means (5). Check. FIGS. 8A and 8B show a part of the correspondence table. In FIG. 8 (a), the feature values fa and fb have not yet been calculated and f
c and fd indicate the already calculated state. FIG. 8b shows a feature quantity fa
It is a table showing that the feature quantities fb, fc, and fd must be already calculated when calculating. That is, whether the feature quantity calculated by the rule selected according to the correspondence table can be directly calculated from the image data,
It can be seen from the correspondence table as shown in FIG. 8 (b) whether or not it should be calculated after calculating other feature amounts. For example, feature quantity
A rule for calculating the feature amount fa is called to calculate fa, but it can be seen from the table in FIG. 8 (b) that the feature amount fb must be calculated in order to calculate the feature amount fa. When it is understood from FIG. 8 (a) that the feature amount fb is not calculated, the feature extraction control means (6)
A rule for finding the feature quantity fb is also called from that need, and fb
Is calculated. When it is necessary to add a new rule because the type of drawing to be input is different, the operator can add the new rule to the feature extraction rule storage means (5) by the rule correction means (11). It is possible and various drawings can be handled flexibly by changing the correspondence table at the same time.

The feature amount of the area information and the line information calculated in this way is stored in the third storage means (7). The determination rule storage means (8) has a rule for determining whether or not the feature amount is the feature amount of the symbol or line to be recognized. The determination rule control means (9) detects a symbol or a line based on a correspondence table of a recognition target and a determination rule necessary for detecting the recognition target similar to the table shown in FIG. Is selected from the judgment rule storage means (8) and taken in, and it is judged whether or not the feature quantity stored in the third storage means (7) is a feature quantity suitable for the recognition target.
Detect the recognition target. For example, with respect to the symbol shown in FIG. 7, the area of the black region is within a certain range, and the distance between the black region and the region is within a certain range. A rule is selected to determine an area that satisfies the condition that the size of the circumscribed rectangle of the combined area is within the range of the size of the symbol as the symbol area. The judgment rule control means (9) sequentially selects the rules to be judged from the judgment rules (8), checks whether or not the characteristic amount necessary for the judgment of the symbol area has already been obtained, and then according to the rule, the symbol area is determined. It is determined whether or not the symbol is detected and separated. Judgment rule control means (9) if the characteristic amount required for judgment has not been obtained.
Returns the control to the feature control means (6) again, and the feature control means (6) extracts from the feature extraction rule storage means (5) a rule for calculating the feature amount necessary for the determination, and according to the rule, the feature amount is calculated. To calculate. Further, as shown in FIG. 9, the judgment rule control means (9) is also used when separating touched characters.
From the third storage means (7) shows that the line length is more than a certain length.
Extract l ₁ , l ₂ and l ₁ , l ₂
The character and the line are separated by extracting from the determination rule storage means (8) a rule for extracting the line having the smallest difference in inclination (θ ₁ and θ ₂ ). In this way, the feature extraction rule (5) is provided separately from the rule for calculating the feature amount necessary for detecting / separating the elements of the symbol or the line, the feature extracting means (6) and the determining rule controlling means (9). , By providing the determination rule (8), a rule or determination for calculating a feature amount necessary for detection / separation even when the type of symbol is changed or the contact condition between a character and a line is changed Targets can be detected or separated simply by adding or deleting rules for doing so.

Information on the detected and separated symbols and lines is stored in the fourth storage means (10). The object recognition means (12) inputs the information of the symbol or line, recognizes the symbol or line, and stores the code of the symbol or the code information of the line in the fifth storage means (13) for each element. The symbol information and line information input in this way are used as maps and equipment management information.

〔The invention's effect〕

As described above, according to the present invention, a rule for calculating and determining a feature amount for detecting and separating an object (for example, a line forming an element such as a symbol, a house, a character, or a road) in a drawing is extracted by the feature extraction. By providing the rules (5) and the judgment rules (8) separately from the feature extraction means (6) and the judgment rule control means (9), the types of symbols to be detected and separated and the target element By adding or deleting rules for extracting the target even if it changes
It is possible to easily input information of another drawing with different rules to be used, and to detect and recognize a specific recognition target, the feature amount of a line, and the above-mentioned feature amount related to this feature amount. A feature amount calculation rule is selected as necessary using a table in which other feature amounts are associated with each other and a table in which the feature amount is calculated based on this table and the feature amount is associated with each other. Since the feature amount is calculated, only the rules necessary for inputting the drawing currently being read are automatically called, so that efficient drawing input processing can be performed and a new type of drawing can be input. Also in the case of doing, a rule called in a table in which the rule for calculating the feature amount and the feature amount are associated with each other, or the feature amount and the other feature amount related to this feature amount. Possible to correspond only by changing the was correspondence table, the various drawings easily and flexibly cope.

[Brief description of drawings]

FIG. 1 is an overall configuration diagram showing an embodiment of the present invention, FIG. 2 is an example of a drawing to be input, and FIGS. 3 (a) and 3 (b) are image information and point sequence information showing the center of a line. Explanatory drawing, FIG. 4 (a)
FIG. 5B is an explanatory diagram for extracting a black area and a white area.
The figure is a figure explaining the feature quantity when detecting the symbol area,
FIG. 6 is a diagram for explaining the feature quantity when detecting line elements,
FIG. 7 is a diagram for explaining an example of determination rules when detecting a symbol area, FIG. 8 is an explanatory diagram for feature extraction rules, FIG. 9 is an explanatory diagram for detecting line elements, and FIG. It is a block diagram of the drawing input device of FIG. In the figure, (1) is an image input means, (2) is a first storage means, (3) is an image processing means, (4) is a second storage means, (5) is a feature extraction rule storage means, ( 6) is feature extraction control means, (7) is third storage means, (8) is determination rule storage means, (9) is determination rule control means, (10) is fourth storage means, and (11) is Rule correction means, (12) object recognition means, (13) fifth storage means, and (14) element detection and separation means. In each figure, the same reference numerals indicate the same or corresponding parts.

Claims (2)

[Claims] 1. A drawing input device for inputting a drawing to detect a specific recognition target in the drawing and recognizing the same, and image input means for reading the drawing to create image information, and an area and a line from the image information. The image processing means extracted as, the area necessary for detecting and recognizing the specific recognition target, the rule for calculating the feature amount of the line, the feature amount, and the other feature amount related to the feature amount. Feature correspondence rule storage means for storing a corresponding table, a table for associating the rule for calculating the feature quantity based on the table with the feature quantity, and the feature extraction rule storage means for storing the feature extraction rule storage means. Based on both tables, a rule for calculating the feature amount of the element extracted by the image processing means is selected from the feature extraction rule storage means and based on the selected rule. Feature extraction control means for calculating the feature amount of the element, rules for determining whether the feature amount of the area or line is the feature amount of the recognition target, and for detecting and recognizing the recognition target and the recognition target Determination rule storage means for storing a correspondence table with a determination rule required for the determination rule storage means, and a determination rule for detecting the specific recognition target based on the correspondence table stored in the determination rule storage means. Determination rule control means for detecting the specific recognition target by determining the feature amount of each element calculated by the feature extraction control means based on the selected determination rule, and the determination rule control means. And a target recognition means for recognizing the recognition target detected by and generating corresponding code information. 2. The drawing input device according to claim 1, further comprising rule correction means for correcting the rules stored in the feature extraction rule storage means and the determination rule storage means.