JPS62274467A - Intellectual conversation processing method - Google Patents

Abstract

PURPOSE:To improve the operability and to shorten the processing time by displaying superposedly a center line vector and a graphic vector over original image data on a display device. CONSTITUTION:A vector processing part 18 extracts the required part of a pertinent center line vector from a center line vector auxiliary memory 16 and displays and emphasizes it on the display device through a display control part 20. When an operator recognizes it, an identifier giving part 19 gives an identifier to the extracted center line vector and stores it in a graphic vector auxiliary memory 12 as a graphic pattern vector. The vector processing part stops displaying and emphasizing the center line vector and displays it as the graphic vector again. When the processing to convert the displayed center line vector to the graphic pattern vector is terminated, the contents of the graphic vector auxiliary memory 12 are transferred to a graphic vector memory 13.

Description

[Detailed Description of the Invention] 3. Detailed Description of the Invention (Technical Field to Which the Invention Pertains) The present invention relates to a figure input method for inputting figures as vector data such as points and line segments into a computer etc. through conversational processing. be.

(Prior art) The first conventional method is to place a drawing or the like on which a figure is drawn on a digitizer (a figure input device that detects specified position coordinates), and from there an operator points out a point on the figure. There is a method of inputting a figure as vector data by specifying .

The second method of the conventional method is to input one drawing as image data into a computer etc. using a scanner, display it on a display, and convert one figure into vector data by having the operator specify points on the figure on the display. There is a way to do it.

FIG. 1 is a diagram illustrating an example of the second conventional method.

This figure is a part of the figure displayed on the display. For ease of explanation, it is assumed that the figure to be input, that is, the original image data, is drawn in black on a white background. In the figure, the area surrounded by solid lines 1.2 represents the drawn area (dot area), and the other areas represent the background area (white area).

The operator selects point p 1 while referring to the original image data.
1, p 21. ..., P,' are input sequentially to generate a graphic vector 3.

FIG. 2 is a diagram showing an example of the configuration of the second conventional method.

The input drawing 4 is scanned by a scanner 5, converted into black and white binary digital two-dimensional original image data, and stored in an original image data memory 6.

Of the original image data, only the portion necessary for display is stored in the original image data sub-memory 7 and displayed on the display 9 via the display control section 8.

While viewing the original image data on the display 9, the operator sequentially inputs necessary points thereon using the indicator 10.

The vector generating section 11 generates a graphic vector using the input points according to the operator's instructions, and stores the generated graphic vector in the one-graphic vector sub-memory 12 . The sub-memory has the function of decoding and storing graphic vectors into rask-type graphic image data for display on the display.

The contents of the graphic vector sub-memory are stored in the graphic vector memory 13 at the stage when the input work on the display screen is completed.

The first conventional graphic input method has the disadvantage that, in order to improve the input accuracy, additional man-hours are required, such as enlarging the input drawing in advance.

Another drawback is that input accuracy varies depending on the skill level of the operator.

Furthermore, the operation is complicated and time-consuming, as it is necessary to alternately compare the drawing on the digitizer with the display screen to verify completed input and confirm that there are no omissions. Ta.

In the conventional second figure input method, when the image data has a line width, the input points of coordinates such as end points and intersection points may differ depending on the operator, and conversely, when the line width is thin, the input points may differ depending on the operator. There is a drawback that the input accuracy varies, such as point P' in FIG. 1, where it protrudes from the line.

Furthermore, inputting one curve has the disadvantage that the input accuracy varies because the line segments are approximated at appropriate intervals based on the intuition of the operator.

(Object of the Invention) The present invention was made to eliminate the above-mentioned drawbacks, and an object of the present invention is to provide an intelligent conversation processing method that inputs figures as vector data accurately and efficiently. It is something.

(Structure of the Invention) A first method of the present invention is to input a drawing into a computer as image data using a scanner such as a facsimile, and display the image data and center line vector in a superimposed manner on a display. Indicate the neighboring points of the input figure using the coordinate indicator above, correct the position of the indicated point to a point on a predetermined center line vector, and set the center line corresponding to one or more of the correction points. Graphics are input as vector data through semi-automatic conversation processing that extracts vectors, assigns identifiers to the extracted vectors, and converts them into graphic vectors.

The second method of the present invention is to perform automatic image processing such as figure recognition processing on the original image data or center line vector in advance to generate a figure vector having an identifier, and then combine the figure vector with the original image data and the center line vector. are displayed in a mutually distinguishable format, and only the center line vector portion that has not been converted into a graphic vector is inputted as vector data through semi-automatic conversation processing similar to the first means.

(Example) Hereinafter, an example of the first method of the present invention will be described in detail using the drawings.

3 to 5 are examples showing a part of the display surface of a display for explaining the concept of the method including the automatic image processing function.

In these figures, the solid lines indicate the boundaries of the drawn portions of the original image data, the broken lines are center line vectors, and for the sake of simplicity, it is assumed that curves are also approximated by line segments.

The black points P4 to P, to P2□, P, and 1 to P3g are the starting points or ending points of the line segments that constitute the center line vector, and are hereinafter referred to as points.

Points Q□ to Q are points indicated by the operator from the indicator.

For example, the drawn portion is displayed in light blue, and the center line vector is displayed in blue.

FIG. 3 is a diagram illustrating an example of extracting a closed loop on a centerline vector near a designated point.

That is, when the designated point Q is input, a point P on the center line vector near Q is detected. Starting from the point P1, for example, look for a closed loop on the center line vector in the direction of looking at the input point Q to the left.

At point P2, the center line vector branches in the direction of point P and in the direction of point P7, but in this case, look at the point immediately before the branching point and select the line segment that first appears in the clockwise direction. do. In the example, line segment P2P is selected.

Do the same below.

Line segment P 3 P 4- P 4 P s -P s P
Select I P s P L.

Extract the closed loop P-work P, P, P4PSP, P1.

The extracted closed loop is highlighted by displaying it in red or blinking so that it can be distinguished from the original image data or center line vector.

If the highlighted automatic extraction result is correct, the operator assigns an identifier such as "building" representing the type of figure to the closed loop and saves it as a figure vector.

Once the identifier has been assigned, the closed loop is highlighted, and if necessary, it is redisplayed in the color that indicates a building or the color that indicates a geometric vector, and the needle that has already been identified as a drawing vector, not just a center line vector, is displayed again. Make it clear to the operator.

FIG. 4 is a diagram showing an example of extracting a smooth curve in which the direction of the back wheel does not change suddenly between two designated points.

In other words, when two pointing points Q2 and Q are input, Q2
, Q, on the center line vector in the vicinity of each point P 0. P
Extract the

Starting from point pit, point Pi. The center line vector is traced in the direction of .

Point P. When you encounter a branch point like this, select the line segment with the smallest angle from the extension direction of the previous line segment. At point P, select line segment P1□P.

In the same manner, smooth curves (in this example, line segment approximation) Pp□pHP13 pz4p□5ptsPpp are extracted and highlighted.

If the automatic extraction result is correct, the smooth curve is given an identifier such as "road" representing the type of figure, and is saved as a figure vector.

FIG. 5 is a diagram illustrating an example of extracting the inner boundary between two designated points.

In other words, when two pointing points Q, , Qs are input.

Point P3□ on the center line vector near each of Q,,Q,
,P,7 is extracted.

Note that the first designated point is, for example, the starting point when the inner boundary is viewed as curving counterclockwise in a general perspective.

Starting from points P and □, line segments are sequentially selected in the vicinity of each point in a clockwise direction from the direction of the immediately previous point.

At point P34, a line segment is locally searched clockwise from direction P34P33, and the first detected line segment P, 4P3G is selected.

In this way, the vector P3XP, □
P33P34P3GP,7 is extracted and highlighted. If the automatic extraction result is correct, its inner boundary.

For example, an identifier "building" indicating the type of figure is assigned and saved as a figure vector.

FIG. 6 shows an example of a block diagram of an embodiment of the first method of the present invention.

In the same figure, 4, 5. 6. 7, 9. 10. 12. 13 are the second
This is the same as explained in the figure.

Reference numeral 14 denotes a center line vector conversion unit which converts the original image data into center line image data with a width of 1 pixel by thinning processing or the like, and further converts the data into a vector by line segment approximation or the like to obtain a center line vector.

Regarding line thinning processing.

Literature “IEICE Technical Research Report, PRL75-66J
This is omitted here as it has been explained in detail elsewhere.

Regarding the line segment approximation process, please refer to the document "1982 Institute of Electronics and Communication Engineers General Conference No. 125".
2J etc., so it will be omitted here.

Reference numeral 15 denotes a center line vector memo. 16 indicates a center line vector sub-memory, which temporarily stores only the part necessary for displaying the center line vector in white. It is also decoded and stored as image data.

Reference numeral 17 denotes a pointing point correction unit, and 18 denotes a vector processing unit, which extracts a center line vector related to a pointed point, and performs editing such as adding, deleting, or creating a new figure vector. an identifier assigning unit that assigns an identifier;
Reference numeral 20 denotes a display control unit that controls the display of original image data, center line vectors, graphic vectors, indicated points, etc. in a state where they can be identified.

The center line vector converter 14 converts the original image data memory 6 into
The original image data is read from , converted into a centerline vector, and stored in the centerline vector memory 15 . This memory stores the centerline vectors of the entire drawing, and the portion necessary for one display is transferred to the centerline vector submemory 16.

If one screen can be displayed on the display at a time, the center line vector memory 15 can be omitted.

The display control unit 20 reads out the contents of the original image data sub-memory 7 and the contents of the centerline vector sub-memory 16 corresponding to the original image data, and displays them on the display 9 in a necessary display format.

The pointing point inputted by the pointing device 10 is also simultaneously displayed on the display 9 by the display control section 20 as a cursor of a necessary shape and color.

The indicated point correction unit 17 searches for a point on the center line vector in the center line vector sub-memory 16 near the indicated point inputted by the indicator 10.

Specifically, the coordinates of the indicated point are (x, y), and d is set to 1 in advance.
0 pixel, etc., and search for a point (Xp, Yp) on the center line vector that satisfies the following conditions: X-d≦Xp≦X+d, Y-d≦Yp≦Y+d.

If there is no point that satisfies the condition, the operator is informed of this fact and prompted to re-enter the designated point.

If there is a point (Xp, Yp), that point is set as the correction point for the designated point, and the operation continues assuming that point has been input.

When one or more indicated points are input and the correction thereof is completed, the vector processing unit 18, for example, as shown in FIGS.
Centerline vector sub-memory 16 in the manner described in the figure.
A necessary part of the corresponding center line vector is extracted from the cursor and highlighted on the display via the display control unit 20.

If the operator recognizes it, the identifier assigning unit 19 assigns an identifier to the extracted center line vector and stores it in the graphic vector sub-memory 12 as a graphic vector.

The vector processing unit 18 stops highlighting the centerline vector and displays it again as a graphic vector.

When the process of converting the center line vector being displayed into a graphic vector is completed, the contents of the graphic vector sub-memory 12 are transferred to the graphic vector memory 13.

Thereafter, the contents of the original image data sub-memory 7, the center line vector sub-memory 16, or the graphic vector sub-memory 12 are written as necessary in order to perform figure input processing in other areas of the input drawing 4. Change it and repeat the above operation.

Next, an embodiment of the second method of the present invention will be described in detail using the drawings.

FIG. 7 is a diagram for explaining an embodiment of the second method of the present invention, and is an example showing a part of the display surface of a display for explaining the concept of the method.

In the figure, the solid line indicates the boundary of the drawing part of the original image data, the dashed line is the center line vector, and the dash-dotted line is the automatic figure! ! It shows the figure vector recognized by the cognitive process.

Sunspot P4°~P4. are the end points of the line segments that make up the centerline vector.

Sunspot P4°~P42. P4s and P□ are also the end points of the line segments that constitute the figure vector.

Points Q, , Q7 are points indicated by the operator using the indicator.

For example, the drawn portion is displayed in light blue, and the center line vector is displayed in blue.

Graphic vector-p4. p, 1p4. and P47P4.・
... is part of the "road".

The graphic vector P42P4sP□... is part of the "contour line".

A method of converting the unrecognized center line vector P4□P43P44 into a "road" graphic vector by semi-automatic conversation processing will be explained.

Instruct to input "road" in advance, and input the recognized "road" graphic vector...P4゜P4, PI
3, and the position is corrected to point P,2.

The neighborhood Q of P44, which is the other end point of the center line vector to be additionally input, is designated, and the position is corrected to point P44.

Using the point P42 as a starting point, curve tracing is performed using the method shown in FIG. 4, etc., and the center line vector P4□P43P44 is extracted and highlighted.

When the operator confirms it, the extracted vector is added to the recognized graphic vectors...P4°P4iP42.

Figure vector...P 40 P 41 P 42 P
43 P 44.

The extraction vector will no longer be highlighted, and the shape vector will no longer be highlighted.
...P4゜P41 P42 P4) Redisplay as P44.

FIG. 8 is a diagram showing the configuration of an embodiment of the second method of the present invention, and the components other than the figure recognition section 21 are the same as those in FIG.

In the first embodiment, all figure vectors were generated by conversation processing, but in the second embodiment, center line vectors are generated to the extent possible by image processing such as automatic figure recognition processing before conversation processing. is given an identifier and converted into a single figure vector.

The figure recognition unit 21 is a part that performs this image processing.

As an example of a specific method of figure recognition, please refer to the document “Preface of Information Systems Division of the 1985 National Conference of the Institute of Electronics and Communication Engineers, No.
56-6'' etc., and detailed explanation will be omitted here.

In FIG. 8, the figure recognition unit 21 reads the original image data or the center line vector (or both data depending on the case) from the original image data memory 6 or the center line vector memory 15, and recognizes buildings, roads, etc. Each identifier is assigned to the identified vector and stored in the graphic vector memory 13 as a graphic vector.

Note that the automatic image processing function of the vector processing section 18 is the same as a part of the automatic image processing function of the figure recognition section 21, and it is also possible to make these parts common.

The indicated point correction unit 17 stores the indicated point and center line vector sub-memory 1.
In addition to correcting the position of the point on the center line vector within 6, if necessary, the position of the indicated point is corrected to the point on the figure vector in the figure vector sub-memory.Other operations are explained in Figure 6. This is the same as in the case of

(Effects of the Invention) As explained above, according to the present invention, the original image data, the center line vector, and the figure vector are displayed in a superimposed manner on the display, which improves operability and reduces processing time. There is.

Further, since vectorization of image data is automatically performed in advance using centerline image data, it is possible to input graphic vectors with extremely high accuracy, regardless of individual differences or skill level of the operator.

Furthermore, since the position of the designated point is automatically corrected to a point on the center line vector, the operator can input the designated point somewhat roughly, which has the effect of allowing high-speed processing with a light burden.

Furthermore, since the recognition function is also used in conversation processing, there is an effect that more vector figures can be input with fewer pointing points.

[Brief explanation of the drawing]

FIG. 1 is a diagram illustrating an example of the conventional second method, FIG. 2 is a diagram showing a configuration example of the conventional second method, and FIGS. 3 to 5 are diagrams illustrating the present invention including an automatic image processing function. FIG. 6 is a block diagram of an embodiment of the first method of the present invention, FIG. 7 is a diagram explaining the second method of the present invention, and FIG. 8 is a diagram of the embodiment of the first method of the present invention. It is a figure which shows the structure of the Example of a 2nd method. 4...Input drawing, 5...Scanner, 6...
・Original image data memory, 7: Original image data sub-memory, 8: Display control unit, 9: Display, 10: Indicator, 1
1... Vector generation part. 12...Graphic vector submemory. 13... Graphic vector memory, 14... Center line vector converter, 15... Center line vector memory, 16... Center line vector sub memory. 17... Indicated point correction section, 18... Vector processing section,
19...Identifier assigning unit, 20...Display control unit. 21...Graphic recognition section. Patent applicant Nippon Telegraph and Telephone Corporation Figure 1 Figure 2 Figure 3 Figure 4 Q2 Figure 5 p.

Claims (2)

[Claims] (1) A method for inputting a figure as vector data in a conversational format, which includes: a first means for accumulating original image data of the figure; a second means for converting and accumulating the original image data into a centerline vector; a third means for superimposing the center line vector corresponding to the original image data in an identifiable state on the same display; a fourth means for indicating an arbitrary point on the display; and an instruction input by the fourth means. A fifth means for obtaining a correction point whose position is corrected to a point on a predetermined centerline vector, and a centerline vector corresponding to one or more correction points obtained by the fifth means. a sixth means for extracting the portion using an automatic image processing function; a seventh means for adding an identifier to the center line vector obtained by the sixth means to make it a figure vector; and a seventh means for making the figure vector recognizable. An intelligent conversation processing method, comprising: an eighth means for displaying the state, and a ninth means for accumulating the graphic vector. (2) A method for inputting a figure as vector data in a conversational format, comprising: a first means for accumulating original image data of the figure; a second means for converting and accumulating the original image data into a centerline vector; and third means for generating a graphic vector having an identifier by performing automatic image processing on the original image data or the center line vector. a fourth means for accumulating graphic vectors; a fifth means for superimposing a center line vector and a graphic vector corresponding to the original image data on the same display in a distinguishable state; and indicating an arbitrary point on the display. a sixth means; a seventh means for obtaining a correction point by correcting the position of the indicated point input by the sixth means to a point on a predetermined centerline vector; an eighth means for extracting a necessary part of the centerline vector corresponding to one or more correction points using the automatic image processing function of the third means; and identifying the centerline vector obtained by the eighth means. An intelligent conversation processing method, comprising: a ninth means for displaying in a possible state; and a tenth means for adding an identifier to the center line vector obtained by the ninth means and converting it into a graphic vector.