JPH04117579A - Hatching area extracting method - Google Patents

Abstract

PURPOSE:To extract a hatching area by parallel lines or points and to process the hatching area as the data of polygonal data by assigning a rectangular area in the area represented by hatching in a graphic and discriminating the pattern of the hatching from an outline vector within the rectangular area. CONSTITUTION:The confirmation of a graphic is executed by displaying an outline vector and a making core line vector generated by an outline vector generating part 3 and a making core line processing part 4 on a display device 21 as drawing. Then, a rectangle is assigned within the graphic by a mouse 22. Next, a pattern information within a rectangle extracting means 23 searches the outline vector within the rectangle from a coordinate value and when the outline vector is searched, the linkage of the outline vector is traced. Then, a closed loop which consists of the outline vector is constituted, it is discriminated as a hatching pattern by parallel lines when the closed loop is larger than the former rectangle and it is discriminated as the hatching pattern by points when the closed loop is smaller than the former rectangle. Thus, processing becomes easy.

Description

[Detailed description of the invention] [Industrial application field]

The present invention relates to a hatching area extraction method for extracting a hatching area composed of parallel lines or points from a vector image. 1 Prior Art] With the advancement of image processing technology, graphic processing technology using computers is being used in various fields. As such graphic processing technology, there is a method for computer processing of line drawings such as maps and illustration images. Research and development is underway. Examples include ``Study on Automatic Extraction of Map Components,'' Information Processing Society of Japan Section 39 National Conference, ``Hatching Area Extraction Algorithm and Its Application to Topographic Maps,'' and 1981 National Conference of the Institute of Electronics and Communication Engineers. [Problems to be Solved by the Invention] The processing technology described above processes figures not as line drawings but as rask image data, which is the same as binary photographic images, etc., so processing such as image enlargement and reduction is easy. Although it is possible to do so, there is a problem in that it is difficult to partially correct or change an image. On the other hand, it has been considered to process line drawings by vectorizing them (for example, [Automatic Recognition of Topographic Maps Using Relationships between Line Segments J, Institute of Electronics, Information and Communication Engineers PRU89-24P,
“Automatic map input system with improved reliability through conversational recognition”, Transactions of the Institute of Electronics, Information and Communication Engineers, 89/4 V
OL, J72-D-2No, 4). However, in the case of line drawings such as hatching patterns, the shapes such as polygons formed by the hatching pattern are more meaningful than the lines or points themselves, and the above-mentioned conventional techniques do not address this point. It wasn't taken into account. SUMMARY OF THE INVENTION An object of the present invention is to provide a hatching area extraction method for extracting a hatching area formed by parallel lines or points and processing the hatching area as polygonal data. [Means for Solving the Problem] In order to achieve the above object, in claim (1), the outline of a binarized input image is vectorized and then converted into a core line,
In a method of outputting line drawing data by converting the input image into a skeleton vector and a connecting vector connecting the skeleton vectors, a rectangular area is specified in the area represented by hatching in the figure, and a contour vector within the rectangular area is In claim (2), wherein the hatching pattern is determined from the contour vector, if the closed loop formed from the contour vector is larger than the rectangle, it is determined that the hatching pattern is formed of parallel lines. In claim (3), if the closed loop is smaller than the rectangle, it is determined as a hatching pattern of points. Claim (4) is characterized in that the parallel line hatching area is extracted by searching for a constituent line, creating a link between adjacent constituent lines, and tracing the end point of the constituent line from the link to obtain a line outside the area. , the coordinates of the end points of the constituent lines are recorded, by searching the contour vector inside the external line constituting the figure and finding a sequence of the end points of the contour vector, It is characterized by extracting a hatched area using points. [Function] According to claims (1) and (2), by specifying a rectangular area in the area represented by the hatching in the figure, the pattern of the hatching is a parallel line based on the contour vector in the rectangular area. It is possible to determine whether the line is a point or a point, and extract information about the slope of the line and the interval between the lines. In claims (3) and (5), since a hatched area formed by parallel lines and points is extracted as polygon data, the area can be processed as a polygon. Furthermore, it is possible to change the hatching pattern within the area, and furthermore, to transform the area itself. In claim (4), the position of the hatching pattern can be detected.

【Example】

Hereinafter, one embodiment of the present invention will be specifically described using the drawings. FIG. 1 is a functional block diagram according to an embodiment of the present invention, in which 1 is an image input unit consisting of a scanner etc. that reads a figure formed by a hatching pattern, and 2 is a unit that binarizes an input image. A binarization processing section 3 is a contour vector generation section that generates a contour vector from the binarized hatching pattern, and 4 is a skeletonization processing section that generates a skeletonized vector from the contour vector. The skeletonized line drawing data 7 is output by the skeletonized processing section 4. Note that since the method of generating a skeletonized vector from a contour vector is well known, a detailed explanation thereof will be omitted. Reference numeral 5 denotes a rectangular area specifying unit that specifies a rectangular area within the vectorized hatching area. 6 is a hatching area extracting unit which outputs hatching area extraction data 8; The configuration of the rectangular area specifying unit 5 is shown in FIG. 2. The rectangular area specifying unit 5 is a C
It is comprised of a display device 21 such as an RT, a pointing device 22 such as a mouse, and an in-rectangle pattern information extraction section 23 that extracts hatching pattern information from within a rectangle designated by the mouse 22. Rectangle designation processing: The contour vector and skeletonization vector generated by the contour vector generation section 3 and skeletonization processing section 4 are displayed as a line drawing on the display device 21 to confirm the figure. Then, a rectangle is designated within the graphic using the mouse 22. This rectangle is specified by specifying two points, the upper left position and the lower right position, on the display screen. However, in the case of a hatching pattern using parallel lines, the line must be specified so that one line falls within the rectangle. In addition, external line 2 representing an area that is not a hatching pattern
It must be specified so that the 4th grade does not fall within the rectangle. Hatching pattern determination process: Next, the rectangular pattern information extraction unit 23 searches for a contour vector within the rectangle from the coordinate values, and when a contour vector is found, traces the connection of the contour vectors. This creates a closed loop consisting of contour vectors. If this closed loop is larger than the previous rectangle, it is determined to be a hatching pattern using parallel lines (Figure 3a), and if the closed loop is smaller than the previous rectangle, it is determined to be a hatching pattern using points (Figure 3b). . Parallel line hatching pattern extraction processing: ■Processing for finding an extension line made up of one or more skeleton lines. Start. At this time, even if the original image is a single hatched line, its contour may be disturbed by noise or the like, so that the core line is not a continuous line but is cut at various parts. Therefore, the core wire is extended by examining both ends of the core wire. That is, as shown in FIG. 4, when the next core vector 43 is connected to the end point 42 of the core vector 41, and as shown in FIG. Okay, vector formed in the surplus part that could not be converted into a core line) 55.56
Accordingly, when two core vectors 51 and 52 are connected, the core line is extended.However, as shown in FIG. 6, when another core vector 62 is connected to the connecting vector 61, does not extend the core wire. Furthermore, since the extension line may be bent while being extended, whether or not to extend it is determined based on the distance and angle. That is, FIG. 7 is a diagram for explaining judgment based on distance, where 71 is a skeleton vector that is the basis of extension, 72 is the first extended skeleton, 73 is the nth extended skeleton, and 74 is the skeleton vector 71. A line 75.76 connecting the end points of 73 is the distance di, d2 between the line 74 and the end point 77.78. If this distance di, d2 is larger than a certain threshold,
The core vector 71 is not extended to the core vector 72. FIG. 8(a) is a diagram for explaining judgment based on angle, where 81 is a skeleton vector on the It is located at the origin of the axis, and 83 is a line that serves as a threshold when determining extension. If the core line vector 84 crosses the threshold line 83 and appears on the right side, no extension is performed. The reason why the threshold line 83 is provided offset from the origin is that the extension core line may change its direction significantly in a minute section as shown in FIG. 8(b), and even in such cases, the extension This is to make it possible. Through the above processing, an extension line serving as a reference for the hatching pattern is obtained, and the slope of the extension line becomes the slope of the hatching pattern. (2) Process for finding spacing between parallel lines FIG. 9(a) is a diagram for explaining the process for finding an extension line constituting a hatching pattern adjacent to a reference hatching pattern. Reference numeral 91 denotes a hatching pattern composition line that serves as a reference (out of extension lines, those determined to constitute a hatching pattern are referred to as pattern composition lines), 92 a search starting point, and 93 a search direction, 94 is core line vector, 95
is an extension of the core line. Using the end of the hatching pattern constituent line 91 as a search starting point 92, an adjacent core line 94 in the normal direction 93 is searched for. If the slope of the extension line 95 formed by connecting these core lines is close to the reference slope, the extension line 95 becomes the second pattern forming line. Then, the distance between these two pattern constituent lines becomes the interval between the hatching patterns. On the other hand, if the slope of the extension line 95 is significantly different from the standard, another extension line must be found. FIG. 9(b) is a diagram for explaining the process of searching for another extension line, in which an end point 96 of an extension line 95 with a greatly different slope is mapped onto the hatching pattern constituent line 91 (end point 97).
, moves the search starting point from the initial starting point 92 to the end point 98 side. From that point 98, the core wire 99 is cast in the same manner as described above.
This process is repeated until an extension line whose slope is close to the reference slope is found. Extraction process of dotted hatching pattern: In the case of a dotted pattern, the size of the dot is determined from the closed loop size of the contour vector. Therefore, the size of the dotted hatching pattern can be determined by averaging the sizes of the closed loops within the rectangular area. FIG. 10 shows the data format of the hatching pattern obtained as described above. The type of pattern represents a parallel line or a point, and pattern data 1 represents the slope of the parallel line, and the size of the point when it is a point.
represents the interval when it is a parallel line, and does not write anything when it is a point. Processing for Extracting Hatching Regions: The operation of the hatching region extracting section 6 to extract regions from the above-mentioned hatching pattern data will be described in detail below. ■When the hatching pattern is parallel lines Figure 11 shows the procedure for extracting a parallel line hatching area.In step 111, extension lines that make up the pattern are searched for, links are created between adjacent constituent lines, and in step 112
Now, trace the end points of the component line from that link to find the area external line. The detailed processing procedure of step 111 is explained in the 12th section.
FIG. 13 shows the detailed processing procedure of step 112. After explaining the definitions of terms used in FIGS. 12 and 13, the processing in FIGS. 12 and 13 will be explained in detail. 14(a) and 14(b) are diagrams for defining the starting point and ending point of the extended line, and the end point of the extended line is on the upper side (142
) or the point on the left side (144) is the starting point, therefore, the end points 143 and 145 are the ending points, and the right side from the ending point to the starting point is the R side and the left side is the L side. Note that θ is an angle with respect to the X axis. Returning to FIG. 12, the process starts from the starting point of the reference line obtained by the hatching pattern extraction process described above (step 1201). From this starting point, Fig. 9(a),
According to the processing procedure shown in FIG. 9(b), an adjacent component line on the R side is searched for (steps 1202 and 1203). However, at this time, even if one component line is found, the process does not end. The search start point is moved as shown in FIG. 9(b), and the process is continued until the search start point is located at the end point. When the processing on the R side is completed, the constituent lines are similarly searched on the L side (steps 1204 and 1205). In this way, the columns of adjacent constituent lines on the R side and L side of one constituent line are determined, and a table of data columns as shown in FIG. 15 is created. The number of constituent lines is written in the part 151, and the number of the constituent lines given in the order found by the search is written in the part 152. Next, in this table, search processing on the R side and L side is similarly performed for the constituent lines for which the above search processing has not been completed (steps 1208 and 1209). Furthermore, if a new constituent line is found, the same search process is performed for that constituent line (steps 1206 and 120).
? ). The search process is performed on all the constituent lines numbered in this way, and the process ends. When the search process is completed and the links between the component lines are determined, the area external lines are extracted according to the processing procedure shown in FIG. FIG. 16 shows an example for explaining the extraction of this area external line. M is the component line numbered N and M, respectively. al, a2. a3 is the starting point of the constituent lines N, M, and bl, b2 . b3 is the end point, and 89 from sl represents the core line vector and the connecting vector. Also, Fig. 17 is a table showing the links between the component lines N and M.
17] is a list on the R side of the configuration line N, and 172 is a list on the L side of the configuration line. Hereinafter, the area external line extraction process will be described in detail with reference to FIGS. 16 and 17. First, the process starts from the starting point of the reference line (step 13
01). Next, the next point constituting the polygon representing the area is searched for on the R side from the starting point of the reference line. That is, with reference to the table (Fig. 17) of the constituent line of the current point of interest (this is referred to as N in Fig. 16), the list on the R side (
It is checked whether a constituent line exists at 171) in FIG. 17 (step 1302). If a constituent line exists (constituent line K), list on the L side of the constituent line (172 in Fig. 17)
In , the order of component line N in the list is checked. At this time, if N is not the head of the list as shown in the L side list 172 in FIG. The contour is connected from al to b2 (steps 1303.130
4), s2 becomes the external line. Then, the list on the L side is examined from this end point (step 1308). When there is no component line M, the process is moved to component line K (step 1305), and the list on the R side is further examined (step 1).
302), repeat the same process as described above. Through this process, the outline of the area is connected from al to 83. Furthermore, in step 1302, if the list on the R side of the component line of interest is empty, that is, if there is no component line K, the process is moved from the starting point a1 of the component line of interest to the end point bl (step 1307), Next, the list on the L side is examined (step 1308). When proceeding to the L side, the process is completely opposite to the case where the process proceeds to the R side described above. In other words, when checking the list on the L side, you will trace the end point of the component line, so check the last number of the L side list of the component line of interest, and
Whether or not processing proceeds from side to R side (opposite to the case in Figure 16) is determined by checking in the list on the R side of the next component line (Figure 17).
(corresponding to the list 172 in FIG. 17), and by checking whether there is another component line (corresponding to the position below N in FIG. 17) after the component line of interest (steps 1308 to 13). In this way, while tracing the end points, the process returns to the point where the process started, and the process ends. By connecting the end points of the constituent lines traced through the above process, a polygon representing the area is extracted. FIG. 18 is a list of point sequences in which the coordinates of the end point of interest are recorded each time the end point moves. In 181, the number of points is entered, and in 182, the coordinates of each point are entered. However, the first and last coordinates of the list are the same. (2) When the hatching pattern is dots FIG. 19 is a diagram for explaining the process of extracting a hatching area using dots. Origin 19 of the rectangle 191 mentioned earlier
2 toward the negative direction of the X-axis, a contour vector 194 inside the external line 193 is searched. The closed loop of the found contour vector 194 is larger than the previous rectangle 191,
If the rectangle 191 is included, the search ends; otherwise, the search is continued in the negative direction of the X axis. A closed loop of contour vectors that encompasses this rectangle becomes the contour inside the external line of the area, so by recording the row of end points 195 of the contour vector obtained by this process in the list shown in FIG. A polygon representing the area is extracted.

【Effect of the invention】

As described above, according to the present invention, the following effects can be achieved. ■By specifying a rectangular area within the hatching area, it is possible to automatically determine whether the hatching pattern is based on parallel lines or points.
At the same time, the inclination and spacing of the lines as well as their position information can be detected. (2) Since the hatched area formed by parallel lines can be extracted from the line data as polygonal data, the hatched area can be processed as a polygon, and the area can therefore be easily transformed. In addition, you can clean up the hatching pattern in the area (replace the hand-drawn hatching pattern with a registered clean pattern), change it (
For example, changing a solid line pattern to a broken line pattern can be easily performed. (2) Similarly, since a hatched area formed by points can be extracted from line data as polygonal data, the hatched area can be processed as a polygon, and therefore the area can be easily transformed.

[Brief explanation of the drawing]

FIG. 1 is a functional block diagram according to an embodiment of the present invention, and FIG.
The figure shows the configuration of the rectangular area specifying section, Figures 3(a) and 3(b) are diagrams for determining a hatching pattern based on parallel lines and a hatching pattern based on points, respectively.
The figure shows an endpoint of a core vector connected to the next core vector, Figure 5 shows two core vectors connected by a connecting vector connecting the end points, and Figure 6 shows a connecting vector and other connections. Diagram where the core line vectors of are connected, 7th
The figure is a diagram for explaining judgment based on distance, Fig. 8 (a)
, 8111 (b) is a diagram for explaining the judgment based on angle, and FIG. 9 (a) is for explaining the process of searching for an extension line constituting the hatching pattern next to the reference hatching pattern. Figure 9(b) is a diagram for explaining the process of searching for another extension line, Figure 10 is a diagram showing the data format of a hatching pattern, and Figure 11 is a diagram for extracting a parallel line hatching area. A flowchart showing the processing procedure, FIGS. 12 and 13 are detailed flowcharts of the extraction processing procedure, and FIGS. 14(a) and 14(b) show the starting point of the extension line,
A diagram for defining the end point, Figure 15 is a diagram showing a table of data columns, Figure 16 is a diagram for explaining extraction of area external lines, and Figure 17 is a link between constituent lines N and M. FIG. 18 is a list of point sequences in which the coordinates of the end point of interest are recorded each time it moves, and FIG. 19 is a diagram for explaining the process of extracting a hatching area using points. l...Image input section, 2...Binarization processing section, 3...
- Contour vector generation unit, 4... Core layering processing unit. 5... Rectangular area designation section, 6... Hatching area extraction section, 21... Display device 22... Mouse, 23... Rectangular pattern information extraction section, 24... External line. Figure 3 (0-) Figure 3 (G) Figure 4 Flies... Figure 6 Figure 7 Figure 8 (Good) Figure S (b) Figure 9 (α) Figure 9 (S) 971 Shinki Figure 11 Figure 10 Figure 14 (α) (b) Figure 15 Figure 16 Figure 17 Figure 18 Figure 19

Claims (5)

[Claims] (1) In a method of outputting line drawing data by vectorizing the outline of a binarized input image, converting it into skeleton lines, and converting the input image into a connecting vector connecting the skeleton vectors, A hatching area extraction method comprising: specifying a rectangular area in the area represented by hatching, and determining the hatching pattern from a contour vector within the rectangular area. (2) If the closed loop composed of the contour vector is larger than the rectangle, it is determined that the hatching pattern is composed of parallel lines; if the closed loop is smaller than the rectangle, it is determined that it is a hatching pattern of points. The hatching area extraction method according to claim 1, characterized in that: (3) Search for the constituent lines that make up the hatching pattern from the slope of the hatching pattern and the spacing between the patterns, create links between adjacent constituent lines, and trace the end points of the constituent lines from the links to find the outside line of the area. A hatching area extraction method characterized by extracting a parallel line hatching area. (4) The hatching area extraction method according to claim (3), further comprising recording coordinates of end points of the constituent lines. (5) A hatching area extraction method characterized by extracting a hatching area by points by searching for a contour vector inside an external line constituting a figure and finding a sequence of end points of the contour vector.