JPS62286176A - Converting/processing method for drawing data - Google Patents

Abstract

PURPOSE:To rapidly process massive data by a software by extracting only picture elements at both sides of a line without performing a fine line processing and obtaining vector data linearly approximated with a constant allowance. CONSTITUTION:An algorithm for converting image data into chain data reads one block data from a binary data file RAM board 4 to the memory of a personal computer 1. The presence and the absence of connection loop data is decided between blocks, when there is the loop data, connection chain data is traced and the data is respectively outputted to a loop chain data file and a chain data file. When there is no loop data, the start point of new loop data is retrieved. When the start point is retrieved, the chain data is traced. When the loop is closed in the block or not or whether the loop reaches a block connection part or not is examined, when the loop is closed, the start chain data is retrieved. When the loop reaches the block connection part, connection data is formed to return to the retrieval of the start chain data again.

Description

Detailed Description of the Invention 3. Detailed Description of the Invention (Technical Field of the Invention) This invention provides an image analysis method for reading various drawings such as one map, nine characters, etc. with an image scanner and processing them as CAD data. Concerning a method for converting data into chain data.

(Technical background of the invention and its problems) OA'lli With the technical progress of CAD+, automatic reading of various drawings has become important. For example, when doing design work, it is rare to create a completely new draft from the beginning, but after laying out the basic design, you may find a similar drawing and make some changes to it, or you may need to make some changes. In most cases, drafting begins with basic design drawings. However, when attempting to perform design and drafting work using a CAD system, the data and human resources required for these existing drawings incurs enormous costs, which is currently a major hindrance to making effective use of CAD. . Furthermore, in some cases, such as circuit design, it is more efficient to draw hand-drawn drawings rather than manually input data using CAD from the beginning. Against this background, automatic reading of existing drawings is the key to making the most of CAD systems. The present invention relates to automatically reading and converting existing drawings into numerical data, and there are two conventional methods. One method is to measure the drawings using a digitizer, which is called hand digitizing. This method has the advantage of being able to manually enter the attributes of a shape or connect broken lines on the spot at the time of measurement, since humans measure the shape while visually checking it.

However, on the other hand, measurement errors and omissions in data are likely to occur, so data checking is required, resulting in extremely high costs. The other method is called auto-digitizing, which uses automatic reading with an image scanner. This method reads the drawing as image data and then converts it into vector data, so there is no data omission or error, and the measurement cost is low. However, in drawings such as mechanical design drawings, which have various symbols and line types, and in which figures are separated by characters, etc., it is difficult to convert them into vectors, and this has not been put to practical use.

On the other hand, conventional drawing reading systems mainly process images of Al size, the system price is expensive because it uses hardware, and most of them are vectorized.
It is difficult to use it as human data for CAD.

Therefore, it is necessary for humans to edit this vector data into data that can be used in CAD using interactive editing, but vectorization and interactive editing are separate systems, and combining both systems is extremely difficult. It becomes expensive. Another drawback is that it requires a great deal of effort to edit data that can be used in CAD.

On the other hand, the thinning method used in the transformation process from image to vectorization is a method that extracts a center line of unit width by narrowing the line width from a given figure, without changing the connectivity of the original figure. In other words, it is required to continuously convert the figure into a line figure without cutting it or creating holes. Since the characteristics of the connection relationships between figures can be determined from the result of line thinning, line thinning has traditionally been an essential process for analyzing line structures from linear figures such as characters and drawings.

Many algorithms have been devised so far, depending on the quality and connectivity of the center line, but basically they take the following form. In other words, the boundary points in the image (4 - 1 in the vicinity)
Among the 11 pixels that have 01 pixels at any time), all pixels that are erasable elements and are not end points of the line (for example, the number of 11 pixels in the 8-neighborhood is 2 or less) are erased. This process is performed on the pixels of the entire image in one operation, and is repeated until there are no pixels left to be erased.

There are also sequential and parallel types of thinning algorithms. Each method has been devised, but overall, just like distance conversion and degeneracy, the principle is easier to think about in the parallel type, and in the sequential type, you can make full use of programming techniques to use clever conditions.

Particularly in line thinning, if a parallel algorithm is applied directly, 2! There is a problem in that a figure with a width of I pixels disappears at once. Either complicate the conditions so that only one side is erased in the case of 2 unit width pixels, or perform thin line processing only on a part of the boundary point that has 01 pixels on the top, bottom, left, and right (subcycle). Measures such as 2-unit width can be prevented from disappearing. In the sequential algorithm, it is easy to set conditions to prevent this disappearance, but on the other hand, it has the disadvantage that the position of the center line tends to depend on the scanning direction.

The thinning algorithm is desired to have the following properties for the purpose of extracting a line figure that preserves the wooden structure of the figure.

■Must be unit line width.

■The position of the line should be approximately at the center of the original shape.

■Thin lines are created while preserving the connectivity of shapes.

■The core wire should not shrink more than necessary during the thinning process.

■Do not cause whiskers due to small irregularities on the boundary.

■Even if the original figure is rotated and input, the shape of the core line should not change significantly.

■The core lines should not be distorted at the intersections of shapes.

Such line thinning processing requires a large amount of time even on large computers, and there has been a strong demand for a simple image processing method.

(Object of the invention) This invention was made under the above circumstances,
The purpose of this invention is to acquire numerical data for CAD at low cost by taking into consideration the advantages and disadvantages of both hand digitizing and automatic digitizing, which acquire numerical data of images from existing drawings, etc., and complement each other. An object of the present invention is to provide a method for converting image data into chain data in an automatic drawing reading system that can be used.

(Summary of the Invention) The present invention relates to a drawing data conversion processing method, in which image data obtained by reading an image of a design drawing etc. with an image scanner is converted into chain data, and then linear approximation 1 is performed. By converting the data into core lines, performing linear approximation 2, and combining the core edges, vector data that can be used for CAD, etc. is obtained. In the drawing data conversion processing method, when converting the image data to chain data, only one block of data is read from the binary data file, and after determining whether the data is complete, the spliced loop data is created between blocks. If the spliced loop data exists, it traces the connected chain data and stores it in a loop data file and a chain data file, and if the spliced loop data is not found or the tracing is completed. In this case, the search for the starting point of new loop data within the current block is performed until the end of the block, and when the block ends, the chain data is traced and whether or not it is closed within the block. It is designed to determine whether or not the block connection section has been reached, and to repeat the creation of connection data until the end of the block.

(Embodiment of the Invention) This invention solves the problem that when digitalizing analog drawings by hunt digitizing, the biggest problem is that it takes a very long time to measure each point by a human. The purpose is to solve the problems of human forgetfulness or error, and the difficulty of separating line drawings and character symbols by automatic digitizing. In this invention, the above part is performed at high speed by measuring image data using an image scanner, thereby improving the efficiency of measurement time. However, since this measurement data is image data, it is necessary to convert it into vector data. General image vector conversion is a process called line thinning as mentioned above, which is a method of removing pixels on both sides of a line and leaving only the center pixel, but due to the amount of data involved, it is extremely difficult to do so. The disadvantage is that it takes time. In the present invention, without performing line thinning processing in vector conversion processing, only pixels on both sides of the line are extracted, and after linear approximation 1 is obtained from this with a certain tolerance, the center is calculated. , by obtaining this as a core vector,
Achieves high-speed vectorization.

First, the terms used to explain this invention will be defined.

■Image data: Image data. Generally refers to shading image data. Image data is made up of dot data one by one. In this system, this specifically refers to binary data of black and white.

■Chain data: Data that saves memory when recording outlines using image data.

From one image to the next, one of the eight directions is always displayed, and these are expressed as a chain.

■Line approximation l...Chain data Char
esM.

Linear approximation using Williams' method.

■ Core line conversion: A method of converting image data to vector data. Calculate the center line (core line) from both sides of the line.

■Line approximation 2: A straight line approximation method that is processed to make straight and curved parts more clear after converting them to core lines.

■Connection loop data: Data created when chain data is extracted from image data, and loops spanning between blocks are managed as part loops.

■Connected chain data...Chain data divided between blocks.

■Loop data file: Loop information file when divided loops are combined into one loop after linear approximation.

02 double vector data: Contour data after linear approximation 1.

[Phase] 0ELT-ALO...The angle of vectors intersecting each other is 3 degrees.

■DELT-AL4...The angle of vectors that intersect with each other is lOo.

@Line relationship matrix table 1.2...A matrix table consisting of line relationships such as parallel and orthogonal lines and line distances.

Zero force rent node point: End point of the vector currently being processed.

[Phase] Search node point: End point of the vector searched centering on the current node point.

■2nd round MSI2E...Search area for searching centered on the current note point.

FIG. 1 shows the overall configuration of a drawing reading system to which the present invention is applied. An image scanner 3 for reading analog drawings, etc. is connected to a personal computer 1 via an interface 2, and an additional RAM is connected to the personal computer 1. Boat 4. A 20MB fixed disk 5 and a tablet 6 are connected. Image reading of image scanner 3 and R
The correspondence relationship with the memory area of AM port 4 is shown in Figure 2, and the decoding of image scanner 3 is EN.
It uses the A signal to determine whether data is currently being transmitted, cuts the 8-bit parallel data into 1-byte data at clock timing, distributes the data in the memory area, and stores it in a binary data file. It looks like this. The interface 2 sequentially calculates 4M addresses for storing data in the RAM board 4, and generates a DMA transfer signal for high-speed storage.

Figure 3 shows the overall flow of the drawing reading system.
This system measures only binary data of image data. When converting image data into vector data, a general problem is that it takes time to thin lines due to the huge amount of data. This invention takes the most time! Instead of converting the Il line data to 2 (1M), only the peripheral area of each line is converted into chain data.This chain data is still connected pixel by pixel, so the amount of data is large, and it is difficult to handle it. is inconvenient.Therefore, we approximate this by a straight line1 with a certain tolerance,
Edit as double vector data. As a result, the data amount becomes approximately 1740 times larger than the initial image data amount. From the double vector data obtained in this manner, the center vectors of both adjacent vectors, that is, the core line vectors are calculated, and the line widths thereof are calculated. In this case, the search for mutually adjacent vectors must be performed quickly.

This system uses BD) to quickly search these vectors.
- Its high speed is achieved by using Lee's data structure. The calculated core line vectors are constructed as continuous vectors up to each branching part, but due to blurring of the lines in the original drawing, combinations of lines, or divisions of lines by characters, etc., they do not necessarily constitute line segments that constitute a figure. incomplete. The joining of these line segments is performed through three steps.

One is linear approximation of a line segment. The purpose of this is, one, to reduce the amount of data, and the other, to clarify the direction of the vector at the branch part. In the next step, points within each range are combined using a range search focusing on these branching parts. However, at this stage, there are still many unconnected parts, so from the information about the inner loop obtained in the double vectorization stage, the core line vectors forming around that loop are searched by range search. At the same time, the divided portion is supplemented and combined considering the double vector. In this process, the core line vector is edited as a closed figure, and in the case of a figure composed of several closed figures, the data of the closed figure on the outer periphery is also extracted at the same time. Through the above processing, the identified closed figure data, skeleton vectors as line data, and double vector data files are finally transmitted from the first image data to the next interactive editing subsystem. It should be noted that the BD tree is binary data obtained by dividing the area into left, right, top and bottom, as shown in FIGS. 4(A) and 4(B), and is suitable for expressing the area. For details, see the Journal of the Institute of Electronics and Communication Engineers '83/l.

Vol. J66-D No. IO (7), pages 1193 to 1199.

Next, a conversion algorithm of the present invention for converting image data obtained by reading an analog drawing using the image scanner 3 into chain data will be described with reference to FIG.

One block of data is read from the binary data file into the memory in step St. At this time, it is determined in step S2 whether the binary data file is finished, and if it is in the middle, the process goes to the next step, and if it is finished, the program ends. In the next step S3, the presence or absence of joint loop data is determined between blocks, and if there is a loop to be connected, the connected chain data is traced in step SlO, and the data is output to the loop data file and chain data file, respectively. do. This process is performed for all connection loop data. If there is no connection loop data, or if the connection loop processing has ended, the process advances to the next step S4, and a search is made for the starting point of new loop data within the current block. During the search, it is determined in step S5 whether the block has ended or not, and if the block has ended, the process returns to step S1 to read the next block data.

If a loop start point is found in any other case, chain data tracking begins in step S6. Here, in step S7 it is determined whether or not the loop is closed within the block, and in step S8 it is determined whether the loop is closed or not! ! 11 cut h<thickness Return to step ψJ again at 7 and when the first block is closed, return to step S4, the step of searching for start chain data. If the block connection part is reached in step S8, connection data is created in step S9 and the process returns to step S4 again.

That is, the image scanner 3 is A as shown in FIG.
Measure the original drawing for the 3rd edition. In this case, since the A3 size is measured at 16 tots/mm, the amount of data is about 4 MB. This greatly exceeds the computer's storage capacity, so it cannot be processed by the computer. Therefore, in this invention, data is divided into blocks (#1.#
2) and process. The purpose of the processing is to extract the outline of a figure, and data must be detected with the outer outline lO clockwise and the inner outline ll counterclockwise. For this reason, the joining process between blocks is performed as follows.

As shown in FIG. 8(A), the part loop 1 is detected as clockwise from the starting point St, and is converted into chain data up to the block joint B4. Next, chain data is converted counterclockwise from the starting point Sl to part loop number 999 in the opposite direction to 81. Furthermore, in the inner loop, part loop l is turned counterclockwise from the starting point S2 to chain data up to B2, and from the starting point S2, it is turned clockwise in the opposite direction.
Chain data is converted to B3 as part loop number 999. Such processing is performed on all data within one block. In this case, the direction of the first part loop in one loop becomes the true direction of that loop, and always 1
start from. In the second block, as shown in FIG. 8(B), first the block joint B1. It is processed sequentially starting from B2.

In this case, at joint B, the loop number and part loop number of the previous block are saved and the bridge is inherited, so
The number is decreased when it is 500 or more, and it is increased when it is 1 or more.

Through the above processing, the loop number, part loop number,
Chain data is output. This data is sorted in the order of loop number and parts loop number, and rearranged as one loop data. Chain data example 9th
As shown in the figure. That is, in conventional data processing, since the amount of data is enormous, processing by hardware is expensive, and only contour data is extracted, with no regularity in the arrangement of the data. On the other hand, this system processes data in blocks, so it can be processed by software, and depending on the size of the blocks, even huge amounts of data can be processed. In addition, not only contour line data, but also regular data such as counterclockwise loop data on the outside and clockwise loop data on the inside is obtained, and since all data is extracted as a closed loop, the maximum area of the figure can be obtained immediately. Can be done.

The chain data obtained as described above is, for example, “
(:OMPIITERGRAP) 目C58ND IMA
GE PROCESSING” 8, pages 286 to 293. The data linearized in this way is
10Vo1. J56-D No. 10” P1193~
It is made into a core wire by the method described in P1200. The core and tJ data are then edited according to the flow shown in FIG. Here, the number of lines (l 1ne-pai
nt). That is, there is a case where the tree has two lines and a case where the number of lines is two or more. Further, this processing flow is recursive, and even if the number of lines is two or more and becomes two trees during the processing process, the same processing is performed thereafter.

Skeleton vectors are read from the skeleton vector file in step S20 and stored in memory.

Next, in step 521, a connected vector from end point to end point of the core vector is searched. At this time, step S
If all the searches have been completed in step S22, the core line vector subjected to the linear approximation 2 is outputted again to the file, and the program ends. If it is in the middle, the process moves to step S23 and the continuous vector is constructed. If there are two points that diverge depending on the number of constituent points, no further processing is performed,
The process returns to step 521 again. If there are 3 points, the process moves to step S24, where the mutual angle of the two vectors and the vector length are calculated, and in step S25, the process returns to step 5 as unprocessed.
It is determined whether to jump 21 to 21, move to step 526 to perform linear approximation 2, and then jump to step 521, or return to step 52+ after performing the branch point processing in step S33. Roughly, 4
If the number of vectors is more than a point, in step S30, the relationship with the next vector is read in advance and linear approximation 2 processing is performed, and step 531 is continued until all vectors in one continuous vector are completed. Next, the process proceeds to step 532, where the number of constituent points after linear approximation 2 is determined, and if it has decreased to 3 points, the process proceeds to step S24, and if not, the process proceeds to step S21 again.

As shown in Figures (A) and (B), the parameter is i+, which is the hector length of dsl and ds2 that sandwich the processing point (xs, ys), and (xp, yp) and (xn, yn).
Processing is performed using these five parameters: the perpendicular length from xs, ys), the angle α, and the total angle (α+β) with the angle β of the next vector. Also, the number of lines is 2
In the case of trees, when calculating core line vectors, short line whisker vectors often occur at some corners of thick lines, as shown in FIG. This process is mainly intended for such cases. Furthermore, when the number of lines is two or more, the vector group constitutes not only a straight line part but also a curved part. It is necessary to find the point where the line changes from a straight line to a curved line and keep the straight line as straight as possible while leaving the curved line as it is. Therefore, in this process, first the angle of α is DEL
Anything below T-ALO is deleted unconditionally, and if DELT-ALO<α<-DELT-Al1, processing is performed that takes into consideration the total angle with the vector length and primary angle β. . Anything else will not be processed. There are also some that still have circular corners. This kind of shape is the 13th tree
Since it should be as shown in the figure, the intersection point of the straight lines on both sides of the small side is found to generate a breaking point.

Linear approximation 2 as described above has the following features compared to the conventional method. That is, in the conventional linear approximation, there is no distinction between a straight line portion and a curved portion, so the approximation is insufficient, and the amount of data is not reduced much.

Furthermore, it is not possible to process bent portions at intersections, orthogonal connections cannot be made at intersections, and straight portions and curved portions cannot be clearly distinguished. On the other hand, according to the second straight line approximation method of the present invention, the straight line portion and the curved portion are processed separately, so the amount of data for the straight line portion can be significantly reduced, and the curved portion at the intersection is also processed. Intersections can also be joined orthogonally,
This has the advantage that straight portions and curved portions are clearly defined.

The linearly approximated core wires are combined according to the flow shown in FIG.

In step 540, a BD tree data structure is constructed from the skeleton vector data, and in step 541, a BD tree data structure is constructed at both end points of continuous vectors composed of the respective skeleton vectors. Step S
At 42, a range search is performed using this end point (node point).

At this time, if all node points have been completed in step 543, the program jumps to step 548 to create a skeleton connection data file and terminate the program. If there is an unprocessed node point, a range search is performed centering on that node point in step S44, and it is determined whether there are other note points. If not, step 550 jumps to execute a range search using already processed note points, and step 551 creates line relationship matrix Table 1. Then, in step 552, the current note point and the search node point are combined based on this matrix table 1. On the other hand, if there is a node point within the range, the process advances to step S45 and a line relationship matrix table 2 is created. Furthermore, based on this matrix table 2, each vector group is classified into groups in step 54B, then in step 547 note points are connected for each group, and the process returns to step 542 to perform the same processing.

By the way, the core wire data obtained through the processing up to now has the following properties.

■The branch part is separated into vectors.

■Short vectors are missing.

■Some data such as characters and symbols remain.

■There are some small-area figures other than letters and symbols that are missing.

■Data is simply a collection of vectors and is not recognized as a figure.

The purpose of this system is not just raster-vector conversion, but to edit and obtain vectorized data as meaningful figures. Therefore, in addition to merging branch parts, obtaining a closed figure or line figure while inserting missing vectors is one of the important processes. This system performs two steps for these processes. One is range coupling, and the other is loop coupling, which will be described next. Below, range coupling will be described first.

A branch part is a part where vectors of several independent figures are concentrated, and includes various elements such as remaining single vectors such as characters and missing vectors. When these are combined using a simple algorithm, erroneous connections occur and incorrect figures are constructed. Therefore, the uncombined parts must be reduced while pursuing completeness.

Range joins group line segments consisting of branch points (nodes) that fall within a certain range, and determine the join pattern for each group based on angles such as parallelism, orthogonality, and distance parameters of each line. do. The range is the area consisting of a square area with side 2 MSTZE centered on the node point, points in front of this point, and the minimum and maximum areas of the same square centered on points left and right away from that point. ,
As shown in FIG. 15, this is a range search area. This range will include multiple nodes, but the relationship between the straight lines formed by these nodes is parallel, orthogonal, or acute angle, as shown in Figure 16. If 160''≦θ≦200°, the opposite is true. Parallel, same parallel if O≦θ≦20°, 340m≦θ
If ≦360°, they are parallel; if 70°≦θ≦110°, they are perpendicular; if 250°≦θ≦290°, they are perpendicular;
In all other cases, the angle is determined as an acute angle. For example, Figure 17 (
In A), the relationship between each line is 0 for lines a to d, 1 for parallel lines in the opposite direction. If you create a matrix with 2 for parallel in the same direction, 3 for perpendicular, 4 for acute angle in the opposite direction, and 5 for acute angle in the same direction, the same figure (
C). Grouping is performed based on this matrix table. The group conditions are as follows.

■When there are lines in the same direction.

■If there is a similar line.

■If none of the above applies, the number of groups will be 1.

At this time, the number of groups is the number obtained by adding 1 to the number in the row where each of the above conditions (2) and (2) is first found. With the number of groups thus determined, which node is closest to the next parent note is determined based on the distance between the lines. In the example shown in FIGS. 17(A) to (C), (a, b) and (c, e, d) are independent groups. Grouped nodes are combined according to their respective patterns. With the current processing, it is possible to combine up to 2, 3, or 4 points within a group.

The two-point connection is processed as follows.

■For parallel lines, combine the end point coordinates of each note.

■For orthogonal lines, calculate their intersection points and combine them.

■If it is an acute angle line, join it to the parent note.

The three-point connection is processed as follows.

■In the case of parallel lines and perpendicular lines, the point of intersection between the line connected to the terminal p- of the parallel line and the perpendicular line is the joining point.

■In the case of orthogonal lines and acute angle lines, calculate the intersection of the orthogonal lines and use this as the connection point.

■For parallel lines and acute angle lines, do the same as ■.

■If all the lines are acute angles, calculate the midpoint between them and use it as the joining point.

The four-point connection is processed as follows.

(2) In the case of parallel lines and orthogonal lines, calculate the intersection points of the orthogonal lines with - with respect to the parallel lines, use these as connection points, and generate a new vector between them.

The nodes that are connected at one end are then excluded from range search targets. However, if you do this, some parts will remain unprocessed depending on the order of note searches and their ranges. Therefore, if the first range search does not find any results, the search is performed again, including existing connection points. If there were a case like the example in FIG. 18 in this case, the line would be divided and a new vector would be generated.

Although the nodes are connected through the above processing, the following cases are not processed.

■When they are parallel to each other in the same direction.

■When the distance between lines is greater than a certain amount.

■If the length of the car vector is smaller than the allowable value.

Furthermore, the limitation of this processing is that it is not possible to handle missing vectors since it is mainly processed on notes. This process is performed in the next loop connection.

Loop join connects endpoints that could not be joined using range join.
Processing is based on the rule that all core vectors with inner loops constitute a closed figure. As shown in FIG. 19, end points are first searched and end point A is found. Then, when a range search is performed around the endpoint A and a vector forming an inner loop is found, it is determined that this endpoint should form a closed figure, but has not yet been connected. Next, centering around this inner loop, a range search is performed for all core line vectors around the inner loop. Then tell them the number of endpoints. At this time, if the number of endpoints is an even number, which endpoint is to be connected is determined based on the distance between the endpoints, and the endpoints are connected.

Furthermore, if the number of end points is an odd number, it is determined which end points the vectors have are unnecessary. This determination is made by performing a range search around this end point and determining whether or not the first inner group is searched.

At the same time as the above is combined, the skeleton vectors are detected clockwise and the data order is adjusted as closed figure data. Also, at this time, closed figures on the outer periphery are also detected by classifying whether the same vector has been created twice or once. Before the core wires are joined, the result is as shown in FIG. 20 (A), for example, and after the core wires are joined as described above, the result is shown in FIG. 20 (B). Conventional core wire connections connect end points that fall within a certain range, resulting in incorrect connections when complex figures are involved. Also,
Because only range combinations were performed, uncombined parts remained, and parts separated by characters, symbols, etc. could not be processed. In contrast, in this invention, each line is grouped by each end point, so even complex figures can be connected more accurately.
Since loop connections are added to range connections, more lines can be connected, resulting in a higher degree of completion. Another advantage is that lines separated by characters or symbols can be combined.

(Effects of the Invention) As described above, in the process of converting a huge amount of data read by an image scanner into vectorized data, the present invention can perform thinning processing by sequentially removing pixels on both sides of a line, which was done in the conventional thinning process. Instead of extracting only the central pixel, we extract only the pixels on both sides of the line and obtain vector data by linearly approximating them with a certain tolerance, so even with a huge amount of data, the software is easy to use. Now it can be processed quickly. In addition, since this data is processed by dividing it into blocks, the size of the blocks makes it possible to process even huge amounts of data. Furthermore, rather than just looking at the contour line data, regular data can be obtained, with the outside as loop data around the left and the inside as loop data around the right. Even if the input figure is incomplete due to the combination of lines or the division of lines by characters or the like, it is possible to obtain effective information for determining the connection of stiff line segments.

Furthermore, since all data is extracted as a closed loop, the maximum area of the figure can be obtained immediately, and it is also possible to separate character data of a certain size.

[Brief explanation of the drawing]

Figure 1 is a connection diagram showing the overall configuration of a drawing reading system to which this invention can be applied, and Figure 2 is an image scanner and RAM.
3 is a diagram showing the data processing flow of the drawing reading system, FIGS. 4 (A) and (B) are diagrams for explaining B[]l-Lee, and The figure is a flowchart showing an example of conversion of image data into chain data according to the present invention, Figures 6 to 9 are diagrams for explaining data blocking, and Figure 10 is a flowchart showing an example of converting image data to chain data according to the present invention. Flowchart showing an example, No. 11
Figures (8) and (B) to Figure 13 are diagrams for explaining core wire formation, Figure 14 is a flowchart showing a processing example of core wire combination, and Figures 15 to 20 (A) and (B) are It is a figure for explaining core wire connection. DESCRIPTION OF SYMBOLS 1...Personal computer, 2...Interface, 3...Image scanner, 4...IIAM board, 5...Fixed disk, 6...Tablet. Applicant's agent Yoshikata Yu Sancha l Reference standard 2 Figure (A) CB) Jin 4
Fig. 6 Fig. 6 Fig. 7 Fig. vine 5 Fig. f, 4) (β) Jin 6 Fig. $IO Fig. (B
) $ 77 Figure 16 Figure 13

Claims (1)

[Claims] The image data obtained by reading the image with an image scanner is converted into chain data and then subjected to linear approximation 1. This linear approximation 1 data is converted to a core line and then linear approximation 2 is applied.
In a drawing data conversion processing method in which vector data is obtained by connecting core lines, when converting the image data to chain data, only one block of data is read from the binary data file, and whether the data is complete or not. After determining whether or not there is, it is determined whether or not there is joint loop data between the blocks, and if there is the joint loop data, tracing the connection chain data and storing it in a loop data file and a chain data file, If the spliced loop data does not exist or if the tracking ends, a search for the starting point of new loop data within the current block is performed until the end of the block, and if the block ends, the chain data is traced. At the same time, it is determined whether the block is closed or not and whether the block connection part is reached or not, and the creation of connection data is repeated until the end of the block. Conversion processing method.