JPH0434654A - Drawing reader - Google Patents

Abstract

PURPOSE:To make it possible to accurately recognize also a segment or character arranged in the vicinity of a boundary between plural adjacent drawings divided due to restriction such as paper size by advancing the recognizing processing of a graphic/character in a drawing to be processed by utilizing the processed result of a its adjacent processed drawing. CONSTITUTION:This drawing reader is constituted of an image input device 1, an image memory 2, a processor 3, and a storing means 4. divided graphics/ characters are joined by using the information (image or a processed result) of adjacent drawings indicated by a drawing discriminator found out from a processing drawing image and the recognition processing or the like of a graphic/character arranged in the vicinity of the jointed picture is advanced. Thereby, a correct recognized result can easily be obtained also from divided graphics/characters (character string).

Description

[Detailed Description of the Invention] [Objective of the Invention] (Industrial Field of Application) The present invention provides a method for processing drawings such as maps that are divided into multiple pages (for example, multiple sheets) and created individually. The present invention relates to a drawing reading device capable of efficiently reading and inputting data while maintaining relationships with adjacent drawings.

(Prior Art) Recently, various attempts have been made to develop drawing reading devices that recognize and process maps and facility information superimposed on maps, convert various types of information into numerical data, and read and input the data.

However, this type of map, or a drawing in which equipment information is superimposed on a map, has a huge amount of information and there are paper size limitations, so it is necessary to write it over multiple pages (multiple sheets). It is often created in pieces. However, in maps and drawings that are divided and created individually in this way, figures and characters may be written across multiple sheets of paper. For example, if a figure or character (character string) hangs over a seam between two facing pages, the figure or character (character string) is divided at the seam and written on each page.

However, in the conventional drawing reading apparatus, which is based on a boat, the above-mentioned plurality of maps and drawings are merely input as image data on a paper (page) basis, and each of these is processed separately. For this reason, it has been extremely difficult to recognize figures divided by so-called map boundaries, which indicate the joints between pages in a plurality of divided maps and drawings.

In addition, even if you check the processing results for maps and drawings processed in this way and correct errors, only the processing results of the maps and drawings that are being processed are displayed.
There was a problem in that it was not possible to confirm the relationship with other divided drawings, for example, adjacent drawings, and it was difficult to perform confirmation and correction work.

(Problems to be Solved by the Invention) As described above, with conventional drawing reading devices, when manually reading information on maps and drawings that are divided into multiple pages,
There is a problem in that it is extremely difficult to accurately recognize figures that are divided by figure boundaries, and furthermore, it is impossible to confirm or correct the figure while checking its relationship with adjacent drawings.

The present invention has been made in consideration of these circumstances, and its purpose is to improve the recognition performance for figures and characters near the outline of maps and drawings that are created in sections due to restrictions such as paper size. It is an object of the present invention to provide a drawing reading device that can improve the processing results and efficiently proceed with checking and correcting the processing results while checking the relationship with adjacent drawings.

[Structure of the Invention] (Means for Solving the Problems) A drawing reading device according to the present invention obtains a drawing identifier adjacent to a drawing from an image of a drawing being processed, and processes an adjacent drawing indicated by this drawing identifier. If the drawing image has already been processed, the stored processing results of the adjacent drawing are used to proceed with the processing of the drawing image being processed.

In addition, when proceeding with the processing of the drawing image being processed using the processing results of the adjacent drawing, the line segments and characters near the outline of the drawing image being processed are extracted, and these line segments and characters are combined with the processing results of the adjacent drawing. At the same time as connecting the line segments and characters indicated by The feature is that it is given as an attribute.

Furthermore, the present invention is characterized in that by connecting the adjacent drawing indicated by the drawing identifier and the drawing image being processed, when processing the image near the drawing area, it is processed together with the image of the connected adjacent drawing. .

Furthermore, the present invention is characterized in that when processing results are displayed for confirmation and correction, images of adjacent drawings indicated by drawing identifiers or processing results of the adjacent drawings are displayed at the same time. ) According to the present invention configured in this way, for example, separated figures and characters can be joined together using information (images and processing results) of adjacent drawings indicated by the drawing identifier obtained from the drawing image being processed. After that, it is possible to proceed with the recognition process for figures and characters near the figure enclosure, so it is possible to easily obtain correct recognition results even for divided figures and characters (character strings). , it becomes possible to improve its recognition performance. In addition, even when displaying the processing results of a drawing image that is being processed for confirmation or correction, information on adjacent drawings indicated by the drawing identifier is also displayed, so you can understand the relationship between these drawings and It becomes possible to proceed with confirmation and correction work easily.

(Embodiment) Hereinafter, an embodiment of the drawing reading device according to the present invention will be described with reference to the drawings.

FIG. 1 is a block diagram showing a schematic configuration of an embodiment device. This device basically consists of an image input unit 1 that inputs drawings to be processed such as maps and drawings as binary image data, an image memory 2 that stores and stores the image data input from this image input unit 1, and an image memory 2 that stores the image data input from this image input unit 1. A processor 3 executes predetermined processing on the drawing image stored in the memory 2, such as recognizing figures and characters written in the drawing, and stores the processing results of the drawing image by the processor 3. It is configured to include storage means 4.

The image input unit l is composed of, for example, a scanner equipped with a line image sensor, and optically scans the drawing information, binarizes the image data, imports it into the device, and stores it in the image memory 2. Store. This image input section 1
This may be a television camera or a drum scanner. Alternatively, drawing image data that has been scanned in advance and converted into an image file may be read.

The drawing image data input in this way is managed, for example, by the page number of the drawing, and stored in the image memory 2. The processor 3 then specifies a drawing to be processed according to instructions from an operation unit such as a keyboard (not shown), performs line segment extraction processing on the drawing image data, and performs figure/character recognition processing based on the extraction results. A series of drawing data processing is executed, and the processing results are stored in the storage means 4. This storage means 4 is constituted by, for example, a secondary storage device such as a disk. Further, the processing results stored in the storage means 4 are displayed on a display (not shown) under the control of the processor 3, and are used for confirmation and correction processing.

2 and 3 are diagrams showing the flow of characteristic drawing processing in the processor 3 of the embodiment apparatus configured as described above, and FIG. 2 also shows the flow of processing during drawing recognition processing. Figure ii3 shows the flow of processing when checking and correcting the recognition processing results.

First, the processor 3 processes the graphic image for the drawing image.
To explain the character recognition process, as shown in FIG. 2, the processor 3 reads the specified drawing number! The drawing image data of is obtained from the image memory 2 (step a). Then, according to the specified drawing number l, it is determined whether or not, for example, drawing J adjacent to drawing I has already been recognized (step b). Specifically, when a drawing number I is specified, the processor 3 stores a drawing image such as the one shown in FIG. 4 in the image memory 3! is read, and it is determined whether or not a drawing adjacent to this drawing, for example, drawing J shown in FIG. 5 adjacent to the left side, has already been processed.

If the adjacent drawing J on the left side has already been processed, the recognition results for that drawing J are stored in the storage means 4, so the processing results for this drawing J can be used to create figures and shapes for drawing I. Starts character recognition processing.

This recognition process begins by extracting line segments and characters from the drawing image data of the drawing to be processed (Step C). This line segment extraction process may be performed by thinning the drawing image data of the drawing to be processed and extracting the line segments that make up the drawing from the thin lines, or by approximating the outline of the drawing image with polygonal lines. , line segments of various figures constituting the drawing may be extracted from the approximate line segments.

After that, the processor 3 processes the processing target drawing I shown in FIG.
The line segment data i1. i2. i3. i4...
and line segment data jl, j stored as processing results in the storage means 4 regarding the already processed adjacent drawing J.
2. j3. The correspondence with j4... is processed (step d). This process uses, for example, the position information on the above-mentioned figure frame of a line segment that intersects with the adjacent figure frame (left side adjacent figure frame) of the drawing to be processed This is done based on the location information on the map outline. Specifically, for example, in each of drawings I and J, line segment data i1. i2. i3. i4..., ,ul,
+2. , +a, j4, and so on, and according to the sorting results, these line segment data are correlated with each other as shown in FIG.

Each of the above-mentioned drawings 1. By matching the line segment data of J, these line segments are logically joined.

After that, the attributes of each line segment of the target drawing engineer are determined according to the attribute information regarding the line segment data obtained for the processed adjacent drawing J, and the figures and characters of the target drawing I are changed according to this attribute information. Recognition processing is performed (step e).

To determine the attributes for this line segment data, the attribute information for each line segment data of the adjacent drawing J required in the storage means 4 is stored in the aforementioned line segment data as shown in FIGS. 7(a) and 7(b). This is done by giving the same attribute to the corresponding line segment data of the drawing I to be processed according to the correspondence relationship. In this manner, the attributes of each line segment data in the drawing I to be processed are determined according to the processing results of the adjacent drawings that have been processed in advance, and then each line segment data is subjected to recognition processing.

Similarly, the character information P obtained from the drawing to be processed (■) is associated with the character information Q obtained from the adjacent drawing J, as shown in FIGS. 8(a) and 8(b). As a result, the character strings divided between adjacent drawings can be integrated as shown in FIG. 9, and these characters can be recognized as one coherent character string.

By actively utilizing the processing results obtained for the processed adjacent drawings as described above to proceed with the recognition process for the figures and characters in the target drawing, each figure and character can be recognized with high accuracy. becomes possible. Furthermore, it is possible to accurately recognize line segments and characters in the vicinity of drawing boundaries between a plurality of adjacent drawings that are separated due to restrictions such as paper size.

Incidentally, if the processing for the adjacent drawing has not been completed, the adjacent drawing J and the drawing object to be processed are joined to each other in step f), and the entire joined image is treated as the processing target image! Recognition processing is performed on the figures and characters inside (step g).

For example, if a drawing adjacent to the right side of the processing target drawing l shown in FIG. 4 is given as shown in FIG. The drawing to be processed 1 and the adjacent drawing on the right side thereof are arranged in such a manner that the drawing to be processed 1 and the drawing adjacent to it on the right side are arranged.

This joining of drawings is achieved by connecting the images near the right side of the drawing to be processed at the image data level based on the image near the right side of the drawing and the image of the left side of the adjacent drawing on the right side. be done.

This kind of image stitching at the image data level is
For example, a black pixel on a diagram border is used as a reference point, and the correspondence between the reference points between adjacent images is determined from the position information. Next, it is realized by applying a predetermined geometric transformation (for example, affine transformation) to each image data based on the result of the correspondence between the reference points.

Line segments are extracted from the drawing images connected in this way, and recognition processing is performed on the extracted line segments. Specifically, when extracting a figure representing a house in the image to be processed shown in Fig. 4, first, as shown in Fig. 11, among the line segments found from the connected images, one that forms a closed loop is extracted. A figure is found and determined as a house figure.

In this way, even if the house figure is cut by the map and does not form a closed loop, it is possible to reliably extract the house figure as a closed loop figure by connecting it with the adjacent drawings mentioned above. becomes. As a result, it becomes possible to reliably recognize the figure without being affected by the division of the figure.

By the way, when displaying the processing results (recognition results for figures and characters) on the drawing I to be processed as described above for confirmation and correction, the displayed image is controlled as shown in FIG. You should do it.

That is, when displaying the processing result of the drawing I to be processed on the display (step A), it is checked whether the drawing J adjacent to the drawing I has been processed (step B). If the adjacent drawing J has been processed, the recognition result of the adjacent drawing J is displayed at the same time when displaying the graphic part of the drawing I to be processed.

In addition, if the adjacent drawing J has not been processed (unprocessed), the image data of the adjacent drawing J is associated with the drawing to be processed and displayed on the display as is (
Step D). Specifically, by displaying the image data of adjacent images near the outline of the drawing as is, it is possible to grasp the relationship with the adjacent drawing.

In this way, when displaying the processing results of the drawing to be processed for confirmation and correction processing, by simultaneously displaying the processing results and image data of the adjacent drawings, the above processing can be done while grasping the relationship with the adjacent drawings. Since it becomes possible to proceed with the confirmation/correction process for the recognition result of the target drawing I, it becomes possible to facilitate the work and perform the confirmation/correction process with high reliability.

Note that the present invention is not limited to the embodiments described above. For example, for adjacent drawings, a drawing identifier that specifies which drawing to be processed should be placed above, below, left, or above the drawing to be processed may be determined by a keyboard instruction from the operator. Also, adjacent drawing information that is written as supplementary information near the drawing border,
For example, it may be possible to recognize a drawing identifier indicating which drawing the drawing is continuous with, and to automatically determine an adjacent drawing to be used for drawing processing based on this drawing identifier. Furthermore, if the reduction ratios, etc. of adjacent drawings are different, it goes without saying that the above-mentioned association of line segment data may be performed by matching the reduction ratios according to the relationships, etc. In addition, the present invention can be implemented with various modifications without departing from the gist thereof.

[Effects of the Invention] As explained above, according to the present invention, when performing processing such as figure/character recognition on a plurality of drawings divided due to restrictions such as paper size, etc., drawing recognition processing in particular around the outline of the drawings is performed. When performing a recognition process, the recognition process is performed using the processing results of adjacent drawings and their image data, so even figures that are separated by figure boundaries can be easily and correctly recognized, and the recognition results Even when confirming and correcting drawings, the process can be carried out while correctly grasping the relationship between adjacent drawings. As a result, it is possible to efficiently and accurately read and input information on drawings that are divided and created across a plurality of sheets, which provides great practical effects.

[Brief Description of the Drawings] Fig. 1 is a schematic configuration diagram of a drawing reading device according to an embodiment of the present invention, Fig. 2 is a diagram showing the flow of figure/character recognition processing procedure in the embodiment device, and Fig. 3 4 is a diagram showing an example of a drawing to be processed, and FIG. 5 is a diagram showing a processed adjacent drawing adjacent to the left of the drawing shown in FIG. 3. It is a figure which shows the example of a drawing. In addition, FIGS. 6 to 9 are diagrams for explaining the process of associating line segments and characters between the drawing to be processed and adjacent drawings, respectively.
Figure 0 is a diagram showing an example of an unprocessed drawing adjacent to the right side of the drawing shown in Figure 4, and Figure 11 is a diagram showing an example of a joined drawing between the drawing shown in Figure 4 and the drawing shown in Figure 10. be. ■...Image input device, 2...Image memory, 3...
Processor, 4... Storage means, b... Means for determining whether or not adjacent drawings have been processed, d... Means for connecting line segments between adjacent drawings, e... Processing results of processed drawings A means for recognizing graphic characters using... A means for joining adjacent drawings at the image data level.

Claims (1)

[Scope of Claims] (1) Means for inputting a drawing as an image, means for storing the input drawing image, means for performing predetermined processing on the stored drawing image, and processing results for the drawing image. means for storing a drawing image being processed, means for obtaining a drawing identifier adjacent to the drawing from the drawing image being processed, and means for checking whether or not the neighboring drawing indicated by the drawing identifier has been processed; Drawing reading characterized in that when an adjacent drawing indicated by a drawing identifier has been processed, the processing result of the adjacent drawing stored in the storage means is used to proceed with processing on the drawing image being processed. Device. (2) Drawing image processing involves extracting line segments and characters near the outline of the drawing image being processed, and connecting these line segments and characters with the line segments and characters shown in the processing results of adjacent drawings. Claim characterized in that, at the same time, the process proceeds while giving attributes of figures and characters in the adjacent drawing as attributes of line segments and characters in the drawing image that is being processed to be connected to line segments and characters in the adjacent drawing. The drawing reading device according to (1). (3) The drawing reading device according to claim (1), further comprising means for connecting an adjacent drawing indicated by a drawing identifier and a drawing image being processed, and an image near the drawing outline of the drawing image being processed. A drawing reading device characterized in that, when processing, images of the connected adjacent drawings are processed together. (4) The drawing reading device according to claim (1), further comprising means for simultaneously displaying the processing results of adjacent drawings that have already been processed, when displaying the processing results for the drawing image being processed for confirmation and correction. A drawing reading device characterized by: (5) The drawing reading device according to claim (1), further comprising means for simultaneously displaying images of adjacent drawings when displaying the processing results for the drawing image being processed for confirmation and correction. drawing reading device.