JPH08262974A - House recognizing method in map input system - Google Patents

Abstract

PURPOSE: To provide a house recognizing method in a map input system capable of reducing an operator's workload and also of quick processing. CONSTITUTION: When an operator specifies a frame on the screen of a CRT display 14 after having a scanner 14 read out a map its binarization data transformed into vectorization data, a graphic of a closed loop in shape larger than a specified size is extracted as a graphic of a house for a recognition object. The graphic of the house for the recognition object is processed in shaping and connecting and the processed result is displayed on the screen of the CRT display 14. The operator decides correct/error of it to see whether or not the results of the process show the actual house, and registers only necessary results.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a house recognition method in a map input system for converting binary image data obtained by inputting a map into vector data and then recognizing and coding house figures. .

[0002]

2. Description of the Related Art Conventionally, various map input systems have been proposed in which a map is read, each figure element is converted into vector data, and a house figure is recognized and coded. For example, in a conventional interactive map input system, first, a map is read by a scanner to obtain binary image data. Next, the operator uses the position input device to pick up the graphic elements one by one and convert them into vector data. And
The house figures are coded by recognizing that they are house figures by picking them one by one. After that, symbol recognition is performed on the symbols included in the map, and structuring processing such as attribute assignment to figures and association between figures is performed to build a structured database.

[0003]

However, in the conventional interactive map input system, when recognizing a house figure, the operator processes the figure by designating the house figures one by one. There is a problem that the work load is imposed and the processing cannot be performed quickly.

The present invention has been made based on the above circumstances, and an object of the present invention is to provide a house recognition method in a map input system which can reduce the work load on an operator and can perform processing quickly. It is what

[0005]

In order to achieve the above object, the invention according to claim 1 reads a map to be processed in advance, converts the obtained binary image data into vector data, and then, In the house recognition method in the map input system for recognizing a figure, when a frame is specified on the screen, a figure that is a closed loop shape and has a predetermined size or more in the frame is extracted as a house figure to be recognized,
It is characterized in that after the extracted house figure to be recognized is shaped into a house and displayed on the screen, whether or not to register the house figure to be recognized is instructed.

A house recognition method in a map input system according to a second aspect of the present invention is the method according to the first aspect of the invention, wherein after the house figure to be recognized is shaped into a house,
When the sides of the recognition target house figure and the already registered house figure are deviated from each other, connection processing is performed to match the deviated sides.

[0007]

According to the invention described in claim 1, when a frame is designated on the screen by the above configuration, a closed loop shape having a size larger than a predetermined size is extracted as a recognition target house figure and shaped. By performing processing and displaying it on the screen, the operator collectively specifies the house figures to be recognized in the frame, and then registers the house figures to be recognized based on the result displayed on the screen. Since it is only necessary to instruct whether or not it is possible to reduce the work load on the operator and to perform the processing quickly.
Further, since the house figures to be recognized can be designated collectively by the frame, the operability is improved as compared with the conventional interactive processing method.

According to the second aspect of the present invention, by performing the connection process after shaping the recognition target house figure into a house shape by the above configuration, the necessary correction is automatically performed on the recognition target house figure. You can

[0009]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a schematic configuration diagram of a map input system using a house recognition method according to an embodiment of the present invention.

The map input system shown in FIG. 1 includes a scanner 12 as an image input device, a CRT display device 14, and
A pointing device 16, a central processing unit 18, an element file 22, and a contour line file 24.
, A region file 26, a vector file 28, and a structured file 32. In this embodiment, for example, consider a case where a database of house maps is created.

The scanner 12 optically scans the drawing,
It is input as binary image data. The CRT display device 14 displays a result of drawing input, a result of house recognition processing described in detail later, and the like. Position input device 1
Reference numeral 6 is for inputting position coordinate data of a figure to be recognized, and a keyboard, a mouse or the like is used. In particular, in the case of performing the house recognition process, it is possible to specify a frame on the screen of the CRT display device 14 using the mouse and select the figures included in the frame all together.

Here, as the binary image data input from the scanner 12, raster data based on the run length format used in facsimiles is adopted. This raster data is, for example, the starting coordinates at the start of the pixel value change from “0” (white) to “1” (black) on the scan line, and the last coordinates at which the pixel value changes from “1” to “0”. Is vector data whose end point is.

The element file 22 stores the feature amount (element data) extracted for each graphic element after the raster data is labeled. Here, the center of gravity, area, circumscribed rectangle, moment, etc. are used as the element data. The contour line file 24 stores the contour line data obtained by extracting the edge of the raster data for each graphic element. The contour line is approximated to a polygon and is composed of a large number of line segments, and one contour line is a closed polygonal line. For this reason, the contour line data has, as vectorized data, the line segments that make up each contour line. The element data is linked to the outline data by a pointer.

The area file 26 stores area data indicating in which area of the entire drawing each line segment forming the outline is included. The area data is managed in a tree structure. In such data management using a tree structure, each branch node represents an area in the middle of division on the drawing, and a leaf stores a pointer to the contour line data included in the area represented by the branch node. Therefore, when a certain area is designated, the leaf corresponding to the designated area can be searched, and the line segment of the contour line included in the designated area can be determined based on the pointer there. As an example, a tree structure of area data for a substantially cross figure will be described. FIG. 2 (a) shows the contour line of the edge of raster data of this cross figure. Each line segment forming the contour line is managed for each circumscribed rectangle. Then, as shown in FIG. 2B, the entire drawing is divided into areas including one circumscribed rectangle for each line segment. This division process is represented as an area management tree as shown in FIG. 2C, and the leaves of the area management tree correspond to the smallest divided areas.

The central processing unit 18 executes predetermined processing based on an interactive drawing input program.
For example, based on the input binary image data (raster data), pre-processing labeling processing, feature amount extraction processing, contour line extraction processing, etc. are performed to create element data, contour line data, area data, In addition, the figure to be recognized is interactively skeletonized, and the raster data is vectorized. The vector file 28 stores vector data after vectorization processing. Further, after this vectorization process, a house recognition process for recognizing and coding the house figure is performed. At the time of this house recognition processing, shaping processing, connection processing, etc. can also be executed. The shaping process is that when a house figure is simply vectorized, one side of the house figure may have multiple polygonal lines, so this can be shaped into a single line, Since the corner may be chamfered, this is a process of shaping the chamfered part into the original corner. The connection process is a process of matching the displaced sides of the recognition target house figure and the already registered house figure when the sides that should originally overlap are displaced.

Further, the central processing unit 18 carries out a structuring process such as giving attributes to the figures to give their meanings and attributes, and associating the figures to define the semantic relationship. Structured data is stored in the structured file 32.
Used as D database. It should be noted that each of these processes is executed by designating a predetermined command from the menu displayed on the screen of the CRT display device 14 using the position input device 16.

Next, a processing procedure in the map input system of this embodiment will be described with reference to FIGS. FIG. 3 is a diagram for explaining the procedure of the skeletonization process, FIG. 4 is a diagram for specifically explaining the skeletonization process, and FIG. 5 is a diagram for explaining a house recognition method in the map input system of the present embodiment. 6 and FIG. 6 are diagrams for explaining the connection processing in the house recognition method in the map input system.

First, the drawing is read by the scanner 12 to obtain raster data (step 2). For example, in FIG.
It is assumed that raster data about four house figures as shown in (a) is obtained. After performing the labeling process based on this raster data, the central processing unit 18 extracts the feature amount for each labeled graphic element, creates the element data, and stores it in the element file 22. Further, the edge of the raster data is extracted for each graphic element to obtain contour line data, and this is stored in the contour line file 24 (step 4). After that, based on the contour line data,
Region data that manages information about the circumscribed rectangle of each line segment that constitutes the contour in a tree structure is created (step
6).

Next, the skeletonizing process is interactively performed on the figure to be recognized. Here, the skeletonization process is a process of obtaining a line connecting the centers of circles inscribed in two line segments. First, the operator uses the position input device 16 to perform a CRT.
A predetermined area on the screen of the display device 14 is picked and a pair of line segments to be recognized is designated (step 8). Here, as shown in FIG. 4A, it is assumed that two substantially parallel line segments S ₁ and S ₂ above the upper left house figure are designated. Then, as shown in FIG. 4B, the area management tree is searched to find a leaf corresponding to the selected predetermined area on the screen (step 12). Then, the contour line data corresponding to the line segments S ₁ and S ₂ is extracted from the contour line file 24 based on the pointer stored in the leaf (step 14). Next, FIG.
As shown in (c), the two extracted line segments S ₁ and S ₂ are subjected to a skeletonization process to generate a skeleton (step 16). After the area is designated in this way and the core line is generated from the two line segments, the two line segments are traced to continuously generate the core line. That is, since the contour lines themselves are connected to each other, they are directly connected from the contour line file 24 to the line segments S ₁ and S ₂ that have been cored in a predetermined direction, for example, to the left in FIG. 4A. The line segment of the contour line is automatically extracted, the core line is generated, and the polygonal line is generated (step 18). Then, for example, at the intersection or the corner of the house figure, the core line is not generated, and the core line conversion process is stopped. After that, when the operator designates the next pair of line segments to be recognized by the drawing input device 16, for example, when the two substantially parallel line segments S ₃ , S ₄ on the left side of the upper left house figure are designated, Skeletonization processing is performed. In this way, skeletonization processing is performed for all raster data. The vector data thus obtained by skeletonizing is stored in the vector file 28 (step 22).

Next, a house recognition process is performed on a specific house figure. Here, it is assumed that the command is selected so that connection processing is also performed during the house recognition processing. First, the operator uses the position input device 16 to display the CRT display device 14
Set a frame on the screen and specify the area within the frame (st
ep32). Here, in order to simplify the explanation, FIG.
As shown in (a), it is assumed that an area including only the upper right house figure H ₄ is designated. It is also assumed that the house figures H ₁ , H ₂ , H ₃ existing near the house figure H ₄ have already been recognized and registered. When the area is designated in step 32, a figure having a closed loop shape and having a predetermined size or more is extracted in the designated area (step 34). In general, a house figure is a simple closed figure such as a quadrangle, and thus the closed loop shape is regarded as a house figure to be recognized. Moreover, since the house figure is drawn with a certain size on the map, only the figure having a size larger than a predetermined size is extracted. As a result, a small figure such as a map symbol included in the frame can be excluded from the beginning, and the house figure to be recognized can be efficiently extracted. After that, the extracted house figure to be recognized is shaped into a house by correcting each side into a straight line or a chamfered part into a corner (step 36).

Next, the house figure to be recognized is connected to the already registered house figure (step 3).
8). First, for each side of the recognition target house figure H ₄ ,
Select the nearest registered house figure within a predetermined range. For example, the registered house figure H ₁ is selected for the side a ₄ of the house figure H _{4 to be} recognized. Then, it is checked whether or not there is a side of the registered house figure H ₁ that is substantially parallel to the side a ₄ of the house figure H _{4 to be} recognized and is a predetermined distance or less. Here, the predetermined distance is set to, for example, several pixels, that is, about 0.3 mm to 0.4 mm. Specifically, this distance is specified as a percentage of the line width, taking into consideration whether it is a thick line or a thin line. Actually, the two sides are 0.3mm-0.4m
Even if it is deviated by about m, there is not much difference as far as human beings can see it, but when the screen is zoomed, for example, a difference appears. This connection process is effective, for example, when calculating the area of land or when it is necessary to actually use the data as numerical values.

Then, the side a ₄ of the house figure H _{4 to be} recognized
On the other hand, if there is a side of the registered house figure H ₁ that is substantially parallel and is a predetermined distance or less, this is determined as the joining side. In this case, the side a ₁ of the registered house figure H ₁ is determined as the joining side. Then, as shown in FIG. 6B, points p ₄ and q ₄ at both ends of the side a ₄ of the house figure H _{4 to be} recognized are points p ₁ and q at both ends of the side a ₁ of the registered house figure H _1. By extending or contracting the side a ₄ of the recognition target house figure H ₄ so as to match ₁ , the recognition target house figure H _4
4 of the sides a ₄ joined to the joining edge. Thereafter, the other side b ₄ of the recognition target house graphic H _4, c _4, sequentially applies d _4,
By performing the same process, the side b ₄ of the recognition target house figure H ₄ is joined to the side b ₃ of the registered house figure H ₃ . On the other hand, the results for the side c _4, d ₄ of the recognition target house graphic H ₄ is not bonded with the sides of which registration house shape, the sides a _4, b ₄ of the recognition target house graphic H ₄ are joined, Stretch automatically,
Be moved. In this way, the recognition target figure H ₄ is obtained as shown in FIG.
As shown in (c), the boundaries come to coincide with the registered house figures H ₁ and H ₃ .

After performing the connection process for one house figure to be recognized, the process result is displayed on the screen of the CRT display device 14. Since the house recognition process is performed on the closed-loop figure, there is a possibility that even a figure that is not actually a house figure, such as a circular figure, may be processed. For this reason, the operator makes a correct / wrong judgment as to whether or not the processed one actually represents a house (step 42), and if it is a house, presses the "register" button displayed on the screen. . As a result, the house figure to be recognized is coded and registered in the vector file 28. On the other hand, if the processed item is not a house or there is no need to register, the operator does not register by pressing the "Cancel" button displayed on the screen. Then, if the area specified by the frame contains other recognition-target house figures, the shaping process (step 36) and the connection process (step 38) are similarly performed for these, and the result is displayed on the CRT. Since it is displayed on the screen of the device 14, the operator judges whether the processing result is right or wrong.

After that, the symbols included in the map are recognized, and the structured processing is performed interactively, and the obtained structured data is stored in the structured file 32. In the symbol recognition processing, the element data and the contour line data are required, but since the contour line data is linked to the element data, the contour line data is extracted from the element data.

In the house recognition method in the map input system of the present embodiment, when a frame is specified on the screen, a closed loop shape having a size equal to or larger than a predetermined size is extracted as a house pattern to be recognized in the frame, By performing shaping processing and connection processing and displaying on the screen, the operator collectively specifies the house figures to be recognized in the frame, and then the house figures to be recognized are displayed in the result displayed on the screen. Since it is only necessary to instruct whether to register based on the above, it is possible to reduce the work load on the operator and to perform the processing quickly. Further, since the house figures to be recognized can be designated collectively by the frame, the operability is improved as compared with the conventional interactive processing method. Further, after the recognition target house figure is shaped into a house shape, the connection processing is also performed, so that the necessary correction can be automatically performed on the recognition target house figure.

The present invention is not limited to the above embodiment, but various modifications can be made within the scope of the invention. For example, in the above embodiment, when the command for connection processing is not selected during the house recognition processing, the house figure to be recognized is shaped into the house shape, and the processing result is displayed on the screen. The operator will make a correctness decision.

[0027]

As described above, according to the first aspect of the present invention, when a frame is specified on the screen, a figure which is a closed loop shape and has a predetermined size or more in the frame is a house figure to be recognized. As a result, after the operator collectively specifies the house figures to be recognized in the frame by displaying them on the screen after performing the shaping process as a result, the house figures to be recognized are displayed on the screen as a result. Since it is only necessary to instruct whether or not to register based on the above, it is possible to provide a house recognition method in a map input system that can reduce the work load on the operator and can perform the processing quickly.

According to the second aspect of the present invention, after the house figure to be recognized is shaped into a house shape and then connection processing is performed, necessary correction can be automatically made to the house figure to be recognized. A house recognition method in a map input system can be provided.

[Brief description of drawings]

FIG. 1 is a schematic configuration diagram of a map input system using a house recognition method according to an embodiment of the present invention.

FIG. 2 is a diagram for explaining the structure of area data.

FIG. 3 is a diagram for explaining a procedure of a core line forming process.

FIG. 4 is a diagram for specifically explaining the skeletonization process.

FIG. 5 is a diagram illustrating a house recognition method in the map input system according to the embodiment of the present invention.

FIG. 6 is a diagram for explaining a connection process in a house recognition method in the map input system.

[Explanation of symbols]

 12 Scanner 14 CRT Display Device 16 Position Input Device 18 Central Processing Unit 22 Element File 24 Contour Line File 26 Area File 28 Vector File 32 Structured File

Claims (2)

[Claims] 1. A map to be processed is read in advance and the obtained binary image data is converted into vector data,
In the house recognition method in the map input system for recognizing house figures, when a frame is specified on the screen, a closed-loop shape within the frame and having a predetermined size or more is extracted as a house figure to be recognized. A method for recognizing a house in a map input system, characterized by shaping the extracted house figure of the recognition target into a house shape and displaying it on a screen, and then instructing whether to register the house figure of the recognition target. . 2. After shaping the recognition target house figure into a house shape, if the sides that should originally overlap with the recognition target house figure and the already registered house figure are deviated The house recognition method in the map input system according to claim 1, wherein connection processing is performed to match existing sides.