JP2002092658A - Three-dimensional digital map forming device and storage medium storing three-dimensional digital map forming program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a device using on-aircraft type LRF data for automatically performing area recognition of a two-dimensional digital map to accurately and efficiently turning a building, a road, and the ground in the existing two-dimensional digital map into three-dimensional ones. SOLUTION: A three-dimensional digital map forming device is provided with a road modeling process 3, a building modeling process 4, and a ground modeling process 5, and the road modeling process 3 forms a road area from a road edge described as a polyline in an existing digital map 1. After a road centerline is found in the formed road area, height information is added in sequence from the road centerline to the road overall area. In the building modeling process 4, the most frequent value of the LRF data within a building outline is set to be a roof height while a minimum value of the LRF data located in the outside of the building outline and within a fixed distance from the building outline is set to be a ground height of the building. In the ground modeling process 5, the ground is modeled by aiming at the vicinity of the building or the road and a flat open space.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to, for example, an airborne laser range finder (LRF).
The present invention relates to a three-dimensional digital map creating apparatus using elevation data of a feature acquired by a Range Finder.

[0002]

2. Description of the Related Art A conventional geographic information system is constructed by a two-dimensional digital map. However, in this two-dimensional digital map, since there is no height information, the three-dimensional real world cannot be sufficiently represented.

[0003] For this reason, there is a limit in using a wireless communication system that requires three-dimensional shape information for planning, disaster prevention simulation, city planning, and the like.

For example, when a relay point of a wireless communication system is installed in a city, the height information of the building is not known.
Sometimes it is not possible to determine whether there is a building that blocks radio waves, and in the disaster prevention field, it may not be possible to understand how the disaster spreads by performing a simulation of the disaster. In addition, there was a possibility that wrong decisions were made.

[0005] In city planning and landscape design based on two-dimensional map data, there are many cases where the will of the designer cannot be sufficiently communicated to the client. This is because the visual and spatial images are different between the designer and the client, but in the worst case, the client could be given a completely different image. .

[0006] These problems are due to lack of three-dimensional shape information as described above. However, there are three-dimensional maps created by the following various methods.

For example, a so-called digital photogrammetry method in which three-dimensional data is obtained from an aerial photograph taken by an operator to create a three-dimensional map, stereo matching of the aerial photograph is automatically performed by a computer, and There has been a method of creating a three-dimensional digital map by setting a temporary height of several meters per floor for a building in the two-dimensional digital map.

In recent years, with the development of laser measurement technology, a laser range finder (LRF) using an eye-safe laser (a laser type that does not cause damage even when directly in the eyes) can be used in urban space. Has become this LR
F is mounted on an aircraft and a method of converting a feature such as a building or a tree into three dimensions based on the data acquired by the LRF, or, for example, collating a two-dimensional digital map with LRF data,
There has also been a method of estimating the height of a building from RF data.

[0009]

However, the photogrammetry by the operator described above is effective and accurate in obtaining three-dimensional information of terrain and buildings, but requires a huge amount of work and is efficient. Was bad.

[0010] In the digital photogrammetry method, there are shadows of a building and portions (occlusions) that are not shown in a photograph, and this causes measurement errors. Also, due to the algorithm and the digitizer accuracy of the pictures, it was a reality that good results were not obtained much.

The method of creating a three-dimensional digital map by estimating the height of a building is not necessarily the same for all buildings because the height per floor is not the same. The accuracy of the three-dimensional digital map was very poor.

In the method of three-dimensionally converting a feature from LRF data, since only LRF data is used, high-density data at intervals of several cm to several tens cm is obtained to model an individual building. Otherwise, there is a problem that an accurate feature outline cannot be obtained.

Further, the method of estimating the building height from the comparison between the two-dimensional digital map and the LRF data has not yet been reported only for the estimation of the roof of the building. Since other features (roads, etc.) are line segments and have not yet been divided into regions, it has not been possible to add height information to these features.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-mentioned problem. The present invention automatically recognizes a region of a two-dimensional digital map, and efficiently and accurately uses data acquired by an airborne LRF. It is an object of the present invention to provide an apparatus for converting a building, a road, and the ground of an existing two-dimensional digital map into three dimensions.

[0015]

According to the present invention, there is provided means for reading a two-dimensional digital map of a predetermined area and LRF data obtained by acquiring the area by an aircraft, defining both data in the same coordinate system, A means for extracting a road edge from the two-dimensional digital map defined in the same coordinate system and defining it in the coordinate system, and a certain condition among a group of triangles generated by generating a TIN from the road edge. Means for searching a triangle belonging to the road surface with the highest possibility as a seed triangle, and sequentially filling the triangle group from the searched seed triangle to an adjacent triangle, and then filling the triangle group. Means for extracting a road region by leaving only the filled triangles by erasing unseeded triangles; and Means for generating a line, and after generating the center line, reads LRF data defined in the same coordinate system as the two-dimensional digital map, and reads LRF data closest to each node constituting the center line of the road area. z
The gist of the invention is to provide a three-dimensional model of a road, comprising: means for setting a value of the center line to a value; and means for giving the height of the center line to a road edge forming the road area. .

The LRF existing in the building outline
The data is extracted, the mode of the Z value is obtained from the LRF data, the mode is defined as the height of the roof of the building, and all the LRF data within a predetermined distance from the building outline are read. Is obtained, and the Z value of the minimum value is defined as the height of the ground of the building.

Further, a TIN is generated from the LRF data, a substantially flat set of triangles is set as an open space, and the TIN is generated again by matching the open space with the three-dimensional model of the road and the contour of the ground of the building. The gist of the invention is to generate a three-dimensional model of the ground.

[0018]

DESCRIPTION OF THE PREFERRED EMBODIMENTS In the present invention, processing is performed in the following procedure to add LRF data to an existing digital map and form a three-dimensional model of roads, buildings, and ground.

First, the existing digital map 1 and the LRF data 2 are converted into the same coordinate system, and a road modeling process 3, a building modeling process 4, and a ground modeling process 5 are performed.

[Road Modeling Process] A road area is generated from the road edge described as a polyline in the digital map 1. Then, for the generated road area,
After obtaining the road center line, height information is added in order from the road center line to the entire road.

[Modeling of Building] In this study, a single flat roof is assumed for simplicity. Since the building outline is described on the two-dimensional map, the mode value of the LRF data in the building outline is set to the roof height, and the LRF data of the LRF data outside the building outline and within a certain distance from the building outline is set. Let the minimum value be the ground level of the building.

[Modeling of the ground] In a city where buildings are densely populated, it is difficult to obtain contour data.
Model the ground based on the data. However,
In order to specify the ground from LRF data of a limited number of points, the details of the ground were ignored, and the ground was modeled by focusing on buildings and roads and flat open spaces.

Next, details of the road modeling process 3, the building modeling process 4, and the ground modeling process will be described.

FIG. 1 is a schematic configuration diagram of a method for creating a three-dimensional digital map according to an embodiment of the present invention.

The three-dimensional digital map creation method shown in FIG. 1 includes a three-dimensional road modeling part 3, a three-dimensional building modeling part 4, and a three-dimensional ground modeling part 5.

(Road Modeling Processing Unit 3) In the road modeling processing unit 3, first, all road data is read from the map, and the road edge of this road is used as a constraint condition, that is, the road edge is always placed on the side of the generated TIN. TIN is generated as follows. Next, in the generated TIN, a seed triangle determined to be highly likely to be in the road area is searched, and a triangle determined to be in the road among adjacent triangles starting from this is searched. Fill it.

Then, by repeating the same painting process on the triangles adjacent to the painted triangle, the road area is represented by a group of painted triangles. This process is repeated until no seed triangle is found.

Then, a road center line is generated in the generated road area and height information is given. Finally, a three-dimensional TIN model of the road is generated by adding height information from the center line having the height information to each vertex of the road area (road edge).

In this description, some concepts will be defined based on FIG. First, a line segment (dotted line) connecting the end point (End Point) of the original road edge and the end point
Is referred to as an end edge of the road. In the target triangle, the side that was the original road edge is the main side (Ke
y Edge), and the vertex paired with the main edge is the main point (Ke
yRoint). Also, the distance from the principal point to the principal side is defined as the height of the triangle, and adjacent triangles having the same principal side are called co-principal triangles.

The three-dimensional road modeling unit 3 is constructed as shown in FIG.
As shown in the figure, a preprocessing 6, a TIN generation processing 7, a triangle analysis processing 8, a seed triangle search processing 9, a road area filling processing 10, a center line generation processing 11, a three-dimensional center line And a road model generation process 13.

In the preprocessing 6, although the actual digital map is formed by polylines, as shown in FIG.
Since fine discontinuous lines and the like are also mixed, in this preprocessing, such an error of the digital map is repaired (dust removal) to obtain the road map data shown in FIG. FIG. 3C shows map data on a road screen after dust removal.

Next, in order to facilitate selection of seed triangles, which will be described later, the map data after error correction in FIG. 3B is reconstructed for all road edges within an allowable range. (D), the length of each segment of the polyline at the road edge is set to be substantially the same. After the completion of the preprocessing 6, the TIN generation processing 7 is started.

Next, a TIN generation process 7 generates a TIN as shown in FIG. 3E using the road edge as a constraint. That is, with the road edge as a constraint, the road edge is always generated by the generated TIN (triangulated Ir).
(regular network),
TIN (triangle) is generated. Then, the triangle analysis processing 8 is started.

Next, the triangle analysis processing 8 is performed as shown in FIG.
As shown in (1), a triangle existing in the generated road area is analyzed. The analysis of this triangle means whether there is a principal side in the target triangle, whether to change the direction of the principal point,
Analyze whether the height of the triangle is within the road edge. Here, the direction of the principal point is the direction of the road edge passing through the principal point.

After performing such an analysis on all triangles, seed triangle search processing 9 is started.

The seed triangle search processing 9 includes the generated TIN
A search is made for triangles (hereinafter referred to as seed triangles) that are determined to be highly likely to be in the road area (FIG. 4B).

This seed triangle always satisfies the following conditions. (A) The road surface has not yet been painted. (B) Having only one main edge. (C) The direction of the principal point is unique and parallel to the principal side. (D) The height of the triangle is within the threshold of the road width. (E) If there is a co-principal side triangle, it is not painted. (F) An angle other than the principal point is an acute angle. (G) The height of the triangular triangle is higher than itself.

Those satisfying these conditions are searched for as seed triangle candidates, and the ratio between the height of the common principal triangle and the height of the seed triangle candidate is determined for each of the various triangle candidates. , And the road region filling process 10 is started with the seed triangle candidate having the highest probability among all the seed triangle candidates as the seed triangle.

In the road area painting process 10, a seed triangle determined to be present on the road is painted, and a triangle adjacent to the seed triangle is painted in the same manner as described above. This is repeated until no triangle is found in the road area.

That is, the seed triangles that have been searched are painted out as shown in FIG.

At this time, the triangle is a seed triangle. With the triangle as the current triangle, another triangle sharing an edge other than the main edge and the end edge of the current triangle is searched for and filled. Using the filled triangle as a new current triangle, a new triangle sharing an edge other than the main edge and the end edge is painted (the number is the order of painting).

Next, the road area painting process 10 generates a road area as shown in FIG. 5A by deleting unfilled triangles and leaving only filled triangles. FIG. 5B shows an example of an actual screen from which unnecessary triangles have been removed.

When this processing is completed, the center line generation processing 11 is started. As shown in (c) of FIG. 5, the center line generation processing 11 performs, for each triangle constituting the road area, a triangle having only one main side, and a triangle aligned with the main side as a center line. pull.

Next, the center lines of all the obtained triangles are connected to generate a road center line. FIG. 5D is an example of a screen when a road center line is drawn.

Then, the center line three-dimensional processing 12 is started. The center line three-dimensional processing 12 searches for nearby LRF data for each node constituting the road center line,
It is a height candidate value. Next, smoothing processing is performed to obtain smooth height information, and the height of the center line is determined.

For example, the bold line in FIG. 6A is the center line of the road, the circle is LRF data, and the dotted line is the search area.
In FIG. 6A, the length of the perpendicular from an arbitrary point on the dotted line to the center line (the distance from the point without the perpendicular to the nearest vertex) is the search distance, in other words, the arbitrary point on the dotted line. The minimum distance from to the center line is constant and is the search distance.

FIG. 6B shows that a perpendicular line is drawn from the LRF point within the dotted line to the center line.

FIG. 6C shows the distance from the foot drop to the start point of the center line on the horizontal axis, and the height of the LRF point on the vertical axis.

Also, the vertex of the center line may be a drop of a plurality of LRF points. For example, as for the start point and end point of the center line, the start point is a drop of two LRF points, and the end point is the end point of three LRF points.

Therefore, as shown in FIG. 6C, there are two small circles at the start point (Begin) and three small circles at the end point (End).

FIG. 7A shows each point of FIG. 6C as an approximate straight line using the least squares method.

FIG. 7D shows the height of the original vertex on the center from the approximate straight line using the distance (horizontal axis) from the approximate vertex to the start point of the center line along the center line. This indicates that it is to be obtained (black dot).

After the height is given in this way, the road model generation processing 13 is started.

In the road model generation process 13, a perpendicular line is drawn from each vertex of each triangle constituting the road area to the road center line, and the height of that point is set to the vertex of each triangle, that is,
Assume the height of the road edge.

In the case of FIG. 8A, since each triangle in the area is directly related to the center line, height information is directly obtained from the center line.

For each vertex of the triangle in the area, height information is sequentially provided from the adjacent triangle to which height information has already been provided. As an example, the height of point C is calculated from points A and B in the figure. For example, the center of the side connecting points A and B of the triangle and the center of the side connecting B and C are determined, and the centers of these sides are connected by a line. The three-dimensional road image shown in FIG. 8B is generated by giving height information in this manner.

(Building Modeling Processing) On the other hand, as shown in FIG. 1, the building modeling processing unit 4 includes a preprocessing 15, a roof height calculation processing 16, and a ground height calculation processing 17 The height of the roof and the height of the ground are obtained as shown in FIG.

Since the outline of the building is described on the digital map, height information on the top and bottom of the building (roof and ground) is added.

In the pre-processing 15, since the digital map is often divided into specific map units, FIG.
As shown in (1), one building may be divided into two adjacent meshes. In order to model a building, it is necessary to connect the divided buildings as shown in FIG. It is also necessary to exclude trash on the map.

For example, as shown in FIG. 10A, a building B extends over a map A and a map B. For this reason, the boundary line of the building B is connected by eliminating the vertices on the boundary line between the map frame A and the map frame B.

For example, the building C has a small number of vertices,
In addition, since the shape is slender, garbage is removed because it is not a building.

After such a process is performed for each building, a calculation process 16 of roof height information is started.

The process 16 for calculating the height information of the roof is based on the premise that the roof is a flat roof and all the LRFs within the building outline
The most frequent value of the Z value of the data is obtained and set as the height of the flat roof ((c) in FIG. 10).

After such processing is performed for each building, the ground height calculation processing 17 is started.

In this processing, as shown in FIG. 11A, the minimum value of the LRF data outside the building outline and within a predetermined distance (fixed value) from the building outline is set as the height of the ground of the building. .

As shown in FIG. 11B, a three-dimensional image of the building is obtained from the height of the building obtained by the above processing. In other words, the two-dimensional original drawing shown in FIG. 11C is changed to a three-dimensional view of the building as shown in FIG. 11D.

Further, the ground modeling processing unit 5
Generation processing 20, processing 21 for deleting points inside buildings and roads,
The apparatus includes a flat triangle search process 22 and a ground model generation process 23, and performs ground modeling based on LRF data. However, in order to estimate the ground from LRF data of a limited number of points, the details of the ground are ignored, and the ground is modeled focusing on buildings and roads and flat open spaces.

The TIN generating step 2 generates a TIN using the LRF data as shown in FIG.

Next, a process 21 for deleting an interior point of a building / road.
Deletes LRF data located in a building or a road and then removes triangles having the vertices as shown in FIG.

Next, the flat triangle search process 22 calculates a normal vector of each generated triangle to detect a flat open space, as shown in FIG. 13A. Only triangles within the threshold are left, and the remaining triangles are painted out (not shown).

Next, the generation process 23 of the ground model is shown in FIG.
As shown in FIG. 3 (b), a TIN is newly generated by adding the ground contour of the building and the contour of the road area to the set of the remaining triangular points as constraint conditions to generate a TIN model of the ground. (Not shown).

Then, the synthesis processing section 14 inputs the road modeling data, the building modeling data, and the ground modeling data obtained by the above-described processing, and synthesizes them. As a result, FIG. A three-dimensional model shown in (c) is obtained.

[0073]

As described above, according to the present invention, L acquired by an airborne laser range finder (LRF) is obtained.
Use RF data to quickly identify roads, buildings, ground, etc.
Since it can be dimensioned, it can contribute to the construction of a three-dimensional geographic information system. By using the three-dimensional geographic information system, it is possible to efficiently realize three-dimensional display and three-dimensional simulation desired in fields such as city planning and disaster prevention.

[Brief description of the drawings]

FIG. 1 is a schematic configuration diagram of a three-dimensional digital map creation device according to an embodiment.

FIG. 2 is an explanatory diagram illustrating a definition related to detection of a road area.

FIG. 3 is an explanatory diagram illustrating a process of creating a road model.

FIG. 4 is an explanatory diagram illustrating a process of creating a road model.

FIG. 5 is an explanatory diagram illustrating a process of creating a road model.

FIG. 6 is an explanatory diagram illustrating a process of creating a road model.

FIG. 7 is an explanatory diagram illustrating a process of creating a road model.

FIG. 8 is an explanatory diagram illustrating a process of creating a road model.

FIG. 9 is an explanatory diagram illustrating generation of a ground area of a building.

FIG. 10 is an explanatory diagram illustrating generation of building modeling.

FIG. 11 is an explanatory diagram illustrating generation of building modeling.

FIG. 12 is an explanatory diagram illustrating generation of ground modeling.

FIG. 13 is an explanatory diagram illustrating generation of ground modeling.

[Explanation of symbols]

 3 3D road modeling processing section 4D building modeling processing section 5 3D ground modeling processing section 14 Synthesis processing section

Continued on the front page (72) Inventor Kazuo Oda 4-2-18 Shinjuku Kofu Building, Shinjuku, Shinjuku-ku, Tokyo Inside Asia Air Survey Co., Ltd. (72) Inventor Lu Wei Shinjuku Kofu Building Asia, 4-2-18 Shinjuku, Shinjuku-ku, Tokyo, Asia Kokuso Co., Ltd. F term (reference) 2C032 HB05 HC09 HC21 HC23 5B050 AA01 BA09 BA17 EA06 EA13 EA30

Claims (12)

[Claims] 1. A means for reading a two-dimensional digital map of a predetermined area and LRF data obtained by acquiring the area by an aircraft, defining both data in the same coordinate system, and means for defining both data in the same coordinate system. A means for extracting a road edge from the two-dimensional digital map and defining it in the coordinate system; a means for generating a TIN with the road edge as a constraint; a predetermined condition among the generated TINs; Means for defining a triangle to be satisfied as a seed triangle and searching for the triangle; and filling the adjacent triangles sequentially from the searched seed triangle in the TIN, and then erasing the unfilled triangles on the TIN. Means for extracting a road area composed of the filled triangles; and means for generating a center line in the extracted road area. After generating the center line, it reads the LRF data defined in the two-dimensional digital map of the same coordinate system, closest to the respective nodes of the center line of the road area LRF
3. A three-dimensional digital map generator, comprising: means for setting the z value of data to the height of a node of the center line; and means for applying the height of the center line to a road edge forming the road area. apparatus. 2. The method according to claim 1, wherein the center line is obtained by obtaining the centers of two sides of the triangle constituting the road area, which are not road edges, and connecting the centers of these sides. 3D digital map creation device. 3. The method according to claim 1, wherein the height of the road edge line is determined by setting a Z value of an intersection point when a perpendicular line is drawn from a vertex of each triangle forming the road area to the center line as a height of the edge point. The three-dimensional digital map creation device according to claim 1, wherein the three-dimensional digital map creation device is provided. 4. A method for extracting the laser data present in the building outline, obtaining a mode of the Z value from the laser data, and determining the mode Z value as the height of the roof of the building. Means for reading all the laser data within a predetermined distance from the building outline, obtaining the minimum value of these laser data, and setting the Z value of this minimum value as the height of the ground of the building. The three-dimensional digital map creation device according to claim 1, wherein: 5. A means for generating triangles in an image of the predetermined area using the laser data of the predetermined area, and a method of removing each of the triangles existing in the road area and the ground of the building after removing the triangles. A means to obtain a predetermined open space by calculating the line vector and leaving only triangles within the threshold value, and adding the ground contour of the building and the contour of the road area as constraints to the set of laser data of the remaining triangles 2. A three-dimensional digital map creation apparatus according to claim 1, further comprising: means for generating a new triangle after the step. 6. A means for generating a three-dimensional image of a road from the height-added center line and the road edge to which the height is applied, and a three-dimensional image of the building from the height of the building outline. Means for generating; a means for generating a three-dimensional image of the ground from the height of the ground of the building; and means for generating a three-dimensional model of the ground based on the heights of the vertices of each triangle in the open space. The three-dimensional digital map creation method according to claim 1, 2, 3, 4, or 5. 7. A step of reading a two-dimensional digital map of a predetermined area and LRF data obtained by acquiring the area by an aircraft, and defining both data in the same coordinate system; Extracting a road edge from the two-dimensional digital map and defining the same in the coordinate system; generating a TIN with the road edge as a constraint; and determining a predetermined condition in the generated TIN. Defining a triangle that satisfies as a seed triangle and searching for it; and, in the TIN, after sequentially filling adjacent triangles from the searched seed triangle, erasing triangles that are not filled on the TIN. Extracting a road area composed of the filled triangles by: generating a center line in the extracted road area A step of, after generating the center line, to read the LRF data defined in the two-dimensional digital map of the same coordinate system,
L closest to each node constituting the center line of the road area
Causing the z-value of the RF data to be the height of the centerline node; providing the height of the centerline to a road edge forming the road area; Generating a three-dimensional model of the road from the road edge to which the height has been assigned. A storage medium storing a program for creating a three-dimensional digital map. 8. The method according to claim 1, further comprising: determining the centers of sides in the road area connecting the vertices of the triangles constituting the road area, and connecting the centers of these sides by lines to generate the center line. A storage medium storing a program for creating a three-dimensional digital map according to claim 7. 9. The method of providing the height of the road edge line, comprising: determining a Z of an intersection of a center line when a perpendicular line is dropped from the vertex of each triangle located on the road edge line constituting the road area to the center line. 9. A storage medium storing a program for creating a three-dimensional digital map according to claim 7, wherein a value is given as a height of a point on the edge. 10. The LR present in the building outline
Extracting the F data, calculating the mode from the Z value of the LRF data, and setting the mode to the height of the building roof; and calculating all the LRF data at a predetermined distance from the building contour. 8. A program for creating a three-dimensional digital map according to claim 7, further comprising the steps of: causing the minimum value of the Z value of the LRF data to be obtained by reading the LRF data; and making the minimum value the height of the ground of the building. Storage medium that stores. 11. A method of generating a triangle in an image of the predetermined area using the laser data of the predetermined area, and removing a triangle existing in the road area and the ground of the building, and calculating a method of each triangle remaining. Calculate the line vector and leave only the triangles within the threshold to obtain a predetermined open space.Add the contours of the building and the ground of the road to the set of laser data of the remaining triangles as constraints. 11. A storage medium storing a program for a method for creating a three-dimensional digital map according to claim 7, further comprising a step of generating a new TIN. 12. A step of generating a three-dimensional image of a road from the height-added center line and the road edge to which the height is applied, and generating a three-dimensional image of the building from the height of the building outline. Generating, a step of generating a three-dimensional image of the ground from the height of the ground of the building, and a step of generating a three-dimensional model of the ground based on the heights of the vertices of each triangle in the open space. 13. The method according to claim 7, wherein the method according to claim 7 is performed.
A storage medium storing a program for a three-dimensional digital map creation method.