Map information processing device

The map information processing device enhances the accuracy and efficiency of labeling road and building boundaries in point cloud maps by using a boundary contour extraction unit to create closed curves from survey maps, addressing the inefficiencies of existing technologies.

JP2024159243APending Publication Date: 2024-11-08HITACHI LTD
View PDF 1 Cites 0 Cited by

Patent Information

Application Number
JP2023075099
Authority / Receiving Office
JP · JP
Patent Type
Applications
Current Assignee / Owner
Filing Date
2023-04-28
Publication Date
2024-11-08

AI Technical Summary

Technical Problem

Existing technologies struggle to accurately recognize road edges in point cloud maps with small changes in height or brightness, leading to increased processing load and inefficiency in labeling structures like roads and buildings.

Method used

A map information processing device that includes a point cloud map acquisition unit, a survey map acquisition unit, a boundary contour extraction unit, and structure point group classification units to create closed curves from survey maps, allowing for accurate extraction and labeling of road and building boundaries in point cloud maps.

Benefits of technology

Enables efficient and accurate labeling of structures on point cloud maps by reducing processing load and improving the recognition of road and building edges, even with small changes in height or brightness.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 2024159243000001_ABST
    Figure 2024159243000001_ABST
Patent Text Reader

Abstract

To provide a map information processing device capable of easily performing labeling for classifying structures such as roads and buildings for a point group map.SOLUTION: According to this disclosure, a map information processing device (100) includes a point group map acquisition part (120) for acquiring a point group map, a survey map acquisition part (110) for acquiring a survey map including outlines of structures, a boundary outline extraction part (140) for creating closed curves of a structure boundary from the survey map, and structure point group extraction parts (150, 160) for extracting a remaining point group obtained by removing point groups within the closed curves of the structure boundary from the point group map as a structure point group.SELECTED DRAWING: Figure 1
Need to check novelty before this filing date? Find Prior Art

Description

[Technical field]

[0001] The present invention relates to a map information processing device that performs labeling of structures on a point cloud map. [Background technology]

[0002] For example, various techniques have been proposed for classifying or extracting desired structures such as roads and buildings from a point cloud map. Patent Document 1 discloses a technique for an information processing device that generates boundary position information indicating the positions of road boundaries such as division lines and road shoulder edges used as map information, in which perpendicular lines are drawn at equal intervals to the road boundaries on the map, and points on the perpendicular lines with large changes in height or brightness are extracted and determined as road ends. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent Publication No. 2022-89828 Summary of the Invention [Problem to be solved by the invention]

[0004] The technology of Patent Document 1 cannot accurately recognize roadsides when the change in height or brightness is small in the point cloud on the perpendicular line. In addition, it is necessary to determine the change in height or brightness for all points in the point cloud, which imposes a large processing load.

[0005] The present invention has been made in consideration of the above points, and an object of the present invention is to provide a map information processing device that can easily perform labeling to classify structures such as roads and buildings on a point cloud map. [Means for solving the problem]

[0006] The map information processing device of the present invention that solves the above problems includes: A point cloud map acquisition unit that acquires a point cloud map; a survey map acquisition unit for acquiring a survey map including boundary information of a structure; a boundary contour extraction unit for generating a structure boundary closed curve from the survey map; a structure point cloud extraction unit that extracts a point cloud by determining whether a point cloud included in the point cloud map exists inside or outside the structure boundary closed curve, and classifies the extracted point cloud into a structure point cloud; The present invention is characterized by having the following. Effect of the Invention

[0007] According to the present invention, it is possible to obtain a map information processing device that can easily perform labeling to classify structures such as roads and buildings on a point cloud map.

[0008] Further features related to the present invention will become apparent from the description of the present specification and the accompanying drawings. In addition, problems, configurations, and effects other than those described above will become apparent from the following description of the embodiments. [Brief description of the drawings]

[0009] [Figure 1] FIG. 1 is a functional block diagram of a map information processing apparatus according to a first embodiment. [Diagram 2] FIG. 1 shows an example of a survey map displayed in plane coordinate format. [Diagram 3] FIG. 13 is a diagram showing an example of a survey map displayed in data format. [Figure 4] A diagram showing the point cloud map in 3D format. [Diagram 5] A diagram showing the point cloud map in data format. [Figure 6] 11 is a flowchart showing the process of a point cloud map area extraction unit. [Figure 7] 11 is a diagram showing an example of extracting road areas from a point cloud map using road boundary information; [Figure 8] 11 is a flowchart showing the contents of a boundary closed polygon extraction process in a boundary contour extraction unit. [Figure 9] 13 is a flowchart showing the contents of a process S630 for extracting a closed building boundary polygon in the boundary contour extraction unit. [Figure 10] FIG. 13 is a diagram illustrating an example of a sequence of building boundary points. [Figure 11] 13 is a flowchart showing the contents of a process S640 for extracting a road boundary closed polygon in a boundary contour extraction unit. [Figure 12] FIG. 13 is a diagram showing an example of a sequence of points of a road boundary extracted in S830. [Figure 13] 13 is a flowchart showing the contents of the process of creating a road boundary closed polygon in S840. [Figure 14] FIG. 13 is a diagram showing an example of creating a road boundary closed polygon. [Figure 15] 11 is a flowchart showing the process of a building point cloud extraction unit. [Figure 16] 13A to 13C are views for explaining a modified example of the processing of the boundary contour extraction unit. [Figure 17] 11 is a flowchart showing the process of a road point cloud extraction unit. [Figure 18] An illustration showing the state on a survey map after removing the points contained within the closed road boundary polygon from the point cloud on the ground. [Figure 19] 11 is a flowchart showing the process of a point cloud reclassification unit. [Figure 20] FIG. 11 is an explanatory diagram illustrating the effect of a point cloud reclassification unit. [Figure 21] FIG. 11 is an explanatory diagram illustrating a second embodiment. [Figure 22] FIG. 11 is an explanatory diagram illustrating a third embodiment. [Figure 23] 13 is a flowchart showing a process of a road point cloud extraction unit in the third embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0010] Next, each embodiment of the present invention will be described in detail with reference to the drawings. In each of the following embodiments, roads and buildings will be described as examples of structures to be labeled, but either one of them may be used, and labeling only other structures or adding other structures is also included in the concept of the present invention.

[0011] [First embodiment] FIG. 1 is a functional block diagram of a map information processing apparatus according to the first embodiment.

[0012] The map information processing device 100 is a device that performs labeling to classify structures such as roads and buildings for each point on a point cloud map. The map information processing device 100 is configured as a computer system that includes a processor such as a CPU (Central Processing Unit), memories such as a RAM (Random Access Memory) and a ROM (Read Only Memory), and further includes input / output devices such as a mouse, a keyboard, and a display, a communication device (none of which are shown), and the like.

[0013] The map information processing device 100 is connected to a survey map database 200 and a point cloud map database 300, and is capable of transmitting and receiving data between these databases. The survey map database 200 stores a survey map 201, and the point cloud map database 300 stores a point cloud map 301.

[0014] Figures 2 and 3 show examples of survey maps stored in the survey map database, where Figure 2 is an example of a survey map displayed in planar coordinate format, and Figure 3 is a diagram showing an example of a survey map displayed in data format.

[0015] The survey map 201 is a two-dimensional (2D) planar map such as a topographical map, and includes boundary information of structures such as a road boundary (road edge contour) 212 of a road 211 and a building boundary (building contour) 222 of a building 221, as shown in FIG. 2. The survey map 201 is stored in a survey map database 200 in a data format such as XML format, as shown in FIG. 3. The data of the survey map 201 includes information such as the type of structure, for example, whether the structure is a road 211 or a building 221. The data of the survey map 201 also includes a section 202 representing a road boundary (here, RoadEdge) and a section 203 representing a building boundary (Building), and each boundary is stored as a point sequence. The data of the survey map 201 also includes the names of the road 211 and the building 221, and attribute information representing the roadway and sidewalk of the road 211.

[0016] Figures 4 and 5 show examples of point cloud maps stored in the point cloud map database, where Figure 4 is a diagram showing the point cloud map in three-dimensional (3D) format, and Figure 5 is a diagram showing the point cloud map in data format.

[0017] The point cloud map 301 is a collection of three-dimensional measurement points that are created by measuring urban areas and city streets with measuring devices such as LiDAR and cameras and using SLAM (Simultaneous Localization and Mapping). The point cloud map 301 includes a plurality of measurement points (hereinafter, point cloud) 302 that measure the three-dimensional shapes of roads 211, buildings 221, roadside trees 231, etc. The measuring devices that measure the three-dimensional shapes are mounted on moving objects such as automobiles, airplanes, and drones, and measurements are performed as these moving objects move. The point cloud map 301 is an environmental map of a three-dimensional (3D) point cloud that uses an orthogonal coordinate system as the reference coordinate system.

[0018] 5, the point cloud map 301 is saved as, for example, a text file, and data on each measurement point of the point cloud 302 may include color information (Rn, Gn, Bn) and information on reflection intensity (In) in addition to position information (Xn, Yn, Zn). Note that the label information output in this embodiment can be retained by assigning label information Ln that corresponds buildings, roads, etc. to integers to each measurement point. In other words, label information Ln indicating the classification of buildings or roads is assigned to each measurement point of the point cloud.

[0019] Various functions are realized in the map information processing device 100 by hardware such as a CPU reading and executing software programs stored in memories such as RAM and ROM. As shown in Fig. 1, the map information processing device 100 includes a survey map acquisition unit 110, a point cloud map acquisition unit 120, a point cloud map area extraction unit 130, a boundary contour extraction unit 140, a building point cloud extraction unit 150, a road point cloud extraction unit 160, and a point cloud reclassification unit 170 as functions realized by executing the software programs.

[0020] The survey map acquisition unit 110 acquires the survey map 201 from the survey map database 200. Then, the point cloud map acquisition unit 120 acquires the point cloud map 301 from the point cloud map database 300. The survey map acquisition unit 110 acquires the survey map in the survey map acquisition area 230. The survey map acquisition unit 110 acquires the attributes of the structures from the survey map. The range of the survey map acquisition area 230 acquired by the survey map acquisition unit 110 is set based on the point cloud acquisition area 320 output by the point cloud map area extraction unit 130, which will be described later. As will be described later, in this embodiment, road boundary information in a wider area than the point cloud acquisition area 320 is required to perform labeling of the point cloud map 301, so the survey map acquisition area 230 needs to be set larger than the point cloud acquisition area 320. The range of the survey map acquisition area 230 depends on the road shape and the map, but it is desirable to set the survey map acquisition area 230 to about 3 to 10 times the point cloud acquisition area 320.

[0021] The point cloud map area extraction unit 130 sets a labeling range 310 in which labeling is performed for each measurement point of the point cloud in the point cloud map 301, and sets a rectangle including the labeling range 310 as a point cloud acquisition area 320.

[0022] The boundary contour extraction unit 140 extracts a closed curve of the structure boundary from the survey map 201. The boundary contour extraction unit 140 creates a closed curve of the structure boundary based on the attributes of the structure in the survey map 201. For example, the boundary contour extraction unit 140 performs a process of extracting a road boundary closed polygon (road boundary closed curve) 210 of the road end contour of a road 211, which is a first structure, and a building boundary closed polygon (building boundary closed curve) 220 of the building contour of a building 221, which is a second structure, as closed curves of the structure boundary from the survey map 201. The boundary contour extraction unit 140 extracts a point sequence of the road boundary 212 from the survey map 201 using the road boundary information of the survey map 201. Then, the boundary contour extraction unit 140 creates a road boundary closed polygon 210 consisting of a closed curve of the road end contour by connecting the end points of the point sequences adjacent to each other of the multiple road boundaries 212. Furthermore, the boundary contour extraction unit 140 extracts a sequence of points relating to buildings from the survey map 201 using the building boundary information of the survey map 201 , and creates a building boundary closed polygon 220 consisting of a closed curve indicating the building contour of a building 221 .

[0023] The building point cloud extraction unit 150 and the road point cloud extraction unit 160 extract a point cloud of a structure using the closed curve of the structure boundary created by the boundary contour extraction unit 140. The building point cloud extraction unit 150 and the road point cloud extraction unit 160 extract a point cloud by determining whether a point cloud 302 included in a point cloud map 301 is inside or outside the closed curve of the structure boundary, and classify the extracted point cloud into a structure point cloud. The building point cloud extraction unit 150 and the road point cloud extraction unit 160 each constitute a structure point cloud extraction unit of claim 1.

[0024] The building point cloud extraction unit 150 performs labeling to classify the point cloud 302 within the labeling range 310 of the point cloud map 301 into a ground surface point cloud 303 and a ground object point cloud 304, using the building boundary information of the survey map 201. This classification may be performed by the road point cloud extraction unit 160, which will be described later.

[0025] The building point cloud extraction unit 150 extracts a point cloud included inside the building boundary closed polygon 220 from the point cloud 304 classified as a ground object, and performs labeling to classify the extracted point cloud into a building point cloud (structure point cloud) 304a. In addition, labeling may be performed to classify a point cloud 304b located outside the building boundary closed polygon 220, which remains after the building point cloud 304a is removed from the ground object point cloud 304, into other structure point clouds such as roadside trees 231.

[0026] The road point cloud extraction unit 160 removes the point cloud 303a inside the road boundary closed polygon 210 contained within the road boundary closed polygon 210 from the point cloud 303 classified as on the earth's surface, extracts the remaining point cloud, that is, the point cloud located outside the road boundary closed polygon 210 from the point cloud 303, and performs labeling to classify the extracted point cloud as a road point cloud (structure point cloud) 303b.

[0027] The point cloud reclassification unit 170 creates a classifier using the point cloud map 301, in which each measurement point of the point cloud 302 is labeled, as training data for machine learning. Then, the entire point cloud map 301 to be labeled is input to the classifier to reclassify the point cloud 302 and output it as a labeled point cloud. As a machine learning method, a method such as Random_Forest or a technology such as Semantic_Segmentation can be used.

[0028] Next, a map information processing method using the map information processing device of this embodiment will be described. FIG. 6 is a flowchart showing the process of the point cloud map area extraction unit.

[0029] The point cloud map area extraction unit 130 acquires the point cloud map 301 from the point cloud map acquisition unit 120 (S510), and sets a labeling range 310 in which labeling is performed in the acquired point cloud map 301 (S520). Then, only the point cloud 302 existing within the labeling range 310 is extracted (S530). Then, a rectangle including the labeling range 310 is set as the point cloud acquisition area 320 (S540). Specifically, the point cloud map area extraction unit 130 calculates the vertices 320a of the four corners of the point cloud acquisition area 320. The point cloud acquisition area 320 is used to calculate the survey map acquisition area 230.

[0030] FIG. 7 is a diagram showing an example of extracting a road area from a point cloud map using road boundary information. In the example shown in Fig. 7, a labeling range 310 is shown on a survey map 201. The point cloud map area extraction unit 130 extracts information on a point cloud 302 existing within the labeling range 310 from the point cloud map 301 and supplies the information to the boundary contour extraction unit 140. The labeling range 310 does not need to be along a road as shown in Fig. 7, and can be set to any shape, and may be, for example, a rectangle equivalent to the point cloud acquisition area 320.

[0031] FIG. 8 is a flowchart showing the contents of the boundary closed polygon extraction process in the boundary contour extraction unit. The boundary contour extraction unit 140 acquires a survey map from the survey map acquisition unit 110 (S610), acquires information on the point cloud within the point cloud acquisition area 320 from the point cloud map area extraction unit 130, and unifies the coordinate systems of the survey map 201 and the point cloud map 301 (S620). Then, it performs a process of extracting a building boundary closed polygon 220 of a building 221 (S630), and a process of extracting a road boundary closed polygon 210 of a road 211 (S640).

[0032] FIG. 9 is a flowchart showing the contents of the extraction process S630 of a closed building boundary polygon in the boundary contour extraction unit. When the boundary contour extraction unit 140 acquires a survey map from the survey map acquisition unit 110 (S710), it extracts a sequence of points related to the building boundary 222 within the labeling range 310 of the point cloud map 301 from the building boundary information of the survey map 201 (S720). The sequence of points related to the building boundary 222 has a closed curve that follows the shape (contour) of the building 221 when viewed in plan, and this closed curve is extracted as the building boundary closed polygon 220.

[0033] FIG. 10 is a diagram illustrating an example of a sequence of points on a building boundary. In many cases, building boundary information in the survey map 201 is stored as a sequence of points in a clockwise or counterclockwise direction. In the example shown in Fig. 10, the boundary contour extraction unit 140 extracts a closed curve formed by a sequence of points B that starts from point B1 of the building boundary information as a starting point, moves clockwise to points B2 to B6, and connects point B6 to point B1 again, as a building boundary closed polygon 220. In the point cloud extraction by the building point cloud extraction unit 150 described later, it is possible to determine whether the point cloud within the labeling range 310 is the point cloud of the building 221 by determining whether the XY coordinates of each point are inside or outside the building boundary closed polygon 220 formed by the sequence of points B of the building 221.

[0034] FIG. 11 is a flow chart showing the contents of the extraction process S640 of the road boundary closed polygon in the boundary contour extraction unit. The boundary contour extraction unit 140 acquires the survey map 201 from the survey map acquisition unit 110 (S810), and acquires the point cloud acquisition area 320 from the point cloud map area extraction unit 130 (S820). Then, it extracts point sequences 213 related to road boundaries 212 in the point cloud acquisition area 320 of the point cloud map 301 from the road boundary information of the survey map 201 (S830). Then, it connects the end points of the point sequences 213 adjacent to each other of the multiple road boundaries 212 to create a road boundary closed polygon 210 consisting of a closed curve of the road edge contour (S840).

[0035] FIG. 12 is a diagram showing an example of a sequence of road boundary points extracted in S830. A road boundary 212 in a survey map 201 is often stored as a substantially linear sequence of points 213 in which boundary points M are arranged in order from a start point Ms to an end point Me. The boundary contour extraction unit 140 extracts a plurality of sequence of points 213 related to the road boundary 212 from the road boundary information of the survey map 201. Each boundary point M of the sequence of points 213 is set with one end point of the sequence of points 213 as a start point Ms and the other end point as an end point Me, and in this embodiment, a plurality of boundary points M are stored as a sequence of points 213a to 213h, with the start point Ms and the end point Me set so as to be continuous clockwise.

[0036] The boundary contour extraction unit 140 judges whether the distance between the mutual end points of the multiple point sequences 213 is less than a threshold value Δth, and when it is less than the threshold value Δth, it joins the mutual end points to form a single point sequence 213 that forms an approximate polygon. Then, it joins the start point Ms and the end point Me of the single point sequence 213 that forms the approximate polygon to create a road boundary closed polygon 210 having a closed shape.

[0037] The boundary contour extraction unit 140 extracts end points corresponding to the start point Ms and the end point Me in each point sequence 213 of the road boundary 212, calculates the distance between the start point Ms and the end point Me of each point sequence 213, or between the start points Ms, or between the end points Me, and performs a process of joining the end points whose distance is less than a threshold value Δth to combine the two point sequences 213 into one point sequence 213. Then, the distance between the start point Ms and the end point Me within one point sequence 213 is calculated. If the distance between the start point Ms and the end point Me of one point sequence 213 is less than the threshold value, it is regarded as a road boundary closed curved polygon 210.

[0038] FIG. 13 is a flowchart showing the process of generating a closed road boundary polygon in S840. The boundary contour extraction unit 140 selects the starting point Mis of a sequence of points 213Mi of a road boundary 212 and the end point Mje of a sequence of points 213Mj of another road boundary 212 (S901), compares the distance between the starting point Mis and the end point Mje with a threshold value Δth to determine whether it is less than the threshold value Δth (S902), and if it is less than the threshold value Δth, combines the starting point Mis of the sequence of points 213Mi and the end point Mje of the sequence of points 213Mj (S903).

[0039] Then, when it is determined that the start points Ms and end points Me have been compared for all the point sequences 213 (S904), the start points Ms are compared with each other, that is, the distance between the start point Mks of the point sequence 213Mk of one road boundary 212 and the start point Mls of the point sequence 213Ml of another road boundary 212 is compared with a threshold value Δth, and a decision is made as to whether or not to combine (S905 to S907). When it is determined that the start points Ms have been compared with each other for all the point sequences 213 (YES in S908), the end points Me are compared with each other, that is, the distance between the end point Mne of the point sequence 213Mn of one road boundary and the end point Mme of the point sequence 213Mm of another road boundary 212 is compared with a threshold value Δth, and a decision is made as to whether or not to combine (S909 to S912). When it is determined that the end points Me have been compared with each other for all the point sequences 213, the process proceeds to S913 and subsequent steps to create a closed curve.

[0040] The start point Mos and end point Moe of a sequence of points 213Mo of a road boundary 212 are selected, and the distance between the start point Mos and the end point Moe is compared with a threshold value Δth to determine whether it is less than the threshold value Δth (S913, S914).If it is less than the threshold value Δth, the start point Mos of the sequence of points Mo and the end point Moe of the sequence of points Mo are joined to create a closed road boundary polygon 210 of the road boundary 212 (S915).

[0041] Figure 14 is a diagram showing an example of creating a road boundary closed polygon, where Figure 14(1) shows the state in which the end points of each sequence of points 213M1, 213M2, 213M3 that make up multiple road boundaries are connected, and Figure 14(2) shows a road boundary closed polygon generated by these multiple road boundaries.

[0042] The boundary contour extraction unit 140 extracts end points M1s, M2s, M3s, M1e, M2e, and M3e, which are the start and end points of the point sequences 213M1 to M3, from the survey map 201. Then, the positions of the end points of each point sequence 213M1 to M3 are confirmed clockwise from the point sequence 213M1, and the distance between the end point M1e of the point sequence 213M1 and the start point M2s of the point sequence 213M2 is calculated. Then, if the distance between the end point M1e and the start point M2s is less than the threshold value Δth, the end point M1e and the start point M2s are joined, and the point sequences 213M1 and 213M2 are connected to each other to form a single point sequence.

[0043] Next, the distance between the end point M2e of the sequence of points 213M2 and the end point M3e of the sequence of points 213M3 is calculated. If the distance between the end point M2e and the end point M3e is less than the threshold value Δth, the end point M2e and the end point M3e are joined to form a single sequence of points 213L that is connected in the order of the sequence of points 213M1, the sequence of points M213M2, and the sequence of points M213M3.

[0044] After performing a process of joining these end points, the distance between the end point L1e and the start point L1s in one point sequence 213L is calculated as shown in Fig. 14(2). If the distance between the start point L1s and the end point L1e is within a threshold, the point sequence L1 is regarded as a road boundary closed curved polygon 210 consisting of the road boundary 212.

[0045] FIG. 15 is a flowchart showing the process of the building point cloud extraction unit. The building point cloud extraction unit 150 acquires the point cloud 302 within the labeling range 310 of the point cloud map 301 from the point cloud map area extraction unit 130 (S1110). Then, the point cloud within the labeling range 310 is separated into a ground point cloud and a ground object point cloud (S1120). Here, the ground point cloud 303 is a point cloud obtained by measuring the ground such as a road, and the ground object point cloud 304 is a point cloud obtained by measuring objects having heights such as buildings 221 and roadside trees 231. A publicly known method can be used to separate the point cloud within the point cloud acquisition area 320 into the ground point cloud and the ground object point cloud. Then, the point cloud included in the building boundary closed polygon 220 is extracted from the ground object point cloud 303, and labeling is performed to classify the extracted point cloud into a building point cloud 303a (S1130).

[0046] FIG. 16 is a diagram for explaining a modified example of the process of the boundary contour extraction unit. As described above, the building point cloud extraction unit 150 extracts the building point cloud 304a from the point cloud map 301 using the building boundary closed polygon 220 extracted from the survey map 201. However, since the building contour of a building 221 such as a building is the same size as the building boundary 222, if the building boundary closed polygon 220 showing the building contour of the building 221 is used as is, there is a risk that a part of the building 221 such as a lobby will be located outside the building boundary 222.

[0047] In this modification, the building boundary closed polygon 220 of the building boundary 222 is enlarged to an enlarged building boundary closed polygon 220' as shown in Fig. 16, and the enlarged building boundary closed polygon 220' is used to extract the building point cloud 304a from the point cloud map 301. This allows the point cloud 304 obtained by measuring building parts that protrude from the building boundary 222, such as the lobby, to be included in the building boundary closed polygon 220' and extracted as the point cloud 304a of the building 221.

[0048] The boundary contour extraction unit 140 performs a process of enlarging the building boundary closed curved polygon 220 of the building boundary 222 created from the survey map 201, centered on the centroid. The building point cloud extraction unit 150 extracts a point cloud that exists within the building boundary closed curved polygon 220' of the building boundary 222 enlarged from the point cloud map 301, and classifies the extracted point cloud into a building point cloud 304a.

[0049] As a method for enlarging the building boundary closed polygon 220, for example, a vector in the direction of each vertex B1 to B6 starting from the centroid B0 of the building boundary closed polygon 220 is calculated and multiplied by a constant. The constant is determined depending on the size of a lobby or the like that extends beyond the building boundary 222, but is set to approximately 1.2 to 1.5 times. The building point cloud extraction unit 150 performs an inside / outside determination as to whether the points are present inside or outside the building boundary closed polygon 220' formed by the point sequence B having the points B1' to B6', and determines whether the point cloud within the labeling range 310 is the point cloud of the building 221.

[0050] FIG. 17 is a flowchart showing the process of the road point cloud extraction unit, and FIG. 18 is an image showing, on a survey map, the state in which the points contained within the road boundary closed polygon have been removed from the points on the ground. The road point cloud extraction unit 160 acquires the point cloud 302 within the labeling range 310 of the point cloud map 301 from the point cloud map area extraction unit 130 (S1010). Then, similar to the process of S1120 shown in FIG. 15 described above, the point cloud within the labeling range 310 is separated into the ground point cloud 303 and the ground object point cloud 304 by a known method (S1020). Then, the point cloud 303a included in the road boundary closed polygon 210 is removed from the ground point cloud 303 (S1030). From the ground point cloud 303, the point cloud 303a included in the road boundary closed polygon 210 is removed, and the remaining point cloud is extracted, and labeling is performed to classify the extracted point cloud into the road point cloud 303b (S1040).

[0051] Figure 18(1) shows the area in which the ground surface point cloud 303 is located on the survey map 201, and Figure 18(2) shows the area in which the road point cloud 303b remaining after removing the point cloud 303a contained within the road boundary closed polygon 210 from the ground surface point cloud 303 is located.

[0052] If the road boundary 212 is regarded as line information based on a sequence of points, it is difficult to determine which side is the road based only on the line information of the road boundary 212. In this embodiment, by connecting a plurality of road boundaries 212 to create a road boundary closed polygon 210, which is a closed curve of the road boundaries 212, it is possible to determine that the inside of the road boundary closed polygon 210 is a non-road and the outside of the road boundary closed polygon 210 is the road 211. Therefore, within the labeling range 310, the area outside the road boundary closed polygon 210 can be regarded as a road area, and the point cloud present in that road area can be determined to be the road point cloud 303b.

[0053] FIG. 19 is a flowchart showing the process of the point cloud reclassification unit. The point cloud reclassification unit 170 acquires a point cloud map 301 having a labeled point cloud (S1210). The point cloud map 301 has a point cloud to which the labels of roads 211 and buildings 221 have been added by labeling. The point cloud reclassification unit 170 creates a classifier using the acquired labeled point cloud as training data for machine learning (S1220). Then, the entire point cloud to be classified is input to the classifier to reclassify the point cloud, and the reclassified labeled point cloud is output (S1230).

[0054] FIG. 20 is an explanatory diagram illustrating the effect of the point cloud reclassification unit. In the boundary contour extraction unit 140 of this embodiment, the road boundary closed curved polygon 210 and the building boundary closed polygon 220 are created by comparing the survey map 201 and the point cloud map 301, so if the survey map 201 does not have data on a new road 211A or a new building 221A, for example, it is not possible to create the road boundary closed curved polygon 210 or the building boundary closed polygon 220. Therefore, it is not possible to label the point clouds of the road 211A and the building 221A as they are.

[0055] In this embodiment, the point cloud is reclassified by the point cloud reclassification unit 170. The point cloud reclassification unit 170 inputs the entire point cloud map 301 to be labeled to a classifier in an unlabeled state, reclassifies the point cloud 302, and outputs it as a labeled point cloud map.

[0056] The point cloud reclassification unit 170 uses a classifier that uses the point cloud map 301 in which the point cloud 302 in the labeling range 310 is labeled as training data to determine the feature amount of the point cloud in the labeling range 310 that is not labeled.

[0057] For example, if there is a difference between the survey map 201 and the point cloud map 301, and a new road 211A or building 221A that is not on the survey map 201 actually exists, but the point cloud for these roads exists on the point cloud map 301, there will be similarities in color and brightness between the point clouds of the new road 211A or building 221A and the point clouds of the surrounding roads 211 and buildings 221 that exist within the same labeling range 310.

[0058] Therefore, the point cloud reclassification unit 170 can classify the points into point clouds of the new road 211A and the new building 221A using a discriminator, and can also perform labeling for the new road 211A and the new building 221A for which there was no data in the survey map 201.

[0059] According to the map information processing device 100 of this embodiment, a road boundary closed polygon 210 is created from a survey map 201, and the remaining points are extracted after removing the points 303a in the road boundary closed polygon 210 from the points 302 of the point cloud map 301, and the extracted points are classified into a road point cloud 303b. Therefore, labeling of roads 211 can be easily performed on the point cloud map 301.

[0060] In addition, a building boundary closed polygon 220 is created from the survey map 201, a point cloud within the building boundary closed polygon 220 is extracted from the point cloud 302 of the point cloud map, and the extracted point cloud is classified into a building point cloud 304a. Therefore, labeling of buildings 221 can be easily performed on the point cloud map 301.

[0061] In this embodiment, an example has been described in which the map information processing device 100 has a point cloud reclassification unit 170, but the configuration may also omit the point cloud reclassification unit 170, and the same effect of easily performing labeling can be obtained.

[0062] [Second embodiment] FIG. 21 is an explanatory diagram illustrating the second embodiment. In this embodiment, first, a point cloud is classified within the teacher data creation range 311 where the road boundary closed curved polygon 210 can be created, and the classified point cloud is used as teacher data for machine learning by a classifier. Next, the classifier is used to classify point clouds in a wider labeling range 310 including the teacher data creation range 311, thereby performing labeling to classify point clouds within an area where the road boundary closed curved polygon 210 cannot be created into a road point cloud.

[0063] As shown in Figure 21, when there is a road 211C within the labeling range 310 for which a road boundary closed polygon 210 cannot be created, such as a single road, the boundary contour extraction unit 140 narrows down the labeling range 310 to a part of the teacher data creation range 311 where the road boundary closed polygon 210 can be created, and creates the road boundary closed polygon 210 within that teacher data creation range 311.

[0064] The road point cloud extraction unit 160 extracts point clouds that exist within the teacher data creation range 311 of the point cloud map 301 by determining whether the point clouds exist inside or outside the road boundary closed polygon 210, and classifies them into a structure point cloud. As shown in Fig. 18(2), the road point cloud extraction unit 160 removes the point clouds that are included within the road boundary closed polygon 210, extracts the point clouds that remain within the teacher data creation range 311, and performs labeling to classify the extracted point clouds into the road point cloud of the road 211B.

[0065] The point cloud reclassification unit 170 acquires a point cloud map 301 having a labeled point cloud of the road 211B, and uses the point cloud as training data for machine learning in a classifier. Then, the entire point cloud within the labeling range 310 for classification is input to the classifier as training data, and reclassification is performed on a wider range of point clouds including the training data creation range 311.

[0066] Roads 211C and 211B are located close to each other, and there is similarity in color and brightness between the point cloud of road 211C and the point cloud of road 211B that exists within the teacher data creation range 311. After creating a classifier, the point cloud reclassification unit 170 performs classification by machine learning using a point cloud in a wider labeling range 310 including the teacher data creation range 311 as an input. Therefore, the point cloud of road 211C in the portion where the road boundary closed polygon 210 cannot be created can be classified as a road point cloud based on the similarity with the point cloud of road 211B.

[0067] [Third embodiment] FIG. 22 is an explanatory diagram illustrating the third embodiment, showing a labeling range 310 of a survey map 201. As shown in FIG.

[0068] This embodiment is characterized in that when the survey map 201 in the survey map database 200 includes boundary information of the road 211D and boundary information of the sidewalk 211E as attribute information of the road 211, the point cloud 303b of the road 211 is classified into a point cloud of the road 211D (roadway point cloud) and a point cloud of the sidewalk 211E (sidewalk point cloud).

[0069] The survey map acquisition unit 110 acquires a survey map 201 including information on the attributes of the road 211, and the boundary contour extraction unit 140 creates a road boundary closed polygon 210 for each attribute of the road 211 from the survey map 201. For example, if the road 211 has the attributes of a roadway 211D and a sidewalk 211E, a road boundary closed polygon for the roadway 211D including the sidewalk 211E and a road boundary closed polygon for only the roadway 211D are created. Then, the road point cloud extraction unit 160 extracts a point cloud for each attribute of the road 211 using the road boundary closed polygon 210 created for each of these attributes.

[0070] The survey map acquisition unit 110 acquires a survey map 201 having attribute information indicating the road edge contour of the road 211 and the type of the road 211 including at least the distinction between a roadway 211D and a sidewalk 211E.

[0071] The boundary contour extraction unit 140 creates a road boundary closed polygon with the same attributes from the survey map 201. The road point cloud extraction unit 160 extracts the remaining points after removing the points inside the road boundary closed polygon from the point cloud map 301, and classifies them into a road point cloud.

[0072] FIG. 23 is a flowchart showing the process of the road point cloud extraction unit in the third embodiment. The road point cloud extraction unit 160 acquires the point cloud 302 within the labeling range 310 of the point cloud map 301 from the point cloud map area extraction unit 130 (S1210). Then, similar to the processing of S1120 shown in Fig. 15 described above, the road point cloud extraction unit 160 separates the point cloud within the point cloud acquisition area 320 into a ground surface point cloud 303 and a ground object point cloud 304 by a known method (S1220).

[0073] Then, the boundary contour extraction unit 140 extracts point sequences of all road boundaries 212 whose attributes are roadway 211D and sidewalk 211E from the survey map 201 to create a first road boundary closed polygon (first road boundary closed curve), and the road point cloud extraction unit 160 removes the points within the first road boundary closed polygon from all point clouds 303 on the earth's surface, and extracts the remaining points as the first road point cloud (S1230).

[0074] Next, the boundary contour extraction unit 140 extracts a point sequence excluding the point sequence of the road boundary of the sidewalk 211E from the point sequence of all the road boundaries 212 described above to create a second road boundary closed polygon (second road boundary closed curve), and the road point cloud extraction unit 160 removes the point cloud within the second road boundary closed polygon from the point cloud 303 of the ground surface and extracts the remaining point cloud as a second road point cloud (S1240). Then, the difference between the first road point cloud and the second road point cloud is extracted, and labeling is performed to classify the difference point cloud as a sidewalk point cloud (S1250).

[0075] According to the map information processing apparatus 100 of this embodiment, the points in the point cloud map 301 can be classified into a road point cloud, and the road point cloud can be further classified into a roadway point cloud and a sidewalk point cloud.

[0076] Although the embodiments of the present invention have been described above in detail, the present invention is not limited to the above-described embodiments, and various design changes can be made without departing from the spirit of the present invention described in the claims. For example, the above-described embodiments have been described in detail to clearly explain the present invention, and the present invention is not necessarily limited to those having all of the configurations described. In addition, it is possible to replace a part of the configuration of one embodiment with the configuration of another embodiment, and it is also possible to add the configuration of another embodiment to the configuration of one embodiment. Furthermore, it is possible to add, delete, or replace a part of the configuration of each embodiment with another configuration. [Explanation of symbols]

[0077] 100...map information processing device, 110...survey map acquisition unit, 120...point cloud map acquisition unit, 130...point cloud map area extraction unit, 140...boundary contour extraction unit, 150...building point cloud extraction unit (structure point cloud extraction unit), 160...road point cloud extraction unit (structure point cloud extraction unit), 170...point cloud reclassification unit, 201...survey map, 210...road boundary closed polygon (structure boundary closed curve), 211...road, 211A...road, 211B...road, 211C...road, 212...road boundary, 212A...building, 212i to 212j...road boundary, 220...building boundary closed polygon (structure boundary closed curve), 221...building, 222...building boundary, 301...point cloud map, 302...point cloud, 303...point cloud, 303b...road point cloud, 304a...building point cloud, 310...labeling range

Claims

1. A point cloud map acquisition unit that acquires a point cloud map; a survey map acquisition unit for acquiring a survey map including boundary information of a structure; a boundary contour extraction unit for generating a structure boundary closed curve from the survey map; a structure point cloud extraction unit that extracts a point cloud by determining whether a point cloud included in the point cloud map exists inside or outside the structure boundary closed curve, and classifies the extracted point cloud into a structure point cloud; A map information processing apparatus comprising:

2. the survey map acquisition unit acquires a survey map including information on a road edge contour of a road, which is a first structure; the boundary contour extraction unit creates a road boundary closed curve from the survey map; 2. The map information processing device according to claim 1, wherein the structure point cloud extraction unit extracts a remaining point cloud after removing the points within the road boundary closed curve from the point cloud of the point cloud map, and classifies the extracted point cloud into a road point cloud.

3. The survey map acquisition unit acquires a survey map including a building outline of a building which is a second structure, the boundary contour extraction unit creates a building boundary closed curve from the survey map; 3. The map information processing device according to claim 2, wherein the structure point cloud extraction unit extracts a point cloud that exists within the building boundary closed curve from the point cloud of the point cloud map, and classifies the extracted point cloud into a building point cloud.

4. The boundary contour extraction unit enlarges the building boundary closed curve created from the survey map with a centroid of the building boundary closed curve as a center, 4. The map information processing apparatus according to claim 3, wherein the structure point cloud extraction unit extracts a point cloud existing within the expanded building boundary closed curve, and classifies the extracted point cloud into the building point cloud.

5. The map information processing device according to claim 1, further comprising a point cloud reclassification unit that creates a classifier through machine learning using the point cloud extracted by the structure point cloud extraction unit as training data, and reclassifies the point cloud of the point cloud map using the classifier.

6. the structure point cloud extraction unit extracts a point cloud by determining whether a point cloud existing within a teaching data creation range of the point cloud map exists inside or outside the structure boundary closed curve, and classifies the point cloud into a structure point cloud; The map information processing device according to claim 5, characterized in that the point cloud reclassification unit trains the classifier on machine learning of the point cloud extracted from within the training data creation range as training data, and uses the classifier to reclassify a wider range of point clouds including the training data creation range of the point cloud map.

7. the survey map acquisition unit acquires a survey map including information on attributes of the structure; the boundary contour extraction unit creates a structure boundary closed curve for each attribute of the structure from the survey map; 2. The map information processing apparatus according to claim 1, wherein the structure point cloud extraction unit extracts the structure point cloud by using the structure boundary closed curve created for each attribute.

8. the survey map acquisition unit acquires a survey map having attribute information indicating a road edge contour of a road, which is a first structure, and a road type including at least a distinction between a roadway and a sidewalk; the boundary contour extraction unit creates a road boundary closed curve having the same attributes from the survey map; 8. The map information processing apparatus according to claim 7, wherein the structure point cloud extraction unit extracts the remaining points after removing the points within the road boundary closed curve from the point cloud map, and classifies the points into a road point cloud.

9. the boundary contour extraction unit extracts from the survey map a sequence of points of all road boundaries whose attributes are roadway and sidewalk, to create a first road boundary closed curve, and extracts a sequence of points excluding a sequence of points of a road boundary of a sidewalk from the sequence of points of all the road boundaries, to create a second road boundary closed curve; The map information processing device according to claim 8, characterized in that the structure point cloud extraction unit removes the points within the first road boundary closed curve from the point cloud map and extracts the remaining points as a first road point cloud, removes the points within the second road boundary closed curve from the point cloud map and extracts the remaining points as a second road point cloud, extracts a difference between the first road point cloud and the second road point cloud, and classifies the difference point cloud into a sidewalk point cloud.

Citation Information

Patent Citations

  • Information processing device

    JP2022089828A