Feature detection device and feature detection program

Abstract

To provide a feature detection device and a feature detection program that appropriately detect a feature.SOLUTION: A feature detection device includes: point group data projection means 103 that sets virtual planes perpendicular to a traveling direction of a trolley (moving body) 3 at a predetermined interval and projects point group data existing between the set virtual planes onto a virtual plane; point group density calculation means 104 that calculates density of point group data per unit area on the projected virtual plane as point group density; and detection means 105 that deletes point group data whose point group density is equal to or higher than a predetermined point group threshold as a wall point group, and detects point group data after the wall point group deletion as an equipment point group.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a feature detection device and a feature detection program for appropriately detecting features. [Background technology]

[0002] In recent years, moving objects such as automobiles and trains have been equipped with laser acquisition devices that emit laser pulses, making it possible to acquire the three-dimensional coordinates (point cloud data) of features in front of or behind the moving object.

[0003] Therefore, there was a strong desire to utilize 3D coordinates (point cloud data) acquired by a laser acquisition device to automatically detect equipment installed on walls, tunnels, etc., and to use this information for equipment inspection.

[0004] Patent Document 1 discloses a data analysis device that divides a space near a cross section, which has a preset width in the depth direction perpendicular to the cutting plane, into a grid to set multiple subspaces, searches for an edge registration area that satisfies predetermined conditions within the cutting plane for each subspace, obtains a direction line that follows the distribution within the cutting plane of the point cloud that belongs to the edge registration area as an edge, and tracks and connects this edge in the direction of travel to detect the cross-sectional shape of a tunnel. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Japanese Patent Application Laid-Open No. 2014-190962 Summary of the Invention [Problem to be solved by the invention]

[0006] However, in the technology described in Patent Document 1, direction lines along the distribution of the point cloud belonging to the edge registration area within the cutting plane are found as edges, and these edges are tracked and connected in the direction of travel to detect the cross-sectional shape of the tunnel. Therefore, if the upper part is open, for example, like a wall surface, that is, if it is not continuous on the cutting plane in the direction of travel of the moving body, it is not possible to find the edges and it is difficult to detect the wall surface.

[0007] Furthermore, with the technology described in Patent Document 1, when equipment is installed on the surface of a tunnel, it may be detected as the surface of the tunnel.

[0008] The present invention has been made in consideration of the above-mentioned problems, and has an object to provide a feature detection device and a feature detection program that can appropriately detect features. [Means for solving the problem]

[0009] In order to achieve the above object, a first feature of the feature detection device according to the present invention is to A feature detection device that acquires point cloud data from a point cloud data generation device that generates three-dimensional coordinates of features around a moving object as point cloud data based on reflected light of laser light irradiated onto the features, and detects the features based on the acquired point cloud data, a point cloud data projection means for setting virtual planes perpendicular to a direction of travel of the moving body at predetermined intervals and projecting the point cloud data existing between the set virtual planes onto the virtual planes; a point cloud density calculation means for calculating a density of the point cloud data per unit area on the projected virtual plane as a point cloud density; a detection means for deleting the point cloud data whose point cloud density is equal to or greater than a predetermined point cloud threshold as a wall point cloud, and detecting the point cloud data after the wall point cloud deletion as an equipment point cloud; The reason is that it is equipped with the following.

[0010] A second feature of the feature detection device according to the present invention is The detection means The point cloud data whose point cloud density is equal to or greater than a predetermined point cloud threshold is determined as a wall candidate point cloud, When a straight line connecting a starting point near the irradiation point of the laser light on the virtual plane to the point cloud data intersects with the wall candidate point cloud, the point cloud data intersecting with the wall candidate point cloud and the wall candidate point cloud are deleted as the wall point cloud; When the straight line does not intersect with the wall candidate point cloud, the point cloud data that does not intersect with the wall candidate point cloud is detected as the facility point cloud.

[0011] A third feature of the feature detection device according to the present invention is The starting point is a position on the center line of the moving body in the left-right direction and a predetermined distance away from the ground surface.

[0012] A fourth feature of the feature detection device according to the present invention is The system further comprises an equipment point cluster identification means for segmenting the equipment point cluster in which the distance between each equipment point is less than a predetermined threshold value and identifying it as an equipment point cluster.

[0013] A fifth feature of the feature detection device according to the present invention is a storage means for storing equipment characteristic conditions indicating the characteristics of the equipment for each equipment name of the equipment; The system further includes an equipment identification means for extracting equipment characteristic conditions for each equipment point collection identified by the equipment point collection identification means, and identifying the equipment name of the equipment point collection by comparing the extracted equipment characteristic conditions with the equipment characteristic conditions stored in the storage means.

[0014] A sixth feature of the feature detection device according to the present invention is The facility characteristic conditions stored for each facility name of the facility are stored in accordance with the traveling direction of the moving object, The facility identification means For each facility point aggregate identified by the facility point aggregate identifying means, a facility characteristic condition is extracted in accordance with the traveling direction of the mobile object.

[0015] A first feature of the feature detection program according to the present invention is: A feature detection program executed by a feature detection device that acquires point cloud data from a point cloud data generation device that generates three-dimensional coordinates of features around a moving object as point cloud data based on reflected light of laser light irradiated onto the features, and detects the features based on the acquired point cloud data, a point cloud data projection step of setting virtual planes perpendicular to a traveling direction of the moving body at predetermined intervals and projecting the point cloud data existing between the set virtual planes onto the virtual planes; a point cloud density calculation step of calculating a density of the point cloud data per unit area on the projected virtual plane as a point cloud density; a detection step of deleting the point cloud data whose point cloud density is equal to or greater than a predetermined point cloud threshold as a wall point cloud, and detecting the point cloud data after the wall point cloud deletion as an equipment point cloud; The object is to have the feature detection device execute the above. [Effects of the Invention]

[0016] According to the feature detection device and feature detection program of the present invention, features can be detected appropriately. [Brief explanation of the drawings]

[0017] [Figure 1] 1 is a schematic diagram showing the schematic configuration of a feature detection system according to an embodiment of the present invention; [Figure 2] FIG. 1 is an explanatory diagram that schematically explains the generation of point cloud data by a point cloud data generation device provided in a feature detection system that is one embodiment of the present invention. [Figure 3] FIG. 2 is an explanatory diagram that schematically explains projection by a point cloud data projection means of the feature detection device provided in the feature detection system that is one embodiment of the present invention. [Figure 4]1A and 1B are diagrams showing point cloud data before and after projection by a point cloud data projection means of a feature detection device provided in a feature detection system according to an embodiment of the present invention, in the case where a carriage travels on rails laid inside a tunnel, where (a) shows the point cloud data before projection and (b) shows the point cloud data after projection. [Figure 5] 1 is an explanatory diagram illustrating the detection of a wall point cloud and a facility point cloud by the detection means of a feature detection device 1 provided in a feature detection system according to one embodiment of the present invention. FIG. [Figure 6] 1A and 1B are explanatory diagrams illustrating the identification of an equipment point cloud group by an equipment point cloud group identification means of a feature detection device provided in a feature detection system according to one embodiment of the present invention, in which (a) is a diagram showing a wall point cloud and an equipment point cloud, and (b) is a diagram showing only the equipment point cloud with the wall point cloud removed. [Figure 7A] 1 is a flowchart showing processing performed by a feature detection device included in a feature detection system according to an embodiment of the present invention. [Figure 7B] 1 is a flowchart showing processing performed by a feature detection device included in a feature detection system according to an embodiment of the present invention. [Figure 8] 10 is a flowchart showing the processing contents of a facility name identification process performed by a feature detection device included in a feature detection system according to one embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0018] Hereinafter, embodiments of the present invention will be described with reference to the drawings. The same or equivalent parts and components are designated by the same or equivalent reference numerals throughout the drawings. However, it should be noted that the drawings are schematic and may differ from the actual product. Furthermore, the drawings may include parts with different dimensional relationships and ratios.

[0019] Furthermore, the embodiments shown below are merely examples of devices that embody the technical concept of the present invention, and the technical concept of the present invention does not limit the arrangement of each component to that shown below. Various modifications can be made to the technical concept of the present invention within the scope of the claims.

[0020] A feature detection system including a feature detection device according to one embodiment of the present invention will be described below.

[0021] FIG. 1 is a schematic diagram showing the schematic configuration of a feature detection system according to one embodiment of the present invention.

[0022] As shown in FIG. 1, the feature detection system 100 includes a feature detection device 1, a point cloud data generation device 2, a cart (mobile body) 3, and a monitor 5.

[0023] The point cloud data generating device 2 is mounted on a carriage 3 that moves along rails 4. While the carriage 3 is moving, the point cloud data generating device 2 irradiates laser light onto features around the carriage 3. The point cloud data generating device 2 then generates three-dimensional coordinates of the features as point cloud data based on the reflected light of the irradiated laser light. Details will be described later. Note that features refer to all objects on the ground, such as tunnels, traffic lights, signs, trees, etc.

[0024] The feature detection device 1 acquires point cloud data from the point cloud data generation device 2, and detects features based on the acquired point cloud data.

[0025] The monitor 5 displays various screens, such as a screen that displays features detected by the feature detection device 1.

[0026] <Generation of point cloud data by point cloud data generator 2> FIG. 2 is an explanatory diagram that schematically explains the generation of point cloud data by the point cloud data generation device 2 provided in the feature detection system 100 according to one embodiment of the present invention.

[0027] As shown in FIGS. 1 and 2, the point cloud data generating device 2 includes a laser scanner 21, a GPS (Global Positioning System) 22, an IMU (Inertial Measurement Unit) 23, a data collecting means 24, and a data temporary storage unit 25.

[0028] The laser scanner 21 is provided at the rear of the carriage 3 and rotates to emit laser light in a 360-degree range. The reach of the emitted laser light is in the range of 80 m to 100 m, and for example, the laser light is emitted in a 360-degree range at a frequency of 200 Hz. The laser scanner 21 receives reflected light of the emitted laser light and obtains laser data (Time, θ, L) that indicates the measurement time of reception and the angle and distance to the feature based on the received reflected light.

[0029] The GPS 22 acquires GPS data (position coordinates) using signals emitted from artificial satellites.

[0030] The IMU 23 can detect the attitude of the point cloud data generation device 2. The attitudes acquired by the IMU 23 are matched with GPS data and supplied to the data collection means 24 as POS (Position and Orientation System) data. That is, the POS data includes data (Time, xi, yi, zi, Roll, Pitch, Yaw) that indicates the reference position and attitude (measurement time, position coordinates, tilt) of the IMU 23, which is a combination of GPS data (position coordinates) from the GPS 22 and the detected attitude. Zi is the altitude value from ground level at which the IMU 23 is attached.

[0031] The position coordinates of the laser scanner 21 relative to the IMU 23 can be expressed as (Δx, Δy, Δz).

[0032] The data collection means 24 generates a collection of three-dimensional coordinates indicating the positions of features as point cloud data (X, Y, Z) based on the POS data acquired from the IMU 23 and the laser data acquired from the laser scanner 21.

[0033] The data temporary storage unit 25 temporarily stores the point cloud data (X, Y, Z) generated by the data collection means 24 in association with movement data (D, V) of the traveling direction D and speed V of the carriage 3. The stored point cloud data and movement data are supplied to the feature detection device 1. Here, an example will be described in which the point cloud data and movement data stored in the data temporary storage unit 25 are supplied from the data temporary storage unit 25 to the feature detection device 1 when the feature detection device 1 and the point cloud data generation device 2 are connected, but this is not limiting. For example, the data may be supplied to the feature detection device 1 in real time via a wireless network.

[0034] <Configuration of feature detection device 1> Returning to Figure 1, the feature detection device 1 has a point cloud data acquisition means 101, a point cloud data storage unit 102, a point cloud data projection means 103, a point cloud density calculation means 104, a detection means 105, an equipment point cluster identification means 106, an equipment characteristic condition storage unit 108, an equipment identification means 109, an equipment data storage unit 110, and a display control means 111.

[0035] The point cloud data acquisition means 101 acquires the point cloud data and movement data supplied from the temporary data storage unit 25 of the point cloud data generation device 2, and stores the acquired point cloud data and movement data in the point cloud data storage unit .

[0036] The point cloud data storage unit 102 is configured by, for example, a hard disk drive, and stores point cloud data (X, Y, Z) and movement data (D, V).

[0037] The point cloud data projection means 103 sets virtual planes perpendicular to the traveling direction of the dolly 3 at predetermined intervals, and projects point cloud data existing between the set virtual planes onto the virtual planes.

[0038] 3 is an explanatory diagram that schematically explains the projection by the point cloud data projection means 103. For the sake of simplicity, this figure shows point cloud data consisting of three pieces of point data indicating the coordinates of measurement points 211, 212, and 213 plotted in three-dimensional space.

[0039] 3, imaginary planes are provided successively at intervals of M so as to be perpendicular to the traveling direction of the carriage 3. Here, imaginary planes 201 and 202 are shown.

[0040] Point data of measurement points 211, 212, and 213, which are point cloud data existing between the set virtual planes 201 and 202, is projected onto the virtual plane 201 to become point data of projected measurement points 211A, 212A, and 213A.

[0041] 4 is a diagram showing the point cloud data before and after projection by the point cloud data projection means 103 when the carriage 3 runs along the traveling direction on rails 4 laid inside a tunnel. (a) shows the point cloud data before projection, and (b) shows the point cloud data after projection.

[0042] In FIG. 4(a), an imaginary plane 201 perpendicular to the traveling direction is provided, and an imaginary plane 202 is provided at a position M behind the imaginary plane 201 in the traveling direction.

[0043] FIG. 4(a) includes point cloud data 301 of the tunnel wall surface, point cloud data 302 of the facilities installed on the tunnel inner wall, and the like.

[0044] When this point cloud data is projected onto the virtual plane 201, areas with high and low point cloud density appear, as shown in Fig. 4(b). Specifically, point cloud data 301 of the tunnel wall surface is plotted continuously along the direction of travel, so when projected onto the virtual plane 201, the density of the point cloud data becomes high at projection point 401 corresponding to point cloud data 301.

[0045] On the other hand, since the point cloud data 302 of the equipment is plotted discontinuously along the direction of travel, when projected onto the virtual plane 201, the density of the point cloud data at the projection point 402 corresponding to the point cloud data 302 is lower than that at the projection point 401.

[0046] By utilizing this difference in density of point cloud data, it is possible to separate point cloud data that indicates tunnels, wall surfaces, etc. from point cloud data that indicates facilities.

[0047] The point cloud density calculation means 104 calculates the density of point cloud data per unit area on the projected virtual plane as the point cloud density. Here, the point cloud density on the virtual plane 201 is calculated, but the point cloud densities are calculated for all the projected virtual planes.

[0048] The detection means 105 determines, as a wall candidate point cloud, point cloud data whose point cloud density calculated by the point cloud density calculation means 104 is equal to or greater than a predetermined point cloud threshold Th1. The wall candidate point cloud is a candidate point cloud of point cloud data indicating a tunnel, a wall, or the like.

[0049] When a straight line connecting the starting point on the virtual plane to the point data intersects with the wall candidate point cloud, the detection means 105 detects the intersecting point data and the wall candidate point cloud as a wall point cloud. On the other hand, when a straight line connecting the starting point near the irradiation point of the laser light on the virtual plane to the point data does not intersect with the wall candidate point cloud, the detection means 105 detects the point data that does not intersect with the wall candidate point cloud as an equipment point cloud.

[0050] FIG. 5 is an explanatory diagram illustrating the detection of the wall point cloud and the facility point cloud by the detection means 105.

[0051] As shown in FIG. 5, a starting point R is set at a position on the center line S in the left-right direction of the carriage 3 and a predetermined distance N away from the ground surface G.

[0052] Then, when a straight line is drawn from the starting point R to each point data (for example, point P) of the point cloud data, it is determined whether or not this straight line intersects with the wall candidate point cloud 501. For example, if the distance between the starting point R and point P is longer than the distance from the starting point R to point Q, it can be determined that the straight line intersects with the wall candidate point cloud 501. Here, point Q is on the wall candidate point cloud 501 and is the closest point to point R.

[0053] Point data where a line connecting from the starting point R intersects with the wall candidate point group 501 can be assumed to be a recessed portion provided on the wall surface, and therefore this point data is detected together with the wall candidate point group 501 as a wall point group.

[0054] On the other hand, for point data for which the line connecting from the starting point R does not intersect with the wall candidate point cloud 501, it can be assumed that the point data is equipment installed inside the wall surface, and therefore this point data is detected as an equipment point cloud.

[0055] In this way, by determining whether or not the straight line to the starting point R for each piece of point data intersects with the wall candidate point cloud 501, the point cloud data can be divided into a wall point cloud and a facility point cloud.

[0056] The facility point aggregate identification means 106 segments the facility point group in which the distance between each facility point is less than a predetermined threshold value, and identifies the facility point group as a facility point aggregate.

[0057] Here, the detection means 105 sets point cloud data whose point cloud density calculated by the point cloud density calculation means 104 is equal to or greater than a predetermined point cloud threshold Th1 as a wall candidate point cloud, and when a straight line connecting a starting point on the virtual plane to the point data intersects with the wall candidate point cloud, detects this intersecting point data and the wall candidate point cloud as a wall point cloud, but these processes may be simplified if it is known in advance that there is no recess on the wall surface. Specifically, the detection means 105 may delete point cloud data whose point cloud density is equal to or greater than a predetermined point cloud threshold as a wall point cloud, and detect the point cloud data after the wall point cloud deletion as an equipment point cloud.

[0058] 6 is an explanatory diagram illustrating the identification of an equipment point group by the equipment point group identification means 106. (a) is a diagram showing a wall point group 601 and an equipment point group 602, and (b) is a diagram showing only the equipment point group 602 with the wall point group 601 removed.

[0059] As shown in FIG. 6(a), a wall point group 601 and a facility point group 602 appear.

[0060] Then, when the wall point cloud 601 is deleted from FIG. 6(a), the facility point cloud 602 remains as shown in FIG. 6(b).

[0061] The facility point cluster identification means 106 calculates the distances between each of the point cloud data included in the facility point cloud 602. Then, the facility point cluster identification means 106 divides the point cloud data in which the distance between each of the facility points is less than a predetermined threshold Th2 into one segment, and identifies each divided segment as a facility point cluster. A facility cluster is a collection of facility point cloud data that indicates a single facility. For example, if the distance between point data P1 in the facility point cloud data and the closest point data P2 is equal to or greater than the threshold Th2, it can be inferred that the point data P1 and the point data P2 are the ends of the respective facilities that constitute the facility, and therefore the point data P1 and the point data P2 are segmented as different facility clusters. Furthermore, if the distance between point data P1 and point data P3 in the equipment point cloud data is less than threshold Th2, and the distance between point data P3 and point data P4 is less than threshold Th2, it can be inferred that point data P1, point data P3, and point data P4 constitute the same equipment, and therefore point data P1, point data P3, and point data P4 are segmented as the same equipment group.

[0062] In this way, by segmenting the equipment point group where the distance between each equipment point is less than a predetermined threshold value Th2 and identifying it as an equipment point collection, the equipment point group 602 can be segmented into equipment point collection 602a and equipment point collection 602b, as shown in Figure 6(b).

[0063] The equipment characteristic condition storage unit 108 is configured, for example, with a hard disk drive or the like, and stores equipment characteristic conditions that indicate the characteristics of the equipment by matching the traveling direction of the bogie 3 for each equipment name such as a traffic light, a gradient marker, an ATS ground coil, etc. For example, the equipment characteristic conditions for a traffic light include (1) the overall shape is a column, (2) the height of the entire column is within a predetermined height range, (3) the column has continuity, (4) there is an attached object at a predetermined position from the ground level, (5) the attached object has the shape of a light box, and (6) the center of gravity of the entire column and the light box are located in a predetermined positional relationship.

[0064] The equipment identification means 109 extracts equipment characteristic conditions for each equipment point collection identified by the equipment point collection identification means 106 in accordance with the traveling direction of the trolley 3. For example, the equipment characteristic conditions extracted include (1) the overall shape is columnar, (2) the height of the entire column is within a predetermined height range, (3) the column has continuity, (4) there is an attached object at a predetermined position from the ground level, (5) the attached object has the shape of a light box, and (6) the center of gravity of the entire columnar object and the light box are located in a predetermined positional relationship.

[0065] Then, the equipment identification means 109 identifies the equipment name of the equipment point collection by comparing the extracted equipment characteristic conditions with the equipment characteristic conditions stored in the equipment characteristic condition storage unit 108. For example, if the extracted equipment characteristic conditions satisfy equipment characteristic conditions (1) to (6) of traffic lights stored in the equipment characteristic condition storage unit 108, the equipment identification means 109 identifies this equipment point collection as a traffic light.

[0066] The equipment can be easily identified by storing the equipment characteristic conditions in the equipment characteristic condition storage unit 108 in accordance with the traveling direction of the carriage 3, and by the equipment identification means 109 extracting the equipment characteristic conditions in accordance with the traveling direction of the carriage 3. Details of the identification of each piece of equipment will be described later.

[0067] The facility data storage unit 110 is configured by, for example, a hard disk drive, and stores the explanatory name identified by the facility identification means 109 in association with the point cloud data as facility data.

[0068] The display control means 111 displays the point cloud data on the monitor 5 based on the equipment data stored in the equipment data storage unit 110.

[0069] <Function of the feature detection device 1> 7A and 7B are flowcharts showing the processing performed by the feature detection device 1 included in the feature detection system according to one embodiment of the present invention.

[0070] As shown in Figures A and B, when feature detection is requested from an input device (not shown) in the feature detection device 1 (step S101; YES), the point cloud data projection means 103 sets virtual planes perpendicular to the traveling direction of the cart 3 at predetermined intervals (step S103).

[0071] Then, the point cloud data projection means 103 projects the point cloud data existing between the set virtual planes onto the virtual plane (step S105).

[0072] The point cloud density calculation means 104 calculates the density of the point cloud data per unit area on the projected virtual plane as the point cloud density (step S107).

[0073] The detection means 105 determines whether the point cloud density calculated by the point cloud density calculation means 104 is equal to or greater than a predetermined point cloud threshold Th1 (step S109).

[0074] In step S109, a point cloud whose point cloud density is determined to be less than the predetermined point cloud threshold Th1 (NO) is set as an equipment point cloud (step S117).

[0075] On the other hand, in step S109, the point clouds whose point cloud density is determined to be equal to or greater than the predetermined point cloud threshold Th1 (YES) are set as wall candidate point clouds (step S111).

[0076] Then, the detection means 105 determines whether or not a straight line connecting the starting point on the virtual plane to each piece of point data in the wall candidate point group intersects with the wall candidate point group (step S113).

[0077] In step S113, if it is determined that the straight line connecting the starting point on the virtual plane to the point data does not intersect with the wall candidate point cloud (NO), the detection means 105 sets the point data as the facility point cloud (step S117).

[0078] On the other hand, if it is determined in step S113 that the straight line connecting the starting point on the virtual plane to the point data intersects with the wall candidate point cloud (YES), the detection means 105 sets the point data as the wall point cloud (step S115).

[0079] Next, the facility point aggregate identification means 106 determines whether or not the distance between each facility point data in the facility point group is less than a predetermined threshold value Th2 (step S119).

[0080] In step S119, if it is determined that the distance between each equipment point data is less than the predetermined threshold Th2 (YES), the equipment point collection identification means 106 sets the equipment point data determined to be less than the predetermined threshold Th2 as an equipment point collection (step S121).

[0081] Then, the facility identification means 109 executes a facility name identification process to identify the facility name of the facility point aggregate (step S123), associates the identified explanatory name with the point cloud data, and stores the result as facility data in the facility data storage unit 110 (step S125). The details of the facility name identification process will be described later.

[0082] <Facility name identification processing> FIG. 8 is a flowchart showing the contents of the facility name identification process performed by the feature detection device 1 included in the feature detection system according to one embodiment of the present invention.

[0083] As shown in Fig. 8, the facility identification means 109 determines whether the facility feature conditions of the facility point set identified by the facility point set identification means 106 match the facility feature conditions of an ATS (Automatic Train Stop) ground coil stored in the facility feature condition storage unit 108 (step S201). An ATS (Automatic Train Stop) ground coil is a type of device that is typically installed on a railroad track and automatically slows down or stops a train in accordance with the signal aspect and track conditions. Since it is installed on a railroad track and has a low height to avoid contact with railway vehicles, the facility feature conditions stored in the facility feature condition storage unit 108 include that it is flat and that the depth and width of the plan are within a predetermined size.

[0084] In step S201, if it is determined that the equipment group matches an ATS ground device (YES), the equipment identification means 109 identifies the equipment group as an ATS ground device (step S203).

[0085] In step S201, if it is determined that there is no match with the ATS ground coil (NO), the equipment identification means 109 extracts equipment collections having the shape of a column from the equipment point collections identified by the equipment point collection identification means 106 (step S205).

[0086] Then, the equipment identification means 109 determines whether the equipment characteristic conditions of the equipment point collection identified by the equipment point collection identification means 106 match the equipment characteristic conditions of the utility pole stored in the equipment characteristic condition memory unit 108 (step S207).

[0087] If it is determined in step S207 that the equipment characteristic conditions for an electric pole are met (YES), the equipment identification means 109 identifies the equipment assembly as an electric pole (step S209).

[0088] If it is determined in step S207 that the equipment characteristic conditions of the utility pole do not match (NO), the equipment identification means 109 determines whether the equipment characteristic conditions of the equipment point collection identified by the equipment point collection identification means 106 match the equipment characteristic conditions of the traffic light stored in the equipment characteristic condition memory unit 108 (step S211).

[0089] If it is determined in step S211 that the equipment characteristic condition of a traffic light is met (YES), the equipment identification means 109 identifies the equipment collection as a traffic light (step S213).

[0090] In step S211, if it is determined that the equipment characteristic conditions of the traffic light are not met (NO), the equipment identification means 109 ends the process. In this case, the equipment will be identified as another piece of equipment by processing in other steps described later.

[0091] If the pole is identified as an electric pole in step S209 or if the traffic light is identified as a traffic light in step S213, it is further determined whether the pole or traffic light satisfies the equipment characteristic conditions of a special attached type (step S215). A special attached type is a type that has a shape attached to the main body as a special signal light.

[0092] If it is determined in step S215 that the equipment characteristic conditions of the accidental attachment type are met (YES), the equipment identification means 109 identifies the equipment assembly as an accidental attachment type (step S217).

[0093] In step S211, if it is determined that the equipment characteristic condition of the spontaneous attachment type is not met (NO), the equipment identification means 109 ends the process.

[0094] Furthermore, the equipment identification means 109 determines whether or not the equipment characteristic conditions of the equipment point collection identified by the equipment point collection identification means 106 match the equipment characteristic conditions of the gradient marker stored in the equipment characteristic condition storage unit 108 (step S221). The equipment characteristic conditions of the gradient marker are set to, for example, (1) the height of the upper end from ground level is less than a predetermined height, (2) the distance between the center and main part of the equipment collection is within a predetermined range, and (3) the angle between the direction of the vector pointing from the main part to the center of the equipment collection and the vector of the traveling direction of the trolley 3 is within a predetermined angle range.

[0095] If it is determined in step S221 that the equipment characteristic conditions of the gradient marker are met (YES), the equipment identification means 109 identifies the equipment collection as a gradient marker (step S223).

[0096] In step S221, if it is determined that the equipment characteristic conditions of the gradient marker are not met (NO), the equipment identification means 109 ends the process.

[0097] Similarly, as shown in steps S225 to S251, the equipment identification means 109 extracts equipment characteristic conditions for each equipment point collection identified by the equipment point collection identification means 106, and identifies the equipment name of the equipment point collection by comparing the extracted equipment characteristic conditions with the equipment characteristic conditions stored in the equipment characteristic condition storage unit 108. Note that a spontaneous freestanding type is a special signal light having a freestanding shape. A stop point is a stop position target.

[0098] As described above, the feature detection device 1 provided in the feature detection system 100, which is one embodiment of the present invention, is equipped with a point cloud data projection means 103 that sets virtual planes perpendicular to the direction of travel of the cart (mobile body) 3 at predetermined intervals and projects point cloud data existing between the set virtual planes onto the virtual planes, a point cloud density calculation means 104 that calculates the density of point cloud data per unit area on the projected virtual plane as a point cloud density, and a detection means 105 that deletes point cloud data whose point cloud density is equal to or greater than a predetermined point cloud threshold as a wall point cloud, and detects the point cloud data after the wall point cloud deletion as an equipment point cloud.

[0099] Therefore, it is possible to detect the wall even if there is an open area on the top, such as a wall, that is, a section that is not continuous on the cutting plane in the direction of travel of the carriage 3 moving body. Also, even if equipment is installed on the surface of a tunnel, it is possible to detect the tunnel wall and the equipment installed on the tunnel wall separately. This allows for proper feature detection.

[0100] The above-described embodiment can also be realized by executing a feature detection program installed on a computer. [Explanation of symbols]

[0101] 1. Feature detection device 2. Point cloud data generator 3. Cart (mobile) 4 Rail 5 monitors 21 Laser scanner 22 GPS 23 IMU 24 Data Collection Methods 25 Temporary data storage unit 100 Feature Detection System 101 Point cloud data acquisition method 102 Point cloud data storage unit 103 Point cloud data projection means 104 Point cloud density calculation means 105 Detection means 106 Equipment point aggregate identification means 108 Equipment characteristic condition memory unit 109 Equipment identification means 110 Equipment data storage unit 111 Display control means

Claims

1. A feature detection device that acquires point cloud data from a point cloud data generation device that generates three-dimensional coordinates of features around a moving object as point cloud data based on reflected light of laser light irradiated onto the features, and detects the features based on the acquired point cloud data, a point cloud data projection means for setting virtual planes perpendicular to a direction of travel of the moving body at predetermined intervals and projecting the point cloud data existing between the set virtual planes onto the virtual planes; a point cloud density calculation means for calculating a density of the point cloud data per unit area on the projected virtual plane as a point cloud density; a detection means for deleting the point cloud data whose point cloud density is equal to or greater than a predetermined point cloud threshold as a wall point cloud, and detecting the point cloud data after the wall point cloud deletion as an equipment point cloud; A feature detection device comprising:

2. The detection means The point cloud data whose point cloud density is equal to or greater than a predetermined point cloud threshold is determined as a wall candidate point cloud, When a straight line connecting a starting point near the irradiation point of the laser light on the virtual plane to the point cloud data intersects with the wall candidate point cloud, the point cloud data intersecting with the wall candidate point cloud and the wall candidate point cloud are deleted as the wall point cloud; When the straight line does not intersect with the wall candidate point cloud, the point cloud data that does not intersect with the wall candidate point cloud is detected as the facility point cloud.

2. The feature detection device according to claim 1.

3. The starting point is a position on the center line of the moving body in the left-right direction and a predetermined distance from the ground surface.

3. The feature detection device according to claim 2.

4. an equipment point cluster identification means for segmenting the equipment point cluster in which the distance between each equipment point is less than a predetermined threshold value and identifying it as an equipment point cluster; 2. The feature detection device according to claim 1, further comprising:

5. a storage means for storing equipment characteristic conditions indicating the characteristics of the equipment for each equipment name of the equipment; an equipment identification means for extracting equipment characteristic conditions for each equipment point collection identified by the equipment point collection identification means, and comparing the extracted equipment characteristic conditions with the equipment characteristic conditions stored in the storage means to identify the equipment name of the equipment point collection; 5. The feature detection device according to claim 4, further comprising:

6. The facility characteristic conditions stored for each facility name of the facility are stored in accordance with the traveling direction of the moving object, The facility identification means For each facility point collection identified by the facility point collection identification means, a facility characteristic condition is extracted in accordance with the traveling direction of the mobile object.

6. The feature detection device according to claim 5.

7. A feature detection program executed by a feature detection device that acquires point cloud data from a point cloud data generation device that generates three-dimensional coordinates of features around a moving object as point cloud data based on reflected light of laser light irradiated onto the features, and detects the features based on the acquired point cloud data, a point cloud data projection step of setting virtual planes perpendicular to a traveling direction of the moving body at predetermined intervals and projecting the point cloud data existing between the set virtual planes onto the virtual planes; a point cloud density calculation step of calculating a density of the point cloud data per unit area on the projected virtual plane as a point cloud density; a detection step of deleting the point cloud data whose point cloud density is equal to or greater than a predetermined point cloud threshold as a wall point cloud, and detecting the point cloud data after the wall point cloud deletion as an equipment point cloud; a feature detection program for causing the feature detection device to execute the above steps;

Images