JP7517675B2 - Map data generating device, map data generating method, and program - Google Patents

Description

The present invention relates to a map data generation device and a map data generation method for creating three-dimensional map data, and further to a program for implementing these.

In recent years, with the advancement of sensing technology, SLAM (Simultaneous Localization and Mapping) has been attracting attention. SLAM is a technology in which a mobile object equipped with a sensor simultaneously estimates its own position and creates map data for an environmental map (see, for example, Patent Document 1 and Non-Patent Document 1). With SLAM, an autonomous robot does not need to move randomly, but can move along an environmental map, improving the efficiency of its movement.

Another example of a sensor that can be mounted on a moving object is LiDAR (Laser Imaging Detection and Ranging). LiDAR emits pulsed laser light to scan the surroundings, and measures the distance to nearby objects by receiving the light reflected or scattered by the surrounding objects.

LiDAR can identify the situation around a moving object over a wide area and obtain map data constructed from a three-dimensional point cloud. Such map data is expected to be effectively used for autonomous driving, etc.

US Patent Application Publication No. 2019/0003836

Ji Zhang, Sanjiv Sing, “LOAM: Lidar Odometry and Mapping in Real-time”, Robotics: Science and Systems 2014, July 12-16, 2014

However, with SLAM, if there are moving objects (other moving objects) other than the moving object equipped with the sensor when creating map data, the data of the other moving objects will also be recorded as map data. For example, when targeting an actual road, cars other than the car equipped with the sensor will also be traveling on the road, and the data of these other cars will also be recorded in the map data.

In this case, the accuracy of the map data is significantly reduced, so the data of other moving objects must be removed, but this data removal must be done manually, significantly increasing the cost of creating the map data.

One example of the objective of the present invention is to provide a map data generation device, a map data generation method, and a program that can solve the above problems and remove data of moving objects included in map data when creating map data using SLAM.

In order to achieve the above object, a map data generating device according to one aspect of the present invention comprises:
A data acquisition unit that acquires three-dimensional point cloud data of a target area outputted at regular intervals from a sensor mounted on the mobile object;
an object extraction unit that extracts an object present in the target area from the acquired three-dimensional point cloud data at a specific time;
a prediction processing unit that predicts a position at a specific time based on a movement amount of the moving body for the object extracted from the three-dimensional point cloud data at a time before the specific time;
a matching processing unit that determines whether the object extracted from the three-dimensional point cloud data at the specific time coincides with the object extracted from the three-dimensional point cloud data at the previous time at a predicted position;
a moving object determination unit that determines whether the object extracted from the three-dimensional point cloud data at the specific time is moving based on a result of the determination of whether the object matches;
a map data creation unit that creates map data of the target area by removing objects determined to be moving from the three-dimensional point cloud data acquired at a specific time;
Equipped with
It is characterized by:

In order to achieve the above object, a map data generating method according to one aspect of the present invention includes:
(a) acquiring three-dimensional point cloud data of a target area outputted at regular intervals from a sensor mounted on a moving object;
(b) extracting an object present in the target area from the acquired three-dimensional point cloud data at a specific time;
(c) predicting a position at the specific time based on a movement amount of the moving body for the object extracted from the three-dimensional point cloud data at a time before the specific time;
(d) determining whether the object extracted from the 3D point cloud data at the specific time coincides with the object extracted from the 3D point cloud data at the previous time at a predicted position ;
(e) determining whether the object extracted from the three-dimensional point cloud data at the specific time is moving based on a result of the determination of whether or not the object matches;
(f) removing objects determined to be moving from the three-dimensional point cloud data acquired at a specific time, and generating map data of the target area;
having
It is characterized by:

Furthermore, in order to achieve the above object, a program according to one aspect of the present invention comprises:
On the computer,
(a) acquiring three-dimensional point cloud data of a target area outputted at regular intervals from a sensor mounted on a moving object;
(b) extracting an object present in the target area from the acquired three-dimensional point cloud data at a specific time;
(c) predicting a position at the specific time based on a movement amount of the moving body for the object extracted from the three-dimensional point cloud data at a time before the specific time;
(d) determining whether the object extracted from the 3D point cloud data at the specific time coincides with the object extracted from the 3D point cloud data at the previous time at a predicted position ;
(e) determining whether the object extracted from the three-dimensional point cloud data at the specific time is moving based on a result of the determination of whether or not the object matches;
(f) removing objects determined to be moving from the three-dimensional point cloud data acquired at a specific time, and generating map data of the target area;
Execute the
It is characterized by:

As described above, according to the present invention, when creating map data using SLAM, data on moving objects included in the map data can be removed.

FIG. 1 is a block diagram showing a schematic configuration of a map data generating device according to an embodiment of the present invention. FIG. 2 is a diagram showing an example of a map data creation system including the map data generating device according to the embodiment of the present invention. FIG. 3 shows an example of three-dimensional point cloud data output by a sensor in the embodiment of the present invention. FIG. 4 is a diagram showing a state in which the ground points have been deleted from the three-dimensional point cloud data shown in FIG. FIG. 5 is a diagram showing a state in which an ROI is generated by performing clustering on the three-dimensional point cloud data shown in FIG. FIG. 6 is a diagram showing an object extracted from the three-dimensional point cloud data shown in FIG. FIG. 7 is a diagram showing a process of predicting the position of an object shown in FIG. FIG. 8 is a diagram showing a combination of the extracted latest object and predicted object shown in FIG. FIG. 9 is a diagram showing a series of processes shown in FIGS. FIG. 10 is a diagram showing the process for the latest object determined to be a moving object. Figures 11(a) and (b) are diagrams showing examples of map data for a target area, where Figure 11(a) shows an example created without removing moving objects, and Figure 11(b) shows an example created with the moving objects removed. FIG. 12 is a flow chart showing the operation of the map data generating device in the embodiment of the present invention. FIG. 13 is a block diagram showing an example of a computer that realizes the map data generating device according to the embodiment of the present invention.

(Embodiment)
A map data generating device according to an embodiment of the present invention will now be described with reference to FIGS.

[Device configuration]
First, a schematic configuration of a map data generating device in an embodiment will be described with reference to Fig. 1. Fig. 1 is a block diagram showing a schematic configuration of a map data generating device in an embodiment of the present invention.

The map data generating device 10 in the embodiment shown in FIG. 1 is a device that generates map data from three-dimensional point cloud data of a target area. As shown in FIG. 1, the map data generating device 10 includes a data acquisition unit 11, an object extraction unit 12, a prediction processing unit 13, a matching processing unit 14, a moving object determination unit 15, and a map data creation unit 16.

The data acquisition unit 11 then acquires three-dimensional point cloud data of the target area output at each time from a sensor mounted on the mobile object. The object extraction unit 12 extracts objects that exist in the target area from the acquired three-dimensional point cloud data at a specific time.

The prediction processing unit 13 predicts the position at a specific time of an object extracted from the 3D point cloud data at a time earlier than the specific time based on the amount of movement of the moving body. The matching processing unit 14 determines whether the object extracted from the 3D point cloud data at the specific time matches an object extracted from the 3D point cloud data at the previous time at the predicted position.

The moving object determination unit 15 determines whether the object extracted from the 3D point cloud data at the specific time is moving based on the result of the match determination. The map data creation unit 16 removes the object determined to be moving from the 3D point cloud data acquired at the specific time, and creates map data for the target area.

In this way, in the embodiment, the map data generation device 10 can delete data of objects corresponding to a moving body from the 3D point cloud data output by a sensor mounted on the moving body without manual intervention. In other words, according to the embodiment, when creating map data by SLAM, data of moving bodies included in the map data can be removed.

Next, the configuration and functions of the map data generating device 10 in the embodiment will be described in more detail with reference to Figures 2 to 11. Figure 2 is a diagram showing an example of a map data creation system configured with the map data generating device in the embodiment of the present invention.

As shown in FIG. 2, in the embodiment, the map data generating device 10 further includes an object storage unit 17 in addition to the configuration shown in FIG. 1. Also, in the embodiment, the moving body 20 is an automobile. The moving body 20, which is an automobile, is equipped with a sensor 21 that outputs three-dimensional point cloud data. The sensor 21 may be any sensor that is capable of acquiring distance information for each pixel, and a specific example is LiDAR.

Figure 3 shows an example of 3D point cloud data output by a sensor in an embodiment of the present invention. In the example of Figure 3, a LiDAR is used as the sensor 21. In addition, in the example of Figure 3, the target area is an urban area where automobiles can travel, and the 3D point cloud data is shown in a bird's-eye view of the target area.

The sensor 21 also outputs the three-dimensional point cloud data at set intervals, i.e., at each time point. Each piece of three-dimensional point cloud data output from the sensor 21 is then transmitted from a communication device mounted on the mobile body 20 to the map data generating device 10 by wireless communication or the like. The communication device of the mobile body 20 also transmits movement information indicating the speed and movement direction of the mobile body 20 at the time the three-dimensional point cloud data is output to the map data generating device 10.

As a result, in the map data generating device 10, the data acquiring unit 11 acquires the 3D point cloud data in the order in which it is output from the sensor 21. The data acquiring unit 11 also sends the acquired 3D point cloud data to the object extracting unit 12 in order.

In addition, in the embodiment, the sensor 21 may be an imaging element that outputs image data. In this case, the sensor 21 or the data acquisition unit 11 generates three-dimensional point cloud data from the image data using an existing image processing technique such as SfM (Structure from Motion).

In the embodiment, when three-dimensional point cloud data is sent to the object extraction unit 12, the object extraction unit 12 first assigns time information indicating the time at which the three-dimensional point cloud data was output from the sensor 21 to the sent three-dimensional point cloud data. Next, as shown in FIG. 4, the object extraction unit 12 executes a process of deleting ground points from the sent three-dimensional point cloud data. The ground points are deleted, for example, by deleting a point cloud whose coordinates in the height direction coincide with the contact surface of the moving body 20. FIG. 4 is a diagram showing the state in which the ground points have been deleted from the three-dimensional point cloud data shown in FIG. 3.

Next, as shown in FIG. 5, the object extraction unit 12 performs clustering on the 3D point cloud data from which the ground points have been removed, and classifies the point cloud into multiple clusters. Then, the object extraction unit 12 generates a 3D region of interest (ROI) based on the coordinates of the point clouds belonging to the same cluster. The object extraction unit 12 also calculates the center of gravity of the generated ROI. FIG. 5 is a diagram showing the state in which an ROI has been generated by performing clustering on the 3D point cloud data shown in FIG. 4. In FIG. 5, each cluster is distinguished by a rectangular boundary line.

Next, as shown in FIG. 6, the object extraction unit 12 identifies an ROI to be extracted as an object from among the generated ROIs, and extracts the identified ROI as an object. FIG. 6 is a diagram showing an object to be extracted from the three-dimensional point cloud data shown in FIG. 5.

Specifically, the object extraction unit 12 calculates the width, height, and depth for each generated ROI, and determines whether the calculated width, height, and depth satisfy the set conditions. The set conditions are set so that moving bodies other than the moving body 20 (cars, bicycles, pedestrians, etc.) can be extracted as objects. The set conditions include the upper and lower limits of at least one of the width, height, and depth, the ratio of any two of them, etc.

If an ROI is found to satisfy the set conditions, the object extraction unit 12 extracts the ROI as an object. The object extraction unit 12 also assigns the time information assigned to the original 3D point cloud data to the extracted object, and sends the object with the time information assigned to the prediction processing unit 13 and the object storage unit 17. The object storage unit 17 stores the sent object together with the time information.

The prediction processing unit 13 first processes the three-dimensional point cloud data output at each time in a time series to create point cloud observation information. The point cloud observation information is information on the amount of parallel movement and rotation angle of the moving body 20 at each time, and is calculated by comparing the estimated value of the self-position of the moving body 20 obtained by SLAM at each time.

Then, the prediction processing unit 13 calculates the direction of movement and amount of movement of the moving body 20 from any past time (for example, the time immediately before the latest time) to the latest time from the point cloud observation information. Next, in the embodiment, the prediction processing unit 13 acquires an object extracted from the 3D point cloud data at any past time from the object storage unit 17. Note that the prediction processing unit 13 may calculate the direction of movement and amount of movement of the moving body 20 based on information obtained from the moving body 20.

Then, as shown in FIG. 7, the prediction processing unit 13 predicts the position of the acquired object at the latest time based on the calculated moving direction and amount of movement of the moving body 20, and generates an object that is the same as the acquired object at the predicted position.

Figure 7 is a diagram showing the process of predicting the position of the object shown in Figure 5. In Figure 7, the position where the selection of the arrow is predicted is shown. In addition, in the following explanation, the object generated at the predicted position is referred to as the "predicted object". Then, the prediction processing unit 13 passes data of all the generated predicted objects to the matching processing unit 14.

The matching processing unit 14 obtains the object at the latest time (hereinafter referred to as the "latest object") sent from the object extraction unit 12 and the predicted object sent from the prediction processing unit 13.

Then, as shown in FIG. 8, the matching processing unit 14 compares the center of gravity of each latest object with the center of gravity of each predicted object, and sets the predicted object with the shortest distance between the centers of gravity as the matching target for the latest object. FIG. 8 is a diagram showing the combination of the extracted latest object and predicted object shown in FIG. 6. In FIG. 8, the predicted object is shown as a rectangle enclosed by dashed lines.

Next, the matching processing unit 14 uses an existing method for aligning three-dimensional shapes, such as the ICP (Iterative Closest Point) algorithm, to determine whether each of the latest objects matches the predicted object to be matched.

Specifically, the matching processing unit 14 rotates, translates, or both of the latest object and the predicted object to be matched to identify the state in which the two are most similar. The matching processing unit 14 then calculates the proportion of the matching portion of the latest object in the identified state, and if this value is equal to or greater than a threshold, it determines that the two match, and if not, it determines that the two do not match.

The processing in the prediction processing unit 13 and the matching processing unit 14 will now be described in detail with reference to FIG. 9. FIG. 9 is a diagram showing the series of processing steps shown in FIGS. 6 to 8.

As shown in the upper left diagram of FIG. 9, the prediction processing unit 13 acquires objects extracted from the 3D point cloud data at the time immediately preceding the most recent time. Next, as shown in the upper right diagram of FIG. 9, the prediction processing unit 13 predicts the position of the acquired object at the most recent time and sets a predicted object. Next, as shown in the lower diagram of FIG. 9, the matching processing unit 14 determines whether each of the most recent objects matches the predicted object to be matched.

The moving object determination unit 15 determines whether the latest object is a moving object, as shown in FIG. 10, based on the determination result by the matching processing unit 14. Specifically, the moving object determination unit 15 determines whether the translation amount and rotation angle when the matching processing unit 14 determines that the two match are within a set range. If the translation amount and rotation angle are within the set range, the moving object determination unit 15 determines that the latest object is not a moving object, and if not, determines that the latest object is a moving object. FIG. 10 is a diagram showing the processing for the latest object determined to be a moving object.

In the embodiment, the data acquisition unit 11, object extraction unit 12, prediction processing unit 13, matching processing unit 14, moving body determination unit 15, and object storage unit 17 are capable of detecting moving bodies, and therefore constitute the moving body detection device 30.

The map data creation unit 16 removes the latest object determined to be a moving body from the 3D point cloud data for the latest time, as shown in FIG. 10. Then, the map data creation unit 16 creates map data for the target area using the 3D point cloud data for each time after the moving body has been removed, as shown in FIG. 11. FIGS. 11(a) and (b) are diagrams showing examples of map data for the target area, with FIG. 11(a) showing an example created without removing the moving body, and FIG. 11(b) showing an example created with the moving body removed.

[Device Operation]
Next, the operation of the map data generation device 10 in the embodiment will be described with reference to FIG. 12. FIG. 12 is a flow diagram showing the operation of the map data generation device in the embodiment of the present invention. In the following description, FIGS. 1 to 11 will be referred to as appropriate. Also, in the embodiment, a map data generation method is implemented by operating the map data generation device 10. Therefore, the description of the map data generation method in the embodiment will be replaced with the description of the operation of the map data generation device below.

As shown in FIG. 12, first, the data acquisition unit 11 acquires the latest 3D point cloud data of the target area output from the sensor 21 mounted on the mobile body 20 (step A1).

Next, the object extraction unit 12 adds time information to the 3D point cloud data acquired in step A1, and performs a process to delete ground points from this 3D point cloud data (step A2).

Next, the object extraction unit 12 performs clustering on the 3D point cloud data from which the ground points have been removed in step A2, and classifies the point cloud into multiple clusters (step A3).

Next, the object extraction unit 12 generates a three-dimensional ROI based on the coordinates of the point cloud belonging to the same cluster. Then, the object extraction unit 12 identifies an ROI to be extracted as an object from among the generated ROIs based on the set conditions, and extracts the identified ROI as the latest object (step A4).

Next, the prediction processing unit 13 calculates the direction and amount of movement of the moving body 20 at any past time, specifically, from the time immediately before the latest time to the latest time, based on the movement information of the moving body 20 sent from the moving body 20 (step A5).

Next, the prediction processing unit 13 acquires from the object storage unit 17 the object extracted in step A4 at the time immediately before the most recent time. The prediction processing unit 13 then predicts the position of the acquired object at the most recent time based on the direction of movement and amount of movement of the moving body 20 calculated in step A5, and generates an object (predicted object) that is the same as the acquired object at the predicted position (step A6).

In the above example, steps A5 and A6 are executed after steps A1 to A4 are executed, but in an embodiment, steps A5 and A6 may be executed simultaneously with steps A1 to A4 or before steps A1 to A4 are executed.

Next, the matching processing unit 14 identifies a predicted object to be matched for each of the latest objects extracted in step A4 (step A7). Specifically, the matching processing unit 14 compares the center of gravity of the latest object extracted in step A4 with the center of gravity of each predicted object, and sets the predicted object with the shortest distance between the centers of gravity as the matching target for the latest object.

Next, the matching processing unit 14 uses the ICP algorithm to determine whether each of the latest objects matches the predicted object to be matched (step A8).

Next, the moving object determination unit 15 determines whether each of the newest objects is a moving object based on the determination result in step A8 (step A9).

The map data creation unit 16 removes the most recent object determined to be a moving body in step A9 from the most recent 3D point cloud data acquired in step A1 (step A10). The map data creation unit 16 then creates map data for the target area using the most recent 3D point cloud data after removal of the moving body in step A10 and the previous 3D point cloud data after step 10 (step A11).

In addition, in the embodiment, the above-described steps A1 to A11 are repeatedly executed each time 3D point cloud data is output from the sensor 21 mounted on the moving body 20. Therefore, the wider the range in which the moving body 20 moves, the wider the range of the map data will be.

[Effects of the embodiment]
As described above, according to the embodiment, when creating map data by SLAM, data of moving objects included in the map data is automatically removed, so that highly accurate map data can be created.

[program]
The program in the embodiment may be a program that causes a computer to execute steps A1 to A11 shown in Fig. 12. By installing and executing this program in a computer, the map data generation device 10 and the map data creation method in the embodiment can be realized. In this case, the processor of the computer functions as a data acquisition unit 11, an object extraction unit 12, a prediction processing unit 13, a matching processing unit 14, a moving object determination unit 15, and a map data creation unit 16, and performs processing.

The program in the embodiment may be executed by a computer system constructed by multiple computers. In this case, for example, each computer may function as any one of the data acquisition unit 11, object extraction unit 12, prediction processing unit 13, matching processing unit 14, moving body determination unit 15, and map data creation unit 16. Furthermore, the computer that executes the program in the embodiment may be a general-purpose computer, or may be an in-vehicle computer mounted on a moving body, a smartphone, or a tablet terminal.

Here, a computer that realizes the map data generating device 10 by executing the program in the embodiment will be described with reference to FIG. 13. FIG. 13 is a block diagram showing an example of a computer that realizes the map data generating device in the embodiment of the present invention.

As shown in FIG. 13, the computer 110 includes a CPU (Central Processing Unit) 111, a main memory 112, a storage device 113, an input interface 114, a display controller 115, a data reader/writer 116, and a communication interface 117. These components are connected to each other via a bus 121 so as to be able to communicate data with each other. Note that the computer 110 may include a GPU (Graphics Processing Unit) or an FPGA (Field-Programmable Gate Array) in addition to or instead of the CPU 111.

The CPU 111 loads the program (code) in this embodiment stored in the storage device 113 into the main memory 112 and executes it in a predetermined order to perform various calculations. The main memory 112 is typically a volatile storage device such as a DRAM (Dynamic Random Access Memory). The program in this embodiment is provided in a state stored in a computer-readable recording medium 120. The program in this embodiment may be distributed over the Internet connected via the communication interface 117.

Specific examples of the storage device 113 include a hard disk drive and a semiconductor storage device such as a flash memory. The input interface 114 mediates data transmission between the CPU 111 and input devices 118 such as a keyboard and a mouse. The display controller 115 is connected to the display device 119 and controls the display on the display device 119.

The data reader/writer 116 mediates data transmission between the CPU 111 and the recording medium 120, reads programs from the recording medium 120, and writes the results of processing in the computer 110 to the recording medium 120. The communication interface 117 mediates data transmission between the CPU 111 and other computers.

Specific examples of the recording medium 120 include general-purpose semiconductor storage devices such as CF (Compact Flash (registered trademark)) and SD (Secure Digital), magnetic recording media such as flexible disks, and optical recording media such as CD-ROMs (Compact Disk Read Only Memory).

Note that the map data generating device 10 in this embodiment can also be realized by using hardware corresponding to each part, rather than a computer on which a program is installed. Furthermore, the map data generating device 10 may be realized in part by a program and the remaining part by hardware.

A part or all of the above-described embodiment can be expressed by (Appendix 1) to (Appendix 12) described below, but is not limited to the following description.

(Appendix 1)
A data acquisition unit that acquires three-dimensional point cloud data of a target area outputted at regular intervals from a sensor mounted on the mobile object;
an object extraction unit that extracts an object present in the target area from the acquired three-dimensional point cloud data at a specific time;
a prediction processing unit that predicts a position at a specific time based on a movement amount of the moving body for the object extracted from the three-dimensional point cloud data at a time before the specific time;
a matching processing unit that determines whether the object extracted from the three-dimensional point cloud data at the specific time coincides with the object extracted from the three-dimensional point cloud data at the previous time at a predicted position;
a moving object determination unit that determines whether the object extracted from the three-dimensional point cloud data at the specific time is moving based on a result of the determination of whether the object matches;
a map data creation unit that creates map data of the target area by removing objects determined to be moving from the three-dimensional point cloud data acquired at a specific time;
Equipped with
A map data generating device comprising:

(Appendix 2)
2. A map data generating device according to claim 1,
the object extraction unit removes a portion corresponding to the ground from the acquired 3D point cloud data at the specific time, performs clustering on the 3D point cloud data from which the portion corresponding to the ground has been removed, classifies the data into a plurality of objects, and extracts an object that satisfies a specific condition from among the obtained plurality of objects;
A map data generating device comprising:

(Appendix 3)
3. A map data generating device according to claim 1,
the matching processing unit calculates a distance between the object extracted from the three-dimensional point cloud data at the specific time and the object extracted from the three-dimensional point cloud data at the previous time at a predicted position, and if the calculated distance satisfies a set condition, determines whether the two match;
A map data generating device comprising:

(Appendix 4)
A map data generating device according to any one of claims 1 to 3,
the matching processing unit rotates, translates, or both of the object extracted from the three-dimensional point cloud data at the specific time and the object extracted from the three-dimensional point cloud data at the previous time at a predicted position, and determines whether the two match;
the moving object determination unit determines whether the object extracted from the three-dimensional point cloud data at the specific time is moving, based on the amount of rotation and the amount of translation, when the two match.
A map data generating device comprising:

(Appendix 5)
(a) acquiring three-dimensional point cloud data of a target area outputted at regular intervals from a sensor mounted on a moving object;
(b) extracting an object present in the target area from the acquired three-dimensional point cloud data at a specific time;
(c) predicting a position at the specific time based on a movement amount of the moving body for the object extracted from the three-dimensional point cloud data at a time before the specific time;
(d) determining whether the object extracted from the 3D point cloud data at the specific time coincides with the object extracted from the 3D point cloud data at the previous time at a predicted position ;
(e) determining whether the object extracted from the three-dimensional point cloud data at the specific time is moving based on a result of the determination of whether or not the object matches;
(f) removing objects determined to be moving from the three-dimensional point cloud data acquired at a specific time, and generating map data of the target area;
having
A map data generating method comprising:

(Appendix 6)
6. A map data generating method according to claim 5, further comprising:
In the step (b), a portion corresponding to the ground is removed from the acquired 3D point cloud data at the specific time, clustering is performed on the 3D point cloud data from which the portion corresponding to the ground has been removed to classify the data into a plurality of objects, and an object that satisfies a specific condition is extracted from the obtained plurality of objects.
A map data generating method comprising:

(Appendix 7)
7. A map data generating method according to claim 5, further comprising:
In the step (d), a distance is calculated between the object extracted from the three-dimensional point cloud data at the specific time and the object extracted from the three-dimensional point cloud data at the previous time at a predicted position, and if the calculated distance satisfies a set condition, it is determined whether the two match.
A map data generating method comprising:

(Appendix 8)
A map data generating method according to any one of Supplementary Note 5 to 7,
In the step (d), rotating, translating, or both of the object extracted from the three-dimensional point cloud data at the specific time and the object extracted from the three-dimensional point cloud data at the previous time at the predicted position is determined to be consistent with the object;
In the step (e), if the two match, it is determined whether or not the object extracted from the three-dimensional point cloud data at the specific time has moved, based on the amount of rotation and the amount of translation.
A map data generating method comprising:

(Appendix 9)
On the computer,
(a) acquiring three-dimensional point cloud data of a target area outputted at regular intervals from a sensor mounted on a moving object;
(b) extracting an object present in the target area from the acquired three-dimensional point cloud data at a specific time;
(c) predicting a position at the specific time based on a movement amount of the moving body for the object extracted from the three-dimensional point cloud data at a time before the specific time;
(d) determining whether the object extracted from the 3D point cloud data at the specific time coincides with the object extracted from the 3D point cloud data at the previous time at a predicted position ;
(e) determining whether the object extracted from the three-dimensional point cloud data at the specific time is moving based on a result of the determination of whether or not the object matches;
(f) removing objects determined to be moving from the three-dimensional point cloud data acquired at a specific time, and generating map data of the target area;
Execute the
A program characterized by:

(Appendix 10)
10. The program according to claim 9,
In the step (b), a portion corresponding to the ground is removed from the acquired 3D point cloud data at the specific time, clustering is performed on the 3D point cloud data from which the portion corresponding to the ground has been removed to classify the data into a plurality of objects, and an object that satisfies a specific condition is extracted from the obtained plurality of objects.
A program characterized by:

(Appendix 11)
The program according to claim 9 or 10,
In the step (d), a distance is calculated between the object extracted from the three-dimensional point cloud data at the specific time and the object extracted from the three-dimensional point cloud data at the previous time at a predicted position, and if the calculated distance satisfies a set condition, it is determined whether the two match.
A program characterized by:

(Appendix 12)
The program according to any one of appendixes 9 to 11,
In the step (d), rotating, translating, or both of the object extracted from the three-dimensional point cloud data at the specific time and the object extracted from the three-dimensional point cloud data at the previous time at the predicted position is determined to be consistent with the object;
In the step (e), if the two match, it is determined whether or not the object extracted from the three-dimensional point cloud data at the specific time has moved, based on the amount of rotation and the amount of translation.
A program characterized by:

As described above, according to the present invention, when creating map data using SLAM, data on moving objects included in the map data can be removed. The present invention is useful for systems that realize SLAM.

REFERENCE SIGNS LIST 10 Map data generating device 11 Data acquisition unit 12 Object extraction unit 13 Prediction processing unit 14 Matching processing unit 15 Moving body determination unit 16 Map data creating unit 17 Object storage unit 20 Moving body 21 Sensor 30 Moving body detection device 110 Computer 111 CPU
112 Main memory 113 Storage device 114 Input interface 115 Display controller 116 Data reader/writer 117 Communication interface 118 Input device 119 Display device 120 Recording medium 121 Bus

Claims (12)

A data acquisition unit that acquires three-dimensional point cloud data of a target area outputted at regular intervals from a sensor mounted on the mobile object;
an object extraction unit that extracts an object present in the target area from the acquired three-dimensional point cloud data at a specific time;
a prediction processing unit that creates point cloud observation information indicating a translation amount and a rotation angle of the moving body for each time from the three-dimensional point cloud data output for each time, calculates a moving direction and a moving amount of the moving body from a time earlier than the specific time to the specific time from the point cloud observation information, and predicts a position of the object at the specific time based on the calculated moving direction and the moving amount of the moving body for the object extracted from the three-dimensional point cloud data for the time earlier than the specific time;
a matching processing unit that determines whether the object extracted from the three-dimensional point cloud data at the specific time coincides with the object extracted from the three-dimensional point cloud data at the previous time at a predicted position;
a moving object determination unit that determines whether the object extracted from the three-dimensional point cloud data at the specific time is moving based on a result of the determination of whether the object matches;
a map data creation unit that creates map data of the target area by removing objects determined to be moving from the three-dimensional point cloud data acquired at a specific time;
Equipped with
A map data generating device comprising: 2. The map data generating device according to claim 1,
the object extraction unit removes a portion corresponding to the ground from the acquired 3D point cloud data at the specific time, performs clustering on the 3D point cloud data from which the portion corresponding to the ground has been removed, classifies the data into a plurality of objects, and extracts an object that satisfies a specific condition from among the obtained plurality of objects;
A map data generating device comprising: 3. The map data generating device according to claim 1,
the matching processing unit calculates a distance between the object extracted from the three-dimensional point cloud data at the specific time and the object extracted from the three-dimensional point cloud data at the previous time at a predicted position, and if the calculated distance satisfies a set condition, determines whether the two match;
A map data generating device comprising: A map data generating device according to any one of claims 1 to 3,
the matching processing unit rotates, translates, or both of the object extracted from the three-dimensional point cloud data at the specific time and the object extracted from the three-dimensional point cloud data at the previous time at a predicted position, and determines whether the two match;
the moving object determination unit determines whether the object extracted from the three-dimensional point cloud data at the specific time is moving, based on the amount of rotation and the amount of translation, when the two match.
A map data generating device comprising: (a) acquiring three-dimensional point cloud data of a target area outputted at regular intervals from a sensor mounted on a moving object;
(b) extracting an object present in the target area from the acquired three-dimensional point cloud data at a specific time;
(c) creating point cloud observation information indicating the translation amount and rotation angle of the moving body for each time from the three-dimensional point cloud data output for each time, calculating the movement direction and movement amount of the moving body from a time earlier than the specific time to the specific time from the point cloud observation information, and predicting a position of the object at the specific time based on the calculated movement direction and movement amount of the moving body for the object extracted from the three-dimensional point cloud data for a time earlier than the specific time;
(d) determining whether the object extracted from the 3D point cloud data at the specific time coincides with the object extracted from the 3D point cloud data at the previous time at a predicted position ;
(e) determining whether the object extracted from the three-dimensional point cloud data at the specific time is moving based on a result of the determination of whether or not the object matches;
(f) removing objects determined to be moving from the three-dimensional point cloud data acquired at a specific time, and generating map data of the target area;
having
A map data generating method comprising: 6. The map data generating method according to claim 5,
In the step (b), a portion corresponding to the ground is removed from the acquired 3D point cloud data at the specific time, clustering is performed on the 3D point cloud data from which the portion corresponding to the ground has been removed to classify the data into a plurality of objects, and an object that satisfies a specific condition is extracted from the obtained plurality of objects.
A map data generating method comprising: 7. A map data generating method according to claim 5, further comprising the steps of:
In the step (d), a distance is calculated between the object extracted from the three-dimensional point cloud data at the specific time and the object extracted from the three-dimensional point cloud data at the previous time at a predicted position, and if the calculated distance satisfies a set condition, it is determined whether the two match.
A map data generating method comprising: A map data generating method according to any one of claims 5 to 7,
In the step (d), rotating, translating, or both of the object extracted from the three-dimensional point cloud data at the specific time and the object extracted from the three-dimensional point cloud data at the previous time at the predicted position is determined to be consistent with the object;
In the step (e), if the two match, it is determined whether or not the object extracted from the three-dimensional point cloud data at the specific time has moved, based on the amount of rotation and the amount of translation.
A map data generating method comprising: On the computer,
(a) acquiring three-dimensional point cloud data of a target area outputted at regular intervals from a sensor mounted on a moving object;
(b) extracting an object present in the target area from the acquired three-dimensional point cloud data at a specific time;
(c) creating point cloud observation information indicating the translation amount and rotation angle of the moving body for each time from the three-dimensional point cloud data output for each time, calculating the movement direction and movement amount of the moving body from a time earlier than the specific time to the specific time from the point cloud observation information, and predicting a position of the object at the specific time based on the calculated movement direction and movement amount of the moving body for the object extracted from the three-dimensional point cloud data for a time earlier than the specific time;
(d) determining whether the object extracted from the 3D point cloud data at the specific time coincides with the object extracted from the 3D point cloud data at the previous time at a predicted position ;
(e) determining whether the object extracted from the three-dimensional point cloud data at the specific time is moving based on a result of the determination of whether or not the object matches;
(f) removing objects determined to be moving from the three-dimensional point cloud data acquired at a specific time, and generating map data of the target area;
Execute the
A program characterized by: The program according to claim 9,
In the step (b), a portion corresponding to the ground is removed from the acquired 3D point cloud data at the specific time, clustering is performed on the 3D point cloud data from which the portion corresponding to the ground has been removed to classify the data into a plurality of objects, and an object that satisfies a specific condition is extracted from the obtained plurality of objects.
A program characterized by: The program according to claim 9 or 10,
In the step (d), a distance is calculated between the object extracted from the three-dimensional point cloud data at the specific time and the object extracted from the three-dimensional point cloud data at the previous time at a predicted position, and if the calculated distance satisfies a set condition, it is determined whether the two match.
A program characterized by: The program according to any one of claims 9 to 11,
In the step (d), rotating, translating, or both of the object extracted from the three-dimensional point cloud data at the specific time and the object extracted from the three-dimensional point cloud data at the previous time at the predicted position is determined to be consistent with the object;
In the step (e), if the two match, it is determined whether or not the object extracted from the three-dimensional point cloud data at the specific time has moved, based on the amount of rotation and the amount of translation.
A program characterized by:

Images