Point cloud processing device, point cloud processing method, and program
The point cloud processing device distinguishes between collision and non-collision risks using millimeter-wave radar, allowing vehicles to navigate through dust without deceleration, enhancing autonomous driving efficiency.
Patent Information
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- TOYOTA JIDOSHA KK
- Filing Date
- 2025-01-16
- Publication Date
- 2026-07-29
AI Technical Summary
Existing autonomous driving systems face inefficiencies when dust is generated, as they may erroneously detect dust as a point cloud, leading to unnecessary vehicle deceleration or inability to navigate through dusty areas, and existing methods to avoid dust do not effectively distinguish between collision risks and non-collision risks.
A point cloud processing device that uses a millimeter-wave radar to generate holding regions for targets in front of the vehicle and distinguishes between point clouds from LiDAR that indicate collision risks (vehicles and obstacles) and non-collision risks (dust, rain, fog) by discarding the latter for automatic driving control.
Enables vehicles to avoid collision risks while driving through dusty environments without unnecessary deceleration, maintaining efficiency by differentiating between relevant and irrelevant point clouds.
Smart Images
Figure 2026122664000001_ABST
Abstract
Description
Technical Field
[0001] This disclosure relates to a point cloud processing device, a point cloud processing method, and a program.
Background Art
[0002] Patent Document 1 describes an autonomous driving mining vehicle that suppresses a decrease in work efficiency when dust is generated at a work site. Also, in Patent Document 1, when dust is generated due to the running of other vehicles or the like, the transmittance of the laser irradiated from LiDAR (Light Detection And Ranging) may decrease and no reflected wave may be obtained. In that case, it is described that an undetected region (a region where point cloud data of LiDAR cannot be obtained) occurs. Furthermore, in the technology described in Patent Document 1, an alternative route is calculated through which the vehicle can run without passing through the undetected region. Patent Document 1 describes that since the planned route is updated by the alternative route, it is possible to suppress a decrease in work efficiency when dust is generated at the work site.
Prior Art Documents
Patent Documents
[0003]
Patent Document 1
Summary of the Invention
Problems to be Solved by the Invention
[0004] When dust is generated, not only may no reflected wave be obtained as described in Patent Document 1, but dust may also be detected as a point cloud. In the technology described in Patent Document 1, when dust is detected as a point cloud, there is a risk that the host vehicle may be erroneously decelerated or the like in order to avoid a collision with the point cloud. Furthermore, in real-world situations, there may be cases where the entire path of the vehicle is covered with dust due to wind, where dust kicked up by surrounding vehicles traveling ahead of the vehicle lingers for a long time, or where there is insufficient width or alternative route for the vehicle to avoid the dusty area. In such cases, methods that involve the vehicle avoiding the dusty area may not be able to prevent a decrease in work efficiency.
[0005] A technology is desired that allows a vehicle to avoid surrounding vehicles and obstacles that pose a collision risk, while driving without having to avoid dust and other debris that do not pose a collision risk.
[0006] In view of the above, the purpose of this disclosure is to provide a point cloud processing device, a point cloud processing method, and a program that enable a vehicle to avoid surrounding vehicles and obstacles that pose a collision risk, and to drive without avoiding dust and other objects that do not pose a collision risk. [Means for solving the problem]
[0007] (1) One aspect of the present disclosure is a point cloud processing device comprising: a generation unit that generates a point cloud holding region which is an area including a target located in front of the vehicle detected by a millimeter-wave radar; and a determination unit that determines whether or not a point cloud indicating the surrounding conditions of the vehicle, generated by a LiDAR that detects the surrounding conditions of the vehicle, exists within the point cloud holding region generated by the generation unit, wherein the determination unit, when it determines that a point cloud indicating the surrounding conditions of the vehicle exists within the point cloud holding region, outputs the point cloud indicating the surrounding conditions of the vehicle as a point cloud corresponding to the target located in front of the vehicle detected by the millimeter-wave radar.
[0008] (2) In the point cloud processing device of (1), if the determination unit determines that there are no point clouds indicating the surrounding conditions of the vehicle within the point cloud holding area, it may discard the point clouds indicating the surrounding conditions of the vehicle as point clouds corresponding to any of the dust, rain, and fog in front of the vehicle.
[0009] (3) In the point cloud processing device of (2), the point cloud showing the surrounding conditions of the vehicle that is discarded by the determination unit is not used for the automatic driving control of the vehicle, and the point cloud showing the surrounding conditions of the vehicle output by the determination unit may be used for the automatic driving control of the vehicle.
[0010] (4) One aspect of the present disclosure is a point cloud processing method comprising: a generation step of generating a point cloud holding region which is an area including a target located in front of the vehicle detected by a millimeter-wave radar; and a determination step of determining whether a point cloud indicating the surrounding conditions of the vehicle, generated by a LiDAR for detecting the surrounding conditions of the vehicle, exists within the point cloud holding region generated in the generation step, wherein in the determination step, if it is determined that a point cloud indicating the surrounding conditions of the vehicle exists within the point cloud holding region, the point cloud indicating the surrounding conditions of the vehicle is output as a point cloud corresponding to the target located in front of the vehicle detected by the millimeter-wave radar.
[0011] (5) One aspect of the present disclosure is a program that causes a processor to perform a generation step of generating a point cloud holding region which is an area including a target located in front of the vehicle detected by a millimeter-wave radar, and a determination step of determining whether a point cloud indicating the surrounding conditions of the vehicle, generated by a LiDAR for detecting the surrounding conditions of the vehicle, exists within the point cloud holding region generated in the generation step, wherein in the determination step, if it is determined that the point cloud indicating the surrounding conditions of the vehicle exists within the point cloud holding region, the point cloud indicating the surrounding conditions of the vehicle is output as a point cloud corresponding to the target located in front of the vehicle detected by the millimeter-wave radar. [Effects of the Invention]
[0012] According to this disclosure, the vehicle can avoid surrounding vehicles and obstacles that pose a collision risk, and can drive without having to avoid dust and other debris that do not pose a collision risk. [Brief explanation of the drawing]
[0013] [Figure 1] This figure shows an example of a vehicle 1 to which the point cloud processing device 16 of the first embodiment is applied. [Figure 2] Figure 1 shows an example of data flow within the vehicle 1. [Figure 3] This is a flowchart illustrating an example of processing performed by the processor 163 of the point cloud processing device 16 of the first embodiment. [Figure 4] This diagram illustrates a specific example of the process shown in Figure 3. [Modes for carrying out the invention]
[0014] Hereinafter, embodiments of the point cloud processing device, point cloud processing method, and program of this disclosure will be described with reference to the drawings.
[0015] <First Embodiment> Figure 1 shows an example of a vehicle 1 to which the point cloud processing device 16 of the first embodiment is applied. Figure 2 shows an example of the data flow within the vehicle 1 shown in Figure 1. In the example shown in Figures 1 and 2, the vehicle 1 is equipped with a LiDAR 11, a millimeter-wave radar 12, an HMI (Human Machine Interface) 13, a vehicle status sensor 14, a position information acquisition device 15, a point cloud processing device 16, a target recognition device 17, a vehicle control device 18, a steering actuator 18A, a braking actuator 18B, and a drive actuator 18C. The LiDAR 11 is positioned at the front of the vehicle 1, as shown in the example in Figure 4. The LiDAR 11 detects the surrounding conditions of the vehicle 1 (e.g., terrain, presence or absence of objects). The LiDAR 11 also generates point clouds PD1 to PD3 (see Figure 4) that show the surrounding conditions of the vehicle 1, and transmits the data of point clouds PD1 to PD3 (LiDAR point cloud (see Figure 2)) to the point cloud processing device 16.
[0016] Through diligent research, the inventors have discovered that when the vehicle 1 is moving forward on an unpaved road, not only are point clouds PD1 and PD2 (see Figure 4) corresponding to the object TG1 (see Figure 4) (surrounding vehicle PV (see Figure 4)) and object TG2 (see Figure 4) (obstacle BT (see Figure 4)) located in front of the vehicle 1 generated by the LiDAR 11, but point cloud PD3 (see Figure 4) corresponding to dust CD (see Figure 4), rain, fog, etc. in front of the vehicle 1. In other words, the inventors have found that LiDAR 11 can not only detect surrounding vehicles PV in front of vehicle 1 that require vehicle 1 to avoid a collision (i.e., control of steering actuator 18A and braking actuator 18B) as part of the surrounding conditions of vehicle 1, or obstacles BT in front of vehicle 1 that require vehicle 1 to avoid a collision as part of the surrounding conditions of vehicle 1, but also detect dust CD, rain, fog, etc. in front of vehicle 1 that do not require vehicle 1 to avoid a collision (i.e., control of steering actuator 18A and braking actuator 18B) as part of the surrounding conditions of vehicle 1. Therefore, in the examples shown in Figures 1 and 2, point clouds PD1 and PD2, which correspond to the targets TG1 (surrounding vehicle PV) and TG2 (obstacle BT) in front of vehicle 1 that vehicle 1 needs to avoid colliding with, and point cloud PD3, which corresponds to dust CD, rain, fog, etc. in front of vehicle 1 that vehicle 1 does not need to avoid colliding with, are implemented to distinguish between these two points.
[0017] In the examples shown in Figures 1 and 2, the millimeter-wave radar 12 is positioned at the front of the vehicle 1, for example, as shown in Figure 4. The millimeter-wave radar 12 detects targets TG1 and TG2 (see Figure 4) located in front of the vehicle 1 and transmits information (sensor data) (Radar target (see Figure 2)) regarding targets TG1 and TG2 to the point cloud processing device 16 and the target recognition device 17. The HMI 13 has the function of receiving various operations from the user of the vehicle 1 and transmits signals indicating the user's operations to the vehicle control device 18. The operations of the user of the vehicle 1 that the HMI 13 receives include, for example, operations to have the vehicle control device 18 execute automatic driving control of the vehicle 1, and operations to switch the vehicle 1 from automatic driving to manual driving. The vehicle state sensor 14 includes, for example, a vehicle speed sensor or the like. The vehicle state sensor 14 transmits information indicating the state of the host vehicle 1 (such as vehicle speed) to the vehicle control device 18.
[0018] The position information acquisition device 15 acquires information indicating the position of the host vehicle 1. The position information acquisition device 15 includes, for example, a GPS (Global Positioning System) device that measures the position of the host vehicle 1. The position information acquisition device 15 may perform well-known self-position estimation processing (localization) to improve the accuracy of the information indicating the position of the host vehicle 1. The position information acquisition device 15 transmits the information indicating the position of the host vehicle 1 to the vehicle control device 18. The point cloud processing device 16 executes processing of the point clouds PD1 to PD3 (see FIG. 4) generated by the LiDAR 11, and transmits the result of the processing of the point clouds PD1 to PD3 (the object-equivalent LiDAR point cloud (see FIG. 2)) to the object recognition device 17.
[0019] The object recognition device 17 executes recognition of the objects TG1 and TG2 based on the sensor data of the millimeter-wave radar 12 (information regarding the objects TG1 and TG2 (see FIG. 4) existing in front of the host vehicle 1 detected by the millimeter-wave radar 12) (specifically, time-series data) and the result of the processing of the point clouds PD1 to PD3 (see FIG. 4) executed by the point cloud processing device 16. The information regarding the objects TG1 and TG2 includes object position information (information indicating the relative positions of the objects TG1 and TG2 with respect to the host vehicle 1) and tracking information. The tracking information is information that can distinguish whether the objects TG1 and TG2 output from the millimeter-wave radar 12 in time series are the same. The object recognition device 17 transmits the recognition results of the objects TG1 and TG2 to the vehicle control device 18. The vehicle control device 18 is constituted by, for example, a vehicle control ECU (Electronic Control Unit). Based on information (signals) transmitted from, for example, the HMI 13, the vehicle state sensor 14, the position information acquisition device 15, and the target recognition device 17, the vehicle control device 18 controls the steering actuator 18A, the braking actuator 18B, and the driving actuator 18C. The vehicle control device 18 has a function of executing automatic driving control of the host vehicle 1. When the automatic driving control of the host vehicle 1 is executed, an automatic driving system (see FIG. 2) is constituted by the target recognition device 17 and the vehicle control device 18.
[0020] The point cloud processing device 16 is constituted by a microcomputer including a communication interface (I / F) 161, a memory 162, and a processor 163. The communication interface 161 has an interface circuit for connecting the point cloud processing device 16 to the LiDAR 11, the millimeter-wave radar 12, the HMI 13, the vehicle state sensor 14, the position information acquisition device 15, the target recognition device 17, and the vehicle control device 18. The memory 162 stores programs and various data used in processes executed by the processor 163. The processor 163 has functions as an acquisition unit 3A, a generation unit 3B, and a determination unit 3C.
[0021] The acquisition unit 3A acquires data (LiDAR point clouds (see FIG. 2)) of point clouds PD1 to PD3 (see FIG. 4) indicating the surrounding situation of the host vehicle 1 generated by the LiDAR 11 from the LiDAR 11. Further, the acquisition unit 3A acquires information regarding targets TG1, TG2 (see FIG. 4) existing in front of the host vehicle 1 detected by the millimeter-wave radar 12 from the millimeter-wave radar 12.
[0022] The generation unit 3B generates point cloud holding regions PA1 and PA2 (see Figure 4), which are areas containing targets TG1 and TG2 (see Figure 4), based on information about targets TG1 and TG2 (see Figure 4) acquired by the acquisition unit 3A. In the example shown in Figure 4, which will be described later, point cloud holding regions of a fixed size and shape are generated for all targets based on the size of the largest expected detection target. In other words, in the example shown in Figure 4, the size and shape of point cloud holding regions PA1 and PA2 are the same. In other examples, the likelihood for each detection target (truck, vehicle, rock, person, animal, etc.) is calculated based on information such as reflected power and speed detected by the millimeter-wave radar 12, and a point cloud holding region corresponding to the size and shape of the target with the highest likelihood may be generated.
[0023] In the examples shown in Figures 1 and 2, the determination unit 3C determines whether the point clouds PD1 to PD3 (see Figure 4), which represent the surrounding conditions of the vehicle 1 and are acquired by the acquisition unit 3A, exist within the point cloud holding regions PA1 and PA2 (see Figure 4) generated by the generation unit 3B. When the determination unit 3C determines that point clouds PD1 and PD2, which indicate the surrounding conditions of the vehicle 1, are located within the point cloud holding regions PA1 and PA2, it outputs the point clouds PD1 and PD2, which indicate the surrounding conditions of the vehicle 1, to the target recognition device 17 as a point cloud (target-equivalent LiDAR point cloud (see Figure 2)) corresponding to targets TG1 and TG2 (see Figure 4) located in front of the vehicle 1, as detected by the millimeter-wave radar 12. The point clouds PD1 and PD2, which indicate the surrounding conditions of the vehicle 1, output to the target recognition device 17 by the determination unit 3C are used by the target recognition device 17 and the vehicle control device 18 for automatic driving control of the vehicle 1. Specifically, the vehicle control device 18 controls the steering actuator 18A, braking actuator 18B, and drive actuator 18C so that collisions between the vehicle 1 and targets TG1 and TG2, which correspond to the point clouds PD1 and PD2, are avoided. On the other hand, if the determination unit 3C determines that the point cloud PD3 (see Figure 4) indicating the surrounding conditions of the vehicle 1 is not present within the point cloud holding areas PA1 and PA2 (see Figure 4), it discards the point cloud PD3 indicating the surrounding conditions of the vehicle 1 as a point cloud corresponding to dust CD (see Figure 4), rain, fog, etc. in front of the vehicle 1 and does not output it to the target recognition device 17. In other words, the point cloud indicating the surrounding conditions of the vehicle 1 that is discarded by the determination unit 3C is not used for the automatic driving control of the vehicle 1. That is, the control of the steering actuator 18A and braking actuator 18B to avoid dust CD (see Figure 4), rain, fog, etc. in front of the vehicle 1 that do not pose a risk of collision with the vehicle 1 is not performed by the vehicle control device 18.
[0024] Figure 3 is a flowchart illustrating an example of processing performed by the processor 163 of the point cloud processing device 16 of the first embodiment. The process shown in Figure 3 is executed, for example, while vehicle 1 is moving (more specifically, while moving forward).
[0025] In the example shown in Figure 3, in step S10, the acquisition unit 3A acquires point cloud data PD1 to PD3 (see Figure 4) from the LiDAR 11, which represent the surrounding conditions of the vehicle 1 generated by the LiDAR 11. In step S11, the acquisition unit 3A acquires information from the millimeter-wave radar 12 regarding targets TG1 and TG2 (see Figure 4) located in front of the vehicle 1 that were detected by the millimeter-wave radar 12. In step S12, the generation unit 3B generates point cloud holding regions PA1 and PA2 (see Figure 4), which are regions containing targets TG1 and TG2, based on the information about targets TG1 and TG2 acquired in step S11.
[0026] In step S13, the determination unit 3C determines whether the point clouds PD1 to PD3 (see Figure 4), which represent the surrounding conditions of the vehicle 1 and were acquired in step S10, are located within the point cloud holding regions PA1 and PA2 (see Figure 4) generated in step S12. If the answer is YES, the process proceeds to step S14; otherwise, the process proceeds to step S15. In step S14, the determination unit 3C outputs point clouds PD1 and PD2, which represent the surrounding conditions of the vehicle 1, to the target recognition device 17 as point clouds corresponding to targets TG1 and TG2 (see Figure 4) located in front of the vehicle 1 as detected by the millimeter-wave radar 12. In step S15, the determination unit 3C discards the point cloud PD3, which represents the surrounding conditions of the vehicle 1, as a point cloud corresponding to dust CD (see Figure 4), rain, fog, etc. in front of the vehicle 1, and does not output it to the target recognition device 17.
[0027] Figure 4 is a diagram illustrating a specific example of the process shown in Figure 3. As shown in Figure 4, when dust CD, surrounding vehicles PV, and obstacles BT are present in front of the vehicle 1, in step S10 of Figure 3, the acquisition unit 3A acquires point cloud PD1 (a point cloud corresponding to surrounding vehicles PV), point cloud PD2 (a point cloud corresponding to obstacles BT), and point cloud PD3 (a point cloud corresponding to dust CD) data from the LiDAR 11, which represent the surrounding conditions of the vehicle 1 generated by the LiDAR 11. In step S11 of Figure 3, the acquisition unit 3A acquires information from the millimeter-wave radar 12 regarding target TG1 (a target corresponding to a surrounding vehicle PV) and target TG2 (a target corresponding to an obstacle BT) located in front of the vehicle 1 that were detected by the millimeter-wave radar 12. In step S12 of Figure 3, the generation unit 3B generates a point cloud holding region PA1 including target TG1 and a point cloud holding region PA2 including target TG2, based on the information about targets TG1 and TG2 acquired in step S11.
[0028] In step S13 of Figure 3, the determination unit 3C determines that the point cloud PD3, which represents the surrounding conditions of the vehicle 1 and was acquired in step S10, is not located within the point cloud holding regions PA1 and PA2 generated in step S12. In step S15 of Figure 3, the determination unit 3C discards the point cloud PD3, which represents the surrounding conditions of the vehicle 1, as a point cloud corresponding to dust CD, rain, fog, etc., in front of the vehicle 1, and does not output it to the target recognition device 17. As a result, the steering actuator 18A and braking actuator 18B, which are controlled to avoid the dust CD, are not executed, and the vehicle 1 passes over the dust CD. In step S13 of Figure 3, which is performed after the vehicle 1 has passed through the dust cloud CD, the determination unit 3C determines that the point cloud PD1, which shows the surrounding conditions of the vehicle 1 and was acquired in step S10, is located within the point cloud holding region PA1 generated in step S12, and that the point cloud PD2, which shows the surrounding conditions of the vehicle 1 and was acquired in step S10, is located within the point cloud holding region PA2 generated in step S12. In step S14 of Figure 3, the determination unit 3C outputs the point clouds PD1 and PD2, which show the surrounding conditions of the vehicle 1, to the target recognition device 17 as point clouds corresponding to targets TG1 and TG2 located in front of the vehicle 1 and detected by the millimeter-wave radar 12. As a result, the vehicle control device 18 controls the steering actuator 18A, braking actuator 18B, and drive actuator 18C to avoid collisions between the vehicle 1 and targets TG1 and TG2.
[0029] As described above, in the examples shown in Figures 1 to 4, the LiDAR 11 is used in an environment where dust particles CD are present in the autonomous driving system, and the dust particles CD in the air are detected as point cloud PD3, which can suppress the risk of the vehicle 1 erroneously decelerating. In the examples shown in Figures 1 to 4, the property that dust particles CD are not detected by the millimeter-wave radar 12 (millimeter waves penetrate dust particles CD) is used. It is assumed that there are no targets at locations where targets are not detected by the millimeter-wave radar 12, and using this assumption, point cloud PD3, which corresponds to dust particles CD, is removed from point clouds PD1 to PD3 generated by LiDAR 11. As a result, the steering actuator 18A, braking actuator 18B, and drive actuator 18C are controlled for surrounding vehicles PV and obstacles BT that need to be avoided in collision with vehicle 1, but the steering actuator 18A and braking actuator 18B are not controlled for dust particles CD that do not need to be avoided in collision with vehicle 1. Consequently, vehicle 1 can pass over dust particles CD without having to avoid them. In detail, in the examples shown in Figures 1 to 4, a portion of the LiDAR 11's laser penetrates the dust CD and reaches the surrounding vehicle PV and obstacle BT, making it possible to recognize surrounding vehicle PV and obstacle BT that would not be recognizable without the LiDAR 11.
[0030] <Second Embodiment> The vehicle 1 to which the point cloud processing device 16 of the second embodiment is applied is configured in the same way as the vehicle 1 to which the point cloud processing device 16 of the first embodiment is applied, except for the points described later.
[0031] As described above, in the vehicle 1 (autonomous vehicle) to which the point cloud processing device 16 of the first embodiment is applied, the vehicle control device 18 performs control to activate the steering actuator 18A and / or braking actuator 18B based on the point clouds PD1 and PD2 output by the determination unit 3C in order to avoid a collision between the vehicle 1 and the targets TG1 and TG2. On the other hand, in the vehicle 1 to which the point cloud processing device 16 of the second embodiment is applied, the vehicle control device 18 causes the HMI 13 to output a warning indicating that an operation is required to avoid a collision between the vehicle 1 and targets TG1 and TG2, based on the point clouds PD1 and PD2 output by the determination unit 3C.
[0032] As described above, embodiments of the point cloud processing device, point cloud processing method, and program of this disclosure have been explained with reference to the drawings. However, the point cloud processing device, point cloud processing method, and program of this disclosure are not limited to the embodiments described above, and can be modified as appropriate without departing from the spirit of this disclosure. The configurations of each example of the embodiments described above may be combined as appropriate. In each example of the embodiments described above, the processing performed by the point cloud processing device 16 was described as software processing performed by executing a program, but the processing performed by the point cloud processing device 16 may also be hardware processing. Alternatively, the processing performed by the point cloud processing device 16 may be a combination of both software and hardware processing. Furthermore, the program stored in the memory 162 of the point cloud processing device 16 (a program that realizes the functions of the processor 163 of the point cloud processing device 16) may be recorded on a computer-readable storage medium such as a semiconductor memory, magnetic recording medium, or optical recording medium and provided and distributed. [Explanation of Symbols]
[0033] 1...Vehicle, 11...LiDAR, 12...Millimeter-wave radar, 13...HMI, 14...Vehicle status sensor, 15...Position information acquisition device, 16...Point cloud processing device, 161...Communication interface, 162...Memory, 163...Processor, 3A...Acquisition unit, 3B...Generation unit, 3C...Determination unit, 17...Target recognition device, 18...Vehicle control device, 18A...Steering actuator, 18B...Brake actuator, 18C...Drive actuator
Claims
1. A generation unit that generates a point cloud holding region which is an area containing a target located in front of the vehicle detected by millimeter-wave radar, The system includes a determination unit that determines whether or not a point cloud representing the surrounding conditions of the vehicle, generated by a LiDAR that detects the surrounding conditions of the vehicle, exists within the point cloud holding region generated by the generation unit. The determination unit determines that a point cloud representing the surrounding conditions of the vehicle exists within the point cloud holding area, and outputs the point cloud representing the surrounding conditions of the vehicle as a point cloud corresponding to the target located in front of the vehicle detected by the millimeter-wave radar.
2. The point cloud processing device according to claim 1, wherein the determination unit determines that no point cloud representing the surrounding conditions of the vehicle exists within the point cloud holding area, and discards the point cloud representing the surrounding conditions of the vehicle as a point cloud corresponding to any of the following: dust, rain, or fog in front of the vehicle.
3. The point cloud representing the surrounding conditions of the vehicle, which is discarded by the determination unit, is not used for the automatic driving control of the vehicle. The point cloud processing device according to claim 2, wherein the point cloud indicating the surrounding conditions of the vehicle output by the determination unit is used for automatic driving control of the vehicle.
4. The point cloud processing device generates a point cloud holding region, which is a region containing a target located in front of the vehicle detected by millimeter-wave radar, and The point cloud processing device includes a determination step of determining whether the point cloud representing the surrounding conditions of the vehicle, generated by a LiDAR that detects the surrounding conditions of the vehicle, exists within the point cloud holding region generated in the generation step. In the determination step, if it is determined that a point cloud representing the surrounding conditions of the vehicle exists within the point cloud holding area, the point cloud representing the surrounding conditions of the vehicle is output as a point cloud corresponding to the target located in front of the vehicle detected by the millimeter-wave radar, in a point cloud processing method.
5. In the processor, A generation step of generating a point cloud holding region which is an area that includes a target located in front of the vehicle detected by millimeter-wave radar, A program for executing a determination step of determining whether a point cloud representing the surrounding conditions of the vehicle, generated by a LiDAR that detects the surrounding conditions of the vehicle, exists within the point cloud holding region generated in the generation step, In the determination step, if it is determined that a point cloud representing the surrounding conditions of the vehicle exists within the point cloud holding area, the program outputs the point cloud representing the surrounding conditions of the vehicle as a point cloud corresponding to the target located in front of the vehicle detected by the millimeter-wave radar.