Recognition system, recognition device, recognition method, recognition program, and recognition data generation method

Abstract

To provide a recognition system the recognition accuracy of which is high.SOLUTION: The processor of a recognition system that recognizes a target mobile body that is movable in a scan space scanned by illumination light from a scanning device in a host vehicle is constituted so as to execute: acquiring scan data Ds that contains a rear point cloud Pb as a scan point cloud at a position rearward of a reflector Or in a scan direction ψs in which the reflector Or whose reflection characteristic to illumination light exists in a range of interest on the high reflection side is scanned; and generating recognition data Dr excluding, from the rear point cloud Pb in the scan data Ds with which the target mobile body is to be recognized, a rear point cloud Pb in which there exists a symmetrical point cloud, as the scanning point cloud of a real image that causes a virtual image to appear at a position rearward of the reflector Or, in an area symmetrical to the rear position.SELECTED DRAWING: Figure 4

Description

[Technical Field]

[0001] The present disclosure relates to a recognition technology for recognizing a moving object. [Background technology]

[0002] Recognition techniques for recognizing a moving object present in a scanning space scanned by irradiation light from a scanning device are widely known. Patent Document 1 discloses a recognition technique for eliminating a situation in which a virtual image generated by reflection of laser light, which is the irradiation light, is mistakenly recognized as a moving object. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent No. 6962365 Summary of the Invention [Problem to be solved by the invention]

[0004] In the recognition technology disclosed in Patent Document 1, pixels in front of the background distance image, which is the scanning result when no moving object is present, are extracted as the moving object to be recognized. However, if a reflector that causes a virtual image transmits the irradiated light, a moving object that actually exists behind the scanning device will not be recognized. This issue is difficult to address with recognition technology that uses a scanning device installed on a train platform, as in Patent Document 1, but is expected to be resolved with recognition technology that uses a scanning device mounted on a vehicle.

[0005] An object of the present disclosure is to provide a recognition system with high recognition accuracy. Another object of the present disclosure is to provide a recognition device with high recognition accuracy. Yet another object of the present disclosure is to provide a recognition method with high recognition accuracy. Yet another object of the present disclosure is to provide a recognition program with high recognition accuracy. [Means for solving the problem]

[0006] The technical means of the present disclosure for solving the problems will be described below. Note that the claims and the reference characters in parentheses in this section indicate the correspondence with the specific means described in the embodiments described later in detail, and do not limit the technical scope of the present disclosure.

[0007] A first aspect of the present disclosure is Processor(12) and storage media (10) A recognition system for recognizing a target moving object (Ot) capable of moving in a scanning space (30) scanned by irradiation light from a scanning device (3) in a host vehicle (2), comprising: The processor In a scanning direction (ψs) in which a reflector (Or) in a range of interest having high reflectivity characteristics with respect to the irradiated light is scanned, scanning data (Ds) including a back point group (Pb) as a scanning point group at a position further back than the reflector is acquired; and generating recognition data (Dr) by excluding a back point cloud, which is a scanning point cloud of a real image (Ia) that makes a virtual image (Iv) appear at a position further back than the reflector, from a back point cloud to be recognized of the target moving body in the scanning data, the back point cloud in which a symmetrical point cloud (Ps) exists in a symmetrical area (As) with respect to the back position of the reflector. 、 The acquisition of scanning data is storing map data (Dm) including identification information (σi) for identifying an object present in the scanning space; and acquiring scan data including a deep point cloud at a position further back than a reflector identified based on the identification information read from the storage medium; The generation of recognition data is This includes excluding from the recognition target a group of points on the far side where a group of symmetrical points exists in an area symmetrical to a position on the far side of a reflector whose identification information represents optical characteristics that allow the transmission of irradiated light.

[0008] A second aspect of the present disclosure is Processor(12) and storage media (10) The recognition device is configured to be mountable on a host vehicle (2) and recognizes a target moving object (Ot) that can move through a scanning space (30) scanned by irradiation light from a scanning device (3) in the host vehicle, The processor In a scanning direction (ψs) in which a reflector (Or) in a range of interest having high reflectivity characteristics with respect to the irradiated light is scanned, scanning data (Ds) including a back point group (Pb) as a scanning point group at a position further back than the reflector is acquired; and generating recognition data (Dr) by excluding a back point cloud, which is a scanning point cloud of a real image (Ia) that makes a virtual image (Iv) appear at a position further back than the reflector, from a back point cloud to be recognized of the target moving body in the scanning data, the back point cloud in which a symmetrical point cloud (Ps) exists in a symmetrical area (As) with respect to the back position of the reflector. 、 The acquisition of scanning data is storing map data (Dm) including identification information (σi) for identifying an object present in the scanning space; and acquiring scan data including a deep point cloud at a position further back than a reflector identified based on the identification information read from the storage medium; The generation of recognition data is This includes excluding from the recognition target a group of points on the far side where a group of symmetrical points exists in an area symmetrical to a position on the far side of a reflector whose identification information represents optical characteristics that allow the transmission of irradiated light.

[0009] A third aspect of the present disclosure is A recognition method executed by a processor (12) for recognizing a target moving object (Ot) capable of moving in a scanning space (30) scanned by illumination light from a scanning device (3) in a host vehicle (2), comprising: In a scanning direction (ψs) in which a reflector (Or) in a range of interest having high reflectivity characteristics with respect to the irradiated light is scanned, scanning data (Ds) including a back point group (Pb) as a scanning point group at a position further back than the reflector is acquired; The scanning point cloud of the real image (Ia) that makes a virtual image (Iv) appear at a position further back than the reflector from the back point cloud to be recognized of the target moving body in the scanning data is a scanning point cloud of the real image (Ia) that makes a virtual image (Iv) appear at a position further back than the reflector, and the recognition data (Dr) is generated by excluding the back point cloud in which a symmetrical point cloud (Ps) exists in a symmetrical area (As) with respect to the back position of the reflector. fruit, The acquisition of scanning data is The method includes acquiring scan data including a deep point cloud at a position farther from a reflector identified based on identification information read from a storage medium (10) that stores map data (Dm) including identification information (σi) for identifying an object present in the scanning space, The generation of recognition data is This includes excluding from the recognition target a group of points on the far side where a group of symmetrical points exists in an area symmetrical to a position on the far side of a reflector whose identification information represents optical characteristics that allow the transmission of irradiated light.

[0010] A fourth aspect of the present disclosure is A recognition program including instructions stored in a storage medium (10) and executed by a processor (12) for recognizing a target moving object (Ot) capable of moving through a scanning space (30) scanned by irradiation light from a scanning device (3) in a host vehicle (2), The command is, In a scanning direction (ψs) in which a reflector (Or) in a range of interest having high reflectivity characteristics with respect to the irradiated light is scanned, scanning data (Ds) including a back point group (Pb) as a scanning point group at a position further back than the reflector is acquired; The scanning point cloud of the real image (Ia) that makes a virtual image (Iv) appear at a position further back than the reflector from the back point cloud to be recognized of the target moving body in the scanning data is a scanning point cloud of the real image (Ia) that makes a virtual image (Iv) appear at a position further back than the reflector, and the recognition data (Dr) is generated by excluding the back point cloud in which a symmetrical point cloud (Ps) exists in a symmetrical area (As) with respect to the back position of the reflector. fruit, The acquisition of scanning data is The method includes storing map data (Dm) including identification information (σi) for identifying an object present in the scanning space, and acquiring scan data including a deep point cloud at a position further back than a reflector identified based on the identification information read from a storage medium, The generation of recognition data is This includes excluding from the recognition target a group of points on the far side where a group of symmetrical points exists in an area symmetrical to a position on the far side of a reflector whose identification information represents optical characteristics that allow the transmission of irradiated light.

[0011] A fifth aspect of the present disclosure is A recognition data generation method executed by a processor (12) for recognizing a target moving object (Ot) capable of moving in a scanning space (30) scanned by irradiation light from a scanning device (3) in a host vehicle (2) and generating recognition data (Dr), comprising: In a scanning direction (ψs) in which a reflector (Or) in a range of interest having high reflectivity characteristics with respect to the irradiated light is scanned, scanning data (Ds) including a back point group (Pb) as a scanning point group at a position further back than the reflector is acquired; The recognition data is generated by excluding a back point cloud, which is a scanning point cloud of a real image (Ia) that makes a virtual image (Iv) appear at a position further back than the reflector, from a back point cloud to be recognized of the target moving body in the scanning data, the back point cloud in which a symmetrical point cloud (Ps) exists in a symmetrical area (As) with respect to the back position of the reflector. fruit, The acquisition of scanning data is The method includes acquiring scan data including a deep point cloud at a position farther from a reflector identified based on identification information read from a storage medium (10) that stores map data (Dm) including identification information (σi) for identifying an object present in the scanning space, The generation of recognition data is This includes excluding from the recognition target a group of points on the far side where a group of symmetrical points exists in an area symmetrical to a position on the far side of a reflector whose identification information represents optical characteristics that allow the transmission of irradiated light.

[0012] In the first to fifth aspects, scanning data is acquired that includes a back point cloud as a scanning point cloud at a position further back than the reflector in the scanning direction of the scanning device that scans a reflector in a range of interest on the high-reflection side, where the reflectivity characteristics of the irradiated light are high. Therefore, according to the first to fourth aspects, recognition data is generated by excluding a back point cloud that is a scanning point cloud of a real image that creates a virtual image at a position further back than the reflector from the back point cloud that is the target of recognition of the target moving object in the scanning data, and a symmetrical point cloud that exists in a symmetrical area with respect to the back position of the reflector as the scanning point cloud of the real image that creates a virtual image at the position further back than the reflector. In this way, a back point cloud observed as a virtual image of the symmetrical point cloud at the back position of the reflector can be excluded from the target of recognition of the target moving object, while a back point cloud observed as a real image at the back position of the reflector can be properly recognized as the target moving object. Therefore, erroneous recognition due to the generation of virtual images can be suppressed, thereby improving the recognition accuracy of the target moving object. [Brief explanation of the drawings]

[0013] [Figure 1] 1 is a block diagram showing the overall configuration of a recognition system according to a first embodiment. [Figure 2] FIG. 2 is a schematic diagram illustrating the relationship between a scanning device of a host vehicle and a target moving object according to the first embodiment. [Figure 3] FIG. 1 is a block diagram showing a functional configuration of a recognition system according to a first embodiment. [Figure 4] 1 is a flowchart showing a recognition flow according to a first embodiment. [Figure 5] FIG. 4 is a schematic plan view for explaining a recognition flow according to the first embodiment. [Figure 6] FIG. 4 is a schematic plan view for explaining a recognition flow according to the first embodiment. [Figure 7] FIG. 4 is a schematic plan view for explaining a recognition flow according to the first embodiment. [Figure 8]FIG. 4 is a schematic plan view for explaining a recognition flow according to the first embodiment. [Figure 9] FIG. 4 is a schematic plan view for explaining a recognition flow according to the first embodiment. [Figure 10] FIG. 4 is a schematic plan view for explaining a recognition flow according to the first embodiment. [Figure 11] FIG. 2 is a schematic perspective view for explaining a recognition flow according to the first embodiment. [Figure 12] FIG. 2 is a schematic perspective view for explaining a recognition flow according to the first embodiment. [Figure 13] FIG. 4 is a schematic plan view for explaining a recognition flow according to the first embodiment. [Figure 14] 10 is a flowchart showing a recognition flow according to a second embodiment. [Figure 15] FIG. 10 is a schematic perspective view for explaining a recognition flow according to a second embodiment. [Figure 16] 10 is a flowchart showing a recognition flow according to a third embodiment. [Figure 17] FIG. 11 is a characteristic diagram for explaining a recognition flow according to a third embodiment. [Figure 18] FIG. 11 is a schematic perspective view for explaining a recognition flow according to a third embodiment. [Figure 19] FIG. 11 is a schematic perspective view for explaining a recognition flow according to a third embodiment. [Figure 20] 10 is a flowchart showing a recognition flow according to a modified example of the third embodiment to which the second embodiment is applied. [Figure 21] 10 is a flowchart showing a recognition flow according to a modified example of the second embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0014] Hereinafter, multiple embodiments of the present disclosure will be described with reference to the drawings. Note that corresponding components in each embodiment are designated by the same reference numerals, and redundant description may be omitted. Furthermore, when only a portion of the configuration is described in each embodiment, the configuration of another previously described embodiment may be applied to the remaining portions of the configuration. Furthermore, in addition to the combinations of configurations explicitly stated in the description of each embodiment, configurations of multiple embodiments may be partially combined together even if not explicitly stated, provided that there is no particular problem with the combination.

[0015] The recognition system 1 of the first embodiment shown in Fig. 1 recognizes a target moving object Ot that can move in a scanning space 30 scanned by light emitted from a scanning device 3 in a host vehicle 2 as shown in Fig. 2. Here, the host vehicle 2 to which the recognition system 1 is applied is a vehicle that can travel on a road with an occupant on board, such as an automobile. The target moving object Ot to be recognized by the recognition system 1 is of multiple types, such as other vehicles other than the host vehicle 2, motorcycles, people, animals, autonomous robots, and remote-controlled robots.

[0016] In the host vehicle 2, an autonomous driving mode is implemented so that the levels are classified according to the degree of manual intervention by the occupant in the dynamic driving task. The autonomous driving mode may be realized by autonomous driving control, such as conditional driving automation, high driving automation, or full driving automation, in which the system performs all dynamic driving tasks when activated. The autonomous driving mode may also be realized by advanced driving assistance control, such as driving assistance or partial driving automation, in which the occupant performs some or all of the dynamic driving tasks. The autonomous driving mode may be realized by either autonomous driving control or advanced driving assistance control, or by a combination of these, or by switching between them.

[0017] The host vehicle 2 is equipped with a sensor system 4, a communication system 5, and an information presentation system 6 shown in Fig. 1. The sensor system 4 acquires sensor information from the external and internal worlds of the host vehicle 2 that can be used for driving control of the host vehicle 2, including recognition control by the recognition system 1. To this end, the sensor system 4 is configured to include an external sensor 40 and an internal sensor 41.

[0018] The external sensor 40 acquires, as sensor information, information about the external environment surrounding the host vehicle 2. The external sensor 40 includes a scanning device 3 that acquires sensor information by scanning a scanning space 30, which is part of the external environment of the host vehicle 2, with irradiated light. Such a scanning device 3 is a three-dimensional LiDAR (Light Detection and Ranging / Laser Imaging Detection and Ranging) that scans the scanning space 30 using infrared laser light as irradiated light. Note that the external sensor 40 other than the scanning device 3 may include at least one type of device that senses the external environment of the host vehicle 2, such as a camera or sonar.

[0019] Here, the scanning device 3 generates sensor information by scanning a scanning space 30 (see FIG. 2 ) determined according to a field of view angle set toward the outside of the host vehicle 2 with illumination light. In particular, the sensor information generated by the scanning device 3 of the first embodiment is scan data Ds, which three-dimensionally represents the state of a group of scanning points observed as reflection points of illumination light from objects in the scanning space 30. The scan data Ds includes state values ​​related to at least one of distance, azimuth angle, position coordinates, velocity, and beam reflection intensity. Among the state values ​​included in the scan data Ds, distance may represent a measurement value obtained by dTOF (direct time of flight) based on the time of flight from the illumination light to receiving a reflected echo. Among the state values ​​included in the scan data Ds, the azimuth angle may represent a scanning direction that changes in at least one of the horizontal and vertical directions relative to the scanning space 30.

[0020] The internal sensors 41 acquire information about the internal environment of the host vehicle 2 as sensor information. The internal sensors 41 may include a physical quantity detection type that detects a specific physical quantity of motion in the internal environment of the host vehicle 2. The physical quantity detection type internal sensors 41 are, for example, at least one of a driving speed sensor, an acceleration sensor, an inertial sensor, etc. The internal sensors 41 may include an occupant detection type that detects a specific state of an occupant in the internal environment of the host vehicle 2. The occupant detection type internal sensors 41 are, for example, at least one of a Driver Status Monitor (registered trademark), a biological sensor, a seating sensor, an actuator sensor, an in-vehicle equipment sensor, etc.

[0021] The communication system 5 wirelessly acquires communication information that can be used for driving control of the host vehicle 2, including recognition control by the recognition system 1. The communication system 5 may include a V2X type that transmits and receives communication signals to and from a V2X system existing outside the host vehicle 2. The V2X type communication system 5 is, for example, at least one of a Dedicated Short Range Communications (DSRC) communication device and a Cellular V2X (C-V2X) communication device. The communication system 5 may include a positioning type that receives positioning signals from a Global Navigation Satellite System (GNSS) satellite existing outside the host vehicle 2. The positioning type communication system 5 is, for example, a GNSS receiver. The communication system 5 may include a terminal communication type that transmits and receives communication signals to and from a terminal existing inside the host vehicle 2. The terminal communication type communication system 5 is, for example, at least one of a Bluetooth (registered trademark) device, a Wi-Fi (registered trademark) device, an infrared communication device, etc.

[0022] The information presentation system 6 presents alarm information to the occupants in the host vehicle 2. The information presentation system 6 may be a visual stimulation type that stimulates the occupants' vision through a display. The visual stimulation type information presentation system 6 is, for example, at least one of a HUD (Head-Up Display), an MFD (Multi-Function Display), a combination meter, a navigation unit, etc. The information presentation system 6 may be an auditory stimulation type that stimulates the occupants' hearing through sound. The auditory stimulation type information presentation system 6 is, for example, at least one of a speaker, a buzzer, a vibration unit, etc.

[0023] The recognition system 1 is connected to a sensor system 4, a communication system 5, and an information presentation system 6 via at least one of, for example, a LAN (Local Area Network) line, a wire harness, an internal bus, or a wireless communication line. The recognition system 1 is configured to include at least one dedicated computer.

[0024] The dedicated computer constituting the recognition system 1 may be a recognition control ECU (Electronic Control Unit) that controls object recognition in the scanning space 30 based on the scan data Ds from the scanning device 3. Here, the recognition control ECU may have a function of integrating sensor information from multiple external sensors 40 including the scanning device 3. The dedicated computer constituting the recognition system 1 may be a driving control ECU that controls driving of the host vehicle 2. The dedicated computer constituting the recognition system 1 may be a navigation ECU that navigates the driving route of the host vehicle 2. The dedicated computer constituting the recognition system 1 may be a locator ECU that estimates the host vehicle 2's own state quantities including its own position. The dedicated computer constituting the recognition system 1 may be an HCU (Human Machine Interface Control Unit) that controls information presentation by the information presentation system 6 in the host vehicle 2. The dedicated computer constituting the recognition system 1 may be a computer other than the host vehicle 2 that constitutes, for example, an external center or a mobile terminal that can communicate with the communication system 5.

[0025] The dedicated computer constituting the recognition system 1 has at least one memory 10 and one processor 12. The memory 10 is at least one type of non-transitory tangible storage medium, such as a semiconductor memory, a magnetic medium, or an optical medium, that non-temporarily stores computer-readable programs and data. The processor 12 includes at least one type of core, such as a central processing unit (CPU), a graphics processing unit (GPU), a reduced instruction set computer (RISC)-CPU, a data flow processor (DFP), or a graph streaming processor (GSP).

[0026] In the recognition system 1, the memory 10 stores map information usable for driving control of the host vehicle 2. The memory 10 acquires and stores the latest map information, for example, by communicating with an external center via a V2X-type communication system 5. In particular, the map information in the first embodiment is map data Dm (see FIG. 3 ), such as a high-precision map or a dynamic map, that represents the driving environment of the host vehicle 2 in three dimensions. Such map data Dm represents the state of a mapping point cloud in which objects at fixed positions existing in the driving environment of the host vehicle 2 are mapped. The map data Dm includes three-dimensional state values ​​related to at least one of the following: position coordinates, distance, azimuth, and shape of the target. The objects mapped in the map data Dm are multiple types of objects located at at least fixed positions, for example, roads, signs, traffic lights, structures, railroad crossings, vegetation, space division objects, space division lines, and marking lines.

[0027] In the recognition system 1, the processor 12 executes a plurality of instructions included in a recognition program stored in the memory 10 in order to recognize a target moving object Ot in a scanning space 30 by the scanning device 3 of the host vehicle 2. As a result, the recognition system 1 constructs a plurality of functional blocks for recognizing the target moving object Ot in the scanning space 30. The plurality of functional blocks constructed in the recognition system 1 include a scanning block 100 and a recognition block 110, as shown in FIG.

[0028] The flow of the recognition method (hereinafter referred to as the recognition flow) in which the recognition system 1 recognizes a target moving object Ot in the scanning space 30 through the cooperation of these blocks 100 and 110 will be described below with reference to Figure 4. The algorithm cycle of the recognition flow is repeatedly executed while the host vehicle 2 is running. Note that each "S" in the recognition flow represents a plurality of steps executed by a plurality of instructions included in the recognition program.

[0029] In S100, the scanning block 100 acquires scan data Ds of the entire scanning space 30 according to the field of view angle from the scanning device 3. At this time, particularly in the recognition flow of the first embodiment, the scan data Ds is acquired so as to include at least the three-dimensional distance and / or three-dimensional position coordinates of the scanning point cloud as state values ​​observed for each of a plurality of pixels in the scanning device 3.

[0030] In S101, the scanning block 100 identifies, in the scan data Ds, the entire scanning direction (i.e., the irradiation direction of the irradiation light) ψs in which a reflector Or, among objects present in the scanning space 30, is scanned by the scanning device 3, as shown in FIGS. 5 to 8. Here, the reflector Or is defined as an object at a fixed position whose reflection characteristics with respect to the irradiation light are within a range of interest on the high-reflectivity side. The range of interest for the reflector Or refers to a range in which at least one of the reflection characteristics, such as reflectance and reflection intensity, is equal to or exceeds a reflection threshold. Therefore, the range of interest is set to a range of reflection characteristics in which the scanning point cloud of the virtual image Iv is observed when the reflection light, which is reflected by another object Oa in a direction different from the scanning direction ψs by the reflector Or, is reflected again by the reflector Or toward the scanning device 3, as shown in FIGS. 5 to 7.

[0031] In S101, the scanning block 100 reads map data Dm, to which identification information σi of objects present in the scanning space 30 is added, from the map storage area 10m of the memory 10 shown in Fig. 3, in order to identify a reflector Or to be identified in the scanning direction ψs. At this time, particularly in the recognition flow of the first embodiment, map data Dm including at least three-dimensional distances and / or three-dimensional position coordinates of the mapping point group as mapped state values ​​is read. Therefore, three-dimensional space information associated with identification information σi for each of a plurality of voxels 300, as described below, may be read as map data Dm acquired from an infrastructure database of an infrastructure system (e.g., an external center) that can communicate with the recognition system 1 via the communication system 5.

[0032] In the map data Dm to be read in S101, identification information σi is included as information for identifying the optical characteristics of an object in the scanning space 30, associated with the position coordinates of the object. In the first embodiment, the optical characteristics identified by the identification information σi include the transmission characteristics of the irradiated light as well as the reflection characteristics described above. Therefore, the identification information σi is added to the map data Dm to represent an object, such as a reflector Or, through which the transmission of the irradiated light is permitted when at least one type of transmission characteristic, such as transmittance or transmission intensity, is equal to or greater than a transmission threshold. Furthermore, the identification information σi is added to the map data Dm to represent an object, such as a reflector Or, through which the transmission of the irradiated light is restricted when at least one type of transmission characteristic is less than or equal to the transmission threshold.

[0033] 5 to 8, the scanning block 100 in S101 identifies the scanning direction ψs in which the scanning device 3 scans the reflector Or in three dimensions. At this time, the self-position of the host vehicle 2, which is the starting point of the scanning direction ψs in which the reflector Or is scanned, is estimated. The self-position estimation is based on at least one of sensor information from a physical quantity detection type internal sensor 41 such as an inertial sensor, communication information from a positioning type communication system 5, and map data Dm.

[0034] In S101, the scanning block 100 determines whether or not the scanning direction ψs in which the reflector Or exists has been identified, as shown in Fig. 4. As a result, if a positive determination is made, the recognition flow proceeds to S102.

[0035] 5 to 8, the scanning block 100 identifies, in the scan data Ds, a back point group Pb as a scanning point group observed at a position on the back side (i.e., the far side) of the reflector Or as viewed from the scanning device 3 in the entire scanning direction ψs in which the reflector Or is scanned. At this time, the back point group Pb is included in the scan data Ds as a scanning point group whose distance from the scanning device 3 in the scanning direction ψs is farther than the reflector Or identified in S101.

[0036] In S102, the scanning block 100 determines whether or not the back point group Pb has been identified, as shown in Fig. 4. If the result is a positive determination, the recognition flow proceeds to S103.

[0037] In S103, the recognition block 110 searches the scan data Ds for a transmission limiting point group Pbl at a position further back than the reflector Or whose identification information σi represents optical characteristics that limit the transmission of irradiated light, as shown in Fig. 5. At this time, the transmission limiting point group Pbl is defined as a back point group Pb that is farther from the scanning device 3 than the reflector Or where at least one type of transmission characteristic based on the identification information σi is less than the transmission threshold or equal to or less than the transmission threshold at a fixed point position.

[0038] In S103, the scanning block 100 determines whether the back point group Pb identified in S102 is a transmission-limited point group Pbl, as shown in Fig. 4. If a positive determination is made, the recognition flow proceeds to S104, and the recognition block 110 excludes the transmission-limited point group Pbl, which is determined to be a scanning point group of the virtual image Iv, from the recognition targets of the target moving object Ot, as shown in Fig. 5. On the other hand, if a negative determination is made, as shown in Fig. 4, the recognition flow proceeds to S105.

[0039] In S105, the recognition block 110 searches the scan data Ds for a quasi-transmission restricted point group Pbpl as a transmission-allowed point group Pbp located at a position behind a reflector Or whose identification information σi represents optical characteristics that allow the transmission of irradiation light, as shown in FIG. 6. In this case, the transmission-allowed point group Pbp is defined as a point group Pb farther from the scanning device 3 than the reflector Or, where at least one type of transmission characteristic based on the identification information σi is equal to or exceeds the transmission threshold at a fixed point position. Therefore, the quasi-transmission restricted point group Pbpl is defined as a point group Pb farther from the scanning device 3 than the fixed object Ol, where at least one type of transmission characteristic based on the identification information σi is less than or equal to the transmission threshold at a fixed point position behind the reflector Or. In other words, the quasi-transmission restricted point group Pbpl is a point group Pb identified at a position farther from the scanning device 3 than the fixed object Ol, where at least one type of transmission characteristic based on the identification information σi is less than or equal to the transmission threshold at a fixed point position behind the reflector Or.

[0040] In S105, the scanning block 100 determines whether the back point group Pb identified in S102 is a quasi-transparency-limited point group Pbpl, as shown in Fig. 4. If a positive determination is made, the recognition flow proceeds to S106, and the recognition block 110 excludes the quasi-transparency-limited point group Pbpl, which is determined to be the scanning point group of the virtual image Iv, from the recognition targets of the target moving object Ot, as shown in Fig. 6. On the other hand, if a negative determination is made, as shown in Fig. 4, the recognition flow proceeds to S107.

[0041] In S107, the recognition block 110 searches the scan data Ds for a virtual image point group Pbpv and a real image point group Pbpa as a transmission-permitted point group Pbp at a position behind the reflector Or, whose identification information σi represents optical characteristics that allow the transmission of irradiation light, as shown in Figures 7 and 8. At this time, as shown in Figures 7 and 9, the virtual image point group Pbpv is defined as a back point group Pb where a symmetric point group Ps exists in a symmetric area As that satisfies symmetry with the back position relative to the reflector Or within the search range, as a scanning point group of a real image Ia that causes a virtual image Iv to appear at a position behind the transmission-permitted reflector Or. On the other hand, as shown in Figures 8 and 10, the real image point group Pbpa is defined as a back point group Pb where a symmetric point group Ps does not exist in a symmetric area As that satisfies symmetry with the back position relative to the transmission-permitted reflector Or within the search range.

[0042] 9 and 10, the point groups Pbpv and Pbpa are determined based on the presence or absence of a symmetric point group Ps within a search range of a symmetric area As that includes a plane-symmetric position with the transmission-permitted point group Pbp, with a symmetry plane being a virtual plane Fv perpendicular to a normal line passing through points in the scanning direction ψs on the reflecting surface of the reflector Or. The search range for searching the symmetric point group Ps as the outermost range of the symmetric area As is set taking into consideration at least one of the scanning error in S100 and the self-localization error in S101. The search for the symmetric point group Ps is performed on the scanning point group included in the scan data Ds of the entire scanning space 30. However, if at least a portion of the symmetric area As is outside the scanning space 30, the mapping point group of the map data Dm may be substituted for the scanning point group.

[0043] As shown in FIGS. 11 and 12, the recognition block 110 in S107 sets a three-dimensional back area Ab in the scanning space 30 to identify the point groups Pbpv and Pbpa, relative to a three-dimensional symmetric area As in which the symmetric point group Ps is searched. The back area Ab and the symmetric area As are assumed to be substantially the same size so that they are plane-symmetric with respect to the transmittance-permitting reflector Or (specifically, the virtual plane Fv). Therefore, the recognition block 110 in S107 further divides each of the back area Ab and the symmetric area As into multiple three-dimensional voxels 300. Each of these voxels 300 is defined as a three-dimensional lattice space, a cube or rectangular parallelepiped, with six sides aligned along the three-dimensional absolute coordinate axes assigned to the scanning space 30. Hereinafter, each voxel 300 in the back area Ab will be referred to as a back voxel 300b, while each voxel 300 in the symmetric area As will be referred to as a symmetric voxel 300s.

[0044] In S107, the recognition block 110 compares the plane-symmetric voxels 300b and 300s. This comparison is based on the similarity of the point cloud distribution between the scanning point clouds of the deep voxel 300b, to which the transparency-permitted point cloud Pbp belongs as the deep point cloud Pb, and the symmetric voxel 300s, for which the symmetric point cloud Ps is searched. The similarity between the deep voxel 300b and the symmetric voxel 300s is calculated using at least one of the following features: an ICP (Iterative Closest Point) algorithm, a Mahalanobis distance, and a SHOT (Signature of Histograms of Orientations) feature.

[0045] In S107, if the similarity between the far voxel 300b and the symmetric voxel 300s falls within the allowable range, i.e., if the voxels 300b and 300s are determined to be similar within the allowable range, the permissible transmission point group Pbp is determined to be a virtual image point group Pbpv in which the symmetric point group Ps exists, as shown in Figures 7, 9, and 12. On the other hand, if the similarity between the far voxel 300b and the symmetric voxel 300s falls outside the allowable range, i.e., if the voxels 300b and 300s are determined to be dissimilar, the permissible transmission point group Pbp is determined to be a real image point group Pbpa in which the symmetric point group Ps does not exist, as shown in Figures 8, 10, and 11.

[0046] In S107, the recognition block 110 may further perform data analysis on the transparency-permitted point group Pbp belonging to the far-side voxel 300b determined to be similar to the symmetric voxel 300s, thereby improving the accuracy of distinguishing between the point groups Pbpv and Pbpa. This data analysis may involve, for example, clustering based on the distance between points and the normal direction of each point, followed by tracking using an extended Kalman filter for the same cluster to estimate the speed and direction of travel. Based on the results of this data analysis, the recognition block 110 recognizes the transparency-permitted point group Pbp moving in the opposite direction to the host vehicle 2 at a relative speed of interest (hereinafter simply referred to as the relative speed of interest) corresponding to the traveling speed of the host vehicle 2 as the true virtual image point group Pbpv, as shown in FIG. 13. In contrast, the recognition block 110 recognizes the transparency-permitted point group Pbp moving in the opposite direction to the host vehicle 2 outside the relative speed of interest and the transparency-permitted point group Pbp moving in the same direction as the host vehicle 2 as the false virtual image point group Pbpv, as the real image point group Pbpa.

[0047] Here, the target relative speed, which is the discrimination criterion for the point groups Pbpv and Pbpa, is defined as the speed difference between the traveling speed of the host vehicle 2 measured by the speed sensor serving as the internal sensor 41 and the estimated speed of the permeable point group Pbp measured by tracking processing. Therefore, the permeable point group Pbp whose target relative speed moving in the opposite direction to the host vehicle 2 is less than or equal to the speed threshold is discriminated as the virtual image point group Pbpv.

[0048] In S107, the recognition block 110 determines whether the deep point group Pb identified in S102 is a virtual image point group Pbpv or a real image point group Pbpa, as shown in Fig. 4. If the determination is that the deep point group Pb is a virtual image point group Pbpv, the recognition flow proceeds to S108, whereby the recognition block 110 excludes the virtual image point group Pbpv, which is determined to be the scanning point group of the virtual image Iv, from the recognition targets of the target moving object Ot, as shown in Fig. 7. On the other hand, if the determination is that the deep point group Pbpa is a real image point group Pbpa, as shown in Fig. 4, the recognition flow proceeds to S109, whereby the recognition block 110 extracts the real image point group Pbpa, which is determined to be the scanning point group of the real image Ia, as the recognition target of the target moving object Ot, as shown in Fig. 8.

[0049] The recognition flow proceeds to S110 after any of S104, S106, S108, and S109 are executed. The recognition flow also proceeds to S110 if a negative determination is made in S101 or S102. In S110, the recognition block 110 excludes the back point group Pb, which was determined not to be a recognition target in the immediately preceding step among the transmission-limited point group Pbl, the semi-transmission-limited point group Pbpl, and the virtual image point group Pbpv, from the recognition targets in the scan data Ds, and then performs recognition processing. For example, a machine learning model such as Point Pillars or a background subtraction method using the map data Dm is used for the recognition processing.

[0050] The recognition block 110 in S110 generates recognition data Dr representing the results of this recognition processing. At this time, if the real image point group Pbpa is determined to be the recognition target in the immediately preceding step S109, recognition data Dr that recognizes the target moving object Ot represented by the real image point group Pbpa is generated. The recognition block 110 in S110 further stores the generated recognition data Dr in a recognition storage area 10r of the memory 10 shown in FIG. 3. The stored recognition data Dr is used, for example, for driving control of the host vehicle 2.

[0051] The recognition block 110 in S110 may control the display of the generated or stored recognition data Dr so that it is displayed by the information presentation system 6 in the host vehicle 2. In the recognition data Dr displayed at this time, the rear point group Pb that is not the recognition target may be controlled not to be displayed, or the rear point group Pb that is not the recognition target may be displayed together with a warning label such as "virtual image." The recognition block 110 in S110 may also control the transmission of the generated or stored recognition data Dr from the host vehicle 2 to an external party (e.g., an external center or another vehicle) via the communication system 5. This completes the current execution of the recognition flow.

[0052] (Action and effect) The effects of the first embodiment described above will be explained below.

[0053] In the first embodiment, in the scanning direction ψs of the scanning device 3 scanning a reflector Or in a range of interest on the high reflectivity side with respect to the irradiated light, scan data Ds containing a far-side point group Pb as a scanning point group at a position farther from the reflector Or is acquired. According to the first embodiment, recognition data Dr is generated by excluding from the far-side point group Pb to be recognized as the target moving object Ot in the scan data Ds the far-side point group Pb, which is a scanning point group of a real image Ia that causes a virtual image Iv to appear at a position farther from the reflector Or and which includes a symmetrical point group Ps in a symmetrical area As with respect to the far-side position of the reflector Or. According to this, the far-side point group Pb observed as the virtual image Iv of the symmetrical point group Ps at the far-side position of the reflector Or can be excluded from the recognition target of the target moving object Ot, while the far-side point group Pb observed as the real image Ia at the far-side position of the reflector Or can be properly recognized as the target moving object Ot. Therefore, it is possible to suppress erroneous recognition due to the occurrence of the virtual image Iv and improve the recognition accuracy of the target moving object Ot.

[0054] According to the first embodiment, scan data Ds containing a back point group Pb is acquired at a back position of a reflector Or specified based on identification information σi, which is read from memory 10 as information contained in map data Dm to identify an object present in the scanning space 30. This allows candidates for the back point group Pb observed as a virtual image Iv caused by the reflector Or of the symmetric point group Ps to be appropriately narrowed down based on the identification information σi. Therefore, it is possible to increase the recognition accuracy of the target moving object Ot while maximizing the speed of processing required to suppress erroneous recognition due to the generation of the virtual image Iv.

[0055] According to the first embodiment, the far-side point group Pb, in which the symmetric point group Ps exists in the symmetric area As with a position farther from the reflector Or, whose identification information σi represents optical characteristics that allow the transmission of irradiated light, is excluded from the recognition target. Accordingly, the far-side point group Pb observed at a position farther from the reflector Or as a virtual image Iv of the symmetric point group Ps can be excluded from the recognition target of the target moving object Ot in accordance with the search for the symmetric point group Ps in a situation where the reflector Or allows the transmission of irradiated light. Therefore, it is possible to suppress erroneous recognition due to the generation of the virtual image Iv and improve the recognition accuracy of the target moving object Ot.

[0056] According to the first embodiment, the far-side point group Pb, which is located behind the reflector Or whose identification information σi represents optical characteristics that allow the transmission of the irradiated light and behind the fixed object Ol whose identification information σi represents optical characteristics that limit the transmission of the irradiated light, is excluded from the recognition target. According to this, the far-side point group Pb observed at the far-side position of the reflector Or as a virtual image Iv of the symmetric point group Ps can be excluded from the recognition target of the target moving object Ot in accordance with the identification of the fixed object Ol that limits the transmission of the irradiated light, even if the reflector Or allows the transmission of the irradiated light. Therefore, it is possible to increase the speed of the processing required to suppress erroneous recognition due to the generation of the virtual image Iv as much as possible while improving the recognition accuracy of the target moving object Ot.

[0057] According to the first embodiment, the back point group Pb located behind the reflector Or, whose identification information σi represents optical characteristics that limit the transmission of irradiated light, is excluded from the recognition target. This allows the back point group Pb observed at a position behind the reflector Or as a virtual image Iv of the symmetric point group Ps to be excluded from the recognition target of the target moving object Ot in accordance with the identification of the reflector Or that limits the transmission of irradiated light. Therefore, it is possible to increase the speed of the processing required to suppress erroneous recognition due to the generation of the virtual image Iv as much as possible while improving the recognition accuracy of the target moving object Ot.

[0058] In the first embodiment, the scanning space 30 is divided into multiple three-dimensional voxels 300, which are defined as a deep voxel 300b to which the deep point group Pb belongs and a symmetric voxel 300s to which the symmetric point group Ps is searched. Therefore, according to the first embodiment, the deep point group Pb belonging to the deep voxel 300b, whose point group distribution is similar to that of the symmetric voxel 300s within an allowable range, is excluded from the recognition target. This allows the deep point group Pb, which is the observed virtual image Iv of the symmetric point group Ps caused by the reflector Or, to be accurately identified based on the similarity of the point group distribution in the local range of the voxels 300b and 300s. This improves the reliability of suppressing erroneous recognition due to the generation of the virtual image Iv, and ultimately the reliability of highly accurate recognition of the target moving object Ot.

[0059] According to the first embodiment, the far-side point group Pb, which moves in the opposite direction to the host vehicle 2 at a relative speed corresponding to the traveling speed of the host vehicle 2, is excluded from the recognition target. This makes it possible to accurately identify the far-side point group Pb, which observes the virtual image Iv of the symmetric point group Ps caused by the reflector Or, based on the relative speed and moving direction of the far-side point group Pb with respect to the host vehicle 2. This makes it possible to increase the reliability in suppressing erroneous recognition caused by the generation of the virtual image Iv, and ultimately the reliability in highly accurate recognition of the target moving object Ot.

[0060] Second Embodiment The second embodiment is a modification of the first embodiment.

[0061] As shown in FIG. 14 , in the recognition flow of the second embodiment, if a negative determination is made in S105, the process proceeds to S2107. In S2107, the recognition block 110 searches for aerial points Pbpx using optical characteristics that allow the transmission of irradiated light as a transmission-permitted point group Pbp located behind the reflector Or represented by the identification information σi (see FIG. 3 of the first embodiment). In this case, the aerial points Pbpx are searched for based on information from the identification information σi included in the map data Dm that identifies the three-dimensional position coordinates of the aerial area Ax located behind the reflector Or as shown in FIG. 15 . In other words, the searched aerial points Pbpx are defined as a scanning point group of the virtual image Iv relative to the real image Ia, that is, a back point group Pb located in the aerial area Ax represented by the identification information σi located behind the reflector Or.

[0062] 14, if a positive determination is made in S2107, in the second embodiment, the recognition flow proceeds to S2108, whereby the recognition block 110 excludes the aerial point group Pbpx from the recognition targets of the target moving body Ot, and then S110 is executed. On the other hand, if a negative determination is made in S2107, the recognition flow proceeds to S107, whereby the transmission-allowed point group Pbp, excluding the quasi-transmission-limited point group Pbpl and the aerial point group Pbpx, is determined to be either the virtual image point group Pbpv or the real image point group Pbpa, as in the first embodiment.

[0063] As described above, according to the second embodiment, the far-side point group Pb present in the aerial area Ax represented by the identification information σi among positions further back than the reflector Or is excluded from the recognition target. Accordingly, the far-side point group Pb observed at a position further back than the reflector Or as a virtual image Iv of the symmetric point group Ps can be excluded from the recognition target of the target moving object Ot in accordance with the identification of its location in the aerial area Ax. Therefore, it is possible to suppress erroneous recognition due to the occurrence of the virtual image Iv and improve the recognition accuracy of the target moving object Ot.

[0064] (Third embodiment) The third embodiment is a modification of the first embodiment.

[0065] 16, in the recognition flow of the third embodiment, S3100 and S3107 are executed instead of S100 and S107, respectively. Specifically, in S3100, the scanning block 100 acquires scan data Ds by receiving a reflected echo Er from the scanning space 30 in response to the irradiation light for each of a plurality of pixels in the scanning device 3. At this time, when at least one echo is received for each pixel in the scanning device 3 as a reflected echo Er having an intensity exceeding a threshold value Rt as shown in FIG. 17, the state value of the scanning point cloud is converted into digital data and stored in the memory 10.

[0066] Therefore, for each pixel in the third embodiment, maximum intensity scan data Dsm is defined as scan data Ds containing a scanning point cloud whose state values ​​corresponding to the reflected echo Erm with the maximum intensity Rm among the received reflected echoes Er, and total intensity scan data Dsa is defined as scan data Ds containing a scanning point cloud whose state values ​​corresponding to the received reflected echoes Er with all intensities, respectively, are three-dimensional distances and / or three-dimensional position coordinates.

[0067] 16, maximum intensity scan data Dsm is acquired by the scanning block 100 to be used in the subsequent processes of S101 to S103, S105, and S110. Meanwhile, full intensity scan data Dsa is acquired by the scanning block 100 in S3100 (FIG. 16 is an example of this) or S3107 to be used in the process of S3107. However, when the full intensity scan data Dsa is acquired before the determination of the virtual image point cloud Pbpv in S3107, it may be acquired as point cloud data of state values ​​in the entire pixel region, or may be acquired as point cloud data of state values ​​in a partial pixel region including voxels 300s and 300b to be compared as described below.

[0068] Furthermore, in S3107, the recognition block 110 uses the transmission-allowed point group Pbp excluding the quasi-transmission-limited point group Pbpl in the full-intensity scan data Dsa shown in FIG. 19 as the discrimination target for the virtual image point group Pbpv and the real image point group Pbpa, instead of the transmission-allowed point group Pbp excluding the quasi-transmission-limited point group Pbpl in the maximum-intensity scan data Dsm shown in FIG. 18. Therefore, in S3107, the recognition block 110 searches for the virtual image point group Pbpv belonging to the deep voxel 300b whose point cloud distribution is within the allowable range and similar to the symmetric voxel 300s, according to the first embodiment, among the transmission-allowed point groups Pb other than the quasi-transmission-limited point group Pbpl in the full-intensity scan data Dsa. As a result, the recognition flow proceeds from S3107 to S108 as shown in FIG. 16, and the discriminated virtual image point group Pbpv is excluded from the recognition targets for the target moving object Ot.

[0069] According to the third embodiment described above, the reflected echo Er from the scanning space 30 in response to the illumination light is received for each of a plurality of pixels in the scanning device 3. Therefore, in the maximum intensity scan data Dsm, which includes the scanning point clouds corresponding to the reflected echo Er of the maximum intensity Rm received for each pixel, if a symmetric point cloud Ps for the far point cloud Pb exists, the full intensity scan data Dsa is further utilized. Specifically, in the full intensity scan data Dsa, which includes the scanning point clouds corresponding to the reflected echoes Er of all intensities received for each pixel, the far point cloud Pb belonging to the far voxel 300b, whose point cloud distribution is similar to the symmetric voxel 300s within an allowable range, is excluded from the recognition target.

[0070] According to the third embodiment, the number of points contained in the far-side voxel 300b, which is the far-side point group Pb obtained by observing the virtual image Iv of the symmetric point group Ps generated by the reflector Or, can be increased in accordance with the number of reflected echoes of all intensities. Therefore, based on the similarity between the point group distributions in the voxels 300b and 300s, the accuracy of identifying the far-side point group Pb obtained by observing the virtual image Iv can be improved. As a result, it is possible to ensure high reliability in suppressing erroneous recognition due to the generation of the virtual image Iv, and therefore high reliability in accurately recognizing the target moving object Ot.

[0071] Furthermore, according to the third embodiment, the deep point group Pb located further back than the object Or, Ol, whose optical properties limit the transmission of irradiated light, can be searched for with high discrimination accuracy based on the identification information σi in a short time using the maximum intensity scan data Dsm, which has a smaller number of points than the full intensity scan data Ds. On the other hand, if a symmetric point group Ps exists for the deep point group Pb located further back than the reflector Or, whose optical properties allow the transmission of irradiated light, the full intensity scan data Dsa is also used. Specifically, the deep point group Pb belonging to the deep voxel 300b, whose point group distribution is similar to the symmetric voxel 300s within the allowable range, can be searched for with high discrimination accuracy using the full intensity scan data Dsa, which has a larger number of points than the maximum intensity scan data Dsm. As a result, it is possible to speed up processing and ensure high reliability in suppressing erroneous recognition due to the generation of virtual images Iv, thereby ensuring high reliability in accurately recognizing the target moving object Ot.

[0072] (Other embodiments) Although several embodiments have been described above, the present disclosure should not be construed as being limited to those embodiments, and can be applied to various embodiments within the scope of the gist of the present disclosure.

[0073] In a modified example, the dedicated computer constituting the recognition system 1 may have at least one of a digital circuit and an analog circuit as a processor. Here, the digital circuit is at least one of the following: an application specific integrated circuit (ASIC), a field programmable gate array (FPGA), a system on a chip (SOC), a programmable gate array (PGA), and a complex programmable logic device (CPLD). Such a digital circuit may also have a memory that stores a program.

[0074] In a modified example, at least one of the set of S103 and S104 and the set of S105 and S106 may be omitted. In a modified example, S103 to S109, S2107, S2108, and S3107 may be executed independently for each scanning direction ψs of multiple reflectors Or identified in S101. In a modified example, the identification of reflectors Or in S101 may be realized based on past scan data Ds instead of or in addition to being based on map data Dm.

[0075] In a modified example, S2107 and S2108 of the second embodiment may be applied to the third embodiment, as shown in Fig. 20. In a modified example, the recognition flow of the second embodiment may proceed to S103 when a negative determination is made in S2107, which is executed following a positive determination in S102, as shown in Fig. 21. However, in S2107 following the positive determination in S102, as a scanning point group at a position behind the reflector Or, a mid-air point group Pbx existing in the aerial area Ax represented by the identification information σi is searched for from among the back point group Pb, which is in an unsearched state for the point groups Pbl, Pbpl, Pbpv, and Pbpa to be searched in the subsequent steps S103, S105, and S107, and the aerial point group Pbx may then be excluded from the recognition target in S2108.

[0076] In a modified example, the host vehicle 2 to which the recognition system 1 is applied may be, for example, an autonomous robot capable of transporting luggage or collecting information by autonomous or remote driving. In addition to the forms described so far, the above-mentioned embodiments and modified examples may be implemented in the form of a processing circuit (e.g., a processing ECU, etc.) or a semiconductor device (e.g., a semiconductor chip, etc.) as a control device that is configured to be mountable on the host vehicle 2 and has at least one processor 12 and one memory 10.

[0077] (Additional remarks) This specification discloses the following technical ideas and combinations thereof.

[0078] (Technical thought 1) A recognition system having a processor (12) for recognizing a target moving object (Ot) capable of moving in a scanning space (30) scanned by irradiation light from a scanning device (3) in a host vehicle (2), comprising: The processor In a scanning direction (ψs) in which a reflector (Or) in a range of interest having high reflectivity characteristics with respect to the irradiated light is scanned, scanning data (Ds) including a back point group (Pb) as a scanning point group at a position further back than the reflector is acquired; A recognition system configured to generate recognition data (Dr) by excluding, from the back point cloud to be recognized of the target moving body in the scanning data, the back point cloud in which a symmetric point cloud (Ps) exists in a symmetric area (As) with respect to the back position of the reflector as the scanning point cloud of the real image (Ia) that makes a virtual image (Iv) appear at a position further back than the reflector.

[0079] (Technical thought 2) a storage medium (10) for storing map data (Dm) containing identification information (σi) for identifying an object present in a scanning space; The acquisition of scanning data is A recognition system according to technical idea 1, which includes acquiring scanning data containing a cloud of points at a position further back than a reflector identified based on identification information read from a storage medium.

[0080] (Technical Thought 3) The generation of recognition data is A recognition system according to technical idea 2, which includes excluding from the recognition target a group of points at the rear side where a group of symmetrical points exists in a symmetrical area with respect to a position at the rear side of a reflector whose identification information represents optical characteristics that allow the transmission of irradiated light.

[0081] (Technical Thought 4) The generation of recognition data is A recognition system according to technical idea 3, which includes excluding from the recognition target a group of points located further back than a reflector whose identification information represents optical characteristics that allow the transmission of irradiated light and further back than a fixed object (O1) whose identification information represents optical characteristics that restrict the transmission of irradiated light.

[0082] (Technical Thought 5) The generation of recognition data is A recognition system described in Technical Idea 3 or 4, which includes excluding from the recognition target a group of points located further back than a reflector whose identification information represents optical characteristics that limit the transmission of irradiated light.

[0083] (Technical Thought 6) The generation of the recognition data includes: A recognition system described in any one of technical ideas 2 to 5, which includes excluding the back point group that exists in the aerial area (Ax) represented by the identification information at a position further back than the reflector from the recognition target.

[0084] (Technical Thought 7) The generation of recognition data is A recognition system described in any one of technical ideas 1 to 6, which includes defining a deep voxel (300b) to which a deep point group belongs and a symmetric voxel (300s) to which a symmetric point group is searched as multiple three-dimensional voxels obtained by dividing the scanning space, and excluding from recognition targets deep point groups belonging to deep voxels whose point group distribution is similar to the symmetric voxel within an acceptable range.

[0085] (Technical Thought 8) The acquisition of the scanning data includes: receiving a reflected echo (Er) from the scanning space in response to the irradiated light for each of a plurality of pixels in the scanning device; Acquiring maximum intensity scan data (Dsm) as the scan data including the scanning point group corresponding to the maximum intensity reflected echo received for each pixel; Acquiring total intensity scan data (Dsa) as the scan data including the scanning point cloud corresponding to the reflected echoes of all intensities received for each pixel, The generation of the recognition data includes: A recognition system according to Technical Idea 7, which includes excluding from the recognition target the back point group belonging to the back voxel whose point group distribution in the full intensity scanning data is similar to the symmetric voxel within an allowable range when the symmetric point group for the back point group exists in the maximum intensity scanning data.

[0086] (Technical Thought 9) The generation of the recognition data includes: Searching for the deep point group at a position deeper than an object (Or, Ol) having optical characteristics that limit the transmission of the irradiation light in the maximum intensity scanning data, and excluding it from the recognition target; A recognition system according to Technical Idea 8, which includes, when a symmetric point group exists for the deep point group at a position further back than the reflector having optical properties that allow the transmission of the irradiation light in the maximum intensity scanning data, searching for the deep point group belonging to the deep voxel whose point group distribution in the full intensity scanning data is similar to the symmetric voxel within an allowable range, and excluding it from the recognition target.

[0087] (Technical Thought 10) The generation of recognition data is A recognition system described in any one of technical ideas 1 to 9, which includes excluding from recognition targets a back point cloud that moves in the opposite direction to the host vehicle at a relative speed corresponding to the traveling speed of the host vehicle.

[0088] (Technical Thought 11) The generation of recognition data is The recognition system according to any one of the technical ideas 1 to 10, further comprising storing the recognition data in a storage medium (10) in the host vehicle.

[0089] (Technical Thought 12) The generation of recognition data is 12. The recognition system according to any one of Technical Ideas 1 to 11, further comprising controlling the display of the recognition data in the host vehicle.

[0090] (Technical Thought 13) The generation of recognition data is The recognition system according to any one of Technical Ideas 1 to 12, further comprising controlling transmission of recognition data from a host vehicle.

[0091] The above-mentioned technical ideas 1 to 13 may be realized in the form of a recognition device, a recognition method, a recognition program, and a recognition data generation method. [Explanation of symbols]

[0092] 1: Recognition system, 2: Host vehicle, 3: Scanning device, 10: Memory, 12: Processor, 30: Scanning space, 300b: Deep voxel, 300s: Symmetric voxel, As: Symmetric area, Ax: Air area, Dm: Map data, Dr: Recognition data, Ds: Scanning data, Dsa: Full intensity scan data, Dsm: Maximum intensity scan data, Ia: Real image, Iv: Virtual image, Ol: Fixed object, Or: Reflector, Ot: Target moving object, Pb: Deep point cloud, Ps: Symmetric point cloud, Px: Air point cloud, σi: Identification information, ψs: Scanning direction

Claims

1. A recognition system having a processor (12) and a storage medium (10), for recognizing a target moving object (Ot) capable of moving in a scanning space (30) scanned by irradiation light from a scanning device (3) in a host vehicle (2), comprising: The processor: In a scanning direction (ψs) in which a reflector (Or) in a range of interest having a high reflectivity characteristic for the irradiated light is scanned, scanning data (Ds) including a back point group (Pb) as a scanning point group at a position further back than the reflector is acquired; and generating recognition data (Dr) by excluding, from the deep point cloud to be recognized by the target moving body in the scanning data, a scanning point cloud of a real image (Ia) that makes a virtual image (Iv) appear at a position deeper than the reflector, the deep point cloud in which a symmetrical point cloud (Ps) exists in a symmetrical area (As) with respect to the deep position of the reflector, and The acquisition of the scanning data includes: storing map data (Dm) including identification information (σi) for identifying an object present in the scanning space; and acquiring the scanning data including the deep point cloud at a position deeper than the reflector identified based on the identification information read from the storage medium; The generation of the recognition data includes: and excluding from the recognition target the back point group, in which the symmetric point group exists in the symmetric area with respect to a position further back than the reflector whose identification information represents optical characteristics that allow the transmission of the irradiated light.

2. The generation of the recognition data includes:

2. The recognition system according to claim 1, further comprising: excluding, from the recognition target, the deep point group located at a position farther than the reflector whose identification information represents optical characteristics that allow the transmission of the irradiated light and at a position farther than a fixed-point object (O1) whose identification information represents optical characteristics that limit the transmission of the irradiated light.

3. The generation of the recognition data includes: The recognition system according to claim 1 , further comprising: excluding, from the recognition target, the back point group located on the back side of the reflector, the identification information of which represents optical characteristics that limit the transmission of the irradiated light.

4. The generation of the recognition data includes: The recognition system according to claim 1 , further comprising excluding the back point group that exists in the aerial area (Ax) represented by the identification information among positions behind the reflector from the recognition target.

5. The generation of the recognition data includes: The recognition system according to any one of claims 1 to 4, further comprising: defining a back voxel (300b) to which the back point group belongs and a symmetric voxel (300s) to which the symmetric point group is searched as a plurality of three-dimensional voxels obtained by dividing the scanning space; and excluding from the recognition target the back point group belonging to the back voxel whose point group distribution is similar to the symmetric voxel within an acceptable range.

6. The acquisition of the scanning data includes: receiving a reflected echo (Er) from the scanning space in response to the irradiated light for each of a plurality of pixels in the scanning device; Acquiring maximum intensity scan data (Dsm) as the scan data including the scanning point group corresponding to the maximum intensity reflected echo received for each pixel; Acquiring total intensity scan data (Dsa) as the scan data including the scan point group corresponding to the reflected echoes of all intensities received for each pixel, The generation of the recognition data includes: The recognition system of claim 5, further comprising: when a symmetric point group exists for the deep point group in the maximum intensity scanning data, excluding the deep point group belonging to the deep voxel whose point group distribution in the full intensity scanning data is similar to the symmetric voxel within an acceptable range from the recognition target.

7. The generation of the recognition data includes: Searching for the deep side point group at a position deeper than an object (Or, Ol) having optical characteristics that limit the transmission of the irradiation light in the maximum intensity scanning data, and excluding the deep side point group from the recognition target; 7. The recognition system of claim 6, further comprising: when a symmetric point group exists for the deep point group at a position further back than the reflector having optical properties that allow the transmission of the irradiation light in the maximum intensity scanning data, searching for the deep point group belonging to the deep voxel whose point group distribution in the full intensity scanning data is similar to the symmetric voxel within an allowable range, and excluding the deep point group from the recognition target.

8. The generation of the recognition data includes: The recognition system according to any one of claims 1 to 4, further comprising excluding the rear point cloud, which moves in the opposite direction to the host vehicle at a relative speed corresponding to the traveling speed of the host vehicle, from the recognition target.

9. The generation of the recognition data includes: A recognition system according to any one of claims 1 to 4, comprising storing said recognition data in a storage medium (10) in said host vehicle.

10. The generation of the recognition data includes: A recognition system according to any preceding claim, comprising controlling the display of the recognition data in the host vehicle.

11. The generation of the recognition data includes:

5. The recognition system of claim 1, further comprising: controlling transmission of the recognition data from the host vehicle.

12. A recognition device having a processor (12) and a storage medium (10), configured to be mountable on a host vehicle (2), for recognizing a target moving object (Ot) capable of moving in a scanning space (30) scanned by irradiation light from a scanning device (3) in the host vehicle, The processor: In a scanning direction (ψs) in which a reflector (Or) in a range of interest having a high reflectivity characteristic for the irradiated light is scanned, scanning data (Ds) including a back point group (Pb) as a scanning point group at a position further back than the reflector is acquired; and generating recognition data (Dr) by excluding, from the deep point cloud to be recognized by the target moving body in the scanning data, a scanning point cloud of a real image (Ia) that makes a virtual image (Iv) appear at a position deeper than the reflector, the deep point cloud in which a symmetrical point cloud (Ps) exists in a symmetrical area (As) with respect to the deep position of the reflector, and The acquisition of the scanning data includes: storing map data (Dm) including identification information (σi) for identifying an object present in the scanning space; and acquiring the scanning data including the deep point cloud at a position deeper than the reflector identified based on the identification information read from the storage medium; The generation of the recognition data includes: A recognition device including: excluding from the recognition target the back point group in which the symmetric point group exists in the symmetric area with respect to a position further back than the reflector whose identification information represents optical characteristics that allow the transmission of the irradiated light.

13. A recognition method executed by a processor (12) for recognizing a target moving object (Ot) capable of moving in a scanning space (30) scanned by illumination light from a scanning device (3) in a host vehicle (2), comprising: In a scanning direction (ψs) in which a reflector (Or) in a range of interest having a high reflectivity characteristic for the irradiated light is scanned, scanning data (Ds) including a back point group (Pb) as a scanning point group at a position further back than the reflector is acquired; generating recognition data (Dr) by excluding, from the deep point cloud to be recognized by the target moving body in the scanning data, a scanning point cloud of a real image (Ia) that makes a virtual image (Iv) appear at a position deeper than the reflector, the deep point cloud having a symmetrical point cloud (Ps) present in a symmetrical area (As) with respect to the deep position of the reflector, The acquisition of the scanning data includes: acquiring the scan data including the deep point cloud at a position farther from the reflector identified based on the identification information read from a storage medium (10) that stores map data (Dm) including identification information (σi) for identifying an object present in the scanning space; The generation of the recognition data includes: A recognition method including: excluding from the recognition target the back point group in which the symmetric point group exists in the symmetric area with respect to a position further back than the reflector whose identification information represents optical characteristics that allow the transmission of the irradiated light.

14. A recognition program including instructions stored in a storage medium (10) and executed by a processor (12) for recognizing a target moving object (Ot) capable of moving in a scanning space (30) scanned by irradiation light from a scanning device (3) in a host vehicle (2), The instruction: In a scanning direction (ψs) in which a reflector (Or) in a range of interest having a high reflectivity characteristic for the irradiated light is scanned, scanning data (Ds) including a back point group (Pb) as a scanning point group at a position further back than the reflector is acquired. generating recognition data (Dr) by excluding, from the deep point cloud to be recognized by the target moving body in the scanning data, a scanning point cloud of a real image (Ia) that makes a virtual image (Iv) appear at a position deeper than the reflector, the deep point cloud having a symmetrical point cloud (Ps) present in a symmetrical area (As) with respect to the deep position of the reflector, The acquisition of the scanning data includes: storing map data (Dm) including identification information (σi) for identifying an object present in the scanning space; and acquiring the scanning data including the deep point cloud at a position deeper than the reflector identified based on the identification information read from the storage medium; The generation of the recognition data includes: A recognition program including: excluding from the recognition target the back point group in which the symmetric point group exists in the symmetric area with respect to a position further back than the reflector whose identification information represents optical characteristics that allow the transmission of the irradiated light.

15. A recognition data generation method executed by a processor (12) for recognizing a target moving object (Ot) capable of moving in a scanning space (30) scanned by irradiation light from a scanning device (3) in a host vehicle (2) and generating recognition data (Dr), comprising: In a scanning direction (ψs) in which a reflector (Or) in a range of interest having a high reflectivity characteristic for the irradiated light is scanned, scanning data (Ds) including a back point group (Pb) as a scanning point group at a position further back than the reflector is acquired; generating the recognition data by excluding, from the deep point cloud to be recognized by the target moving body in the scanning data, a scanning point cloud of a real image (Ia) that makes a virtual image (Iv) appear at a position deeper than the reflector, the deep point cloud having a symmetrical point cloud (Ps) present in a symmetrical area (As) with respect to the deep position of the reflector, The acquisition of the scanning data includes: acquiring the scan data including the deep point cloud at a position farther from the reflector identified based on the identification information read from a storage medium (10) that stores map data (Dm) including identification information (σi) for identifying an object present in the scanning space; The generation of the recognition data includes: A recognition data generation method including: excluding from the recognition target the back point group in which the symmetric point group exists in the symmetric area with respect to a position further back than the reflector whose identification information represents optical characteristics that allow the transmission of the irradiated light.

Images