Object recognition device and method thereof

Abstract

The invention relates to an object recognition apparatus and a method thereof. An apparatus may include a first LIDAR, a second LIDAR, and a processor. The processor may be based on obtaining a signal from a map system indicative of a confined space in which the host vehicle is located, the contour points of the object are identified by using or changing the points obtained by the first LIDAR and included in the specific frame and the points obtained by any one of the first LIDAR and the second LIDAR among the points obtained by the second LIDAR and included in the specific frame. A point obtained by the first LIDAR and included in the specific frame may be obtained from the same object as a point obtained by the second LIDAR and included in the specific frame.

Description

[0001] Related Applications

[0002] This application claims the benefit of priority from Korean Patent Application No. 10-2023-0107849 filed in the Korean Intellectual Property Office on August 17, 2023, the entire contents of which are incorporated herein by reference. Technical Field

[0003] The present disclosure relates to object recognition apparatus and methods, and more particularly, to techniques for recognizing contour points of an object based on points obtained by one or more sensors (e.g., light detection and ranging (LIDAR) and / or any other object sensor). Background Art

[0004] Various studies have been conducted to develop autonomous vehicle technology. In order for a vehicle to respond to unexpected situations and adjust its driving path without driver manipulation, it may be necessary to obtain and utilize more accurate information about the surrounding environment.

[0005] The vehicle may obtain data indicating the position of an object around the vehicle through one or more sensors (e.g., LIDAR). The LIDAR may obtain the distance from the LIDAR to the object through the interval between the time point of emitting laser light and the time point of receiving the laser light reflected on the object. The vehicle may identify the position of a point included in the object in the space where the vehicle is located based on the angle of the transmitted laser light and the distance to the object.

[0006] For a wide range of measurements, multiple LIDARs may be mounted on an autonomous vehicle, and therefore, how to process data obtained by the multiple LIDARs has been discussed. Summary of the invention

[0007] The following summary presents a simplified overview of certain features. This summary is not an exhaustive overview and is not intended to identify key or critical elements.

[0008] Aspects of the present invention provide an object recognition apparatus and method for improving the accuracy of a position of a recognized object by processing data obtained by a plurality of LIDARs.

[0009] Another aspect of the present disclosure provides an object recognition apparatus and method for improving the accuracy of a position of a host vehicle recognized in a specific space (eg, a covered passage such as a tunnel) by using data obtained and processed by a plurality of LIDARs.

[0010] Another aspect of the present disclosure provides an object recognition apparatus and method for reducing the amount of calculation by processing data obtained by a plurality of LIDARs only when an object to be recognized is a designated object.

[0011] The technical problems to be solved by the present disclosure are not limited to the above-mentioned problems, and any other technical problems not mentioned herein will be clearly understood by those skilled in the art to which the present disclosure belongs from the following description.

[0012] A device may include: a first sensor; a second sensor; a processor; and a memory storing instructions, which, when executed by the processor, cause the device to: identify multiple first candidate contour points of a stationary object based on sensing information of the first sensor, wherein the multiple first candidate contour points correspond to specific frames; determine multiple second candidate contour points of the stationary object based on sensing information of the second sensor, wherein the multiple second candidate contour points correspond to specific frames; when a vehicle is traveling in an area associated with a predefined type, determine the contour points of the stationary object based on a direction from the vehicle to the stationary object, wherein the determined contour points of the stationary object are also based on: an adjustment of the second candidate contour points and the multiple first candidate contour points; or an adjustment of the first candidate contour points and the multiple second candidate contour points; and generate a signal representing the determined contour points of the stationary object to display the location of the stationary object on a map.

[0013] The area associated with the predefined type may include a tunnel, and wherein the instructions, when executed by the processor, cause the apparatus to determine that the stationary object is not a dynamic object based on the contour points not being representative of a side view mirror and not being representative of an exhaust.

[0014] When the instructions are executed by a processor, the device can: identify a shape frame representing a stationary object, wherein the shape frame includes at least one of a plurality of first candidate contour points and a plurality of second candidate contour points, and determine the contour points of the stationary object based on the length of the line segments of the shape frame in the length direction being greater than or equal to a first threshold length, and the length of the line segments of the shape frame in the horizontal direction being greater than or equal to a second threshold length.

[0015] The instructions, when executed by a processor, may cause the device to identify a shape box in a specific frame, wherein the shape box includes at least one of: a plurality of first candidate contour points and a plurality of second candidate contour points, and the contour points are determined based on that a length of a line segment of the shape box in a longitudinal direction is longer than a length of the line segment of the shape box in a transverse direction, wherein the longitudinal direction is parallel to a direction of travel of the vehicle.

[0016] The instructions, when executed by a processor, may cause the device to determine contour points based on: a plurality of second candidate contour points; a distance between the second sensor and the stationary object being closer than a distance between the first sensor and the stationary object, and adjustment of the first candidate contour points comprising excluding at least one of the first candidate contour points.

[0017] The instructions, when executed by a processor, may cause the device to determine contour points based on: a plurality of first candidate contour points; a stationary object located in a lateral direction of the vehicle, a first sensor located in a lateral direction of the vehicle, and wherein the adjustment of the second candidate contour points includes excluding at least one of the second candidate contour points.

[0018] The instructions, when executed by a processor, may cause the device to, based on the distance between the second sensor and the stationary object being closer than the distance between the first sensor and the stationary object: move at least one of the plurality of first candidate contour points toward a traveling direction of the vehicle, and determine at least one contour point, wherein an amount of movement of at least one of the plurality of first candidate contour points is determined according to a distance identified based on at least one of the following: coordinates of a first point of the plurality of first candidate contour points, coordinates of a second point of the plurality of second candidate contour points corresponding to the first point, a difference between a time when a specific frame is acquired by the first sensor and a time when a specific frame is acquired by the second sensor, and a speed of the vehicle.

[0019] An adjusted distance between a first point obtained via a first sensor and an adjusted point corresponding to the first point obtained via a second sensor may be smaller than an unadjusted distance between the first point and an unadjusted point corresponding to the first point obtained via the second sensor, wherein the first point and the unadjusted point may be obtained from the same part of the stationary object.

[0020] When the instructions are executed by the processor, the device: based on identifying that the stationary object is in a side direction of the vehicle, moves a point corresponding to another side direction of the vehicle different from the one side direction and obtained by one of the first sensor and the second sensor toward the driving direction of the vehicle; and determines the contour points of the stationary object based on one of the first sensor and the second sensor corresponding to the one side direction.

[0021] The stationary object may include a vehicle having an emergency escape route within a tunnel, and the first sensor and the second sensor are mounted on the vehicle and arranged to be spaced apart from each other.

[0022] A device may include: a first sensor; a second sensor; a processor; and a memory storing instructions that, when executed by the processor, cause the device to: identify a plurality of first candidate contour points of a stationary object based on sensing information of the first sensor, wherein the plurality of first candidate contour points correspond to a specific frame; identify a plurality of second candidate contour points of the stationary object based on sensing information of the second sensor, wherein the plurality of second candidate contour points correspond to a specific frame; and when a vehicle travels in an area associated with a predefined type, adjust the plurality of first candidate contour points or the plurality of second candidate contour points to reduce a spacing between the plurality of first candidate contour points and the plurality of second candidate contour points, wherein one of the plurality of first candidate contour points and one of the plurality of second candidate contour points represent a same portion of the stationary object, and the spacing is determined based on a distance between one of the plurality of first candidate contour points and one of the plurality of second candidate contour points.

[0023] The instructions, when executed by a processor, may cause the device to move a plurality of first candidate contour points along a travel direction of the vehicle based on the distance between the second sensor and the stationary object being closer than the distance between the first sensor and the stationary object, and wherein the movement amounts of the plurality of first candidate contour points are determined based on distances identified based on at least one of the following items: coordinates of a first point among the plurality of first candidate contour points, coordinates of a second point among the plurality of second candidate contour points corresponding to the first point, a difference between a time when the first sensor acquires a specific frame and a time when the second sensor acquires the specific frame, and a speed of the vehicle.

[0024] When the instructions are executed by the processor, the device adjusts multiple first candidate contour points or multiple second candidate contour points through the following operations: based on the identification of a stationary object in a first lateral direction of the vehicle, one or more candidate contour points identified based on points corresponding to a second lateral direction of the vehicle different from the first lateral direction of the vehicle are moved in the driving direction of the vehicle, wherein the movement amount of the one or more candidate contour points is determined based on a distance identified by at least one of the following: coordinates of the one or more candidate contour points, coordinates of the one or more candidate contour points identified based on points corresponding to the first lateral direction, a difference between a time when a first sensor acquires a specific frame and a time when a second sensor acquires a specific frame, and a speed of the vehicle.

[0025] These and other features and advantages are described in greater detail below. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] The above and other objects, features and advantages of the present disclosure will become more apparent from the following detailed description in conjunction with the accompanying drawings:

[0027] Figure 1 is a block diagram showing a configuration of an object recognition device;

[0028] Figure 2 is a diagram showing an example of points obtained in a specific frame by LIDAR in an object recognition device or an object recognition method;

[0029] Figure 3 A flow chart showing the operation of an object recognition device for recognizing contour points of an object in an object recognition device or an object recognition method;

[0030] Figure 4 An example of a condition for recognizing that an object is a stationary object in an object recognition device or an object recognition method is shown;

[0031] Figure 5 An example of a condition for recognizing whether an object is a stationary object based on the size of a shape frame in an object recognition device or an object recognition method is shown;

[0032] Figure 6 A flowchart showing the operation of an object recognition device for recognizing contour points of an object by using points obtained by a LIDAR in the object recognition device or the object recognition method;

[0033] Figure 7 An example of points obtained by a plurality of LIDARs in an object recognition device or an object recognition method is shown;

[0034] Figure 8 An example of contour points based on multiple LIDAR recognitions and contour points based on one LIDAR recognition in an object recognition device or object recognition method is shown;

[0035] Fig. 9 A flowchart showing the operation of an object recognition device for recognizing contour points of an object by changing points obtained by a LIDAR in an object recognition device or an object recognition method;

[0036] Fig.10 An example of contour points recognized based on multiple LIDARs and points obtained by one LIDAR in an object recognition device or an object recognition method is shown;

[0037] Fig.11 A flow chart showing the operation of an object recognition device for changing contour points of an object in an object recognition device or an object recognition method; and

[0038] Fig.12 A computing system for an object recognition apparatus and an object recognition method is shown. DETAILED DESCRIPTION

[0039] Hereinafter, various embodiments of the present disclosure will be described in detail with reference to the exemplary drawings. In addition, when reference numerals are added to the constituent parts of each drawing, the same reference numerals are also marked when the same or equivalent constituent parts are shown in other drawings. In addition, when describing the embodiments of the present disclosure, detailed descriptions of well-known features or functions will be excluded so as not to unnecessarily obscure the main purpose of the present disclosure.

[0040] When describing components according to the present disclosure, terms such as first, second, "A", "B", (a), (b), etc. may be used. These terms are intended only to distinguish one component from another, and these terms do not limit the nature, order, or sequence of the constituent components. Unless otherwise defined, all terms (including technical or scientific terms) used herein have the same meaning as those commonly understood by those skilled in the art to which the present disclosure belongs. Those terms as defined in commonly used dictionaries should be interpreted as having the same meaning as the contextual meaning in the relevant art, and should not be interpreted as having an ideal or overly formal meaning unless explicitly defined in this application as having an ideal or overly formal meaning.

[0041] In the following, reference will be made to Figures 1 to 10 Various examples of the present disclosure are described in detail.

[0042] Figure 1 is a block diagram showing the configuration of an object recognition device.

[0043] The object recognition device 101 may be implemented in a vehicle. In this case, the object recognition device 101 may be formed integrally with an internal control unit of the vehicle, or may be implemented as a separate device and connected to the control unit of the vehicle via a connector.

[0044] Reference Figure 1 , the object recognition device 101 may include a first sensor (eg, a first LIDAR 103), a second sensor (eg, a second LIDAR 105), and a processor 107. The object recognition device 101 and / or the processor may also include one or more storage devices (eg, memories).

[0045] The first LIDAR 103 and / or the second LIDAR 105 may emit laser light (e.g., pulsed light waves) around the main vehicle (which may be a vehicle including the object recognition device 101) to search the environment around the main vehicle and recognize one or more objects in the surrounding environment. The first LIDAR 103 and / or the second LIDAR 105 may recognize the distance from the LIDAR that has emitted the laser to the target object based on the time interval from the light emission time (e.g., the time when the laser is emitted) to the light reception time (e.g., the time when the emitted laser is received after being reflected from the target object) and the speed of the laser. The first LIDAR 103 and / or the second LIDAR 105 may obtain a point indicating the position of an object (e.g., at least a portion of one or more surfaces of the object) around the main vehicle based on the distance to the target object and the emission angle of the emitted laser.

[0046] The processor 107 of the object recognition device 101 can recognize the contour points of the object to recognize the shape of the object. The processor 107 of the object recognition device 101 can recognize the contour points of the object based on the points obtained by the first LIDAR 103 and / or the second LIDAR 105. The contour points may include points representing the external shape of the object. The processor 107 of the object recognition device may recognize some points that can represent the shape of the object among the points representing the object as the contour points of the object to reduce the amount of calculation.

[0047] The first LIDAR 103 and the second LIDAR 105 may be spaced apart from each other. When the first LIDAR 103 and the second LIDAR 105 are spaced apart from each other, a time point at which the laser emitted from the first LIDAR 103 is reflected and a time point at which the laser emitted from the second LIDAR 105 is reflected may be different at a specific point of an object around the vehicle.

[0048] When the vehicle including the first LIDAR 103 and the second LIDAR 105 is in operation, the position of the object recognized by the first LIDAR 103 may be different from the position of the object recognized by the second LIDAR 105. This difference may be caused by the fact that when the vehicle is in operation, the relative position of the host vehicle and the object when the first LIDAR 103 recognizes a specific point of the object may be different from the relative position of the host vehicle and the object when the second LIDAR 105 recognizes the specific point of the object. Therefore, the processor 107 of the object recognition device 101 may obtain the position of the object recognized by the second LIDAR 105, which is different from the position of the object recognized by the first LIDAR 103.

[0049] The processor 107 of the object recognition device 101 may use or change the points obtained by any one of the first LIDAR 103 and the second LIDAR 105 to accurately specify the position of the object. For example, the processor 107 of the object recognition device 101 may obtain the contour points of the object based on the points obtained by any one of the first LIDAR 103 and the second LIDAR 105. For example, the processor 107 of the object recognition device 101 may obtain the contour points of the object based on the points obtained and changed by any one of the first LIDAR 103 and the second LIDAR 105.

[0050] The processor 107 of the object recognition device 101 may obtain contour points of the object based on the obtained points, and change at least one of the obtained contour points without changing the obtained points (e.g., points obtained by any one of the first LIDAR 103 and the second LIDAR 105). The processor 107 may obtain a new set of contour points of the object based on the changed contour points. For example, the processor 107 of the object recognition device 101 may move one of the first contour points obtained by the first LIDAR 103 and identified based on the points included in the specific frame and the second contour points obtained by the second LIDAR 105 and identified based on the points included in the specific frame.

[0051] Figure 2 is a diagram showing an example of points obtained by LIDAR in a specific frame in an object recognition device or an object recognition method.

[0052] Reference Figure 2 , the picture 201 may represent the surrounding environment recognized by the first LIDAR, the surrounding environment recognized by the second LIDAR, and the surrounding environment recognized by the precise map included in a specific frame. The first line 203 may represent the surrounding environment recognized by the precise map. The second line 205 may represent the line connecting the points obtained by the first LIDAR, and the first LIDAR may be a LIDAR on one side of the main vehicle. The third line 207 may represent the line connecting the points obtained by the second LIDAR, and the second LIDAR may be a LIDAR on the front side of the vehicle or on another side different from one side of the main vehicle. The top view 211 may represent a one side direction 213 and another side direction 215 distinguished based on the main vehicle, and the main vehicle may be a vehicle including an object recognition device.

[0053] On the screen 201, a first line 203 may represent a surrounding environment through a precise map. A second line 205 may represent a point obtained by a first LIDAR and included in a specific frame. A third line 207 may represent a point obtained by a second LIDAR different from the first LIDAR and included in a specific frame.

[0054] The host vehicle may be operated in a specific location (eg, a covered passage such as a tunnel and an underground passage). The point obtained by LIDAR may represent a preset object in the specific location (eg, an emergency evacuation route in a tunnel).

[0055] The host vehicle can improve the reliability of the information about the surrounding environment obtained through the LIDAR and identify the precise position of the host vehicle by matching the information about the surrounding environment obtained through the precise map with the surrounding environment obtained through the LIDAR.

[0056] It may be difficult for the processor of the object recognition device to find a point (e.g., a sign, a road intersection) formed laterally relative to a road boundary in a specific space (e.g., a tunnel, an underpass). Therefore, it may be difficult to set a reference point to match the surrounding environment obtained through the precise map with the surrounding environment obtained through the LIDAR. Therefore, it may be difficult to identify the precise position of the host vehicle or to improve the reliability of information about the surrounding environment.

[0057] Therefore, the need to accurately identify objects formed laterally in a confined space (e.g., a covered passage such as a passage, an underpass) (e.g., an emergency evacuation route in a passage) may be higher than the need to accurately identify objects formed laterally in an unconfined space (e.g., an uncovered passage).

[0058] The object formed laterally may be obtained by the first LIDAR and the second LIDAR. Depending on the difference between the position where the first LIDAR is placed and the position where the second LIDAR is placed and the driving speed of the main vehicle, the second line 205 representing the point obtained by the first LIDAR and the third line 207 representing the point obtained by the second LIDAR may not match. In order to accurately identify the position of the object, the processor of the object recognition device may identify the contour points of the object by using and / or changing the points obtained by any one of the first LIDAR and the second LIDAR. In other words, the object recognition device may identify the contour points of the object by using or changing the points represented by one of the second line 205 and the third line 207. The processor of the object recognition device may improve the accuracy of the position of the object around the main vehicle and the accuracy of the position of the main vehicle by comparing the contour line connecting the identified contour points of the object with the first line 203 of the accuracy map.

[0059] The processor of the object recognition device may recognize the contour points of the object by using the points obtained by one of the first LIDAR and the second LIDAR. For example, the processor of the object recognition device may recognize the contour points of the object by using the points represented by the second line 205. For example, the processor of the object recognition device may recognize the contour points of the object by using the points represented by the third line 207.

[0060] The processor of the object recognition device may recognize the contour points of the object by using the points obtained by the LIDAR that satisfies the specified conditions. For example, the specified conditions may include a condition that the LIDAR that obtains the points used is a LIDAR set in the same direction as the direction from the main vehicle to the object (for example, the left direction of the main vehicle or the right direction of the main vehicle). For example, the specified conditions may include a condition that the LIDAR that obtains the points used is a LIDAR with a shorter distance to the object than other LIDARs. For example, the specified conditions may include a condition that the LIDAR that obtains the points used is a LIDAR that obtains a frame containing points earlier than other LIDARs.

[0061] The processor of the object recognition device can recognize the contour points of the object by changing the points obtained by any one of the first LIDAR and the second LIDAR. For example, the processor of the object recognition device can recognize the contour points of the object by moving the points represented by the third line 207 using the points represented by the second line 205 or the points represented by the moved third line 207. In other words, the processor of the object recognition device can recognize some of the points represented by the second line 205 as the contour points of the object, and recognize some of the points represented by the third line 207 as the contour points of the object.

[0062] The processor of the object recognition device may recognize the contour point of the object by moving the point obtained by the LIDAR that satisfies the specified condition. For example, the specified condition may include a condition that the LIDAR that obtains the moving point is a LIDAR arranged in a direction (e.g., the right direction of the main vehicle or the left direction of the main vehicle) different from the direction from the main vehicle to the object (e.g., the left direction of the main vehicle or the right direction of the main vehicle). For example, the specified condition may include a condition that the LIDAR that obtains the moving point is a LIDAR that is farther away from the object than other LIDARs. For example, the specified condition may include the following condition: compared with other LIDARs, the LIDAR that obtains the moving point is a LIDAR with a later frame identification time point for obtaining the point.

[0063] As reference Figure 1 As described, the processor of the object recognition device may obtain contour points of the object based on the obtained points and change the obtained contour points without changing the obtained points (e.g., points obtained by any one of the first LIDAR 103 and the second LIDAR 105) and obtain a new set of contour points of the object based on the changed contour points. For example, the processor of the object recognition device may move a set of contour points of a first contour point obtained by the first LIDAR and identified based on points included in a specific frame and a second contour point obtained by the second LIDAR and identified based on points included in the specific frame.

[0064] The processor of the object recognition device may move a point obtained by one LIDAR to be closer to a point obtained by a LIDAR different from the one LIDAR.

[0065] Therefore, the spacing between a point obtained by one LIDAR after movement (e.g., one of the moving points represented by the third line 207) and a point corresponding to a point obtained by one LIDAR and obtained by a LIDAR different from the one LIDAR (e.g., represented by the second line 205 and corresponding to the point represented by the third line 207) can be smaller than the spacing between a point obtained by one LIDAR before movement and a point corresponding to a point obtained by one LIDAR and obtained by a LIDAR different from the one LIDAR.

[0066] Points for identifying a gap obtained by one LIDAR and points corresponding to the points obtained by the one LIDAR and obtained by a LIDAR different from the one LIDAR can be obtained from the same portion.

[0067] The top view 211 may represent an example of a direction from the host vehicle to an object. The object may be in one lateral direction 213 or another lateral direction 215. One lateral direction 213 may include the left side of the host vehicle. The other lateral direction 215 may include the right side of the host vehicle. Figure 6 and Fig. 9 Describes the conditions used to specify the LIDAR based on the direction an object is located.

[0068] exist Figure 2 In the embodiment, one lateral direction is described as including only the left direction, and the other lateral direction is described as including only the right direction, but the terms used in the present disclosure may not be limited thereto. In an example, one lateral direction may include the right direction of the host vehicle, and the other lateral direction may include the left direction of the host vehicle.

[0069] Figure 3 A flow chart showing the operation of an object recognition device for recognizing contour points of an object in an object recognition device or an object recognition method.

[0070] In the following, it is assumed Figure 1 The object recognition device 101 performs Figure 3 However, aspects of the present disclosure are not limited thereto. Figure 3 In the description of , operations described as being performed by a device or a processor may be understood to be controlled by the processor 107 of the object recognition device 101 .

[0071] Reference Figure 3In the first operation 301, the processor of the object recognition device may obtain a signal indicating a specific space (e.g., a limited space, a covered passage, etc.) where the host vehicle is located. Only when the host vehicle is in the limited space, the processor of the object recognition device may use or change a point obtained by one of the LIDARs spaced apart from each other to reduce the amount of calculation. When the vehicle including the object recognition device enters a specific space (e.g., a tunnel, an underpass), the processor of the object recognition device may obtain a signal indicating that the vehicle is in the specific space from the map system.

[0072] In the second operation 303, the processor of the object recognition device may recognize contour points of the object by using or changing points obtained by any one of the first LIDAR and the second LIDAR. The second operation 303 may be performed to increase the accuracy of the position of the object.

[0073] The processor of the object recognition device may recognize the contour points of the object based on the points obtained by one of the first LIDAR and the second LIDAR. For example, the processor of the object recognition device may remove or exclude the points obtained by one LIDAR and recognize the contour points of the object using only the points obtained by the other LIDAR.

[0074] The processor of the object recognition device may change the points obtained by any one of the first LIDAR and the second LIDAR to recognize the contour points of the object based on the changed points. For example, the processor of the object recognition device may change the points obtained by one LIDAR and recognize a new set of contour points of the object based on the changed points.

[0075] Although in Figure 3 , and obtaining contour points of an object based on the changed points. However, aspects of the present disclosure may not be limited thereto. Fig.11 Description: Contour points of an object may be obtained based on the obtained points, and the obtained contour points may then be changed.

[0076] Figure 4 An example of a condition for recognizing that an object is a stationary object in an object recognition device or an object recognition method is shown.

[0077] If the target object is an emergency evacuation route in a tunnel, the need for the present disclosure to improve the accuracy of the contour points of the object may be increased. If the object is an emergency evacuation route in a tunnel, one or more operations described herein may be applied. It should be noted that aspects of the present invention may not be limited to this. The accuracy of the contour points of the object may be increased and the amount of calculation may be reduced. In the following, conditions for determining whether an object is an emergency evacuation route in a tunnel are described.

[0078] Reference Figure 4 , the processor of the object recognition device may recognize a tunnel identifier 403 indicating whether the host vehicle (which may be a vehicle including the object recognition device) is in the tunnel on the screen 401. The processor of the object recognition device may recognize at least one point 413 representing a side-view mirror among points included in the first frame 411 and representing an object. The processor of the object recognition device may recognize at least one point 423 representing exhaust (e.g., vehicle exhaust) among points included in the second frame 421 and representing an object.

[0079] The processor of the object recognition device may recognize whether the host vehicle is in a tunnel. If the host vehicle is not located in a tunnel, the probability that the object recognized around the host vehicle is an emergency evacuation route in the tunnel is low. The processor of the object recognition device may obtain a signal indicating that the host vehicle is in a tunnel from a map system. If a signal indicating that the host vehicle is in a tunnel is received, the processor of the object recognition device may store an identifier indicating that the signal has been received. For example, the tunnel identifier 403 may be stored as a first value (e.g., "1").

[0080] The processor of the object identification device may obtain a signal from the map system indicating that the host vehicle is outside the tunnel. If a signal indicating that the host vehicle is outside the tunnel is received, the processor of the object identification device may store an identifier indicating that the signal has been received. For example, tunnel identifier 403 may be stored as a second value (e.g., "0").

[0081] The processor of the object recognition device may recognize whether the object is not a dynamic object. If the object is a dynamic object, the probability that the object recognized around the host vehicle is an emergency evacuation route in the tunnel is low. If at least one point 413 corresponding to the side mirror and at least one point 423 corresponding to exhaust gas among the points included in a specific frame (e.g., frame 411 or frame 421) are recognized, the processor of the object recognition device may recognize that the object represented by the points included in the specific frame is a dynamic object (e.g., a vehicle).

[0082] The processor of the object recognition device can recognize an identifier representing a specific space, and recognize a contour point of the object by using or changing the points obtained by the first LIDAR or the second LIDAR based on the fact that at least one point representing a side-view mirror and an exhaust among the points representing the object and included in the specific frame is not recognized.

[0083] The processor of the object recognition device may recognize an identifier representing a specific space, and recognize contour points of the object by changing the recognized contour points based on points obtained by the first LIDAR or the second LIDAR based on the fact that at least one point representing the side mirror and the exhaust is not recognized.

[0084] Figure 5An example of a condition for identifying whether an object is a stationary object based on the size of a shape frame in an object recognition device or an object recognition method is shown. Figure 4 As shown, conditions for determining whether an object is an emergency evacuation route in a tunnel are described.

[0085] Reference Figure 5 , the first shape box 501 may include the points 503 obtained from the LIDAR and included in the specific frame. The second shape box 511 may include the points obtained from the LIDAR and included in the specific frame.

[0086] The size of the first shape frame 501 may be expressed as a length in a longitudinal direction (eg, a direction parallel to a road boundary direction) and a length in a lateral direction (eg, a direction perpendicular to the road boundary direction).

[0087] The processor of the object recognition device can recognize the contour points of the object by using or changing the points obtained by either the first LIDAR or the second LIDAR or by changing the contour points based on the fact that the length of the first shape frame 501 in the longitudinal direction is greater than or equal to a specified length (e.g., about 3.5 m) and the length in the transverse direction is greater than or equal to a specified length (e.g., about 4.5 m).

[0088] If the length of the first shape frame 501 in the longitudinal direction is less than the specified length or the length in the transverse direction is less than the specified length, the probability that the object corresponding to the first shape frame 501 is an emergency evacuation route in the tunnel is low. If the length of the first shape frame 501 in the longitudinal direction is less than the specified length or the length in the transverse direction is less than the specified length, the probability that the object represented by the point is a dynamic object (e.g., a vehicle) is high.

[0089] The processor of the object recognition device can identify the contour points of the object by using or changing the points obtained by either the first LIDAR or the second LIDAR based on the fact that the length of the second shape frame 511 in the longitudinal direction (e.g., the direction parallel to the road boundary direction) is longer than the length of the second shape frame 511 in the transverse direction (e.g., the direction perpendicular to the road boundary direction).

[0090] If the length of the second shape frame 511 in the longitudinal direction is shorter than the length of the second shape frame 511 in the transverse direction, the probability that the object corresponding to the second shape frame 511 is an emergency evacuation route in the tunnel is low. If the length of the second shape frame 511 in the longitudinal direction is shorter than the length of the second shape frame 511 in the transverse direction, the probability that the object represented by the point is a dynamic object (e.g., a vehicle) is high.

[0091] Figure 6 A flowchart showing the operation of an object recognition device for recognizing contour points of an object by using points obtained by a LIDAR in the object recognition device or the object recognition method.

[0092] In the following, it is assumed Figure 1 The object recognition device 101 performs Figure 6 However, aspects of the present disclosure are not limited thereto. Figure 6 In the description of , operations described as being performed by a device or a processor may be understood to be controlled by the processor 107 of the object recognition device 101 .

[0093] Reference Figure 6 In a first operation 601, the processor of the object recognition device may obtain a signal representing a specific space (e.g., a tunnel or an underground passage) from a map system. In order to reduce the amount of calculation, the processor of the object recognition device may use or change the points obtained by one of the LIDARs spaced apart from each other (e.g., only when the main vehicle is in a specific space such as a covered passage). Specifically, if the object is an emergency evacuation route in a tunnel, the need to improve the accuracy of the contour points of the object may increase. Therefore, the processor of the object recognition device may apply one or more features of the present disclosure (e.g., if the object is an emergency evacuation route or any other stationary object). As has been mentioned above, Figure 4 The method of determining whether a target is an emergency evacuation route has been described, and therefore, repeated description is omitted below.

[0094] In the second operation 603, the processor of the object recognition device may recognize whether the object is recognized as being in one side direction. If the object is recognized as being in one side direction (e.g., in the left direction of the host vehicle or the right direction of the own vehicle), the processor of the object recognition device may perform a third operation 605. If the object is not recognized as being in one side direction, the processor of the object recognition device may perform a fourth operation 607.

[0095] The side direction in which the object is located may be identified by LIDAR. The processor of the object identification device may obtain an identifier indicating the left side or an identifier indicating the right side according to the direction identified by the LIDAR point together with the point representing the object.

[0096] In the third operation 605, the processor of the object recognition device may recognize the contour points of the object according to the points obtained by the LIDAR corresponding to one side direction. This may be at least because the points obtained by the LIDAR corresponding to the one side direction where the object is located are more accurate than the points obtained by the LIDAR corresponding to the other side direction.

[0097] The processor of the object recognition device may exclude or remove points obtained through the LIDAR corresponding to another side direction different from the one side direction, and recognize contour points of the object based on the points obtained through the LIDAR corresponding to the one side direction.

[0098] In the fourth operation 607, the processor of the object recognition device may recognize contour points of the object based on points obtained by the LIDAR corresponding to the other side direction and included in the specific frame. If the object is recognized as being in the other side direction, the processor of the object recognition device may exclude or remove points obtained by the LIDAR corresponding to the one side direction and recognize contour points of the object based on points obtained by the LIDAR corresponding to the direction in which the object is located.

[0099] Figure 7 An example of points obtained by a plurality of LIDARs in an object recognition device or an object recognition method is shown.

[0100] Reference Figure 7 , the points included in the first picture 701 and the points included in the second picture 711 may correspond to the same object in the left direction (e.g., an emergency evacuation route in a tunnel). The first line 703 of the first picture 701 may represent the points obtained by the LIDAR corresponding to the left direction. The second line 705 of the first picture 701 may represent the points obtained by the LIDAR corresponding to the right direction. The first picture 701 may include points obtained earlier than the points of the second picture 711. The first line 713 of the second picture 711 may indicate the points obtained by the LIDAR corresponding to the left direction. The second line 715 of the second picture 711 may represent the points obtained by the LIDAR corresponding to the right direction.

[0101] The distance between the object represented by the point and the host vehicle when acquiring the point included in the first screen 701 may be greater than the distance between the object represented by the point and the host vehicle including the object recognition device when acquiring the point included in the second screen 711. As the distance between the host vehicle and the object decreases, the number of points acquired by the LIDAR corresponding to the left direction where the object is located may increase. In other words, the accuracy of the points acquired by the LIDAR corresponding to the left direction where the object is located may be higher than the accuracy of the points acquired by the LIDAR corresponding to the right direction, which is different from the LIDAR corresponding to the left direction.

[0102] Therefore, in order to improve the accuracy of the position of the object, the processor of the object recognition device may recognize the contour points of the object by using points obtained via the LIDAR corresponding to the side direction where the object is located and included in a specific frame.

[0103] Figure 8Examples of contour points recognized based on a plurality of LIDARs and contour points recognized based on one LIDAR in an object recognition device or an object recognition method are shown.

[0104] Reference Figure 8 , in the first picture 801, the existing object recognition device may recognize the contour points of the object by using the points obtained by the first LIDAR and included in the specific frame, or the points obtained by the second LIDAR and included in the specific frame. The first line 803 of the first picture 801 may represent the points obtained by the LIDAR corresponding to one side direction and included in the specific frame. The second line 805 of the first picture 801 may represent the points obtained by the LIDAR corresponding to another side direction different from the one side direction and included in the specific frame. The third line 807 of the first picture 801 may represent the contour points of the object.

[0105] In the second screen 811, the processor of the object recognition device may recognize the contour points of the object by using the points obtained by any one of the first LIDAR and the second LIDAR and included in the specific frame. The first line 813 of the second screen 811 may represent the points obtained by the LIDAR corresponding to one side direction and included in the specific frame. The second line 815 of the second screen 811 may represent the points obtained by the LIDAR corresponding to the other side direction and included in the specific frame. The third line 817 of the second screen 811 may represent the contour points of the object.

[0106] In the first screen 801 and the second screen 811, points obtained by the LIDAR corresponding to one side direction and included in a specific frame may be separated from points obtained by the LIDAR corresponding to another side direction different from the one side direction and included in a specific frame. This may be because at least the time when the laser emitted from the LIDAR corresponding to one side direction hits the object may be different from the time when the laser emitted from the LIDAR corresponding to the other side direction hits the object. When the vehicle is running, when the laser emitted from the LIDAR corresponding to one side direction hits the target, the relative position of the host vehicle and the object may be different from the relative position of the host vehicle and the object when the laser emitted from the LIDAR corresponding to the other side direction hits the target.

[0107] If one or more contour points of an object are identified using one of a point obtained by LIDAR corresponding to one side direction and included in a specific frame and a point obtained by LIDAR corresponding to the other side direction and included in a specific frame, the contour points of the object may be identified as being spaced apart from each other similarly to the third line 807 of the first screen 801. The existing object recognition apparatus may identify some points represented by the first line 803 of the first screen 801 as contour points of the object, and identify some points represented by the second line 805 of the first screen 801 as contour points of the object.

[0108] If the contour points of the object are recognized using points obtained by one of the LIDAR corresponding to one side direction and the LIDAR corresponding to the other side direction, the contour points of the object may be recognized as not being spaced apart from each other like the third line 817 of the second screen 811. Thus, it may be determined that the accuracy of the position of the object recognized by the processor of the object recognition device has increased.

[0109] Fig. 9 A flowchart showing the operation of an object recognition device for recognizing contour points of an object by changing points obtained by a LIDAR in an object recognition device or an object recognition method.

[0110] Reference Fig. 9 In a first operation 901, a processor of an object recognition device may obtain a signal indicating a specific space (e.g., a covered passage, such as a tunnel, or a dark space without sufficient ambient light or natural light) from a map system. Figure 6 The first operation 901 is performed in the same manner as the operation 601, and thus, redundant description will be omitted hereinafter.

[0111] In the second operation 903, the processor of the object recognition device may recognize whether the object is recognized as being in the one-side direction. If the object is recognized as being in the one-side direction, the processor of the object recognition device may perform a third operation 905. If the object is not recognized as being in the one-side direction, the processor of the object recognition device may perform a fourth operation 907.

[0112] The side direction in which the object is located may be recognized by LIDAR. The processor of the object recognition device may obtain an identifier indicating the left side or an identifier indicating the right side together with the point according to the direction in which the point indicating the object is recognized by LIDAR.

[0113] In a third operation 905 , the processor of the object recognition device may move the point obtained by the LIDAR corresponding to the other side direction.

[0114] The processor of the object recognition device may move the point obtained by the LIDAR corresponding to the other side direction in the driving direction of the host vehicle, or move the point obtained by the LIDAR corresponding to the other side direction in the direction of the point obtained by the LIDAR corresponding to the one side direction different from the LIDAR corresponding to the other side direction. The point obtained by the LIDAR corresponding to the other side direction may be obtained from the same part of the object as the point obtained by the LIDAR corresponding to the one side direction different from the LIDAR corresponding to the other side direction. This may be because at least the direction from the point obtained by the LIDAR corresponding to the other side direction to the point obtained by the LIDAR corresponding to the one side direction where the object is located may be the driving direction of the host vehicle.

[0115] The processor of the object recognition device may move the point obtained by the LIDAR corresponding to the other side direction by a distance, wherein the distance is recognized based on at least one of the following items: the coordinates of the point obtained by the LIDAR corresponding to the other side direction, the coordinates of the points corresponding to the points obtained by the LIDAR corresponding to the one side direction and the points obtained by the LIDAR corresponding to the other side direction, the difference between the time when the LIDAR corresponding to the other side direction acquires a specific frame and the time when the LIDAR corresponding to the one side direction acquires the specific frame, the speed of the host vehicle, or any combination thereof.

[0116] For example, the processor of the object recognition device may move the point obtained by the LIDAR corresponding to the other side direction by a distance identified based on the product of the difference between the time when the LIDAR corresponding to the other side direction acquires a specific frame and the time when the LIDAR corresponding to the one side direction acquires a specific frame and the speed of the host vehicle.

[0117] In a fourth operation 907 , the processor of the object recognition device may move a point corresponding to one side direction.

[0118] The processor of the object recognition device may move the point obtained by the LIDAR corresponding to one side direction in the traveling direction of the host vehicle, or move the point obtained by the LIDAR corresponding to one side direction in the direction of the point obtained by the LIDAR different from the LIDAR corresponding to one side direction and the LIDAR corresponding to the other side direction. This may be because the direction from the point obtained by the LIDAR corresponding to the one side direction where the object is located to the point obtained by the LIDAR corresponding to the other side direction may be the traveling direction of the host vehicle.

[0119] The processor of the object recognition device may move a point obtained by the LIDAR corresponding to one side direction by a distance, wherein the distance is recognized based on at least one of the following items: coordinates of the point obtained by the LIDAR corresponding to one side direction, coordinates of a point corresponding to the point obtained by the LIDAR corresponding to one side direction and obtained by the LIDAR corresponding to the other side direction, a difference between a time when the LIDAR corresponding to one side direction acquires a specific frame and a time when the LIDAR corresponding to the other side direction acquires the specific frame, a speed of the host vehicle, or any combination thereof.

[0120] For example, the processor of the object recognition device may move the point obtained by the LIDAR corresponding to one side direction by a distance recognized based on the product of the difference between the time when the LIDAR corresponding to one side direction acquires a specific frame and the time when the LIDAR corresponding to the other side direction acquires a specific frame and the speed of the host vehicle.

[0121] In the fifth operation 909, the processor of the object recognition device may recognize contour points of the object. The processor of the object recognition device may recognize contour points of the object based on points that have moved (e.g., points obtained by LIDAR corresponding to the other side direction) or points that have not moved (e.g., points obtained by LIDAR corresponding to one side direction).

[0122] For example, the processor of the object recognition device may determine moved points and non-moved points as points representing an object, and recognize contour points of the object based on the moved points in some parts, and recognize contour points of the object based on the non-moved points in other parts.

[0123] In the sixth operation 911, the processor of the object recognition device may recognize contour points of the object. As in the fifth operation 909, the processor of the object recognition device may determine a moved point (e.g., a point obtained by a LIDAR corresponding to the other side direction) or a non-moved point (e.g., a point obtained by a LIDAR corresponding to one side direction) as a point representing an object, and recognize the contour points of the object based on the moved points in some parts, and recognize the contour points of the object based on the non-moved points in some other parts.

[0124] Fig.10 An example of contour points recognized based on a plurality of LIDARs and points obtained by one LIDAR in an object recognition device or an object recognition method is shown.

[0125] Reference Fig.10, the first picture 1001, the second picture 1011, and the third picture 1021 may represent the surrounding environment recognized by the first LIDAR and the second LIDAR. The first picture 1001 may include a first contour line 1003 based on the point recognition obtained by the first LIDAR and the second LIDAR. The second picture 1011 may include a second contour line 1013 based on the point recognition obtained by the first LIDAR. The third picture 1021 may include a third contour line 1023 based on the point recognition obtained by the second LIDAR. The first picture 1001, the second picture 1011, and the third picture 1021 may represent an emergency evacuation route and a walking path in the tunnel.

[0126] In the first screen 1001, the existing object recognition device may recognize contour points based on points obtained by the first LIDAR or points obtained by the second LIDAR. The first contour line 1003 may be a line segment connecting contour points. As shown in the second screen 1011 and the third screen 1021, the points obtained by the first LIDAR are spaced apart from the points corresponding to the points obtained by the first LIDAR and obtained by the second LIDAR, so that the first contour line 1003 may be recognized discontinuously. The accuracy of the position of the object according to the discontinuously recognized contour may be lower than the accuracy of the position of the object according to the continuously recognized contour.

[0127] In order to improve the accuracy of the position of the object on the first screen 1001, the processor of the object recognition device may move the point obtained by the second LIDAR in the direction of the point obtained by the first LIDAR.

[0128] Fig.11 A flow chart showing the operation of an object recognition device for changing contour points of an object in an object recognition device or an object recognition method.

[0129] Reference Fig.11 In a first operation 1101, a processor of the object recognition device may obtain a signal indicating that the host vehicle is located in a specific space (e.g., a covered passage such as a tunnel, or a dark space without sufficient ambient light or natural light). Figure 6 The first operation 1101 is performed in the same manner as operation 601, and thus, repeated description will be omitted below.

[0130] In the second operation 1103, the processor of the object recognition device may recognize first contour points of the object based on the points obtained by the first LIDAR, and recognize second contour points of the object based on the points obtained by the second LIDAR. The contour points may include points representing the outer shape of the object. In order to reduce the amount of calculation, the processor of the object recognition device may recognize some points representing the shape of the object among the points representing the object as contour points of the object.

[0131] In a third operation 1105 , the processor of the object recognition device may change a contour point of the object recognized based on a point obtained by any one of the first LIDAR and the second LIDAR.

[0132] like Figure 1 As shown in , although the first LIDAR and the second LIDAR recognize the same object when the host vehicle travels in a state in which the first LIDAR and the second LIDAR are spaced apart from each other, the processor of the object recognition device can obtain the position of the object recognized by the second LIDAR, which may be different from the position of the object recognized by the first LIDAR.

[0133] The processor of the object recognition device may change the recognized contour points based on points obtained by any one of the first LIDAR and the second LIDAR to accurately specify the position of the object.

[0134] Based on recognizing that the distance between the second LIDAR and the object is closer than the distance between the first LIDAR and the object, the processor of the object recognition device can move the first contour point in the travel direction of the host vehicle or in the direction of the second contour point.

[0135] The processor of the object recognition device can move the first contour point by a distance, and the distance is identified based on at least one of the following items: the coordinates of the first contour point, the coordinates of the second contour point, the difference between the time when the first LIDAR acquires the specific frame and the time when the second LIDAR acquires the specific frame, the speed of the host vehicle, and any combination thereof.

[0136] For example, the processor of the object recognition device may move the first contour point by a distance recognized based on the product of the difference between the time when the first LIDAR acquires a specific frame and the time when the second LIDAR acquires the specific frame and the speed of the host vehicle, based on the fact that the distance between the second LIDAR and the object is closer than the distance between the first LIDAR and the object.

[0137] The processor of the object recognition device can move the third contour point in the driving direction of the main vehicle or in the direction of a fourth contour point based on recognition of the object located in one side direction of the main vehicle, wherein the third contour point is recognized based on a point corresponding to another side direction different from the one side direction and obtained by any one of the first LIDAR and the second LIDAR, and the fourth contour point is recognized based on a point corresponding to the one side direction and obtained by the other LIDAR of the first LIDAR and the second LIDAR different from the above-mentioned one LIDAR.

[0138] The processor of the object recognition device may move the third contour point by a distance, the distance being identified based on at least one of the following items: a coordinate of the third contour point, a coordinate of the fourth contour point, a difference between a time when one LIDAR acquires a specific frame and a time when another LIDAR acquires the specific frame, a speed of the host vehicle, and any combination thereof.

[0139] For example, the processor of the object recognition device moves the contour points identified based on the points obtained by the LIDAR corresponding to the other side direction by a distance identified based on the product of the difference between the time when the LIDAR corresponding to the other side direction acquires a specific frame and the time when the LIDAR corresponding to one side direction acquires a specific frame and the speed of the host vehicle.

[0140] According to one aspect of the present disclosure, an object recognition device includes a first LIDAR, a second LIDAR, and a processor.

[0141] According to an example, the processor may identify contour points of an object by using or changing points obtained by any one of the first LIDAR and the second LIDAR among points obtained by the first LIDAR and included in a specific frame and points obtained by the second LIDAR and included in a specific frame based on a signal obtained from a map system indicating a limited space in which the host vehicle is located. The points obtained by the first LIDAR and included in the specific frame may be obtained from the same object as the points obtained by the second LIDAR and included in the specific frame.

[0142] According to an example, the limited space may include a tunnel. Based on not identifying at least one point representing a side mirror and at least one point representing exhaust gas among the points included in the specific frame and representing the object, the processor may identify contour points of the object by using or changing points obtained by any one of the first LIDAR and the second LIDAR.

[0143] According to an example, the processor may identify a shape box that is included in a specific frame and includes points representing an object, and identify contour points of the object by using or changing points obtained by either the first LIDAR or the second LIDAR based on identifying that a length of a line segment of the shape box in a longitudinal direction is greater than or equal to a specified length and a length of a line segment of the shape box in a transverse direction is greater than or equal to a specified length.

[0144] According to an example, the processor may identify a shape box including points included in a specific frame representing an object, and identify contour points of the object by using or changing points obtained by either the first LIDAR or the second LIDAR based on the length of a line segment of the identified shape box in a longitudinal direction being greater than a length of a line segment of the shape box in a transverse direction.

[0145] According to an example, the processor may identify contour points of the object using points obtained by the second LIDAR close to the object based on recognizing that the distance between the second LIDAR and the object is closer than the distance between the first LIDAR and the object.

[0146] According to an example, the processor may identify contour points of the object based on identifying the object in a one side direction of the host vehicle by using points corresponding to the one side direction and obtained through any one of the first LIDAR and the second LIDAR.

[0147] According to an example, the processor may move the point obtained by the first LIDAR in the direction of travel of the host vehicle or in a direction related to the point corresponding to the point obtained by the first LIDAR and the point obtained by the second LIDAR, and identify the contour point of the object based on the point obtained by the first LIDAR or the point obtained by the second LIDAR based on recognizing that the distance between the second LIDAR and the object is closer than the distance between the first LIDAR and the object. The movement may be performed by a distance, and the distance is identified based on at least one of the following items: the coordinates of the point obtained by the first LIDAR, the coordinates of the point corresponding to the point obtained by the first LIDAR and the point obtained by the second LIDAR, the difference between the time when the first LIDAR obtains a specific frame and the time when the second LIDAR obtains a specific frame, the speed of the host vehicle, and any combination thereof. The point obtained by the first LIDAR and the point corresponding to the point obtained by the first LIDAR and the point obtained by the second LIDAR may be obtained from the same part of the object.

[0148] According to an example, a spacing between a point obtained by one LIDAR and a point corresponding to a point obtained by one LIDAR and obtained by a LIDAR different from one LIDAR after the change may be smaller than a spacing between a point obtained by one LIDAR and a point corresponding to a point obtained by one LIDAR and obtained by a LIDAR different from one LIDAR before the change. A point obtained by one LIDAR and a point corresponding to a point obtained by one LIDAR and obtained by a LIDAR different from one LIDAR may be obtained from the same part of an object. A spacing may be identified by a distance between a point obtained by one LIDAR and a point corresponding to a point obtained by one LIDAR and obtained by a LIDAR different from one LIDAR.

[0149] According to an example, the processor may identify the contour point of the object based on the recognition that the object moves in a side direction of the main vehicle, in the driving direction of the main vehicle, or in a direction related to a point corresponding to one side direction and obtained by one of the first LIDAR and the second LIDAR (different from one LIDAR) and corresponding to another side direction different from the one side direction and obtained by one of the first LIDAR and the second LIDAR, and based on the point obtained by one LIDAR or the point obtained by a different LIDAR corresponding to the one side direction. The movement may be performed by a distance, and the distance is identified based on at least one of the following items: the coordinates of the point obtained by one LIDAR, the coordinates of the point corresponding to the point obtained by one LIDAR and obtained by a different LIDAR, the difference between the time when one LIDAR obtains a specific frame and the time when a different LIDAR obtains a specific frame, the speed of the main vehicle, and any combination thereof. The point obtained by one LIDAR and the point corresponding to the point obtained by one LIDAR and obtained by a LIDAR different from one LIDAR may be obtained from the same part of the object.

[0150] According to an example, the object may include an emergency escape route in a tunnel. The side of the host vehicle may include one of a left side of the host vehicle and a right side of the host vehicle. The first LIDAR and the second LIDAR may be disposed to be spaced apart from the host vehicle.

[0151] According to another aspect of the present disclosure, an object recognition device may include a first LIDAR, a second LIDAR, and a processor.

[0152] According to an example, the processor may change a contour point of a group of the first contour point and the second contour point based on a signal obtained from a map system indicating a limited space in which the vehicle is located, so as to reduce a distance between a first contour point of an object identified based on a point obtained by a first LIDAR and included in a specific frame and a second contour point of an object identified based on a point obtained by a second LIDAR and included in a specific frame. One contour point corresponding to one of the first contour points and one of the second contour points may represent the same part of the object, and the distance may be identified by a distance between the one contour point and a contour point corresponding to the one contour point of the second contour points.

[0153] According to an example, the processor may move the first contour point in a traveling direction of the host vehicle or in a direction related to the second contour point based on recognizing that the distance between the second LIDAR and the object is closer than the distance between the first LIDAR and the object. The movement may be performed by a distance, the distance being identified based on at least one of the following items: a coordinate of a point obtained by the first LIDAR, a coordinate of a point corresponding to the point obtained by the first LIDAR and obtained by the second LIDAR, a difference between a time when the first LIDAR obtains a specific frame and a time when the second LIDAR obtains the specific frame, a speed of the host vehicle, and any combination thereof.

[0154] According to an example, the processor may move a third contour point in a traveling direction of the host vehicle or in a direction related to a fourth contour point, the third contour point being identified based on a point corresponding to another side direction different from the one side direction and obtained by one of the first LIDAR and the second LIDAR, the fourth contour point being identified based on a point corresponding to the one side direction and obtained by a LIDAR different from the one of the first LIDAR and the second LIDAR, or move the third contour point by a distance identified based on at least one of the following items: a coordinate of the third contour point, a coordinate of the fourth contour point, a difference between a time when one LIDAR acquires the specific frame and a time when a different LIDAR acquires the specific frame, a speed of the host vehicle, and any combination thereof.

[0155] According to another aspect of the present disclosure, an object recognition method may include: based on obtaining a signal indicating a limited space in which a vehicle is located from a map system, a processor may recognize contour points of an object by using or changing points obtained by any one of the first LIDAR and the second LIDAR among points obtained by the first LIDAR and included in a specific frame and points obtained by the second LIDAR and included in the specific frame. The points obtained by the first LIDAR and included in the specific frame may be obtained from the same object as the points included in the specific frame and obtained by the second LIDAR.

[0156] According to an example, the limited space may include a tunnel. Identifying the contour points of the object by using or changing the points obtained by any one of the first LIDAR and the second LIDAR among the points obtained by the first LIDAR and included in the specific frame and the points obtained by the second LIDAR and included in the specific frame may include: the processor identifying the contour points of the object by using or changing the points obtained by any one of the first LIDAR and the second LIDAR based on at least one point representing a side mirror and at least one point representing exhaust gas not being identified among the points included in the specific frame and identifying the object.

[0157] According to an example, identifying contour points of an object by using or changing points obtained by any one of the first LIDAR and the second LIDAR among points obtained by the first LIDAR and included in a specific frame and points obtained by the second LIDAR and included in the specific frame may include identifying, by a processor, a shape box that is included in the specific frame and includes points representing the object, and identifying, by the processor, the contour points of the object by using or changing points obtained by any one of the first LIDAR and the second LIDAR based on recognizing that lengths of line segments of the shape box in a longitudinal direction are greater than or equal to a specified length and lengths of line segments of the shape box in a transverse direction are greater than or equal to a specified length.

[0158] According to an example, identifying contour points of an object by using or changing points obtained by any one of the first LIDAR and the second LIDAR among points obtained by the first LIDAR and included in a specific frame and points obtained by the second LIDAR and included in the specific frame may include identifying, by a processor, a shape box that is included in the specific frame and includes points representing the object, and identifying, by the processor, the contour points of the object by using or changing the points obtained by any one of the first LIDAR and the second LIDAR based on recognizing that lengths of line segments of the shape box in a longitudinal direction are longer than lengths of line segments of the shape box in a transverse direction.

[0159] According to an example, identifying contour points of an object by using or changing points obtained by any one of the first LIDAR and the second LIDAR among points obtained by the first LIDAR and included in a specific frame and points obtained by the second LIDAR and included in a specific frame may include: identifying, by the processor, contour points of the object by using points obtained by the second LIDAR that are closer to the object based on recognizing that a distance between the second LIDAR and the object is closer than a distance between the first LIDAR and the object.

[0160] According to an example, identifying contour points of an object by using or changing points obtained by the first LIDAR and included in a specific frame and points obtained by the second LIDAR and included in a specific frame, including: identifying contour points of the target by using points corresponding to the one side direction and obtained by any one of the first LIDAR and the second LIDAR based on identifying that the object is in a side direction of the main vehicle.

[0161] According to an example, identifying the contour point of an object by using or changing the point obtained by any one of the first LIDAR and the second LIDAR among the points obtained by the first LIDAR and included in a specific frame and the points obtained by the second LIDAR and included in a specific frame may include: moving the point obtained by the first LIDAR by the processor in the driving direction of the main vehicle or in a direction related to the point obtained by the second LIDAR corresponding to the point obtained by the first LIDAR, and identifying the contour point of the object based on the point obtained by the first LIDAR or the point obtained by the second LIDAR based on recognizing that the distance between the second LIDAR and the object is closer than the distance between the first LIDAR and the object. The movement may be performed by a distance, and the distance may be identified based on at least one of the following items: the coordinates of the point obtained by the first LIDAR, the coordinates of the point obtained by the second LIDAR corresponding to the point obtained by the first LIDAR, the difference between the time when the first LIDAR obtains the specific frame and the time when the second LIDAR obtains the specific frame, the speed of the main vehicle, and any combination thereof. The point obtained by the first LIDAR and the point obtained by the second LIDAR corresponding to the point obtained by the first LIDAR may be obtained from the same part of the object.

[0162] Fig.12 A computing system for an object recognition apparatus and an object recognition method is shown.

[0163] Reference Fig.12 The computing system 1200 may include at least one processor 1210 , a memory 1230 , a user interface input device 1240 , a user interface output device 1250 , a storage device 1260 , and a network interface 1270 connected to each other via a bus 1220 .

[0164] The processor 1210 may be a central processing unit (CPU) or a semiconductor device that processes instructions stored in the memory 1230 and / or the storage device 1260. The memory 1230 and the storage device 1260 may include various types of volatile or non-volatile storage media. For example, the memory 1230 may include a ROM (Read Only Memory) 1231 and a RAM (Random Access Memory) 1232.

[0165] Thus, the operations of the methods or algorithms described in conjunction with the features disclosed herein may be directly embodied in hardware or software modules, or a combination thereof, executed by the processor 1210. The software modules may reside on a storage medium (i.e., memory 1230 and / or storage device 1260) such as RAM, flash memory, ROM, EPROM, EEPROM, registers, hard disk, removable disk, and CD-ROM.

[0166] An exemplary storage medium may be coupled to the processor 1210, and the processor 1210 may read information from the storage medium and may record information in the storage medium. Alternatively or additionally, the storage medium may be integrated with the processor 1210. The processor and the storage medium may exist in an application specific integrated circuit (ASIC). The ASIC may reside in a user terminal. In another case, the processor and the storage medium may reside in a user terminal as separate components.

[0167] The above description is merely an illustration of the technical concept of the present disclosure, and those skilled in the art to which the present disclosure belongs may make various modifications and changes without departing from the essential characteristics of the present disclosure.

[0168] Therefore, the exemplary features disclosed in the present disclosure are not intended to limit the technical concept of the present disclosure, but are intended to describe the present disclosure, and the scope of the technical concept of the present disclosure is not limited by the illustrated embodiments. The protection scope of the present disclosure should be interpreted by the attached claims, and all technical concepts within the scope equivalent thereto should be interpreted as included within the scope of the present disclosure.

[0169] The present technology can improve the accuracy of the position of an object identified by LIDAR by proposing a reference for processing points obtained by a plurality of LIDARs spaced apart from each other.

[0170] Furthermore, the present technique may improve the accuracy of the position of a host vehicle identified in a confined space by proposing a reference for passing through multiple LIDAR processing points.

[0171] Furthermore, the present technology can reduce the amount of calculation by processing points obtained by a plurality of LIDARs only when the object to be recognized is a specified object.

[0172] Furthermore, various effects directly or indirectly understood through the present disclosure can be provided.

[0173] Although the present disclosure has been described above with reference to some embodiments and drawings, the present disclosure is not limited thereto, but various modifications and changes may be made by those skilled in the art without departing from the spirit and scope of the present disclosure as claimed in the appended claims.

Claims

1. An object recognition device, comprising: First sensor; Second sensor; processor; as well as a memory storing instructions that, when executed by the processor, cause the apparatus to: Identify a plurality of first candidate contour points of a stationary object based on the sensing information of the first sensor, wherein the plurality of first candidate contour points correspond to a specific frame; determine a plurality of second candidate contour points of the stationary object based on the sensing information of the second sensor, wherein the plurality of second candidate contour points correspond to the specific frame; When a vehicle is traveling in an area associated with a predefined type, contour points of the stationary object are determined based on a direction from the vehicle to the stationary object, wherein the determined contour points of the stationary object are further based on: The adjustment of the second candidate contour points and the plurality of first candidate contour points; or The adjustment of the first candidate contour points and the plurality of second candidate contour points; and Signals representing the determined contour points of the stationary object are generated, and the location of the stationary object is displayed on a map.

2. The device according to claim 1, wherein: The area associated with the predefined type comprises a tunnel, and Wherein, when the instructions are executed by the processor, the device determines that the stationary object is not a dynamic object based on the contour points not representing a side view mirror and not representing exhaust gas.

3. The device according to claim 1, in, The instructions, when executed by the processor, cause the device to: identifying a shape box representing the stationary object, the shape box comprising at least one of: the plurality of first candidate contour points and the plurality of second candidate contour points, and Determining contour points of the stationary object based on that the length of the line segments of the shape frame in the longitudinal direction is greater than or equal to a first threshold length, and the length of the line segments of the shape frame in the transverse direction is greater than or equal to a second threshold length, The longitudinal direction includes a direction parallel to the road boundary direction, The lateral direction includes a direction perpendicular to the road boundary direction.

4. The device according to claim 1, wherein: The instructions, when executed by the processor, cause the apparatus to identify a shape box in the specific frame, wherein the shape box includes at least one of: the plurality of first candidate contour points and the plurality of second candidate contour points, and determining the contour point based on that the length of the line segment of the shape frame in the longitudinal direction is longer than the length of the line segment of the shape frame in the transverse direction, The longitudinal direction includes a direction parallel to the road boundary direction, The lateral direction includes a direction perpendicular to the road boundary direction.

5. The device according to claim 1, wherein: The instructions, when executed by the processor, cause the device to determine the contour points based on: the plurality of second candidate contour points; and The second sensor is closer to the stationary object than the first sensor is to the stationary object, and The adjusting of the first candidate contour points includes excluding at least one of the first candidate contour points.

6. The device according to claim 1, wherein: The instructions, when executed by the processor, cause the device to determine the contour points based on: the plurality of first candidate contour points; and The stationary object is located in a lateral direction of the vehicle, the first sensor is located in the lateral direction of the vehicle, and The adjusting of the second candidate contour points includes excluding at least one of the second candidate contour points.

7. The device according to claim 1, wherein: The instructions, when executed by the processor, cause the device to: moving at least one of the plurality of first candidate contour points toward a traveling direction of the vehicle, and determining the at least one contour point, The movement amount of at least one candidate contour point among the plurality of first candidate contour points is determined according to a distance identified based on at least one of the following items: The coordinates of the first point among the plurality of first candidate contour points, coordinates of a second point corresponding to the first point among the plurality of second candidate contour points, a difference between a time when the first sensor acquires the specific frame and a time when the second sensor acquires the specific frame, and The speed of the vehicle.

8. The object recognition device according to claim 1, wherein: an adjusted spacing between a first point obtained via the first sensor and an adjusted point obtained via the second sensor corresponding to the first point is smaller than an unadjusted spacing between the first point and an unadjusted point obtained via the second sensor corresponding to the first point, and The first point and the unadjusted point are obtained from the same portion of the stationary object.

9. The device according to claim 1, wherein: The instructions, when executed by the processor, cause the device to: Based on recognizing that the stationary object is in a side direction of the vehicle, moving a point corresponding to another side direction of the vehicle different from the one side direction and obtained by one of the first sensor and the second sensor toward a traveling direction of the vehicle; as well as Based on one of the first sensor and the second sensor corresponding to the one side direction, a contour point of the stationary object is determined.

10. The device according to claim 1, wherein: The stationary objects include emergency escape routes in tunnels, and The first sensor and the second sensor are mounted on the vehicle and are arranged to be spaced apart from each other.

11. An object recognition device, comprising: First sensor; Second sensor; processor; as well as a memory storing instructions that, when executed by the processor, cause the apparatus to: Identifying a plurality of first candidate contour points of a stationary object based on the sensing information of the first sensor, wherein the plurality of first candidate contour points correspond to a specific frame; identifying a plurality of second candidate contour points of the stationary object based on the sensing information of the second sensor, wherein the plurality of second candidate contour points correspond to the specific frame; and When the vehicle is traveling in an area associated with a predefined type, adjusting the plurality of first candidate contour points or the plurality of second candidate contour points to reduce a distance between the plurality of first candidate contour points and the plurality of second candidate contour points, wherein a first candidate contour point among the plurality of first candidate contour points and a second candidate contour point among the plurality of second candidate contour points represent the same part of the stationary object, and The distance is determined based on a distance between the one first candidate contour point among the plurality of first candidate contour points and the one second candidate contour point among the plurality of second candidate contour points.

12. The device according to claim 11, wherein: The instructions, when executed by the processor, cause the apparatus to move the plurality of first candidate contour points along a traveling direction of the vehicle based on the distance between the second sensor and the stationary object being closer than the distance between the first sensor and the stationary object, and The movement amounts of the plurality of first candidate contour points are determined according to a distance identified based on at least one of the following items: The coordinates of the first point among the plurality of first candidate contour points, coordinates of a second point corresponding to the first point among the plurality of second candidate contour points, a difference between a time when the first sensor acquires the specific frame and a time when the second sensor acquires the specific frame, and The speed of the vehicle.

13. The device according to claim 11, wherein: When the instructions are executed by the processor, the apparatus adjusts the plurality of first candidate contour points or the plurality of second candidate contour points by: Based on recognizing that the stationary object is in a first lateral direction of the vehicle, one or more candidate contour points identified based on points corresponding to a second lateral direction of the vehicle different from the first lateral direction of the vehicle are moved in a traveling direction of the vehicle, The movement amount of the one or more candidate contour points is determined based on at least one of the following identified distances: the coordinates of the one or more candidate contour points, coordinates of one or more candidate contour points identified based on the point corresponding to the first side direction, a difference between a time when the first sensor acquires the specific frame and a time when the second sensor acquires the specific frame, and The speed of the vehicle.

14. A method for object recognition, comprising: identifying, by a processor, a plurality of first candidate contour points of a stationary object based on sensing information of a first sensor, wherein the plurality of first candidate contour points correspond to a specific frame; identifying a plurality of second candidate contour points of the stationary object based on sensing information of a second sensor, wherein the plurality of second candidate contour points correspond to the specific frame; When a vehicle travels in an area associated with a predefined type, a contour point of the stationary object is determined based on a direction from the vehicle to the stationary object, wherein the determined contour point of the stationary object is further based on: adjustment of the second candidate contour points and the plurality of first candidate contour points; or The adjustment of the first candidate contour points and the plurality of second candidate contour points; and Signals representing the determined contour points of the stationary object are generated, and the location of the stationary object is displayed on a map.

15. The method according to claim 14, wherein: The area associated with the predetermined type comprises a tunnel, and The method further includes determining that the stationary object is not a dynamic object based on contour points that do not represent a side mirror and do not represent exhaust gas.

16. The method according to claim 14, further comprising: determining, by the processor, a shape box representing the stationary object, wherein the shape box includes at least one of: the plurality of first candidate contour points and the plurality of second candidate contour points; and The processor determines the contour points of the stationary object based on that the length of the line segments of the shape box in the longitudinal direction is greater than or equal to a first threshold length, and the length of the line segments of the shape box in the transverse direction is greater than or equal to a second threshold length, The longitudinal direction includes a direction parallel to the road boundary direction, The lateral direction includes a direction perpendicular to the road boundary direction.

17. The method according to claim 14, further comprising: identifying, by the processor, a shape box in the specific frame, wherein the shape box includes at least one of: the plurality of first candidate contour points and the plurality of second candidate contour points; and determining, by the processor, the contour point based on that a length of a line segment of the shape frame in a longitudinal direction is longer than a length of a line segment of the shape frame in a transverse direction, The longitudinal direction includes a direction parallel to the road boundary direction, The lateral direction includes a direction perpendicular to the road boundary direction.

18. The method according to claim 14, wherein: Determining the contour points comprises: The processor determines the contour points based on: the plurality of second candidate contour points; and The second sensor is closer to the stationary object than the first sensor is to the stationary object, and The adjusting of the first candidate contour points includes excluding at least one of the first candidate contour points.

19. The method according to claim 14, wherein: Determining the contour points comprises: The processor determines the contour points based on: the plurality of first candidate contour points; and The stationary object is located at a side direction of the vehicle, wherein the first sensor is located at the side direction of the vehicle, and The adjusting of the second candidate contour points includes excluding at least one of the second candidate contour points.

20. The method according to claim 14, wherein: Determining the contour point includes: based on the distance between the second sensor and the stationary object being closer than the distance between the first sensor and the stationary object: moving, by the processor, at least one candidate contour point of the plurality of first candidate contour points toward a traveling direction of the vehicle; and determining, by the processor, the at least one contour point, The movement amount of at least one candidate contour point among the plurality of first candidate contour points is determined according to a distance identified based on at least one of the following items: The coordinates of the first point among the plurality of first candidate contour points, coordinates of a second point corresponding to the first point among the plurality of second candidate contour points, a difference between a time when the first sensor acquires the specific frame and a time when the second sensor acquires the specific frame, and The speed of the vehicle.

Images