Device for object recognition and method therefor

Abstract

The present disclosure relates to an apparatus and method for object detection. The object detection device comprises a sensor (e.g., LIDAR) and a processor.The processor may extract a plurality of representative points, determine a first reference contour point, a first representative point, a second representative point, and a third representative point from the plurality of representative points, determine a first distance between the first reference contour point and the second representative point; determine a first angle between a first ray extending from the first reference contour point to the first representative point and a second ray extending from the second representative point to the third representative point, determine a fourth representative point, and determine as a contour point the first representative point, the second representative point, or the fourth representative point.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to an apparatus and method for object detection, and more particularly to a technique for obtaining a contour point of an object based on a point cloud obtained by a LIDAR (Light Detection and Ranging) device. BACKGROUND

[0002] Technologies for environmental detection and obstacle avoidance are essential for autonomous vehicles.

[0003] A vehicle can obtain data from a LIDAR indicating the position of an object in the vehicle's surroundings. The distance from a LIDAR to an object can be measured based on the interval between the time the laser is emitted by the LIDAR and the time the laser reflected from the object is received. A vehicle is able to identify the location of a point contained within the object in a space where the vehicle is located based on the angle of the emitted laser and the distance to the object.

[0004] The data obtained by LIDAR is characterized by high resolution and a large number of points in the data. The importance of this technology for identifying the shape of an object in a vehicle's surroundings from the data by efficiently performing calculations with limited resources is increasing. SUMMARY

[0005] The present disclosure has been made to solve the above-mentioned problems encountered in the prior art while maintaining the advantages of the prior art.

[0006] One aspect of the present disclosure provides an object recognition device and method for identifying contour points among a plurality of representative points.

[0007] One aspect of the present disclosure provides an apparatus and method for object detection to increase the accuracy of contour points.

[0008] One aspect of the present disclosure provides an apparatus and method for object recognition to reduce the computational effort required to identify contour points.

[0009] 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 here will be clearly understood by those skilled in the art to which the present disclosure relates from the following description.

[0010] According to one or more embodiments of the present disclosure, an object recognition device may include: a sensor and a processor, wherein the processor may be configured to: extract, from a plurality of points obtained via the sensor and representing an object, a plurality of representative points comprising at least a portion of surface points of the object; determine a first reference contour point that is one of two outermost endpoints of the plurality of representative points; determine a first representative point that is adjacent to the first reference contour point among the plurality of representative points; determine a second representative point that is different from the first reference contour point among the plurality of representative points; determine a third representative point that is a point,which is further away from the first reference contour point than two representative points on either side of the second representative point; determine a first distance between the first reference contour point and the second representative point; determine a first angle between a first ray extending from the first reference contour point to the first representative point and a second ray extending from the second representative point to the third representative point; determine, as a contour point, one of: the second representative point, based on the first representative point coinciding with the second representative point, and further based on determining that the first distance is greater than a threshold distance or the first angle is greater than or equal to a threshold angle, one of the first representative point or the second representative point,based on the first representative point not coinciding with the second representative point, further based on the determination that the first distance is greater than the threshold distance or the first angle is greater than or equal to the threshold angle, and further based on the determination that no other representative point exists, from the plurality of representative points, between the first representative point and the second representative point, or either the second representative point or the fourth representative point, based on the first representative point not coinciding with the second representative point, further based on the determination that at least one other representative point exists from the plurality of representative points between the first representative point and the second representative point, and further based on the determination,that the first distance is greater than the threshold distance or the first angle is greater than or equal to the threshold angle. The fourth representative point may be a point that is closer to the first reference contour point among the two representative points on either side of the second representative point; and to control a vehicle based on the contour point.

[0011] The processor may be further configured to determine as the contour point one of the following points: the second representative point, based on the determination that the first distance is greater than the threshold distance, and further based on the agreement of the first representative point with the second representative point, the first representative point based on the determination that the first distance is greater than the threshold distance, and further based on the determination that no other representative point exists between the first representative point and the second representative point, or the fourth representative point based on the determination that the first distance is greater than the threshold distance, and further based on the determination that the at least one other representative point exists between the first representative point and the second representative point.

[0012] The processor may be further configured to determine one of: the second representative point as the contour point based on the second representative point being one of the two outermost endpoints, and further based on determining that the first distance is less than or equal to the threshold distance and the first angle is less than the threshold angle, or the contour point that does not exist between the first representative point, the second representative point, and the fourth representative point based on the second representative point not being one of the two outermost endpoints.

[0013] The processor may be further configured to determine the second representative point as a contour point if it determines that the first distance is less than or equal to the threshold distance and the first angle is greater than or equal to the threshold angle.

[0014] The processor may be further configured to determine as the contour point one of: the second representative point, based on the determination that the second representative point is one of the two outermost endpoints, further based on the determination that the first distance is greater than the threshold distance, and further based on the determination that the first representative point does not coincide with the second representative point, the first representative point, based on the determination that the second representative point is one of the two outermost endpoints, further based on the determination that the first distance is greater than the threshold distance, further based on the determination that the first representative point does not coincide with the second representative point, and further based on the determination that there is no other representative point from the plurality of representative points,exists between the first representative point and the second representative point or the fourth representative point, based on the determination that the second representative point is one of the two outermost endpoints, further based on the determination that the first distance is greater than the threshold distance, further based on the determination that the first representative point does not coincide with the second representative point, and further based on the determination that at least one other representative point from the plurality of representative points exists between the first representative point and the second representative point.

[0015] The processor may be further configured to: determine the second representative point as a contour point based on determining that the second representative point is one of the two outermost endpoints, and further based on determining that the first distance is less than or equal to the threshold distance.

[0016] The processor may be further configured to: determine a fifth representative point, which is a point that is farther from the first reference contour point than two representative points on both sides of the third representative point, based on determining that the contour point does not exist among the first representative point, the second representative point, and the fourth representative point; and determine one of the second representative point or the third representative point as the contour point based on determining that a second distance between the first reference contour point and the third representative point is greater than the threshold distance or a second angle between the first ray and a third ray extending from the third representative point to the fifth representative point is greater than or equal to the threshold angle.

[0017] The processor may be further configured to: determine a second reference contour point and a third reference contour point. The third reference contour point is one of two contour points on either side of the second reference contour point; determine a sixth representative point that is not between the second reference contour point and the third reference contour point; determine a seventh representative point that is one of two representative points on either side of the sixth representative point that is farther from the second reference contour point; determine a second distance between the third reference contour point and the seventh representative point;Determining a second angle between a third ray extending from the second reference contour point to the third reference contour point and a fourth ray extending from the sixth representative point to the seventh representative point; and determining, as the contour point, one of the seventh representative points based on determining that the second distance is greater than the threshold distance or the second angle is greater than or equal to the threshold angle, and further based on the third reference contour point coinciding with the sixth representative point;or one of the sixth representative point or the seventh representative point, based on determining that the second distance is greater than the threshold distance or the second angle is greater than or equal to the threshold angle, and further based on the third reference contour point not coinciding with the sixth representative point;

[0018] The processor may be further configured to determine as the contour point one of the following points: the seventh representative point based on the second distance being greater than the threshold distance and further based on the third reference contour point matching the sixth representative point, or the sixth representative point based on the second distance being greater than the threshold distance and further based on the third reference contour point not matching the sixth representative point.

[0019] The processor may be further configured to: determine whether the seventh representative point is one of the two outermost endpoints based on determining that the second distance is less than or equal to the threshold distance and the second angle is less than the threshold angle; and determine whether: the seventh representative point is the contour point based on the seventh representative point being one of the two outermost endpoints, or whether the contour point does not exist between the sixth representative point and the seventh representative point based on the seventh representative point not being one of the two outermost endpoints.

[0020] The processor may be further configured to determine the seventh representative point as a contour point if it determines that the second distance is less than or equal to the threshold distance and the second angle is greater than or equal to the threshold angle.

[0021] The processor may be further configured to: determine an eighth representative point, which is a point that is farther from the second reference contour point than two representative points on either side of the seventh representative point, based on determining that the contour point does not exist between the sixth representative point and the seventh representative point; and determine one of the seventh representative point and the eighth representative point as the contour point based on determining that a third distance between the third reference contour point and the eighth representative point is greater than the threshold distance or a third angle between the third ray and a fifth ray extending from the seventh representative point to the eighth representative point is greater than or equal to the threshold angle.

[0022] According to one or more embodiments of the present disclosure, an object recognition method may include: extracting a plurality of points obtained via a sensor and representing an object from a plurality of representative points including at least a portion of surface points of the object; determining a first reference contour point that is one of two outermost endpoints of the plurality of representative points; determining a first representative point that is adjacent to the first reference contour point among the plurality of representative points; determining a second representative point that is different from the first reference contour point among the plurality of representative points; determining a third representative point that is a pointwhich is further away from the first reference contour point from two representative points on either side of the second representative point; determining a first distance between the first reference contour point and the second representative point; determining a first angle between a first ray extending from the first reference contour point to the first representative point and a second ray extending from the second representative point to the third representative point; determining, as a contour point, one of: the second representative point, based on the first representative point coinciding with the second representative point, and further based on determining that the first distance is greater than a threshold distance or the first angle is greater than or equal to a threshold angle, one of the first representative point or the second representative point, based onthat the first representative point does not coincide with the second representative point, further based on determining that the first distance is greater than the threshold distance or the first angle is greater than or equal to the threshold angle, further based on determining that no other representative point of the plurality of representative points exists between the first representative point and the second representative point, or one of the second representative point or a fourth representative point, based on the first representative point does not coincide with the second representative point, further based on determining that at least one other representative point of the plurality of representative points exists between the first representative point and the second representative point, and further based on determiningthat the first distance is greater than the threshold distance or the first angle is greater than or equal to the threshold angle. The fourth representative point may be a point that is closer to the first reference contour point among the two representative points on either side of the second representative point; and controlling a vehicle based on the contour point.

[0023] The object detection method may further comprise determining one of the following as the contour point: the second representative point based on determining that the first distance is greater than the threshold distance and further based on the first representative point matching the second representative point, the first representative point based on determining that the first distance is greater than the threshold distance and further based on determining that no other representative point exists between the first representative point and the second representative point, or the fourth representative point based on determining that the first distance is greater than the threshold distance and further based on determining that the at least one other representative point exists between the first representative point and the second representative point.

[0024] The object detection method may further include determining one of: the second representative point as a contour point based on the second representative point being one of the two outermost endpoints, and further based on the determination that the first distance is less than or equal to the threshold distance and the first angle is less than the threshold angle, or the contour point that does not exist between the first representative point, the second representative point, and the fourth representative point based on the second representative point not being one of the two outermost endpoints.

[0025] The object detection method may further include determining the second representative point as a contour point based on determining that the first distance is less than or equal to the threshold distance and the first angle is greater than or equal to the threshold angle.

[0026] The object detection method may further include determining, as the contour point, one of: the second representative point, based on determining that the second representative point is one of the two outermost endpoints, further based on determining that the first distance is greater than the threshold distance, and further based on the first representative point matching the second representative point, the first representative point, based on determining that the second representative point is one of the two outermost endpoints, further based on determining that the first distance is greater than the threshold distance, further based on the first representative point not matching the second representative point, and further based on determining that no other representative point is present,exists between the first representative point and the second representative point or the fourth representative point, based on determining that the second representative point is one of the two outermost endpoints, further based on determining that the first distance is greater than the threshold distance, further based on the first representative point not coinciding with the second representative point, and further based on at least one other representative point among the plurality of representative points existing between the first representative point and the second representative point.

[0027] The object detection method may further include determining the second representative point as a contour point based on determining that the second representative point is one of the two outermost endpoints and further based on determining that the first distance is less than or equal to the threshold distance.

[0028] The object detection method may further include: determining a fifth representative point, which is a point farther from the first reference contour point, among two representative points on both sides of the third representative point, based on determining that the contour point does not exist between the first representative point, the second representative point, and the fourth representative point; and determining either the second representative point or the third representative point as the contour point based on determining that a second distance between the first reference contour point and the third representative point is greater than the threshold distance or a second angle between the first ray and a third ray extending from the third representative point to the fifth representative point is greater than or equal to the threshold angle.

[0029] The object recognition method may further include: determining a second reference contour point and a third reference contour point. The third reference contour point may be one of two contour points on both sides of the second reference contour point; determining a sixth representative point that is not between the second reference contour point and the third reference contour point; determining a seventh representative point that is a point that is farther from the second reference contour point than two representative points on both sides of the sixth representative point; determining a second distance between the third reference contour point and the seventh representative point;Determining a second angle between a third ray extending from the second reference contour point to the third reference contour point and a fourth ray extending from the sixth representative point to the seventh representative point; and determining, as the contour point, one of the seventh representative points based on determining that the second distance is greater than the threshold distance or the second angle is greater than or equal to the threshold angle, and further based on the third reference contour point coinciding with the sixth representative point;or one of the sixth representative point or the seventh representative point, based on determining that the second distance is greater than the threshold distance or the second angle is greater than or equal to the threshold angle, and further based on the third reference contour point not coinciding with the sixth representative point; BRIEF DESCRIPTION OF THE DRAWINGS

[0030] The above and other objects, features and advantages of the present disclosure will become more apparent from the following detailed description taken in conjunction with the accompanying drawings: Fig. 1 is a block diagram showing an object detection apparatus; Fig. Figure 2 is an example showing the change in the direction of movement of an object identified by selecting contour points; Fig. 3 shows a flowchart of the operation of an object recognition device for identifying contour points according to a specified angle and a specified distance in the object recognition device or an object recognition method; Fig. 4 shows a flowchart of the operation of the object recognition device for identifying a contour point by a reference contour point in the object recognition device or the object recognition method; Fig. 5 shows examples of a reference contour point for identifying contour points and identified representative points in an object recognition device or method; Fig. 6 shows a flowchart of the operation of the object recognition device for identifying a contour point by a reference contour point in the object recognition device or the object recognition method; Fig. 7 is a flowchart showing the operation of the object recognition device for identifying a specific representative point as a contour point by a reference contour point in the object recognition device or the object recognition method; Fig. 8 shows examples of a plurality of reference contour points for identifying contour points and identified representative points in an object recognition device or method; Fig. 9 is a flowchart showing the operation of the object recognition device for identifying a contour point by a plurality of reference contour points in the object recognition device or the object recognition method; Fig. 10 shows examples of identified contour points in an object recognition device or method; Fig. 11 shows an example of reliability for indicating the accuracy of the identified contour points in an object recognition device or method; and Fig. 12 shows a computer system for an object recognition device or an object recognition method. DETAILED DESCRIPTION

[0031] One or more exemplary embodiments of the present disclosure will now be described in detail with reference to the exemplary drawings. When adding reference numerals to the components in each drawing, it should be noted that the identical or equivalent component will be designated by the identical numeral even if it is shown in other drawings. Furthermore, in describing the exemplary embodiments of the present disclosure, a detailed description of well-known features or functions will be omitted in order not to unnecessarily obscure the essence of the present disclosure.

[0032] In describing the components of the exemplary embodiments of the present disclosure, terms such as first, second, "A," "B," (a), (b), and the like may be used. These terms are used merely to distinguish one component from another, and the terms do not limit the type, order, or sequence of the components. Unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure relates.Such terms, as defined in a commonly used dictionary, are to be construed as having a meaning consistent with contextual meanings in the relevant field of art and not as having an ideal or overly formal meaning, unless they are clearly defined as such in the present application.

[0033] Furthermore, the terms “unit”, “device”, “element”, “body” or the like used below may refer to at least one shape structure or a unit for performing a function.

[0034] Furthermore, in the examples of the present disclosure, the terms "greater than" or "less than" may be used to indicate whether a particular condition is met or satisfied; however, they are for illustrative purposes only and do not exclude "greater than or equal to" or "less than or equal to." A condition indicating "greater than or equal to" may be replaced with "greater than," a condition indicating "less than or equal to" may be replaced with "less than," and a condition indicating "greater than or equal to and less than" may be replaced with "greater than and less than or equal to." Furthermore, "A" to "B" means at least one of the elements from A (including A) to B (including B).

[0035] Hereinafter, one or more embodiments of the present disclosure will be described with reference to the Fig. 1 to 12 described in detail.

[0036] Fig. 1 is a block diagram showing an object detection device.

[0037] Referring to Fig. 1, an object detection device 101 may be installed in a vehicle. In this case, the object detection device 101 may be integrated into the vehicle's internal control units or implemented as a separate device and connected to the vehicle's control units by separate connecting means.

[0038] Referring to Fig. 1, the object detection device 101 may include a sensor (e.g., LIDAR) 103 and a processor 105.

[0039] The processor 105 of the object recognition device 101 can obtain position information of points of an object around a vehicle including the object recognition device 101 via the sensor (e.g., LIDAR) 103. The processor 105 of the object recognition device 101 can acquire points representing the object through the sensor (e.g., LIDAR) 103. The points representing the object obtained by the sensor (e.g., LIDAR) may be referred to as a point cloud, but the present disclosure is not limited thereto.

[0040] The processor 105 of the object recognition device 101 may extract a plurality of representative points that include some or all of the exterior points (e.g., surface points) among the points that represent the object.

[0041] The processor 105 of the object recognition device 101 can identify contour points from the plurality of representative points. The contour points can represent the external shape of an object. The contour points can include representative points capable of representing the external shape of the object from the plurality of representative points.

[0042] The processor 105 of the object recognition device 101 can obtain information about the object, such as the shape of the object and the direction of movement of the object, from the contour points.

[0043] Because an autonomous driving system or a driver assistance system must be able to detect an environment instantly, the amount of memory available to an object detector and the available execution time may be limited. The object detector can identify an object within a limited amount of memory and execution time by reducing the number of contour points. This is because the computational effort required to obtain information about the object may decrease as the number of contour points decreases.

[0044] However, as the number of contour points decreases, the accuracy of the information about the object may decrease. To compensate for the decrease in information accuracy, the processor 105 of the object recognition device 101 may efficiently identify contour points that can represent the shape of the object among the plurality of representative points to improve the accuracy of the information about the object.

[0045] The processor 105 of the object recognition device 101 may identify a contour point based on a reference contour point according to a first method. The processor 105 of the object recognition device 101 may identify one of the two outermost (e.g., most protruding or farthest from the object center) endpoints among the plurality of representative points as a first reference contour point.The processor 105 of the object recognition device 101 can identify a first representative point, which is the representative point adjacent to the first reference contour point, a second representative point, a third representative point, which is the representative point farther from the first reference contour point among the representative points on both sides of the second representative point, and a fourth representative point, which is a representative point closer to the first reference contour point among the representative points on both sides of the second representative point. The processor 105 of the object recognition device 101 can identify a contour point different from the first reference contour point based on the first reference contour point, the first representative point, the second representative point, and the third representative point, according to the first method.

[0046] The processor 105 of the object recognition device 101 can determine that a contour point does not exist among the first representative point, the second representative point, the third representative point, and the fourth representative point based on the first reference contour point, the first representative point, the second representative point, the third representative point, and the third representative point according to the first method.The processor 105 of the object recognition device 101 may identify a fifth representative point, which is the representative point farther from the first reference contour point, among the representative points on both sides of the third representative point, and in turn identify a contour point different from the first reference contour point based on the first reference contour point, the first representative point, the third representative point, and the fifth representative point, according to the first method.

[0047] If at least one contour point different from the first reference contour point is identified by the first method, the processor 105 of the object recognition device 101 may identify a contour point different from the plurality of contour points according to a second method. The processor 105 of the object recognition device 101 may identify a second reference contour point and a third reference contour point that is one of two contour points on either side of the second reference contour point.

[0048] The processor 105 of the object recognition device 101 may identify a sixth representative point that is not located between the second reference contour point and the third reference contour point and a seventh representative point that is a representative point that is farther away from the second reference contour point among two representative points on both sides of the sixth representative point.

[0049] The processor 105 of the object recognition device 101 may identify a contour point different from the second reference contour point and the third reference contour point based on the second reference contour point, the third reference contour point, the sixth representative point, and the seventh representative point according to a second method.

[0050] The processor 105 of the object recognition device 101 can determine that no contour point exists between the sixth representative point and the seventh representative point based on the second reference contour point, the third reference contour point, the sixth representative point, and the seventh representative point according to the second method. The processor 105 of the object recognition device 101 can identify an eighth representative point, which is a representative point farther from the second reference contour point, among the representative points on both sides of the seventh representative point, and identify the contour point different from the second reference contour point and the third reference contour point based on the first reference contour point, the second reference contour point, the seventh representative point, and the eighth representative point according to the second method.

[0051] The first method is described below with reference to the Fig. 4 to 7 and the second method with reference to the Fig. 8 to 9.

[0052] Fig. Figure 2 shows an example showing the change in the direction of movement of an object identified by selecting contour points.

[0053] Referring to Fig. 2, a first screen 201 may display the direction of movement of an object identified based on contour points. A first situation 203 of the first screen 201 may display the direction of movement of an object identified based on contour points, including a first contour point 205. A second situation 207 of the first screen 201 may display the direction of movement of an object identified based on contour points, including a second contour point 209. A second screen 211 may display the shape of an object identified based on contour points.

[0054] On the first screen 201, the identified representative points of the first situation 203 and the identified representative points of the second situation 207 may be identical. Even if the representative points are the same, the movement direction of the identified object may vary depending on which representative point is designated as a contour point by an existing object recognition device. For example, in the first situation 203, if the object recognition device identifies the first contour point 205 from a plurality of representative points as a contour point, the movement direction of the object may be identified based on the gradient L2 of the first situation 203.For example, in the second situation 207, when the object recognition device identifies a second contour point 209 from a plurality of representative points as a contour point, the movement direction of the object may be identified based on the gradient L2 of the second situation 207. Referring to the first screen 201, the movement direction of the object identified in the first situation 203 and the movement direction of the object identified in the second situation 207 may be different.

[0055] Depending on which representative point is identified as a contour point, the accuracy of the information about the object, including the object's movement direction, may vary. In particular, the movement direction of the identified object may be significantly affected depending on which representative point is set as a contour point among the representative points located in a range where a line connecting the representative points is curved. For example, the θ value indicating the degree to which a mold box of the first situation 203 is inclined may be different from the θ value indicating the degree to which a mold box of the second situation 207 is inclined.

[0056] The processor of the object recognition device can increase the number of contour points identified at the portion where the line connecting the representative points is curved, compared with the existing object recognition device, thereby improving the accuracy of the information about the object (e.g., the moving direction of an object) identified from the contour points.

[0057] The second screen 211 can display the shape of the identified object according to the specified contour points. Since the outer representative points among the representative points of the section where the line connecting the representative points is curved are not identified as contour points, the error in the information about the object identified from the contour points (e.g., the moving direction of the object) may be larger than the error in the information about the object identified from the contour points including the contour point that is an outer representative point.

[0058] A flowchart of the operation of the object recognition device which identifies contour points to reduce errors in the information about the object identified by the contour points is presented with reference to Fig. 3 described.

[0059] Fig. 3 shows a flowchart of the operation of an object recognition device for identifying contour points according to a certain angle and a certain distance in the object recognition device or an object recognition method.

[0060] In the following, it is assumed that the object recognition device 101 of Fig. 1 the process of Fig. 3. In addition, the description of Fig. 3 the operations described as being performed by the device are understood to be controlled by the processor 105 of the object recognition device 101.

[0061] Referring to Fig. 3, in the first operation 301, the processor of the object recognition device may define a first reference contour point, a first representative point, a second representative point, and a third representative point, or alternatively, a second reference contour point, a third reference contour point, a sixth representative point, and a seventh representative point.

[0062] The processor of the object recognition device may identify a first reference contour point, a first representative point, a second representative point, and a third representative point to identify a contour point by a first method using one reference contour point. Alternatively, the processor of the object recognition device may identify a second reference contour point, a third reference contour point, a sixth representative point, and a seventh representative point to identify a contour point by a second method using two reference contour points.

[0063] In a second operation 303, the processor of the object recognition device may identify a contour point. In the first method, for example, a first angle, which is a reference angle for identifying contour points, may comprise an angle between a half line (e.g., a ray) extending from the first reference contour point (e.g., a starting point) to the first representative point and a half line (e.g., a ray) extending from the second representative point (e.g., a starting point) to the third representative point. In the first method, a first distance, which is a reference distance for identifying contour points, may comprise a distance between the first reference contour point and the second representative point.

[0064] In the second method, for example, a second angle, which is a reference angle for identifying contour points, may include an angle between a half line (e.g., a ray) extending from the second reference contour point to the third representative point and a half line (e.g., a ray) extending from the sixth representative point to the seventh representative point. In the second method, the second distance, which is a reference distance for identifying contour points, may include a distance between the third reference contour point and the seventh representative point.

[0065] A first condition 311 may include that an angle (e.g., first angle, second angle) is less than a certain angle (e.g., about 10°) and a distance (e.g., first distance, second distance) is less than or equal to a certain distance (e.g., about 3 m).

[0066] In the first method, when the first condition 311 is satisfied, the processor of the object recognition device may determine that there is no contour point between the first representative point, the second representative point, and the fourth representative point. The fourth representative point may include a representative point that is closer to the first reference contour point than two representative points on either side of the second representative point. The processor of the object recognition device may identify the fifth representative point, which is the representative point farther from the first reference contour point, among two representative points on either side of the third representative point.The processor of the object recognition device may re-identify the contour point in the first method based on the first reference contour point, the first representative point, the third representative point, and the fifth representative point.

[0067] In the second method, when the first condition 311 is satisfied, the processor of the object recognition device may determine that there is no contour point between the sixth representative point and the seventh representative point. The processor of the object recognition device may identify the eighth representative point, which is the representative point farther from the second reference contour point, among two representative points on both sides of the seventh representative point. The processor of the object recognition device may re-identify the contour point in the second method based on the second reference contour point, the third reference contour point, the seventh representative point, and the eighth representative point.

[0068] A second condition 313 may include a condition that an angle (e.g., first angle, second angle) is smaller than a specific angle (e.g., about 10°) and a distance (e.g., first distance, second distance) is greater than a specific distance (e.g., about 3 m). A third condition 315 may include the condition that an angle is greater than or equal to a specific angle and a distance is greater than the specific distance. A fourth condition 317 may include the condition that an angle is greater than or equal to the specified angle and a distance is less than or equal to the specified distance.

[0069] In the first method, if one of the second condition 313, the third condition 315, or the fourth condition 317 is satisfied, the processor of the object recognition device may identify the first representative point, the second representative point, the third representative point, or the fourth representative point as a contour point.

[0070] In the second method, when the first condition 311 is satisfied and when one of the second condition 313, the third condition 315, or the fourth condition 317 is satisfied, the processor of the object recognition device may identify one of the sixth representative point and the seventh representative point as a contour point.

[0071] The first method for identifying contour points is described below with reference to the Fig. 4 to 7. The second method for identifying contour points is described below with reference to the Fig. 8 to 9.

[0072] Fig. 4 shows a flowchart of the operation of the object recognition device for identifying a contour point by a reference contour point in the object recognition device or the object recognition method.

[0073] In the following, it is assumed that the object recognition device 101 of Fig. 1 the process of Fig. 4. In addition, the description of Fig. 4, the operations described as being performed by the device are understood to be controlled by the processor 105 of the object recognition device 101.

[0074] Referring to Fig. 4, in a first operation 401, the processor of the object recognition device may identify one of the two outermost (e.g., most protruding or farthest from the center of the object) endpoints from a plurality of extracted representative points as a first reference contour point. To extract a plurality of representative points, the processor of the object recognition device may obtain points representing an object from a sensor (e.g., LIDAR). The processor of the object recognition device may identify some or all of the outermost points (e.g., surface points) among the points obtained from the sensor (e.g., LIDAR) as a plurality of representative points.

[0075] In a second operation 403, the processor of the object recognition device may identify a first representative point, a second representative point, and a third representative point. The first representative point may include a representative point adjacent to the first reference contour point among the plurality of representative points. The second representative point may include a representative point that is not between the first reference contour point and the first representative point. The third representative point may include a representative point that is further away from the first reference contour point than two representative points on either side of the second representative point.

[0076] In a third operation 405, the processor of the object recognition device may determine a distance between the first reference contour point and the second representative point.

[0077] In a fourth operation 407, the processor of the object recognition device can determine an angle between a half line (e.g., a ray) extending from the first reference contour point to the first representative point and a half line (e.g., a ray) extending from the second representative point to the third representative point. The angle can be used to determine how much a line segment connecting the second representative point and the third representative point deviates from a line segment connecting the first reference contour point and the first representative point.

[0078] In a fifth operation 409, the object recognition device processor may determine whether the first representative point matches the second representative point. If the first representative point matches the second representative point, the object recognition device processor may perform a sixth operation 411. If the first representative point does not match the second representative point, the object recognition device processor may perform a seventh operation 413.

[0079] In the sixth process 411, the processor of the object recognition device may identify the second representative point as a contour point when the distance is greater than or equal to a specified distance (e.g., about 3 m) or the angle is greater than or equal to a specified angle (e.g., about 10 degrees). This is because when the distance is greater than or equal to the specified distance, the accuracy of the information represented by the contour points may decrease as the distance between the representative points increases. This is also because the importance of the representative points may be high when the angle is greater than the specified angle because the line segment connecting the representative points is curved. The processor of the object recognition device may identify the second representative point as a contour point because there is no representative point closer to the first reference contour point than the second representative point.

[0080] In the seventh operation 413, the object recognition processor may determine whether at least one other representative point exists between the first representative point and the second representative point. If at least one other representative point exists between the first representative point and the second representative point, the object recognition processor may perform an eighth operation 415. If at least one other representative point does not exist between the first representative point and the second representative point, the object recognition processor may perform a ninth operation 417.

[0081] If the distance between the first reference contour point and the second representative point is greater than the specified distance, the processor of the object recognition device can identify the first representative point as the contour point. This is because there is no representative point closer to the first reference contour point than the first representative point.

[0082] The processor of the object recognition device may identify the second representative point as a contour point if an angle between a half line (e.g., a ray) extending from the first reference contour point to the first representative point and a half line (e.g., a ray) extending from the second representative point to the third representative point is greater than a certain angle (e.g., about 10°).

[0083] In the eighth operation 415, the processor of the object recognition device may identify the first or second representative point as a contour point based on the distance being greater than the specified distance or the angle being greater than or equal to the specified angle.

[0084] If the distance between the first reference contour point and the second representative point is greater than the specified distance, the processor of the object recognition device may identify the first representative point, which is closer to the first reference contour point than the second representative point, as a contour point.

[0085] The processor of the object recognition device may identify the second representative point as a contour point if an angle between a half line (e.g., a ray) extending from the first reference contour point to the first representative point and a half line (e.g., a ray) extending from the second representative point to the third representative point is greater than a certain angle (e.g., about 10°).

[0086] In the ninth operation 417, the processor of the object recognition device may identify the second representative point or the fourth representative point as a contour point if the distance is greater than the specified distance or the angle is greater than or equal to the specified angle.

[0087] The fourth representative point may comprise a representative point that is closer to the first reference contour point than two representative points on either side of the second representative point.

[0088] If the distance between the first reference contour point and the second representative point is greater than the specified distance, the processor of the object recognition device may identify the fourth representative point, which is closer to the first reference contour point than the second representative point, as a contour point. This is because the fourth representative point is the representative point farthest from a reference contour point among the representative points within the specified distance.

[0089] The processor of the object recognition device can identify the second representative point as a contour point if an angle between a half line (e.g., a ray) extending from the first reference contour point to the first representative point and a half line (e.g., a ray) extending from the second representative point to the third representative point is greater than a certain angle (e.g., approximately 10°). After one or more contour points have been identified, a vehicle can be controlled based on the identified contour points (e.g., avoiding an object, braking to avoid a collision, etc.).

[0090] Fig. 5 shows examples of a reference contour point for identifying contour points and identified representative points in an object recognition device or method.

[0091] Referring to Fig. 5, the processor of the object recognition device may identify representative points that include some or all of the exterior points (e.g., surface points) among the points obtained by the sensor (e.g., LIDAR). Representative points in a first example 501 may include a first point 503, a second point 505, and a third point 507. Representative points in a second example 511 may include a first point 513, a second point 515, a third point 517, and a fourth point 519. Representative points in a third example 521 may include a first point 523, a second point 525, a third point 527, a fourth point 529, and a fifth point 531.

[0092] The first example 501 shows a method for identifying contour points using a reference contour point when the first representative point and the second representative point match. The first point 503 can be identified as a first reference contour point. The second point 505 can be identified as a first representative point and a second representative point. The third point 507 can be identified as a third representative point.

[0093] The second example 511 shows a method for identifying contour points using a reference contour point when the first representative point and the second representative point are different and at least one other representative point is not located between the first representative point and the second representative point. The first point 513 can be identified as a first reference contour point. The second point 515 can be identified as a first representative point. The third point 517 can be identified as a second representative point. The fourth point 519 can be identified as a third representative point.

[0094] The third example 521 shows a method for identifying contour points using a reference contour point when the first representative point and the second representative point are different and at least one other representative point lies between the first representative point and the second representative point. The first point 523 can be identified as a first reference contour point. The second point 525 can be identified as a first representative point. The third point 527 can be identified as a fourth representative point. The fourth point 529 can be identified as a second representative point. The fifth point 531 can be identified as a third representative point.

[0095] In the first example 501, the processor of the object recognition device may identify the second point 505, which is the first representative point, as a contour point when the distance between the first point 503, which is the first reference contour point, and the second point 505, which is the second representative point, is greater than a certain distance (e.g., about 3 m).

[0096] In the first example 501, if an angle between a half line (e.g., a ray) extending from the first point 503, which is the first reference contour point, to the second point 505, which is the first representative point, and a half line (e.g., a ray) extending from the second point 505, which is the second representative point, to the third point 507, which is the third representative point, is greater than or equal to a certain angle (e.g., about 10°), the processor of the object recognition device can identify the second point 505, which is the first representative point, and the second representative point, as a contour point.

[0097] In the first example 501, the processor of the object recognition device may determine that there is no contour point between the second point 505, which is the first representative point and the second representative point, and the third point 507, which is the third representative point, based on that a distance between the first point 503, which is the first reference contour point, and the second point 505, which is the second representative point, is less than or equal to the specified distance and the angle between the half line (e.g., a ray) extending from the first point 503, which is the first reference contour point, to the second point 505, which is the first representative point, and the half line (e.g., a ray) extending from the second point 505, which is the second representative point, to the third point 507, which is the third representative point, is less than the specified angle.

[0098] The processor of the object recognition device may re-identify a contour point as in the second example 511 by determining that there is no contour point between the second point 505, which is the first representative point and the second representative point, and the third point 507, which is the third representative point.

[0099] In the second example 511, a distance between the first point 513, which is the first reference contour point, and the third point 517, which is the second representative point, is greater than a specified distance (e.g., about 3 m), and the processor of the object recognition device can identify the second point 515, which is the first representative point, as a contour point.

[0100] In the second example 511, if an angle between a half line (e.g., a ray) extending from the first point 513, which is the first reference contour point, to the second point 515, which is the first representative point, and a half line (e.g., a ray) extending from the third point 517, which is the second representative point, to the fourth point 507, which is the third representative point, is greater than or equal to a certain angle (e.g., about 10°), the processor of the object recognition device can identify the third point 517, which is the second representative point, as a contour point.

[0101] In the second example 511, the processor of the object recognition device may identify that there is no contour point from the second point 515, which is the first representative point, the third point 517, which is the second representative point, and the fourth point 519, which is the third representative point, based on identifying that a distance between the first point 513, which is the first reference contour point, and the third point 517, which is the second representative point, is less than or equal to the specified distance and the angle between the half line (e.g., a ray) extending from the first point 513, which is the first reference contour point, to the second point 515, which is the first representative point, and the half line (e.g., a ray) extending from the third point 517, which is the second representative point, to the fourth point 519, which is the third representative point, is less than the specified angle.

[0102] The processor of the object recognizer may re-identify a contour point as in the third example 521 by determining that there is no contour point among the second point 515, which is the first representative point, the third point 517, which is the second representative point, and the fourth point 519, which is the third representative point.

[0103] In the third example 521, when a distance between the first point 523, which is the first reference contour point, and the fourth point 529, which is the second representative point, is greater than a certain distance (e.g., about 3 m), the processor of the object recognition device may identify the third point 527, which is the fourth representative point, as a contour point.

[0104] In the third example 521, if an angle between a half line (e.g., a ray) extending from the first point 523, which is the first reference contour point, to the second point 525, which is the first representative point, and a half line (e.g., a ray) extending from the fourth point 529, which is the second representative point, to the fifth point 531, which is the third representative point, is greater than or equal to a certain angle (e.g., about 10°), the processor of the object recognition device can identify the fourth point 529, which is the second representative point, as a contour point.

[0105] In the third example 521, the processor of the object recognition device may determine that there is no contour point between the second point 525, which is the first representative point, the third point 527, which is the fourth representative point, the fourth point 529, which is the second representative point, and the fifth point 531, which is the third representative point, based on the identification that a distance between the first point 523, which is the first reference contour point, and the fourth point 529, which is the second representative point, is less than or equal to the specified distance, and the angle between the half line (e.g., a ray) extending from the first point 523, which is the first reference contour point, to the second point 525, which is the first representative point, and the half line (e.g.,a ray) extending from the fourth point 529, which is the second representative point, to the fifth point 531, which is the third representative point, is smaller than the specified angle.

[0106] Fig. 6 shows a flowchart of the operation of the object recognition device for identifying a contour point by a reference contour point in the object recognition device or the object recognition method.

[0107] In the following, it is assumed that the object recognition device 101 of Fig. 1 the process of Fig. 6. In addition, the description of Fig. 6 the operations described as being performed by the device are understood to be controlled by the processor 105 of the object recognition device 101.

[0108] Referring to Fig. 6, in a first operation 601, the processor of the object recognition device may identify a first reference contour point and a first representative point. The first reference contour point may comprise one of the two outermost (e.g., most protruding or farthest from the center of the object) endpoints of the plurality of representative points. The first representative point may comprise a representative point adjacent to the first reference contour point.

[0109] In a second operation 603, the processor of the object recognition device may identify a second representative point and a third representative point. The second representative point may coincide with the first representative point. The third representative point may include a representative point that is further from the first reference contour point than two representative points on either side of the second representative point.

[0110] In a third operation 605, the processor of the object recognizer may determine whether the second representative point is one of the two endpoints. If the second representative point is one of the two endpoints, the processor of the object recognizer may perform a fourth operation 607. If the second representative point is not one of the two endpoints, the processor of the object recognizer may perform a fifth operation 609. In other words, if the second representative point is one of the two endpoints, the processor of the object recognizer may identify a contour point through an extraordinary operation because it is difficult to identify the third representative point.

[0111] In a fifth operation 609, the processor of the object recognition device can determine which of the second representative point and the fourth representative point is a contour point. A method for identifying a contour point when the second representative point is one of two endpoints is described below (see FIG. Fig. 7).

[0112] In the fourth operation 607, the processor of the object recognition device may determine whether a distance between the first reference contour point and the second representative point is greater than a specified distance. If the distance between the first reference contour point and the second representative point is greater than the specified distance, the processor of the object recognition device may perform a sixth operation 611. If the distance between the first reference contour point and the second representative point is not greater than the specified distance, the processor of the object recognition device may perform a seventh operation 613.

[0113] This is because, when the second representative point is identified as a contour point, the distance between the first reference contour point and a contour point identified based on the first reference contour point may be greater than the specified distance. If the distance between the contour points is greater than the specified distance, the accuracy of comparing the shape of the object identified by the contour points with a precision map may decrease.Therefore, in order to ensure that the reliability of the contour line of an object identified by contour points is greater than a specified reliability value, the processor of the object recognition device may identify the second representative point or the fourth representative point as a contour point when the distance between the first reference contour point and the second representative point is greater than a certain distance.

[0114] In a sixth operation 611, the object recognizer processor may determine whether the first representative point and the second representative point match. If the first representative point and the second representative point match, the object recognizer processor may identify an eighth operation 615. If the first representative point and the second representative point do not match, the object recognizer processor may identify a ninth operation 617.

[0115] In the eighth operation 615, the processor of the object recognition device may identify the second representative point as a contour point. If the first representative point and the second representative point match, there is no representative point closer to the first reference contour point than the second representative point.

[0116] In the ninth operation 617, the processor of the object recognition device may identify the fourth representative point as a contour point. This is because the fourth representative point is the representative point farthest from a reference contour point among the representative points located within the specified distance.

[0117] In the seventh operation 613, the object recognition processor may identify whether an angle identified based on the first reference contour point, the first representative point, the second representative point, and the third representative point is smaller than a specified angle. If the angle identified based on the first reference contour point, the first representative point, the second representative point, and the third representative point is smaller than the specified angle, the object recognition processor may perform a tenth operation 619. If the angle identified based on the first reference contour point, the first representative point, the second representative point, and the third representative point is greater than or equal to the specified angle, the object recognition processor may perform a thirteenth operation 625.

[0118] This is because the importance of the second representative point can be high because the second representative point is located in a region where the angle of a line segment connecting the representative points is curved when the angle is greater than or equal to the specified angle. An angle identified based on the first reference contour point, the first representative point, the second representative point, and the third representative point may include an angle between a half line (e.g., a ray) extending from the first reference contour point to the first representative point and a half line (e.g., a ray) extending from the second representative point to the third representative point.

[0119] In the tenth operation 619, the object recognizer processor may determine whether the second representative point is one of two endpoints. If the second representative point is one of the two endpoints, the object recognizer processor may perform an eleventh operation 621. If the second representative point is not one of the two endpoints, the object recognizer processor may perform a twelfth operation 623.

[0120] Even if the distance between the first reference contour point and the second representative point is less than or equal to the specified distance and the angle is smaller than the specified angle, the processor of the object recognition device can identify the second representative point as a contour point if the second representative point is one of the two end points.

[0121] In the eleventh operation 621, the processor of the object recognition device may identify the second representative point as a contour point.

[0122] In the twelfth operation 623, the processor of the object recognition device may identify the second representative point and the third representative point as representative points rather than contour points. In other words, the processor of the object recognition device may determine that there is no contour point between the first representative point, the second representative point, and the third representative point. Furthermore, the processor of the object recognition device may perform the second operation 603 again.It should be noted that when performing the second operation 603 again, the processor of the object recognition device may identify a previous third representative point as the second representative point, and may identify a representative point farther from the first reference contour point among two representative points on both sides of a previous second representative point as the third representative point. In other words, the processor of the object recognition device may identify a representative point farther from the first reference contour point among two representative points on both sides of the third representative point as the fifth representative point based on identifying that there is no contour point among the first representative point, the second representative point, and the fourth representative point.The processor of the object recognition device may be configured to identify either the second representative point or the fourth representative point as a contour point if it determines that the distance between the first reference contour point and the third representative point is greater than the specified distance or that the angle between the half line (e.g., a ray) extending from the first reference contour point to the first representative point and the half line (e.g., a ray) extending from the third representative point to the fifth representative point is greater than or equal to the specified angle.

[0123] In a thirteenth operation 625, the processor of the object recognition device may identify the second representative point as a contour point.

[0124] Fig. 7 shows a flowchart of the operation of the object recognition device for identifying a specific representative point as a contour point by a reference contour point in the object recognition device or the object recognition method.

[0125] In the following, it is assumed that the object recognition device 101 of Fig. 1 the process of Fig. 7. In addition, the description of Fig. 7, the operations described as being performed by the device are understood to be controlled by the processor 105 of the object recognition device 101.

[0126] Referring to Fig. 7, the processor of the object recognition device can identify a first reference contour point and a first representative point in a first operation 701. Since the first operation 701 is carried out in the same way as the first operation 601 of Fig. 6, an overlapping description is omitted below.

[0127] In a second operation 703, the processor of the object recognition device may identify a second representative point and a third representative point. Since the second operation 703 is performed in the same manner as the second operation 603 of Fig. 6, an overlapping description is omitted below.

[0128] In a third operation 705, the processor of the object recognizer may determine whether the second representative point is one of the two endpoints. If the second representative point is one of the two outermost (e.g., most protruding or farthest from the center of the object) endpoints of a plurality of representative points, the processor of the object recognizer may identify a contour point through an exception operation because it is difficult to identify the third representative point.

[0129] In the fourth operation 707, the processor of the object recognition device may determine whether a distance between the first reference contour point and the second representative point is greater than a specified distance. If the distance between the first reference contour point and the second representative point is greater than the specified distance, the processor of the object recognition device may perform a fifth operation 709. If the distance between the first reference contour point and the second representative point is less than or equal to the specified distance, the processor of the object recognition device may perform a sixth operation 711. Since the fourth operation 707 is performed in the same manner as the fourth operation 607 of Fig. 6, an overlapping description is omitted below.

[0130] In a sixth operation 711, the processor of the object recognition device may identify the second representative point as a contour point. Since a half line (e.g., a ray) extending from the second representative point to the third representative point cannot be identified if the second representative point is one of the two outermost (e.g., most protruding or farthest from the center of the object) endpoints of the plurality of representative points, an angle between the half line (e.g., a ray) extending from the first reference contour point to the first representative point and the half line (e.g., a ray) extending from the second representative point to the third representative point may not be identified.Accordingly, the processor of the object recognition device can identify the second representative point as a contour point even if the distance between the first reference contour point and the second representative point is not greater than the specified distance, without determining an angle condition.

[0131] In a fifth operation 709, the processor of the object recognition device may determine whether the first representative point and the second representative point match. If the first representative point matches the second representative point, the processor of the object recognition device may perform a seventh operation 713. If the first representative point does not match the second representative point, the processor of the object recognition device may perform an eighth operation 715. Since the fifth operation 709 is performed in the same manner as the sixth operation 611 of Fig. 6, an overlapping description is omitted below.

[0132] In the seventh operation 713, the processor of the object recognition device can identify the second representative point as a contour point. Since the seventh operation 713 is performed in the same manner as the eighth operation 615 of Fig. 6, an overlapping description is omitted below.

[0133] In the eighth operation 715, the processor of the object recognition device can identify the fourth representative point as a contour point. Since the eighth operation 715 is performed in the same manner as the ninth operation 617 of Fig. 6, a comprehensive description is omitted below.

[0134] Fig. 8 shows examples of a plurality of reference contour points for identifying contour points and identified representative points in an object recognition device or an object recognition method.

[0135] Referring to Fig. 8, the processor of the object recognition device may identify representative points, including some or all of the exterior points (e.g., surface points) from the points obtained by the sensor (e.g., LIDAR). Representative points in a first example 801 may include a first point 803, a second point 805, and a third point 807. Representative points in a second example 811 may include a first point 813, a second point 815, a third point 817, and a fourth point 819.

[0136] The first example 801 shows a method for identifying contour points using two reference contour points when a third reference contour point and a sixth representative point match. The first point 803 can be identified as a second reference contour point. The second point 805 can be identified as a third reference contour point and as a sixth representative point. The third point 807 can be identified as a seventh representative point.

[0137] The second example 811 shows a method for identifying contour points using two reference contour points when a third reference contour point and a sixth representative point do not coincide with each other. The first point 813 can be identified as a second reference contour point. The second point 815 can be identified as a third reference contour point. The third point 817 can be identified as a sixth representative point. The fourth point 819 can be identified as a seventh representative point.

[0138] In the first example 801, if the distance between the second point 805, the third reference contour point, and the third point 807, the seventh representative point, is greater than a specified distance (e.g., about 3 m), the processor of the object recognition device may identify the third point 807, which is the seventh representative point, as a contour point.

[0139] In the first example 801, if an angle between a half line (e.g., a ray) extending from the first point 803, which is the second reference contour point, to the second point 805, which is the third reference contour point, and a half line (e.g., a ray) extending from the second point 805, which is the sixth representative point, to the third point 807, which is the seventh representative point, is greater than or equal to a specified angle (e.g., about 10°), the processor of the object recognition device can identify the third point 807, which is the seventh representative point, as a contour point.

[0140] In the first example 801, the processor of the object recognition device may determine that there is no contour point between the second point 805, which is the sixth representative point, and the third point 807, which is the seventh representative point, based on determining that a distance between the second point 805, which is the third reference contour point, and the third point 807, which is the seventh representative point, is less than or equal to the specified distance and the angle between the half line (e.g., a ray) extending from the first point 803, which is the second reference contour point, to the second point 805, which is the third reference contour point, and the half line (e.g.,a ray) extending from the first point 803, which is the third reference contour point, to the second point 805, which is the seventh reference contour point, a ray) extending from the first point 803, which is the second reference contour point, to the second point 805, which is the third reference contour point, and half the line (e.g., a ray) extending from the second point 805, which is the sixth representative point, to the third point 807, which is the seventh representative point, is smaller than the specified angle.

[0141] The processor of the object recognition device may re-identify a contour point as in the second example 811 by determining that there is no contour point between the second point 805, which is the sixth representative point, and the third point 807, which is the seventh representative point.

[0142] In the second example 811, if a distance between the second point 815, the third reference contour point, and the fourth point 819, the seventh representative point, is greater than a certain distance (e.g., about 3 m), the processor of the object recognition device may identify the third point 817, the sixth representative point, as a contour point.

[0143] In the second example 811, if an angle between a half line (e.g., a ray) extending from the first point 813, which is the second reference contour point, to the second point 815, which is the third reference contour point, and a half line (e.g., a ray) extending from the third point 817, which is the sixth representative point, to the fourth point 819, which is the seventh representative point, is greater than or equal to a certain angle (e.g., about 10°), the processor of the object recognition device can identify the fourth point 819, which is the seventh representative point, as a contour point.

[0144] In the second example 811, the processor of the object recognition device may determine that there is no contour point between the third point 817, which is the sixth representative point, and the fourth point 819, which is the seventh representative point, based on determining that a distance between the second point 815, which is the third reference contour point, and the fourth point 819, which is the seventh representative point, is less than or equal to the specified distance and the angle between the half line (e.g., a ray) extending from the first point 813, which is the second reference contour point, to the second point 815, which is the third reference contour point, and the half line (e.g.,a ray) extending from the first point 813, which is the third reference contour point, to the fourth point 819, which is the seventh representative point, a ray) extending from the first point 813, which is the second reference contour point, to the second point 815, which is the third reference contour point, and the half line (e.g., a ray) extending from the third point 817, which is the sixth representative point, to the fourth point 819, which is the seventh representative point, is smaller than the specified angle.

[0145] Fig. 9 shows a flowchart of the operation of the object recognition device for identifying a contour point by a plurality of reference contour points in the object recognition device or the object recognition method.

[0146] In the following, it is assumed that the object recognition device 101 of Fig. 1 the process of Fig. 9. In addition, the description of Fig. 9 the operations described as being performed by the device are understood to be controlled by the processor 105 of the object recognition device 101.

[0147] Referring to Fig. 9, in a first operation 901, the processor of the object recognition device may identify a second reference contour point and a third reference contour point. The second reference contour point may comprise one of a plurality of representative points. The third reference contour point may comprise one of two contour points on either side of the second reference contour point.

[0148] In a second operation 903, the processor of the object recognition device may identify a sixth representative point and a seventh representative point. The sixth representative point and the seventh representative point may not be located between the second reference contour point and the third reference contour point. The seventh representative point may include a representative point that is further away from the second reference contour point than two representative points on either side of the sixth representative point.

[0149] In the third operation 905, the object recognition device processor may determine whether the distance between the third reference contour point and the seventh representative point is greater than a specified distance. If the distance between the third reference contour point and the seventh representative point is greater than the specified distance, the object recognition device processor may perform a fourth operation 907. If the distance between the third reference contour point and the seventh representative point is less than or equal to the specified distance, the object recognition device processor may perform a fifth operation 909.

[0150] This is because, when the seventh representative point is identified as a contour point, the distance between the third reference contour point and a contour point identified based on the second reference contour point and the third reference contour point may be greater than the specified distance. If the distance between the contour points is greater than the specified distance, the accuracy of comparing the shape of the object identified by the contour points with a precision map may decrease.Therefore, in order to ensure a reliability greater than a specified reliability value for the contour line of an object identified by contour points, a distance between the third contour point and the seventh representative point can be taken into account when identifying contour points.

[0151] In a fourth operation 907, the object recognition device processor may determine whether the sixth representative point and the third reference contour point match. If the sixth representative point and the third reference contour point match, the object recognition device processor may perform a sixth operation 911. If the sixth representative point and the third reference contour point do not match, the object recognition device processor may perform a seventh operation 913.

[0152] In the fifth operation 909, the object recognition device processor may determine whether an angle identified based on the second reference contour point, the third reference contour point, the sixth representative point, and the seventh representative point is smaller than a specified angle. If the angle identified based on the second reference contour point, the third reference contour point, the sixth representative point, and the seventh representative point is smaller than the specified angle, the object recognition device processor may perform an eighth operation 915. If the angle identified based on the second reference contour point, the third reference contour point, the sixth representative point, and the seventh representative point is greater than or equal to the specified angle, the object recognition device processor may perform a ninth operation 921.

[0153] An angle determined based on the second reference contour point, the third reference contour point, the sixth representative point, and the seventh representative point may include an angle between a half line (e.g., a ray) extending from the second reference contour point to the third reference contour point and a half line (e.g., a ray) extending from the sixth representative point to the seventh representative point. This is because the importance of the seventh representative point may be high because the seventh representative point lies in a range where the angle of a line segment connecting the representative points is curved if the angle is greater than or equal to the specified angle.Accordingly, the processor of the object recognition device may consider the angle identified based on the second reference contour point, the third reference contour point, the sixth representative point, and the seventh representative point when identifying contour points.

[0154] In the eighth operation 915, the object recognizer processor may identify whether the seventh representative point is one of two endpoints. If the seventh representative point is one of the two endpoints, the object recognizer processor may perform a tenth operation 917. If the seventh representative point is not one of the two endpoints, the object recognizer processor may perform an eleventh operation 919.

[0155] Even if the distance between the third reference contour point and the seventh representative point is less than or equal to the specified distance and the angle is smaller than the specified angle, the processor of the object recognition device can identify the seventh representative point as a contour point if the seventh representative point is one of the two end points.

[0156] In the tenth operation 917, the processor of the object recognition device may identify the seventh representative point as a contour point.

[0157] In the eleventh operation 919, the object recognition device processor may identify the sixth representative point and the seventh representative point as representative points rather than contour points. Furthermore, the object recognition device processor may repeat the second operation 903.

[0158] The processor of the object recognition device may identify a representative point that is farther from the second reference contour point than two representative points on either side of the seventh representative point as the eighth representative point based on the identification that there is no contour point between the sixth representative point and the seventh representative point. The processor of the object recognition device may be configured to identify the sixth representative point or the seventh representative point as a contour point if it determines that the distance between the third reference contour point and the seventh representative point is greater than the specified distance or that the angle between the half line (e.g., a ray) extending from the second reference contour point to the third reference contour point and the half line (e.g.,a ray) extending from the sixth representative point to the seventh representative point is greater than or equal to the specified angle.

[0159] In the ninth operation 921, the processor of the object recognition device may identify the seventh representative point as a contour point.

[0160] Fig. 10 shows examples of identified contour points in an object recognition device or method.

[0161] Referring to Fig. 10, a first screen 1001 may display representative points and contour points identified by an existing object recognizer. A second screen 1011 may display representative points and contour points identified by an object recognizer. Empty pixels on the first screen 1001 and the second screen 1011 may be representative points that are not identified as contour points. Solid pixels on the first screen 1001 and the second screen 1011 may be representative points that have been identified as contour points.

[0162] If a contour point is identified on the first screen 1001 by the existing object recognition device, a comparison of the angle between the first reference line segment and the subsequent reference line segment for identifying the contour point cannot be performed.

[0163] Since the angle between the identified contour points and the representative point at which it is to be determined whether the representative point is a contour point can be taken into account when a contour point is identified by the object recognition device, representative points whose angle changes by more than a specified angle can be identified as contour points in the second screen 1011.

[0164] Compared with existing object recognition devices, the object recognition device can therefore identify representative points whose angles change by more than a specified angle as contour points.

[0165] Fig. Figure 11 shows an example of reliability in indicating the accuracy of the identified contour points in an object recognition device or method.

[0166] Referring to Fig. 11, a screen 1100 may display a first distribution 1103 of contour points identified by an existing object recognizer and a second distribution 1105 of contour points identified by an object recognizer.

[0167] In the first distribution 1103, the reliability of the contour points identified by the existing object recognizer may be approximately 66.5. In the second distribution 1105, the reliability of the contour points identified by the object recognizer may be approximately 71.5. Accordingly, the reliability of the contour points identified by the object recognizer may be higher than the reliability of the contour points identified by the existing object recognizer.

[0168] The processor of the object recognition device may calculate the reliability of the contour points based on at least a first variable value, a second variable value, a third variable value, a fourth variable value, a fifth variable value, or any combination thereof.

[0169] The first endpoint and the second endpoint can correspond to the two outermost (e.g., most protruding or farthest from the center of the object) endpoints of the contour points. A vertex point can be a contour point that is farthest from a line segment connecting the first endpoint and the second endpoint of the contour points.

[0170] The first line segment may include a line segment connecting the first endpoint and the apex point. The second line segment may include a line segment connecting the second endpoint and the apex point.

[0171] The reliability of the contour points may be determined based on the sum of at least one of the following values: a value obtained by multiplying the first variable value by a weight, a value obtained by multiplying the second variable value by a weight, a value obtained by multiplying the third variable value by a weight, a value obtained by multiplying the fourth variable value by a weight, a value obtained by multiplying the fifth variable value by a weight, or any combination thereof.

[0172] The first variable value may include a value determined based on the variance of the contour points. For example, the processor of the object recognition device may identify the first variable value based on a variance value for distances between the first line segment and contour points near the first line segment and a variance value for distances between the second line segment and contour points near the second line segment.

[0173] The second variable value can be determined based on the distribution of contour points for the first line segment and the distribution of contour points for the second line segment.

[0174] For example, the distribution of contour points for the first line segment may involve identifying whether all contour points near the first line segment are distributed in only one direction relative to the first line segment. The distribution of contour points for the second line segment may involve determining whether all contour points near the second line segment are distributed in only one direction.

[0175] The third variable value may be determined based on the average of the angles of the contour points. The angle of each contour point may comprise an angle between half a line (e.g., a ray) extending from a particular contour point to one of the contour points on either side of the particular contour point and half a line (e.g., a ray) extending from a particular contour point to the other of the contour points on either side of the particular contour point.

[0176] The fourth variable value can be determined based on the distribution of contour points for the first line segment and the distribution of contour points for the second line segment.

[0177] The fifth variable value may be determined based on the ratio between the unidirectional length of a contour box containing contour points in a specific layer and the unidirectional length of a cluster box containing contour points in a plurality of layers.

[0178] Fig. 12 illustrates a computer system for an object recognition device or method.

[0179] Referring to Fig.12, a computer system 1200 may include at least a processor 1210, a memory 1230, a user interface input device 1240, a user interface output device 1250, a mass storage 1260, and a network interface 1270 interconnected via a bus 1220.

[0180] Processor 1210 may be a central processing unit (CPU) or a semiconductor device that processes instructions stored in memory 1230 and / or mass storage 1260. Memory 1230 and mass storage 1260 may include various types of volatile or non-volatile storage media. For example, memory 1230 may include a ROM (Read Only Memory) 1231 and a RAM (Random Access Memory) 1232.

[0181] Thus, the operations of the method or algorithm described in connection with one or more example embodiments disclosed herein may be embodied directly in hardware or a software module executed by processor 1210, or a combination thereof. The software module may be located on a storage medium (i.e., memory 1230 and / or mass storage 1260), such as RAM, flash memory, ROM, EPROM, EEPROM, register, hard disk, removable disk, and CD-ROM.

[0182] The example mass storage medium may be connected to processor 1210, and processor 1210 may read information from the storage medium and store information in the mass storage medium. Alternatively, the mass storage medium may be integrated with processor 1210. The processor and the storage medium may be encapsulated in an application-specific integrated circuit (ASIC). The ASIC may be encapsulated in a user terminal. Alternatively, the processor and the mass storage medium may be encapsulated as separate components in the user terminal.

[0183] The above description is merely illustrative of the technical idea of ​​the present disclosure, and various modifications and variations can be made by those skilled in the art of the present disclosure without departing from the essential features of the present disclosure.

[0184] Accordingly, the embodiments disclosed in the present disclosure are not intended to limit the technical idea of ​​the present disclosure, but to describe the present disclosure, and the scope of the technical idea of ​​the present disclosure is not limited by the embodiments. The scope of the present disclosure should be interpreted by the following claims, and all technical ideas within the equivalent scope should be construed as being included within the scope of the present disclosure.

[0185] The present technology can reduce the amount of computation required for object detection by identifying some of the representative points detected by a sensor (e.g., LIDAR) as contour points.

[0186] Furthermore, the present technology can increase the accuracy of the contour point representing the object by using angle conditions and length conditions.

[0187] Furthermore, the present technology increases the accuracy of representation of an object compared to the number of contour points by increasing the distance between the contour points of parts where the contour of the object does not change and decreasing the distance between the contour points of parts where the contour of the object changes.

[0188] Furthermore, the present technology can increase the accuracy of the movement direction of the identified object by improving the accuracy of the representation of the object by contour points.

[0189] Furthermore, the present technology can improve the user experience by increasing the accuracy of the contour points that represent an object.

[0190] Furthermore, the present technology can reduce the risk of accidents in autonomous vehicles or vehicles with activated driver assistance systems by increasing the accuracy of the direction of movement of the detected object.

[0191] Furthermore, current technology can improve the performance of autonomous driving or driver assistance devices by increasing the accuracy of the movement direction of the detected object.

[0192] In addition, various effects can be provided that are understood directly or indirectly through the disclosure.

[0193] Although the present disclosure has been described with reference to one or more embodiments and the accompanying drawings, the present disclosure is not limited thereto, but may be modified and altered in various ways by those skilled in the art to which the present disclosure relates without departing from the spirit and scope of the present disclosure as claimed in the following claims.

Claims

[1] Object recognition device, comprising: a sensor; and a processor configured to: extracting from a plurality of points obtained via the sensor and representing an object a plurality of representative points comprising at least a portion of the surface points of the object; to determine a first reference contour point which is one of the two extreme endpoints of the plurality of representative points; to determine a first representative point adjacent to the first reference contour point from the plurality of representative points; to determine from the plurality of representative points a second representative point which is different from the first reference contour point; to determine a third representative point, which is a point further away from the first reference contour point than two representative points on either side of the second representative point; to determine a first distance between the first reference contour point and the second representative point; determine a first angle between a first ray extending from the first reference contour point to the first representative point and a second ray extending from the second representative point to the third representative point; to determine as a contour point one of: the second representative point, based on the agreement of the first representative point with the second representative point, and further based on the determination that the first distance is greater than a threshold distance or the first angle is greater than or equal to a threshold angle, one of the first representative point or the second representative point, based on the determination that the first representative point does not coincide with the second representative point, further based on the determination that the first distance is greater than the threshold distance or the first angle is greater than or equal to the threshold angle, and further based on the determination that no other representative point of the plurality of representative points exists between the first representative point and the second representative point, or one of the second representative point or a fourth representative point, based on the determination that the first representative point does not coincide with the second representative point, further based on the determination that at least one other representative point from the plurality of representative points exists between the first representative point and the second representative point, and further based on the determination that the first distance is greater than the threshold distance or the first angle is greater than or equal to the threshold angle, wherein the fourth representative point is a point that is closer to the first reference contour point among the two representative points on either side of the second representative point; and to control a vehicle based on the contour point. [2] The object recognition device of claim 1, wherein the processor is further configured to determine as the contour point one of: the second representative point, based on the determination that the first distance is greater than the threshold distance, and further based on the agreement of the first representative point with the second representative point, the first representative point, based on the determination that the first distance is greater than the threshold distance, and further based on the determination that no other representative point exists between the first representative point and the second representative point, or the fourth representative point, based on the determination that the first distance is greater than the threshold distance, and further based on the determination that the at least one other representative point exists between the first representative point and the second representative point. [3] The object recognition device of claim 1, wherein the processor is further configured to determine one of: the second representative point as the contour point on the basis that the second representative point is one of the two outermost endpoints, and further on the basis of determining that the first distance is less than or equal to the threshold distance and the first angle is less than the threshold angle, or the contour point that is not below the first representative point, the second representative point, and the fourth representative point, based on the second representative point not being one of the two extreme endpoints. [4] The object recognition device of claim 1, wherein the processor is further configured to determine the second representative point as the contour point based on determining that the first distance is less than or equal to the threshold distance and the first angle is greater than or equal to the threshold angle. [5] The object recognition device according to claim 1, wherein the processor is further configured to determine as a contour point one of: the second representative point, based on the determination that the second representative point is one of the two outermost endpoints, further based on the determination that the first distance is greater than the threshold distance, and further based on the agreement of the first representative point with the second representative point, the first representative point, based on the determination that the second representative point is one of the two outermost endpoints, further based on the determination that the first distance is greater than the threshold distance, further based on the determination that the first representative point does not coincide with the second representative point, and further based on the determination that no other representative point from the plurality of representative points exists between the first representative point and the second representative point, or the fourth representative point, based on the determination that the second representative point is one of the two outermost endpoints, further based on the determination that the first distance is greater than the threshold distance, further based on the determination that the first representative point does not coincide with the second representative point, and further based on the determination that at least one other representative point from the plurality of representative points exists between the first representative point and the second representative point. [6] The object recognition device of claim 1, wherein the processor is further configured to: to determine the second representative point as a contour point based on determining that the second representative point is one of the two outermost endpoints and further based on determining that the first distance is less than or equal to the threshold distance. [7] The object recognition device of claim 1, wherein the processor is further configured to: to determine a fifth representative point, which is a point further away from the first reference contour point than two representative points on either side of the third representative point, based on the determination that the contour point does not exist between the first representative point, the second representative point, and the fourth representative point; and to determine the second representative point or the third representative point as a contour point based on determining that a second distance between the first reference contour point and the third representative point is greater than the threshold distance or a second angle between the first ray and a third ray extending from the third representative point to the fifth representative point is greater than or equal to the threshold angle. [8] The object recognition device of claim 1, wherein the processor is further configured to: to determine a second reference contour point and a third reference contour point, wherein the third reference contour point is one of two contour points on either side of the second reference contour point; to determine a sixth representative point which is not located between the second reference contour point and the third reference contour point; to determine a seventh representative point, which is a point further away from the second reference contour point than two representative points on either side of the sixth representative point; to determine a second distance between the third reference contour point and the seventh representative point; to determine a second angle between a third ray extending from the second reference contour point to the third reference contour point and a fourth ray extending from the sixth representative point to the seventh representative point; and to determine as contour point one of: the seventh representative point, based on the determination that the second distance is greater than the threshold distance or the second angle is greater than or equal to the threshold angle, and further based on the third reference contour point matching the sixth representative point; or one of the sixth representative point or the seventh representative point, based on determining that the second distance is greater than the threshold distance or the second angle is greater than or equal to the threshold angle, and further based on the third reference contour point not coinciding with the sixth representative point. [9] An object recognition device according to claim 8, wherein the processor is further configured to use as a contour point a point determined by: the seventh representative point based on the second distance which is greater than the threshold distance, and further based on the third reference contour point which coincides with the sixth representative point, or the sixth representative point, based on the second distance being greater than the threshold distance, and further based on the third reference contour point not coinciding with the sixth representative point. [10] The object recognition device of claim 8, wherein the processor is further configured to: to determine whether the seventh representative point is one of the two extreme endpoints based on determining that the second distance is less than or equal to the threshold distance and the second angle is less than the threshold angle; and to determine one of: the seventh representative point as the contour point, where the seventh representative point is one of the two extreme endpoints, or the contour point that does not lie between the sixth representative point and the seventh representative point, based on the seventh representative point not being one of the two extreme endpoints. [11] The object recognition device according to claim 8, wherein the processor is further configured to determine the seventh representative point as the contour point based on determining that the second distance is less than or equal to the threshold distance and the second angle is greater than or equal to the threshold angle. [12] The object recognition device of claim 8, wherein the processor is further configured to: to determine an eighth representative point, which is a point further away from the second reference contour point than two representative points on either side of the seventh representative point, based on the determination that the contour point does not exist between the sixth representative point and the seventh representative point; and to determine the seventh representative point or the eighth representative point as a contour point based on determining that a third distance between the third reference contour point and the eighth representative point is greater than the threshold distance or that a third angle between the third ray and a fifth ray extending from the seventh representative point to the eighth representative point is greater than or equal to the threshold angle. [13] Object recognition methods, comprising: Extracting a plurality of points obtained via a sensor and representing an object from a plurality of representative points comprising at least a portion of surface points of the object; Determining a first reference contour point which is one of the two outermost endpoints of the plurality of representative points; Determining a first representative point adjacent to the first reference contour point from the plurality of representative points; Determining a second representative point from the plurality of representative points that is different from the first reference contour point; Determining a third representative point which is a point further away from the first reference contour point than two representative points on either side of the second representative point; Determining a first distance between the first reference contour point and the second representative point; Determining a first angle between a first ray extending from the first reference contour point to the first representative point and a second ray extending from the second representative point to the third representative point; Determine, as contour point, one of: the second representative point, based on the agreement of the first representative point with the second representative point, and further based on the determination that the first distance is greater than a threshold distance or the first angle is greater than or equal to a threshold angle, one of the first representative point or the second representative point, based on the first representative point not coinciding with the second representative point, further based on the determination that the first distance is greater than the threshold distance or the first angle is greater than or equal to the threshold angle, and further based on the determination that no other representative point among the plurality of representative points exists between the first representative point and the second representative point, or one of the second representative point or a fourth representative point, based on the first representative point not coinciding with the second representative point, further based on the determination that at least one other representative point among the plurality of representative points exists between the first representative point and the second representative point, and further based on the determination that the first distance is greater than the threshold distance or the first angle is greater than or equal to the threshold angle, wherein the fourth representative point is a point that is closer to the first reference contour point among the two representative points on either side of the second representative point; and Controlling a vehicle based on the contour point. [14] An object recognition method according to claim 13, further comprising determining, as a contour point, one of: the second representative point, based on the finding that the first distance is greater than the threshold distance, and further based on the agreement of the first representative point with the second representative point, the first representative point, based on the determination that the first distance is greater than the threshold distance, further based on the determination that no other representative point exists between the first representative point and the second representative point, or the fourth representative point, based on the determination that the first distance is greater than the threshold distance, and further based on the determination that the at least one other representative point exists between the first representative point and the second representative point. [15] The object recognition method of claim 13, further comprising determining one of: the second representative point as the contour point based on the second representative point being one of the two outermost endpoints, and further based on the determination that the first distance is less than or equal to the threshold distance and the first angle is less than the threshold angle; or the contour point that is not below the first representative point, the second representative point, and the fourth representative point, based on the second representative point not being one of the two extreme endpoints. [16] The object recognition method according to claim 13, further comprising: Determining the second representative point as a contour point based on determining that the first distance is less than or equal to the threshold distance and the first angle is greater than or equal to the threshold angle. [17] An object recognition method according to claim 13, further comprising determining, as a contour point, one of: the second representative point, based on the determination that the second representative point is one of the two outermost endpoints, further based on the determination that the first distance is greater than the threshold distance, and further based on the agreement of the first representative point with the second representative point, the first representative point, based on the determination that the second representative point is one of the two outermost endpoints, further based on the determination that the first distance is greater than the threshold distance, further based on the determination that the first representative point does not coincide with the second representative point, and further based on the determination that no other representative point from the plurality of representative points exists between the first representative point and the second representative point, or the fourth representative point, based on the determination that the second representative point is one of the two outermost endpoints, further based on the determination that the first distance is greater than the threshold distance, further based on the determination that the first representative point does not coincide with the second representative point, and further based on the determination that at least one other representative point from the plurality of representative points exists between the first representative point and the second representative point. [18] The object recognition method according to claim 13, further comprising: Determining the second representative point as a contour point based on determining that the second representative point is one of the two outermost endpoints and further based on determining that the first distance is less than or equal to the threshold distance. [19] The object recognition method according to claim 13, further comprising: Determining a fifth representative point, which is a point further away from the first reference contour point from two representative points on both sides of the third representative point, based on the determination that the contour point does not exist between the first representative point, the second representative point, and the fourth representative point; and Determining the second representative point or the third representative point as a contour point based on determining that a second distance between the first reference contour point and the third representative point is greater than the threshold distance or a second angle between the first ray and a third ray extending from the third representative point to the fifth representative point is greater than or equal to the threshold angle. [20] The object recognition method according to claim 13, further comprising: Determining a second reference contour point and a third reference contour point, wherein the third reference contour point is one of two contour points on either side of the second reference contour point; Determining a sixth representative point that is not located between the second reference contour point and the third reference contour point; Determining a seventh representative point which is a point further away from the second reference contour point than two representative points on both sides of the sixth representative point; Determining a second distance between the third reference contour point and the seventh representative point; Determining a second angle between a third ray extending from the second reference contour point to the third reference contour point and a fourth ray extending from the sixth representative point to the seventh representative point; and Determine, as contour point, one of: the seventh representative point, based on the determination that the second distance is greater than the threshold distance or the second angle is greater than or equal to the threshold angle, and further based on the third reference contour point coinciding with the sixth representative point; or one of the sixth representative point or the seventh representative point, based on determining that the second distance is greater than the threshold distance or the second angle is greater than or equal to the threshold angle, and further based on the third reference contour point not coinciding with the sixth representative point.

Images