Apparatus and method for inspecting angular resolution of lidar device
Patent Information
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2026-02-05
- Publication Date
- 2026-08-13
Smart Images
Figure KR2026002153_13082026_PF_FP_ABST
Abstract
Description
Inspection device and method for each resolution of a LiDAR device
[0001] The present invention relates to a lidar device. More specifically, the present invention relates to performance testing of a lidar device, for example, testing of each resolution of a lidar device.
[0002] LiDAR devices are important sensors for detecting the shape and distance of objects and are utilized in various industrial fields, such as autonomous vehicles, drones, and robots. LiDAR devices scan the surrounding environment using light and generate a point cloud, which is a set of data points in 3D space, based on the collected data. At this time, the quality of the point cloud depends significantly on the performance of the LiDAR device. One such performance characteristic of a LiDAR device is Angular Resolution.
[0003] Angular resolution refers to the minimum angular difference that a LiDAR device can detect within a specific angle, and LiDAR devices with high angular resolution can detect objects in detail. Therefore, the angular resolution of a LiDAR device can serve as an important indicator for evaluating its performance.
[0004] Each resolution of a LiDAR device can be measured based on data points derived from light reflected back from the edges of a target object being scanned. In this case, it is common for the operator to visually determine whether the data point is located at the edge of the target object. However, since the manner in which a data point exists near the edge of a target object can be implemented in various ways, there is a problem in that the criteria for determining under what circumstances a data point is located at the edge of the target object are unclear.
[0005] Furthermore, the process of measuring the angular resolution of a LiDAR device typically involves rotating the device using a rotating mechanism. In this case, if the angular resolution of the rotating mechanism itself is not sufficiently small, there is a possibility that data points may not be generated at the edges of the target object.
[0006] In addition, if distortion occurs in the point cloud itself generated by the LiDAR device, there is also a problem in that the measured resolution may vary depending on which of the multiple data points is used as the reference for measuring each resolution.
[0007] Therefore, there is a need for a method to measure each resolution using more objective and consistent criteria, rather than relying on the subjective judgment of the measurer.
[0008] The objective of the present invention is to provide a device and method for inspecting each resolution of a LiDAR device that does not rely on the subjective judgment of the measurer.
[0009] In addition, the objective of the present invention is to provide a device and method for inspecting each resolution of a lidar device that is not affected by the each resolution of the rotating device itself that rotates the lidar device.
[0010] In addition, the objective of the present invention is to provide a device and method for inspecting each resolution of a LiDAR device that is not affected by distortion of the point cloud itself produced by the LiDAR device.
[0011] To solve the problem of the present invention, in a method for inspecting each resolution of a LiDAR (Light Detection And Ranging) device, the method comprises the steps of: preparing a LiDAR device to be inspected and at least one target; rotating the LiDAR device to be inspected in n steps in an azimuth direction or an elevation direction, and recording m frames for the target through the LiDAR device to be inspected at each rotation angle of the n steps; and selecting at least one data point from a point cloud calculated by the LiDAR device to be inspected through the recorded frames. The present invention provides a method for inspecting each resolution of a lidar device, comprising the step of determining each resolution of the lidar device to be inspected based on depth information included in each of the selected data points; wherein each of the data points corresponds to each of a plurality of super-pixels constituting a photodetector of the lidar device to be inspected, and the depth information included in the data points is calculated by the corresponding super-pixels based on a mean value-based method over m times, which is the number of frames, at each of the n rotation angles.
[0012] According to one embodiment of the present invention, the super pixel includes a plurality of individual pixels, and the depth information calculated by the super pixel may be calculated by processing the information calculated by each of the individual pixels in an average value-based manner.
[0013] According to one embodiment of the present invention, the step of determining each resolution of the inspection target lidar device based on depth information included in the selected data point comprises the step of calculating a partial average depth for the data point based on the depth information and generating a suitable depth graph for the data point; and each resolution may be determined based on the generated suitable depth graph and the calculated partial average depth.
[0014] According to one embodiment of the present invention, the method further comprises the step of positioning the inspection target lidar device and the target; wherein the target includes a main target and a sub-target, the main target is positioned such that its edge faces the optical center of the inspection target lidar device in a frontal manner, and the sub-target may be positioned parallel to the main target at the rear of the main target in a direction away from the inspection target lidar device.
[0015] According to one embodiment of the present invention, the inspection target lidar device can rotate by an angle predetermined in each of the n steps.
[0016] According to one embodiment of the present invention, the selected data points may be sequentially adjacent in a horizontal or vertical direction within the calculated point cloud.
[0017] According to one embodiment of the present invention, the step of calculating a partial average depth for the data point based on the depth information and generating a suitable depth graph for the data point includes the step of calculating an average depth for the data point as an average of the depth information at each of the n rotation angles; wherein the partial average depth is calculated as an average of at least some of the calculated average depths, and the suitable depth graph may be generated by applying curve fitting to a graph plotting the calculated average depths against the rotation angles.
[0018] According to one embodiment of the present invention, the step of determining each resolution of the lidar device to be inspected based on depth information included in the selected data points may further include: identifying a pair of data points among the selected data points in which the partial average depth values are mutually identical; calculating, for each of the identified data points, the rotation angle corresponding to the partial average depth on the suitable depth graph for each of the data points; identifying the number of data points sequentially adjacent in the horizontal or vertical direction between the identified data points; and determining each resolution of the lidar device to be inspected based on the calculated rotation angle and the identified number.
[0019] To solve the problem of the present invention, a resolution inspection device for a LiDAR (Light Detection And Ranging) device is provided, comprising: a support member for a device to be inspected for supporting the device to be inspected; a rotation adjustment member for rotating the support member for the device to be inspected and the device to be inspected; and a processor member for controlling the operation of the device to be inspected and the rotation adjustment member and determining each resolution of the device to be inspected based on information output by the device to be inspected and the rotation adjustment member.
[0020] According to one embodiment of the present invention, the processor unit performs the steps of: rotating the inspection target lidar device in n steps in an azimuth direction or an elevation direction, and recording m frames for a target through the inspection target lidar device at each of the n steps of rotation angles; selecting at least one data point from a point cloud calculated by the inspection target lidar device through the recorded frames; and determining each resolution of the inspection target lidar device based on depth information included in each of the selected data points; wherein each of the data points corresponds to each of a plurality of super-pixels constituting a photodetector of the inspection target lidar device, and the depth information included in the data points may be calculated by the corresponding super-pixels based on a mean value-based method over m times, which is the number of frames, at each of the n rotation angles.
[0021] According to one embodiment of the present invention, the super pixel includes a plurality of individual pixels, and the depth information calculated by the super pixel may be calculated by processing the information calculated by each of the individual pixels in an average value-based manner.
[0022] According to one embodiment of the present invention, the step of determining each resolution of the inspection target lidar device based on depth information included in the selected data point comprises the step of calculating a partial average depth for the data point based on the depth information and generating a suitable depth graph for the data point; and each resolution may be determined based on the generated suitable depth graph and the calculated partial average depth.
[0023] According to one embodiment of the present invention, the target includes a main target and a sub-target, wherein the main target is positioned so that the optical center and edge of the lidar device to be inspected face each other directly, and the sub-target may be positioned parallel to the main target behind the main target in a direction away from the lidar device to be inspected.
[0024] According to one embodiment of the present invention, the inspection target lidar device can rotate by an angle predetermined in each of the n steps.
[0025] According to one embodiment of the present invention, the selected data points may be sequentially adjacent in a horizontal or vertical direction within the calculated point cloud.
[0026] According to one embodiment of the present invention, the step of calculating a partial average depth for the data point based on the depth information and generating a suitable depth graph for the data point includes the step of calculating an average depth for the data point as an average of the depth information at each of the n rotation angles; wherein the partial average depth is calculated as an average of at least some of the calculated average depths, and the suitable depth graph may be generated by applying curve fitting to a graph plotting the calculated average depths against the rotation angles.
[0027] According to one embodiment of the present invention, the step of determining each resolution of the lidar device to be inspected based on depth information included in the selected data points may further include: identifying a pair of data points among the selected data points in which the partial average depth values are mutually identical; calculating, for each of the identified data points, the rotation angle corresponding to the partial average depth on the suitable depth graph for each of the data points; identifying the number of data points sequentially adjacent in the horizontal or vertical direction between the identified data points; and determining each resolution of the lidar device to be inspected based on the calculated rotation angle and the identified number.
[0028] The device and method for inspecting each resolution of a lidar device according to the present invention can determine each resolution of a lidar device based on depth information, etc., for a plurality of data points calculated over a plurality of frames by rotating the lidar device in a plurality of steps.
[0029] FIG. 1 is a schematic diagram illustrating the operation of a light sensing and distance measuring device or a lidar device to which the present invention is applied.
[0030] FIG. 2 is a drawing showing each resolution inspection device of a lidar device according to one embodiment of the present invention.
[0031] FIG. 3 is a schematic diagram illustrating the arrangement between each resolution inspection device and at least one target of a lidar device according to one embodiment of the present invention.
[0032] FIG. 4 is a conceptual diagram illustrating the relationship between the arrival point of light emitted from a device to be inspected according to the present invention on a target and the data point calculated based on the light reflected from the target.
[0033] FIG. 5 illustrates a plurality of data points calculated by the inspection target device from the light reaching the edge of the target and the light reflected from the target at a rotation angle in which the light emitted from the inspection target device according to the present invention reaches the edge of the target.
[0034] FIG. 6 is a diagram showing the average depth calculated based on a plurality of data points calculated by the inspection target device in FIG. 5, displayed on a coordinate plane.
[0035] FIG. 7 is a diagram illustrating a method for calculating depth information from data points according to the present invention.
[0036] FIG. 8 is a conceptual diagram illustrating the process of generating a suitable depth graph and calculating a partial average depth according to the present invention.
[0037] Figure 9 shows multiple depth-of-fit graphs for the data points shown in Figure 8.
[0038] FIG. 10 is a flowchart illustrating a method for inspecting each resolution of a lidar device according to one embodiment of the present invention.
[0039] Hereinafter, a method and system for inspecting each resolution of a LiDAR device according to an embodiment of the present invention will be described in detail with reference to the attached drawings. However, it will be readily apparent to those skilled in the art that the attached drawings are provided merely to facilitate the disclosure of the contents of the present invention, and that the scope of the present invention is not limited to the scope of the attached drawings.
[0040] Introduction
[0041] FIG. 1 is a schematic diagram illustrating the operation of a light sensing and distance measuring device or a lidar device to which the present invention is applied.
[0042] Referring to FIG. 1, a light detection and distance measuring device (100, hereinafter also referred to as a LiDAR device) to which the present invention is applied may include a light emitter (110) for emitting light, a light detector (120) for detecting reflected light that is reflected back from an object (200) after the emitted light, and an optical device (130) provided in the light path radiated and received from the light emitter (110) and the light detector (120). Here, the light emitter (110) may be a diode and a laser light source. The LiDAR device can calculate the range or property of an object (200) using the reflected light that is reflected back from the object (200).
[0043] In this specification, the device subject to inspection may be interchangeably referred to as a lidar device.
[0044] A point cloud refers to a set of data points in 3D space. A set of data points calculated by a LiDAR device to which the present invention is applied can also be called a point cloud. Since the distances between data points constituting a point cloud are generally non-uniform, it is desirable to specifically encode all three coordinates (orthogonal coordinates or spherical coordinates) for each point.
[0045] According to the present invention, the device subject to testing is referred to as the Device Under Test (DUT).
[0046] Among results where the measurement value is positive, the case where the measurement value is accurate—that is, where both the measurement value and the result are positive—is called a True Positive (TP). Among results where the measurement value is positive, the case where the measurement value is inaccurate—that is, where the measurement value is positive but the result is negative—is called a False Positive (FP).
[0047] The probability of valid points in a single measurement and / or multiple accumulated measurements for a single target is called the Probability of Detection (PoD) or True Positive Rate. The probability of detection may depend on background noise, the reflectivity of the target, the tolerance of the range, and other attributes. The PoD can be calculated by the following [Equation 1], where True (TP) represents the scan points that hit the target in total at distance (actual) ±Δ. The probability of detection is calculated as the ratio of the number of valid points to the number of theoretical points.
[0048]
[0049]
[0050] In a point cloud of a LiDAR device, the angle formed by the connection between two adjacent detection points (hereinafter also referred to as data points) and the 3D coordinate origin of the point cloud in terms of azimuth angle and elevation angle is called angular resolution. The angular resolution of a LiDAR device can be divided into azimuth resolution and elevation resolution.
[0051] In the point cloud of a LiDAR device, the angle between two outermost effective points where the PoD exceeds 50% (Lambertian target reflectance 50%) is called the field of view (FOV). The FOV includes a horizontal FOV range and a vertical FOV range.
[0052] Capturing the entire FOV (horizontal / vertical) is called a frame.
[0053]
[0054] Overall configuration of each resolution inspection device of the LiDAR device
[0055] FIG. 2 is a drawing showing each resolution inspection device of a lidar device according to one embodiment of the present invention.
[0056] As shown in FIG. 2, the inspection device (10) may include a support part (11) for the device to be inspected, a rotation adjustment part (12), a processor part (13), and a memory part (14).
[0057] The support member (11) of the device to be inspected is intended to support the device to be inspected. More specifically, the support member (11) of the device to be inspected can be combined with the device to be inspected to support the device to be inspected, and can rotate together with the device to be inspected by the rotation adjustment member (12) as described below.
[0058] The rotation adjustment unit (12) is for rotating the device to be inspected and the device to be inspected support unit (11). The rotation adjustment unit (12) can be connected to the device to be inspected support unit (11) and can rotate the device to be inspected support unit (11) and / or the device to be inspected.
[0059] The processor unit (13) can control the operation of the rotation adjustment unit (12). Additionally, the processor unit (13) can control the operation of the inspection target device connected to the inspection device (10). As described below, the processor unit (13) can process information output by the inspection target device connected to the inspection device (10) and the rotation adjustment unit (12), and can determine each resolution of the inspection target device by processing the above information.
[0060] The memory unit (14) can store information output by the device to be inspected connected to the inspection device (10). The memory unit (14) can store information generated by the processor unit (14) based on the information output by the device to be inspected. The memory unit (14) can have a program recorded therein to enable the processor unit (13) to perform an operation to control the operation of the rotation adjustment unit (12) and to process the information output by the device to be inspected to determine each resolution of the device to be inspected.
[0061] An inspection device (10) according to one embodiment of the present invention can rotate an inspection target device and generate multiple data points across multiple frames, and can acquire and / or calculate depth data, average depth, fitted depth graph, and partial average depth for the multiple data points. The depth data, average depth, fitted depth graph, and partial average depth will be described later in relation to FIGS. 4 and FIGS. 5.
[0062] FIG. 3 is a schematic diagram illustrating the arrangement between each resolution inspection device and at least one target of a lidar device (100) according to one embodiment of the present invention. FIG. 3 (a), (b), and (c) are a perspective view, a front view, and a plan view illustrating the arrangement between the inspection device and the target.
[0063] According to the present invention, each resolution inspection device (10) of the lidar device (100) is for determining each resolution of the device to be inspected. At this time, the device to be inspected may be the lidar device (100).
[0064] As illustrated in FIG. 3, the inspection device (10) may be positioned with at least one target. The at least one target may include a main target (20) and a sub-target (21). The main target (20) and the sub-target (21) may each be positioned at a certain distance from the inspection device (10).
[0065] When the inspection device (10) performs scanning in a horizontal direction, the main target (20) and the sub-target (21) may be spaced apart from the inspection device (10) by different predetermined distances and arranged side by side in a horizontal direction.
[0066] When the inspection device (10) performs scanning in a vertical direction, the main target (20) and the sub-target (21) may be spaced apart from each other by a different predetermined distance from the inspection device (10) and arranged side by side in a vertical direction.
[0067] Although not illustrated in FIG. 3, as described below, the inspection device (10) may be connected to the device to be inspected. For example, the inspection device (10) may be connected to a lidar device (100).
[0068] As illustrated in FIG. 3(a), the distance (d1) between the inspection device (10) and the main target (20) may be smaller than the distance (d2) between the inspection device (10) and the sub-target (21). Here, the distance between the inspection device (10) and the target may refer to the straight-line distance between the inspection target device and the target connected to the inspection device (10). For example, the distance between the inspection device (10) and the target may refer to the length of a line segment perpendicular to the plane of the target, in the line segment connecting the optical origin of the LiDAR device (100) connected to the inspection device (10) and any point located on the target.
[0069] As illustrated in FIGS. 3(b) and (c), the main target (20) may be positioned so that the optical center and edge of the inspection target device connected to the inspection device (10) face each other directly. Additionally, the sub-target (21) may be positioned behind the main target (20) in a direction away from the inspection device (10) and parallel to the main target (20). The main target (20) and the sub-target (21) may be positioned so that at least a portion of the sub-target (21) is obscured from view when viewed from the front of the inspection device (10). The main target (20) and the sub-target (21) may be positioned so that when viewed from the front of the inspection device (10), the edge of the main target (20) appears to overlap with the sub-target (21) positioned behind it. Alternatively, the main target (20) and the sub-target (21) may be arranged so that when viewed from the front of the inspection device (10), the respective edges of the main target (20) and the sub-target (21) appear to overlap. In the former case, when viewed from the front of the inspection device (10), the edge of the main target (20) may form a boundary surface (a-a'). In the latter case, when viewed from the front of the inspection device (10), the edges of the main target (20) and the sub-target (21) may form a boundary surface (a-a'). In the following description, the former case is described as an example, but the concept of the present invention may also be applied to the latter case.
[0070] The main target (20) and the sub-target (21) may be positioned so that a portion of the edge of the main target (20) overlaps on the sub-target (21) positioned behind it. The main target (20) and the sub-target (21) may be positioned to clarify the reference for depth information measured according to the location of each calculated data point. Accordingly, the main target (20) and the sub-target (21) may be positioned so that a portion of the main target (20) and the sub-target (21) overlap.
[0071] According to the present invention, the target may include only the main target (20) and not include the sub-target (21). In this case, the background located behind the main target (20) may take the place of the sub-target (21).
[0072] Hereinafter, with reference to FIGS. 4 to 9, each resolution inspection device (10) of a lidar device (100) according to one embodiment of the present invention will be described in detail.
[0073]
[0074] Inspection environment
[0075] The inspection device (10) and method may be conditioned on the target reflectivity being included in a specific numerical range. For example, the target reflectivity is preferably 10% or more and 90% or less, but is not limited thereto.
[0076] The inspection device (10) and method may be conditioned on the target dimension being included within a specific numerical range. Specifically, the target dimension may be a size sufficient to measure each resolution according to the specifications of the device to be inspected.
[0077] The inspection device (10) and method may be subject to the condition that the temperature of the space where the inspection is performed falls within a specific numerical range. Specifically, it is preferable that the temperature of the space where the inspection is performed be room temperature. For example, it is preferable that the temperature of the space where the inspection is performed be 22.5 degrees or higher and 23.5 degrees or lower, but is not limited thereto.
[0078]
[0079] Movement of arrival points and data points due to rotation
[0080] FIG. 4 is a conceptual diagram illustrating the relationship between the arrival point of light emitted from a device to be inspected according to the present invention on a target and the data points calculated based on the light reflected from the target. Specifically, FIG. 4(a) illustrates the movement of the arrival point of light emitted from the device to be inspected on the target. FIG. 4(b) illustrates the movement of multiple data points calculated in overlap based on the emitted light. Meanwhile, FIG. 4(a) and FIG. 4(b) respectively illustrate the light arrival point and data points at different rotation angles. That is, in FIG. 4(a) and FIG. 4(b), the light arrival point at the first rotation angle and n data points are superimposed and illustrated at the top, and the light arrival point at the second rotation angle and n data points are superimposed and illustrated at the bottom.
[0081] In FIG. 4, only a total of 25 lights having a square arrangement of 5 in the transverse direction and 5 in the longitudinal direction are illustrated as an example among the lights emitted by the device under inspection. In the following description, only the total of 25 lights exemplified in FIG. 4 are used as examples, but the concept of the present invention is not to be interpreted as being limited by the number or arrangement of emitted lights.
[0082] As illustrated in FIG. 4, for each of the total 25 lights emitted from the inspection target device according to the present invention, n data points are calculated from a total of n lights reflected from the target over n frames. In FIG. 4(b), the n data points calculated from the n lights reflected from the target are superimposed.
[0083] The processor unit (13) can control the operation of the rotation adjustment unit (12) to rotate the inspection target device connected to the inspection device (10) in the azimuth direction or the elevation direction. For example, if the inspection target device is a LiDAR device (100), the processor unit (13) can rotate the LiDAR device (100) in the azimuth direction or the elevation direction.
[0084] According to the present invention, the processor unit (13) can control the rotation adjustment unit (12) so that the rotation adjustment unit (12) repeatedly rotates the device to be inspected. The angle at which the device to be inspected repeatedly rotates may be a preset angle.
[0085] According to the present invention, as the processor unit (13) rotates the lidar device (100) in an azimuth direction or an elevation direction, the point of arrival (31) on the target of the light emitted from the lidar device (100) may move in a horizontal direction or a vertical direction. The point of arrival (31) may refer to a point on the target where the light emitted from the inspection target device actually reaches.
[0086] For example, as illustrated in FIG. 4(a), when the processor unit (13) rotates the lidar device (100) in an azimuth direction, the arrival point (31) may move in a horizontal direction. The arrival point (31) may move to penetrate the boundary surface (a-a') formed by the edge of the main target (20) in a horizontal direction. When the lidar device (100) emits a plurality of lights, there may be a plurality of arrival points (31) of the emitted light. For example, as illustrated in FIG. 4(a), the arrival point (31) may include a first arrival point (31a), a second arrival point (31b), a third arrival point (31c), and a fourth arrival point (31d).
[0087] According to the present invention, at least some of the reaching points (31) may move to penetrate the boundary surface (a-a') formed by the edge of the main target (20) in a horizontal or vertical direction as the device to be inspected rotates. The depth information (50) of the data points (30) calculated by the device to be inspected based on light reflected from the main target (20) or the sub-target (21) upon reaching the corresponding reaching points (31) may differ from one another between a plurality of data points (30) and between a plurality of frames.
[0088] For example, as illustrated in FIG. 4(a), the first arrival point (31a), the second arrival point (31b), the third arrival point (31c), and the fourth arrival point (31d) can move to penetrate the boundary surface (a-a') formed by the edge of the main target (20) in a horizontal direction as the inspection target device rotates in an azimuth direction. At this time, the depth information (50) of the data points (30) calculated by the LiDAR device (100) connected to the inspection device (10) based on the light reflected from the main target (20) or sub-target (21) by the light reaching the first arrival point (31a), the second arrival point (31b), the third arrival point (31c), and the fourth arrival point (31d) may be different from each other.
[0089] According to the present invention, a LiDAR device (100) connected to an inspection device (10) can detect light that reaches a point (31) on a target and is reflected back, and can calculate a data point (30) based thereon. According to the present invention, n data points (30) can be calculated over a total of n frames for a set rotation angle. FIG. 4(b) shows these n data points (30) superimposed. Each point (31) can correspond to each data point (30). For example, referring to FIG. 4, the first point (31a), second point (31b), third point (31c), and fourth point (31d) shown in FIG. 4(a) can correspond to the first data point (30a), second data point (30b), third data point (30c), and fourth data point (30d), respectively.
[0090] According to the present invention, the processor unit (13) can control the lidar device (100) connected to the inspection device (10) so that when the inspection target device rotates in stages while changing the azimuth angle or elevation angle in the azimuth direction or elevation direction, the lidar device (100) calculates data points (30) across multiple, i.e., n frames, at each stage. The number of frames n recorded by the lidar device (100) at each stage may be of a size sufficient to measure each resolution according to the specifications of the lidar device (100). For example, the number of multiple frames n may be 10, but is not limited thereto.
[0091] For example, as illustrated in FIG. 4(b), the processor unit (13) can control the lidar device (100) so that the lidar device (100) connected to the inspection device (10) records n frames at each rotational step.
[0092] As illustrated in FIG. 4(b), multiple data points (30) calculated over multiple frames can be represented by overlapping each other. Each location within the point cloud of the multiple data points (30) calculated by the lidar device (100) over multiple frames may differ between the multiple data points (30) and between the multiple frames.
[0093]
[0094] Depth information and average depth
[0095] FIG. 5 illustrates a plurality of data points calculated by a device to be inspected from light reaching the edge of a target and light reflected from a target at a rotation angle in which light emitted from the device to be inspected according to the present invention reaches the edge of a target. FIG. 5(a) illustrates a case where the proportion of data points located on the main target among the calculated plurality of data points is high, and FIG. 5(b) illustrates a case where the proportion of data points located on the sub-target among the calculated plurality of data points is high.
[0096] As shown in FIG. 5, the point of arrival (31) of the light reaching the edge of the target at the rotation angle at which the light emitted from the device to be inspected reaches the edge of the target can be located at the boundary surface (a-a') formed by the edge of the main target (20).
[0097] As illustrated in FIG. 5, at the corresponding rotation angle, the lidar device (100) can record a plurality of frames. In some of the plurality of frames, the lidar device (100) can calculate a data point (30) located in an area corresponding to the main target (20) based on light that reaches and is reflected from a point (31) located at the boundary surface (a-a'). Additionally, in some of the remainder of the plurality of frames, the lidar device (100) can calculate a data point (30) located in an area corresponding to the sub-target (21) based on light that reaches and is reflected from a point (31) located at the boundary surface (a-a').
[0098] For example, as illustrated in FIG. 5(a), the ratio of frames in which the lidar device (100) produces data points (30) located in an area corresponding to the main target (20) may be higher.
[0099] Also, for example, as shown in FIG. 5(b), the ratio of frames in which the lidar device (100) calculates data points (30) located in an area corresponding to the sub-target (21) may be higher.
[0100] According to the present invention, the processor unit (13) can control the lidar device (100) connected to the inspection device (10) so that the lidar device (100) calculates data points (30) over a plurality of frames. Each data point (30) may have depth information (50).
[0101] For example, when a LiDAR device (100) calculates a data point (30) located in an area corresponding to the main target (20) shown in FIG. 5, the data point (30) may include depth information (50) having a predetermined value. The predetermined value may be close to the distance between the LiDAR device (100) connected to the inspection device (10) and the main target (20).
[0102] Additionally, for example, when the lidar device (100) calculates a data point (30) located in an area corresponding to the sub-target (21) shown in FIG. 5, the lidar device (100) may include depth information (50) having a predetermined value for the data point (30). The predetermined value may be close to the distance between the lidar device (100) connected to the inspection device (10) and the sub-target (21).
[0103] According to the present invention, the processor unit (13) can receive depth information (50) from a LiDAR device (100) connected to an inspection device (10). The processor unit (13) can calculate an average depth (60) based on the depth information (50). The average depth (60) refers to the average value of the depth information (50) of the data points (30) calculated by the LiDAR device (100) over a plurality of frames based on light reflected from a predetermined arrival point (31) at a specific rotation angle.
[0104] FIG. 6 is a diagram showing the average depth calculated based on multiple data points calculated by the inspection target device in FIG. 5, displayed on a coordinate plane. FIG. 6a shows the average depth when the proportion of data points located at the main target among the multiple calculated data points is high, and FIG. 6b shows the average depth when the proportion of data points located at the sub-target among the multiple calculated data points is high.
[0105] As illustrated in FIG. 6, the average depth (60) calculated by the inspection target device in relation to a predetermined reach point (31) on the target at a specific rotation angle can be varied even though the rotation angle is the same.
[0106] For example, referring to point (P1) illustrated in FIG. 6(a), if the ratio of frames in which the lidar device (100) produces data points (30) located in the area corresponding to the main target (20) is higher (see FIG. 5(a)), the average depth (60) for the corresponding rotation angle may be A. A may be a value closer to d1 than to d2.
[0107] Also, for example, referring to point (P2) illustrated in FIG. 6(b), if the ratio of frames in which the lidar device (100) produces data points (30) located in the area corresponding to the sub-target (21) is higher (see FIG. 5(b)), the average depth (60) for the corresponding rotation angle may be B. B may be a value closer to d2 than to d1.
[0108]
[0109] Depth information calculation method
[0110] FIG. 7 is a diagram illustrating a method for calculating depth information from data points according to the present invention. Specifically, FIG. 7(a) illustrates Graph 1, which shows the frequency of each depth information calculated by individual pixels included in a super pixel based on light reflected from near the boundary between the main target and the sub-target. In addition, FIG. 7(b) illustrates Graph 2, which shows the value of the final depth information calculated by the super pixel based on a mode-based method or an average-based method according to the rotation angle.
[0111] According to one embodiment of the present invention, a photodetector (120) of a lidar device (100) to be inspected may include a plurality of superpixels arranged in a predetermined pattern. Specifically, a superpixel may include a plurality of adjacent pixels and may be treated as a single unit of analysis.
[0112] According to one embodiment of the present invention, each super pixel may correspond to each data point (30) within a point cloud calculated by a lidar device (100). Specifically, each super pixel may calculate a predetermined data point (30) based on light reflected from a predetermined light arrival point (31). That is, each data point (30) may correspond to each light arrival point (31) existing within the entire field of view of the lidar device (100), and each data point (30) may correspond to each of a plurality of super pixels.
[0113] According to one embodiment of the present invention, when a super pixel receives light reflected from a predetermined arrival point (31), individual pixels included in the super pixel may receive the reflected light. Each individual pixel may generate a predetermined depth information based on the received light. A data point (30) corresponding to a super pixel may include depth information determined in a predetermined manner based on a plurality of depth information generated by individual pixels. For example, the depth information of the data point (30) corresponding to a super pixel may be determined by a mode value-based method or a mean value-based method, but is not limited thereto.
[0114] As illustrated in FIG. 7(a), when a super pixel receives light reflected from near (A) the boundary between the main target (20) and the sub-target (21), the depth information calculated by each individual pixel included in the super pixel may have different values. For example, the depth information calculated by each individual pixel may include one value selected from a group including a distance (d1) corresponding to the main target (20) and a distance (d2) corresponding to the sub-target (21). Meanwhile, the relationship between the frequency of depth information including a value corresponding to distance (d1) and the frequency of depth information including a value corresponding to distance (d2) may change.
[0115] As shown in Fig. 7(b), the value of the final depth information for the superpixel may change depending on the method of determining the depth information of the superpixel.
[0116] Specifically, in the case of a mode-based method, the depth information with the highest frequency among the depth information generated by the individual pixels of the superpixel can be determined as the final depth information of the superpixel.
[0117] Specifically, in the case of a method based on average values, the overall average of the depth information generated by the individual pixels of the superpixel can be determined as the final depth information of the superpixel.
[0118] Meanwhile, when based on an average-value-based method, compared to a method based on a mode-value-based method, the final depth information of the superpixel according to the rotation angle can appear with higher continuity.
[0119] Conversely, when based on a mode-based method, compared to a mean-based method, the final depth information of the superpixel according to the rotation angle may appear with higher discontinuity.
[0120] According to one embodiment of the present invention, in order to generate a suitable depth graph, it may be preferable for the final depth information of the superpixel to follow an average value-based method, but is not limited thereto.
[0121]
[0122] Fit depth graph and partial average depth
[0123] FIG. 8 is a conceptual diagram illustrating the process of generating a suitable depth graph and calculating a partial average depth according to the present invention. FIG. 8 (a), (b), and (c) show multiple data points calculated over multiple frames based on light emitted by a device under inspection, superimposed and plotted by rotation angle. FIG. 8 (d) is an example of a graph plotting the average depth calculated for the multiple data points shown in FIG. 8 (a), (b), and (c) against the rotation angle. FIG. 8 (e) is an example of a suitable depth graph generated by fitting the graph shown in FIG. 8 (d).
[0124] As illustrated in FIG. 8 (a), (b) and (c), the processor unit (13) can output an average depth (60) for data points (30) that the lidar device (100) calculates over a plurality of frames, according to the rotation angle. The processor unit (13) can output an average depth (60) for data points (30) whose position changes from the area of the main target (20) to the area of the sub-target (21) as the lidar device (100) rotates. For example, the processor unit (13) can calculate an average depth (60) for a first data point (30a) according to the rotation angle.
[0125] According to the present invention, the average depth (60) for a data point (30) can be plotted on a planar coordinate system with the degree of rotation and the n-frame average depth as axes, respectively. The n-frame average depth may refer to the average depth (60) calculated from n data points obtained by the processor unit (13) recording n frames for each rotation angle. For example, the average depth (60) for a first data point (30a) can be plotted on a coordinate system as shown in the graph (d) of FIG. 8.
[0126] According to the present invention, the processor unit (13) can generate a suitable depth graph (80) by performing curve fitting on a graph of average depth (60) for a data point (30). For example, as shown in FIG. 8 (e), the processor unit (13) can generate a suitable depth graph (80) (C1) by performing curve fitting on a graph of average depth (60) for a first data point (30a) (see FIG. 8 (d)). An algorithm for performing curve fitting can be determined by considering the expected characteristics of the LiDAR device (100) connected to the inspection device (10).
[0127] According to the present invention, the processor unit (13) may select at least a portion of the average depth (60) for a data point (30) and calculate an average for that portion. The average is referred to as a partial average depth (70). For example, the processor unit (13) may select the average depth (60) from the rotational stage where the data point (30) begins to enter the sub-target (21) from the main target (20) to the rotational stage where it completely moves to the sub-target (21), and calculate a partial average depth (70) for that portion. Alternatively, the processor unit (13) may select a section of the rotational stage where the average depth (60) is calculated as a value between a value corresponding to the main target (20) (e.g., d1 shown in FIG. 6) and a value corresponding to the sub-target (21) (e.g., d2 shown in FIG. 6), and calculate a partial average depth (70) for that portion. However, it is not limited thereto.
[0128] The rotational step in which the data point (30) begins to move from the main target (20) to the sub-target (21) may mean a rotational step in which the data point (30) is calculated to be located at the sub-target (21) in at least some of the multiple frames recorded by the lidar device (100). Additionally, the rotational step in which the data point (30) moves completely to the sub-target (21) may mean a rotational step in which the data point (30) is calculated to be located at the sub-target (21) in all of the multiple frames recorded by the lidar device (100).
[0129] For example, as illustrated in FIG. 8 (d), the processor unit (13) can calculate a partial average depth (70) for the first data point (30a) based on the average depth (60) corresponding to points existing between the reference line (L1) and the reference line (L2). The average depth (60) corresponding to a point located on the reference line (L1) may represent the average depth (60) in the rotational stage where the first data point (30a) begins to enter the sub-target (21) from the main target (20). Additionally, the average depth (60) corresponding to a point located on the reference line (L2) may represent the average depth (60) in the rotational stage where the first data point (30b) moves completely to the sub-target (21).
[0130] According to the present invention, the partial average depth (70) can be illustrated together with the suitable depth graph (80). Specifically, the partial average depth (70) can be represented as a point on the suitable depth graph (80) where the function value of the n-frame average depth axis is equal to the partial average depth (70). The processor unit (13) can identify the value of the rotation angle axis (i.e., rotation angle) corresponding to the point.
[0131] For example, as illustrated in FIG. 8(e), the partial average depth (70) for the first data point (30a) can be illustrated together with the fit depth graph (80) for the first data point (30a). The partial average depth (70) for the first data point (30a) can be represented as a point (M1) on the fit depth graph (80) (C1) for the first data point (30a) where the function value of the n-frame average depth axis is equal to the partial average depth (70). The processor unit (13) can identify the value of the rotation angle axis corresponding to the point (M1).
[0132] According to the present invention, the processor unit (13) may select a plurality of data points (30) to perform the above-described operation. The plurality of data points (30) may be data points (30) that are sequentially adjacent in a horizontal direction or a vertical direction. For example, referring to FIG. 8 (a), (b) and (c), the processor unit (13) may select a first data point (30a), a second data point (30b), a third data point (30c), and a fourth data point (30d) that are sequentially adjacent in a horizontal direction to determine each resolution (azimuth angle resolution) in an azimuth direction. Additionally, the processor unit (13) may perform the above-described operation for each of the first data point (30a), the second data point (30b), the third data point (30c), and the fourth data point (30d).
[0133] When determining the azimuth angle resolution of the device to be inspected, the processor unit (13) may select a plurality of data points (30) that are sequentially adjacent in the horizontal direction to perform the operation described above. Additionally, when determining the elevation angle resolution of the device to be inspected, the processor unit (13) may select a plurality of data points (30) that are sequentially adjacent in the vertical direction to perform the operation described above.
[0134] According to the present invention, the processor unit (13) can determine each resolution of the device to be inspected based on the fit depth graph (80) and the partial average depth (70).
[0135] In this case, the influence that uncertainty may occur during the process of identifying depth information (50) for data points (30) located near the edge of the target may have on the accurate determination of each resolution can be eliminated.
[0136] In addition, even if the data point (30) is not output near the edge of the target due to the angle of the angle of the rotation adjustment unit (12) itself, it is possible to determine the angle of the inspection target device.
[0137] In addition, in this case, since multiple data points (30) are selected to determine each resolution of the device to be inspected, the problem of each resolution value determined based on a single data point (30) differing from each selected data point (30) due to distortion of the point cloud itself does not occur.
[0138]
[0139] Method for determining each resolution
[0140] Figure 9 shows multiple depth-of-fit graphs for the data points shown in Figure 8.
[0141] As described above, according to the present invention, the processor unit (13) can perform the above-described operation on a plurality of data points (30) that are sequentially adjacent in a horizontal or vertical direction.
[0142] The processor unit (13) can calculate and / or identify an average depth (60), a graph of the average depth (60), a suitable depth graph (80), a partial average depth (70), and a rotation angle corresponding to the partial average depth (70) on the suitable depth graph (80) for a plurality of data points (30) sequentially adjacent in the horizontal or vertical direction.
[0143] The processor unit (13) can identify a pair of data points (30) that have the same partial average depth (70) value among a plurality of data points (30) that are sequentially adjacent in the horizontal or vertical direction.
[0144] According to the present invention, the value of the partial average depth (70) being identical may include not only cases where the value of the partial average depth (70) is numerically identical, but also cases where it is similar within an allowable error range set in the device to be inspected itself. The allowable error range may be the precision of the device to be inspected. Alternatively, the allowable error range may be determined based on the precision of the device to be inspected.
[0145] According to the present invention, when the minimum value of the angle that the rotation adjustment unit (12) can adjust in a single operation is greater than a specific threshold value, there cannot be more than one pair of data points (30) with the same value of partial average depth (70). The present invention is intended to overcome technical limitations related to the minimum adjustable angle of the rotation adjustment unit (12) and to determine the angular resolution of the device to be inspected.
[0146] The processor unit (13) can identify the rotation angle corresponding to the partial average depth (70) on each suitable depth graph (80) for a pair of data points (30) in which the value of the partial average depth (70) is the same.
[0147] The processor unit (13) can identify the number of data points (30) that exist sequentially in a horizontal or vertical direction between the pair of data points (30).
[0148] The processor unit (13) can determine each resolution of the device to be inspected based on the rotation angle corresponding to each of the partial average depths (70) of the pair of data points (30) and the number of data points (30) that exist sequentially in the horizontal or vertical direction between the pair of data points (30).
[0149] The processor unit (13) can determine each resolution through the following [Equation 2].
[0150]
[0151]
[0152] X1, X2: Rotation angles corresponding to the respective partial average depth (70) of a pair of data points with the same average depth (60) (where X1 < X2)
[0153] N: The number of data points existing between the above pair of data points (including both ends)
[0154]
[0155] For example, referring to FIG. 9, the processor unit (13) can calculate / identify the rotation angle corresponding to the partial average depth (70) on the average depth (60), the graph of the average depth (60), the suitable depth graph (80) (C1, C2, C3, C4), the partial average depth (70) (see M1, M2, M3, M4), and the suitable depth graph (80) (C1, C2, C3, C4) for each of the first data point (30a), the second data point (30b), the third data point (30c), and the fourth data point (30d) shown in FIG. 8.
[0156] As illustrated in FIG. 9, the processor unit (13) can identify the first data point (30a) and the fourth data point (30d) that have the same partial average depth (70) value among the first data point (30a), the second data point (30b), the third data point (30c), and the fourth data point (30d) that are sequentially adjacent in the horizontal direction.
[0157] As illustrated in FIG. 9, the processor unit (13) can identify rotation angles (X1, X2) corresponding to the partial average depth (70) on each suitable depth graph (80) (C1, C4) for the first data point (30a) and the fourth data point (30d) having the same value of partial average depth (70).
[0158] As illustrated in FIG. 9, the processor unit (13) can identify the second data point (30b) and the third data point (30c) that exist sequentially in the horizontal direction between the first data point (30a) and the fourth data point (30d) on the frame. That is, the processor unit (13) can identify that there are four data points (30) sequentially in the horizontal direction from the first data point (30a) to the fourth data point (30d).
[0159] As illustrated in FIG. 9, the processor unit (13) can determine each resolution of the inspection target device through the above-described [Equation 2]. More specifically, each resolution can be determined in the form of (X2 - X1) / (4 - 1).
[0160]
[0161] Method for inspecting each resolution of a LiDAR device
[0162] FIG. 10 is a flowchart illustrating a method for inspecting each resolution of a lidar device according to one embodiment of the present invention.
[0163] As illustrated in FIG. 10, each resolution inspection method of the lidar device (100) may include the following steps.
[0164] The inspection method may include the step (S100) of preparing a device to be inspected and at least one target. For example, the device to be inspected may be a LiDAR device (100).
[0165] The inspection method may include the step (S200) of placing a device to be inspected and at least one target. Detailed information regarding the placement between the device to be inspected and the target is as described above.
[0166] The inspection method may include a step (300) of rotating the device to be inspected and a step (S400) of recording a plurality of frames for the target through the device to be inspected. More specifically, steps (S300) and (S400) may be repeated n times from i=1 to i=n. n is a natural number. That is, for each rotation step performed sequentially, a plurality of frames may be recorded.
[0167] The inspection method may include a step of selecting a data point (30) (S500), a step of obtaining depth information (50) of the data point (30) from a plurality of frames (S600), and a step of calculating the average depth (60) of the data point (30) from the depth information (50) of the data point (30) (S700). Steps (S600) and (S700) may be repeated n times from h=1 to h=n. For each rotation step performed sequentially by a preset angle, the average depth (60) of the data point (30) may be calculated n times. Detailed information regarding the depth information (50) and the average depth (60) is as described above.
[0168] The inspection method may include the step (S800) of generating a fitting depth graph (80) for a data point (30), the step (S900) of calculating a partial average depth (70) for at least some of the average depths (60) for the data point (30), and the step (S1000) of calculating a rotation angle corresponding to the partial average depth (70) for the data point (30) on the fitting depth graph (80). The details regarding the fitting depth graph (80) and the partial average depth (70) are as described above.
[0169] Steps (S500), (S600), and (S700) are repeated n times, and steps (S800), (S900), and (1000) can be repeated m times from j=1 to j=m. m is a natural number greater than 1. That is, for m data points (30), steps (S500), (S600), (S700), (S800), (S900), and (1000) can be repeated sequentially.
[0170] According to the present invention, m data points (30) may be data points (30) that are adjacent to each other in the horizontal or vertical direction, penetrating the boundary surface (a-a') formed by the edge of the main target (20) while the inspection target device is rotating. For example, when determining the azimuth angle resolution of the inspection target device, m data points (30) that are sequentially adjacent in the horizontal direction may be selected. Also, for example, when determining the elevation angle resolution of the inspection target device, m data points (30) that are sequentially adjacent in the vertical direction may be selected.
[0171] The inspection method may include a step (S1100) of identifying a pair of data points (30) having the same value of partial average depth (70), and a step (S1200) of determining each resolution of the device to be inspected based on the pair of data points (30). Detailed information on the method for determining each resolution is as described above.
[0172] According to the present invention, a method for inspecting each resolution of a lidar device (100) (100) may include the steps of: placing at least one target spaced apart from the lidar device (100) to inspect; observing a point cloud through the lidar device (100) to inspect; rotating the lidar device (100) to shift at least one of a point and a row or column point from the edge of the target; acquiring depth information at each rotation angle until at least one of the point and the row or column point moves away from the target; rotating the lidar device (100) to inspect until the same depth information is acquired within the precision of the lidar device (100); and acquiring each resolution by considering the number of data points between the same depth information within the precision of the lidar device (100).
[0173] According to the present invention, a method for inspecting each resolution of a lidar device (100) may include the steps of: placing the lidar device (100) to be inspected on a rotating stage spaced apart from at least one target; irradiating light from the lidar device (100) to be inspected into an edge region of the target so that a point cloud is observed; rotating the rotating stage according to a predetermined rotation angle in an azimuth direction or an elevation direction, and recording a plurality of frames for the target through the lidar device (100) to be inspected at each rotation angle; obtaining a suitable depth graph at each rotation angle from the frames; obtaining depth information included within an allowable error range according to the precision of the lidar device (100) to be inspected and the rotation angle having the depth information included within the allowable error range from the suitable depth graph; and determining each resolution of the lidar device (100) to be inspected based on the rotation angle having the depth information.
[0174]
[0175] Although the present invention described above has been explained with reference to the embodiments illustrated in the drawings, this is merely illustrative, and those skilled in the art will understand that various modifications and variations of the embodiments are possible therefrom. However, such modifications should be considered to be within the technical scope of protection of the present invention. Accordingly, the true technical scope of protection of the present invention should be determined by the technical spirit of the appended claims.
[0176] [Explanation of the symbol]
[0177] 10: Each resolution inspection device of the LiDAR device
[0178] 11: Support for the device under inspection
[0179] 12: Rotation adjustment section
[0180] 13: Processor section
[0181] 14: Memory section
[0182] 20: Main Target
[0183] 21 : Sub-target
[0184] 30: Data points
[0185] 30a: 1st data point
[0186] 30b: 2nd data point
[0187] 30c: 3rd data point
[0188] 30d: 4th data point
[0189] 31 : Reach point
[0190] 31a : First point of arrival
[0191] 31b : Second arrival point
[0192] 31c: Third point of arrival
[0193] 31d: 4th point of arrival
[0194] 50: Depth information
[0195] 60: Average depth
[0196] 70: Partial average depth
[0197] 80: Fit depth graph
[0198] 100: Light sensing and distance measuring device
[0199] 110 : Photoeer
[0200] 120 : Photodetector
[0201] 130 : Optical device
[0202] 200 : Object
[0203] S100: A step of preparing a device to be inspected and at least one target.
[0204] S200: Step of placing a device to be inspected and at least one target
[0205] S300: Step of rotating the device to be inspected
[0206] S400: A step of recording multiple frames for a target through a device under inspection.
[0207] S500: Step of selecting data points (30)
[0208] S600: A step of obtaining depth information (50) of data points (30) from multiple frames.
[0209] S700: A step of calculating the average depth (60) of a data point (30) from the depth information (50) of a data point (30).
[0210] S800: A step of generating a depth of fit graph (80) for data points (30).
[0211] S900: A step of calculating a partial average depth (70) for at least some of the average depths (60) for data points (30).
[0212] S1000: A step of calculating a rotation angle corresponding to the partial average depth (70) for a data point (30) on a suitable depth graph (80).
[0213] S1100: A step of identifying a pair of data points (30) having the same value of partial average depth (70).
[0214] S1200: A step of determining each resolution of the device to be inspected based on a pair of data points (30).
Claims
1. In a method for inspecting each resolution of a LiDAR (Light Detection And Ranging) device, A step of preparing a lidar device to be inspected and at least one target; A step of rotating the above-mentioned inspection target lidar device in an azimuth direction or an elevation direction in n steps, and recording m frames of the target through the inspection target lidar device at each of the n rotation angles of the above steps; A step of selecting at least one data point from a point cloud calculated by the inspection target lidar device through the above-mentioned recorded frame; and A step of determining each resolution of the inspected LiDAR device based on depth information included in each of the selected data points; Includes, Each of the above data points corresponds to each of the plurality of superpixels constituting the photodetector of the lidar device under inspection, and The depth information included in the above data point is calculated by the corresponding super pixel based on a mean value-based method over a number of times equal to the number of frames (m) at each of the n rotation angles. Method for inspecting each resolution of a LiDAR device.
2. In Paragraph 1, The above super pixel includes a plurality of individual pixels, and The depth information calculated by the super pixel is calculated by processing the information calculated by each of the individual pixels in an average-value-based manner. Method for inspecting each resolution of a LiDAR device.
3. In Paragraph 2, The step of determining each resolution of the inspected lidar device based on depth information included in the selected data points is A step of calculating the partial average depth for the data points based on the depth information and generating a suitable depth graph for the data points; Includes, Each of the above resolutions is determined based on the generated fit depth graph and the calculated partial average depth, Method for inspecting each resolution of a LiDAR device.
4. In Paragraph 1, Step of placing the above-mentioned inspection target lidar device and the above-mentioned target; Includes more, The above target includes a primary target and a secondary target, and The above main target is positioned so that its edge faces the optical center of the inspection target lidar device head-on, and The above sub-target is positioned parallel to the main target at the rear of the main target in a direction away from the inspection target lidar device. Method for inspecting each resolution of a LiDAR device.
5. In Paragraph 1, The above-mentioned inspection target lidar device rotates by a preset angle in each of the n steps, Method for inspecting each resolution of a LiDAR device.
6. In Paragraph 1, The selected data points are sequentially adjacent in the horizontal or vertical direction within the calculated point cloud, Method for inspecting each resolution of a LiDAR device.
7. In Paragraph 3, The step of calculating the partial average depth for the data point based on the depth information and generating a suitable depth graph for the data point is, A step of calculating the average depth for the data point as the average of the depth information at each of the n rotation angles; Includes, The above partial average depth is calculated as the average of at least some of the above calculated average depths, and The above-mentioned fitting depth graph is generated by applying curve fitting to a graph plotting the above-mentioned average depth against the above-mentioned rotation angle. Method for inspecting each resolution of a LiDAR device.
8. In Paragraph 7, The step of determining each resolution of the inspected lidar device based on depth information included in the selected data points is A step of identifying a pair of data points among the selected data points in which the partial average depth values are mutually identical; For each of the identified data points, a step of calculating the rotation angle corresponding to the partial average depth on the suitable depth graph for each of the data points; A step of identifying the number of data points that exist sequentially adjacent to each other in a horizontal or vertical direction between the identified data points; and A step of determining the angular resolution of the lidar device to be inspected based on the calculated rotation angle and the identified number; including, Method for inspecting each resolution of a LiDAR device.
9. In each resolution inspection device of a LiDAR (Light Detection And Ranging) device, A support for a device to be inspected for supporting the device to be inspected; A support member for the device to be inspected and a rotation adjustment member for rotating the device to be inspected; and A processor unit for controlling the operation of the above-mentioned inspection target device and the above-mentioned rotation adjustment unit, and for determining each resolution of the above-mentioned inspection target device based on information output by the above-mentioned inspection target device and the above-mentioned rotation adjustment unit; including, Each resolution inspection device of the LiDAR device.
10. In Paragraph 9, The above processor unit is, A step of rotating the above-mentioned inspection target lidar device in an azimuth direction or an elevation direction in n steps, and recording m frames of the target through the inspection target lidar device at each of the n rotation angles of the above steps; A step of selecting at least one data point from a point cloud calculated by the inspection target lidar device through the above-mentioned recorded frame; and A step of determining each resolution of the inspected LiDAR device based on depth information included in each of the selected data points; It is intended to perform, Each of the above data points corresponds to each of the plurality of superpixels constituting the photodetector of the lidar device under inspection, and The depth information included in the above data point is calculated by the corresponding superpixel based on a mean value-based method over a number of times equal to the number of frames (m) at each of the n rotation angles. Each resolution inspection device of the LiDAR device.
11. In Paragraph 10, The above super pixel includes a plurality of individual pixels, and The depth information calculated by the super pixel is calculated by processing the information calculated by each of the individual pixels in an average-value-based manner. Each resolution inspection device of the LiDAR device.
12. In Paragraph 11, The step of determining each resolution of the inspected lidar device based on depth information included in the selected data points is A step of calculating the partial average depth for the data points based on the depth information and generating a suitable depth graph for the data points; Includes, Each of the above resolutions is determined based on the generated fit depth graph and the calculated partial average depth, Each resolution inspection device of the LiDAR device.
13. In Paragraph 10, The above target includes a primary target and a secondary target, and The above main target is positioned so that the optical center and edge of the lidar device under inspection face each other directly, and The above sub-target is positioned parallel to the main target at the rear of the main target in a direction away from the inspection target lidar device. Each resolution inspection device of the LiDAR device.
14. In Paragraph 10, The above-mentioned inspection target lidar device rotates by a preset angle in each of the n steps, Each resolution inspection device of the LiDAR device.
15. In Paragraph 10, The selected data points are sequentially adjacent in the horizontal or vertical direction within the calculated point cloud, Each resolution inspection device of the LiDAR device.
16. In Paragraph 12, The step of calculating the partial average depth for the data point based on the depth information and generating a suitable depth graph for the data point is, A step of calculating the average depth for the data point as the average of the depth information at each of the n rotation angles; Includes, The above partial average depth is calculated as the average of at least some of the above calculated average depths, and The above-mentioned fitting depth graph is generated by applying curve fitting to a graph plotting the above-mentioned average depth against the above-mentioned rotation angle. Each resolution inspection device of the LiDAR device.
17. In Paragraph 16, The step of determining each resolution of the inspected lidar device based on depth information included in the selected data points is A step of identifying a pair of data points among the selected data points in which the partial average depth values are mutually identical; For each of the identified data points, a step of calculating the rotation angle corresponding to the partial average depth on the suitable depth graph for each of the data points; A step of identifying the number of data points that exist sequentially adjacent to each other in a horizontal or vertical direction between the identified data points; and A step of determining the angular resolution of the lidar device to be inspected based on the calculated rotation angle and the identified number; including, Each resolution inspection device of the LiDAR device.