Object recognition method and object recognition device
The object recognition method and device accurately differentiate cylindrical road structures from other objects by analyzing point cloud data and circularity differences, enhancing detection accuracy even when structures are close together.
Patent Information
- Application Number
- PCT/JP2023/019172
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2023-05-23
- Publication Date
- 2024-11-28
AI Technical Summary
Existing object recognition technologies fail to accurately distinguish cylindrical road structures from other objects when they are close to each other, often detecting them as a single entity.
An object recognition method and device that utilize a ranging sensor to generate point cloud data, extract arc-shaped components, calculate circularity, and determine cylindrical structures by analyzing the difference in circularity and other characteristics of circular arc components across multiple layers.
Prevents cylindrical road structures from being detected as a single entity with other objects, improving detection accuracy of columnar structures and nearby objects by determining cylindrical structures based on circularity and variance thresholds.
Smart Images

Figure JP2023019172_28112024_PF_FP_ABST
Abstract
Description
Object recognition method and object recognition device
[0001] The present disclosure relates to an object recognition method and an object recognition device.
[0002] A technology has been proposed in which, based on a plurality of detection points detected by a laser radar, a set of detection points that are within a predetermined distance from each other is detected as a cluster candidate, and the detected cluster candidate is associated with the previously detected cluster candidate at predetermined time intervals, and a cluster candidate with a high degree of shape match among the associated cluster candidates is identified as an object (see Patent Document 1).
[0003] JP 2013-228259 A
[0004] The technology described in Patent Document 1 may not be able to identify multiple objects in close proximity as individual objects. In particular, when a cylindrical road structure such as a bollard or pole is placed close to another object, the road structure and the other object may be detected as a single unit.
[0005] The present disclosure has been made in consideration of the above-mentioned problems, and aims to provide an object recognition method and an object recognition device that can prevent a cylindrical road structure and another object from being detected as one object, even when the cylindrical road structure is located close to the other object.
[0006] In order to solve the above-mentioned problems, an object recognition method and an object recognition device according to one aspect of the present disclosure process point cloud data relating to the positions of ranging points located in a plurality of different vertical directions and a plurality of different horizontal directions on the surface of an object, extracting arc-shaped components for each of the plurality of different vertical positions based on the ranging point data for the plurality of different vertical positions and the plurality of different horizontal positions, calculating the circularity of each of the plurality of arc-shaped components when the arc-shaped components are extracted, and determining that the specific object corresponding to the ranging point related to the plurality of arc-shaped components is a cylindrical structure when the difference between the plurality of circularities is less than a predetermined value.
[0007] According to the present disclosure, even if a cylindrical road structure is placed close to other objects, it is possible to prevent the road structure and other objects from being detected as one object.
[0008] Fig. 1 is a block diagram showing the configuration of an object recognition device according to an embodiment of the present disclosure. Fig. 2 is a flowchart showing processing of the object recognition device according to an embodiment of the present disclosure. Fig. 3 is a schematic diagram showing the positional relationship between a distance measurement sensor and distance measurement points on the surface of a cylindrical structure. Fig. 4A is a first example showing the arrangement of distance measurement points on the surface of a cylindrical structure. Fig. 4B is a second example showing the arrangement of distance measurement points on the surface of a cylindrical structure.
[0009] Next, embodiments of the present disclosure will be described in detail with reference to the drawings. In the description, the same components are designated by the same reference numerals and redundant description will be omitted.
[0010] [Configuration of Object Recognition Device] An example configuration of an object recognition device 1 will be described with reference to Fig. 1. Fig. 1 is a block diagram showing the configuration of an object recognition device according to this embodiment. As shown in Fig. 1, the object recognition device 1 includes a distance measurement sensor 10 (sensor) and a controller 20.
[0011] The object recognition device 1 may be mounted on a vehicle with an automatic driving function, or on a vehicle without an automatic driving function. Furthermore, the object recognition device 1 may be mounted on a vehicle capable of switching between automatic driving and manual driving. Furthermore, the automatic driving function may be a driving assistance function that automatically controls only some of the vehicle control functions, such as steering control, braking force control, and driving force control, to assist the driver in driving. In this embodiment, the object recognition device 1 is described as being mounted on a vehicle with an automatic driving function.
[0012] 1, the object recognition device 1 may control various actuators such as a steering actuator, an accelerator pedal actuator, and a brake actuator based on the recognition results (position, shape, attitude, etc. of the object), thereby realizing highly accurate autonomous driving.
[0013] The distance measurement sensor 10 generates point cloud data relating to the positions of distance measurement points located in a plurality of different vertical directions and a plurality of different horizontal directions on the surface of an object.
[0014] For example, the distance measurement sensor 10 may include a sensor that measures the distance and direction to objects around the vehicle by emitting electromagnetic waves from an emission point around the vehicle and detecting the positions of reflection points based on the waves reflected from the emitted electromagnetic waves. One example of such a sensor is a LIDAR (Laser Imaging Detection and Ranging). A LIDAR is a sensor that emits light (laser light) from an emission point around a predetermined range around the vehicle, detects the positions of reflection points, which are distance measurement points, based on the reflected waves, and generates point cloud data related to the distance measurement points.
[0015] Lidar measures the distance and direction to an object and recognizes the shape of the object by measuring the time it takes for the light (reflected wave) to bounce back after being emitted. Lidar can also obtain the positional relationship of objects in three dimensions. Mapping is also possible using the intensity of the reflected wave.
[0016] For example, a lidar scans the surroundings of a vehicle in the main scanning direction and the sub-scanning direction by changing the direction of light irradiation. In particular, the lidar acquires point cloud data related to measurement points located on the sides of a cone with the emission point as its apex by changing the angle from the horizontal plane (the horizontal plane in the lidar's coordinate system) when emitting electromagnetic waves from the emission point. One side of the cone corresponds to one angle when the lidar emits electromagnetic waves. Hereinafter, one side of the cone will be referred to as a "layer."
[0017] 3 is a schematic diagram showing the positional relationship between the distance measurement sensor 10 and distance measurement points on the surface of a cylindrical structure. In Fig. 3, the area through which the electromagnetic waves emitted from the emission point of the distance measurement sensor 10 pass is represented as a cone BL. The emission point of the distance measurement sensor 10 is located at the apex of the cone BL, and the emitted electromagnetic waves pass through the surface of the cone BL. In addition, multiple distance measurement points P located at the intersections of the cone BL and the surface of the cylindrical structure LS are indicated by black dots.
[0018] For example, the LIDAR acquires point cloud data related to the ranging points included in one "layer" by scanning within the layer (scanning in the main scanning direction). The LIDAR then acquires point cloud data for multiple "layers" by scanning in the sub-scanning direction, changing the angle from the horizontal plane when emitting electromagnetic waves from the emission point. In this way, the LIDAR sequentially irradiates multiple ranging points around the vehicle with light. The LIDAR emits electromagnetic waves and generates point cloud data related to the ranging points located on the surface of the object based on the reflected waves of the electromagnetic waves from the surface of the object.
[0019] The LIDAR repeatedly emits light to all of the distance measurement points around the vehicle at predetermined time intervals. The LIDAR generates information (distance measurement point information) for each distance measurement point obtained by the light irradiation. The LIDAR then outputs point cloud data consisting of the distance measurement point information to the controller 20.
[0020] The ranging point information includes position information of the ranging point. The position information is information indicating the position coordinates of the ranging point. The position coordinates may use a polar coordinate system represented by the direction from the lidar to the ranging point (yaw angle, pitch angle) and the distance from the lidar to the ranging point (depth). The position coordinates may use a three-dimensional coordinate system represented by x, y, and z coordinates with the installation position of the lidar as the origin. For example, the x and y coordinates may be coordinates on a horizontal plane in the lidar coordinate system, and the z coordinate may be a coordinate on an axis in the height direction perpendicular to the horizontal plane. Note that the x and y coordinates may be coordinates on a plane parallel to the road surface on which the vehicle is traveling, and the z coordinate may be a coordinate on an axis in the height direction above the road surface.
[0021] The ranging point information may also include time information of the ranging point. The time information is information indicating the time when the position information of the ranging point was generated (when the reflected electromagnetic wave was received). Additionally, the ranging point information may also include information on the intensity of the reflected wave from the ranging point (intensity information).
[0022] Alternatively, the distance measurement sensor 10 may be a stereo camera, which generates point cloud data relating to the positions of distance measurement points using the principles of trigonometry based on a plurality of image data captured by the stereo camera.
[0023] The controller 20 is a general-purpose microcomputer equipped with a CPU (Central Processing Unit), memory, and input / output units. A computer program for causing the microcomputer to function as the object recognition device 1 is installed in the microcomputer. By executing the computer program, the microcomputer functions as multiple information processing circuits equipped in the object recognition device 1. The controller 20 processes data acquired from the distance measurement sensor 10.
[0024] Although an example is shown here in which the multiple information processing circuits provided in the object recognition device 1 are realized by software, it is of course also possible to configure the information processing circuits by providing dedicated hardware for executing each of the information processes described below.Furthermore, the multiple information processing circuits may be configured by individual hardware.
[0025] The controller 20 includes, as examples of a plurality of information processing circuits (information processing functions), a point cloud acquisition unit 21, an arc-shaped component extraction unit 23, a cylinder candidate extraction unit 25, a calculation unit 27, a determination unit 29, and an output unit 31. The controller 20 may also be expressed as an ECU (Electronic Control Unit).
[0026] The point cloud acquisition unit 21 acquires point cloud data from the distance measurement sensor 10. Then, based on the point cloud data, the point cloud acquisition unit 21 extracts distance measurement points of objects present around the vehicle, excluding distance measurement points on the road surface on which the vehicle is traveling.
[0027] The arc-shaped component extraction unit 23 extracts arc-shaped components included in the arrangement of the ranging points based on the positions of the ranging points on a horizontal plane. More specifically, the arc-shaped component extraction unit 23 refers to the coordinate points of the ranging points on the horizontal plane. Note that the arc-shaped component extraction unit 23 may project the ranging points onto a horizontal plane and extract the coordinate points of the projected points on the road surface as the coordinate points of the ranging points on the horizontal plane. Alternatively, the arc-shaped component extraction unit 23 may project the ranging points onto the road surface on which the vehicle is traveling.
[0028] The arc-shaped component extraction unit 23 then applies a Hough transform to extract, as an arc-shaped component, an arc that passes through the coordinate points of the ranging points in the horizontal plane. The Hough transform calculates, for each ranging point, candidate parameters (center position and radius of the circle) that characterize a circle that passes through the coordinate points of the ranging points in the horizontal plane. The calculated parameters are then plotted in a three-dimensional space having axes corresponding to the x- and y-coordinates indicating the center position of the circle and the radius r of the circle. In the three-dimensional space, one two-dimensional curved surface is associated with one ranging point.
[0029] Then, in the Hough transform, after the plotting of parameters for all ranging points is completed, the shared point (most likely point) through which the greatest number of two-dimensional surfaces pass is calculated from the shared points of the two-dimensional surfaces associated with the ranging points in the three-dimensional space.
[0030] The coordinates of the most likely point are parameters that characterize a circle that passes through the coordinate points in the horizontal plane of the most distance measurement points in the original two-dimensional plane. Therefore, the arc-shaped component extraction unit 23 extracts the circle characterized by the coordinates of the most likely point as an arc-shaped component.
[0031] The cylinder candidate extraction unit 25 extracts, as cylinder candidates, multiple arc-shaped components whose centers are "close" among multiple arc-shaped components across multiple layers. For example, the cylinder candidate extraction unit 25 extracts, as cylinder candidates, multiple arc-shaped components that overlap each other. Here, "overlapping each other" means that the multiple circles or the interiors of the circles extracted as arc-shaped components intersect in a horizontal plane. Furthermore, the arc-shaped components to be determined as overlapping may be two arc-shaped components extracted in different layers, or two arc-shaped components extracted in the same layer.
[0032] The cylinder candidate extraction unit 25 may determine that the centers of multiple circles extracted as arc-shaped components are "close" when the distance between them is equal to or less than a predetermined distance. Here, the predetermined distance may be the sum of the radii of the two circles for which overlap is to be determined, or may be a value obtained by multiplying the sum by a predetermined positive coefficient less than 1. The closer the coefficient is to 0, the closer the centers of the circles extracted as arc-shaped components are to each other.
[0033] That is, the cylinder candidate extraction unit 25 groups multiple arc-shaped components based on whether they are “close” to each other. The cylinder candidate extraction unit 25 extracts the grouped multiple arc-shaped components as cylinder candidates.
[0034] The calculation unit 27 calculates the characteristics of the cylinder candidate. For example, the calculation unit 27 calculates the circularity of the cylinder candidate based on the ranging points related to the cylinder candidate, and calculates the variance of the circularity as the first variance.
[0035] More specifically, the calculation unit 27 calculates the circularity based on the ranging points associated with the arc-shaped components grouped as cylinder candidates. When calculating the circularity, the calculation unit 27 calculates the distance between the coordinate points in the horizontal plane of the ranging points associated with the arc-shaped components and the center positions of the circles extracted as the arc-shaped components, and calculates the root mean square of the difference between the calculated distance and the radius of the circles extracted as the arc-shaped components. The calculation unit 27 regards the root mean square as the circularity. The calculation unit 27 calculates the circularity for each arc-shaped component. Then, the calculation unit 27 calculates the variance of the circularity obtained for each arc-shaped component as a first variance.
[0036] When only one arc-shaped component is extracted in each layer, the coordinate point in the horizontal plane of the distance measurement point associated with the arc-shaped component passes through the circle extracted as the arc-shaped component, and so in principle the calculated circularity will be 0. However, in practice, multiple distinct arc-shaped components that are close to each other may be extracted in one layer, and the circularity will not necessarily be 0. However, if the cylinder candidate is a cylindrical structure such as a bollard, pole, or utility pole that actually exists on the road surface, the circularity will be a value close to 0. This property can be used to determine whether the cylinder candidate is a cylindrical structure.
[0037] The calculation unit 27 may calculate the center position of an arc-shaped component of the cylinder candidate as a feature of the cylinder candidate. More specifically, when one arc-shaped component is extracted from each layer, the calculation unit 27 may set the center position of the circle extracted as the arc-shaped component as the center position of the arc-shaped component of the cylinder candidate. Furthermore, when multiple arc-shaped components are extracted from each layer, the calculation unit 27 may calculate the average of the center positions of the circles extracted as the arc-shaped components.
[0038] The calculation unit 27 may calculate the radius of an arc-shaped component of the cylinder candidate as a feature of the cylinder candidate, and calculate the variance of the radius as the second variance. More specifically, the calculation unit 27 may calculate the average of the radii of the arc-shaped components of the cylinder candidate across multiple layers, and calculate the root mean square of the difference between the calculated average and the radius of the arc-shaped component as the second variance.
[0039] The calculation unit 27 may calculate the radius of an arc-shaped component of the cylinder candidate as a feature of the cylinder candidate, and calculate the rate of change of the radius in the height direction as the radius change rate. For example, the calculation unit 27 may calculate the change in the radius of the arc-shaped component when the height from the road surface on which the vehicle is traveling increases by a unit length as the radius change rate. More specifically, the calculation unit 27 may calculate the difference in the radii of the arc-shaped components between two adjacent layers, and divide the difference by the distance between the layers to obtain the radius change rate.
[0040] The calculation unit 27 may calculate the radius of an arc-shaped component of the cylinder candidate as a feature of the cylinder candidate, and calculate the rate of change in the height direction of the curvature of the curve as the curvature change rate based on a curve that indicates the change in radius along the height direction.
[0041] The determination unit 29 determines whether the cylinder candidate satisfies the condition for being a cylindrical structure based on the characteristics of the cylinder candidate. For example, the determination unit 29 determines a specific object corresponding to a ranging point related to multiple arc-shaped components as a cylindrical structure when the difference between multiple circularities is less than a predetermined value. In other words, the determination unit 29 determines a specific object as a cylindrical structure when an arc-shaped component whose difference in circularity is less than a predetermined value is found. Alternatively, the determination unit 29 may determine a specific object corresponding to a ranging point related to the cylinder candidate as a cylindrical structure when the first variance is less than a first threshold. Here, the first threshold may be predetermined based on the shape of a cylindrical structure (such as a bollard, a pole, or a utility pole) that may exist in the environment in which the vehicle travels.
[0042] On the other hand, the determination unit 29 may determine that the specific object is not a cylindrical structure when the difference between the multiple roundnesses is equal to or greater than a predetermined value.
[0043] The determination unit 29 may determine that the specific object is a cylindrical structure when the difference between the multiple circularities is less than a predetermined value and the center positions are located on the same straight line.
[0044] For example, the determination unit 29 arbitrarily selects a first layer, a second layer, and a third layer arranged in order in the height direction from the multiple layers. Then, the determination unit 29 determines whether the angle formed by a line connecting the center position of the first layer to the center position of the second layer and a line connecting the center position of the second layer to the center position of the third layer is equal to or less than a predetermined angle. For the arbitrarily selected first layer, second layer, and third layer, the determination unit 29 may determine that the center positions are located on the same line if the angle formed by the two lines is equal to or less than a predetermined angle. Alternatively, the determination unit 29 may draw an approximate line for the center position and calculate a variance related to the deviation of the center position from the approximate line. Then, the determination unit 29 may determine that the center positions are located on the same line if the variance related to the deviation of the center position from the approximate line is equal to or less than a predetermined threshold.
[0045] The conditions for determining that a specific object is a cylindrical structure take into account that many cylindrical structures that may exist in the environment in which a vehicle is traveling extend in a straight line (especially extending vertically) and that the center positions of the arc-shaped components are aligned on a straight line.
[0046] On the other hand, the determination unit 29 may determine that the specific object is not a cylindrical structure when the center positions are not located on the same straight line.
[0047] The determination unit 29 may determine that the specific object is a cylindrical structure when the difference between the multiple circularity measurements is less than a predetermined value and the second variance is less than a second threshold. Here, the second threshold may be determined in advance based on the shape of a cylindrical structure that may exist in the environment in which the vehicle travels. The second variance is the variance of the radius of the arc-shaped component of the cylinder candidate. For cylindrical structures that may exist in the environment in which the vehicle travels, the second variance tends to be relatively small. On the other hand, the determination unit 29 may determine that the specific object is not a cylindrical structure when the second variance is equal to or greater than the second threshold.
[0048] The determination unit 29 may determine that the specific object is a cylindrical structure when the difference between the multiple roundnesses is less than a predetermined value and the radius change rate is less than a third threshold. Here, the third threshold may be determined in advance based on the shape of a cylindrical structure that may exist in the environment in which the vehicle travels. On the other hand, the determination unit 29 may determine that the specific object is not a cylindrical structure when the radius change rate is equal to or greater than the third threshold.
[0049] The determination unit 29 may determine that the specific object is a cylindrical structure when the difference between the multiple roundness measurements is less than a predetermined value and the curvature change rate is less than a fourth threshold. Here, the fourth threshold may be determined in advance based on the shape of a cylindrical structure that may exist in the environment in which the vehicle travels. This determination condition takes into account, for example, that the curvature is constant on the surface of a spherical bollard.
[0050] On the other hand, the determination unit 29 may determine that the specific object is not a cylindrical structure when the curvature change rate is equal to or greater than a fourth threshold value.
[0051] The output unit 31 outputs the determination result by the determination unit 29. Alternatively, the output unit 31 may output point cloud data from which data of the ranging points related to the specific object determined to be a cylindrical structure has been removed.
[0052] [Processing Procedure of Object Recognition Apparatus] Next, a processing procedure of the object recognition apparatus 1 according to this embodiment will be described with reference to the flowchart of Fig. 2. Fig. 2 is a flowchart showing the processing of the object recognition apparatus according to this embodiment. The processing of the object recognition apparatus 1 shown in Fig. 2 may be repeatedly executed at a predetermined cycle.
[0053] First, in step S101, the point cloud acquisition unit 21 acquires point cloud data from the distance measurement sensor 10. Then, based on the point cloud data, the point cloud acquisition unit 21 extracts distance measurement points of objects present around the vehicle, excluding distance measurement points on the road surface on which the vehicle is traveling.
[0054] In step S103, the arc-shaped component extraction unit 23 selects an unprocessed layer from among a plurality of layers in which point cloud data exists.
[0055] In step S107, the arc component extracting unit 23 extracts an arc passing through the coordinate point of the distance measurement point in the horizontal plane as an arc component.
[0056] In step S109, the arc component extraction unit 23 determines whether or not all of the layers containing point cloud data have been processed.
[0057] That is, the arc-shaped component extraction unit 23 determines whether or not the arc-shaped components have been extracted in all layers, regardless of whether or not the arc-shaped components have been extracted.
[0058] If it is determined that all layers have not been processed (NO in step S109), the process returns to step S103.
[0059] If it is determined that all layers have been processed (YES in step S109), in step S111, the cylinder candidate extraction unit 25 extracts, from among the multiple arc-shaped components, multiple adjacent arc-shaped components as cylinder candidates.
[0060] In step S113, the calculation unit 27 calculates the characteristics of the cylinder candidate.
[0061] In step S115, the determination unit 29 determines whether or not the cylinder candidate satisfies the condition for being a columnar structure, based on the characteristics of the cylinder candidate.
[0062] If it is determined that the cylinder candidate satisfies the condition for being a columnar structure (YES in step S115), in step S117, the determining unit 29 determines that the cylinder candidate is a columnar structure.
[0063] If it is determined that the cylinder candidate does not satisfy the condition for being a cylindrical structure (NO in step S115), in step S119, the determining unit 29 determines that the cylinder candidate is something other than a cylindrical structure.
[0064] In step S121, the determination result is output by the determination unit 29. After that, the flow chart of FIG.
[0065] As described above in detail, the object recognition method and object recognition device according to this embodiment processes point cloud data relating to the positions of ranging points located in a plurality of different vertical directions and a plurality of different horizontal directions on the surface of an object. In this process, arc-shaped components are extracted for each of the plurality of different vertical positions based on the ranging point data for the plurality of different vertical positions and the plurality of different horizontal positions. When arc-shaped components are extracted, the circularity of each of the plurality of arc-shaped components is calculated. If the difference between the circularities is less than a predetermined value, the specific object corresponding to the ranging points related to the plurality of arc-shaped components is determined to be a cylindrical structure.
[0066] This prevents the road structure and other objects from being detected as one object, even if the road structure is located close to the other object. In particular, since the road structure is determined to be a cylindrical structure based on circularity, it is possible to extract cylindrical road structures that have a constant cross-sectional shape in the height direction. As a result, it is also possible to improve the detection accuracy of other objects close to the cylindrical road structure.
[0067] Furthermore, the object recognition method and object recognition device according to this embodiment may calculate the variance of circularity as the difference in circularity as a first variance, and if the first variance is less than a first threshold, determine that the specific object corresponding to the ranging points related to the multiple arc-shaped components is a cylindrical structure. This makes it possible to prevent the road structure and the other object from being detected as one object, even if the cylindrical road structure is located close to the other object.
[0068] Fig. 4A is a first example showing the arrangement of ranging points on the surface of a cylindrical structure. Fig. 4B is a second example showing the arrangement of ranging points on the surface of a cylindrical structure. Figs. 4A and 4B show the arrangement of ranging points aligned in the height direction AR from the road surface RD. It is assumed that each ranging point is located on the surface of the cylindrical structure.
[0069] In Figure 4A, the cylinder candidate GR1 corresponds to a cylindrical structure extending from the road surface RD in the height direction AR. Because the cross-sectional shape of the cylindrical structure does not change along the height direction AR, the arc-shaped components extracted from each layer of the cylinder candidate GR1 are approximately the same. Therefore, the variance of the circularity is small. Therefore, since this corresponds to a case where the variance of the circularity is less than the first threshold, it is possible to determine that the specific object corresponding to the ranging point associated with the cylinder candidate GR1 is a cylindrical structure.
[0070] Furthermore, the object recognition method and object recognition device according to this embodiment may calculate the center positions of a plurality of arc-shaped components, and determine that a specific object is a cylindrical structure if the difference in circularity between the plurality of components is less than a predetermined value and the center positions are located on the same straight line. This improves the accuracy of determining that the specific object is a cylindrical structure.
[0071] In Figure 4A, the cylinder candidate GR2 corresponds to a cylindrical structure that has a cylindrical shape and extends obliquely from the road surface RD in a direction inclined from the height direction AR. Because the centers of the arc-shaped components are located on the same straight line, it is possible to accurately determine even cylindrical structures that extend obliquely from the road surface RD. Furthermore, it is possible to accurately determine even cylindrical structures that have become tilted due to vehicle contact or the like.
[0072] Furthermore, the object recognition method and object recognition device according to this embodiment may calculate the radius of each of the plurality of arc-shaped components, calculate the variance of the radii as a second variance, and determine that the specific object is a cylindrical structure if the difference between the plurality of circularities is less than a predetermined value and the second variance is less than a second threshold value. This improves the accuracy of determining that the specific object is a cylindrical structure.
[0073] Because the radius of the cross section of the cylindrical structure does not change along the height direction AR, the radius of the arc-shaped components extracted in each layer in the cylinder candidate GR1 and the cylinder candidate GR2 is approximately the same. Therefore, the variance of the radii is small. Therefore, since this corresponds to a case where the variance of the radii is less than the second threshold, it is possible to determine that the specific object corresponding to the ranging point related to the cylinder candidate GR1 is a cylindrical structure.
[0074] Furthermore, the object recognition method and object recognition device according to this embodiment may calculate the radius of each of the multiple arc-shaped components, calculate the rate of change of the radius in the height direction as the radius change rate, and determine that the specific object is a cylindrical structure if the difference between the multiple circularities is less than a predetermined value and the radius change rate is less than a third threshold value. This improves the accuracy of determining that the specific object is a cylindrical structure.
[0075] 4B , the radius of the arc-shaped component of the cylinder candidate GR3 gradually decreases as the cylinder candidate GR3 moves from the road surface RD in the height direction AR. Even if the cylindrical structure has a conical shape such as a road cone or a pylon, the specific object corresponding to the ranging point related to the cylinder candidate GR3 can be determined to be a cylindrical structure.
[0076] Furthermore, the object recognition method and object recognition device according to this embodiment may calculate the radii of each of the plurality of arc-shaped components, calculate the rate of change in the height direction of the curvature of the curve as a curvature change rate based on a curve that indicates the change in the radii along the height direction, and determine that the specific object is a cylindrical structure when the difference between the plurality of circularities is less than a predetermined value and the curvature change rate is less than a fourth threshold value. This improves the accuracy of determining that the specific object is a cylindrical structure.
[0077] 4B, the radius of the arc-shaped component of the cylinder candidate GR4 gradually increases and then decreases as it moves from the road surface RD to the height direction AR. For example, this change in radius occurs when the cylindrical structure is a spherical bollard.
[0078] When a cross section of a spherical bollard is obtained along a plane parallel to the height direction AR and passing through the center of the spherical bollard, the curvature of the surface of the spherical bollard is constant (the reciprocal of the radius of the spherical bollard). Therefore, the curvature does not change. Therefore, even if the cylindrical structure is spherical, the specific object corresponding to the ranging point related to the cylinder candidate GR4 can be determined to be a cylindrical structure.
[0079] Furthermore, the object recognition method and object recognition device according to this embodiment may determine that a specific object is not a cylindrical structure when the difference between the multiple circularities is equal to or greater than a predetermined value. This makes it possible to distinguish between cylindrical structures and objects other than cylindrical structures. As a result, it is also possible to improve the detection accuracy of other objects close to the cylindrical structure.
[0080] Each of the functions described in the above embodiments may be implemented by one or more processing circuits, including programmed processors, electrical circuits, and even devices such as application specific integrated circuits (ASICs), or circuit components arranged to perform the described functions.
[0081] Although the contents of the present disclosure have been described above based on the embodiments, the present disclosure is not limited to these descriptions, and various modifications and improvements are possible, which will be apparent to those skilled in the art. The descriptions and drawings that form part of this disclosure should not be understood as limiting the present disclosure. Various alternative embodiments, examples, and operating techniques will be apparent to those skilled in the art from this disclosure.
[0082] Of course, the present disclosure includes various embodiments not described herein. Therefore, the technical scope of the present disclosure is defined only by the invention-specifying matters according to the scope of the claims that are appropriate from the above description.
[0083] REFERENCE SIGNS LIST 1 object recognition device 10 distance measurement sensor 20 controller 21 point cloud acquisition unit 23 arc-shaped component extraction unit 25 cylinder candidate extraction unit 27 calculation unit 29 determination unit 31 output unit
Claims
1. An object recognition method for an object recognition device comprising: a sensor that generates point cloud data regarding the positions of ranging points located in a plurality of different vertical directions and a plurality of different horizontal directions on the surface of an object; and a controller that processes the point cloud data, wherein the controller extracts arc-shaped components for each of the plurality of different vertical positions based on the ranging point data for the plurality of different vertical positions and the plurality of different horizontal positions; when the arc-shaped components are extracted, calculates the circularity of each of the plurality of arc-shaped components; and when the difference between the plurality of circularities is less than a predetermined value, determines that a specific object corresponding to the ranging points related to the plurality of arc-shaped components is a cylindrical structure.
2. The object recognition method described in claim 1, characterized in that the controller calculates the variance of the circularity as a difference in the circularity as a first variance, and if the first variance is less than a first threshold, determines that a specific object corresponding to the ranging points related to the multiple arc-shaped components is a cylindrical structure.
3. The object recognition method according to claim 1, characterized in that the controller calculates the center positions of a plurality of arc-shaped components, and determines that the specific object is a cylindrical structure if the difference in circularity between the plurality of components is less than the predetermined value and the center positions are located on the same straight line.
4. The object recognition method described in claim 1, characterized in that the controller calculates the radius of each of the multiple arc-shaped components, calculates the variance of the radii as a second variance, and determines that the specific object is a cylindrical structure if the difference between the multiple circularities is less than the specified value and the second variance is less than a second threshold value.
5. The object recognition method according to claim 1, characterized in that the controller calculates the radius of each of the multiple arc-shaped components, calculates the rate of change of the radius in the height direction as a radius change rate, and determines that the specific object is a cylindrical structure if the difference between the multiple circularities is less than the specified value and the radius change rate is less than a third threshold value.
6. The object recognition method according to claim 1, characterized in that the controller: calculates the radius of each of the multiple circular arc components; calculates the rate of change in the height direction of the curvature of the curve as a curvature change rate based on a curve indicating the change in the radius along the height direction; and determines that the specific object is a cylindrical structure when the difference between the multiple circularities is less than the specified value and the curvature change rate is less than a fourth threshold value.
7. The object recognition method according to any one of claims 1 to 6, characterized in that the controller determines that the specific object is not a cylindrical structure when the difference between multiple circularities is equal to or greater than the predetermined value.
8. An object recognition device comprising: a sensor that generates point cloud data regarding the positions of ranging points located in a plurality of different vertical directions and a plurality of different horizontal directions on the surface of an object; and a controller that processes the point cloud data, wherein the controller extracts arc-shaped components for each of the plurality of different vertical positions based on the ranging point data for the plurality of different vertical positions and the plurality of different horizontal positions; when the arc-shaped components are extracted, calculates the circularity of each of the plurality of arc-shaped components; and when the difference between the plurality of circularities is less than a predetermined value, determines that a specific object corresponding to the ranging points related to the plurality of arc-shaped components is a cylindrical structure.
Citation Information
Patent Citations
Centering method of cylindrical hedge trimmer based on multi-line laser radar
CN111025323A
metal matrix composite wire with metal clad
JP2007524977A
System and method for controlling surface texturing of metal substrates by low pressure rolling
JP2020528355A
