Lane boundary line detection method and lane boundary line detection device

Abstract

To provide a lane boundary line detection method and a lane boundary line detection device, capable of efficiently detecting a lane boundary line.SOLUTION: A lane boundary line detection method includes acquiring position information of a feature point of an object indicating a lane boundary on a road on which a vehicle V is traveling, and detecting the lane boundary line based on a relative position of the feature point with respect to a reference line Z that represents the road and has a continuously changing tangent inclination.SELECTED DRAWING: Figure 3

Description

Technical Field

[0001] The present invention relates to a lane boundary detection method and a lane boundary detection device.

Background Art

[0002] There is known a travel path shape recognition device that calculates a first travel path shape approximated from feature points indicating a travel path boundary, detects a deterioration position where the degree of coincidence with the feature points deteriorates in the first travel path shape, calculates a second travel path shape approximated from the shape of the travel path boundary farther than the deterioration position, recognizes the travel path boundary nearer than the deterioration position by the first travel path shape, and recognizes the travel path boundary farther than the deterioration position by the second travel path shape (Patent Document 1).

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] In the above prior art, since the calculation of the travel path shape approximated from the shape of the travel path boundary such as the lane boundary is performed before detecting the deterioration position and after detecting the deterioration position, there is a problem that the travel path boundary cannot be efficiently detected.

[0005] The problem to be solved by the present invention is to provide a lane boundary detection method and a lane boundary detection device capable of efficiently detecting a lane boundary.

Means for Solving the Problems

[0006] The present invention solves the above problems by acquiring position information of feature points of an object indicating a lane boundary of a road on which a vehicle is traveling, representing the road, and detecting a lane boundary from the relative position of the feature points with respect to a reference line whose tangent slope changes continuously.

Advantages of the Invention

[0007] According to the present invention, a lane boundary line can be efficiently detected.

Brief Description of the Drawings

[0008]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Modes for Carrying Out the Invention

[0009] Hereinafter, embodiments of the present invention will be described with reference to the drawings.

[0010] [Configuration of Driving Support System] FIG. 1 is a block diagram showing an example of an embodiment of a driving support system according to the present invention. The driving support system is a group of devices that execute driving support for a vehicle, and driving support includes driving the vehicle by autonomous driving control and presenting information to the driver of the vehicle to support the driver's driving operation. For example, the driving support system generates a driving route to a destination set by a passenger and autonomously drives the vehicle along the driving route. The driving support system may be an in-vehicle system, or a part of the constituent devices may be provided outside the vehicle.

[0011] Autonomous driving control refers to the autonomous control of the driving operations of a vehicle. The driving operations include all driving operations such as acceleration, deceleration, starting, stopping, and steering. The autonomous control of the driving operations is executed by a controller mounted on the vehicle using the vehicle's devices. The controller controls the driving operations within a predetermined range. For driving operations not controlled by the controller, manual operations by the driver are performed. When the vehicle is driven manually by the driver without relying on autonomous driving control, the controller does not execute the autonomous control of the driving operations, and the driving operations of the vehicle are controlled by the driver's operations.

[0012] As shown in FIG. 1, the driving assistance system 1 includes an imaging device 11, a map database 12, a vehicle control device 13, and a controller 20. These devices are connected by a CAN (Controller Area Network) or other in-vehicle LAN and can exchange information with each other. The devices provided outside the vehicle exchange information via a network such as the Internet or a LAN (local area network).

[0013] The imaging device 11 is a device that images objects around the vehicle and is a camera equipped with an imaging element such as a CCD, or a camera such as an infrared camera. The objects are the road and the objects existing around it, including lane boundary lines, center lines, road surface markings, median strips, guardrails, curbs, road signs, traffic signals, etc. In addition, the objects also include obstacles that can affect the driving of the vehicle, and examples of the obstacles include other vehicles, motorcycles, bicycles, pedestrians, etc. A plurality of imaging devices 11 are installed on the front grille, door mirrors, rear bumper, etc. of the vehicle to reduce the blind spots in the imaging range. The controller 20 acquires image information from the imaging device 11 at predetermined time intervals (for example, every 0.1 to 1 millisecond).

[0014] The map database 12 is a storage medium storing map information and is provided inside or outside the vehicle. The map information is information used for generating a driving route and includes information on nodes corresponding to points (such as intersections and branch points) on a road where the traveling direction of the vehicle changes, and information on links corresponding to road sections connecting the nodes. Examples of the node information include position information (such as latitude and longitude) and information regarding entry and exit at intersections and branch points. Examples of the link information include road width, radius of curvature of the road, and road traffic regulations. The controller 20 acquires map information from the map database 12 as necessary. Note that the map information may be high-precision map information capable of grasping the movement locus of each lane.

[0015] The vehicle control device 13 is an in-vehicle computer such as an electronic control unit (ECU: Electronic Control Unit) and electronically controls in-vehicle devices that regulate the driving of the vehicle. The vehicle control device 13 autonomously controls the operations of the driving device and the steering device of the vehicle according to a control signal input from the controller 20.

[0016] The controller 20 is a device that controls and coordinates the devices constituting the driving support system 1 to execute driving support. The controller 20 is, for example, a computer and includes a CPU (Central Processing Unit) as a processor, a ROM (Read Only Memory) storing a program, and a RAM (Random Access Memory) functioning as an accessible storage device. The CPU executes the program stored in the ROM and is an operation circuit for realizing the functions of the controller 20.

[0017] [Functions of the Controller] In the ROM of the controller 20, a program for vehicle driving support is stored, and the CPU of the controller 20 executes the program to perform driving support. In FIG. 1, as functional blocks for performing driving support, a feature acquisition unit 21, a boundary detection unit 22, a trajectory generation unit 23, and a travel control unit 24 are extracted for convenience. Note that the controller 20 includes, as a part thereof, a lane boundary line detection device including the feature acquisition unit 21 and the boundary detection unit 22.

[0018] The feature acquisition unit 21 acquires information on feature points indicating an object from the image acquired by the imaging device 11. A feature point is a characteristic part of an object used for detecting the object, and examples thereof include an edge part (edge) that captures the contour of the object, a corner part (corner) of the object, and a part where the luminance changes in the object. To acquire feature points, a known algorithm such as SIFT (Scale-Invariant Feature Transform) is used.

[0019] The boundary detection unit 22 detects the lane boundary lines of the road on which the vehicle V travels from the information on the feature points acquired by the feature acquisition unit 21, the map information acquired from the map database 12, and the like. A lane boundary line is a road surface marking and a structure that define a lane, and includes a white line, a curb, a median strip, and the like.

[0020] The trajectory generation unit 23 generates a travel trajectory to be followed when the vehicle V travels by autonomous driving control. A travel trajectory is a passage (route) of the vehicle, and is generated based on the detection result of a vehicle detection device (including the imaging device 11), the map information acquired from the map database 12, the travel route generated by a navigation device (not shown), and the like.

[0021] The travel control unit 24 generates a control signal for the vehicle to travel along the travel trajectory. The vehicle control device 13 controls the drive device and the steering device according to the control signal, and thereby the vehicle autonomously travels along the lane.

[0022] FIG. 2 is a plan view showing an example of a driving scene in which the driving support system 1 executes driving support. The road shown in FIG. 2 has a lane W1 defined by white lines L1 and L2, and a lane W2 defined by white lines L2 and L3. It is assumed that the vehicle V travels from the left side to the right side (in the positive X-axis direction) along the lane W1.

[0023] In the scene shown in FIG. 2, the boundary detection unit 22 detects the white lines L1 and L2 that define the lane W1, and the trajectory generation unit 23 generates a driving trajectory D along which the vehicle V travels in the lane W1. When the information on the white lines L1 and L2 is not included in the map information, the boundary detection unit 22 detects the white lines L1 and L2 from the information on the feature points acquired by the feature acquisition unit 21 and recognizes the lane W1.

[0024] In this case, as the vehicle speed of the vehicle V increases, it is necessary to execute the recognition of the lane W1 and the generation of the driving trajectory D in a short time. Depending on the processing capacity of the controller 20, there is a possibility that stable autonomous driving control cannot be executed. Therefore, the boundary detection unit 22 realizes efficient detection of the lane boundary line and stable autonomous driving control by detecting the lane boundary line from the relative position of the feature points with respect to a reference line that represents the road and whose tangent slope changes continuously.

[0025] The reference line is, for example, a line along the road that shows the shape of the road corresponding to the road section between the nodes on the map. Also, when viewed in plan, the tangent slope continuously changes at each position on the reference line (specifically, the portion corresponding to the road on which the vehicle travels). That is, the tangent slope is continuous at each position on the reference line. As an example, the reference line is located on the center line in the width direction of the road, and one reference line is set for each road section between the nodes. The information on the reference line is included in the link information (map information).

[0026] Nodes are arranged at predetermined intervals (for example, every 1 to 3 m) on the reference line. For example, when a plurality of nodes are arranged on a map along a road, the boundary detection unit 22 performs polynomial interpolation (for example, the spline method) on the plurality of nodes to generate a reference line. The nodes are arranged based on, for example, information on the travel history when the vehicle and / or another vehicle other than the vehicle travels. Further, the boundary detection unit 22 may generate a reference line based on the travel trajectory of the vehicle and / or another vehicle other than the vehicle.

[0027] The nodes on the reference line have position information. Examples of the position information include information on coordinates in the global coordinate system used in map information, information on the angle formed by a predetermined direction and the direction (travel direction) of the reference line, and information on the travel distance from one node to another on the reference line. For example, as one node that is the starting point, a node closest to the starting point of the travel route is set.

[0028] Examples of the predetermined direction include the north direction of the map and the direction along the axis of the global coordinate system. The direction of the reference line is, for example, the tangent direction of the reference line at the nodes on the reference line. The travel distance between nodes is calculated as the travel distance when moving from one node to another along the reference line.

[0029] FIG. 3 shows an example of a reference line set for the road shown in FIG. 2. The reference line Z shown in FIG. 3 is a line along the road (lanes W1, W2) shown in FIG. 2 and shows the curvature of the lane W1. Nodes X1 to X13 are provided on the reference line Z, and each of the nodes X1 to X13 has information on coordinates in the global coordinate system, information on the angle (tangent inclination) formed by the X-axis direction and the tangent direction of the reference line Z, and information on the travel distance from the node X1 which is the starting point. Between the nodes X1 to X13, the inclination of the tangent is continuous.

[0030] The relative position of the feature point with respect to the reference line Z is not particularly limited as long as it is a position based on the reference line Z. For example, the position of the feature point in the local coordinate system along the reference line Z may be used as the relative position. The local coordinate system is a coordinate system (hereinafter also referred to as the reference line coordinate system) that represents the position of the feature point using the length (travel distance) of the reference line Z from a predetermined first position on the reference line Z to the second position on the reference line Z that is closest to the feature point, and the distance from the second position to the feature point. The first position is, for example, the position of a certain node that is the starting point, and the second position is, for example, the position of the intersection of the normal line of the reference line Z passing through the feature point and the reference line Z. Also, the local coordinate system may be a rectangular coordinate system composed of the tangent direction and the normal direction at a certain position on the reference line Z.

[0031] When detecting the lane boundary line from the relative position of the feature point with respect to the reference line Z, the feature acquisition unit 21 acquires, for example, the position information of the feature points of the object indicating the lane boundary line of the road on which the vehicle V is traveling from the image acquired by the imaging device 11 mounted on the vehicle V. The object indicating the lane boundary line is a road surface marking and a structure that define the lane, and the information of the feature points of the object is, for example, the position information of the edge of the white line, the information of the length in the traveling direction of the curbstone, etc.

[0032] In the driving scene shown in FIG. 2, as shown in FIG. 3, the feature points A1 to A5 corresponding to the lane L1 and the feature points B1 to B5 corresponding to the lane L2 are acquired. The position information of the feature points A1 to A5, B1 to B5 is, for example, the information of the coordinates of the feature points, and is represented by the global coordinate system used in the map information. The global coordinate system is not particularly limited, but is, for example, a rectangular coordinate system composed of the Y axis along the north direction of the map and the X axis orthogonal to the Y axis, as shown in FIGS. 2 and 3.

[0033] When the feature acquisition unit 21 outputs information on feature points, the boundary detection unit 22 converts the coordinate system of the position information from the global coordinate system used for map information to the local coordinate system along the reference line Z in order to detect the lane boundary line from the position information in the local coordinate system along the reference line Z. The boundary detection unit 22 generates a first approximation line that approximates the shape of the lane boundary line based on the position information in the local coordinate system and the slope of the tangent line of the reference line Z. Then, the coordinate system of the first approximation line is converted from the local coordinate system to the global coordinate system, and the lane boundary line is detected based on the first approximation line in the global coordinate system. The coordinate conversion between the global coordinate system and the local coordinate system is performed by a known coordinate conversion method using formulas such as the Frenet-Serret formulas.

[0034] As an example, in the driving scene shown in FIG. 2, the boundary detection unit 22 clusters the feature points A1 to A5 and B1 to B5 represented in the local coordinate system and classifies them into a cluster of feature points A1 to A5 and a cluster of feature points B1 to B5. When using the reference line coordinate system, clustering can be performed relatively easily based on the distance from the reference line Z. Then, polynomial interpolation such as the spline method is performed for each cluster to generate an approximation line C1 corresponding to the lane L1 and an approximation line C2 corresponding to the lane L2. In polynomial interpolation, the slope of the tangent line of the reference line Z and the slopes of the tangent lines of the approximation lines C1 and C2 may be set to the same value. In this case, the approximation lines C1 and C2 correspond to the first approximation line. The boundary detection unit 22 converts the coordinate systems of the approximation lines C1 and C2 from the local coordinate system to the global coordinate system, and detects the lanes L1 and L2 from the approximation lines C1 and C2 in the global coordinate system.

[0035] When representing the relative positions of the feature points by the reference line coordinate system, the boundary detection unit 22 may specify a target section that includes the second position among the sections delimited by two adjacent nodes arranged on the reference line Z, and perform numerical analysis on the target section to calculate the second position.

[0036] For example, when converting the position of a feature point represented as coordinates (x p , y p ) in the global coordinate system to the reference line coordinate system, on the reference line Z, the coordinates (xp , y p ), the position closest to it is the coordinate (x i , y i ). Assuming that it is), the target interval is the interval [s, s i , s i+1 or the interval [s i-1 , s i , and the inner product value F(s) using the interpolation function and its derivative of each interval is expressed as in the following equations (1) and (2).

[0037]

Equation

[0038] F i-1 (s) = 0 or F i (s) = 0, and the path length s at which this occurs corresponds to the second position. However, when using a cubic function for polynomial interpolation, F(s) becomes a fifth-degree equation with respect to s, and the solution cannot be obtained analytically. Therefore, from the intermediate value theorem, the condition for the existence of a solution in the interval [s a , s b is F(s a ) × F(s b ) ≤ 0. Using this, the existence of a solution is determined at the start and end points of each interval, and then the solution is numerically obtained by the ordinary bisection method.

[0039] As an example, when calculating the second position corresponding to the feature point B3 shown in FIG. 3, the existence of a solution is determined at the start and end points of each interval for each of the interval delimited by the nodes X5, X6 and the interval delimited by the nodes X6, X7, and numerical analysis by the bisection method is executed. As a result, the intersection point Y of the perpendicular line T passing through the feature point B3 and the reference line Z is calculated as the second position.

[0040] The boundary detection unit 22 determines whether the feature point exists on the same side of the vehicle V as the reference line Z. When it is determined that the feature point exists on the same side of the vehicle V as the reference line Z, the distance from the second position to the feature point may be represented by a positive value. On the other hand, when it is determined that the feature point does not exist on the same side of the vehicle as the reference line Z, the boundary detection unit 22 represents the distance from the second position to the feature point by a negative value. Thereby, the position of the feature point with respect to the reference line Z can be easily recognized. As an example, since the feature point B3 shown in FIG. 3 exists on the same side of the vehicle V as the reference line Z, the distance between the second position (intersection point Y) and the feature point B3 is represented by a positive value.

[0041] When the boundary detection unit 22 detects the lane boundary line from the relative position of the feature point with respect to the reference line Z, the road on which the vehicle V travels is divided into a driving section where the vehicle V exists, a front section adjacent to the driving section and existing in front of the traveling direction of the vehicle V, and a rear section adjacent to the driving section and existing behind the traveling direction of the vehicle V. In each section, a second approximation line approximating the shape of the lane boundary line may be generated, and the lane boundary line may be detected from the second approximation line.

[0042] For example, the boundary detection unit 22 sets a front section S1, a driving section S2, and a rear section S3 shown in FIG. 3 for the road shown in FIG. 2. For setting the driving section S2, for example, the current position information of the vehicle V acquired from a positioning system (not shown) such as GPS and map information are used. The length of each section can be set to an appropriate value within a range where the lane boundary line can be appropriately detected, for example, 10 to 100 m.

[0043] When generating the second approximation line, the boundary detection unit 22 may consider that at the boundary U1 between the driving section S2 and the front section S1 and the boundary U2 between the driving section S2 and the rear section S3, the second approximation lines of two adjacent sections have the same value, the slope of the tangent line is continuous, and the curvature is continuous. Further, when the relative position of the feature point is represented by the reference line coordinate system, the boundary detection unit 22 may consider that in each of the driving section S2, the front section S1, and the rear section S3, the slope of the tangent line of the reference line Z at the second position coincides with the slope of the tangent line of the second approximation line at the position corresponding to the second position. The position corresponding to the second position is, for example, the position of the feature point closest to the second position on the second approximation line.

[0044] For example, the reference line Z is a two-dimensional curve specified with the path length s as a mediating variable

Number

Number

Number

[0045] In the above formulas (3) to (5), w is the width of the lanes W1, W2, and α i , β i , γ i , δ i are respectively the coefficients of the polynomials. The first derivative and the second derivative of each approximation line are commonly represented as in the following formulas (6) and (7).

Number

[0046] The first derivative of the normal component with respect to the path length s in the reference line coordinate system is expressed as the following equation (8).

Number

[0047] In the above equation (8), k r (s) is the curvature of the reference line Z, and Δθ(s) is the difference in the inclination between the tangent of the lane and the reference line Z. When it is considered that the inclination of the tangent of the reference line Z at the second position coincides with the inclination of the tangent of the second approximation line at the position corresponding to the second position, when a coordinate value s = s * of the detection section of the lane boundary line is sampled, the following equation (9) holds.

Number

[0048] Regarding the total number N of sections to which the runway boundary estimation is applied, for the i-th section [s i , s i+1 , the position coordinates a(i) of the N

Number

Number

Number

Number

[0049] For the section [s i , s i+1and the interval [s i+1 , s i+2 , the boundary point s i+1 On this, since the second approximation line has the same value, the slope of the tangent line is continuous, and the curvature is continuous, the connection conditions according to the following equations (11) to (13) are satisfied.

Number

[0050] The least - squares estimation problem with linear equality constraints incorporating the connection conditions is formulated as a Lagrange function as shown in the following equation (14).

Number

[0051] Since the parameter estimation problem by the Lagrange function of equation (14) is a least - squares method with linear constraint conditions, it can be rewritten as shown in the following equation (15).

Number

[0052] The optimization problem of equation (15) is based on the stationary - point conditions for the Lagrange function shown below

Number

Number

[0053] The equation of equation (16) can be solved for the vector

Number

Number

[0054] Thus, using the relational expression between the global coordinate system and the local coordinate system, the position of the center line of the lane, the left lane, and the right lane that follow the approximate line of the parameter vector and the local coordinate system can be obtained.

[0055] As an example, when the above-described operation is performed on the feature points A1 to A5, B1 to B5 shown in FIG. 3 and the reference line Z, in the driving section S2, an approximate line C0 of the center line of the lane, an approximate line C1 of the left lane, and an approximate line C2 of the right lane are generated. Also, in the front section S1, an approximate line C0a of the center line of the lane, an approximate line C1a of the left lane, and an approximate line C2a of the right lane are generated, and in the rear section S3, an approximate line C0b of the center line of the lane, an approximate line C1b of the left lane, and an approximate line C2b of the right lane are generated. In this case, the approximate lines C0, C1, C2, C0a, C1a, C2a, C0b, C1b, C2b correspond to the second approximate lines.

[0056] In this way, by setting the conditions for generating the second approximate line, the position (shape) of the lane boundary line can be estimated in the front section S1 of the driving section S2 in which the vehicle V travels. That is, since the position of the lane boundary line can be estimated also for the portion of the lane W1 that is not included in the imaging device 11, the distance of the travel trajectory D generated by the trajectory generation unit 23 is extended, and stable autonomous driving control can be realized.

[0057] [Processing in the Driving Support System] With reference to FIGS. 4 to 6, the procedure when the controller 20 processes information will be described. The processing described below is executed by a processor (CPU) provided in the controller 20 at predetermined time intervals (for example, every 0.1 to 1 millisecond).

[0058] FIG. 4 is a flowchart showing an example of a processing procedure executed in the driving support system 1. First, in step S1, the feature acquisition unit 21 acquires the position information of the feature points of the object indicating the lane boundary line from the image acquired from the imaging device 11. In step S2, the boundary detection unit 22 calculates the relative position of the feature points with respect to the reference line Z, and in the subsequent step S3, the lane boundary line is detected from the relative position of the feature points.

[0059] Next, FIG. 5 is a flowchart showing another example of the processing procedure executed in the driving support system 1. First, in step S11, the feature acquisition unit 21 acquires an image in front of the vehicle V from the imaging device 11, and in the subsequent step S12, the position information of the feature points of the object indicating the lane boundary line is acquired from the acquired image. In step S13, the boundary detection unit 22 acquires the information of the reference line Z from the map database 12, and in the subsequent step S14, the relative position of the feature points with respect to the reference line Z is calculated, and in the subsequent step S15, the lane boundary line is detected from the relative position of the feature points. In step S16, the trajectory generation unit 23 generates a travel trajectory D along which the vehicle V travels in the lane, and in step S17, the travel control unit 24 autonomously drives the vehicle V along the travel trajectory D.

[0060] FIG. 6 is a flowchart showing an example of a subroutine of step S15 in FIG. 5. First, in step S21, the boundary detection unit 22 converts the coordinate system of the position information of the feature points from the global coordinate system to the local coordinate system, and in the subsequent step S22, an approximate line of the lane boundary line in the local coordinate system is generated. In the subsequent step S23, the boundary detection unit 22 converts the coordinate system of the generated approximate line from the local coordinate system to the global coordinate system, and in the subsequent step S24, the lane boundary line is detected from the approximate line converted to the global coordinate system.

[0061] [Embodiments of the Present Invention] According to the present embodiment, in a lane boundary line detection method executed by a controller 20 of a vehicle V, the controller 20 acquires position information of feature points of an object indicating a lane boundary line of a road on which the vehicle V is traveling, and detects the lane boundary line from the relative position of the feature points with respect to a reference line Z that represents the road and whose tangent slope changes continuously. Thus, the lane boundary line can be detected efficiently.

[0062] In the lane boundary line detection method of the present embodiment, when the controller 20 acquires the position information in a global coordinate system used for map information from an image acquired by an imaging device 11 mounted on the vehicle V, and detects the lane boundary line from the position information in a local coordinate system along the reference line Z, the controller 20 converts the coordinate system of the position information from the global coordinate system to the local coordinate system, generates a first approximation line that approximates the shape of the lane boundary line based on the position information in the local coordinate system and the tangent slope of the reference line Z, and detects the lane boundary line based on the first approximation line in the global coordinate system. Thus, the lane boundary line can be detected more efficiently.

[0063] In the lane boundary line detection method of the present embodiment, when the controller 20 represents the relative position using the length of the reference line Z from a predetermined first position on the reference line Z to a second position closest to the feature point on the reference line Z and the distance from the second position to the feature point, the controller 20 specifies a target section including the second position among sections delimited by two adjacent nodes arranged on the reference line Z, and executes numerical analysis on the target section to calculate the second position. Thus, the path length in the reference line coordinate system can be accurately calculated.

[0064] In the lane boundary line detection method of the present embodiment, the controller 20 determines whether the feature point exists on the same side of the vehicle V as the reference line Z. When it is determined that the feature point exists on the same side of the vehicle V as the reference line Z, the distance from the second position to the feature point is represented by a positive value. When it is determined that the feature point does not exist on the same side of the vehicle V as the reference line Z, the distance from the second position to the feature point is represented by a negative value. Thereby, the position of the feature point with respect to the reference line Z can be easily recognized.

[0065] In the lane boundary line detection method of the present embodiment, the reference line Z is a line generated by polynomial interpolation for a plurality of nodes arranged on a map. Thereby, the relative position of the feature point in the local coordinate system along the reference line Z can be calculated smoothly.

[0066] In the lane boundary line detection method of the present embodiment, the information of the reference line Z is included in the map information. Thereby, accurate information of the reference line Z can be used.

[0067] In the lane boundary line detection method of the present embodiment, the reference line Z is a line generated based on the information of the travel history when the vehicle V and / or other vehicles other than the vehicle V travel. Thereby, the reference line Z can be generated without relying on map information.

[0068] In the lane boundary line detection method of the present embodiment, when the controller 20 detects the lane boundary line from the relative position, the road is divided into a travel section where the vehicle V exists, a front section adjacent to the travel section and existing in front of the traveling direction of the vehicle V, and a rear section adjacent to the travel section and existing behind the traveling direction of the vehicle V. In each of the travel section, the front section, and the rear section, a second approximation line approximating the shape of the lane boundary line is generated, and the lane boundary line is detected from the second approximation line. Thereby, the position (shape) of the lane boundary line can be estimated in the front section S1 of the travel section S2 where the vehicle V travels.

[0069] In the lane boundary line detection method of the present embodiment, when the controller 20 generates the second approximate line, at the boundary between the driving section and the front section and the boundary between the driving section and the rear section, it is considered that the second approximate lines of two adjacent sections have the same value, the slope of the tangent line is continuous, and the curvature is continuous. Thereby, the position (shape) of the lane boundary line can be estimated in the front section S1 of the driving section S2 where the vehicle V travels.

[0070] In the lane boundary line detection method of the present embodiment, when the controller 20 represents the relative position using the length of the reference line Z from a predetermined first position on the reference line Z to a second position closest to the feature point on the reference line Z and the distance from the second position to the feature point, in each of the driving section, the front section, and the rear section, it is considered that the slope of the tangent line of the reference line Z at the second position coincides with the slope of the tangent line of the second approximate line at the position corresponding to the second position. Thereby, the position (shape) of the lane boundary line can be estimated in the front section S1 of the driving section S2 where the vehicle V travels.

[0071] Further, according to the present embodiment, there is provided a lane boundary line detection device including a feature acquisition unit 21 that acquires position information of a feature point of an object indicating a lane boundary line of a road on which the vehicle V is traveling, and a boundary detection unit 22 that detects the lane boundary line from the relative position of the feature point with respect to a reference line Z representing the road and having a continuously changing tangent slope. Thereby, the lane boundary line can be detected efficiently.

Explanation of Reference Numerals

[0072] 1... Driving support system 11... Imaging device, 12... Map database, 13... Vehicle control device 20... Controller, 21... Feature acquisition unit, 22... Boundary detection unit, 23... Trajectory generation unit, 24... Driving control unit A1, A2, A3, A4, A5, A6, B1, B2, B3, B4, B5, B6... characteristic points, C0, C1, C2... approximate lines, D... travel track, L1, L2, L3... white lines, S1... rear section, S2... travel section, S3... front section, T... vertical line, U1, U2... boundaries, V... vehicle, W1, W2... lanes, X1, X2, X3, X4, X5, X6, X7, X8, X9, X10, X11, X12, X13... nodes, Y... intersection point, Z... reference line

Claims

1. In a lane boundary detection method executed by a vehicle controller, the controller: acquires position information of feature points of an object indicating a lane boundary of a road on which the vehicle is traveling; and detects the lane boundary from a relative position of the feature points with respect to a reference line representing the road and having a continuously changing tangent slope. A lane boundary detection method.

2. The controller: acquires the position information in a global coordinate system used for map information from an image acquired by an imaging device mounted on the vehicle; when detecting the lane boundary from the position information in a local coordinate system along the reference line, converts the coordinate system of the position information from the global coordinate system to the local coordinate system; generates a first approximation line approximating the shape of the lane boundary based on the position information in the local coordinate system and the tangent slope of the reference line; and detects the lane boundary based on the first approximation line in the global coordinate system. The lane boundary detection method according to claim 1.

3. The controller: when representing the relative position using the length of the reference line from a predetermined first position on the reference line to a second position closest to the feature point on the reference line and the distance from the second position to the feature point, identifies a target section including the second position among sections delimited by two adjacent nodes arranged on the reference line; and performs numerical analysis on the target section to calculate the second position. The lane boundary detection method according to claim 1 or 2.

4. The controller: determines whether the feature point exists on the same side of the vehicle as the reference line; when it is determined that the feature point exists on the same side of the vehicle as the reference line, represents the distance from the second position to the feature point as a positive value, When it is determined that the feature point does not exist on the same side of the vehicle as the reference line, the method for detecting a lane boundary line according to claim 3, wherein the distance from the second position to the feature point is represented by a negative value. **Claim 5** The method for detecting a lane boundary line according to claim 1 or 2, wherein the reference line is a line generated by polynomial interpolation for a plurality of nodes arranged on a map. **Claim 6** The method for detecting a lane boundary line according to claim 1 or 2, wherein the information of the reference line is included in map information. **Claim 7** The method for detecting a lane boundary line according to claim 1 or 2, wherein the reference line is a line generated based on information on a driving history when the vehicle and / or another vehicle other than the vehicle travels. **Claim 8** The controller When detecting the lane boundary line from the relative position, the road is divided into a driving section where the vehicle exists, a front section adjacent to the driving section and existing in front of the traveling direction of the vehicle, and a rear section adjacent to the driving section and existing behind the traveling direction of the vehicle. In each of the driving section, the front section, and the rear section, a second approximation line approximating the shape of the lane boundary line is generated. The method for detecting a lane boundary line according to claim 1 or 2, wherein the lane boundary line is detected from the second approximation line. **Claim 9** The controller When generating the second approximation line, at the boundary between the driving section and the front section and at the boundary between the driving section and the rear section, it is considered that the second approximation lines of two adjacent sections have the same value, the slope of the tangent line is continuous, and the curvature is continuous. The method for detecting a lane boundary line according to claim 8. **Claim 10** When the controller represents the relative position by using the length of the reference line from a predetermined first position on the reference line to a second position closest to the feature point on the reference line and the distance from the second position to the feature point, in each of the driving section, the front section, and the rear section, the slope of the tangent line of the reference line at the second position is considered to be the same as the slope of the tangent line of the second approximate line at the position corresponding to the second position. The lane boundary line detection method according to claim 8.

11. A feature acquisition unit that acquires position information of a feature point of an object indicating a lane boundary line of a road on which a vehicle is traveling; A lane boundary line detection device including a boundary detection unit that represents the road and detects the lane boundary line from the relative position of the feature point with respect to a reference line whose tangent slope changes continuously.

Images