KR102674613B1 - Vehicle position correction method and vehicle position correction device - Google Patents

Abstract

In the lane determination method and lane determination device, the navigation position of the vehicle is obtained, lane information adjacent to the navigation position is extracted from the high-precision map data, and a plurality of lanes within the position error range of the navigation position are based on the navigation position, lane information, and camera parameters. A plurality of virtual lane images are generated at vehicle locations, an image in front of the vehicle is captured, a lane detection image is generated by detecting a lane in the front image, and matching the lane detection image with the plurality of virtual lane images By performing, a virtual lane image matching the lane detection image is selected from among the plurality of virtual lane images, the navigation position is corrected according to the vehicle position corresponding to the selected virtual lane image among the plurality of vehicle positions, and the corrected navigation position is performed. Based on this, the current lane of the vehicle is determined.

Description

Lane determination method and lane determination device {VEHICLE POSITION CORRECTION METHOD AND VEHICLE POSITION CORRECTION DEVICE}

The present invention relates to a method and device for a vehicle, and more specifically, to a navigation device, a lane determination method and a lane determination device using high-precision map data and cameras.

Technology for estimating the location of a vehicle is widely used to provide route guidance to vehicles. For example, Global Positioning System (GPS) technology and Dead Reckoning (DR) GPS technology are being used. However, GPS location using GPS technology may have location errors due to factors such as being affected by noise and interference from the surrounding environment or multiple paths occurring due to surrounding buildings. Accordingly, this GPS technology, which is currently widely used, has a problem in that it cannot obtain the precise location of the vehicle.

In particular, in order to perform autonomous driving, the precise location of the vehicle must be obtained, and it must also be accurately determined which lane the vehicle is in on the current road using a high definition (HD) map. However, when acquiring the location of a vehicle using the above-described GPS technology, there is a problem that the current lane of the vehicle may be incorrectly determined.

One object of the present invention is to provide a lane determination method that can accurately determine the lane of a vehicle.

Another object of the present invention is to provide a lane determination device that can accurately determine the lane of a vehicle.

However, the purpose of the present invention is not limited to the above-mentioned purposes, and may be expanded in various ways without departing from the spirit and scope of the present invention.

In order to achieve an object of the present invention, in the method for determining the lane of a vehicle according to embodiments of the present invention, the navigation device acquires the navigation position of the vehicle, and the control unit acquires the navigation position adjacent to the navigation position from high-precision map data. Lane information is extracted, and the control unit extracts a plurality of virtual lane images by the line of sight of the camera at a plurality of vehicle positions within a position error range of the navigation position based on the navigation position, the lane information, and camera parameters of the camera. The camera captures a front image of the vehicle, the control unit detects a lane in the front image to generate a lane detection image, and the control unit captures the lane detection image and the plurality of virtual lane images. performs matching to select a virtual lane image that matches the lane detection image among the plurality of virtual lane images, and the control unit selects the virtual lane image according to the vehicle position corresponding to the selected virtual lane image among the plurality of vehicle positions. The navigation position is corrected, and the control unit determines the current lane of the vehicle based on the corrected navigation position.

In one embodiment, the camera parameters may include intrinsic parameters and extrinsic parameters of the camera.

In one embodiment, the internal product parameters include a focal length, principal point coordinates, and asymmetry coefficient of the camera, and the external product parameters may include a rotation matrix value and a relative position of the camera. .

In one embodiment, the control unit generates a virtual camera model using the camera parameters to generate the plurality of virtual lane images at the plurality of vehicle locations, and the control unit generates a virtual camera model in the virtual camera model to generate the plurality of virtual lane images at the plurality of vehicle locations. Each of the plurality of virtual lane images at each of the plurality of vehicle locations can be generated while changing the relative position.

In one embodiment, the control unit converts each lane point coordinate of the lane information into virtual lane point coordinates using the virtual camera model to generate each of the plurality of virtual lane images at each of the plurality of vehicle locations. After converting, the control unit may generate each of the plurality of virtual lane images including a virtual lane point from the virtual lane point coordinates. The relative position of the camera is changed to correspond to each of the plurality of vehicle positions until the plurality of virtual lane images at the plurality of vehicle positions within the position error range of the navigation position are generated. Each of the virtual lane images may be generated.

In one embodiment, the control unit uses the equation " Convert each lane point coordinate of the lane information into the virtual lane point coordinates using ", where s represents a scale factor, x and y multiplied by the scale factor represent the virtual lane point coordinates, f _x and f _y represent the focal length of _the camera, c _x and c _y represent _the principal point coordinates _of the camera, skew_cf values, t ₁ to t ₃ may represent the relative position of the camera, and X, Y, and Z may represent coordinates of each lane point of the lane information.

In one embodiment, the controller may change the relative position of the camera by 0.1 m to generate the plurality of virtual lane images at the plurality of vehicle positions with an interval of 0.1 m.

In one embodiment, the control unit selects the virtual lane image that matches the lane detection image among the plurality of virtual lane images, normalized mutual information between the lane detection image and the plurality of virtual lane images ( Normalized Mutual Information (NMI) values are calculated, and the control unit may select the virtual lane image with the largest NMI value among the NMI values.

In one embodiment, the control unit may determine the corrected navigation position as the vehicle position corresponding to the selected virtual lane image to correct the navigation position according to the vehicle position corresponding to the selected virtual lane image.

In one embodiment, the control unit determines a lane center line adjacent to the corrected navigation position among a plurality of lane center lines included in the high-precision map data to determine the current lane of the vehicle based on the corrected navigation position. And, the control unit may determine the current lane of the vehicle according to the determined lane center line.

In order to achieve an object of the present invention, a lane determination device for determining the lane of a vehicle according to embodiments of the present invention includes a navigation device for acquiring the navigation position of the vehicle, a camera for capturing a front image of the vehicle, and a high-precision It includes a high-precision map storage unit that stores map data, and a control unit that controls the operation of the lane determination device. The control unit extracts lane information adjacent to the navigation position from the high-precision map data, and selects a plurality of vehicle positions within a position error range of the navigation position based on the navigation position, the lane information, and camera parameters of the camera. Generate a plurality of virtual lane images by the line of sight of the camera, generate a lane detection image by detecting a lane in the front image, and perform matching between the lane detection image and the plurality of virtual lane images. Selecting a virtual lane image that matches the lane detection image among a plurality of virtual lane images, correcting the navigation position according to a vehicle position corresponding to the selected virtual lane image among the plurality of vehicle positions, and Determine the current lane of the vehicle based on the navigation position.

In one embodiment, the camera parameters may include intrinsic parameters and extrinsic parameters of the camera.

In one embodiment, the internal product parameters include a focal length, principal point coordinates, and asymmetry coefficient of the camera, and the external product parameters may include a rotation matrix value and a relative position of the camera. .

In one embodiment, the control unit generates a virtual camera model using the camera parameters, and changes the relative position of the camera in the virtual camera model while changing the plurality of virtual lane images at each of the plurality of vehicle positions. Each of them can be created.

In one embodiment, the control unit converts each lane point coordinate of the lane information into virtual lane point coordinates using the virtual camera model, and the plurality of virtual lane images including virtual lane points in the virtual lane point coordinates. each of which is generated, and the control unit determines the relative position of the camera to the plurality of vehicles until the plurality of virtual lane images are generated at the plurality of vehicle positions within the position error range of the navigation position. Each of the plurality of virtual lane images can be generated while changing the positions to correspond to each other.

In one embodiment, the control unit uses the equation " Convert each lane point coordinate of the lane information into the virtual lane point coordinates using ", where s represents a scale factor, x and y multiplied by the scale factor represent the virtual lane point coordinates, f _x and f _y represent the focal length of _the camera, c _x and c _y represent _the principal point coordinates _of the camera, skew_cf values, t ₁ to t ₃ may represent the relative position of the camera, and X, Y, and Z may represent coordinates of each lane point of the lane information.

In one embodiment, the controller may change the relative position of the camera by 0.1 m to generate the plurality of virtual lane images at the plurality of vehicle positions with an interval of 0.1 m.

In one embodiment, the control unit calculates normalized mutual information (NMI) values between the lane detection image and the plurality of virtual lane images, and calculates the normalized mutual information (NMI) values between the lane detection images and the plurality of virtual lane images, and You can select a virtual lane image.

In one embodiment, the controller may determine the corrected navigation position as the vehicle position corresponding to the selected virtual lane image.

In one embodiment, the control unit may determine a lane center line adjacent to the corrected navigation position among a plurality of lane center lines included in the high-precision map data, and determine the current lane of the vehicle according to the determined lane center line. there is.

In the lane determination method and lane determination device according to embodiments of the present invention, lane information adjacent to the navigation position is extracted from high-precision map data, and the navigation position is determined based on the navigation position, the lane information, and the camera parameters of the camera. A plurality of virtual lane images are generated at a plurality of vehicle positions within the position error range, a lane detection image is generated by detecting a lane in the front image, and the lane detection image and the plurality of virtual lane images are matched. The navigation position is corrected, and the current lane of the vehicle can be determined based on the corrected navigation position. Accordingly, the precise location of the vehicle can be obtained, and the current lane of the vehicle can be accurately determined.

However, the effects of the present invention are not limited to the effects described above, and may be expanded in various ways without departing from the spirit and scope of the present invention.

1 is a flowchart showing a method for determining a lane according to embodiments of the present invention.
Figure 2 is a diagram showing an example of lane information adjacent to a navigation position.
Figure 3 is a diagram showing an example of a virtual camera model.
FIG. 4 is a diagram illustrating an example of multiple virtual lane images at multiple vehicle locations.
Figure 5 is a diagram for explaining an example of generating a lane detection image by detecting a lane in a front image.
FIG. 6 is a diagram illustrating an example of correcting the navigation position by matching a lane detection image and a plurality of virtual lane images.
FIG. 7 is a diagram illustrating an example of calculating a normalized mutual information (NMI) value between a lane detection image and each virtual lane image.
FIG. 8 is a diagram illustrating an example of determining the current lane of a vehicle based on the corrected navigation position.
Figure 9 is a block diagram showing a lane determination device according to embodiments of the present invention.

Regarding the embodiments of the present invention disclosed in the text, specific structural and functional descriptions are merely illustrative for the purpose of explaining the embodiments of the present invention, and the embodiments of the present invention may be implemented in various forms. It should not be construed as limited to the embodiments described in.

Since the present invention can be subject to various changes and can have various forms, specific embodiments will be illustrated in the drawings and described in detail in the text. However, this is not intended to limit the present invention to a specific disclosed form, and should be understood to include all changes, equivalents, and substitutes included in the spirit and technical scope of the present invention.

Terms such as first, second, etc. may be used to describe various components, but the components should not be limited by the terms. The above terms may be used for the purpose of distinguishing one component from another component. For example, a first component may be referred to as a second component, and similarly, the second component may also be referred to as a first component without departing from the scope of the present invention.

When a component is said to be "connected" or "connected" to another component, it is understood that it may be directly connected to or connected to that other component, but that other components may also exist in between. It should be. On the other hand, when it is mentioned that a component is “directly connected” or “directly connected” to another component, it should be understood that there are no other components in between. Other expressions that describe the relationship between components, such as "between" and "immediately between" or "neighboring" and "directly adjacent to" should be interpreted similarly.

The terms used in this application are only used to describe specific embodiments and are not intended to limit the invention. Singular expressions include plural expressions unless the context clearly dictates otherwise. In this application, terms such as “comprise” or “have” are intended to designate the existence of a described feature, number, step, operation, component, part, or combination thereof, but are not intended to indicate the presence of one or more other features or numbers. It should be understood that this does not exclude in advance the possibility of the existence or addition of steps, operations, components, parts, or combinations thereof.

Additionally, terms such as “~unit” used in the text refer to a unit that processes at least one function or operation, and this may be implemented through hardware, software, or a combination of hardware and software.

Unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by a person of ordinary skill in the technical field to which the present invention pertains. Terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with the meaning in the context of the related technology, and unless clearly defined in the present application, should not be interpreted as having an ideal or excessively formal meaning. .

In addition, the terms used in the detailed description and claims of the present invention should be understood to include those commonly understood by those skilled in the art in the technical field to which the present invention pertains, as well as equivalents or substitutes thereof.

Hereinafter, embodiments of the present invention will be described in more detail with reference to the attached drawings. The same or similar reference numerals are used for identical components in the drawings.

FIG. 1 is a flowchart showing a method for determining a lane according to embodiments of the present invention, FIG. 2 is a diagram showing an example of lane information adjacent to a navigation position, and FIG. 3 is a diagram showing an example of a virtual camera model. FIG. 4 is a diagram illustrating an example of a plurality of virtual lane images at a plurality of vehicle locations, FIG. 5 is a diagram illustrating an example of generating a lane detection image by detecting a lane in a front image, and FIG. 6 is a diagram to explain an example of correcting the navigation position by matching the lane detection image and a plurality of virtual lane images, and Figure 7 shows the normalized mutual information between the lane detection image and each virtual lane image. This is a diagram for explaining an example of calculating a Mutual Information (NMI) value, and FIG. 8 is a diagram for explaining an example of determining the current lane of a vehicle based on the corrected navigation position.

Referring to FIG. 1, in the method for determining the lane of a vehicle according to embodiments of the present invention, a navigation device mounted on the vehicle may acquire the navigation position of the vehicle (S110). In one embodiment, the navigation device is a Global Navigation Satellite System (GNSS) that determines the three-dimensional position of the vehicle in real time using a triangulation method by measuring the satellite's orbit information and the arrival time difference of the received signal. and/or an inertial navigation system (INS) that determines a direction reference through a gyroscope and measures movement displacement using an accelerometer, and the navigation position is determined by the GNSS and/or the INS. It may be the location of .

The lane determination device for the vehicle includes a high-precision map storage unit that stores high-definition (HD) map data, and the control unit of the lane determination device can extract information on lanes adjacent to the navigation position from the high-precision map data. (S120). Figure 2 shows an example 200 of the lane information adjacent to the navigation location. As shown in FIG. 2, the lane information representing lanes indicated by white and yellow lines adjacent to the navigation position indicated by a red dot can be extracted from the high-precision map data.

The control unit determines a plurality of vehicle positions within a position error range of the navigation position based on the navigation position acquired by the navigation device, the lane information extracted from the high-precision map data, and the camera parameters of the camera mounted on the vehicle. A plurality of virtual lane images can be generated by the camera's line of sight (S130, S140). In one embodiment, the camera parameters may include intrinsic parameters and extrinsic parameters of the camera. For example, the intrinsic parameter is an intrinsic parameter for the camera, including the focal length of the camera, the principal point coordinates, and the asymmetry coefficient obtained by calibration of the camera. It can be included. Also, for example, the external parameter is an external parameter for the camera and may include a rotation matrix value determined by the posture of the camera and a relative position of the camera from the navigation position.

In one embodiment, the control unit generates a virtual camera model using the camera parameters (S130), and changes the relative position of the camera in the virtual camera model while changing the plurality of vehicle locations at each of the plurality of vehicle locations. Each of the virtual lane images can be generated (S140). Here, the virtual camera model may represent a mathematical equation for converting the coordinates of each lane point in the lane information into the coordinates of the virtual lane point in each virtual lane image.

For example, as shown in FIG. 3, the virtual camera model may be Equation 300, and the control unit may be Equation 300, that is, " " can be used to convert each lane point coordinate of the lane information into the virtual lane point coordinates. Here, s represents a scale factor, and x and y multiplied by the scale factor are the coordinates of each virtual lane point in each virtual lane image. represents the virtual lane point coordinates, f _x and f _y represent the focal distances (in the x- _axis and y-axis) of the camera, and _c represents the coordinates _of the principal point ₍ on _the axis ₎ , _{skew_cf} represents the relative position of the camera from the navigation position, and .

The control unit converts each lane point coordinate of the lane information into the virtual lane point coordinates using the virtual camera model, and generates each of the plurality of virtual lane images including a virtual lane point from the virtual lane point coordinates. can do. Additionally, the relative position of the camera is changed to correspond to each of the plurality of vehicle positions until the plurality of virtual lane images are generated at the plurality of vehicle positions within the position error range of the navigation position. Each of the plurality of virtual lane images may be generated. For example, as shown in FIG. 4, the control unit converts each lane point coordinate of the lane information extracted from the high-precision map data at the navigation location into the virtual lane point coordinates using the virtual camera model. , a virtual lane image 410 at the navigation position can be generated as the vehicle position. In addition, when the position error range of the navigation position is about +2m to about -2m, the control unit changes the vehicle location in the lateral direction perpendicular to the traveling direction of the vehicle, that is, the camera A virtual lane image 420 is generated at each vehicle location while changing the relative position (for example, at least one of t1 to t3) at a certain interval from about 0m to about -2m, and the relative position of the camera A virtual lane image 430 at each vehicle location may be generated while changing (for example, at least one of t1 to t3) at regular intervals from about 0 m to about +2 m. In one embodiment, the control unit may change the relative position of the camera by about 0.1 m to generate a plurality of virtual lane images 400 at the plurality of vehicle positions with an interval of about 0.1 m. . Meanwhile, each of the plurality of virtual lane images 400 may be an image of lanes from the perspective of the camera at the corresponding vehicle location. For example, the virtual lane image 410 located in the middle of FIG. 4 may be an image of lanes viewed from the camera's line of sight when the vehicle is located at the navigation position. Additionally, the virtual lane image 420 located on the left in FIG. 4 may be an image of lanes viewed from the camera's line of sight when the vehicle is located 2 m to the left along the lateral direction from the navigation position. Additionally, the virtual lane image 430 located on the right in FIG. 4 may be an image of lanes viewed from the camera's line of sight when the vehicle is located 2 m to the right along the lateral direction from the navigation position.

The camera mounted on the vehicle captures a front image of the vehicle (S150), and the control unit detects a lane in the front image to generate a lane detection image (S160). For example, as shown in FIG. 5 , the camera captures a front image 510 of the vehicle, and the control unit may detect a lane in the front image 510 as indicated by 530 . In one example, the controller may perform Hough Transform on the front image 510 to obtain a straight line and detect a lane using the vanishing point. Accordingly, a lane detection image 550 representing a lane in the front image 510 may be generated.

The control unit may perform matching between the lane detection image and the plurality of virtual lane images and select a virtual lane image that matches the lane detection image among the plurality of virtual lane images (S170). For example, as shown in FIG. 6, the control unit performs matching between the lane detection image 550 and a plurality of virtual lane images 400 to match the lane detection image 550 with a virtual lane image ( 420) can be obtained. Meanwhile, when the virtual lane image 420 matching the lane detection image 550 and the front image 510 of FIG. 5 overlap, as shown at 650 in FIG. 6, the lanes of the virtual lane image 420 The lanes of the front image 510 may coincide.

In one embodiment, the control unit calculates normalized mutual information (NMI) values between the lane detection image and the plurality of virtual lane images, and the virtual lane having the largest NMI value among the NMI values. A lane image may be selected as the virtual lane image matching the lane detection image. For example, the control unit uses equation 710 shown in FIG. 7, that is, " ", and equation (730), i.e. " The NMI value of the lane detection image and each virtual lane image can be calculated using ". Here, Y is the virtual lane image, C is the lane detection image, and NMI(Y,C) is the lane detection image. and the NMI value of each virtual lane image, and the “H (variable)” function may represent the entropy of the variable. However, matching the lane detection image and the plurality of virtual lane images is performed using the equations in FIG. It is not limited to using 710, 730).

The control unit may correct the navigation position according to the matching result, that is, according to the vehicle position corresponding to the selected virtual lane image among the plurality of vehicle positions (S180). In one embodiment, the controller may determine the corrected navigation position as the vehicle position corresponding to the selected virtual lane image. Meanwhile, the vehicle location determined in this way may be a precise vehicle location that is substantially the same as the actual vehicle location. For example, as shown at 600 in FIG. 6, the navigation position indicated by a red dot may be distant from the actual vehicle position indicated by a green dot, but the navigation position indicated by a blue dot according to the matching result is indicated by a green dot. It may be identical to or overlap with the actual vehicle location.

The control unit may determine the current lane of the vehicle based on the corrected navigation position (S190). In one embodiment, the control unit determines a lane center line adjacent to the corrected navigation position among a plurality of lane center lines included in the high-precision map data, and determines the current lane of the vehicle according to the determined lane center line. . For example, at 800 in FIG. 8, green lines represent lanes, and pink lines represent lane center lines. The control unit may select the lane center line closest to the corrected navigation position among the lane center lines represented by the pink lines, and determine the lane where the lane center line is located as the current lane of the vehicle.

As described above, in the lane determination method according to embodiments of the present invention, the lane information adjacent to the navigation position is extracted from the high-precision map data, and based on the navigation position, the lane information, and the camera parameters. The plurality of virtual lane images are generated at the plurality of vehicle positions within the position error range of the navigation position, the lane detection image is generated by detecting the lane in the front image, and the lane detection image and the The navigation position may be corrected by performing matching on a plurality of virtual lane images, and the current lane of the vehicle may be determined based on the corrected navigation position. Accordingly, the precise location of the vehicle can be obtained, and the current lane of the vehicle can be accurately determined.

Figure 9 is a block diagram showing a lane determination device according to embodiments of the present invention.

Referring to FIG. 9, the lane determination device 900 that determines the lane of the vehicle according to embodiments of the present invention includes a navigation device 910 that acquires the navigation position of the vehicle, and a camera that captures a front image of the vehicle. 950, a high-precision map storage unit 930 that stores high-precision map (HD-MAP) data, and a control unit 970 that controls the operation of the lane determination device 100. In one embodiment, the navigation device 910 may include a Global Navigation Satellite System (GNSS) and/or an Inertial Navigation System (INS).

The control unit 970 extracts information on lanes adjacent to the navigation position from the high-precision map data, and selects a plurality of vehicles within the position error range of the navigation position based on the navigation position, the lane information, and the camera parameters of the camera 950. A plurality of virtual lane images can be generated based on the line of sight of the camera 950 at the locations. In one embodiment, the camera parameters include intrinsic parameters and extrinsic parameters of the camera 950, the intrinsic parameters include the focal length, principal point coordinates, and asymmetry coefficient of the camera 950, and the extrinsic parameters include the camera ( 950) of rotation matrix values and relative positions.

In one embodiment, the control unit 970 creates a virtual camera model using the camera parameters, and changes the relative position of the camera 950 in the virtual camera model while changing the plurality of positions at each of the plurality of vehicle locations. Each of the virtual lane images can be generated. For example, the control unit 970 converts each lane point coordinate of the lane information into virtual lane point coordinates using the virtual camera model, and selects the plurality of virtual lanes including virtual lane points in the virtual lane point coordinates. Each of the images can be created. In one embodiment, the control unit 970 uses the equation " " can be used to convert each lane point coordinate of the lane information into the virtual lane point coordinates, where s represents a scale factor, and x and y multiplied by the scale factor represent the virtual lane point coordinates. _, f _x and f _y represent the focal length of the camera 950, c _x and c _y represent the coordinates of the principal point of the camera 950, _{skew_cf} r ₃₃ represents rotation matrix values of the camera 950, t ₁ to t ₃ represent the relative position of the camera 950, and X, Y, and Z may represent coordinates of each lane point of the lane information.

Meanwhile, the control unit 970 adjusts the relative position of the camera 950 to the plurality of vehicle positions until the plurality of virtual lane images are generated at the plurality of vehicle positions within the position error range of the navigation position. Each of the plurality of virtual lane images can be generated while changing the positions to correspond to each other. In one embodiment, the controller 970 may change the relative position of the camera 950 by 0.1 m to generate the plurality of virtual lane images at the plurality of vehicle positions with an interval of 0.1 m.

In addition, the control unit 970 generates a lane detection image by detecting a lane in the front image captured by the camera 950, and performs matching between the lane detection image and the plurality of virtual lane images to determine the plurality of lanes. Among the virtual lane images, a virtual lane image that matches the lane detection image can be selected. In one embodiment, the control unit 970 calculates normalized mutual information (NMI) values between the lane detection image and the plurality of virtual lane images, and has the largest NMI value among the NMI values. The virtual lane image can be selected.

Additionally, the control unit 970 may correct the navigation position according to the vehicle location corresponding to the selected virtual lane image among the plurality of vehicle locations. In one embodiment, the controller 970 may determine the corrected navigation position as the vehicle position corresponding to the selected virtual lane image. Meanwhile, the corrected navigation position determined in this way may be a vehicle precise position corresponding to the actual position of the vehicle.

Additionally, the control unit 970 may determine the current lane of the vehicle based on the corrected navigation position. In one embodiment, the control unit 970 determines a lane center line adjacent to the corrected navigation position among a plurality of lane center lines included in the high-precision map data stored in the high-precision map storage unit 930, and determines the lane center line adjacent to the determined lane center line. Accordingly, the current lane of the vehicle can be determined.

As described above, the lane determination device 100 according to embodiments of the present invention extracts the lane information adjacent to the navigation position from the high-precision map data and based on the navigation position, the lane information, and the camera parameters. generating the plurality of virtual lane images at the plurality of vehicle positions within the position error range of the navigation position, detecting the lane in the front image to generate the lane detection image, and generating the lane detection image and The navigation position may be corrected by matching the plurality of virtual lane images, and the current lane of the vehicle may be determined based on the corrected navigation position. Accordingly, the lane determination device 100 according to embodiments of the present invention can obtain the precise location of the vehicle and accurately determine the current lane of the vehicle.

Above, the lane determination method and lane determination device according to the embodiments of the present invention have been described with reference to the drawings. However, the above description is illustrative and is based on common knowledge in the relevant technical field without departing from the technical spirit of the present invention. It may be modified and changed by those who have it.

The present invention can be applied to methods and devices for vehicles. Although the present invention has been described above with reference to embodiments, those skilled in the art may modify and change the present invention in various ways without departing from the spirit and scope of the present invention as set forth in the claims below. You will understand that it is possible.

900: Lane decision device
910: navigation device
930: High-precision map storage unit
950: Camera
970: Control unit

Claims (20)

In the method of determining the lane of a vehicle,
A navigation device acquiring the navigation position of the vehicle;
A control unit extracting lane information adjacent to the navigation location from high-precision map data;
The control unit generating a plurality of virtual lane images by the line of sight of the camera at a plurality of vehicle positions within a position error range of the navigation position based on the navigation position, the lane information, and camera parameters of the camera;
capturing an image of the front of the vehicle by the camera;
generating a lane detection image by the control unit detecting a lane in the front image;
The control unit performing matching between the lane detection image and the plurality of virtual lane images to select a virtual lane image that matches the lane detection image among the plurality of virtual lane images;
The control unit correcting the navigation position according to a vehicle location corresponding to the selected virtual lane image among the plurality of vehicle locations; and
The control unit determines the current lane of the vehicle based on the corrected navigation position,
Generating the plurality of virtual lane images at the plurality of vehicle locations includes:
generating a virtual camera model by the control unit using the camera parameters; and
A method for determining a lane, wherein the control unit generates each of the plurality of virtual lane images at each of the plurality of vehicle positions while changing the relative position of the camera in the virtual camera model. The method of claim 1, wherein the camera parameters include an intrinsic parameter and an extrinsic parameter of the camera. The method of claim 2, wherein the inner product parameters include a focal length, principal point coordinates, and asymmetry coefficient of the camera,
The external parameter includes a rotation matrix value and a relative position of the camera. delete The method of claim 1, wherein generating each of the plurality of virtual lane images at each of the plurality of vehicle locations comprises:
The control unit converting each lane point coordinate of the lane information into virtual lane point coordinates using the virtual camera model; and
The control unit generating each of the plurality of virtual lane images including a virtual lane point in the virtual lane point coordinates,
The relative position of the camera is changed to correspond to each of the plurality of vehicle positions until the plurality of virtual lane images at the plurality of vehicle positions within the position error range of the navigation position are generated. A lane determination method characterized in that each of the virtual lane images is generated. The method of claim 5, wherein the control unit,
Mathematical expression " Converting each lane point coordinate of the lane information into the virtual lane point coordinates using ",
Here, s represents the scale factor, x and y multiplied by the scale factor represent the virtual lane point coordinates, f _x and f _y represent the focal length of the camera, and c _x and c _y represent the focal length of the camera. _{represents the} principal _point coordinates _{, skew_cf} represents the coordinates of each lane point of the lane information. The method of claim 5, wherein the control unit changes the relative position of the camera by 0.1 m to generate the plurality of virtual lane images at the plurality of vehicle positions with an interval of 0.1 m. method. The method of claim 1, wherein selecting the virtual lane image that matches the lane detection image from among the plurality of virtual lane images comprises:
The control unit calculating normalized mutual information (NMI) values between the lane detection image and the plurality of virtual lane images; and
A method for determining a lane, wherein the control unit selects the virtual lane image having the largest NMI value among the NMI values. The method of claim 1, wherein correcting the navigation position according to the vehicle position corresponding to the selected virtual lane image comprises:
A lane determination method comprising the step of the control unit determining the corrected navigation position as the vehicle position corresponding to the selected virtual lane image. The method of claim 1, wherein determining the current lane of the vehicle based on the corrected navigation position comprises:
The control unit determining a lane center line adjacent to the corrected navigation position among a plurality of lane center lines included in the high-precision map data; and
A method for determining a lane, wherein the control unit determines the current lane of the vehicle according to the determined lane center line. In the lane determination device for determining the lane of a vehicle,
a navigation device that acquires the navigation position of the vehicle;
A camera that captures an image of the front of the vehicle;
A high-precision map storage unit that stores high-precision map data; and
A control unit that controls the operation of the lane determination device,
The control unit,
Extract lane information adjacent to the navigation location from the high-precision map data,
Generate a plurality of virtual lane images by the line of sight of the camera at a plurality of vehicle positions within a position error range of the navigation position based on the navigation position, the lane information, and the camera parameters of the camera,
Generate a lane detection image by detecting a lane in the front image,
Matching the lane detection image and the plurality of virtual lane images to select a virtual lane image that matches the lane detection image among the plurality of virtual lane images,
Correcting the navigation position according to a vehicle position corresponding to the selected virtual lane image among the plurality of vehicle positions,
Determine the current lane of the vehicle based on the corrected navigation position,
The control unit,
Create a virtual camera model using the camera parameters,
A lane determination device that generates each of the plurality of virtual lane images at each of the plurality of vehicle positions while changing the relative position of the camera in the virtual camera model. The lane determination device according to claim 11, wherein the camera parameters include an intrinsic parameter and an extrinsic parameter of the camera. 13. The method of claim 12, wherein the inner product parameters include a focal length, principal point coordinates, and asymmetry coefficient of the camera,
The external parameter includes a rotation matrix value and a relative position of the camera. delete According to claim 11,
The control unit converts each lane point coordinate of the lane information into virtual lane point coordinates using the virtual camera model, and generates each of the plurality of virtual lane images including a virtual lane point from the virtual lane point coordinates, ,
The control unit sets the relative position of the camera to correspond to the plurality of vehicle positions, respectively, until the plurality of virtual lane images at the plurality of vehicle positions within the position error range of the navigation position are generated. A lane determination device characterized in that it generates each of the plurality of virtual lane images while changing them. The method of claim 15, wherein the control unit:
Mathematical expression " Converting each lane point coordinate of the lane information into the virtual lane point coordinates using ",
Here, s represents the scale factor, x and y multiplied by the scale factor represent the virtual lane point coordinates, f _x and f _y represent the focal length of the camera, and c _x and c _y represent the focal length of the camera. _{represents the} principal _point coordinates _{, skew_cf} represents the coordinates of each lane point of the lane information. The method of claim 15, wherein the control unit changes the relative position of the camera by 0.1 m to generate the plurality of virtual lane images at the plurality of vehicle positions with an interval of 0.1 m. Device. The method of claim 11, wherein the control unit:
Calculate normalized mutual information (NMI) values between the lane detection image and the plurality of virtual lane images,
A lane determination device characterized in that the virtual lane image having the largest NMI value among the NMI values is selected. The method of claim 11, wherein the control unit:
Lane determination device, characterized in that determining the corrected navigation position as the vehicle position corresponding to the selected virtual lane image. The method of claim 11, wherein the control unit:
Determining a lane center line adjacent to the corrected navigation position among a plurality of lane center lines included in the high-precision map data,
A lane determination device that determines the current lane of the vehicle according to the determined lane center line.

Images