Travel route recognition device

Abstract

To provide a travel route recognition device capable of accurately recognizing even relatively distant travel routes.SOLUTION: A travel route recognition device 10 includes: a distant parking zone line recognition unit 11 that recognizes a parking zone line PL2 that exists at a distance in the direction of travel using a distant image in the direction of travel of a vehicle acquired by a distant image capturing device 21 mounted on the vehicle; a neighborhood information recognition unit 12 that recognizes a linear structure existing in the neighborhood of the vehicle using at least one of an image taken in the neighborhood of the vehicle acquired by the neighborhood imaging device 22 mounted on the vehicle and sensor information in the neighborhood of the vehicle acquired by the sensor 32 mounted on the vehicle; a correction unit 14 that corrects the recognition result of the parking zone line recognized by the distant parking zone line recognition unit using the recognition result of the linear structure recognized by the neighborhood information recognition unit; and a travel route setting unit 15 that sets a travel route using the information of the parking zone lines after the correction is made by the correction unit.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present disclosure relates to a path recognition device. [Background technology]

[0002] For example, Patent Document 1 describes a driving assistance device that supports vehicle driving so as to drive the vehicle more safely in a parking lot. This driving assistance device detects the ends of parking lane lines using the image capture results around the vehicle, sets an approximate line connecting the detected ends as a virtual driving lane line, and supports the vehicle driving so as to maintain a certain distance from the driving lane line based on the set driving lane line. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2019-66934 A Summary of the Invention [Problem to be solved by the invention]

[0004] As described above, in order to recognize the driving route that a vehicle should take in a parking lot, it is effective to use information on parking space lines. However, in general, a near-field camera that captures the surroundings of a vehicle can accurately recognize parking space lines, but has the disadvantage that the recognition range is narrow. Since the range in which a near-field camera can recognize parking space lines is limited to the vicinity of the vehicle, it is difficult to obtain information on a relatively distant range where parking space lines continue. Although it is possible to recognize parking space lines from a distance by using a far-field camera that can capture images up to several tens of meters ahead, it is difficult to accurately recognize distant parking space lines because of low positional accuracy. In view of the above problems, a driving path recognition device that can accurately recognize driving paths up to a relatively distant distance is desired. [Means for solving the problem]

[0005] The present disclosure can be realized in the following forms.

[0006] According to one embodiment of the present disclosure, a lane recognition device is provided, which is mounted on a vehicle (1) and recognizes a lane for the vehicle, and includes a distant parking space line recognition unit (11) that recognizes a parking space line (PL2) located far away in the traveling direction of the vehicle using a distant captured image in the traveling direction of the vehicle acquired by a distant imaging device (21) mounted on the vehicle, a nearby information recognition unit (12) that recognizes a linear structure located near the vehicle using at least one of a captured image in the vicinity of the vehicle acquired by a nearby imaging device (22) mounted on the vehicle and sensor information in the vicinity of the vehicle acquired by a sensor (32) mounted on the vehicle, a correction unit (14) that corrects the recognition result of the parking space line recognized by the distant parking space line recognition unit using the recognition result of the linear structure recognized by the nearby information recognition unit, and a lane setting unit (15) that sets the lane using information on the parking space line corrected by the correction unit. According to this configuration, the distant parking space lines recognized by the distant parking space line recognition unit are corrected by the correction unit using the recognition results of the linear structures recognized by the nearby information recognition unit.Then, the path setting unit sets the vehicle's path using the corrected parking space lines.That is, the distant parking space lines recognized using the captured image acquired by the distant imaging device are corrected using the recognition results, which are nearby information with relatively high recognition accuracy, and then the path setting is performed.Therefore, the driving route up to a relatively distant place that can be captured by the distant imaging device can be recognized with high accuracy. [Brief description of the drawings]

[0007] [Figure 1] 1 is a functional block diagram showing a schematic configuration of a lane recognition device according to a first embodiment of the present disclosure. [Diagram 2] FIG. 2 is an explanatory diagram illustrating a schematic diagram of a state in which a distant lane marking is corrected based on a nearby lane marking in the first embodiment. [Diagram 3]4 is a flowchart showing a procedure of a lane recognition process in the first embodiment. [Figure 4] 10 is a flowchart showing a procedure of a lane recognition process in a second embodiment of the present disclosure. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0008] Hereinafter, several embodiments will be described with reference to the drawings. A. First embodiment: A1. Configuration of the road recognition device 10: The configuration of the first embodiment will be described with reference to Figs. 1 to 3. The lane recognition device 10 of the first embodiment is mounted on a vehicle 1 (see Fig. 2) and recognizes the lane of the vehicle 1. In this embodiment, the vehicle 1 mounted with the lane recognition device 10 is also called the "own vehicle 1". The own vehicle 1 is a vehicle 1 capable of automatic driving. The own vehicle 1 may be either a vehicle 1 that switches between automatic driving and manual driving (driving by a driver) and performs only automatic driving out of these two types of driving. As shown in Fig. 1, the lane recognition device 10 is configured as an ECU (Electronic Control Unit) equipped with a CPU 16 and a memory 17. The lane recognition device 10 is electrically connected to an imaging unit 20, a millimeter wave radar 31, a sonar 32, a vehicle speed sensor 33, a yaw rate sensor 34, and a vehicle control unit 40 mounted on the vehicle 1.

[0009] In the following description, left and right directions are based on left and right when viewed in the traveling direction of vehicle 1. Imaging unit 20 includes distant camera 21 and nearby camera 22. Distant camera 21 is composed of a single camera, and captures an image of the area ahead of vehicle 1. As shown in FIG. 2, distant camera 21 is disposed, for example, above the front window of vehicle 1. Distant camera 21 has a viewing angle of, for example, about 60° to 100°, and can capture an image up to about several tens to 100 meters ahead. Distant camera 21 corresponds to an example of a "distant imaging device."

[0010] The nearby camera 22 includes a front camera (FR camera) 23, a right camera (RT camera) 24, a left camera (LT camera) 25, and a rear camera (RR camera) 26. The front camera 23 is provided, for example, near the grill or bumper of the vehicle 1. The right camera 24 is provided, for example, on the right rearview mirror (side portion) of the vehicle 1 so as to face to the right. The left camera 25 is provided, for example, on the left rearview mirror (side portion) of the vehicle 1 so as to face to the left.

[0011] The rear camera 26 is provided, for example, near the rear bumper of the vehicle 1. The cameras 23, 24, 25, and 26 constituting the proximity camera 22 are each configured with a fisheye camera. The proximity camera 22 is disposed on the vehicle 1 facing slightly downward, and captures an image of the vicinity of the vehicle 1. The proximity camera 22 can capture an image of the vicinity up to about 5 m with a viewing angle of about 180°, for example. The proximity camera 22 corresponds to an example of a "proximity imaging device." When the ignition of the vehicle 1 is turned on, the imaging unit 20 repeatedly executes acquisition of captured images.

[0012] In Fig. 2, the area enclosed in a rectangle by a thin line is the vicinity area of ​​the host vehicle 1 that can be recognized by the vicinity camera 22. In Fig. 2, the right side area of ​​the host vehicle 1 that can be recognized by the vicinity camera 22 is illustrated as a right vicinity area AnR. Also, the left side area of ​​the host vehicle 1 that can be recognized by the vicinity camera 22 is illustrated as a left vicinity area AnL. Note that the right vicinity area AnR and the left vicinity area AnL include areas on the sides of the vehicle 1 that are longer than the length of the vehicle 1 in the front-to-rear direction. However, simply "the sides of the vehicle 1" means the left and right areas of the part that corresponds to the length of the vehicle 1 in the front-to-rear direction.

[0013] 2, the area surrounded by a thin line in a fan shape is the area recognizable by distant camera 21. That is, the region recognizable by distant camera 21 is a fan-shaped area that extends about 100° around the front of distant camera 21, centered on a line extending from the installation position of distant camera 21 in the traveling direction, and is illustrated as distant area Af.

[0014] Here, "the vicinity of the vehicle 1" refers to an area in front of and to the sides of the vehicle 1, within a predetermined distance from the vehicle 1. In this embodiment, the predetermined distance that defines the vicinity of the vehicle 1 is 5 m. However, this is not limited to 5 m, and may be any distance.

[0015] Referring again to FIG. 1, the millimeter wave radar 31 senses an object by irradiating it with millimeter waves, and detects the object's position information such as the distance and angle, and the relative speed of the object. The sonar 32 detects objects at a relatively short distance by utilizing the bounce of sound waves. The millimeter wave radar 31 and the sonar 32 function as peripheral sensors that detect linear structures. Note that other sensors may be provided as peripheral sensors. Information about the vicinity of the vehicle 1 acquired by the peripheral sensors corresponds to "sensor information." The vehicle speed sensor 33 measures the moving speed of the host vehicle 1. The yaw rate sensor 34 detects the yaw rate (rotational angular velocity) of the host vehicle 1. Each sensor is activated when the ignition of the host vehicle 1 is turned on, and each sensor repeatedly performs a predetermined detection.

[0016] The vehicle control unit 40 controls the movement of the vehicle 1. The vehicle control unit 40 is configured to have, for example, an ECU that controls the operation of a drive mechanism such as an internal combustion engine or an electric motor, an ECU that controls an electronically controlled brake system (ECB), an ECU that controls the steering angle of the wheels, and a display unit such as a meter.

[0017] The lane recognition device 10 includes a distant parking space line recognition unit 11, a nearby information recognition unit 12, a determination unit 13, a correction unit 14, and a lane setting unit 15. The CPU 16 included in the lane recognition device 10 executes a control program stored in the memory 17, so that the lane recognition device 10 functions as the distant parking space line recognition unit 11, the nearby information recognition unit 12, the determination unit 13, the correction unit 14, and the lane setting unit 15.

[0018] The distant parking space line recognition unit 11 recognizes parking space lines (hereinafter referred to as "distant parking space lines PL2" (see FIG. 2)) that exist far away in the traveling direction of the vehicle 1 using a distant image captured by the distant camera 21 in the traveling direction of the vehicle 1. "Far away in the traveling direction" refers to a range excluding the range near the vehicle 1 in the traveling direction. Specifically, in this embodiment, it means a range farther than 5 m from the vehicle 1. The vicinity information recognition unit 12 recognizes linear structures that exist near the vehicle 1. The linear structures refer to, for example, nearby parking space lines PL1 that exist near the vehicle 1, as well as walls, curbs, guardrails, etc. that exist near the vehicle 1. In this embodiment, the vicinity information recognition unit 12 recognizes parking space lines as linear structures.

[0019] The determination unit 13 determines whether the distant parking stall line PL2 recognized by the distant parking stall line recognition unit 11 is identical to the nearby parking stall line PL1 recognized by the nearby information recognition unit 12. "Identity" means that the stall lines constitute the same parking row. Hereinafter, "the distant parking stall line PL2 recognized by the nearby information recognition unit 12 using the distant image in the traveling direction of the vehicle 1 acquired by the distant camera 21" is simply referred to as "the distant parking stall line PL2". "The nearby parking stall line PL1 recognized by the distant parking stall line recognition unit 11 using the image in the vicinity of the vehicle 1 acquired by the nearby camera 22" is simply referred to as "the nearby parking stall line PL1".

[0020] The correction unit 14 corrects the recognition result of the distant parking stall line PL2 recognized by the distant parking stall line recognition unit 11 using the recognition result of the nearby parking stall line PL1 recognized by the nearby information recognition unit 12. The path setting unit 15 sets the path of the vehicle 1 using the distant parking stall line PL2 after correction by the correction unit 14. The distant parking stall line PL2 after correction is hereinafter referred to as the "corrected parking stall line CPL2."

[0021] A2. Processing of the road recognition device 10: Next, the process executed by the lane recognition device 10 described above in detail will be described with reference to the flowchart shown in Fig. 3. In this embodiment, the lane recognition by the lane recognition device 10 described above in detail is used for driving support of the vehicle 1 in a large parking lot such as a shopping mall. In the parking lot, many parking areas where the vehicle 1 should be parked are defined by parking lane lines. When the host vehicle 1 enters the parking lot that is the subject of driving support, the process of the flowchart shown in Fig. 3 is executed.

[0022] As described above, the distant camera 21 can recognize parking space lines that are farther away than the nearby camera 22, but the position accuracy is lower than that of the nearby camera 22. This is due to the difference in performance of the mounted cameras, and also due to the fact that, even if the same type of camera is used, in the image of the front, the lateral straight line components extending along the traveling direction are obliquely captured on the image. In this respect, in the image acquired by the nearby camera 22 such as the right camera 24 or the left camera 25, the lateral straight line components extending along the traveling direction can be recognized as nearly horizontal straight line components extending to the left and right. Therefore, the recognition result using the captured image acquired by the nearby camera 22 is more accurate than the recognition result using the captured image acquired by the distant camera 21. In particular, the accuracy is higher for the angle component.

[0023] In this embodiment, when performing distant lane recognition, the recognition result of the distant parking stall line PL2 is not used as is, but is corrected by the correction unit 14 using the recognition result of the nearby parking stall line PL1, and the lane of the vehicle 1 is set using the corrected recognition result of the distant parking stall line PL2, i.e., the corrected lane line CPL2. Details will be described below.

[0024] 3, when the lane recognition process starts, in S101, it is determined whether or not the distant parking space line PL2 is recognized by the distant camera 21. If the distant parking space line PL2 is recognized by the distant camera 21 (S101: Yes), the process proceeds to S102, where it is determined whether or not the nearby parking space line PL1 is recognized by the nearby camera 22. If the nearby parking space line PL1 is recognized by the nearby camera 22 (S102: Yes), the process proceeds to S103, where the determination unit 13 performs the same parking row determination.

[0025] In the same parking row determination, the identity of the nearby parking stall line PL1 and the distant parking stall line PL2 is determined. Specifically, in FIG. 2, the angle θ between the straight line C1 and the straight line C2 is determined based on whether it is within a predetermined angle range. Here, the straight line C1 is a straight line that connects the sides facing the side of the vehicle 1 or the sides that are scheduled to face each other in the nearby parking stall line PL1 detected continuously along the traveling direction. In the example of FIG. 2, an example in which two parking stall lines are detected is illustrated, but the straight line C1 may be determined using one stall line or three or more stall lines. In addition, if the detected stall lines are not on a straight line, the straight line C1 may be determined as an approximate line as appropriate. For example, a straight line connecting the midpoints of the sides facing the vehicle 1 or the sides that are scheduled to face each other may be determined as the straight line C1. The straight line C2 is a straight line that connects the distant parking stall lines PL2 detected substantially continuously along the traveling direction.

[0026] If the angle θ between the straight lines C1 and C2 is within a predetermined angle range, it is determined that the angle is within the variation range, and the nearby parking stall line PL1 and the distant parking stall line PL2 are stall lines that make up the same parking row, i.e., the same parking row determination is established. On the other hand, if the angle θ is outside the predetermined angle range, it is determined that the nearby parking stall line PL1 and the distant parking stall line PL2 are not stall lines that make up the same parking row, i.e., the same parking row determination is not established.

[0027] Referring again to FIG. 3, after the same parking row determination (S103), in S104, it is determined whether the same parking row determination is established. If it is determined in S103 that they are the same and the same parking row determination is established (S104: Yes), the process proceeds to S105, where the position and angle of the distant parking space line PL2 captured by the distant camera 21 are corrected. Specifically, the position and angle of the distant parking space line PL2 are corrected so as to correspond to the position and angle of the nearby parking space line PL1 captured by the nearby camera 22. For example, the angle is first corrected by rotating the camera position as a starting point, and then the lateral position is corrected.

[0028] Then, in S106, the position reliability of the distant parking space line PL2 is set to "High". The position reliability is an index showing the degree of detection accuracy of the distant parking space line PL2, and in this embodiment, one of two levels, "High" and "Low", is set as the index. When the position reliability is "High", it means that the detection accuracy is high. When the position reliability is "Low", it means that the detection accuracy is low.

[0029] If the nearby parking space line PL1 is not recognized by the nearby camera 22 in S102 (S102: No), and if the same parking row determination is not established in S104 (S104: No), the process proceeds to S107, where the position reliability of the distant parking space line PL2 is set to "Low." After the position reliability is set in S106 and S107, and if the distant parking space line PL2 is not recognized by the distant camera 21 in S101 (S101: No), the process proceeds to S108, where the running path is set. This is the end of this processing routine.

[0030] In setting the lane, the position coordinates of the vehicle 1 passing through are set, for example, at predetermined intervals, up to the position where the lane marking is recognized in the direction of travel. A specific driving control is determined so that the vehicle 1 runs while maintaining a position at a certain lateral distance (for example, 0.5 m to 1 m) from the recognized lane marking. In S106 and S107, the position reliability is set, and then in S108, the lane is set using the corrected lane marking CPL2 set to the position reliability "High". At this time, it is possible to set a relatively distant position, up to several tens of meters ahead, as a target. In FIG. 2, the position of the vehicle 1 at the target using the corrected lane marking CPL2 in the above case is illustrated by a two-dot chain line as vehicle 1T.

[0031] When the process of S108 is reached after the process of S107, that is, when the distant parking stall line PL2 is recognized but no correction based on the nearby parking stall line PL1 by the nearby camera 22 is performed and the position reliability is "Low", the driving control is performed to maintain the current course with a target distance of a relatively short distance of several meters from the vehicle 1 as a target. Also, when the distant parking stall line PL2 is not recognized by the distant camera 21 in S101 (S101: No), the process proceeds to S108. At this time, the driving control may be performed to maintain the current course with a target distance of several meters from the vehicle 1 as described above, or, if the nearby parking stall line PL1 is recognized, the driving control may be determined to target the recognized course.

[0032] According to the control flow shown in Fig. 3 described above, when setting a driving path, the distant parking stall line PL2 corrected based on the nearby parking stall line PL1 has a high position reliability, so the corrected stall line CPL2 is used. On the other hand, if the distant parking stall line PL2 is recognized (S101: Yes), but the nearby parking stall line PL1 is not recognized by the nearby camera 22 (S102: No), and correction based on the nearby parking stall line PL1 is not performed, the distant parking stall line PL2 is not used because the position reliability is low.

[0033] [effect] (1) According to the lane recognition device 10 of the first embodiment, the correction unit 14 corrects the distant parking space line PL2 using the nearby parking space line PL1, and the lane setting unit 15 sets the lane of the vehicle 1 using the corrected lane line CPL2. Since the distant parking space line PL2 has low detection accuracy as described above, if the lane is set using the detected information as it is, there is a risk of it lacking in reliability. However, according to the above configuration, the distant parking space line PL2 is corrected using the nearby parking space line PL1, which has high detection accuracy, to obtain the corrected lane line CPL2, so that the lane can be set farther away with high accuracy.

[0034] (2) According to the lane recognition device 10 of the first embodiment, the determination unit 13 determines whether the nearby parking stall line PL1 and the distant parking stall line PL2 are identical in the same parking row determination. If they are determined to be identical, the position and angle of the distant parking stall line PL2 are corrected to correspond to the position and angle of the nearby parking stall line PL1. This makes it possible to prevent the distant parking stall line PL2 from being inappropriately corrected based on the nearby parking stall line PL1 that does not correspond to each other. In addition, the correction can be easily performed by adjusting the position and angle.

[0035] (3) According to the lane recognition device 10 of the first embodiment, the vicinity information recognition unit 12 recognizes the nearby parking stall line PL1 as a linear structure existing in the vicinity of the vehicle 1. That is, as a recognition result of a linear structure, a stall line that is the same index as the distant parking stall line PL2 is used, and the subsequent identity determination and correction are performed based on the nearby parking stall line PL1.

[0036] Therefore, by using the same indicator, the accuracy of the identity determination and the correction accuracy can be improved compared to identity determination by comparing the distant parking space line PL2 with indicators such as walls, curbs, guardrails, etc., which are other linear structures. Also, since the parking space line is basically a straight line drawn with a white line or the like in a parking lot, it is easy and accurate to detect it as a linear structure. Therefore, the accuracy of the identity determination and the correction accuracy can be further improved.

[0037] (4) In the lane recognition device 10 of the first embodiment, the nearby camera 22 can capture the right nearby area AnR and the left nearby area AnL, which are areas on the sides and near the vehicle 1, as the captured image of the vicinity. The nearby parking stall line PL1 on the side extending along the traveling direction can be recognized as a nearly horizontal straight line component extending to the left and right in the captured image. In the example shown in FIG. 2, the left camera 25 can recognize the parking stall line PL1 on the left side as a straight line component. Therefore, the nearby parking stall line PL1 can be recognized with high accuracy, especially in the angle component, compared to the case where it is detected as an oblique component in the captured image.

[0038] B. Second embodiment: Next, a lane recognition device 10 according to a second embodiment of the present disclosure will be described with reference to FIG. 4. The same components as those in the first embodiment are denoted by the same reference numerals, and the description thereof will be omitted. The lane recognition device 10 according to the second embodiment differs from the lane recognition device 10 according to the first embodiment in the procedure of lane recognition processing. The functional block configuration of the device is the same as that of the first embodiment.

[0039] In the second embodiment, the "straight-line structure" recognized by the vicinity information recognition unit 12 includes not only the nearby parking stall line PL1 but also "stationary solid objects" such as walls, curbs, and guardrails. In cases where the nearby parking stall line PL1 is not recognized, the linear components of the stationary solid objects are used as substitutes for the parking stall lines to determine whether the same straight line is determined (S111) and whether the same straight line direction is determined (S112). Hereinafter, a detailed description will be given focusing on the above points that are different from the first embodiment.

[0040] As shown in Fig. 4, in the lane recognition process executed by the lane recognition device 10 of the second embodiment, the processes of S101 to S108 of the first embodiment are executed in the same manner. In the lane recognition procedure of the second embodiment, if the nearby parking space line PL1 is not recognized (S102: No) and if the same parking row determination is not established (S104: No), the process proceeds to S110, where it is determined whether or not a straight line component is extracted from the stationary three-dimensional object recognized by the nearby camera 22.

[0041] Here, the term "stationary three-dimensional object" does not include movable objects, such as parked vehicles. Movable objects are also referred to simply as "movable objects" below. A parked vehicle can be recognized as a linear structure consisting of a part of the body of the vehicle 1, but depending on the design of the vehicle 1 and the parking style of the driver, the linear component of the line of the vehicle 1 does not necessarily follow the parking space line, so it is not preferable to use it as a target element for the same parking row judgment.

[0042] More specifically, depending on the design of the vehicle 1, the outline may be angular, or the driver may park the vehicle at an angle to the parking space line, so that the straight line components constituting the body of the vehicle 1 may be oblique to the parking space line rather than parallel to it. Therefore, the straight line components extracted from a parked vehicle are not preferably used as a substitute for the parking space line, particularly because the reliability of the angle is low, so the parked vehicle is not included in the stationary three-dimensional object. Note that whether or not an object is a movable object can be determined by well-known image recognition processing using the captured image acquired by the nearby camera 22.

[0043] In S110, if a straight line component is extracted from the stationary three-dimensional object recognized by the nearby camera 22 (S110: Yes), the process proceeds to S111, where the determination unit 13 executes a same-line determination. In the same-line determination, the identity of the straight line component of the stationary three-dimensional object recognized in S110 and the distant parking stall line PL2 is determined. Specifically, similar to the same parking row determination in S103, the nearby parking stall line PL1 can be replaced with the straight line component of the stationary three-dimensional object to make the determination.

[0044] After the same line determination (S111), in S112, it is determined whether the same line determination is established. If it is determined in S111 that they are the same and the same line determination is established (S112: Yes), the process proceeds to S105, where the position and angle of the distant parking space line PL2 captured by the distant camera 21 are corrected. Specifically, the position and angle of the distant parking space line PL2 are corrected so as to correspond to the position and angle of the straight line component captured by the nearby camera 22.

[0045] If no straight line component is extracted from the stationary solid object recognized by the near camera 22 in S110 (S110: No), and if the same straight line determination is not established in S112 (S112: No), the process proceeds to S107, and the position reliability of the distant parking stall line PL2 is set to "Low". If the distant parking stall line PL2 is not recognized by the distant camera 21 in S101 (S101: No), and this processing routine ends via S107 or S108. As in the first embodiment, if the distant parking stall line PL2 is not recognized by the distant camera 21 in S101 (S101: No), the process proceeds to S108, and a predetermined running path may be set.

[0046] According to the lane recognition device 10 of the second embodiment, it is possible to achieve the same effects as those of the first embodiment. Furthermore, even if the nearby parking stall line PL1 is not recognized, the distant parking stall line PL2 can be corrected using the straight line components of the stationary three-dimensional object recognized by the nearby camera 22.

[0047] Furthermore, image recognition processing using the nearby camera 22 prevents the distant parking space line PL2 from being corrected by comparison with uncertain straight line components and used to set the driving path, since stationary three-dimensional objects do not include movable objects.

[0048] C. Other embodiments: (C1) In each of the above embodiments, distant camera 21 and nearby camera 22 are different types of cameras, but they may be the same type of cameras.

[0049] (C2) The lane recognition device 10 in each of the above embodiments is used to support the driving of the vehicle 1 in a parking lot, but is not limited to being used only in a parking lot. For example, it may be used to support the driving of the vehicle 1 while it is driving on a road on the side of the parking lot.

[0050] (C3) In each of the above embodiments, the correction unit 14 corrects the position and angle of the distant parking stall line PL2 to correspond to the position and angle of the nearby parking stall line PL1 captured by the nearby camera 22. Instead of this, for example, only either the position or the angle may be corrected according to the amount of deviation between the nearby parking stall line PL1 and the distant parking stall line PL2.

[0051] (C4) In the second embodiment, the stationary three-dimensional object does not include a movable object, but it may be a movable object as long as the reliability of the linear component is high enough to be used as a basis for correction. In this case, the recognition result from the surrounding sensor may be used instead of the recognition from the image captured by the nearby camera 22.

[0052] (C5) In the first embodiment, the vicinity information recognition unit 12 uses a captured image of the vicinity of the vehicle 1, and in the second embodiment, the captured image of the vicinity of the vehicle 1 and sensor information of the vicinity of the vehicle 1 are used. The vicinity information recognition unit 12 only needs to be able to recognize linear structures existing in the vicinity of the vehicle, and may use at least one of the captured image and the sensor information. In other words, the linear component may be extracted using only the sensor information.

[0053] The present disclosure is not limited to the above-mentioned embodiment, and can be realized in various configurations without departing from the spirit of the present disclosure. For example, the technical features in each embodiment corresponding to the technical features in the form described in the Summary of the Invention column can be appropriately replaced or combined in order to solve some or all of the above-mentioned problems or to achieve some or all of the above-mentioned effects. Furthermore, if the technical feature is not described as essential in this specification, it can be appropriately deleted. [Explanation of symbols]

[0054] 1...vehicle (own vehicle), 10...roadway recognition device, 11...distant parking space line recognition unit, 12...proximal information recognition unit, 13...judgment unit, 14...correction unit, 15...roadway setting unit, 16...CPU, 17...memory, 20...imaging unit, 21...distant camera, 22...proximal camera, 23...front camera, 24...right camera, 25...left camera, 26...rear camera, 31...millimeter wave radar, 32...sonar, 33...vehicle speed sensor, 34...yaw rate sensor, 40...vehicle control unit, Af...distant area, AnL...left nearby area, AnR...right nearby area, CPL2...corrected parking space line, PL1...proximal parking space line, PL2...distant parking space line

Claims

1. A lane recognition device that is mounted on a vehicle (1) and recognizes a lane of the vehicle, a distant parking space line recognition unit (11) that recognizes a parking space line (PL2) that exists far away in the traveling direction of the vehicle by using an image captured far away in the traveling direction of the vehicle acquired by a far-field imaging device (21) mounted on the vehicle; a neighborhood information recognition unit (12) that recognizes linear structures present in the vicinity of the vehicle using at least one of an image of the vicinity of the vehicle acquired by a neighborhood imaging device (22) mounted on the vehicle and sensor information of the vicinity of the vehicle acquired by a sensor (32) mounted on the vehicle; a correction unit (14) that corrects the recognition result of the parking space line recognized by the distant parking space line recognition unit using the recognition result of the linear structure recognized by the nearby information recognition unit; A route setting unit (15) that sets the route using the information of the parking space line after the correction unit has been performed; A road recognition device comprising:

2. a determination unit (13) that determines the identity between the parking space line recognized by the distant parking space line recognition unit and the linear structure recognized by the nearby information recognition unit; Further equipped with The correction unit is 2. The lane recognition device according to claim 1, wherein when the determination unit determines that the parking stall lines are identical, the position and angle of the parking stall line recognized by the distant parking stall line recognition unit are corrected so as to correspond to the position and angle of the linear structure.

3. The lane recognition device according to claim 1 , wherein the linear structure is a parking space line recognized using the captured image acquired by the nearby imaging device.

4. The lane recognition device according to claim 1 , wherein the linear structure is a stationary three-dimensional object recognized using information acquired by the nearby imaging device or the sensor.

5. The lane recognition device according to claim 4 , wherein the stationary three-dimensional object is recognized using at least the captured image acquired by the nearby imaging device, and does not include a movable object.

6. the near-field imaging device is provided on a side of the vehicle so as to face the side of the vehicle, The lane recognition device according to claim 1 , wherein the captured image of the vicinity of the vehicle acquired by the vicinity imaging device is an image of at least a side of and vicinity of the vehicle.

7. The road recognition device according to any one of claims 1 to 6, which is used to assist driving of the vehicle in a parking lot where a parking area in which the vehicle is to be parked is defined by the parking space lines.

Images