JP2010146470A - Traveling-lane detector and traveling-lane detection program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a technique that allows a traveling-lane detector that detects traveling-lanes, to accurately detect traveling-lanes. <P>SOLUTION: In a traveling-lane detection system, traveling-lane recognition processing determines whether a road for vehicle traveling is an expressway or an open road based on a traveling speed as the traveling conditions of the vehicle and based on a pedestrian crossing width and on a traveling-lane width as environmental conditions of the road for vehicle traveling (S130) and a parameter is set, with respect to the shape allowed range of the traveling lane (S140). Then, edge components that match the parameter setting among the extracted edge components (edge points) are detected as contour components representing a part of the contour of the traveling lane, and an area closer to the center than the contour components in an image is specified as the traveling lane (S150). <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a travel lane detection device that is mounted on a vehicle and detects a travel lane on which the vehicle travels, and a travel lane detection program.

Conventionally, as the travel lane detection device, the road type on which the vehicle is traveling is determined, the white line drawn on both sides of the travel lane is predicted according to the road type, and the front of the vehicle is imaged A technique for recognizing a driving lane by extracting only a white line having an expected thickness from an image and excluding a white line such as a pedestrian crossing is known (see, for example, Patent Document 1).
JP 2007-264712 A

  However, in the above-described traveling lane detection device, when a white line with an expected thickness exists in the captured image in addition to the white line drawn on both sides of the traveling lane, the white line is mistaken as a white line drawn on both sides of the traveling lane. There was a risk of detection.

  Therefore, in view of such problems, it is an object of the present invention to provide a technique that can detect a travel lane with high accuracy in a travel lane detection device that is mounted on a vehicle and detects a travel lane.

  The travel lane detection device according to claim 1, wherein the road type determination means is based on at least one of a travel state of the vehicle and an environment on a road on which the vehicle travels. The road type on the road on which the vehicle travels is determined, and the parameter setting means sets a parameter relating to the allowable range of the shape of the travel lane on which the vehicle travels according to the road type. Then, the specifying unit detects an edge component that matches the parameter setting among the extracted edge components as a contour component that represents a part of the contour of the traveling lane, and is a region closer to the center than the contour component in the captured image. Is identified as the driving lane.

  That is, the position of a contour component such as a white line representing the shape of the travel lane and part of the contour of the travel lane is predicted according to the road type, and the edge component that matches the prediction is detected as the contour component. Here, the “road type” determined by the road type determination means is information for predicting the shape of the travel lane. For example, the road on which the vehicle travels is an uphill, downhill, right curve, Indicates a left curve, straight road, highway, general road, intersection, etc.

  The parameters set by the parameter setting means correspond to, for example, a road width range, a curve radius range, and the like. Furthermore, the “edge component” is a pixel in which a luminance component, a color difference component in each adjacent pixel, or a difference between each component of RGB is larger than a certain value among a large number of pixels constituting a captured image, or these This shows an extracted set of pixels.

  More specifically, the parameter setting of the present invention will be described. For example, if the road type is an expressway, there is no pedestrian crossing or intersection in the captured image, and the travel lane width is defined by the law. It will never be less than the lane width. Also, if the road type is a right curve, the contour curve formed by connecting adjacent contour components bends to the upper side of the captured image (that is, farther away from the vehicle) within the range of the curve. Is not possible. Therefore, when setting parameters, the parameters are set so that only edge components that match the shape of the travel lane estimated from the road type (the shape of the contour curve) can be detected as the contour components.

  Therefore, according to such a travel lane detection device, an edge component that matches the parameter setting according to the road type can be detected as a contour component, and thus the travel lane can be detected with high accuracy.

  By the way, in the travel lane detection device according to claim 1, as described in claim 2, in the captured image, an edge component that continues for a predetermined distance or more in the travel direction of the vehicle in the left-right direction of the captured image. A pedestrian crossing detection unit is provided that determines that a pedestrian crossing has been detected when a predetermined number or more are detected, and the road type determination unit determines the road type based on the presence or absence of a pedestrian crossing in the captured image. Also good.

  According to such a traveling lane detection device, the road type can be determined according to the presence or absence of a pedestrian crossing. Specifically, for example, when a pedestrian crossing is detected, it can be determined that the road type is an intersection, or it can be determined that the road type is a general road that is not a highway.

  Further, in the travel lane detection device according to claim 2, as described in claim 3, information on the vehicle speed of the vehicle is acquired, and vehicle speed determination for determining whether or not the vehicle speed is equal to or higher than a reference vehicle speed. Lane width determining means for acquiring information on the lane width of the traveling lane in which the vehicle is traveling and determining whether the lane width is equal to or larger than the reference lane width. In this case, the road type determining means determines that the pedestrian crossing is not detected by the pedestrian crossing detecting means, and the lane width determining means determines that the lane width of the traveling lane in which the vehicle is traveling is equal to or larger than the reference lane width. When the vehicle speed determining means determines that the vehicle speed is equal to or higher than the reference vehicle speed, the road type is determined to be a highway, and in other cases, the road type is determined to be a general road. You just have to do it.

  In the present invention, “the case other than the above” means that a pedestrian crossing is detected by the pedestrian crossing detection means, or the lane width of the traveling lane in which the vehicle travels is determined by the lane width determination means. It is determined that the vehicle speed is determined to be less than the lane width, or the vehicle speed determination means determines that the vehicle speed of the vehicle is less than the reference vehicle speed.

  Further, the “reference lane width” is a value for identifying a general road and a highway by utilizing the fact that the lane width of an expressway is generally wider than that of a general road. Therefore, the value of the “reference lane width” may be set to a value about the lane width determined by the law as an expressway.

  Further, the “reference vehicle speed” is a value for distinguishing a general road and a highway by using the fact that the speed limit on the expressway is higher than the speed limit on the general road. Therefore, the value of the “reference vehicle speed” may be set to the value of the speed limit determined by law as an expressway.

  According to such a traveling lane detection device, the presence / absence of a pedestrian crossing, the lane width of the traveling lane, and the vehicle speed are used in combination, so whether the road type is a general road or a highway is improved. Can be determined.

  Further, in the travel lane detection device according to claim 3, the parameter setting means, as described in claim 4, when the road type is a highway, the width of the travel lane as one of the parameters The lane width range that is the allowable range may be set wider than when the road type is a general road.

  According to such a travel lane detection device, an appropriate travel lane width range can be set according to the road type, so even if there is an edge component that is not a contour component in the vicinity of the contour component in the captured image. The contour component can be detected with high accuracy.

  Further, in the travel lane detection device according to any one of claims 1 to 4, when the specifying unit repeatedly detects the travel lane, the travel lane detected by the specifying unit as described in claim 5 There may be provided recording means for recording shape information which is information relating to the shape of the vehicle, and correction means for correcting the shape of the current travel lane based on the shape information recorded in the recording means. Specifically, when the specifying unit can identify the shape of the current traveling lane, the shape of the current traveling lane is corrected based on the shape information recorded in the recording unit, and the identifying unit corrects the current traveling lane. If the shape of the vehicle cannot be specified, correction may be made to complement the current shape of the driving lane based on the shape information recorded in the recording means.

  According to such a travel lane detection device, even if the shape of the travel lane is erroneously detected, the current travel lane shape is determined based on the shape information that is information related to the past travel lane shape. It can be corrected. Even if the shape of the travel lane cannot be detected, the current shape of the travel lane can be supplemented based on the shape information. Therefore, the traveling lane can be detected with high accuracy.

  The shape information recorded in the recording means corresponds to, for example, information on the coordinates of each contour component detected by the specifying means.

  Further, in the travel lane detection device according to claim 5, the shape determination means for determining whether or not the shape of the travel lane is within the reference shape range, and the shape of the travel lane is determined by the shape determination means. And use prohibiting means for prohibiting use of the travel lane shape corrected by the correcting means when it is determined that the vehicle is outside.

  According to such a travel lane detection device, when it is estimated that the accuracy of detecting the travel lane is deteriorated by the correction using the past shape information, the use of the past shape information can be prohibited.

  Note that the shape determination unit may determine whether the same parameter as the parameter set by the parameter setting unit falls within the reference shape range, or may determine another parameter.

  Further, in the travel lane detection device according to claim 6, as described in claim 7, shape range setting means for setting the reference shape range according to the road type may be provided.

  According to such a travel lane detection device, the reference shape range can be set according to the shape of the travel lane estimated by the road type, so that the accuracy of detecting the travel lane can be improved.

  Further, in the travel lane detection device according to claim 6 or claim 7, as described in claim 8, when the use prohibiting means prohibits the use of the shape of the travel lane corrected by the correcting means, A discarding unit that discards the recording contents of the recording unit may be provided.

  According to such a travel lane detection device, it is possible to avoid using the shape information recorded in the recording means in the future processing with a simple processing.

  Further, as a specific configuration of the travel lane detection device according to claim 8, as described in claim 9, the shape determination means includes a right and left of the vehicle with respect to the traveling direction of the vehicle in the specified travel lane. Yaw angle determination means for determining whether the yaw angle representing the inclination angle of the direction is within the reference yaw angle range, and the discarding means at least determines the yaw angle of the vehicle to be within the reference yaw angle range by the yaw angle determination means. When it is determined that it is outside, the recording content by the recording means may be discarded.

  According to such a travel lane detection device, it is possible to set whether or not to perform correction using past shape information depending on whether or not the yaw angle is within the reference yaw angle range.

  Furthermore, in the travel lane detection device according to claim 9, as described in claim 10, an edge that continues a predetermined distance or more while forming a predetermined angle with respect to the traveling direction of the vehicle in the captured image. When a predetermined number or more components are detected in the left-right direction of the captured image, a safety zone detection unit that determines that a safety zone has been detected, and a discarding unit that records when a safety zone is detected by the safety zone detection unit And a first discard prohibiting unit that prohibits discarding the recorded contents by the unit.

  Here, the “safe zone” in the present invention refers to the area where the vehicle should not travel, such as where the lane width is changed while maintaining a smooth flow of the vehicle, with a plurality of diagonal lines. Indicates the area. Such a “safe zone” is often arranged at a point where the number of lanes on the road changes, such as near an intersection.

  Thus, since the lane width changes in the vicinity where the safety zone is detected, a contour component that is not parallel to the traveling direction of the vehicle is detected. Therefore, when the yaw angle determination means compares the yaw angle with the reference yaw angle range, the yaw angle may be out of the reference yaw angle range if it is detected to be abnormally large. Therefore, in the present invention, in order to ignore the abnormality of the yaw angle when the safety zone is detected, the discarding unit is prohibited from discarding the recording contents by the recording unit.

  Therefore, according to such a travel lane detection device, when an abnormality in the yaw angle at a point where the number of lanes changes is detected, it is possible to hold the recording content by the recording means as not being abnormal.

  Further, in the travel lane detection device according to any one of claims 8 and 9, as described in claim 11, a safety zone detection means and a safety zone detection means similar to those described in claim 10 are provided. An intersection detection unit that determines that an intersection has been detected when the identification unit cannot identify a travel lane within a certain time after the detection of the safety zone, and a discard unit when an intersection is detected by the intersection detection unit May include a second discarding prohibiting unit that prohibits discarding the recorded content by the recording unit.

  According to such a travel lane detection device, even if the travel lane shape estimation result is outside the reference shape range when the travel lane cannot be specified because the vehicle is traveling at an intersection, Since the contents recorded by the recording means are not discarded, the process for estimating the shape of the traveling lane can be continued.

  Note that the travel lane detection device according to the present invention is employed when detecting a travel lane in an alarm device that warns the departure of the travel lane, a steering torque control device that controls the steering torque so as not to deviate from the travel lane, and the like. can do.

  Next, the invention according to claim 12 is a traveling lane detection program for realizing a function as each means constituting the traveling lane detection device according to any one of claims 1 to 11 by a computer. It is characterized by.

  According to such a travel lane detection program, at least the same effects as those of the travel lane detection device according to claim 1 can be obtained.

  Embodiments according to the present invention will be described below with reference to the drawings.

[Configuration of this embodiment]
FIG. 1A is a block diagram illustrating a schematic configuration of a traveling lane detection system 1 to which the present invention is applied, and FIG. 1B is a block diagram illustrating a configuration of a traveling lane detection ECU 10. The traveling lane detection system 1 is mounted on a vehicle such as a passenger car, for example, and is a system that detects a traveling lane in which the vehicle travels from a captured image obtained by capturing the front of the vehicle.

  More specifically, as shown in FIG. 1A, the travel lane detection system 1 includes a travel lane detection ECU (ECU is an electronic control unit) 10, a vehicle speed sensor 20, a navigation ECU 30, and a vehicle control ECU 40. For example, each is connected to a communication line 5 on which communication is performed by a protocol such as CAN (Controller Area Network), and each is configured to be communicable. The travel lane detection ECU 10 is connected to a camera 15 that captures a road surface ahead of the vehicle, and the travel lane detection ECU 10 is configured to acquire an image captured by the camera 15 at an arbitrary timing.

  The navigation ECU 30 is connected to an ETC (Electronic Toll Collection) vehicle-mounted device 35. The navigation ECU 30 receives information indicating that the ETC vehicle-mounted device 35 has passed the toll gate and toll gate passing information such as a fee settlement result. It is configured so that it can be acquired. Further, the navigation ECU 30 can provide toll gate passage information in response to a request from the travel lane detection ECU 10.

  As shown in FIG. 1B, the travel lane detection ECU 10 is configured as a known microcomputer including a CPU 11, a ROM 12, a RAM 13 (recording means), and the like, and travels via a communication interface (I / F) 14. A process of acquiring information necessary for the process of detecting the lane from another ECU or sensor, a process of transmitting the detection result of the traveling lane to another ECU (in particular, the vehicle control ECU 40 in this case), and the like are performed. These processes are performed based on a program (running lane detection program) stored in the ROM 12.

  Note that the hardware configuration of the navigation ECU 30 and the vehicle control ECU 40 is the same as that of the travel lane detection ECU 10, and thus the description thereof is omitted. Further, the vehicle control ECU 40 functions as an alarm device for warning the departure of the travel lane by using the detection result of the travel lane by the travel lane detection ECU 10, or controls the steering torque so as not to deviate from the travel lane. Or function as a control device.

[Processing in this embodiment]
[Running lane recognition process]
In the travel lane detection system 1 configured as described above, processing for recognizing a travel lane will be described with reference to FIG. FIG. 2 is a flowchart showing a travel lane recognition process executed by the travel lane detection ECU 10, FIG. 3A is a flowchart showing an edge extraction process in the travel lane recognition process, and FIG. 3B is a travel lane recognition process. It is a flowchart which shows the road classification determination process.

  4 is a flowchart showing individual determination processing in road type determination processing, FIG. 5 is a flowchart showing road shape constant processing in travel lane recognition processing, and FIG. 6 is lane detection processing in travel lane recognition processing. It is a flowchart which shows. 8A is a flowchart showing the safe zone detection process in the lane detection process, FIG. 9 is a flowchart showing the intersection determination process in the lane detection process, and FIG. 10 is a guard process in the lane detection process. It is a flowchart to show.

  The travel lane recognition process is a process that is started when the power source of a vehicle such as an ignition switch is turned on, and then repeatedly performed at regular intervals (for example, every 100 ms). Specifically, first, a captured image in which the traveling direction of the vehicle is captured by the camera 15 is acquired from the camera 15 (S110: image acquisition means).

  And based on at least one of the edge extraction process (S120: edge component extraction means) which is a process which extracts the edge component in the acquired captured image, the driving | running state of the said vehicle, and the environment in the road where the said vehicle drive | works. , Road type determination processing (S130: road type determination means) for determining the road type on the road on which the vehicle travels, and the allowable range (reference shape range) of the shape of the travel lane on which the vehicle travels according to the road type A road shape constant setting process for setting a parameter (S140: parameter setting means, shape range setting means), and an edge component that matches the parameter setting among the extracted edge components represents a part of the contour of the traveling lane This is detected as a contour component, and a region for identifying a region closer to the center than the contour component as a travel lane in the captured image. Emission detection processing (S150: specifying means), the carried out sequentially. These various processes will be described in detail later.

  Subsequently, shape information that is information relating to the shape of the traveling lane detected in the processing of S150 is output (S160: storage means). Here, the shape information includes each coordinate value of the edge point detected as the contour component, an offset value indicating the position of the vehicle with respect to the center in the width direction of the traveling lane, a pitch angle indicating the vertical inclination of the vehicle, traveling The yaw angle that represents the inclination angle of the vehicle in the left-right direction with respect to the traveling direction of the vehicle in the lane, the radius of the travel lane, the travel lane width, and the like are included.

  Further, the shape information is stored in the RAM 13 during the process of S160. When such a process ends, the travel lane recognition process ends.

  Next, details of edge extraction processing (S120), road type determination processing (S130), road shape constant setting processing (S140), and lane detection processing (S150) will be described.

[Edge extraction processing]
In the edge extraction process, as shown in FIG. 3A, first, a differential value is calculated using a differential filter for each horizontal line (all pixels having the same vertical coordinate value) in the captured image (S210). ). That is, the change rate of the luminance value of an adjacent pixel is calculated for a plurality of pixels constituting the horizontal line. When the camera 15 is a color camera, the RGB signal output from the color camera or the change rate of the color difference signal when the RGB signal is converted into a luminance signal and a color difference signal may be calculated. .

  Subsequently, it is determined whether or not the calculated differential value is equal to or greater than a predetermined threshold value. If the differential value is equal to or greater than the threshold value, it is assumed that the luminance value of the adjacent pixel has greatly changed, and the coordinate value of the corresponding pixel is determined as an edge. A point (edge component) is registered in the RAM 13 (S220). After performing the processing of S210 and S220 for all the pixels in the captured image, the edge extraction processing ends.

[Road type judgment processing]
Next, in the road type determination process, as shown in FIG. 3B, first, an individual determination process is performed (S310). In the individual determination process, as shown in FIG. 4, a pedestrian crossing detection process (S400: pedestrian crossing detection means) for detecting a pedestrian crossing from a captured image, information on the vehicle speed of the vehicle is acquired, and the vehicle speed is equal to or higher than a reference vehicle speed. Vehicle speed detection processing (S450: vehicle speed determination means) for determining whether or not the vehicle lane width information of the traveling lane in which the vehicle is traveling is acquired, and whether or not the lane width is equal to or greater than the reference lane width The traveling lane width detecting process (S500: lane width determining means) is sequentially performed.

  In the pedestrian crossing detection process (S400), the number of edge points is calculated for each horizontal line (left and right direction) in the captured image (S410), and a horizontal line having a predetermined number (for example, 10) or more edge points travels on the vehicle. It is determined whether or not a predetermined distance (for example, about 2 m) is extended in the direction (S420). If a horizontal line having a predetermined number or more of edge points extends a predetermined distance or more (S420: YES), the pedestrian crossing flag in the RAM 13 indicating that a pedestrian crossing has been detected is set to ON (S430), The detection process ends. On the other hand, if the horizontal lines having a predetermined number or more edge points extend less than the predetermined distance (S420: NO), the pedestrian crossing flag is set to the OFF state (S440), and the pedestrian crossing detection process is terminated.

  In the vehicle speed detection process (S450), information on the vehicle speed of the vehicle is acquired from the vehicle speed sensor 20, and the vehicle speed is compared with a reference vehicle speed (for example, 80 km / h) on the highway (S460). If the vehicle speed is less than the reference vehicle speed (S460: YES), the vehicle speed flag in the RAM 13 indicating that the vehicle speed is less than the reference vehicle speed is set to ON (S470), and the vehicle speed detection process is terminated. If the vehicle speed is equal to or higher than the reference vehicle speed (S460: NO), the vehicle speed flag is set to the OFF state (S480), and the vehicle speed detection process is terminated.

  In the travel lane width detection process (S500), information on the travel lane width on the road on which the vehicle is traveling is acquired, and a reference lane width (for example, 3 m) that is a value near the upper limit value of the travel lane width and the lane width of the general road. The degree is compared (S510). In this process, the travel lane width information may refer to the detection result (shape information) by the previous travel lane recognition process, or the travel lane width information on the road on which the vehicle travels from the navigation ECU 30. May be obtained.

  If the travel lane width is less than the reference lane width (S510: YES), the lane width flag in the RAM 13 indicating that the travel lane width is less than the reference lane width is set to ON (S520), and the travel lane width is detected. The process ends. If the travel lane width is equal to or greater than the reference lane width (S510: NO), the lane width flag is set to the OFF state (S530), and the travel lane width detection process is terminated. When the travel lane width detection process ends, the individual determination process ends.

  When the individual determination process ends, the process returns to FIG. 3B, and the state of the general road mode flag is determined (S320). Here, the “general road mode flag” is a flag indicating that the type of road on which the vehicle is currently traveling is a general road, and is set to ON at the first activation of this process. . When the “general road mode flag” is in an OFF state, it is indicated that the road type is an expressway.

  If the general road mode flag is ON (S320: YES), the state of the pedestrian crossing flag, vehicle speed flag, and lane width flag is determined (S330). If all of the pedestrian crossing flag, the vehicle speed flag, and the lane width flag are in the OFF state (S330: YES), the general road mode flag is set to the OFF state, assuming that the road on which the vehicle is traveling is an expressway. (S340), the road type determination process is terminated.

  If any one of the pedestrian crossing flag, the vehicle speed flag, and the lane width flag is ON (S330: NO), the road on which the vehicle is traveling is assumed to be a general road, and the general road mode flag is set. The road type determination process is terminated without changing the state (while keeping the ON state).

  In the process of S320, if the general road mode flag is in the OFF state (S320: NO), the states of the pedestrian crossing flag, the vehicle speed flag, and the lane width flag are determined (S350). If any of the pedestrian crossing flag, the vehicle speed flag, and the lane width flag is on (S350: YES), the road on which the vehicle is traveling is assumed to be a general road, and the general road mode flag is on. (S360), and the road type determination process ends.

  If all of the pedestrian crossing flag, the vehicle speed flag, and the lane width flag are in an OFF state (S350: NO), the state of the general road mode flag is changed assuming that the road on which the vehicle is traveling is an expressway. If there is no (leave in the OFF state), the road type determination process is terminated.

[Road shape constant setting process]
Next, in the road shape constant setting process, as shown in FIG. 5, it is first determined whether or not it is in a reset state (S610). Here, the “reset state” refers to a state in which all data (shape information) relating to the shape of the driving lane in the past that should be stored in the RAM 13 is discarded.

  If not in the reset state (S610: NO), the process immediately proceeds to S650, which will be described later. If it is in the reset state (S610: YES), an initial value for road parameter calculation is set. This initial value is used as a reference value when detecting the shape of the traveling lane as a computer process, and is used as a value indicating the shape of the traveling lane in the past during a Kalman filter process described later.

  Also, different values are set as initial values depending on the road type. In this case, first, the state of the general road mode flag is determined (S620). If the general road mode flag is ON (S620: YES), the driving lane width is set to 2.8 m, the yaw angle is 0 degree, the curve radius is 15000 m (that is, a straight line), and the pitch angle is 0 degree (S630).

  If the general road mode flag is OFF (S620: NO), the yaw angle, curve radius, and pitch angle are set to the same values as when the general road mode flag is ON. The width is set to 3.5 m, which is a larger value than when the road type is a general road (S640).

  Subsequently, each threshold value used when determining the shape of the traveling lane is set. In this process, first, the state of the general road mode flag is determined again (S650). If the general road mode flag is ON (S650: YES), the values of various parameters are set so as to match the design specifications of the general road and the running state of the vehicle on the general road.

  That is, the set lane width that is the target value of the travel lane width is set to 2.8 m (S660), and the minimum lane width that is the minimum value allowed as the travel lane width on a general road is set to 2.2 m (S670). . Further, the range of the yaw angle is set to ± 14 degrees (S680), and the minimum curve radius is set to 80 m (S690). When these processes are finished, the road shape constant setting process is finished.

  On the other hand, if the general road mode flag is in the OFF state (S650: NO), the values of various parameters are set so as to match the design specifications of the highway and the running state of the vehicle on the highway.

  That is, the set lane width is set to 3.5 m which is wider than the setting of the general road (S710), and the minimum lane width on the expressway is set to 2.8 m (S720). In consideration of the fact that the running speed is faster than that of ordinary roads, the yaw angle range is set to ± 3 degrees, which is narrower than the value of ordinary roads (S730), and the minimum curve radius is larger than that of ordinary roads. It is set to 250 m (S740). When these processes are finished, the road shape constant setting process is finished.

  Note that the value of the set lane width set in the road shape constant setting process is used in the process of S820 of the lane detection process, and the values of the minimum lane width, the yaw angle range, and the curve radius are guard processing (see FIG. 10).

[Lane detection processing]
Next, in the lane detection process, a well-known Hough conversion process is performed based on the registered edge points to calculate a straight line pattern (S810), and the settings set in the road shape constant setting process A combination of left and right linear patterns closest to the lane width is selected (S820). At this time, an area closer to the center of the captured image than the left and right linear patterns is specified as an area indicating a travel lane.

  Here, the process of S820 will be described with reference to FIG. FIG. 7A is an explanatory diagram illustrating an example of a captured image on a general road, and FIG. 7B is an explanatory diagram illustrating an example of a captured image on an expressway.

  As shown in FIG. 7 (a), the actual driving lane width is 2.8 m as if the vehicle is driving on a general road, and the guardrail is positioned 3.5 m from the center line (shown by a broken line). Consider the case where exists. If the travel lane detection system 1 does not know the type of road (in particular, the width of the travel lane here) on which the vehicle travels, the position of 3.5 m from the center line, which is the travel lane width of the expressway, is processed in S820. It is conceivable that the combination of the guard rail and the center line existing in the vehicle is detected as the outline (width) of the traveling lane.

  However, in the case of this embodiment, the set lane width is set to 2.8 m, which is a general road condition, and a road close to this condition can be detected preferentially. It will not be detected as the outline of the travel lane.

  Next, as shown in FIG. 7 (b), the actual driving lane width is 3.5 m as if the vehicle is driving on a highway, and 2.8 m from the center line (indicated by a broken line). Consider a case where a bright area exists between the shadow of the guardrail at the position.

  If the traveling lane detection system 1 does not know the type of road on which the vehicle travels, the shadow and shadow of the guardrail existing at a position 2.8 m from the center line, which is the traveling lane width of a general road, are processed in S820. It is conceivable that a combination of a bright area between the center line and the center line is detected as the outline of the traveling lane.

  However, in the case of the present embodiment, the set lane width is set to 3.5 m which is a condition of the expressway, and those close to this condition can be detected preferentially. A bright area between the shadows is not detected as the outline of the traveling lane.

  That is, in the process of S820, an appropriate edge point can be detected as the outline of the traveling lane based on the road type detected in advance.

  Subsequently, it is determined whether or not a combination of linear patterns located on both sides of the vehicle has been detected (S830). If a combination of straight line patterns can be detected (S830: YES), the edge points constituting the straight line pattern are set as Kalman input points as values indicating the shape of the current driving lane (S840). Then, the lost counter indicating the number of times the shape of the traveling lane on at least one side cannot be identified continuously is cleared (that is, set to 0) (S850), and the process proceeds to S1090 described later.

  If the combination of straight line patterns is not detected in the process of S830 (S830: NO), the state of the general road mode flag is determined (S860). If the general road mode flag is OFF (S860), the lost counter is incremented (S870).

  Normally, on highways, there are generally white lines and other paint that define the shape of the lane on both sides. In this case, however, this paint cannot be detected, so the shape of the lane is It is processed as if it could not be recognized. Information about whether or not the shape of the travel lane has been recognized is stored in the RAM 13 for at least 10 frames (10 cycles).

  Subsequently, the value of the lost counter is compared with a predetermined comparison value (for example, 10) (S880). If the value of the lost counter is equal to or greater than the comparison value (S880: YES), the reset state is set (S890). That is, all the data relating to the shape of the driving lane in the past stored in the RAM 13 is discarded (deleted).

  Then, the lost counter is cleared (S900), and the process proceeds to S1090 described later. On the other hand, if the value of the lost counter is less than the comparison value in the process of S880 (S880: NO), the process proceeds to S940 described later.

  If the general road mode flag is in the ON state in the process of S860 (S860: NO), a safe zone determination process for determining the presence or absence of a safe zone in the captured image is performed (S910: safe zone detection means). . Here, the safety zone refers to a region where the interior of a region where the vehicle should not travel is painted with a plurality of diagonal lines (such as in FIG. The right hatched portion in b) is shown.

  In the safety zone determination process, as shown in FIG. 8A, the detection result of the linear pattern calculated in the process of S810 is read (S1110), and the clockwise direction with respect to the traveling direction of the vehicle is set to be positive. Sometimes, it is parallel and has a certain length (for example, the number of pixels corresponding to 1 m) having an angle within a certain range of positive or negative with respect to the traveling direction of the vehicle (for example, within a range of 30 to 60 degrees). The number of the straight line patterns is calculated (S1120).

  If the number of the straight line patterns is equal to or greater than the number (for example, 3) as a criterion for determination (S1130: YES), a safety zone flag indicating that a safety zone has been detected is set to ON (S1140). The determination process ends. If the number of straight line patterns is less than the reference number (S1130: NO), the safety zone flag is set to the OFF state (S1150), and the safety zone determination process ends. The state of the safe zone flag is accumulated in the RAM 13 for at least 10 frames (10 cycles).

  When such a safety zone determination process is completed, an intersection determination process for determining the presence or absence of an intersection in the captured image is performed (S920: intersection detection means). In the intersection determination process, as shown in FIG. 9, data relating to the shape of the driving lane for the past 5 frames (for the past 5 cycles) is read from the RAM 13 (S1210), and the safety zone determination results for the past 10 frames are stored in the RAM 13 (S1220).

  Then, the state of the intersection flag is determined (S1230). If the intersection flag is in the OFF state (S1230: YES), the intersection counter is cleared (S1240), whether or not the safety zone flag is set to the ON state even once within the past 10 frames (S1250), and the past 5 frames. It is determined whether or not the detection result of the shape of the traveling lane is lost (missed) continuously for 3 frames or more (S1260).

  If the safety zone flag is set to ON within the past 10 frames (S1250: YES), and the detection result of the shape of the traveling lane is lost continuously for 3 frames or more within the past 5 frames (S1260: (YES), an intersection flag indicating that an intersection has been detected is set to an ON state (S1270), and the intersection determination process is terminated. In addition, the safety zone flag has not been set to the ON state within the past 10 frames (S1250: NO), or the travel lane shape detection result is continuously lost within less than 3 frames within the past 5 frames. If not (S1260: NO), the intersection determination process is terminated with the intersection flag set to the OFF state.

  On the other hand, if the intersection flag is in the ON state in the processing of S1230 (S1230: NO), the intersection counter is incremented (S1310), and the intersection counter and a predetermined comparison value (for example, when passing the intersection in normal driving) Is compared with a counter value corresponding to the required time required for (30) (S1320). If the intersection counter is equal to or smaller than the comparison value (S1320: YES), the intersection determination process is terminated while the intersection flag remains set to ON. If the intersection counter is larger than the comparison value (S1320: NO), the intersection flag is set to OFF (S1330), and the intersection determination process is terminated.

  When such an intersection determination process ends, the process returns to FIG. 6 to determine the state of the intersection flag and the state of the safety zone flag (S930, S960: second discard prohibiting means). If the intersection flag is ON (S930: YES), the data (shape information) indicating the shape of the previous traveling lane is read from the RAM 13, and the read value is used as the value indicating the shape of the current traveling lane as the Kalman input point. (S940).

  If the intersection flag is OFF and the safety zone flag is ON (S930: NO, S960: YES), the straight line on the side where the safety zone is not detected in the straight line pattern calculated in S810. A pattern (one-side straight line pattern) is extracted (S970).

  At this time, assuming that the vehicle is traveling in the center of the traveling lane, a linear pattern close to a distance corresponding to half of the set lane width set in the road shape constant setting process is extracted from the center of the captured image. Alternatively, based on the offset value included in the data (shape information) indicating the shape of the previous traveling lane, the position of the traveling lane of the vehicle is identified, and the vehicle travels at the identified position. As a result, a straight line pattern in the vicinity of the position where the outline (white line) of the traveling lane is expected is extracted.

  Note that when the process of S970 is performed, discarding data indicating the shape of the traveling lane recorded in the RAM 13 is prohibited.

  Subsequently, the extracted straight line pattern is set to the Kalman input point as a value indicating the shape of the current traveling lane (S980). When this process ends, the process proceeds to S1090.

  If the intersection flag is OFF and the safety zone flag is OFF in the processes of S930 and S960 (S930: NO, S960: NO), either the left or right side of the straight line pattern calculated in the process of S810 At the same time, an attempt is made to extract a straight line pattern (one-side straight line pattern) (S1010), and it is determined whether or not the one-side straight line pattern is detected (S1020). When extracting the one-side straight line pattern, for example, the same method as the processing in S970 can be employed.

  If the one-sided linear pattern can be extracted (S1020: YES), the one-side lost counter indicating the number of times the one-sided linear pattern cannot be extracted continuously is cleared (S1030), and the above-described processing of S980 is performed. If the one-sided linear pattern has not been extracted (S1020: NO), the one-side lost counter is incremented (S1040), and the value of the one-side lost counter is compared with a predetermined comparison value (for example, 10) (S1050). If the value of the one-side lost counter is equal to or greater than the comparison value (S1050: YES), the reset state is set (S1060), the one-side lost counter is cleared (S1070), and the process proceeds to S1090.

  If the value of the one-side lost counter is less than the comparison value (S1050: NO), data (shape information) indicating the shape of the previous travel lane is read from the RAM 13, and the read data is used as the current travel lane shape. A Kalman input point is set as the indicated value (S1080), and the process proceeds to S1090.

  Subsequently, in the process of S1090, data (shape information) indicating the shape of the past driving lane is read from the RAM 13, and a known Kalman filter process is performed using the read shape information and the value set at the current Kalman input point. Implement (S1090: correction means). When the reset state is set (when S890 and S1060 are performed), the processing of S1090 and S1100 may be omitted.

  By the process of S1090, the current detection result of the shape of the traveling lane is corrected. Then, a guard process is performed on the corrected traveling lane detection result (S1100). The guard process is a process for determining whether or not the corrected travel lane shape falls within the parameter range (reference shape range) set in the road shape constant setting process. In the guard process, the processes of S1410 to S1440 correspond to the shape determining means in the present invention.

  That is, as shown in FIG. 10, first, the travel lane width after the Kalman filter processing is compared with the minimum lane width set in the road shape constant setting processing (S1410). If the traveling lane width is equal to or smaller than the minimum lane width (S1410: NO), the reset state is set (S1450), and the guard process is terminated.

  If the travel lane width is larger than the minimum lane width (S1410: YES), the state of the safety zone flag is determined (S1420: first discard prohibiting means). If the safety zone flag is OFF (S1420: NO), it is determined whether the yaw angle after the Kalman filter process is within the yaw angle range set in the road shape constant setting process (S1430: yaw angle determination). means).

  If the yaw angle is outside the yaw angle range (S1430: NO), the reset state is set (S1450), and the guard process is terminated. If the yaw angle is within the yaw angle range (S1430: YES), the process proceeds to S1440.

  On the other hand, if the safety zone flag is ON in S1420 (S1420: YES), the process proceeds to S1440 without determining the yaw angle. That is, when the safety zone flag is in the ON state, the reset state is not set based on the determination result of the yaw angle.

  Subsequently, in the process of S1440, the curve radius of the travel lane after the Kalman filter process is compared with the minimum curve radius set in the road shape constant setting process (S1440). If the curve radius is larger than the minimum curve radius (S1440: YES), the guard process is terminated as it is. If the curve radius is equal to or smaller than the minimum curve radius (S1440: NO), the reset state is set (S1450: use prohibiting means, discarding means), and the guard process is terminated.

  In addition, when it resets in guard processing, it will be in a missing measurement state without data about the shape of the current traveling lane. That is, the use of data after the Kalman filter processing is prohibited. At this time, when the current Kalman input point is set, the data before the Kalman filter processing may be adopted as the shape of the current traveling lane.

  When such guard processing ends, the lane detection processing (see FIG. 6) also ends.

[Effect of this embodiment]
In the travel lane detection system 1 described in detail above, the travel lane detection ECU 10 detects the travel speed that is the travel state of the vehicle and the pedestrian crossing that is the environment on the road on which the vehicle travels. Based on the travel lane width, it is determined whether the road type on the road on which the vehicle travels is an expressway or a general road, and parameters relating to the allowable range of the shape of the travel lane on which the vehicle travels are set according to the road type. . Then, the traveling lane detection ECU 10 detects an edge component that matches the parameter setting of the extracted edge components (edge points) as a contour component that represents a part of the contour of the traveling lane, and uses the contour component in the captured image. Is also specified as a travel lane.

  Therefore, according to the traveling lane detection system 1 as described above, an edge component that matches the parameter setting according to the road type can be detected as a contour component, and thus the traveling lane can be detected with high accuracy.

  The travel lane detection ECU 10 determines that a pedestrian crossing has been detected when a predetermined number or more edge components are detected in the captured image in the lateral direction of the captured image. Then, the traveling lane detection ECU 10 determines the road type based on the presence or absence of a pedestrian crossing in the captured image.

  According to such a traveling lane detection system 1, the road type can be determined according to the presence or absence of a pedestrian crossing. Specifically, for example, when a pedestrian crossing is detected, it can be determined that the road type is an intersection, or it can be determined that the road type is a general road that is not a highway.

  Further, the travel lane detection ECU 10 acquires information on the vehicle speed of the vehicle from the vehicle speed sensor 20, and determines whether the vehicle speed is equal to or higher than the reference vehicle speed. Further, the traveling lane detection ECU 10 acquires information on the lane width of the traveling lane in which the vehicle is traveling, and determines whether the lane width is equal to or larger than the reference lane width. The traveling lane detection ECU 10 determines that no pedestrian crossing has been detected, determines that the lane width of the traveling lane is greater than or equal to the reference lane width, and determines that the vehicle speed of the vehicle is greater than or equal to the reference vehicle speed. In addition, it is determined that the road type is a highway, and in cases other than the above, it is determined that the road type is a general road.

  Therefore, according to the traveling lane detection system 1 as described above, since the conditions of the presence / absence of a pedestrian crossing, the lane width of the traveling lane, and the vehicle speed are used in combination, whether the road type is a general road or a highway Can be determined satisfactorily.

  Further, when the road type is a highway, the travel lane detection ECU 10 indicates a travel lane width range that is an allowable range of the travel lane width as one of the parameters, and a case where the road type is a general road. Compare and set wider.

  According to such a travel lane detection system 1, an appropriate travel lane width range can be set according to the road type, so that there is an edge component that is not a contour component in the vicinity of the contour component in the captured image. In addition, the contour component can be detected with high accuracy.

  The travel lane detection ECU 10 is configured to repeatedly detect travel lanes, records shape information, which is information related to the detected shape of the travel lane, in the RAM 13, and based on the shape information recorded in the RAM 13, the current travel lane is detected. Correct the shape of the lane. Specifically, when the travel lane detection ECU 10 can identify the shape of the current travel lane, the travel lane detection ECU 10 corrects the current travel lane shape based on the shape information recorded in the RAM 13, and If the shape cannot be specified, correction is performed to complement the current travel lane shape based on the shape information recorded in the RAM 13.

  Therefore, according to such a traveling lane detection system 1, even if the shape of the traveling lane is erroneously detected, the current traveling lane is based on the shape information that is information on the past traveling lane shape. Can be corrected. Even if the shape of the travel lane cannot be detected, the current shape of the travel lane can be supplemented based on the shape information. Therefore, the traveling lane can be detected with high accuracy.

  Further, the travel lane detection ECU 10 determines whether or not the shape of the travel lane is within the reference shape range, and when it is determined that the shape of the travel lane is outside the reference shape range, the corrected travel lane The use of shapes is prohibited.

  According to such a traveling lane detection system 1, when it is estimated that the accuracy of detecting a traveling lane is deteriorated by the correction using the past shape information, the use of the past shape information can be prohibited. .

  Further, the travel lane detection ECU 10 sets the reference shape range according to the road type.

  According to such a traveling lane detection system 1, the reference shape range can be set according to the shape of the traveling lane estimated by the road type, so that the accuracy of detecting the traveling lane can be improved.

  Further, the travel lane detection ECU 10 discards the content recorded by the RAM 13 when the use of the corrected travel lane shape is prohibited.

  According to such a traveling lane detection system 1, it is possible to prevent the shape information recorded in the RAM 13 from being used in a future process with a simple process.

  In addition, the travel lane detection ECU 10 determines whether or not a yaw angle representing a left-right inclination angle of the vehicle with respect to the traveling direction of the vehicle in the identified travel lane is within a reference yaw angle range. When it is determined that the yaw angle is outside the reference yaw angle range, the contents recorded by the RAM 13 are discarded.

  According to such a traveling lane detection system 1, it is possible to set whether correction is performed using past shape information depending on whether the yaw angle is within the reference yaw angle range.

  Furthermore, the travel lane detection ECU 10 has a predetermined number or more of edge components in the left-right direction in the captured image, and these edge components extend a predetermined distance or more while forming a certain angle with respect to the travel direction of the vehicle. In this case, it is determined that the safe zone has been detected, and when the safe zone is detected, it is prohibited to discard the contents recorded by the RAM 13.

  Therefore, according to such a traveling lane detection system 1, when an abnormality in the yaw angle at a point where the number of lanes changes is detected, it is possible to retain the recorded content by the RAM 13 as not abnormal.

  In addition, the travel lane detection ECU 10 determines that an intersection has been detected when the travel lane cannot be specified within a certain time after the safety zone is detected. Then, the traveling lane detection ECU 10 prohibits discarding the recorded contents of the RAM 13 when an intersection is detected.

  According to such a traveling lane detection system 1, even when the traveling lane cannot be specified because the vehicle is traveling at an intersection, the estimation result of the traveling lane shape is out of the reference shape range. Since the contents recorded by the RAM 13 are not discarded, the process of estimating the shape of the traveling lane can be continued.

[Other Embodiments]
Embodiments of the present invention are not limited to the above-described embodiments, and can take various forms as long as they belong to the technical scope of the present invention.

  For example, in the above embodiment, as the “road type”, it is detected whether the road on which the vehicle travels is an expressway or a general road other than an expressway. Any information may be detected as long as the information is used. For example, the road on which the vehicle travels is an uphill, downhill, right curve, left curve, straight road, etc., using information from an acceleration sensor, a yaw rate sensor, or a navigation device. You may make it detect that there exists.

  In the above embodiment, the determination result of whether or not it is an intersection and the determination result of the presence or absence of a safety zone are used only for the process of detecting the driving lane, but it is used for the process of determining the road type. May be.

1 is a block diagram (a) showing a schematic configuration of a traveling lane detection system 1 to which the present invention is applied, and a block diagram (b) showing a configuration of a traveling lane detection ECU 10. FIG. It is a flowchart which shows a driving | running | working lane recognition process. It is the flowchart (a) which shows the edge extraction process among driving | running | working lane recognition processes, and the flowchart (b) which shows a road classification determination process. It is a flowchart which shows an individual determination process. It is a flowchart which shows a road shape constant process. It is a flowchart which shows a lane detection process. It is explanatory drawing (a) which shows an example of the captured image in a general road, and explanatory drawing (b) which shows an example of the captured image in a highway. It is flowchart (a) which shows the safe zone detection process among lane detection processes, and explanatory drawing (b) which shows an example of a safe zone. It is a flowchart which shows an intersection determination process. It is a flowchart which shows a guard process.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... Travel lane detection system, 5 ... Communication line, 10 ... Travel lane detection ECU, 11 ... CPU, 13 ... RAM, 15 ... Camera, 20 ... Vehicle speed sensor, 30 ... Navigation ECU, 35 ... ETC onboard equipment, 40 ... Vehicle Control ECU.

Claims (12)

A travel lane detection device that is mounted on a vehicle and detects a travel lane in which the vehicle travels,
Image acquisition means for acquiring a captured image in which the traveling direction of the vehicle is captured;
Edge component extraction means for extracting edge components in the acquired captured image;
Road type determination means for determining a road type on a road on which the vehicle is driven based on at least one of a traveling state of the vehicle and an environment on a road on which the vehicle is driven;
Parameter setting means for setting a parameter relating to the allowable range of the shape of the travel lane on which the vehicle travels according to the road type;
An edge component that matches the setting of the parameter among the extracted edge components is detected as a contour component that represents a part of the contour of the travel lane, and a region on the center side of the contour component in the captured image is detected. A specifying means for specifying the travel lane;
A travel lane detection apparatus comprising: In the traveling lane detection device according to claim 1,
A pedestrian crossing detection unit is provided that determines that a pedestrian crossing has been detected when a predetermined number or more of edge components that are longer than a predetermined distance in the travel direction of the vehicle in the captured image are detected in the left-right direction of the captured image. ,
The road lane detection device, wherein the road type determination means determines the road type based on the presence or absence of a pedestrian crossing in the captured image. In the traveling lane detection device according to claim 2,
Vehicle speed determination means for acquiring vehicle speed information of the vehicle and determining whether the vehicle speed is equal to or higher than a reference vehicle speed;
Lane width determination means for acquiring information on the lane width of a traveling lane in which the vehicle is traveling, and determining whether the lane width is equal to or greater than a reference lane width;
With
The road type determining means determines that no pedestrian crossing has been detected by the pedestrian crossing detecting means, and the lane width determining means determines that the lane width of the traveling lane in which the vehicle is traveling is greater than or equal to the reference lane width. When the vehicle speed determining means determines that the vehicle speed is equal to or higher than the reference vehicle speed, the road type is determined to be a highway. In other cases, the road type is a general road. A traveling lane detection device characterized by determining that there is a vehicle. In the traveling lane detection device according to claim 3,
When the road type is a highway, the parameter setting means indicates a driving lane width range that is an allowable range of the driving lane width as one of the parameters, and the road type is a general road. A traveling lane detection device characterized in that it is set wider than the case. In the traveling lane detection device according to any one of claims 1 to 4,
The specifying means is configured to repeatedly detect a traveling lane,
Storage means for storing shape information, which is information relating to the shape of the traveling lane detected by the specifying means, in the recording means;
Correction means for correcting the shape of the current traveling lane based on the shape information recorded in the recording means;
A travel lane detection apparatus comprising: In the traveling lane detection device according to claim 5,
Shape determining means for determining whether or not the shape of the travel lane after correction by the correcting means falls within a reference shape range;
Use prohibiting means for prohibiting use of the shape of the travel lane corrected by the correcting means when the shape determining means determines that the shape of the travel lane is outside the reference shape range;
A travel lane detection apparatus comprising: In the traveling lane detection device according to claim 6,
Shape range setting means for setting the reference shape range according to the road type,
A travel lane detection apparatus comprising: In the traveling lane detection device according to claim 6 or 7,
A travel lane detection apparatus comprising: discarding means for discarding the recorded content by the recording means when the use prohibiting means prohibits the use of the travel lane shape corrected by the correcting means. In the traveling lane detection device according to claim 8,
The shape determination means includes
A yaw angle determination means for determining whether or not a yaw angle representing an inclination angle of the vehicle in the left-right direction with respect to the traveling direction of the vehicle in the identified travel lane is within a reference yaw angle range;
The discarding means discards the content recorded by the recording means when the yaw angle of the vehicle is determined to be outside the reference yaw angle range by at least the yaw angle determining means. apparatus. In the traveling lane detection device according to claim 9,
In the captured image, when a predetermined number or more of edge components that form a predetermined angle with respect to the traveling direction of the vehicle and continue for a predetermined distance or more are detected in the left-right direction of the captured image, a safety zone is detected. Safety zone detection means for judging,
A first discard prohibiting unit for prohibiting the discarding unit from discarding the recorded content by the recording unit when a safe zone is detected by the safe zone detecting unit;
A travel lane detection apparatus comprising: In the traveling lane detection apparatus according to claim 8 or 9,
In the captured image, when a predetermined number or more of edge components that form a predetermined angle with respect to the traveling direction of the vehicle and continue for a predetermined distance or more are detected in the left-right direction of the captured image, a safety zone is detected. Safety zone detection means for judging,
An intersection detection means for determining that an intersection has been detected when the identification means cannot identify a driving lane within a certain time after the safety zone detection means detects a safety zone;
A second discard prohibiting unit for prohibiting the discarding unit from discarding the recorded content by the recording unit when an intersection is detected by the intersection detecting unit;
A traveling lane detection device comprising:   A travel lane detection program for a computer to realize functions as respective means constituting the travel lane detection device according to any one of claims 1 to 11.