JP2015064735A - Apparatus and method for estimating vehicle position - Google Patents

Abstract

PROBLEM TO BE SOLVED: To estimate a position and attitude of a vehicle between lane markings even if the lane markings cannot be observed in some of captured images.SOLUTION: An edge extraction section 107 extracts an edge of a lane marking from an image obtained by capturing left and right lane markings for forming lanes. An edge coordinate holding section 109 holds the extracted edge in association with coordinates on the image. An edge coordinate update section 110 moves the held edge opposite to a vehicle traveling direction by a vehicle moving amount corresponding to a vehicle speed, for update. A lane marking recognition section 113 recognizes a lane marking from the held edge. A position attitude estimation section 114 estimates a position and attitude of a vehicle 101 between the lane markings.

Description

  The present invention relates to a host vehicle position estimation device and a host vehicle position estimation method.

  Conventionally, a technique for recognizing a white line on a traveling road using cameras mounted on the left and right sides of the vehicle and estimating the lateral position and yaw angle of the vehicle in the white line has been disclosed (Patent Document 1). In Patent Document 1, the distance from the front and rear wheels to the white line is obtained based on the white line included in the imaging area of the camera and the images of the front and rear wheels of the vehicle, and the lateral position and yaw angle of the vehicle within the white line are determined. Estimated.

JP 2001-283390 A

  However, since the above-described prior art estimates the lateral position and yaw angle of the vehicle based on the white line and wheels in the image, it is necessary that the white line is always shown in all captured images. There is a problem that it cannot be estimated when the white line cannot be observed in the image of the part.

  The present invention has been made in view of the above problems, and the object of the present invention is to determine the position and posture of the vehicle in the white line even when the white line cannot be observed in some of the captured images. Is to estimate.

  The host vehicle position estimation apparatus according to an aspect of the present invention extracts white line edges from an image obtained by imaging the left and right white lines that divide a traveling lane. Further, the extracted edge is held in association with the coordinates on the image, and the held edge is updated by moving to the opposite side in the vehicle traveling direction by the amount of movement of the vehicle according to the vehicle speed. Then, the white line is recognized from the held edge, and the position and orientation of the vehicle in the recognized white line are estimated.

  According to the present invention, even if the white line cannot be observed in some of the captured images, the white line can be recognized from the held edge, so that the position and orientation of the vehicle within the recognized white line can be estimated. it can.

FIG. 1 is a system configuration diagram of the host vehicle position estimation apparatus according to the first embodiment of the present invention. FIG. 2 is a system block diagram of the host vehicle position estimation apparatus according to the first embodiment of the present invention. FIG. 3 is a diagram illustrating an imaging region of the camera. FIG. 4 is a diagram illustrating an example of an image captured by a camera using a fisheye lens. FIG. 5 is a diagram for extracting an edge from a captured image. FIG. 6 is a diagram for converting edge coordinates from an image coordinate system to a world coordinate system. FIG. 7 is a diagram for explaining the update of edge coordinates. FIG. 8 is a diagram for explaining white line recognition. FIG. 9 is a diagram for explaining edge coordinates necessary for recognizing a white line during high-speed traveling. FIG. 10 is a diagram for explaining edge coordinates necessary for recognizing a white line during low-speed traveling. FIG. 11 is a diagram for explaining a method for deleting edge coordinates. FIG. 12 is a flowchart showing an example of the vehicle position estimation method according to the first embodiment of the present invention. FIG. 13 is a diagram for explaining an edge coordinate deletion method according to the second embodiment of the present invention. FIG. 14 is a system block diagram of the host vehicle position estimation apparatus according to the third embodiment of the present invention.

  Hereinafter, specific embodiments of the present invention will be described in detail with reference to the drawings. In the description of the drawings, the same parts are denoted by the same reference numerals and description thereof is omitted.

[First Embodiment]
FIG. 1 is a system configuration diagram of the vehicle position estimation apparatus according to the first embodiment. The own vehicle position estimation device includes a front camera 102, a right camera 103, a left camera 104, a computer 105, and a lane departure prevention alarm system 106.

  The front camera 102 is mounted in front of the vehicle 101 at a position / posture of a height h1 from the ground and an angle θ1 from the horizontal to the downward direction. The right camera 103 and the left camera 104 are attached to the side of the vehicle 101 at a position / posture of a height h2 from the ground and an angle θ2 downward from the horizontal. These cameras are cameras having an image sensor such as a CCD or a CMOS, for example, and take images continuously in time and output the taken images. These cameras are fisheye cameras in which a fisheye lens (or a wide-angle lens) is attached to the front surface of the image sensor in order to ensure a wide detection area. A similar camera may be mounted not only on the front and side of the vehicle 101 but also on the rear.

  The computer 105 includes a microcomputer including a CPU, a RAM, a ROM, various operators, and the like. The computer 105 takes in images captured by the front camera 102, the right camera 103, and the left camera 104, and executes predetermined image processing.

  The lane departure prevention warning system 106 warns the driver according to the image processing result of the computer 105. For example, the alarm generates an alarm sound or various voice guidance, or displays a warning screen on the car navigation device.

  FIG. 2 is a system block diagram of the vehicle position estimation apparatus. As shown in FIG. 2, the host vehicle position estimation apparatus includes a front camera 102, a right camera 103, a left camera 104, a calculator 105, a vehicle speed sensor 112, and a lane departure prevention alarm system 106. Further, the computer 105 includes an edge extraction unit 107, a coordinate conversion unit 108, an edge coordinate holding unit 109, an edge coordinate update unit 110, an edge coordinate deletion unit 111, a white line recognition unit 113, and a position and orientation estimation unit 114. And comprising.

  The edge extraction unit 107 extracts edges by performing image processing on images captured by the front camera 102, the right camera 103, and the left camera 104 (hereinafter referred to as each camera).

  The coordinate conversion unit 108 converts the coordinates on the edge image extracted by the edge extraction unit 107 into the world coordinate system.

  The edge coordinate holding unit 109 holds the edge coordinates extracted from the image captured by each camera and the edge coordinates updated by the edge coordinate updating unit 110.

  The edge coordinate updating unit 110 moves and updates the edge coordinates held by the edge coordinate holding unit 109 by the vehicle speed measured by the vehicle speed sensor 112.

  The edge coordinate deleting unit 111 deletes the edge coordinates updated by the edge coordinate updating unit 110 based on a predetermined condition.

  The vehicle speed sensor 112 is a device for observing the vehicle speed and is attached to the vehicle 101.

  The white line recognition unit 113 recognizes a white line from the edge coordinates held by the edge coordinate holding unit 109.

  The position / orientation estimation unit 114 estimates the position and orientation of the vehicle 101 within the white line recognized by the white line recognition unit 113.

  FIG. 3 is a diagram illustrating an imaging area of each camera. As shown in FIG. 3, the front camera 102 images the imaging area 201 in front of the vehicle at a predetermined angle of view. Similarly, the right camera 103 and the left camera 104 respectively capture the image areas 202 and 203 on the side of the vehicle at a predetermined angle of view. At this time, the left and right white lines that divide the lane in which the vehicle 101 travels can be imaged within the angle of view of the front camera 102. The white line here includes any white line that divides a lane such as a solid line or a broken line.

  FIG. 4 is a diagram illustrating an example of an image captured by each camera. 4A to 4C are images captured by the left camera 104, the front camera 102, and the right camera 103 in order. Since each camera is equipped with a fisheye lens, these images are images in which the vicinity of the camera is stretched, as shown in FIGS.

  FIG. 5 is a diagram for extracting an edge in an image captured by each camera. The edge extraction unit 107 performs image processing on the images shown in FIGS. 5A to 5C to extract edges existing in the image. The extracted edges are represented by pixels of .largecircle., .DELTA., And .quadrature. As shown in FIGS. In addition, the extracted edge is associated with the extracted time. Note that, when extracting an edge, it is preferable that an edge extraction inner region is determined in advance. For example, since the lens edge of the camera is a fixed position in the captured image, the edge of the camera lens edge is not detected by removing the lens edge of the camera from the edge detection inner area in advance. Further, it is determined that an edge far from the camera is not processed. By setting in this way, it is possible to prevent the edge extraction unit 107 from extracting unnecessary edges.

FIG. 6 is a diagram for converting the image coordinate system to the world coordinate system. As shown in FIGS. 6A to 6C, the edge extraction unit 107 extracts edges in the images captured by the cameras, and extracts the horizontal direction of the image as the x axis and the vertical direction of the image as the y axis. The edge is output in association with the coordinates (x _i , y _i ) on the image. Since there are usually a plurality of extracted edges, there are also a plurality of edge coordinates. The plurality of edge coordinates are referred to as an edge coordinate group. The coordinate conversion unit 108 converts the edge coordinate group (x _i , y _i ) output from the edge extraction unit 107 into the world coordinate system (XW, YW, ZW). When converted into the world coordinate system, the edge pixels shown in FIGS. 6A to 6C are associated with the vehicle 101 as shown in FIG. The world coordinate system of this embodiment is a right-handed coordinate system, and the XW axis is perpendicular to the vehicle traveling direction. The YW axis is the vehicle vertical direction, and the ZW axis is the vehicle traveling direction. The coordinate conversion from the image coordinate system to the world coordinate system can be performed using, for example, the method described in “Computer Vision -Geometry of Viewing Angles” (author: Atsushi Sato). Specifically, if the camera internal parameters and external parameters (camera mounting height, posture (yaw angle, pitch angle, roll angle)) are measured in advance, arbitrary coordinates in the image are converted to world coordinates. it can. The edge coordinate group of the world coordinate system thus converted is held in the edge coordinate holding unit 109. Note that since the conversion from the two-dimensional image coordinate system to the three-dimensional world coordinate system is performed, the YW value corresponding to the vertical direction of the world coordinate system is restored to zero. YW = 0 means a road surface point.

  In addition, when converting from the image coordinate system to the world coordinate system, the conversion calculation may be executed every time, but it is more preferable if the calculation time can be shortened. As the method, for example, assuming that the internal parameters and external parameters of the camera are fixed, a conversion table from the image coordinate system to the world coordinate system is created. Here, the internal parameters of the camera are always fixed, but the external parameters change depending on the vehicle behavior (particularly, the pitch angle). Therefore, if the external parameters are handled as fixed, an error may occur. However, if it can be assumed that the vehicle 101 is moving straight as in the case of an expressway, even if this error is ignored, the influence is small. Therefore, when it can be assumed that the vehicle 101 is moving straight, a conversion table can be created to shorten the conversion calculation time from the image coordinate system to the world coordinate system.

  FIG. 7 is a diagram illustrating the update of the edge coordinate group. FIG. 7A illustrates the vehicle 101 and the imaging regions 202 and 203 of the right camera 103 and the left camera 104 as an overhead view. Here, in order to simplify the description, the imaging region 201 of the front camera 102 is omitted. Further, it is assumed that the vehicle 101 moves straight in the positive direction of the ZW axis at a constant speed V [km / h]. As shown in FIG. 7A, the edge extraction unit 107 extracts edges from images captured at times t = −2, −1, and 0, respectively. Edges extracted at times t = −2, −1, and 0 are respectively represented by white circles, gray circles, and black circle pixels. Since each camera captures an image at 30 fps, the time interval between times t = −2, −1, 0 is about 30 ms.

  Next, the update of the edge coordinate group will be described with reference to FIGS. FIGS. 7B to 7D correspond to the time t = −2, −1, 0 shown in FIG. First, as illustrated in FIG. 7B, the edge extraction unit 107 extracts an edge from an image captured at time t = −2. The extracted edge is represented by white circle pixels. Thereafter, the coordinate conversion unit 108 converts the image coordinates of the extracted edge from the image coordinate system to the world coordinate system. The converted edge coordinate group of the world coordinate system is (XW1, YW1, ZW1). The edge coordinate holding unit 109 holds these coordinate groups.

  Subsequently, as shown in FIG. 7C, when the time t = −1, which is about 30 ms after the time t = −2, the same processing as that at the time t = −2 is performed, and the edge coordinate group To extract. The edge coordinate group extracted at time t = −1 is (XW2, YW2, ZW2). Here, if the movement amount of the vehicle 101 from time t = −2 to time t = −1 is d [m], d = (V × 1000 ÷ 3600) ÷ 30. The edge coordinate update unit 110 updates the edge coordinate group using the movement amount d. Specifically, the edge coordinate updating unit 110 moves the edge coordinate group (XW1, YW1, ZW1) extracted at time t = −2 by the movement amount d at time t = −1 to obtain (XW1 , YW1, ZW1-d). The reason why only the value of ZW is updated is that the vehicle 101 is moving straight ahead. By updating in this way, at the time t = −1, the edge coordinate group (XW1, YW1, ZW1) moves from the edge coordinate group (XW2, YW2, ZW2) to the opposite side in the vehicle traveling direction by the movement amount d. Will do. The edge coordinate holding unit 109 holds the edge coordinate group (XW2, YW2, ZW2) extracted at time t = −1 and the updated edge coordinate group (XW1, YW1, ZW1-d).

  Subsequently, as shown in FIG. 7D, the same processing as that at time t = −2 is performed at time t = 0, which is about 30 ms after time t = −1. Extract. The edge coordinate group extracted at time t = 0 is defined as (XW3, YW3, ZW3). Further, since the movement amount of the vehicle 101 from the time t = −1 to the time t = 0 is d described above, the edge coordinate update unit 110 extracted at the time t = −1 when the time t = 0. The edge coordinate group (XW2, YW2, ZW2) and the updated edge coordinate group (XW1, YW1, ZW1-d) are moved by the movement amount d, and each edge coordinate group is moved to (XW2, YW2, ZW2-d). ), (XW1, YW1, ZW1′-d). Here, “ZW1 ′” will be described. The coordinate group (XW1, YW1, ZW1) extracted at time t = -2 was updated to (XW1, YW1, ZW1-d) at time t = -1. If the updated “ZW1-d” is replaced with ZW1 ′, it can be expressed as (XW1, YW1, ZW1′-d) when the edge coordinate group is updated at time t = 0. That is, by replacing in this way, the coordinate values of the coordinate group can be expressed in a form that is moved by the movement amount d at the time of update. The edge coordinate holding unit 109 extracts the edge coordinate group (XW3, YW3, ZW3) extracted at time t = 0, the updated edge coordinate group (XW2, YW2, ZW2-d), (XW1, YW1, ZW1 ′). -D). As described above, the edge coordinate updating unit 110 updates the edge coordinate group by moving the edge coordinate group to the opposite side in the vehicle traveling direction by the movement amount d in accordance with the elapsed time for each imaging process. Then, the edge coordinate holding unit 109 holds the updated edge coordinate group and accumulates it for each imaging. The updated edge coordinate group is associated with the updated time.

FIG. 8 is a diagram illustrating white line recognition. FIGS. 8B to 8D correspond to the time t = −2, −1, 0 shown in FIG. As shown in FIG. 8D, the white line recognition unit 113 recognizes a white line by linearly complementing the edge coordinate group held by the edge coordinate holding unit 109. In addition, optimization processing can be executed when linearly complementing edge coordinate groups. The optimization process is disclosed in, for example, Japanese Patent No. 3424334, so please refer to it if necessary. In this embodiment, the optimization process is executed as follows. As shown in FIG. 8D, assuming that the white line is a straight line model and the left white line 401 and the right white line 402 are parallel, the left white line 401 is modeled as Z = aX + b + w, and the right white line 402 is modeled as Z = aX + b. be able to. At this time, since the values of XW and ZW in the edge coordinate group are known, a, b, and w are unknown parameters to be estimated. a and b correspond to the yaw angle and lateral position of the vehicle 101 within the white line in this order, and w corresponds to the width of the left white line 401 and the right white line 402. These parameters can be obtained using existing methods such as the Simplex method and the Powell method. Here, the recognition of the white line using the edge coordinate group of the world coordinate system has been described. However, when the edge is extracted, the extracted edge is output in association with the coordinates (x _i , y _i ) on the image. Therefore, the edge coordinate group corresponding to the xy coordinate is held and updated, and the white line is You may recognize it.

  The position / orientation estimation unit 114 estimates the position and orientation of the vehicle 101 within the white line recognized by the white line recognition unit 113. In the present embodiment, the yaw angle and lateral position of the vehicle 101 are estimated when the white line recognition unit 113 executes an optimization process in which the yaw angle and lateral position are incorporated into the model. The position / orientation estimation unit 114 estimates the position and orientation of the vehicle 101 using the estimated values as they are.

  In this embodiment, the edge coordinate updating unit 110 updates and accumulates the edge coordinate group for each imaging process. The white line recognition unit 113 recognizes a white line from the accumulated edge coordinate group, but depending on the amount of the accumulated edge coordinate group, the white line recognition unit 113 may affect the accuracy of the recognized white line. Therefore, when recognizing the white line, it is necessary to appropriately erase the accumulated edge coordinate group so as not to affect the accuracy.

  The distance between the edge coordinate group necessary for recognizing the white line and the vehicle 101 differs depending on the vehicle speed. The distance between the edge coordinate group necessary for recognizing the white line and the vehicle 101 will be described by taking high speed traveling and low speed traveling as an example. FIG. 9 is a diagram for extracting an edge coordinate group during high-speed traveling. FIGS. 9B to 9D correspond to the time t = −2, −1, 0 shown in FIG. As shown in FIG. 9A, when traveling at a high speed, compared to when traveling at a low speed, which will be described later, since the moving distance for each imaging is large, there are fewer edges that can be extracted for the traveling distance. Therefore, it is preferable to recognize the white line using an edge coordinate group extracted relatively backward from the vehicle position at the current time. That is, when the edge coordinate group extracted in the past based on the distance from the vehicle position at the current time is deleted during high speed traveling, the value of the distance is large.

  FIG. 10 is a diagram for extracting an edge coordinate group during low-speed traveling. FIGS. 10B to 10D correspond to the time t = −2, −1, 0 shown in FIG. As shown in FIG. 10A, when traveling at a low speed, the moving distance for each imaging is smaller than when traveling at a high speed, so that more edges can be extracted with respect to the traveling distance. For this reason, when a white line is recognized using an edge coordinate group extracted relatively rearward during high-speed traveling, the accuracy may be deteriorated. Further, during low-speed traveling, it is assumed that the vehicle is traveling on a curve or changing lanes. If a white line is recognized in consideration of the edge coordinate group extracted in such a case, the broken line in FIG. The accuracy may be deteriorated. Therefore, when the edge coordinate group extracted in the past based on the distance from the vehicle position at the current time is deleted during low speed traveling, the value of the distance is smaller than that during high speed traveling. Thus, the distance between the edge coordinate group required for recognizing the white line and the vehicle 101 differs depending on the vehicle speed. For this reason, it is not possible to delete the edge coordinate group based on the distance from the vehicle 101.

Therefore, the distance between the edge coordinate group necessary for recognizing the white line and the vehicle 101 may be changed according to the vehicle speed. For example, on a highway, white lines are approximately 15 m apart, so if the vehicle speed exceeds 60 km / h, there may be an edge coordinate group from the current vehicle position to 15 m behind. On the other hand, on general roads, white lines are spaced at an interval of about 6 m, so if the vehicle speed is 60 km / h or less, there may be an edge coordinate group from the vehicle position at the current time to 6 m behind. Therefore, as shown in FIG. 11D, when the vehicle speed exceeds 60 km / h at time t = 0, the edge coordinate erasing unit 111 starts from the vehicle position at that time 15m behind (ZW = −15). Delete the back edge coordinate group. If the vehicle speed is 60 km / h or less, the edge coordinate group behind 6 m behind (ZW = −6) from the vehicle position at that time is deleted. Such an erasing process may be performed as shown below using, for example, If-Then. That is, if the vehicle speed at the current time detected by the vehicle speed sensor 112 is V [km / h],
If (V> 60 && ZW <-15)
The delete
Else if (V <= 60 && ZW <-6)
The delete
Else
Then Save
It becomes. As described above, the edge coordinate erasure unit 111 erases an edge coordinate group that is unnecessary when recognizing a white line according to the vehicle speed. Note that the processing shown here is merely an example, and various modifications can be made by simulation or experiment.

  Further, the edge coordinate erasure unit 111 presets the data upper limit of the edge coordinate group that can be recorded in the memory of the computer 105, and erases the extracted time in order from the oldest data so that the number of data does not exceed the upper limit. May be.

  Next, the vehicle position estimation process will be described with reference to the flowchart shown in FIG. This process starts when the system is turned on.

  First, in step S201, the computer 105 acquires an image captured by each camera.

In step S202, the edge extraction unit 107 extracts the edge in the image acquired in step S201, and outputs the extracted edge in association with the coordinate group (x _i , y _i ) on the image.

Next, in step S203, the coordinate conversion unit 108 performs coordinate conversion from the image coordinate system to the world coordinate system for the edge coordinate group (x _i , y _i ) output in step S202.

  Next, in step S204, the edge coordinate holding unit 109 holds the edge coordinate group subjected to coordinate conversion in step S203. Note that the edge coordinate holding unit 109 initializes the edge coordinate group held before starting the system when the system is started.

  Next, in step S205, the edge coordinate updating unit 110 updates the edge coordinate group held in step S204 based on the vehicle speed detected by the vehicle speed sensor 112.

  Next, in step S206, the edge coordinate deleting unit 111 deletes the edge coordinate group updated in step S205 based on the vehicle speed detected by the vehicle speed sensor 112.

  Next, in step S207, the edge coordinate holding unit 109 re-holds the edge coordinate group remaining without being erased in step S206.

  In step S208, the white line recognition unit 113 recognizes a white line from the edge coordinate group held in step S207.

  Next, in step S209, the position / orientation estimation unit 114 estimates the position and orientation of the vehicle 101 within the white line recognized in step S208. If the position / orientation estimation unit 114 estimates that the vehicle 101 has deviated from the white line, the position / orientation estimation unit 114 outputs a lane departure signal to the lane departure prevention warning system 106 and proceeds to step S210. On the other hand, when it is estimated that the vehicle 101 has not deviated from the white line, the process proceeds to step S211.

  In step S210, the lane departure prevention warning system 106 performs a predetermined warning.

  In step S211, the computer 105 ends the process when the system power is turned off, and repeats the operation from step S201 if the system power is turned on.

  As described above, according to the host vehicle position estimation apparatus of the present embodiment, the edge of the white line is extracted from the image obtained by imaging the left and right white lines that divide the lane in which the vehicle 101 travels. The extracted edge is output in association with the coordinate group on the image. The output edge coordinate group is converted into world coordinates, and the converted edge coordinate group is held. The held edge is updated by moving to the opposite side of the vehicle traveling direction by the amount of vehicle movement corresponding to the vehicle speed, and the updated edge coordinate group is held again. A white line is recognized from the held edge coordinate group. And the position and attitude | position of the vehicle 101 in the recognized white line are estimated. As a result, even if the white line cannot be observed in some of the captured images, the white line can be recognized from the held edge coordinate group, so the position and orientation of the vehicle within the recognized white line can be estimated. Can do.

  Moreover, according to the display apparatus for vehicles of this embodiment, the edge coordinate deletion part 111 deletes the edge coordinate group extracted in the past. Thereby, when recognizing a white line, an unnecessary edge coordinate group can be deleted.

  Further, according to the vehicle display device of the present embodiment, the edge coordinate erasure unit 111 erases the edge coordinate group extracted in the past based on the vehicle speed. Thereby, when recognizing a white line, an unnecessary edge coordinate group can be deleted.

  Moreover, according to the display apparatus for vehicles of this embodiment, the edge coordinate deletion part 111 deletes an edge coordinate group in an order from the oldest extracted time. Thereby, when recognizing a white line, an unnecessary edge coordinate group can be deleted.

  Further, according to the vehicle display device of the present embodiment, the coordinate conversion unit 108 converts the edge coordinate group from the image coordinate system to the world coordinate system. This facilitates the association between the vehicle 101 and the edge coordinate group.

  Further, according to the vehicle display device of the present embodiment, the white line recognition unit 113 performs an optimization process when linearly complementing the edge coordinate group. Thereby, the accuracy of the recognized white line is improved.

  Further, according to the vehicle display device of the present embodiment, fisheye cameras are attached to the front and left and right of the vehicle 101. Thereby, since an edge can be extracted from each image imaged with three cameras, an edge can be extracted efficiently.

  Furthermore, a fisheye camera is attached not only to the front and left and right of the vehicle 101 but also to the rear of the vehicle. Thereby, since an edge can be extracted from each image imaged with four cameras, an edge can be extracted efficiently.

[Second Embodiment]
Next, a second embodiment of the present invention will be described. The second embodiment is different from the first embodiment in that, when extracting / updating an edge coordinate group, in addition to associating the extracted / updated time, processing time indicating the extracted / updated time is simplified. Is to set. The setting of the processing time will be described with reference to FIG. FIGS. 13B to 13D correspond to the time t = −2, −1, 0 shown in FIG. As illustrated in FIG. 13B, the edge extraction unit 107 sets the processing time of the edge coordinate group extracted at time t = −2 as T = 0. Subsequently, as illustrated in FIG. 13C, the edge extraction unit 107 sets the processing time of the edge coordinate group extracted at time t = −1 as T = 0. Further, the edge coordinate update unit 110 sets the processing time of the edge coordinate group updated at time t = −1 as T = −1. Subsequently, as shown in FIG. 13D, the edge extraction unit 107 sets the processing time of the edge coordinate group extracted at time t = 0 as T = 0. Further, the edge coordinate update unit 110 sets the processing time of the edge coordinate group updated at time t = 0 as T = −1 and T = −2, respectively. In this way, the processing time of the edge coordinate group extracted at a certain time is set as T = 0, and each time the edge coordinate group is updated, it is incremented by −1 such that T = −1 and T = −2. Set the time. By setting the processing time in this way, the edge coordinate group extracted and updated at a certain time can be simplified and shown as a relative time instead of a real time. Note that the setting method is not limited to this, and a time variable indicating how far the edge coordinate group has been detected from the current time may be updated each time.

  In the first embodiment, the threshold value of the distance for deleting the edge coordinate group is set according to the vehicle speed, but in this embodiment, the edge coordinate group is deleted based on the processing time. Specifically, the distance D [m] that the vehicle 101 moves each time the edge coordinate group is updated is D = vehicle speed × time (= current time−updated time). For this reason, if the criterion for erasing the edge coordinate group is “current time−updated time”, the distance for erasing the edge coordinate group changes according to the vehicle speed. At this time, the edge coordinates are not based on the processing time indicating the time when the edge coordinate group is updated, but the time when the edge coordinate group is updated (actual time) such as “current time−updated time”. If the group is deleted, the edge coordinate group can be deleted without using real time. That is, as shown in FIG. 13D, if the edge coordinate group related to the processing time T = −2 is set to be erased, it is not necessary to calculate the distance using the real time, and further, according to the first embodiment. Thus, the edge coordinate group can be deleted without the need to create if-then processing. The processing time threshold value is obtained in advance by simulation or experiment.

  As described above, according to the vehicle position estimation apparatus of the present embodiment, the edge coordinate group extracted before the predetermined time is deleted. Thereby, when recognizing a white line, an unnecessary edge coordinate group can be deleted.

[Third Embodiment]
The vehicle position estimation device shown as the third embodiment is different from the first embodiment described above in that it includes a steering angle sensor 115 and a yaw rate sensor 116 as shown in FIG.

  In the first embodiment, when updating the edge coordinate group, the edge coordinate updating unit 110 moves and updates the edge coordinate group only by the movement amount d of the vehicle 101 only on the opposite side of the vehicle traveling direction. Since the host vehicle position estimation apparatus according to the present embodiment includes the steering angle sensor 115 and the yaw rate sensor 116, it can detect movement in the vehicle width direction. For this reason, when the edge coordinate updating unit 110 updates the edge coordinate group, the movement amount d is not only moved from the edge coordinate group to the opposite side in the vehicle traveling direction, but is detected by the rudder angle sensor 115 and the yaw rate sensor 116. The amount of movement in the vehicle width direction is also updated by moving from the edge coordinate group. Thus, since the edge coordinate group is updated based on the movement of the vehicle 101 in the traveling direction and the vehicle width direction, the movement of the vehicle 101 can be captured more accurately than in the first embodiment. As described above, according to the vehicle position estimation apparatus of the present embodiment, the accuracy of the updated edge coordinate group is increased, and therefore the accuracy of the white line recognized using the edge coordinate group is also increased. Thereby, the estimation accuracy of the position and orientation of the vehicle 101 within the white line is improved.

DESCRIPTION OF SYMBOLS 101 Vehicle 102 Front camera 103 Right camera 104 Left camera 105 Computer 106 Lane departure prevention alarm system 107 Edge extraction part 108 Coordinate conversion part 109 Edge coordinate holding part 110 Edge coordinate update part 111 Edge coordinate deletion part 112 Vehicle speed sensor 113 White line recognition part 114 Position and orientation estimation unit 115 Rudder angle sensor 116 Yaw rate sensor

Claims (7)

Vehicle speed detecting means for detecting the vehicle speed of the vehicle;
Imaging means mounted on the vehicle and capable of imaging the left and right white lines that divide the lane in which the vehicle travels;
Edge extraction means for extracting the edge of the white line from the image captured by the imaging means;
Edge holding means for holding the edge extracted by the edge extraction means in association with coordinates on the image;
Edge updating means for updating the edge held by the edge holding means by moving to the opposite side in the vehicle traveling direction by the amount of movement of the vehicle according to the vehicle speed;
White line recognition means for linearly complementing the edge held by the edge holding means to recognize a white line;
A vehicle position estimation device comprising: position and orientation estimation means for estimating the position and orientation of the vehicle in the white line recognized by the white line recognition means.   The vehicle position estimation apparatus according to claim 1, further comprising an edge erasing unit that erases the edge held by the edge holding unit.   The edge erasing means erases an edge behind the rear of the first distance from the current vehicle position when the vehicle speed is greater than a predetermined vehicle speed, and when the vehicle speed is equal to or lower than the predetermined vehicle speed, 3. The own vehicle position estimating device according to claim 2, wherein an edge behind a second distance shorter than the first distance from the position is deleted.   The vehicle position estimation apparatus according to claim 2, wherein the edge erasing means erases the edges in order of the extracted time.   The own vehicle position estimation apparatus according to any one of claims 1 to 4, further comprising coordinate conversion means for converting the coordinates of the edge from image coordinates to world coordinates. Rudder angle detecting means for detecting the rudder angle of the vehicle;
A yaw rate detecting means for detecting the yaw rate of the vehicle;
The edge update means updates the edge held by the edge holding means by moving the vehicle according to the vehicle speed according to the vehicle speed, and detects the steering angle and yaw rate detection means detected by the steering angle detection means. 6. The own vehicle position estimating apparatus according to claim 1, wherein the vehicle position is estimated to be updated by moving in the vehicle width direction based on the yaw rate detected in step 1. Detect the vehicle speed,
Extracting an edge of the white line from an image obtained by imaging the left and right white lines that divide the lane in which the vehicle travels;
Holding the extracted edge in association with the coordinates on the image;
The held edge is moved and updated in the vehicle traveling direction opposite side by the amount of movement of the vehicle according to the vehicle speed,
Recognize white lines by linearly complementing the held edges,
A vehicle position estimation method, wherein the position and orientation of the vehicle in a recognized white line are estimated.