JPH1139464A - Image processor for vehicle - Google Patents

Abstract

PROBLEM TO BE SOLVED: To accurately estimate a road shape by processing an image of circumferential roads, detecting the position of the horizontal surface, current position, and traveling direction of a vehicle, detecting the position of a road node in the traveling direction by retrieving three-dimensional road map data, and calculating a road slope and correcting the whole line position. SOLUTION: An image process controller 3 is equipped with an image memory 31 which stores the video signal from a camera 2 and an image process part 32 which detects the white line by processing the image. A navigation controller 4 is equipped with a storage device 41 for storing 3D map data and a road slope calculation part 42 which calculates a road slope on the basis of the 3D map data. A vehicle controller 6 is provided with a synchronizing signal generation part 62 and a vehicle control part 62 which controls a steering actuator and runs a control program to generate a synchronizing signal and also corrects the white line found through the image processing with the road slope, thereby controlling the steering angle so that the vehicle travels within the white line.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus for a vehicle for recognizing an environment around a host vehicle, and more particularly, to a follow-up traveling apparatus that analyzes an image and recognizes a preceding vehicle while following the vehicle at a fixed inter-vehicle distance, Used in lane keeping devices that automatically detect and maintain the vehicle in the lane by detecting white lines, and in lane departure warning devices that detect whether the vehicle is maintaining the lane by detecting road white lines. is there.

[0002]

2. Description of the Related Art When a road image captured by a camera is processed to recognize the shape of the road, an accurate shape may not be recognized due to the gradient of the road. For example, as shown in FIG. 1, even when the vehicle is actually traveling on the same point on a road having the same shape, the road on the camera screen is more uphill than in the case where the road is flat (a). (B) is recognized as a gentle curve, and the position of the white line at the gazing point (indicated by a broken line in the figure) set at a predetermined distance is recognized as being on the right side in the case of (b) on the uphill. You. As a result, the own vehicle position relative to the white line is erroneously recognized to be on the left side in the case of (b) on an uphill. Therefore, when such an image processing device for a vehicle is used for a lane keeping device, the vehicle cannot run accurately along the lane.

The above problem is a problem that arises when trying to estimate a three-dimensional road shape from an image that is plane information. For example, in the case of the road shown in FIG. 1, it cannot be determined whether the road is a flat road or an uphill unless there is some additional information.

Various methods for estimating a road shape in consideration of the influence of such a road gradient have been proposed. The first method is
The three-dimensional shape of the road is estimated from the image of one road using the condition that the road width is constant, that is, the interval between the white lines on both sides is constant (for example, Japanese Patent Application Laid-Open No. 6-2018).
No. 9). However, according to this method, the white line interval may change before and after a curved road, a junction, or a branch point, and the estimation error due to this may not be eliminated. In some cases, white lines on both sides cannot be detected. At that time, estimation becomes impossible.

A second method is to stereoscopically view a road with parallax in the same manner as human eyes using two cameras (for example, Japanese Patent Application Nos. 4-70377 and 4-322).
174, Japanese Patent Application No. 4-76499). However, even with this method, since the parallax is reduced in inverse proportion to the distance, in order to accurately stereoscopically view a distant place, the mounting accuracy, calibration, and resolution of the camera must be increased, which is practically difficult. . In addition, two cameras must be used, and images taken by the two cameras must be processed. This requires a high-speed microcomputer that performs enormous image processing, which increases the cost of the apparatus. .

An object of the present invention is to provide a vehicular image processing apparatus for accurately estimating a road shape.

[0007]

The principle of the present invention will be described. Taking the case of approaching an uphill as an example, consider a coordinate system as shown in FIG. This coordinate system is an XYZ coordinate system in which the XY plane is a plane horizontal to the vehicle and the Z axis passes through the installation position of the camera, and is hereinafter referred to as a vehicle fixed coordinate system. Point P in the figure is a point on the white line of the actual road.
o is a projection point of the point P on the XY plane. The projection center is the camera installation point C (0, 0, Hc). Point O is the vehicle position. As for the gradient of the road, the height Z of the road
Is determined only by the distance X, and the cant (lateral inclination of the road) is not considered. That is,

## EQU1 ## Z = Z (X)

Here, all points on a straight line connecting the projection center C and the projection point Po overlap on the screen. That is,
As the road gradient changes, the position of point P is changed to projection center C and projection point P.
Even if it moves on the straight line connecting o, the movement of the point P cannot be determined on the screen, and is recognized as the point Po on the horizontal plane. Therefore, in the present invention, the change in the gradient of the road is ignored in the image processing stage, and only the point P (X, Y, Z) on the road and its projection point Po (Xo, Yo, 0) on the XY plane are ignored. I will ask for. The projection can be expressed as:

## EQU2 ## Yo = Yo (X)

Next, a point P (X, Y, Z) on the actual white line
And its projection point Po (Xo, Yo, 0).
As is clear from FIG. 2, the following relationship is established regardless of the change in road gradient.

X / Xo + Z / Hc = 1, X / Xo = Y / Yo Here, the following relationship is obtained by substituting Equation 1 representing the gradient change into Equation 3.

Xo = X / {1-Z (X) / Hc}

Y = Yo {1−Z (X) / Hc} Further, by differentiating Expressions 4 and 5, the following relationship regarding the yaw angle is obtained.

DY / dX = dY / dXo + ｛(Xo / Hc)
(DYo / dXo) -Yo / Hc {dZ (X) / dX} By further differentiating, the following relationship regarding the curvature is also obtained.

Ρ (X) = ρ ｛Xo, Yo, dYo / dXo, ρo
(Xo), dZ (X) / dX, d 2 Z (X) / dX 2}

[0010] By using the above four equations, if the gradient change is given in the form of Equation 1, the point Po on the projected white line is obtained.
From (Xo, Yo, 0) and the slope and curvature at that point, the corresponding actual point P (X, Y, Z) on the white line and the slope and curvature at that point can be obtained.

In the above description, the case where the cant component of the road gradient is not taken into consideration for easy understanding, but more generally, the height Z of the road is

## EQU8 ## The same theory can be applied to the case where Z = Z (X, Y).

Next, a point P (X, Y, Z) on the actual road
And the point (x, y) on the screen corresponding to the point P is determined. The coordinate system X'Y'Z 'shown in FIG. 3 has the origin O' set at the center of the lens of the camera, the X axis set at the optical axis of the lens, and the Y axis set parallel to the road plane. . Assuming that the position of the point P in this coordinate system is (X ′, Y ′, Z ′), the corresponding point (x, y) on the screen is given by the following equation.

X = −fc · Y ′ / X ′, y = −fc · Z ′ / X ′

Next, the point (X ', Y', Z ') is changed to the point (X,
Y, Z) is translated by Hc in the Z-axis direction and rotated by θc about the Y-axis.

X ′ = Xcosθc− (Z−Hc) sinθc, Y ′ = Y, Z ′ = Xsinθc + (Z−Hc) cosθc. Substituting this into Equation 9 gives

X = −fc · Y / {Xcosθc− (Z−Hc) sinθc}, y = −fc × {Xsinθc + (Z−Hc) cosθc} / ΔXco
sθc− (Z−Hc) sin θc｝.

In particular, when the point P (X, Y, Z) is on a flat road, Z = 0, so that equation 11

X = −fc · Y {Xcosθc + Hcsinθc}, y = −fc · {Xsinθc−Hccosθc} / {Xcosθc + Hc
sinθc｝. This is the relationship between the point (X, Y) on a flat road and the corresponding point (x, y) on the screen. Also, solving Equation 11 for (X, Y) gives

X = −Hc · (ysinθc−fccosθc) / (ycosθc + f
csinθc), Y = −Hc × x / (ycosθc + fcsinθc), and the point (x, y) on the screen and the point (X,
It is also possible to inversely calculate Y).

(1) An image pickup means mounted on a vehicle for picking up an image of a surrounding road, a white line position detecting means for processing an image to detect a position of a white line on a horizontal plane of the vehicle, and Position / direction detecting means for detecting the current position and the traveling direction; node position detecting means for searching three-dimensional road map data to detect the position of the road node in the traveling direction from the current position; There is provided a gradient calculating means for calculating the gradient, and a correcting means for correcting the white line position based on the road gradient. (2) The image processing apparatus for a vehicle according to claim 2, wherein the white line position detected by the white line position detecting means includes a horizontal position of the white line,
Includes direction and curvature. (3) The image processing apparatus for a vehicle according to claim 3, wherein the gradient calculating means calculates a road gradient at the time of imaging by the imaging means. (4) The vehicular image processing device according to claim 4 is configured to perform the white line detection by the white line position detecting means and the gradient calculation by the position / direction detecting means, the node position detecting means and the gradient calculating means in parallel. Things. (5) An image processing apparatus for a vehicle according to claim 5, wherein the current position and the traveling direction are detected by the position / direction detection means by satellite navigation and / or autonomous navigation. (6) The image processing apparatus for a vehicle according to the sixth aspect includes a vehicle control unit that controls the vehicle to travel in the lane based on the white line position corrected by the correction unit. (7) The image processing apparatus for a vehicle according to the seventh aspect includes a traveling monitoring unit that monitors whether the vehicle is traveling in the lane based on the white line position corrected by the correcting unit. (8) An image processing apparatus for a vehicle according to an eighth aspect of the present invention includes a follow-up control unit that performs a follow-up control on a preceding vehicle while maintaining a constant inter-vehicle distance based on the white line position corrected by the correction unit.

[0016]

【The invention's effect】

(1) According to the first aspect of the present invention, an image around a vehicle is processed to detect a position of a white line on a horizontal plane of the vehicle, and three-dimensional road map data is searched for a road node in a traveling direction from a current position. And calculates the road gradient based on the node position. Since the position of the white line is corrected based on the road gradient, the position of the white line can be accurately detected in a short time without performing complicated image processing.
Of the three-dimensional information related to roads, only planar information such as distance, lateral position, and yaw angle requires high accuracy when performing vehicle control. According to the first aspect of the present invention, plane information of a road is accurately detected by image processing, and a road gradient is detected based on three-dimensional road map data. ,
The amount of calculation in image processing can be reduced, and a high-speed microcomputer is not required, so that the cost of the apparatus can be reduced. (2) According to the second aspect of the present invention, the lateral position, the direction, and the curvature of the white line on the vehicle horizontal plane are detected by the image processing, so that in addition to the above-mentioned effects of the first aspect, lane keeping control, lane departure warning Thus, an accurate road shape can be provided for the preceding vehicle following control and the like. (3) According to the third aspect of the present invention, since the road gradient at the time of capturing the image is calculated, the white line detected by the image processing can be accurately corrected. (4) According to the invention of claim 4, since the white line detection and the gradient calculation are performed in parallel, the white line of the road can be detected in a short time after imaging the road around the vehicle. . (5) According to the fifth aspect of the present invention, since the current position and the traveling direction are detected by the satellite navigation and / or the autonomous navigation, the accurate current position and the traveling direction of the vehicle can be detected. (6) According to the invention of claim 6, since the vehicle is controlled to run in the lane based on the white line position corrected by the gradient change, control accuracy and reliability in lane keeping control can be improved. it can. (7) According to the seventh aspect of the present invention, whether or not the vehicle is traveling in the lane is monitored based on the white line position corrected by the gradient change, thereby improving the accuracy and reliability in the lane departure warning control. be able to. (8) According to the invention of claim 8, since the following control is performed on the preceding vehicle while keeping the inter-vehicle distance constant based on the corrected white line position, the control accuracy and reliability in the following vehicle following control can be improved. it can.

[0017]

DESCRIPTION OF THE PREFERRED EMBODIMENTS A vehicle image processing apparatus according to the present invention
An embodiment applied to a lane keeping device that detects a white line on a road and automatically maintains the own vehicle in a lane will be described.

FIG. 4 shows the configuration of one embodiment. Vehicle 1
Includes sensors such as a camera 2, an image processing controller 3, a navigation controller 4, a data communication unit 5, a vehicle control controller 6, a GPS receiver and a geomagnetic sensor (not shown), a vehicle speed sensor, a steering angle sensor, and a yaw rate sensor ( (Not shown). The camera 2 captures an image of the front of the vehicle and outputs a video signal to the image processing controller 3.

The image processing controller 3 includes an image memory 31 for storing a video signal from the camera 2 and an image processing unit 32 for processing an image to detect a white line. The image processing unit 32 includes a microcomputer and peripheral components, and executes a control program described later to perform image processing.

The navigation controller 4 has 3
A storage device 41 for storing D map data and a road gradient calculation unit 42 for calculating a road gradient based on 3D map data are provided. The road gradient calculating section 42 is composed of a microcomputer and its peripheral components, and calculates a road gradient by executing a control program described later.

The data communication unit 5 includes controllers 3 and 4
And 6 are transmitted and received. The vehicle control controller 6 includes a synchronization signal generator 61 that generates a synchronization signal.
And a vehicle control unit 62 that controls the steering actuator. The synchronization signal generation unit 61 and the vehicle control unit 62 are constituted by a microcomputer and its peripheral parts, generate a synchronization signal by executing a control program described later, and correct the white line obtained by the image processing by the road gradient to correct the vehicle. Controls the steering angle so that the vehicle travels in the lane.

According to the flowcharts shown in FIGS.
The operation of this embodiment will be described. FIG. 5 is a flowchart showing the image processing. Image processing controller 3
Performs this process in synchronization with a synchronization signal sent from the vehicle control controller 6 every 100 ms. In step 1, the camera 2 captures an image, inputs image data, and stores it in the image memory 31. Note that the image data is luminance information A for each pixel of the image sensor of the camera 2.
It is stored as [x, y]. In step 2, a point (xi, yi) whose luminance is equal to or higher than the threshold value in the input image data
Is extracted as a white line candidate point.

In step 3, the above-described inverse perspective transformation is performed on each of the extracted white line candidate points, and the extracted white line candidate points are projected on the XY plane of the vehicle fixed coordinate system, that is, on the horizontal plane of the vehicle.
The projection point (Xi, Yi) of each white line candidate point (xi, yi) is as follows when the camera specifications are height Hc, pitch angle θc, and focal length fc with respect to the road.

Xi = −Hc (yisin θc−fccos θc) / (yicos θc +
fcsin θc), Yi = −Hc · xi / (yicos θc + fcsin θc)

Next, in step 4, it is determined whether the projection point (Xi, Yi) is the white line on the left or right side of the road, and only the projection point on the left white line is extracted. The white line exists on the left and right of each lane, but if we detect both the left and right white lines,
In addition to an increase in the amount of calculation, if one of the white lines cannot be detected, the white line detection result becomes unstable. Therefore, in this embodiment, only the white line on the left side of the lane is extracted, and the vehicle travels on the right side at a constant interval based on the white line on the left side. Of course, only the right white line may be detected.

In step 5, the projected line of the white line on the left side of the road from the projected point of the left white line is converted into a quadratic curve by the least square method.

## EQU15 ## Identification is made by Y = aYX ² + bYX + cY. That is, each coefficient (aY, bY,
cY) is a simultaneous equation

(Equation 16) Is obtained as the solution of These coefficients (aY, bY, c
Y) is hereinafter referred to as white line identification parameter or road identification parameter. In the subsequent step 6, the white line identification parameters (aY, bY, cY) are
Send to The vehicle control controller 6 temporarily stores the received white line identification parameters in the buffer memory.

FIG. 6 is a flowchart showing the operation of the navigation controller 4. Note that, among the operations of the navigation controller 4, a description of a process such as a route calculation to a destination not directly related to the present invention will be omitted. In step 11, the current position in the GPS coordinate system is calculated based on the GPS signal from the GPS receiver and the azimuth signal from the geomagnetic sensor, and the current traveling road is specified by map matching to correct the current position. . In a succeeding step 12, a yaw angle indicating the traveling direction of the vehicle on the XY plane and a pitch angle with respect to the XY plane are obtained as the attitude of the vehicle in the GPS coordinate system by using the in-vehicle three-axis geomagnetic sensor. In step 13, the 3D map data in the storage device 41 is searched, and the position (Xni, Yni, Z) of the GPS coordinate system is determined for all road nodes within 100 m of the traveling direction of the vehicle from the current position.
ni) (i = 1, 2, ...) is input.

In step 14, the position of each node is converted into the vehicle fixed coordinate system shown in FIG. 2 using the calculated vehicle attitude, and is set as (Xni ', Yni', Zni '). In a succeeding step 15, a change in road height (change in road gradient) Z is obtained as a function of the distance X from each node coordinate converted into the vehicle fixed coordinate system. Although there is no restriction on the structure of the function, in this embodiment, a cubic expression,

Assuming that Z = azX ³ + bzX ² + cz, each coefficient is identified by the least squares method. Hereinafter, these coefficients (az, bz, cz) are referred to as road gradient identification parameters. The primary coefficient was set to 0 because
At the origin, the fact that the road is in contact with the XY plane is used. The specific method of the least-squares method is the same as that when the white line is identified. In step 16, the calculated road gradient identification parameters (az, bz, cz) are transmitted to the vehicle controller 6. The vehicle control controller 6 temporarily stores the received road gradient identification parameter in the buffer memory.

FIGS. 7 and 8 are flowcharts showing a vehicle control program. The vehicle control controller 6 executes this processing at every sampling time of 10 ms of the built-in timer counter. Steps 21 to 25 are processing of the synchronization signal generator 61. In step 21, the count value of the built-in timer counter is incremented. However, clear when the count value becomes 0. That is,

## EQU18 ## Cnt [n] = Cnt [n-1] +1 (when Cnt [n-1] ≠ 9), = 0 (when Cnt [n-1] = 9)

Next, at step 22, it is confirmed whether the count value is 0, that is, whether 100 ms has elapsed.
After the elapse of ms, the processing of steps 23 to 25 is performed.
That is, the processing of steps 23 to 25 is performed every 100 ms. In step 23, a synchronization signal is sent to the navigation controller 4 and the vehicle control controller 6 every 100 ms. In step 24, white line identification parameters (aY, bY, cY) are read from the buffer memory, and in step 25, road road gradient identification parameters (az, bz, cz) are read from the buffer memory. The white line identification parameters (aY, bY, cy) and the road gradient identification parameters (az, bz, cz) are values calculated from the image taken 100 ms before.

The processing of the following steps 26 to 36 is the processing of the vehicle control unit 62. In step 26, in order to estimate the yaw rate and the lateral speed, the vehicle speed V from the vehicle speed sensor, the yaw rate (dψ / dt) from the yaw rate sensor, and the control input from the steering angle sensor are input as the observation quantities used by the observer. Enter the angle δ. In the following step 27, the yaw rate (dψc / dt) and the lateral speed VY of the vehicle are estimated from the observer designed using the two-wheel model.

Equation 19] ^{_{x 1 S = A 1 x 1}} S + B 1 u 1 -K 1 C 1 (y 1 S -y 1) where, x ₁ ^S is an estimate of the state quantity x _1.

X ₁ = [γ β] ^T = [dψc / dt VY / V] ^{T Further} , y ₁ ^S is an estimated value of the state quantity y ₁ .

Y ₁ = [dψ / dt] Further, u ₁ is an input, and δ is a front wheel steering angle.

## EQU22 ## u ₁ = [δ 0] ^T A ₁ , B ₁ , and C ₁ represent the two-wheel model as a state equation,

Equation 23] is a coefficient matrix when written in _{dx 1 / dt = A 1 x} 1 + B 1 u 1, y 1 = C 1 x 1, K 1 is the observer gain.

In step 28, step 2
The white line identification parameters (aY, b) read out in steps 4 and 25
Y, cY) and road gradient identification parameters (az, bz,
It calculates how many milliseconds before cz) is the image input. The processing delay time Td is the sum of the image processing time Tv and the data update waiting time Th, as shown in FIG.

Td = Tv + Th In FIG. 9, the image processing time Tv is based on the white line identification parameters (aY, bY, cY) which are the processing results after the camera 2 captures an image.
, Which is 100 ms in this embodiment as described above. The data update waiting time T
h is the holding time from when the vehicle controller 40 receives the white line identification parameters (aY, bY, cY) to when the values are actually used. The fixation point distance Lo shown in FIG. 9 indicates a change when the vehicle speed V increases, and the fantasy distance of the image processing time Tv increases as the vehicle speed V increases.

In step 29, based on the processing delay time Td and the vehicle speed V, as shown in FIG.
A gazing point distance Lo to (X, Y, Z) is obtained.

L = Ts２５V [n−k] Here, Σ represents a sum operation from k = 0 to (Td−1), and Ts is a sampling interval of 10 ms. That is, the gazing point distance Lo is equal to the idle running distance of the vehicle 1 during the processing delay time Td. In the following step 30, the gazing point distance Lo is corrected to obtain the pseudo gazing point distance Lo 'in order to remove the influence of the gradient change of the road. Specifically, in Expression 4, Xo → Lo ′, X → Lo, Z (X) → azLo ³
By substituting + bzLo ² + cz,

Lo ′ = Lo / {1− (azLo ³ + bzLo ² + cz) / Hc}

In step 31, based on the white line identification parameters (aY, bY, cY), the horizontal position YL ', direction φL', and curvature ρL 'of the white line at the pseudo fixation point are obtained.

YL ′ = aYLo ′ ² + bYLo ′ + cY, tan ψL ′ = 2aYLo ′ + bY, ρL ′ = 2aY / ｛1+ (2aYLo ′ + bY) ² ｝ ^{3/2 In} step 32, the horizontal position YL of the white line just obtained is obtained. The correction values (YL, ψL, ρL) are obtained by removing the influence of the road gradient change on the 、, the direction ψL ', and the curvature ρL'. Specifically, it is calculated as follows using Expressions 5 to 7.

YL = YL ′ {1- (azLo ³ + bzLo ² + cz) / Hc}, tanψL = tanψL ′ + ｛(Lo ′ / Hc) tanψL′−YL ′
/ Hc｝ (3azLo ³ + 2bzLo), ρL = ρL '(Lo', YL ', tanψL', ρL ', 3azL
o ² + 2bzLo, 6azLo + 2bz)

In step 33, a lateral movement amount ΔYc and a yaw angle change amount Δψc are obtained as movement amounts of the own vehicle during the processing delay time. As a specific method, the lateral speed VY and the yaw rate (dψc / dt) are obtained by integrating each.

２９Yc = TsΣVY [nk], Δψc = TsΣ (dψc / dt) [nk] Here, Σ represents a sum operation of k = 0 to (Td−1). In step 34, the lateral position and the yaw angle of the current vehicle with respect to the white line are calculated as the difference between the moving amount of the vehicle and the position / direction of the white line at the point of interest.

Yc = ΔYc−YL, ψc = Δψc−ψL

In step 35, the steering angle is controlled so that the vehicle runs on the right side of the white line at a constant interval W. That is, as the target values, the lateral position is -W and the yaw angle is 0. Feedforward control and feedback control are performed as control rules. The feedforward term is determined so as to be proportional to the road curvature, and the feedback term is determined by PI control.

Δ ^* = Kρ (V) ρL− (KYP + KYI / s)
(Yc + W)-(KψP + KψI / s) ψc where Ki represents various gains and s represents a Laplace operator. In particular, the gain (Kρ (V)) related to the curvature in the feedforward term is a value determined by the vehicle speed. In the following step 36, the position of the steering actuator is controlled so as to obtain the target steering angle just obtained.

In the above-described embodiment, the method of correcting the gradient change with respect to the position of the white line detected by the image processing by applying to the lane keeping device has been described. However, the application of the present invention is limited to the lane keeping device. Instead, the present invention can be widely applied to devices that require accurate lane position information, such as a lane departure warning device. Further, in the above-described embodiment, an example has been described in which the height Z is expressed as a cubic expression of the distance X to represent the road gradient, but an arbitrary function Z = Z (X, Y) may be used. Further, in the above-described embodiment, an example in which the navigation system is used as an input source of the road gradient information has been described, but road-vehicle communication such as a beacon may be used as the input source.

In the configuration of the above embodiment, the camera 2 serves as an image pickup means, the image processing controller 3 and the vehicle control controller 6 serve as white line position detecting means, the navigation controller 4 serves as a position / direction detecting means, and the node position detecting means. The vehicle control controller 6 constitutes the correction means and the vehicle control means.

[Brief description of the drawings]

FIG. 1 is a diagram for explaining a change in road shape on a screen due to a change in road gradient.

FIG. 2 shows a point P on an actual road viewed from an imaging camera in a vehicle fixed coordinate system XYZ and a projection point Po on an XY plane.
FIG.

FIG. 3 is a diagram illustrating a relationship between a vehicle fixed coordinate system XYZ and a coordinate system X′Y′Z ′ of an imaging camera.

FIG. 4 is a diagram showing a configuration of an embodiment.

FIG. 5 is a flowchart illustrating image processing according to an embodiment.

FIG. 6 is a flowchart showing a navigation operation according to one embodiment.

FIG. 7 is a flowchart showing vehicle control according to one embodiment.

FIG. 8 is a flowchart following FIG. 7, showing vehicle control of one embodiment.

FIG. 9 is a diagram for explaining a processing delay time.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Vehicle 2 Camera 3 Image processing controller 4 Navigation controller 5 Data communication unit 6 Vehicle control controller 31 Image memory 32 Image processing unit 41 Storage device 42 Road gradient calculation unit 61 Synchronization signal generation unit 62 Vehicle control unit

Claims (8)

[Claims] 1. An image pickup means mounted on a vehicle for imaging surrounding roads; a white line position detection means for processing the image to detect a position of a white line on a horizontal plane of the vehicle; and detecting a current position and a traveling direction of the vehicle. A position / direction detecting means, a node position detecting means for searching three-dimensional road map data and detecting a position of a road node in the traveling direction from the current position, and calculating a road gradient based on the road node position. An image processing apparatus for a vehicle, comprising: a gradient calculating unit; and a correction unit configured to correct the white line position based on the road gradient. 2. The image processing device for a vehicle according to claim 1, wherein the white line position detected by the white line position detecting means includes:
An image processing apparatus for a vehicle, comprising a horizontal position, a direction, and a curvature of a white line. 3. The image processing apparatus for a vehicle according to claim 1, wherein said gradient calculating means calculates a road gradient at the time of image pickup by said image pickup means. . 4. The image processing apparatus for a vehicle according to claim 1, wherein a white line is detected by said white line position detector, said position / direction detector, said node position detector, and said gradient. An image processing apparatus for a vehicle, wherein the gradient calculation by the calculation means is performed in parallel. 5. The vehicular image processing apparatus according to claim 1, wherein the position / direction detection unit detects a current position and a traveling direction by satellite navigation and / or autonomous navigation. An image processing device for a vehicle, comprising: 6. The vehicle image processing apparatus according to claim 1, wherein the vehicle is controlled to run in the lane based on the white line position corrected by the correction unit. An image processing apparatus for a vehicle, comprising: means. 7. The vehicular image processing apparatus according to claim 1, wherein whether the vehicle is traveling in a lane is monitored based on the white line position corrected by the correction unit. An image processing apparatus for a vehicle, comprising: a traveling monitoring unit. 8. The vehicle image processing apparatus according to claim 1, wherein the following control is performed on the preceding vehicle while maintaining a constant inter-vehicle distance based on the white line position corrected by the correction unit. An image processing apparatus for a vehicle, comprising a tracking control unit.