JPH08329224A - Road white line detector - Google Patents

Abstract

PURPOSE: To obtain the shape of a road with a high precision even in the case of the road whose change of curvature is large. CONSTITUTION: A preprocessor 12 extracts a white line on the road from the picture signal picked up with a camera 11. A tangent calculation device 13 for road model calculates a tangent in the window set start position to the road model based on the preceding picked-up picture, and the distance between the tangent and the road model is successively calculated in the vanishing point direction of the road model with a difference calculation device 14. A line detection area determining device 15 determines the range, which ends in the position where the distance between the road model and the tangent is a prescribed value, as the window of white line detection in the present picture and determines this position as the adjacent window set start position. A straight line detector 16 detects white lines as the straight lines in these windows and a road shape estimating device 17 estimates the road shape from this group of straight lines to determine a new road model. The distance to the tangent is limited within a prescribed value to determine the window which has a size in accordance with the change of curvature.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a road white line detecting device for detecting a white line indicating a lane area of a vehicle.

[0002]

2. Description of the Related Art Conventionally, in a situation where a vehicle is traveling on a road such as a highway in which the lane area is indicated by a white line, an image pickup means is used to detect the traveling direction of the vehicle. There is known a road white line detection device that picks up an image of a traveling lane area in front of a vehicle and detects the white line as a feature indicating the traveling course of the host vehicle in the image. The detection result is obtained as a road model (white line model), that is, an approximate expression expressing the three-dimensional shape of the white line.

The white line model uses road parameters representing the three-dimensional shape of the road, the vehicle position, and the vehicle attitude, respectively.
The white line is, for example, a quadratic equation in the horizontal plane (X-Z), a vertical plane (Z
-Y) is approximated by a linear expression. FIG. 12 shows a white line model in the camera coordinate system, and road parameters are determined as shown in the figure, and an approximate expression can be expressed as in Expression (1). In FIG. 12, (a) shows a horizontal plane and (b) shows a vertical plane. X = BZ ² + CZ + A−iE Y = DZ−H 0 (1) In FIG. 12A, the white lines are numbered 0, 1, 2, ... The white lines 0 to i are described by the common formula (1). Camera coordinate system (X, Y, Z)
The origin of is moved forward as the vehicle advances, and the parameters A to E in the equation (1) are changed.

Parameter A is the distance (hereinafter referred to as deviation) between the white line located on the left side of the vehicle and the center of the vehicle (attachment position of the image pickup device), parameter B is the road curvature in front of the vehicle, and parameter C is the white line at Z = 0. The yaw angle of the vehicle with respect to the tangential direction, the parameter D corresponds to the pitch angle of the vehicle with respect to the road plane (the relative angle between the road and the Z axis), and the parameter E corresponds to the distance between the white lines (lane width for a straight road or Z = 0). Perspective transformation from three-dimensional to two-dimensional is performed on this white line model to create a white line model on the road surface image described by a multi-dimensional expression in the coordinate system (x, y) of the road surface image. On the contrary, by detecting the white line on the road surface image in the xy coordinate system and determining the parameters of the white line model formula, each road parameter such as the road curvature, the yaw angle, and the pitch angle can be obtained.

To obtain the white line model on the road surface image, a plurality of straight line detection small areas (windows) are set on the screen, the white line is extracted as a straight line in each of them, and the white line model is extracted from the plurality of straight lines. Assemble.
The setting of the straight line detection small area (window) on the screen in the conventional road white line detection apparatus is disclosed in, for example, Japanese Patent Laid-Open No. 4-291405.

Here, as shown in FIG.
With respect to the vehicle traveling direction, the ranges S1 to S5 are set according to the vehicle speed and the distance from the host vehicle to the actual road surface, and in the vehicle left-right direction, the range S1 closest to the host vehicle is fixed,
Regarding the ranges S2 to S5, the position of the window and the size of the vehicle in the left-right direction are set based on the extended line of the straight line detected in the adjacent window on the side closer to the own vehicle. That is, for example, the setting of the window 202 adjacent to the window 201 set in the range S1 is performed by setting the end point coordinates P2 (x (P2), y (of the straight line detected in the window 201 as shown in FIG. 13B. P
2)), the inclination θ of the detection straight line, and the constant value ω for positioning in the x direction, and S2 corresponding to the size and position in the y direction.
Is determined by the vehicle speed.

[0007]

However, in such a conventional road white line detecting device, as shown in FIG. 13 (c), the change in the curvature of the road shape on the screen becomes large. In this case, the size of the window in the y direction does not take into consideration the road shape, and thus the window may not be set at an appropriate position. That is, in the window 212 corresponding to the portion where the curvature changes greatly, the difference between the straight line extension 213 detected in the window 211 adjacent to the side close to the vehicle and the actual road shape 214 is too wide. However, there is a possibility that it may not be possible to set an appropriate small area for straight line detection. The present invention has been made in view of such a conventional problem, and sets a straight line detection area suitable for an actual road shape even when the change in the curvature of the road shape on the screen is large. It is an object of the present invention to provide a road white line detection device capable of detecting a driving lane region with high accuracy.

[0008]

Therefore, according to the present invention, as shown in FIG. 1, an image pickup means 1 for sequentially picking up a road condition in front of a vehicle, and a road white line from a road surface image picked up by the image pickup means 1. In the road white line detection apparatus that includes the image processing unit 2 that extracts the road shape as a road model for each of the road surface images, a plurality of small areas for detecting the white lines in the road surface image are set to the curvature of the road model. A straight line detection area setting means 3 to be set accordingly, a straight line detection means 4 to detect the white line existing in the small area set by the straight line detection area setting means 3 as a straight line, and the straight line detection means 4 The road shape estimating means 5 for estimating the road shape and determining a new road model based on the straight line detected for each of the small areas.

The straight line detection area setting means 3 calculates the tangent line of the road model to the previous road model at the straight line detection area setting start position, and the distance between the previous road model and the tangent line to the straight line. A road model-tangent line difference calculating means for sequentially calculating from the detection area setting start position toward the vanishing point of the road model, and a position where the distance between the road model and the tangent line of the previous time from the straight line detection area setting start position becomes a predetermined value. Can be configured as a straight line detection region determining unit that sets the range up to as a straight line detection region.

The straight line detection area setting means 3 also calculates the plane curvature change of the road model, which sequentially calculates the curvature of the previous road model from the straight line detection area setting start position toward the vanishing point of the road model. Curvature change curve integrating means for integrating the curvature of the previous road model from the straight line detection area setting start position until its integral value reaches a predetermined value, and the integral value from the straight line detection area setting start position to the predetermined value. It can also be configured with a straight line detection area determining unit that sets a range up to a position where the value becomes a straight line detection area.

Furthermore, the straight line detection area setting means 3 calculates the curvature of the previous road model from the start position of the straight line detection area setting toward the vanishing point of the road model, and a plane curvature change calculation means of the road model; The curvature change curve integrating means for integrating the curvature of the previous road model from the straight line detection area setting start position to the vicinity of the vanishing point and the integral value integrated from the straight line detection area setting start position to the vicinity of the vanishing point are set in advance. It may have a straight line detection region determining means for determining a straight line detection region by obtaining a coordinate value that is divided by the set number of straight line detection regions and divided into even areas.

[0012]

The image pickup means 1 sequentially picks up the road situation in front of the vehicle, and the image processing means 2 extracts the white line of the road from the picked-up image signal. The straight line detection area setting means 3 sets a plurality of small areas in the road surface image, and the straight line detection means 4 detects a white line existing in the small area as a straight line. Then, the road shape estimation means 5 estimates the road shape based on the straight lines detected for each of the plurality of small areas, and a road model in which the road shape is expressed by a multi-dimensional expression is obtained. The straight line detection area setting means sets the position and size of the small straight line detection area according to the change in the curvature of the road model. Therefore, a small straight line detection small area is set in a range where the curvature of the road shape is large on the screen, and a large straight line detection small area is set in a range where the curvature is small. Therefore, the accuracy of the road shape estimation is improved. .

[0013]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described below with reference to the drawings. FIG. 2 is a block diagram showing the configuration of the first exemplary embodiment of the present invention. A camera 11 that captures an image of road conditions is installed toward the front of the vehicle. The image signal picked up by the camera 11 is sent to a pre-processing device 12 as an image processing means for extracting a white line from the image as a feature indicating the course in which the vehicle is going to travel. A road model tangent calculation device 13, a road model-tangential difference calculation device 14, and a straight line detection area determination device 15 are connected in series to the pre-processing device 12.

The tangent calculation device 13 for the road model, based on the road model (white line model) on the previous imaging screen estimated by the road shape estimation device 17, which will be described later, detects a straight line detection small area (hereinafter referred to as a window) on the currently processed screen. The tangent of the road model at the setting start position is calculated. The road model / tangent line difference calculation device 14 sequentially calculates the distance between the tangent line on the processing screen and the road model from the window setting start position toward the vanishing point. Then, the straight line detection area determination device 15 determines a range from the window setting start position to a position where the difference (distance) between the road model and the tangent becomes a predetermined value or more as a new window,
The position that becomes equal to or greater than the predetermined value is set as the next window setting start position.

The straight line detection area determining device 15 is connected with a straight line detecting device 16 for detecting the road white line in the set window as a straight line. Further, the road shape estimating device 17 detects the straight line detecting device 16 by the straight line detecting device 16. Using the start point and end point coordinates of the straight line, the road shape of the route in which the vehicle is going to travel is estimated, a new road model is determined, and the road model is updated. This road model estimated by the road shape estimation device 17 and the straight line detection area determination device 15
The new window set in step 3 is input to the tangent line calculation device 13 of the previous road model.

FIG. 3 is a flow chart showing the overall flow of the operation of the above apparatus. First, step 31
Then, in starting the control operation, after resetting the memory contents and other initial settings, step 32
At, the camera 11 inputs an image signal obtained by imaging the road condition in front of the host vehicle. In step 33, the pre-processing device 12 extracts a white line indicating the route on which the vehicle is going to travel from the input image signal as pre-processing for the window setting operation. Next, in step 34, the road model tangent calculation device 13 resets the window setting count to i = 0, and then in step 35, the window setting for each extracted white line is performed. In step 36, the window setting frequency is set to i = i + 1.

Then, in step 37, it is checked whether or not the number of times the window has been set reaches a predetermined value, for example, 5, and when it reaches the predetermined value, the process proceeds to step 38, and when it does not reach the predetermined value, the step proceeds. Return to 35.
In step 38, the straight line detection device 16 detects the white line existing in each window set in the previous step 35.
Detect as a straight line. In step 39, the road shape estimation device 17 estimates the road shape of the route on which the vehicle is going to travel from the coordinates of the start point and the end point of the straight line detected in the previous step, and updates this as a new road model formula. . Then, in step 40, other processing such as road parameter output is executed, and the new image signal in step 32 is returned to the input. In this way, the detection of the road shape by the new images is sequentially repeated. Step 32 above
Constitutes image pickup means of the invention, step 33 constitutes image processing means, step 38 constitutes straight line detection means, and step 39 constitutes road shape estimation means.

FIG. 4 shows details of the straight line detection small area (window) setting for each white line in step 35.
Here, first, in step 51, the tangent line calculation device 13 of the road model calculates the tangent line 21 to the road model R at the window setting start coordinates P (x0, y0) as shown in FIG. In step 52, the road model R and the tangent 21 are calculated by the difference calculation device 14 between the road model and the tangent.
The position of the y coordinate for calculating the distance d is set to 1 in the vanishing point P direction.
The pixel is moved by a pixel amount, and in the next step 53, the distance d between the road model R and the tangent 21 at this new y coordinate position is calculated. In step 54, it is checked whether or not the distance d is equal to or greater than a predetermined value, and if the distance d is less than the predetermined value, the process returns to step 52, and the vanishing point Pe is sequentially deleted from the window setting start position P (x0, y0). The calculation of d is repeated in the direction. Here, d is set to an arbitrary value that is equal to or less than the magnitude of deviation that is experimentally allowed.

When the distance d becomes equal to or larger than a predetermined value, the process proceeds to step 55, where the position P (x
Position P (x where the distance d is greater than or equal to a predetermined value from 0, y0)
1, y1) are set as the height range of the window 25. In step 56, it is determined how far the upper surface of the window 25 set in step 55 is from the vanishing point Pe, and if it is larger than a predetermined value, the process proceeds to step 57 to set the next window 26. If it is within the predetermined value, the window setting control operation is terminated, and the process proceeds to step 36 in the flowchart of FIG. The above steps 54 to 56 are executed in the straight line detection area determination device 15.

In step 57, in the road model tangent calculation device 13, the position P (x1, y1) of the upper surface of the window 25 set in step 55 is set as the next window setting start position and the tangent 22 to the road model R is set from that position. As a tangent line for judging the distance d from the road model R when setting the next window 26
Return to.

After step 52, which is returned, similarly, in step 52, the position of the y coordinate for calculating the distance d between the road model R and the tangent 22 is moved by one pixel in the vanishing point Pe direction, and in step 53. , The distance d between the road model R and the tangent 22 of the road model R at the window 26 setting start position P (x1, y1) is set at the setting start position P (x1, y1).
Is sequentially calculated in the direction of the vanishing point P, and when it becomes equal to or larger than the predetermined value in step 54, the range is set as the window 26 in step 55. After that, the same control operation is repeated until the vanishing point Pe is set, the windows 27 to 29 are set, and when it is determined that the distance from the vanishing point Pe is within a predetermined value in step 56 and the vanishing point Pe is near, the above-mentioned operation is performed. Thus, the control operation of the window setting is ended, and the process proceeds to step 36 in the flowchart of FIG. Step 51 to above
57 and steps 36 and 37 constitute the straight line detection area setting means of the invention.

The present embodiment is configured as described above, draws a tangent line to the road model at each window setting start position, calculates the distance between the tangent line and the road model, and the distance becomes a predetermined value or more. The position is calculated, and the range from the window setting start position to the position where the window value is equal to or greater than the predetermined value is set as a window. Therefore, a small window is set in a range where the curvature of the road shape on the screen is large, and a range is set in a range where the curvature is small A large window is set up. in this way,
According to the present embodiment, since the position and size of the window are set according to the curvature of the road model, the effect of improving the accuracy of road shape estimation can be obtained.

FIG. 6 is a block diagram showing the configuration of the second embodiment of the present invention. A camera 11 that captures a road condition in front of the host vehicle and a preprocessing device 12 that extracts a road white line from an image signal captured by the camera 11 are provided, and the output of the preprocessing device 12 is a plane curvature change calculation device 63 for a road model. Sent to. Road model plane curvature change calculation device 6
3 sequentially calculates the curvature of the road model from the window setting start position to the vanishing point direction based on the road model on the previous imaging screen estimated by the road shape estimation device 17, and obtains a curvature change curve indicating a change in plane curvature. Ask.

The above calculation result is the curvature change curve integrating device 6
4, the plane curvature along the curvature change curve is integrated from the window setting start position toward the vanishing point. Then, in the straight line detection area determination device 65 connected to the curvature change curve integration device 64, when the integrated value becomes a predetermined value or more, a range of the position from the window setting start position to the predetermined value or more And set the window. Further, the position that is equal to or more than the predetermined value is set as the next window setting start position.

The straight line detection area determining device 65 is connected to a straight line detecting device 16 for detecting a road white line in the set window as a straight line. Further, the road shape estimating device 17 detects the straight line detecting device 16 by the straight line detecting device 16. Using the start point and end point coordinates of the straight line, the road shape of the route in which the vehicle is going to travel is estimated, a new road model is determined, and the road model is updated. This road model estimated by the road shape estimation device 17 is input to the plane curvature change calculation device 63 of the previous road model.

Next, the control operation of the window setting according to this embodiment will be described. The overall flow is the same as the flow in the first embodiment shown in the flowchart of FIG. In this embodiment, the straight line detection area setting in step 35 in the flowchart of FIG. 3 is performed according to the flowchart shown in FIG. First, at step 71, the curvature of the road model on the screen is calculated in the plane curvature change calculating device 63. Then, for a road model R on the screen as shown in FIG. 8A, for example, the vanishing point Pe direction (y
By repeating the calculation in the (direction), a curvature change curve K as shown in FIG. 8B is obtained corresponding to the change in the curvature.

In the next step 72, the curvature change curve integrator 64 executes the integration of the change in curvature in the y direction from the window setting start position. And step 7
In step 3, the straight line detection area determining device 65 checks whether or not the integrated value calculated in the previous step has become a predetermined value S1 or more as shown in (c) of FIG. While the integrated value is less than the predetermined value S1, the position of the y coordinate for calculating the integral of the curvature change is moved by one pixel in the vanishing point Pe direction in step 74, and the process returns to step 72. The calculation of d is sequentially repeated from the window setting start position P (x0, y0) in the vanishing point Pe direction.

When the integrated value reaches the predetermined value S1, the routine proceeds to step 75, where the range from the window setting start position y0 to the position y1 at which the integrated value reaches the predetermined value S1 is set as the window 83 closest to the host vehicle. In addition, the predetermined value S1
The position y1 that has become is the next window setting start position. In step 76, it is checked whether or not the number of set straight line detection small areas (windows) has reached a predetermined value, for example, 5 in the case of FIG. Repeat the operation of. That is, again in step 72, the change in curvature in the y direction is integrated from the next window setting start position y1, and step 73,
This integration is repeated until the integrated value reaches a predetermined value S2 (here, S2 = S1) at 74, and then the next window 84 is set at step 75.

Thus, the five windows 83, 84,
..., 87 is set. During this period, the predetermined value for the integrated value is S1 = S2 = S3 = S4 = S5. When the integrated values S1 to S5 of the windows 83 to 87 are equal, as shown in (c) of FIG. 8, the value in the y direction of the window becomes smaller as the curvature changes on the screen. When the set number of windows reaches a predetermined value in the check in step 76, the window setting control operation is terminated and the process proceeds to step 36 in the flowchart of FIG. In this embodiment, in particular steps 71-76,
Steps 36 and 37 constitute the straight line detection area setting means of the invention.

The present embodiment is configured as described above. The plane curvature change calculating device 63 of the road model calculates the curvature on the screen of the road model, and the curvature change curve integrating device 64 calculates the integrated value of the curvature in the y direction. Since the range up to the specified value is set as the window, the size of the window in the y direction, which is directly related to the accuracy of the road model, becomes small in the range where the curvature of the road shape changes greatly on the screen. Since the window is very small, the road shape can be estimated with high accuracy even on a road whose curvature changes.

FIG. 9 shows a third embodiment of the present invention. This has a configuration similar to that shown in FIG. 6 of the previous embodiment, and has a curvature change curve integrating device 94 and a straight line detection area determining device 95 which are sequentially connected to the plane curvature change calculating device 63 of the road model. The processing modes of the curvature change curve integration device 94 and the straight line detection area determination device 95 are different.

FIG. 10 is a flowchart of window setting in this embodiment. Similar to the previous embodiment, when the overall control operation of road shape estimation reaches step 35 of the flowchart shown in FIG. 3, window setting is performed according to the flow shown in FIG. First, in step 101, the curvature of the road model in the y direction on the screen is calculated from the window setting start position by the road model plane curvature change calculating device 63. In step 102, as shown in FIG. 11, in the curvature change curve integrating device 94, the straight line detection small area (window) setting range, that is, the window setting start position y0 closest to the host vehicle to the vanishing point Pe vicinity position y.
A value obtained by integrating the range L up to 5 is calculated.

Then, in step 103, the y-coordinate for dividing the integration region calculated in the previous step 72 into equal areas is calculated in the straight line detection region determination device. That is,
In the example shown in FIG. 11, the number of windows is 5, and the y-coordinate y1 is the same area obtained by equally dividing the total area into 1/5.
~ Y4 is calculated. Then, in step 104,
These areas are equally divided into y coordinates y1 to y4, and y0 to y5
The respective areas obtained by dividing the ranges up to are set as straight line detection small areas (windows) 111 to 115. As a result, the window setting control operation is terminated, and thereafter, the process proceeds to step 36 in the flowchart of FIG. In this embodiment, in particular, steps 101 to 104, steps 36 and 3
Reference numeral 7 constitutes the linear detection area setting means of the invention.

Also in this embodiment, as shown in FIG. 11, the value in the y direction of the window becomes smaller in the range in which the change in curvature on the screen is larger. In this way, since the window is set according to the change in the curvature of the road model, it is possible to obtain the same effect as in the previous embodiment, and particularly when the number of windows is predetermined, the excellent effect is exhibited. To do.

In each of the above-described embodiments, an example in which a white line is extracted as a feature indicating a route has been described. However, the present invention is not limited to this. The shape can be determined. Therefore, in the present invention, the white line includes other lane markers, road reflectors, and other index installations arranged along the road.

[0036]

As described above, according to the present invention, the road condition in front of the vehicle is sequentially picked up by the image pickup means, and in the road surface image, a plurality of small areas set by the straight line detection area setting means according to the curvature of the road model. Since the white line of the road is detected as a straight line in the inside, and the road shape is estimated and a new road model is determined based on the straight line detected for each small area,
A small area for detecting a white line is set to be small in a range where the curvature of the road shape changes largely on the screen, and a small area is set to be large in a range where the curvature change is small, so that the road shape can be estimated with high accuracy.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of the present invention.

FIG. 2 is a block diagram illustrating a configuration of a first exemplary embodiment of the present invention.

FIG. 3 is a flowchart showing the overall flow of road shape estimation according to the first embodiment.

FIG. 4 is a flowchart showing a window setting control operation in the first embodiment.

FIG. 5 is an explanatory diagram showing a window setting procedure.

FIG. 6 is a block diagram showing a configuration of a second exemplary embodiment.

FIG. 7 is a flowchart showing a window setting control operation of the second embodiment.

FIG. 8 is an explanatory diagram of window setting for partitioning into predetermined values of integral values of curvature change.

FIG. 9 is a block diagram showing a configuration of a third exemplary embodiment.

FIG. 10 is a flowchart showing a window setting control operation of the third embodiment.

FIG. 11 is an explanatory diagram of a window setting that equally divides a total integral value of curvature changes.

FIG. 12 is a diagram showing a white line model in a camera coordinate system.

FIG. 13 is a diagram showing a window setting method in a conventional example.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 Image pickup means 2 Image processing means 3 Straight line detection area setting means 4 Straight line detection means 5 Road shape estimation means 11 Camera 12 Pre-processing device 13 Road model tangent line calculation device 14 Road model and tangent line difference calculation device 15 Straight line detection region determination device 16 straight line detection device 17 road shape estimation device 21, 22 tangent lines 25-29, 83-87, 111-115 straight line detection small area (window) 63 plane curvature change calculation device 64, 94 curvature change curve integration device 65, 95 Straight Line Detection Area Determining Device 81 Curve Showing Road Model on Screen 82 Curve Showing Curvature Change on Screen Pe Vanishing Point R Road Model K Curvature Change Curve

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification code Office reference number FI Technical display location G08G 1/16 G05D 1/02 K // G05D 1/02 9061-5H G06F 15/70 330Z

Claims (4)

[Claims] 1. An image pickup means for sequentially picking up a road condition in front of a vehicle, and an image processing means for extracting a white line of a road from a road surface image picked up by the image pickup means, wherein a road shape is used as a road model for each road surface image. In the road white line detection device to be obtained, in the road surface image, a plurality of small areas for detecting the white line are set according to the curvature of the road model, and a straight line detection area setting means is set by the straight line detection area setting means. Based on a straight line detecting means for detecting the white line existing in a small area as a straight line, and a straight line detected for each of the plurality of small areas by the straight line detecting means,
A road white line detection device, comprising: a road shape estimating means for estimating a road shape and determining a new road model. 2. The straight line detection area setting means calculates a tangent line of a road model for calculating a tangent line to a previous road model at a straight line detection area setting start position, and a distance between the previous road model and the tangent line. A road model-tangent line difference calculating means for sequentially calculating from the straight line detection area setting start position toward the vanishing point of the road model, and a distance between the previous road model and tangent line from the straight line detection area setting start position becomes a predetermined value. The road white line detection device according to claim 1, further comprising a straight line detection area determining unit that sets a range up to a position as a straight line detection area. 3. The straight line detection area setting means sequentially calculates the curvature of the previous road model from the straight line detection area setting start position toward the vanishing point of the road model, and the plane curvature change calculation means of the road model, The curvature change curve integrating means for integrating the curvature of the previous road model from the straight line detection area setting start position until the integral value reaches a predetermined value, and the integral value from the straight line detection area setting start position 2. The road white line detection device according to claim 1, further comprising a straight line detection area determining unit that sets a range up to a position having a predetermined value as a straight line detection area. 4. The straight line detection area setting means calculates the curvature of the previous road model from a straight line detection area setting start position toward a vanishing point of the road model, and a plane curvature change calculation means of the road model. A curvature change curve integrating means that integrates the curvature of the previous road model from the straight line detection area setting start position to near the vanishing point, and a straight line preset with an integrated value integrated from the straight line detection area setting start position to near the vanishing point The road white line detection apparatus according to claim 1, further comprising: a straight line detection area determining unit that determines a straight line detection area by obtaining a coordinate value that is divided by a set number of detection areas and divided into equal areas.