JPH10320559A - Roadway detection device for vehicles - Google Patents

Abstract

(57) [Problem] For a vehicle capable of detecting a direction of a traveling road with high accuracy without asking a curved road or a straight road and without asking whether a white line is a solid line or a broken line. A travel path detection device is provided. An image input means mounted on a vehicle to capture a road image ahead of the vehicle, a preprocessing means for inputting a signal from the image input means to extract an edge point, Curve detecting means for detecting the degree of curve
106, a window setting means 104 for setting a window near a straight line or a curve expression at the time of the previous calculation, a candidate point extracting means 103 for calculating coordinates of an edge point in the window set by the window setting means 104,
A travel path calculation means 107 for calculating the shape of the travel path by referring to the coordinates of the edge points extracted by the candidate point extraction means 103 and connecting the pixels representing the edge points to form a straight line or a curve; It is a traveling path detection device for a vehicle. The window setting unit 104 switches the window setting direction to the horizontal direction or the vertical direction according to the degree of the curve detected by the curve detection unit 106.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an automatic traveling technique in an automobile or an automatic guided vehicle, and more particularly to a vehicle traveling road detecting apparatus for recognizing the shape of a road area by taking an image of a road with an on-vehicle camera and processing the image. About.

[0002]

2. Description of the Related Art As a conventional road area recognition apparatus of this type, for example, an apparatus described in Japanese Patent Application Laid-Open No. 6-225308 is known. In this device, a front image is input by a camera installed in a vehicle, a plurality of windows are set so as to be equally spaced in a road surface plan view, and these windows can be moved in a horizontal direction. The position is set to the position of the lane marking (white line) predicted based on the past data, the vehicle speed, and the steering angle, and the data obtained in each window is approximated by a quadratic curve, and the The direction has been detected.

[0003]

However, in the above-mentioned prior art, the traveling path is detected by moving the horizontally long window in the horizontal direction, so that the traveling path is a curved road and a distant white line is displayed on the image. When approaching a horizontal line, white line information may be present in the entire horizontally long window, and it may not be possible to accurately detect the position of the white line.

When the white line of the traveling road is drawn by a broken line, the influence of noise such as that the window set at the broken portion of the broken line erroneously detects that the road surface is dirty or the like is a white line. There was also a problem of receiving.

SUMMARY OF THE INVENTION The present invention has been made in view of the above-mentioned problems, and it is possible to detect the direction of a traveling road with high accuracy without questioning a curved road or a straight road, and without asking whether a white line is a solid line or a broken line. It is an object of the present invention to provide a vehicle travel path detection device that can be performed with high accuracy.

[0006]

According to a first aspect of the present invention, there is provided an apparatus for detecting a traveling path of a vehicle, comprising: an image input unit mounted on the vehicle for capturing a road image ahead of the vehicle;
A preprocessing means for inputting a signal from the image input means to extract an edge point; a curve detecting means for detecting a degree of a curve of a traveling road in front of the vehicle; Window setting means for setting a window, a candidate point extracting means for calculating coordinates of an edge point in the window set by the window setting means, and referring to coordinates of the edge points extracted by the candidate point extracting means, Travel path calculation means for calculating the shape of the travel path by connecting the pixels representing the edge points to form a straight line or a curve, the window setting means,
The setting direction of the window is switched between a horizontal direction and a vertical direction according to the degree of the curve detected by the curve detecting means.

In the vehicle traveling path detecting device according to the first aspect, when the traveling path is determined to be a curved road, the shape of the window is changed in an area at a certain distance or more from the vehicle according to the degree of the curve. The vertical line is set to be vertically long, and the horizontal edge is scanned in the vertical direction to detect the position of the white line. On the near side near the vehicle, the window is set horizontally long in the horizontal direction, and the vertical edge is scanned in the horizontal direction.

In this case, it is desirable to provide a white line type determining means for determining whether the white line of the traveling road is a solid line or a broken line. It is preferable that the window setting means be controlled so as not to set a window at a broken portion of the broken line according to the output from the white line type determining means.

In the vehicle travel path detecting device according to the second and third aspects, the type of the white line is determined, and the detection window is not set in a place where the white line does not exist. This makes it possible to detect a traveling road that is resistant to road surface dirt and noise.

[0010]

According to the vehicle traveling path detecting device of the present invention, when the traveling path is a curved road, the window is set to be horizontally long or vertically long according to the degree of the curve. The setting makes it possible to detect the traveling road with high accuracy even on a curved road.

According to the vehicle travel path detecting device of the present invention, it is possible to detect whether the type of white line is a solid line or a broken line, so that the lane position on a plurality of lanes can be determined or the adjacent lanes can be determined. It is possible to determine the presence or absence, and to provide global road information to subsequent devices.

According to the third aspect of the present invention, the window is not set in the area where the white line does not exist, so that it is affected by noise factors such as dirt on the road surface where the broken line is broken. And the traveling path can be detected with high accuracy.

[0013]

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a perspective view showing a vehicle traveling path detecting apparatus according to an embodiment of the present invention.

FIG. 1 is a block diagram showing a vehicle traveling road detecting device.

The image input means 101 captures a road image in front of the vehicle and converts the road image into an electric signal, and is a video camera or the like installed in front of the vehicle in a forward direction.

The pre-processing means 102 comprises the image input means 1
The edge point is extracted by inputting the signal from the input unit 01. Here, the edge point is a point at which the brightness changes from "dark to bright" or "bright to dark", and is considered to correspond to a lane marking or a road edge. Note that the travel road division line is a linear mark indicating the center line or the end of the travel road, such as a white line drawn on the road. The road edge refers to a median strip, a groove, a step, or the like that forms a boundary between the edge of the road and another area, and is one that can be distinguished between bright and dark on an image.

The edge detection in the preprocessing means 102 is specifically performed as follows. A (x, y) (x = 1 to 256, y = 1 to 240)
Assuming that the pre-processed image is a vertical edge image Bx (x, y) and a horizontal edge image By (x, y), the pre-processed image is obtained as follows.

## EQU1 ## Bx (x, y) = A (x-1, y-1) + 2A (x-1, y)
+ A (x-1, y + 1)-｛A (x + 1, y-1) + 2 · A (x + 1, y) + A (x +
1, y + 1)｝ By (x, y) = A (x-1, y-1) + 2A (x, y-1) + A (x +
1, y-1)-｛A (x-1, y + 1) + 2A (x, y + 1) + A (x + 1, y +
1)｝ Since such preprocessing can be executed at a high speed by a pipeline type image processing apparatus, it is preferable to perform arithmetic processing simultaneously with image input.

The curve detecting means 106 detects the degree of curve ahead of the traveling road. Specifically, for example, the inclination of a white line on an image is directly obtained by obtaining the edge direction of an edge point by image processing by a method as described in JP-A-5-151340, or the curvature information of a navigation system is obtained. The inclination of the white line on the image is obtained by perspective transformation using the above.

The white line type determining means 105 determines whether the white line is a solid line or a dashed line based on a time-series change of the white line data of the monitoring area set in advance, assuming that the white line remains in the same state to some extent. I do. The window setting unit 104 sets a window near a straight line or a curved line in the previous calculation. The candidate point extraction means 103
A candidate point is searched while scanning the preprocessed image in each set window.

The travel path calculation means 107 refers to the coordinates of the edge points extracted by the candidate point extraction means 103 and approximates the pixels representing the edge points by, for example, the least squares method so as to form a straight line or a curve. The straight line or the curve output by the travel path calculation means 107 corresponds to the travel path division line or the road edge, and is given to a subsequent automatic traveling apparatus, a display device, or the like as an indication of the shape of the travel path ahead of the vehicle. Can be

FIG. 2 is a block diagram schematically showing the entire configuration of the vehicle traveling path detecting device.

In the figure, an image pickup section 1 is provided with the image input means 1.
Equivalent to 01. Image input means 101 of the above means
Each of the means 102 to 107 except for the above is constituted by a microcomputer as the image processing unit 2. The image processing unit 2 performs image processing on the video signal from the imaging unit 1 to detect a traveling road. The detection result in the image processing unit 2 is sent to an external device such as a vehicle control device or an alarm device, and is used as information on automatic traveling control. The display unit 3 is a CRT display device or a liquid crystal display device for displaying a detection result of the image processing unit 2.

Next, the operation of the image processing section 2 will be described with reference to the flowchart of FIG.

Step S1 corresponds to the window setting means 104 of FIG. 1, and sets a scanning window based on the result of detection of the shape of the traveling path in the previous calculation. The scanning window is for detecting a white line in the window, and the setting direction in the image is switched between the horizontal direction and the vertical direction depending on the shape of the traveling path ahead.

Here, the manner of setting each scanning window will be described. Each scanning window is set at equal intervals along the traveling path, and on the image after the perspective transformation, as shown in FIG. The distance between the vehicles is large near the vehicle and becomes smaller as the vehicle gets further away. The interval of the scanning window 6 on the traveling path 5 is set to, for example, 1 / N (N is a natural number of 2 or more) or less of the solid line portion of the broken line in consideration of the case where the white line indicating the boundary of the traveling path 5 is a broken line. Set. Then, at least N windows can be set on one broken line from near the vehicle to far away. The window can be set on the broken line even if the interval between the scanning windows 6 on the traveling path 5 is set to a relatively small value of, for example, 2 m, irrespective of the length of the solid line portion of the broken line.

The center position of the scanning window 6 is represented by a quadratic curve y = ax ² + bx + c representing the shape of the traveling path detected last time.
And a straight line x = L (L is the distance from the vehicle to the scanning window 6). The shorter width W of each scanning window 6 is fixed to, for example, one pixel. Scanning window 6
Is the number of pixels corresponding to an area having a width of 3 m around the coordinates of the intersection obtained as described above, for example. Therefore, the longer width of the scanning window 6 is relatively large near the vehicle and becomes smaller as the distance increases.

In the second and subsequent scanning windows 6, n is an integer, r is the interval between the scanning windows 6 on the traveling road, and a straight line x = L + nr is used instead of the straight line x = L.
It is set by finding the intersection with the quadratic curve y = ax ² + bx + c. At this time, if the previous determination result of the form of the white line is a broken line, the setting position of the scanning window 6 is restricted. That is, when the dashed line was detected last time, it is assumed that the form of the white line has a certain degree of continuity and is assumed to be a dashed line even at the present time, and as shown in FIG. Do not set 6. Further, as shown in FIG. 6, the setting direction of the scanning window 6 is switched according to the degree of the curve based on the shape of the traveling road calculated last time. Note that the initial conditions at the time of the first processing are such that the scanning line 6 is set on the assumption that the white line is a solid line and the traveling road is a straight road.

Step S2 is a candidate point extracting means 10 shown in FIG.
The search for white line candidate points is performed for each scanning window 6 set in S1. Several methods for extracting white line candidate points in each scanning window 6 are conceivable. For example, there is a method of extracting a white or yellow candidate area by binarization or a one-dimensional filter and obtaining the white or yellow candidate area as a center point of the area. As another method, using the edge images Bx and By created in the preprocessing, from the area of the edge image Bx when the scanning window 6 is horizontally long, and from the area of the edge image By when the scanning window 6 is vertically long. The white line candidate point can also be obtained by searching the maximum value or the minimum value, respectively.

Step S3 is the white line type determining means 1 in FIG.
It is determined whether the white line of the traveling road is a solid line or a broken line. The area closest to the vehicle in the scanning window 6 set in step S1 is set as the white line type determination area, and the type of the white line is determined at this position. If there are always white line candidate points in the white line type determination area, it is determined that the white line on the traveling path is a solid line, and all the scanning windows are set. On the other hand, when the timing at which the white line candidate point exists and the timing at which the white line candidate point does not exist are repeated at a constant period, the white line is determined to be a broken line, and the pattern with the white line candidate point and the pattern without the white line candidate point are analyzed. In the next processing, the place where the broken solid line portion exists is estimated.

The setting position of the white line type determination area does not necessarily have to be the scanning window closest to the vehicle, and may be newly set. It is preferable that the type of the white line is determined separately for each of the left and right white lines, so that it is possible to cope with the case where the types of the left and right white lines are different.

Step S4 corresponds to the curve detecting means 10 shown in FIG.
6, which detects the degree of curve ahead of the traveling road. In the present embodiment, a method of detecting the degree of a curve by directly calculating the inclination of a white line on an image using image processing will be described.

Assuming that the coordinates of the white line candidate points detected in step S2 on the original image A are A (x, y) from among the scanning windows 6 set in step S1, a value g indicating the edge direction (X, y) is a vertical edge image Bx (x, y).
y) and the horizontal edge image By (x, y) are shown as follows.

G (x, y) = sgn {Bx (x, y)} tan ^-1 {By (x, y) / Bx (x, y)} where sgn {Bx (x, y)} = 1 (when Bx (x, y) ≧ 0) −1 (when Bx (x, y) <0) The value of g (x, y) is 0
゜ 360 °, but are actually converted into appropriate numerical values and stored. In the next process of step S1, referring to this value, the scanning window 6 is set to be vertically long when the white line is inclined near the horizontal such as far away on a curved road,
When the white line is nearly vertical as in a straight road, the scanning window 6 is set to be horizontally long.

Note that the method of detecting the degree of the curve ahead of the traveling road is not necessarily limited to the method of detecting the inclination of the white line by image processing. For example, the method of detecting the degree of the curve based on the curvature information of the navigation system is used. May be detected, or the curve may be estimated by detecting a roadside delineator with a laser radar.

Step S5 corresponds to the travel path calculating means 10 in FIG.
7, which calculates the shape of the traveling road based on the white line candidate points detected in step S2. Assuming that the road surface is flat and that the vertical movement, rolling, and pitching of the vehicle are minute, the white line candidate points in each scanning window 6 are inversely perspective-transformed onto the road surface plane, and then, for example, a quadratic least square method Is applied to perform curve approximation. Thereby, the direction of the white line 4 on the traveling path is obtained, and the shape of the traveling path 5 is obtained.

The embodiments described above are described for facilitating the understanding of the present invention, and are not described for limiting the present invention. Therefore, each element disclosed in the above embodiment is intended to include all design changes and equivalents belonging to the technical scope of the present invention.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a vehicle traveling path detection device according to an embodiment of the present invention.

FIG. 2 is a block diagram showing an overall configuration of a vehicle traveling path detection device according to an embodiment of the present invention.

FIG. 3 is a flowchart illustrating an operation of the vehicle traveling path detection device according to the embodiment;

FIG. 4 is a setting diagram of a scanning window when a white line is a solid line.

FIG. 5 is a setting diagram of a scanning window when a white line is a broken line.

FIG. 6 is a diagram illustrating a setting of a scanning window in a case where the front of a traveling road is a curved road.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 ... Imaging part 2 ... Image processing part 3 ... Display part 4 ... White line 5 ... Road 6 ... Scanning window

Claims (3)

[Claims] 1. An image input means mounted on a vehicle for capturing a road image ahead of the vehicle, a preprocessing means for inputting a signal from the image input means to extract an edge point, and a curve of a traveling road in front of the vehicle. Curve detection means for detecting the degree of the window, window setting means for setting a window near a straight line or a curve expression in the previous calculation, and a candidate for calculating the coordinates of an edge point in the window set by the window setting means Point extraction means, and a travel path calculation means for calculating the shape of the travel path by referring to the coordinates of the edge points extracted by the candidate point extraction means and connecting the pixels representing the edge points to form a straight line or a curve. Wherein the window setting means sets a window setting direction to a horizontal direction or a vertical direction in accordance with a degree of a curve detected by the curve detecting means. Vehicular travel path detection device, characterized in that switching to. 2. The vehicle travel path detecting device according to claim 1, further comprising white line type determination means for determining whether the white line of the travel path is a solid line or a broken line. 3. The vehicle according to claim 2, wherein the window setting means is controlled in accordance with an output from the white line type judging means so as not to set a window at a broken portion of the broken line. Roadway detection device.