JPH06348991A - Traveling environment recognizer for traveling vehicle - Google Patents

Abstract

PURPOSE:To provide a traveling environment recognizer which can accurately recognize the traveling lane of a vehicle, the preceding vehicles on the lane, the traffic control signs, etc. CONSTITUTION:A data converting part 2 projects the picture elements of a color image supplied from a color input part 1 of a color image pickup device which is mounted on a traveling vehicle into a three-dimensional space different from that which defines the R(red), G(green) and B(blue) values as three axes based on these three values. Then the three-dimensional space where each picture element is projected is divided into plural areas by an dividing part 3. Then the part 3 defines the center value of each divided area as the picture element value that belongs to the corresponding divided area and divides the color image into plural area based on each picture element value valculated by the part 2. An image recognizing part 4 recognizes a traveling lane and the preceding vehicles and the traffic control signs existing on the lane for each divided area in response to the actual views. The objects recognized by the part 4 are shown at a display part 5.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention is mounted on a moving vehicle or the like, and recognizes a driving lane such as a road, a white line, or a preceding vehicle on the basis of a color image obtained by capturing a scene in front of the traveling area of the moving vehicle Driving environment recognition device.

[0002]

2. Description of the Related Art A means for assisting a safe driving operation of a vehicle driver by developing an image of a scene in front of a traveling vehicle, recognizing the traveling lane of the vehicle based on the captured image, has been developed. JP-A-4-134503, JP-A-62-221800). In Japanese Patent Laid-Open No. 4-134503, a vehicle is mounted.
The image in front of the car captured by a video camera such as a CCD camera is subjected to filtering and color extraction to extract the white line area, which is binarized and then stored in the frame memory. The technology for detecting the traveling road area and traveling road direction is disclosed. In addition, JP-A-62-221800
In the gazette, a constant conditional expression is used from the same captured image, pixels satisfying this are extracted, and white lines extracted respectively,
A technique for identifying the driving lane of an automobile according to the yellow line area is disclosed.

[0003]

In the conventional device of this type, in order to extract the driving lane of the automobile, road indications such as white lines or yellow lines are extracted, and the area sandwiched by the extracted road indications is extracted. It is specified as the driving lane. However, on an actual road, the road display may not be recognized due to damage to the road surface or the like, and in this case, there is a problem that the traveling lane cannot be specified.

On the other hand, in the technique of making a judgment using a constant conditional expression relating to the density value in order to recognize the driving lane, it is not possible to make a stable judgment when, for example, the sunshine conditions change, and white and yellow When a vehicle is shown in the image, it may be mistakenly recognized as a road display.In the conventional technology, it is virtually impossible to individually recognize the vehicles around the vehicle and road signs in the captured image. There was also the problem of being impossible.

[0005]

A traveling environment recognition apparatus for a moving vehicle according to the present invention includes a color image pickup apparatus for picking up the front of the moving vehicle to obtain a color image, and a color image obtained by the color image pickup apparatus. It is characterized by comprising an area dividing section for dividing a pixel into a plurality of areas based on color similarity, and an image recognition section for recognizing a traveling environment based on the area division result.

[0006]

According to the present invention, by dividing the pixels of the color image based on the color similarity, it is possible to directly recognize the presence / absence of the white line, the yellow line, and the sign accurately without erroneous recognition. Becomes

[0007]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be specifically described below with reference to the drawings showing the embodiments. FIG. 1 is a block diagram showing the main configuration of a traveling environment recognition device for a moving vehicle such as an automobile according to the present invention, in which 1 denotes a color image input section constituted by a color image pickup device such as a CCD camera. . A color image picked up by a color image pickup device forming a color image input section is input to the data conversion section 2 as R (red), G (green), and B (blue) light three primary color values for each pixel. . Figure 2 is a CCD onboard
It is explanatory drawing of the color image imaged with the camera, 11 is a driving lane of the own vehicle in the figure, 12 is an adjacent lane, 13 is the white line which shows the boundary of both lanes. The preceding vehicle 14 on the driving lane 11 of the own vehicle
Is imaged with its shadow area 14a. 15 and 16 are soundproof walls provided along the left end of the driving lane 11 of the own vehicle and the right end of the adjacent lane 12. The data conversion unit 2 outputs R, G, B of each pixel of the color image input from the color image input unit 1.
3 which is different from the 3D coordinate system in which the values are R, G and B as the 3 axes
Dimensional space, that is, a parameter representing luminance related to color similarity and two parameters for facilitating separation of saturation and lightness, a total of three parameters are projected onto a three-dimensional coordinate system having three axes, Region division is performed, and each region is replaced with a representative value determined for the divided region and output to the region division unit 3.

FIG. 3 is an explanatory diagram showing the contents of processing in the data conversion unit 2, and FIG.
It is the figure which projected on the three-dimensional coordinate system which makes G and B values each coordinate axis. In Fig. 3 (a), R, G, and B axes have R,
Pixel P (10,10,10) whose G and B values are 10 and 20 respectively
The pixel Q (20,20,20) is The data conversion unit 2 converts each pixel on the O-RGB coordinate system into a value on the O-XYZ coordinate system as shown in FIG. 3 (b). This will be specifically described below. On the O-RGB three-dimensional coordinate system of FIG. 3 (a), the R, G, and B values are the same on the line connecting the origin O and the pixels P and Q, like the pixels P and Q. Pixels are different. Therefore, if the X axis is determined so as to coincide with this line, pixels having different brightness are projected on the X axis. At this time, pixels having different lightness and saturation are projected on the Y axis and the Z axis. OR
The conversion from the GB coordinate system to the O-XYZ coordinate system may be performed as follows. That is, as shown in FIG. 3 (b), the R, G, and B axes in the O-RGB coordinate system are rotated about the B axis by 45 °, and the R axis is rotated by 45 °. , G axis
The axis obtained by rotating by 45 ° is the g-axis, and this is the O-rgB coordinate system. Next, in this O-rgB coordinate system, the B axis and the r axis are each rotated by -θ ° about the g axis, and r
An O-XYZ coordinate system can be obtained by letting the axis obtained by rotating the axis by −θ be the X axis and the axis obtained by rotating the B axis by −θ ° be the Z axis. Where θ is tan θ = 1 / sqr (2) (however, sqr
(n) indicates the square root of n).

As a result, X in the O-XYZ coordinate system
The axis passes through point P shown in FIG. 3 (a), and the R value, B value, and G value of the pixel at each point on the X axis become equal, and the difference in the position of each pixel in the X axis direction causes the difference in brightness. Show. Also the Y axis,
A large absolute value of the Y value and the Z value on the Z axis indicates that the pixel is chromatic, and a small absolute value of the Y value and the Z value indicates that the pixel is achromatic. Further, the area dividing unit 3 divides the three-dimensional space of the O-XYZ coordinate system onto which each input pixel is projected into a plurality of areas. FIG. 4 is an explanatory diagram showing the processing contents of the area dividing unit 3.
Divides the X-axis, Y-axis, and Z-axis into L, M, and N pieces, respectively, and L × M
Divide into N regions. For example, the R, G, B3 primary color values are 256 gradations from 0 to 255, respectively, and the corresponding value range in the X-axis direction is 0 to 255 * sqr (3).

Next, the value of 0 to 255 × sqr (3) is converted into a logarithmic value. For example, the range with base 10 is −∞ to log ₁₀ (255 ×
By dividing sqr (3)) into (L-1) equal parts, it is possible to divide the whole into L in the X-axis direction. The same applies to the directions of the Y and Z axes, and division is performed based on the logarithmic values of the Y and Z axes. Thereby, for example, the X, Y, and Z values of the pixel P are (10 × sqr (3), 0,
It can be expressed as 0).

Pixels projected in such an XYZ space have X values in the range of 5 to 20, as shown in FIG. 3 (c).
If the Y and Z values are located within the range of -5 to 2, respectively, the X, Y, and Z values of the X axis are 12 and Y as the representative value of the median value of 5 and 20, respectively. , Z values are replaced by -3 with the median values of 2 and -5 as representative values. Therefore, the pixel P has X, Y, and Z values of (12, -3, -3).

The color image picked up by this is L ×
Represented as a set of M × N pixels with different values,
It is output to the area dividing unit 3. The region dividing unit 3 divides the color image into regions based on the X, Y, and Z values of each pixel obtained by the data converting unit 2, and the region-divided image is the image recognizing unit 4.
Is output to. The image recognition unit 4 collates the characteristics of the divided areas with the known data held by the image recognition unit 4, and associates them with the actual structure of each area. For example, in the case of a color image as shown in FIG. 2, the following knowledge is used to associate each area with a structure in the real world. (1) The vehicle lane is below the horizon and is larger than any other area below the horizon. (2) The area having the same X, Y, and Z values as the vehicle lane is the road surface. (3) Areas with areas above the horizon are solid objects. (4) The moving vehicle in the image has a shadow at the boundary with the road surface. (5) White lines and yellow lines have parallel lines extending upward in the image. (6) The X value of the pixel in the white line portion is larger than the X value of the pixel in the runway area. In addition, it is possible to recognize the traveling environment in front of the own vehicle by also using information such as the size of the real world area estimated from the position in the image. The image recognition unit 4 recognizes the environment in front of the own vehicle based on the data possessed by the image recognition unit 4 and the value of each region divided by the region division unit 3 in this way. The recognition result is displayed to the driver through the display unit 5 in a manner as shown in FIG.

FIG. 5 is an explanatory diagram of an image displayed on the display unit 5. In the figure, reference numeral 21 indicates the traveling lane area of the own vehicle, and 22 indicates the existence area of the preceding vehicle. Although the three-dimensional space in which a color image is photographed is divided into L × M × N pieces in the above-described embodiment, L, M, and N are not particularly limited, and may be as necessary. And set it.

[0014]

As described above, in the apparatus of the present invention, the R, G, B3 primary color values of each pixel of the color image obtained by photographing are divided into a plurality of areas based on the color similarity, and the area division is performed. The present invention has an excellent effect such that the traveling environment can be easily recognized from the result.

[Brief description of drawings]

FIG. 1 is a block diagram showing a traveling environment recognition device for a mobile vehicle according to the present invention.

2 is an explanatory diagram of an image captured by the color image capturing apparatus shown in FIG.

FIG. 3 is an explanatory diagram showing a conversion mode of a data conversion unit shown in FIG. 1.

4 is an explanatory diagram showing a division aspect of a data conversion unit shown in FIG. 1. FIG.

5 is an explanatory diagram of an image displayed on the display unit shown in FIG.

[Explanation of symbols]

 1 color image input unit 2 data conversion unit 3 area division unit 4 image recognition unit 5 display unit

Continuation of the front page (51) Int.Cl. ⁵ Identification code Office reference number FI Technical display area H04N 7/18 G

Claims (3)

[Claims] 1. A color image pickup device for picking up a color image by picking up an image in front of a moving vehicle, and an area dividing section for dividing each pixel of the color image obtained by the color image pickup device into a plurality of regions based on color similarity. And a traveling environment recognition device for a moving vehicle, comprising: an image recognition unit that recognizes a traveling environment based on the area division result. 2. The area dividing unit projects each pixel of a color image into a value projected in a three-dimensional space different from a three-dimensional space having R (red), G (green) and B (blue) values as three axes. The driving environment recognition device for a mobile vehicle according to claim 1, wherein the image is divided into a plurality of regions based on the image. 3. The area dividing unit projects the R, G, B values of each pixel of a color image onto a three-dimensional space having three axes related to color similarity as three axes, and the three-dimensional space of the three-dimensional space Each axis is divided into L, M, and N, and the three-dimensional space is L × M ×
A method in which the image is divided into N areas, the median value of each divided area is used as a representative value of the area, and the representative value of the area where each pixel is projected is replaced, and the image is divided into areas based on the replaced value. Item 2. A traveling environment recognition device for a moving vehicle according to item 2.