JP3257199B2 - White line recognition device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a white line recognition device, and more particularly to a white line recognition device for recognizing a white line on a road.

[0002]

2. Description of the Related Art Conventionally, there is a white line recognition device for recognizing a white line drawn as a lane on a road in an image of the road captured by an image sensor such as a video camera.

For example, LANELOK: DETECTI
ON OF LANE BOUNDARIES AND
VEHICLE TRACKING USING I
MAGE-PROCESSING TECHNIQUE
S Part II: TEMPLATE MATCHIN
G ALGORITHMS by Surender
K. The method described in Kenue uses the total sum of the absolute values of the luminance levels of the pixels between the predetermined search area and the white line template in the image obtained by the image sensor as an index of the degree of correlation between the search area and the white line template. It recognizes the position of the white line. In this case, if the luminance levels of the white line portion and the asphalt portion that is the background of the white line in the template and the search area are not substantially the same, there is a risk of erroneous recognition. For this reason,
The brightness level of the white line portion and the asphalt portion are obtained by obtaining the brightness distribution or the brightness maximum and minimum values in the search area, and the brightness levels of the white line portion and the asphalt portion of the search area are the brightness of the white line portion and the asphalt portion of the template. The luminance levels of all the pixels in the search area are corrected so as to match the levels.

[0004]

Normally, the number of pixels in the search area must be several times larger than the number of pixels in the template. For this reason, there is a problem that the time required for correcting and rewriting the luminance levels of all the pixels in the search area as in the conventional device becomes longer, and the processing speed of white line recognition becomes slower.

[0005] The present invention has been made in view of the above points,
By correcting the brightness level of each of the white line portion and the background portion of the template to match the brightness level of each of the white line portion and the background portion of the search area, the number of pixels for correcting the brightness level is reduced, and the processing speed is increased. It is an object of the present invention to provide a white line recognition device that performs

[0006]

FIG. 1 shows the principle of the present invention.

In FIG. 1, a correlation calculating means M1 calculates a correlation between a predetermined search area M2 of a road image obtained by an image sensor and a template M3 in which a white line portion and a background portion are set in advance.

[0008] Recognition means M4 recognizes a portion having a high correlation as a white line position.

The search means M5 searches the brightness level of each pixel in the search area M2 for the brightness level of each of the white line portion and the background portion.

The correcting means M6 controls the brightness level of each pixel of the template M3 so that the brightness level of each of the white line portion and the background portion of the template matches the brightness level of each of the white line portion and the background portion of the search area obtained by the search means. Is corrected.

[0011]

In the present invention, in order to correct the luminance levels of the white line portion and the background portion of the template so as to match the respective luminance levels of the white line portion and the background portion of the search region, the conventional brightness level of the search region is corrected. The number of pixels for which the luminance level is to be corrected is reduced as compared with a device that performs the processing, and the processing speed is increased.

[0012]

FIG. 2 is a block diagram showing an embodiment of the apparatus according to the present invention. In the figure, an image signal of an image of a road taken by a video camera 10 is supplied to an image input circuit 11, where the A / D-converted image input circuit 11 supplies image data to an interface circuit 12 in accordance with an instruction from a CPU 20. And
The image data is supplied to and stored in a video RAM 15 via a bus 14.

The CPU 20 reads out image data of a search area from image data of one screen stored in the video RAM 15 based on a processing program stored in the ROM 18 and creates a luminance histogram for each pixel to generate a luminance. The luminance level of each of the white line portion and the asphalt portion is obtained by obtaining the distribution or the maximum value and the minimum value of the luminance in pixel units. Also, the CPU 20 has a RAM
The luminance level of the template is corrected so that the luminance levels of the image data of the white line portion and the asphalt portion of the template stored in advance in 16 become the luminance levels of the white line portion and the asphalt portion of the search area, respectively. . Thereafter, the image data of the search area stored in the video RAM 15 and the image data of the template stored in the RAM 16 are transferred to the correlation operation processor 21. Further, the position and the inclination of the white line in the search area are recognized from the degree of correlation notified from the correlation operation processor 21, and the position and the inclination of the white line are output to an external system such as an automatic operation control system through the interface circuit 17.

The correlation operation processor 21 corresponding to the correlation operation means M1 moves the template on a pixel-by-pixel basis in the search area, and calculates the absolute difference between the pixel data of the corresponding search area and the pixel data of the template at each position. The sum of the values is determined as a correlation value, and the correlation between the search area and the template is considered to be the highest at the position where the correlation value is minimum,
The CPU 20 is notified of the center position of the white line portion of the template at that position together with the correlation value.

FIG. 3 shows a flowchart of the white line recognition process executed by the CPU 20 of the apparatus of the present invention. In the figure, in step S10, an initial setting of a search area is performed. Here, FIG. 4 search area 33 ₁ ~ 33 _2n against the white line 31 in the image 30 of a road shown in (A) (n is, for example, 9) is provided, each search region 33 ₁ ~ 33 _2n image The vertical position within 30 is fixed and the horizontal position is variable. By default, the horizontal position of the people each search area 33 ₁ ~ 33 _n respectively are set to typical values as shown in FIG. 4 (A).

When one screen has 512 × 512 pixels,
Each search area 33₁~ 33_2nIs, for example, 32 x vertical
It is composed of 23 pixels in the direction. In addition, the search area 33₁~ 3
3_2n2n templates corresponding to each are stored in the RAM 16
Each template is stored in the horizontal direction 16 × vertical direction
FIG. 4B shows a search area 33. ₁
Shows a template corresponding to. However, in contrast to FIG.
35 is white line, 30 is asphalt
G (background part).

In step S20, the CPU 20 causes the image input circuit 11 to supply image data for one screen, and fetches the image data into the video RAM 15. Next, step S3
0 is set to 1 i indicating the number of the 2n search area 33 ₁ ~ 33 _2n.

Next, step S corresponding to the search means M5
At 40, the image data of the i-th search area is stored in the video RAM1.
5, the maximum value and the minimum value of the luminance are searched, and the maximum value of the luminance is set as the luminance level of the white line portion, and the minimum value of the luminance is set as the luminance level of the asphalt portion. Step S40
In creating the i-th search area luminance histogram as shown in FIG. 5 for the luminance level of the white line portion of the luminance level B ₁ to the number of pixels with high luminance to form a peak, luminance pixel number at low intensity to form a peak The level B ₀ may be set as the luminance level of the asphalt portion.

Next, step S5 corresponding to the correction means M6
At 0, the luminance levels of the white line portion 35 and the asphalt portion 36 of the i-th template are obtained at step S40.
Correction is made so that the brightness levels of the white line portion and the asphalt portion of the second search area are respectively set.

Thereafter, in step S60, the CPU 20
All the image data of the i-th template and all the image data of the i-th template are transferred to the correlation operation processor 21 to perform the correlation operation. In step S70, the position of the center of the white line portion and the correlation value notified from the correlation operation processor 21 are stored in the RAM 16.

In step S80, it is determined whether or not i is 2n or more. If i is less than 2n, i is set to 1 in step S90.
Then, the process proceeds to step S40, and a white line in the next search area is detected. When the white line detection is performed for the 2n search area 33 ₁ ~ 33 _2n all proceeds to step S100.

Step S100 corresponding to the recognition means M4
In the white line portion detected in the search area 33 ₁ ~ 33 _2n, respectively, white line correlation value to select only those deemed white line below the predetermined reference value, selected from the among the search area 33 ₁ ~ 33 _n Based on the position of the part, the position and inclination of the white line 31 are obtained by, for example, the least square method, and similarly, the position and inclination of the white line 32 are obtained based on the position of the white line part selected from the search areas 33 _{n + 1 to} 332 _n . In this way, the selection of the one whose correlation value is equal to or smaller than the reference value is made by the white lines 31 and 32 of the road.
May be broken.

Next, in step S110, it is determined whether or not the position of the vanishing point 34 where the extended lines of the white lines 31 and 32 intersect on the image 30 is within a predetermined range from the position of the previous vanishing point. Proceeding to S120, it is determined that there is no misrecognition of the white lines 31 and 32.
The position and inclination of each of the first and second 32
Output to more external systems. Further, step S
130 in the search area 33 ₁ ~ 33 _2n respective horizontal center position in the direction position to adjust the horizontal position of the people next search area 33 ₁ ~ 33 _2n respectively so as to overlap the people white line 31, 32 respectively to set, in step S20 move on.

If the vanishing point position is not within a predetermined range from the previous vanishing point position in step S110, the white line 31,
Since there is a high possibility that the recognition of No. 32 was erroneous, the process proceeds to step S20 without adopting this.

As described above, the brightness level of each pixel of the template is corrected so that the brightness level of each of the white line portion and the asphalt portion of the template matches the brightness level of each of the white line portion and the asphalt portion of the search area. The number of pixels to be corrected is 16 × 8 pixels, which is greatly reduced from the number of pixels to be corrected of 32 × 23 pixels when correcting the luminance level of each pixel in the search area as in the related art. For this reason, the time required for the entire recognition process is reduced, and the speed of the process is increased.

[0026]

According to the white line recognition device of the present invention, the luminance level of each of the white line portion and the background portion of the template is corrected so as to match the luminance level of each of the white line portion and the background portion of the search area. Is reduced, the processing speed is increased, and this is extremely useful in practice.

[Brief description of the drawings]

FIG. 1 is a principle diagram of the present invention.

FIG. 2 is a block diagram of the device of the present invention.

FIG. 3 is a flowchart of a white line recognition process.

FIG. 4 is a diagram for explaining a road image and a template.

FIG. 5 is a diagram showing a luminance histogram.

[Explanation of symbols]

 Reference Signs List 10 video camera 11 image input circuit 12, 17 interface circuit 15 video RAM 16 RAM 18 ROM 20 CPU 21 correlation calculation processor M1 correlation calculation means M2 search area M3 template M4 recognition means M5 search means M6 correction means

Continuation of the front page (58) Field surveyed (Int.Cl. ⁷ , DB name) G06T 1/00 G06T ⁷ /00-7/60 H04N 7/18 G05D 1/00-1/12 G08G 1/00-9 / 02

Claims (1)

(57) [Claims] 1. A white line that calculates a correlation between a predetermined search area of a road image obtained by an image sensor and a template whose white line portion and background portion are set in advance, and recognizes a portion having a high correlation as a white line position. In the recognition device, a search unit for searching the brightness level of each of the white line portion and the background portion from the brightness level of each pixel in the search area, and the brightness level of each of the white line portion and the background portion of the template are obtained by the search unit. Correction means for correcting the brightness level of each pixel of the template so as to match the brightness level of each of the white line portion and the background portion of the search area; and calculating the correlation between the search area and the corrected template to obtain a white line A white line recognition device characterized by recognizing a position.