JP4526963B2 - Lane mark extraction device - Google Patents

Description

  The present invention relates to a lane mark extraction device that extracts lane marks of a white line or a yellow line in a road area.

  Research is underway to recognize the lane marks laid on the road and assist the driver. As a lane mark detection method, a method is known in which the luminance of an image is compared with a predetermined value, and if the luminance value is equal to or higher than a predetermined value, it is determined as a lane mark. That is, many lane marks are shown in white, and white has higher brightness than surrounding roads. From this, the brightness of the image is detected, and if the brightness value of the image is a predetermined value or more, it can be determined as a white lane mark.

  However, lane marks are not only white but also yellow lane marks. In such a conventional lane mark detection method, the white lane mark and the yellow lane mark cannot be distinguished and detected.

As a lane mark detection device capable of detecting a white lane mark and a yellow lane mark, there is a device as shown in Patent Document 1. That is, Patent Document 1 discloses an imaging device that outputs a color signal that constitutes a road surface image, and a lane mark enhancement conversion that converts a color signal into a grayscale signal so as to obtain a grayscale image in which a lane mark portion of the color image is enhanced. It is determined by investigating the white or yellow lane mark in the color space and the distribution status other than the lane mark, and statistically determining the axis that most separates the lane mark and the other distribution in the color space. The conversion coefficient is calculated to extract white and yellow lane marks.
JP 2002-123819 A

  However, the lane mark detection apparatus disclosed in Patent Document 1 is an apparatus that improves the extraction accuracy of white and yellow lane marks by emphasizing the lane marks. There is a problem that it is impossible to extract a yellow line that has disappeared, and a white lane mark that has turned yellow under the influence of yellow headlights and yellow road lights at night is erroneously extracted as a yellow lane mark.

  Accordingly, an object of the present invention is to provide a lane mark detection device that can detect and distinguish a white lane mark and a yellow lane mark without being affected by a light source of a chromatic road light or a headlight. To do.

In order to solve the above-mentioned problem, a lane mark detection device according to the invention of claim 1 is based on an imaging device that captures an image including a road region in the vicinity of a vehicle, and an image signal from the imaging device. based on the average value of the color signal calculated for each of the plurality of color signals included in the image signal indicating an area to calculate the brightness values of the road area, for each of the color signal, the color lightness value of the road area by subtracting the average value of each signal, and road region color measurement processing unit that measures a color component of the road area deviation of the color signal of the road area indicating the tendency of the color of the road area, the road region color the color components of the road area measured by the measurement processing unit, for each of the color signals, and an image correction processing unit that performs color correction to be added to the image signal from the imaging device, the color corrected image signal Use the lane Characterized by being adapted to extract over click.

Lane mark detecting apparatus according to the invention of claim 2 is the invention of claim 1, wherein the road region color measurement processing unit, the average value of the RGB signal is the color signal of the image signal indicative of the road area RGB was calculated for each signal, calculated by averaging the average value of the RGB signal is calculated brightness value of the road area for each of the RGB signals, for each of the RGB signal, the R GB lightness value of the road area by subtracting the signal of the average value, respectively, characterized in that so as to obtain the color components of the road area.

  According to the present invention, the color component of the road region is measured from the image signal from the imaging device, the color correction of the image signal from the imaging device is performed according to the measured color component of the road region, and the color correction is performed. The lane mark is extracted using the processed image signal. For this reason, it is possible to distinguish and detect the yellow lane mark and the white lane mark without being affected by the light source of the chromatic road light or the headlight.

  Further, according to the present invention, the white lane mark and the yellow lane mark can be distinguished and detected, so that a center line recognition function and a vehicle lane change detection function can be realized.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. In FIG. 1, an imaging device 1 captures a landscape including a road area in the vicinity of a vehicle. As shown in FIG. 2, the imaging device 1 outputs an R (red) G (green) B (blue) image signal including a road area AR1 in the vicinity of the vehicle.

  The RGB image signal from the imaging device 1 is sent to a lane mark extraction processing unit 2 surrounded by a wavy line. The lane mark extraction processing unit 2 performs color correction according to the color component of the road area AR1 measured by the road area color measurement processing unit 11 and the road area color measurement processing unit 11 that measures the color component of the road area AR1. An image correction processing unit 12, a color conversion processing unit 13 for converting an RGB image signal into an HSV image signal, and a yellow line color for setting a threshold value for estimating a color range of a yellow line from the converted HSV image signal The range estimation processing unit 14, the white line range estimation processing unit 15 that sets a threshold for estimating the color range of the white line from the converted HSV image signal, and the threshold set by the yellow line color range estimation processing unit 14 The color extraction processing unit 16 extracts a line binary image and the white line extraction processing unit 17 extracts a white line binary image based on the threshold set by the white line range estimation processing unit 15.

  The RGB signals from the imaging device 1 are sent to the road area color measurement processing unit 11. The road area color measurement processing unit 11 measures the color component of the road area AR1 by the light source of the chromatic road light or the headlight as follows.

First, average values Rmean, Gmean, and Bmean of RGB signals of the road area AR1 input from the imaging device 1 are calculated. Then, the average values Rmean, Gmean, and Bmean of the calculated RGB signals are averaged, and the lightness value Gray of the road region is calculated as in the following equation.
Gray = (Rmean + Gmean + Bmean) / 3 (1)

If the road area AR1 is achromatic, the signal values of the color components of the RGB signals of the road area AR1 from the imaging device 1 are equal,
Rmean = Gmean = Bmean
It becomes. For this reason, if the brightness value Gray of the road area AR1 is calculated as shown in the above equation,
Gray = Rmean = Gmean = Bmean (2)
Should be.

  However, the actually photographed road area AR1 is discolored by the color of the chromatic road light or the light source of the headlight and does not strictly satisfy the relational expression expressed by the expression (2). For example, when the road area AR1 is discolored by a yellow headlight, yellow is an additive color mixture of red and green, so the average value Rmean of the R signal and the average value Gmean of the G signal are the average of the B signal. It becomes larger than the value Bmean.

Since the brightness value Gray is obtained by averaging the average values Rmean, Gmean, and Bmean of the RGB signals as shown in the equation (1), the brightness value Gray determined by the equation (1) is actually Differences from the average values Rmean, Gmean, and Bmean of the respective color components of the RGB signal obtained from the imaging device 1 are the color components of the road area AR1. That is, the color components (Roff, Goff, Boff) of the road area AR1 are
(Roff, Goff, Boff) = (Gray-Rmean, Gray-Gmean, Gray-Bmean) (3)
Can be calculated as

  As described above, the road area color measurement processing unit 11 measures the color components of the road area AR1 due to the light source of the chromatic road light or the headlight. FIG. 3 is a flowchart showing the measurement processing of the color component of such a road area.

  In FIG. 3, the sum Rsum, Gsum, and Bsum of RGB signal values of the image signal of the input road area AR1 are calculated (step S1), and the sum Rsum, Gsum, and Bsum of each image signal are calculated as the number of pixels in the road area AR1. The average values Rmean, Gmean, and Bmean are calculated by dividing by (step S2). Then, the lightness value Gray is calculated from the calculated average values Rmean, Gmean, and Bmean of the RGB signals by the calculation shown in the equation (1) (step S3). Then, as shown in equation (3), the color components (Roff, Goff, Boff) of the road area AR1 are obtained (step S4).

In FIG. 1, the color component of the road area AR1 measured by the road area color measurement processing unit 11 and the light source of the chromatic color road light or headlight is sent to the image correction processing unit 12. In the image correction processing unit 12, the signal value of the input image signal is corrected by the color component of the road area AR1. That is, the color components (Roff, Goff, Boff) of the road area AR1 obtained by the equation (3) are added to the signal values (Rsource, Gsource, Bsource) of the input image signal, and the following correction is made. Image signal values (Rmodify, Gmodify, Bmodify) are obtained.
(Rmodify, Gmodify, Bmodify) = (Rsource + Roff, Gsource + Goff, Bsource + Boff)

  As a result of this image correction processing, the RGB signal values (Rmodify, Gmodify, Bmodify) input from the imaging device 1 are converted into color component amounts of the road area AR1, as shown in FIGS. 4 (A) to 4 (C). Will only be slid up and down. 4A to 4C, the horizontal axis represents the input signal value, and the vertical axis represents the output signal value. By adding the color component amounts of the road area AR1, the RGB signal values slide up and down as indicated by modify-line.

  In FIG. 1, the RGB signal whose image has been corrected by the image correction processing unit 12 is sent to the color conversion processing unit 13. The color conversion processing unit 13 converts the RGB signal value into a color signal in the HSV space composed of three components of lightness, saturation, and hue. Here, the RGB signal values are converted into the HSV color system, but may be converted into another color system such as CIE Lab.

  The HSV space signal from the color conversion processing unit 13 is sent to the yellow line color range estimation processing unit 14 and the white line range estimation processing unit 15.

  The yellow line color range estimation processing unit 14 estimates the existence range of the yellow line from the color signal in the HSV space. That is, the yellow line is around the yellow (60 degrees) hue angle in the HSV color space. Since the chromatic yellow line is on the achromatic road, it is higher than the saturation value of the road area. Therefore, the yellow line color range estimation processing unit 14 obtains a saturation value having a yellow hue angle among the pixels of the road area AR1, thereby obtaining a threshold value of the yellow saturation value for dividing the road area and the yellow line area. calculate.

Specifically, in the road area AR1, the standard deviation σs of the saturation values and average value Smean of the pixels whose hue angle satisfies a preset yellow range (around 60 degrees) are calculated. In this embodiment, the yellow line saturation limit value Slimit is set to
Slimit = Smean + 3σs
The lower limit of the yellow line saturation is calculated.

  The white line range estimation processing unit 15 estimates the existence range of the white line from the color signal in the HSV space. That is, the white line is achromatic and has high luminance. Therefore, the white line range estimation processing unit 15 calculates the brightness of the achromatic pixel of the road area AR1 as a threshold for dividing the road area and the white line area.

  The threshold value set by the yellow line color range estimation processing unit 14 is sent to the color extraction processing unit 16. The color extraction processing unit 16 performs yellow line extraction processing from the input image with the set threshold value. Thereby, a yellow line binary image is formed.

  The threshold set by the white line range estimation processing unit 15 is sent to the white line extraction processing unit 17. The white line extraction processing unit 17 performs white line extraction processing from the input image with the set threshold value. Thereby, a white line binary image is formed.

  As described above, a yellow line binary image and a white line binary image are formed. A yellow line mark and a white line mark are extracted from the yellow line binary image and the white line binary image.

  As described above, in the embodiment of the present invention, the color component of the road region is measured from the image signal from the imaging device, and the image signal from the imaging device is measured according to the measured color component of the road region. Color correction is performed, and a lane mark is extracted using the color-corrected image signal. For this reason, it is possible to distinguish and detect the yellow lane mark and the white lane mark without being affected by the light source of the chromatic road light or the headlight.

  The present invention is not limited to the above-described embodiments, and various modifications and applications can be made without departing from the gist of the present invention.

  INDUSTRIAL APPLICABILITY The present invention can be used for laying on a road, recognizing a lane mark, giving a warning to a driver that the vehicle is operating along a traveling lane, and performing steering control.

It is a block diagram of an embodiment of the present invention. It is explanatory drawing of embodiment of this invention. It is a flowchart used for description of a road area color measurement process. It is a graph used for description of image correction processing.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Imaging device 2 Lane mark extraction process part 11 Road area color measurement process part 12 Image correction process part 13 Color conversion process part 14 Yellow line color range estimation process part 15 White line range estimation process part 16 Color extraction process part 17 White line extraction process part

Claims (2)

An imaging device that captures an image including a road area in the vicinity of the vehicle;
Based on an image signal from the imaging device, a lightness value of the road region is calculated based on an average value of the color signals calculated for each of a plurality of color signals included in an image signal indicating the road region, and the color for each signal, by subtracting the average value for each of the color signal from the brightness values of the road area, measuring the difference of the color signal of the road area indicating the tendency of the color of the road area as the color component of the road area A road area color measurement processing unit,
The color components of the road area measured by the road area color measurement processing unit, for each of the color signals, and an image correction processing unit that performs color correction to be added to the image signal from the imaging device,
A lane mark extraction device, wherein a lane mark is extracted using the color-corrected image signal. The road area color measurement processing unit
An average value of RGB signals that are the plurality of color signals included in the image signal indicating the road area is calculated for each RGB signal ,
By averaging the average value of the RGB signal calculated for each of the RGB signals is calculated as the brightness value of the road area,
Lane mark according to claim 1, wherein for each RGB signal, wherein the by subtracting the average value of the R GB signals from each of the brightness values of the road area and to determine the color component of the road area Extraction device.

Images