KR20090061923A - Apparatus and method for detecting lane color - Google Patents

Abstract

A lane color recognition apparatus and a method thereof for enhancing the recognition accuracy of a color of a lane from the various road environment are provided to recognize the color about the lane by using color information around the lane. A lane recognizing unit(20) recognizes driving road and all lanes from the forward image of the front of the traveling car photographed by a photographing unit(10). In the image information of the interest region of the front of the running vehicle, a color information output unit produces the color information and the color information of the neighborhood road. A color recognition unit recognizes clearly the color about lane based on the color information of the calculated lane and color information of the calculated road.

Description

Apparatus and method for detecting lane color}

The present invention relates to an apparatus and method for recognizing lane colors, and more particularly, to an apparatus and method for recognizing lane colors from an image in front of a driving vehicle.

The present invention is derived from the research conducted as part of the IT growth engine technology development project of the Ministry of Information and Communication and the Ministry of Information and Communication Research and Development. [Task Management No .: 2005-S-114-03, Project Name: Building / Managing Realistic Content for Telematics] Technology Development for Construction and Management of Tangible Content for Telematics Services].

At present, various techniques have been proposed to recognize lanes and prevent lane departures (commonly referred to as "lane departures" but "lane departures" is a more accurate expression). There are technologies that recognize driving lanes and warn drivers when they are leaving a lane due to drowsiness or other tasks. BACKGROUND ART Techniques for recognizing information with a car driving for navigation have been proposed.

In order to recognize the lane accurately, it is necessary to accurately recognize whether the lane is the rightmost lane or the center line (also referred to as the center lane) through color information as well as the lane line. Here, the color of the lane is calculated using color information, but color distortion occurs due to various factors in the road environment. As a result, the driving lane may not be recognized correctly. In particular, conventionally, there is a problem in that it is not possible to reflect the road environment by converting the color image of the front of the vehicle to obtain the color of the lane using the threshold information.

The present invention has been proposed to solve the above-described problems, and an object of the present invention is to provide a lane color recognition apparatus and method for increasing the accuracy of lane color recognition from various road environments.

In order to achieve the above object, the lane color recognition method according to the preferred embodiment of the present invention, the color information calculation unit, the color information of the lane in the image information of the selected region of interest in front of the running vehicle and the surrounding of the lane A color information calculating step of calculating color information of the roadway; And a lane color recognition step of recognizing a color of the lane based on the calculated color information of the lane and the calculated color information of the road.

The color information calculating step may include selecting a region of interest including a lane in the image information; The values of the red (R), green (G), and blue (B) components at each point on the lane within the selected area of interest, and the red (R), green (G), And calculating an average value of each of the blue (B) components; And the values of the red (R), green (G), and blue (B) components at each calculated point and the red (R), green (G), and blue (B) components of the portion excluding the calculated lanes. Correlating each average value with each other to calculate each difference.

The lane color recognition step may include determining whether a condition based on each calculated difference is satisfied; And if the condition is satisfied, the color of the lane is recognized as yellow, and if the condition is not satisfied, the color of the lane is recognized as white.

The condition is that the difference between the value of the red (R) component at each point on the calculated lane and the average value of the red (R) component at the portion except the calculated lane is calculated from the blue (B) component at each point on the calculated lane. A value greater than the difference between the average value of the blue component (B) except for the calculated lane, and the value of the green component (G) at each point on the calculated lane and the calculated green component (excluding the calculated lane) Information that satisfies the difference between the average value of the value of the blue (B) component at each point on the calculated lane and the average value of the blue (B) component of the portion except the calculated lane is included.

In accordance with another aspect of the present invention, a lane color recognition apparatus includes: a color information calculation unit configured to calculate color information of a lane and color information of a road around a lane from image information of a vehicle in front of a driving vehicle; And a color recognition unit recognizing a color of the lane based on the calculated color information of the lane and the calculated color information of the road.

The color information calculation unit selects a region of interest including a lane from the image information, and calculates values of the red (R), green (G), and blue (B) components at each point on the lane within the selected region of interest. Include as color information.

The color information calculation unit selects a region of interest including a lane in the image information, and calculates an average value of each of the red (R), green (G), and blue (B) components of the selected region of interest except for the lane. Includes color information of the surrounding roads.

The color information calculation unit is configured to calculate the values of the red (R), green (G), and blue (B) components at each calculated point, and the red (R), green (G), and blue (B) portions except the calculated lanes. Each difference of the components is correlated to each other to calculate the difference.

The color recognition unit determines whether the condition based on each calculated difference is satisfied. If the condition is satisfied, the color recognition unit recognizes the color of the lane as yellow, and recognizes the color of the lane as white when the condition is not satisfied.

According to the present invention having such a configuration, the color information of the lanes that are recognized from the image information captured by the photographing apparatus in front of the vehicle is obtained, and together with the color information of the roads around the lanes, the colors of the lanes are finally used by using them. As it is recognized (determined), more accurate color recognition is achieved.

Since it is possible to correctly recognize the color of the lane that is incorrectly recognized by the road environment, it is possible to provide accurate lane information when driving or leaving the right lane of the road.

Hereinafter, a lane color recognition apparatus and method according to an embodiment of the present invention will be described with reference to the accompanying drawings.

1 is a block diagram illustrating a configuration of a lane color recognition apparatus according to an exemplary embodiment of the present invention, and includes a photographing unit 10, a lane recognizing unit 20, and a lane color recognizing unit 30.

The photographing unit 10 is installed in at least one of the front of the vehicle to take a picture of the front of the driving vehicle. The photographing unit 10 may be installed at an angle at which the driving lane, lanes, and all lanes on the road in the driving direction may be photographed. Here, the driving lane means a lane on which the vehicle currently runs. The photographing unit 10 may be installed inside the vehicle due to the influence of weather such as rain and wind, but may be installed outside the vehicle in a good weather day.

The lane recognizing unit 20 recognizes the driving lane and all lanes from an image of the front of the driving vehicle photographed by the photographing unit 10. Various methods are known for an algorithm for recognizing a lane from an image in front of a traveling vehicle. For example, in Korean Patent Application No. 10-2004-0096729 (name; lane departure warning method of a vehicle), the input image is black and white, and the black and white lane image is filtered with a gold filter, and then the filtered image is edged. I use an algorithm that extracts lanes by filtering. In addition to the above-described algorithm, there are various algorithms capable of recognizing lanes. Accordingly, the lane recognizing unit 20 may be considered to use a lane recognizing algorithm well known in the art.

The lane color recognizer 30 detects the color of the lane recognized by the lane recognizer 20 and the color of the road around the lane and finally recognizes (determines) the color of the lane based on the detected color. On the road, the center line (also called the center lane) and the rightmost lane are yellow. In a road environment, it is difficult to distinguish between yellow and white lanes due to backlight or sunlight. In general, in an RGB image, yellow is obtained by (R, G, B) = (255,255,0), blue is obtained by (R, G, B) = (0,0,255), and white is (R Is obtained by (G, B) = (255, 255, 255). The above-described numerical values "0, 255" refer to brightness values. Yellow is greater than R, G and B is blue, B is greater than R and G. By using these points, the lane color recognizer 30 recognizes the recognized lane color.

The lane color recognizer 30 includes a color information calculator 32 and a color recognizer 34.

The color information calculator 32 selects an area of interest (AOI) including a lane from the image information in front of the driving vehicle based on the information from the lane recognizing unit 20. The color information calculator 32 calculates brightness values of the red (R), green (G), and blue (B) components at each point L (x, y) on the lane within the selected region of interest. Here, the brightness values of the red (R), green (G), and blue (B) components at the respective points L (x, y) on the obtained lanes are one of the color information of the lanes according to claim 1 of the present invention. Spring is preferred, for example.

The color information calculator 32 obtains an average value of brightness values of the red (R), green (G), and blue (B) components of the selected region of interest except for the lanes. Here, the average value of the brightness values of the red (R), green (G), and blue (B) components except the obtained lane is preferably an example of the color information of the road around the lane according to claim 1 of the present invention. Do.

The color information calculation unit 32 includes the red (R) and green (G) portions of the calculated red (R), green (G), and blue (B) components at each calculated point except for the calculated lanes. Comparing the average value of the brightness value of the and blue (B) component with each other to calculate the difference. That is, the color information calculator 32 calculates the brightness value L (x, y, R) of the red component R at each calculated point and the brightness value of the red component R except for the calculated lane. The difference (V (x, y, R)) from the average value A (R) of is calculated. The color information calculator 32 calculates an average value of brightness values of the green (G) components at the calculated points, and the brightness values of the green (G) components except the calculated lanes. Calculate the difference V (x, y, G) from (A (G)). The color information calculator 32 calculates an average value of brightness values of the blue (B) component of each of the calculated points (L (x, y, B)) and blue (B) components of the portion excluding the calculated lane. Calculate the difference V (x, y, B) from (A (B)).

The color recognizer 34 determines whether the condition based on each calculated difference V (x, y, R), V (x, y, G), V (x, y, B) is satisfied. . The color recognizer 34 recognizes a yellow or white color of the lane recognized by the lane recognizer 20 according to the determination result. On the other hand, if the brightness of the red (R), green (G) and blue (B) components at each point on the lane calculated by the color information calculator 32 is zero, the color recognizer 34 is zero. Recognize lane color as black.

Here, the condition is that the calculated difference V (x, y, R) is greater than the calculated difference V (x, y, B), and the calculated difference V (x, y, G) is calculated. Contains information that satisfies greater than the difference V (x, y, B). Therefore, the color recognizer 34 recognizes the color of the recognized lane as yellow when the condition is satisfied, and recognizes the color of the recognized lane as white when the condition is not satisfied.

In the above description, it is assumed that the lane color is yellow or white. When the color of the lane is blue (for example, both lanes of the bus lane), it is not described separately. To recognize the blue lanes, blue uses the value of B greater than that of R and G. In other words, the calculated difference V (x, y, B) is greater than the calculated difference V (x, y, R) and the calculated difference V (x, y, B) is calculated If it is larger than (V (x, y, G)), the recognized lane color is recognized as blue.

2 and 3 are diagrams showing examples of lanes that are incorrectly recognized. If only the color information of the lane is used without considering the road environment, the following error occurs.

Referring to FIG. 2, although the rightmost lane (that is, the part indicated by the square box) in the left picture is yellow, when the RGB image is distributed as shown in the graph on the right, it is recognized that the RGB is uniformly distributed in white. .

Referring to FIG. 3, although the second lane from the right in the picture on the left is white, when the RGB image is distributed as shown in the graph on the right, R and G are recognized as yellow larger than B.

The reason for the error as shown in Figs. 2 and 3 is that the color distortion due to various factors in the road environment could not be reflected. In order to solve this problem, the color of the lane is finally determined using the color of the road as well as the color of the lane.

Next, an operation of the lane color recognition apparatus according to an exemplary embodiment of the present invention will be described with reference to the flowchart of FIG. 4.

An image of the front of the driving vehicle photographed by the photographing unit 10 is input to the lane recognizing unit 20 (S10).

The lane recognizing unit 20 recognizes a lane from an input image by using a normal lane recognizing algorithm (S12). The lane recognizing unit 20 sends information of the recognized lane and each point L (x, y) on the obtained lane to the lane color recognizing unit 30.

The color information calculator 32 of the lane color recognizer 30 selects an area of interest (AOI) including a lane (S14).

The color information calculator 32 may calculate an average value A (R) of the brightness values of the red (R) components of the image excluding the lanes in the selected region of interest, and green (G) components of the image excluding the lanes in the selected region of interest. The average value (A (G)) of the brightness value of and the average value (A (B)) of the brightness value of the blue (B) component of the image excluding the lane in the selected region of interest are calculated (S16, S18, S20).

The color information calculating unit 32 is configured to calculate the brightness values L (x, y, R) of the red (R) component at each point L (x, y) on the lane within the selected region of interest, and green G. The brightness value L (x, y, G) of the component and the brightness value L (x, y, B) of the blue (B) component are obtained (S22, S24, S26). Here, the operations of S16, S18, and S20 may be performed after the operations of S22, S24, and S26 are performed.

Thereafter, the color information calculator 32 calculates an average value of the brightness values of the R, G, and B components at each point L (x, y) and the brightness values of the R, G, and B components of the image excluding the lanes. Calculate the difference (S28, S30, S32). That is, the color information calculation unit 32 calculates the brightness value L (x, y, R) of the red component R at each calculated point and the brightness value of the red component R except for the calculated lane. The difference (V (x, y, R)) from the average value A (R) of is calculated. The color information calculator 32 calculates an average value of brightness values of the green (G) components at the calculated points, and the brightness values of the green (G) components except the calculated lanes. Calculate the difference V (x, y, G) from (A (G)). The color information calculator 32 calculates an average value of brightness values of the blue (B) component of each of the calculated points (L (x, y, B)) and blue (B) components of the portion excluding the calculated lane. Calculate the difference V (x, y, B) from (A (B)).

The color information calculator 32 calculates the calculated brightness values L (x, y, R), L (x, y, G), L (x, y, B), and difference (V (x, y, R), V (x, y, G) and V (x, y, B)) are sent to the color recognizer 34.

The color recognizer 34 receives the input brightness values L (x, y, R), L (x, y, G), L (x, y, B), and difference V (x, y, R). Recognize the color of the lane based on V), V (x, y, G), V (x, y, B)).

That is, the color recognizer 34 determines whether the input brightness values L (x, y, R), L (x, y, G), and L (x, y, B) are all zero. If the determination is zero (YES in S34), the color of the lane is recognized as black (S36).

The color recognizer 34 determines whether the input brightness values L (x, y, R), L (x, y, G), and L (x, y, B) are all zero. If not zero (No in S34), the difference V (x, y, R) is greater than the difference V (x, y, B) and the difference V (x, y, G) is the difference (V (x, y, B)) is determined to be greater than (S38).

As a result of the determination, the difference V (x, y, R) is greater than the difference V (x, y, B) and the difference V (x, y, G) is the difference V (x, y, B) Greater than (YES in S38), the color recognition unit 34 recognizes the color of the lane as yellow (S40). Otherwise (NO in S38), the color recognition unit 34 recognizes the color of the lane as white (S42).

On the other hand, the present invention is not limited only to the above-described embodiment, but can be modified and modified within the scope not departing from the gist of the present invention, the technical idea to which such modifications and variations are also applied to the claims Must see

1 is a block diagram showing the configuration of a lane color recognition apparatus according to an embodiment of the present invention.

2 and 3 are diagrams showing examples of lanes that are incorrectly recognized.

4 is a flowchart illustrating a lane color recognition method according to an embodiment of the present invention.

Claims (10)

A color information calculating step of calculating, by a color information calculating unit, color information of a lane and color information of a road around the lane from image information of a selected region of interest in front of a driving vehicle; And And a lane color recognition step of recognizing a color of the lane based on the calculated color information of the lane and the calculated color information of the road. The method according to claim 1, The color information calculating step, Selecting a region of interest including a lane from the image information; Values of red (R), green (G), and blue (B) components at each point on the lane within the selected region of interest and red (R), green (G) of the portion of the selected region of interest excluding lanes. ) And a mean value of each of the blue (B) components; And The red (R), green (G), and blue (B) components of the portions excluding the calculated lanes and the values of the red (R), green (G), and blue (B) components at the calculated points. And calculating each difference by correlating each average value of each other with each other. The method according to claim 2, The lane color recognition step, Determining whether a condition based on each of the calculated differences is satisfied; And And recognizing the color of the lane as yellow when the condition is satisfied, and recognizing the color of the lane as white when the condition is not satisfied. The method according to claim 3, The condition is The difference between the value of the red (R) component at each point on the calculated lane and the average value of the red (R) component at the portion except the calculated lane is the blue (B) component at each point on the calculated lane. Is greater than the difference between the value of and the mean value of the blue (B) components in the The difference between the value of the green (G) component at each point on the calculated lane and the average value of the green (G) component at the portion except the calculated lane is the blue (B) component at each point on the calculated lane. And information satisfying a value larger than a difference between a value of and a mean value of blue (B) components of the portion excluding the calculated lane. A color information calculator configured to calculate color information of a lane and color information of a road around the lane from image information of a vehicle in front of the vehicle which is running; And And a color recognizer configured to recognize a color of the lane based on the calculated color information of the lane and the calculated color information of the road. The method according to claim 5, The color information calculation unit may select a region of interest including a lane from the image information and calculate values of red (R), green (G), and blue (B) components at each point on the lane within the selected region of interest, Lane color recognition apparatus comprising the color information for the lane. The method according to claim 6, The color information calculation unit selects a region of interest including a lane from the image information, and calculates an average value of each of red (R), green (G), and blue (B) components of the selected region of interest except for the lane. And color information of a road around the lane. The method according to claim 7, The color information calculation unit, The values of the red (R), green (G) and blue (B) components at each of the calculated points and the red (R), green (G) and blue (B) components of the parts excluding the calculated lanes, respectively. And calculating a difference between the average values of the lanes. The method according to claim 8, The color recognizing unit determines whether the condition based on each of the calculated differences is satisfied, and if the condition is satisfied, the color recognition unit recognizes the color of the lane as yellow, and if the condition is not satisfied, the color of the lane Lane color recognition device, characterized in that recognized in white. The method according to claim 9, The condition is The difference between the value of the red (R) component at each point on the calculated lane and the average value of the red (R) component at the portion except the calculated lane is the blue (B) component at each point on the calculated lane. Is greater than the difference between the value of and the mean value of the blue (B) components in the The difference between the value of the green (G) component at each point on the calculated lane and the average value of the green (G) component at the portion except the calculated lane is the blue (B) component at each point on the calculated lane. And information satisfying a value greater than a difference between a value of and a mean value of blue (B) components of the portion excluding the calculated lane.

Images