JP2735033B2 - Lane change detection apparatus and method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an IT installed on a road.
The present invention relates to a lane change detection device and method for detecting a lane change of a vehicle by performing image processing on a video signal from a V (Industrial TV) camera or the like.

[0002]

2. Description of the Related Art Hitherto, means for detecting various types of traffic information have been proposed in order to construct a traffic information system for realizing a smooth traffic flow. As one of them, there is a method in which a video signal from an ITV camera is subjected to image processing to detect a lane change of a vehicle to obtain traffic information.

[0003] As a lane change detection method for detecting a lane change of a vehicle by performing image processing on a video signal from an ITV camera, a moving object (vehicle) is detected by using an inter-frame difference.
There is a method of detecting a lane change by tracking the detected moving object (vehicle). FIG. 6 is a conceptual diagram for explaining the conventional lane change detection method. FIG. 6A shows images (frames 1 to 3) showing a state in which a vehicle has changed lanes from a traveling lane to an overtaking lane. FIG. 2B shows the difference between frames, and FIG. 2B shows frames 1 to 3 shown in FIG.
FIG. 3 is a diagram showing the XY coordinates of the center of gravity (P1 to P3) of the moving body (vehicle) in FIG. In the figures, the images of frames 1 to 3 have the traveling lane as area A and the passing lane as area B. In frames 1 and 2, vehicle a1 moves to vehicle a2 in area A, and in frame 3, vehicle a2 moves. The vehicle is moving to the vehicle a3 in the area B (change of lane). Less than,
This lane change detection method will be specifically described with reference to FIG.

In a conventional lane change detection method, for example, FIG.
When the frame 2 shown in (a) is the current frame and the frame 1 is the previous frame, when the difference between these frames is obtained, it is extracted as an area where the vehicle a1 has disappeared and an area where the vehicle a2 has increased, and these extracted. The X and Y coordinates of the position of the center of gravity of each moving body are determined using the areas as moving bodies. Then, the position of the center of gravity of the moving body in the current frame 2 and the previous frame 1
And the position of the center of gravity of the moving object is determined, and the moving point is obtained. For example, the vehicle a in the front frame 1 closest to the point P2 (x ₂ , y ₂ ) which is the position of the center of gravity of the vehicle a 2 in the current frame 2
The point P1 (x ₁ , y ₁ ) which is the position of the center of gravity of point 1 is the point P2
(X ₂ , y ₂ ), and it is determined that the point P 1 (x ₁ , y ₁ ) has moved to the point P ₂ (x ₂ , y ₂ ) (FIG. 6).
(B)).

Similarly, when the frame 3 is the current frame and the frame 2 is the previous frame, the vehicles a2 and a3 are extracted as moving objects from the difference between the frames, and the point P2 ( x ₂ , y ₂ ), point P
3 (x ₃ , y ₃ ) is obtained, and the point P ₂ (x ₂ , y ₂ ) is
3 (x ₃ , y ₃ ) (see FIG. 6B).

When the moving point is detected as described above,
It is determined whether or not the movement is a movement to another area. If it is determined that the movement is to another area, the lane is changed. The point P1 (x ₁ ,
y ₁ )) to point P ₂ (x ₂ , y ₂ )
It is not a lane change because it is a movement within. On the other hand, the point in the frame _{2,3 P2 (x 2, y 2} ) from the point P3 (x
_Since the movement to ( ₃ , y ₃ ) is from area A to area B, the lane change is performed.

As described above, in the conventional lane change detection method, the moving body is obtained from the difference between the current frame and the previous frame, the movement of the obtained center of gravity of the moving body is traced (trace processing), and the center of gravity is calculated. The lane change is detected by determining whether the movement of the position is to another area.

[0008]

However, the above-described conventional lane change detection method has the following problems.

The detection result of the lane change is used as traffic information. For example, when it is detected that lane changes occur intensively in a certain direction, information that there is an obstacle on the road (for example, an accident vehicle, a falling object, etc.) is provided as traffic information. Is done. Therefore, real-time processing is required to detect a lane change. However, in the conventional lane change detection method, a trace process is used in which the center of gravity of the moving object extracted by the inter-frame difference is obtained and the position of the center of gravity is tracked, and this trace process is difficult to process. However, there is a problem that real-time processing cannot be performed. Further, there is a problem that the hardware has a complicated configuration.

An object of the present invention is to provide a lane change detection device and a detection method which can solve the above-mentioned problem, can detect lane change by simple processing, and have a simple hardware configuration.

[0011]

SUMMARY OF THE INVENTION A lane change detecting device according to the present invention is provided.
Is a lane change detecting means for detecting a lane change on a road divided into a first lane and a second lane, and is displayed on the basis of individual frames taken successively in time on the road. for the display image, the first lane side area a, the first lane and the center portion of the area B relative to the line that divides the second lane, the second area lane side C Dividing means for dividing into three areas, a difference between the current frame and the previous frame, and an inter-frame difference detecting means for binarizing the result of the difference, and an inter-frame difference for each scanning line forming the display image. A scanning line in which the difference result binarized by the difference detecting means is “0” in the area A and the area C and “1” in the area B is “1”.
And the other scanning lines are pattern processing means for performing pattern processing of “0”, and the sum of the results of pattern processing performed by the pattern processing means is calculated. Lane change determining means for determining lane change.

A lane change detecting method according to the present invention is a lane change detecting method for detecting a lane change on a road divided into a first lane and a second lane. for a display image displayed on the basis of the individual frames, the area of the first lane side a, the area of the central portion of the first lane and the reference line for dividing the second lane B, the second A first step of dividing the lane side portion into three areas having an area C, a second step of obtaining a difference between the current frame and the previous frame, and binarizing the difference result, For each scanning line to be formed, the difference result binarized in the second step is “0” in area A and area C, and “1” in area B.
Is set to “1”, and the other scanning lines are set to “0”.
A third step of performing a pattern process,
Take the sum of the results of pattern processing in step
And determining a lane change when the value is equal to or greater than a predetermined threshold.

[0013]

According to the lane change detection apparatus and method of the present invention, a road is divided into three areas A to C on a display image. Since the vehicle always passes through the area B when changing lanes, the lane change can be detected by detecting that the vehicle has passed the area B. In the present invention, the difference result obtained by the inter-frame difference is binarized, and for each scanning line related to the display image, the binarized difference result is “0” in area A and area C, and “1” in area B. The moving object in the area B is detected by performing a pattern process of setting the scanning line to be “1” and setting the other scanning lines to “0”. In this pattern processing, if there is a moving body in the area B, the scanning line of the moving body portion is set to “1”. You can find the size of your body. Therefore, if the size that can be recognized as a vehicle is set as the threshold value, it is possible to detect the passage of the vehicle in the area B when the sum of the scanning lines that have become “1” in the pattern processing is equal to or larger than the threshold value. Thus, the lane change can be detected. in this way,
According to the present invention, the lane change can be detected by a simple process without using a tracing process that is complicated and takes a long time as in the related art, so that real-time processing can be performed.

[0014]

Next, embodiments of the present invention will be described with reference to the drawings.

FIG. 1 shows lane change detection according to one embodiment of the present invention.
FIG. 2 is a block diagram illustrating a schematic configuration of the device .

In FIG. 1, reference numeral 1 denotes a CPU (Central Processing Unit) to which input means 2 such as a keyboard and a mouse, a frame buffer 4, and a lane change detection processing section 6 are connected.

A video signal from an ITV camera (not shown) installed on the road is input to the frame buffer 4 via an A / D converter 5, and individual framed images taken successively in time are received. A plurality of images (that is, frames) are held. The individual images held in the frame buffer 4 are displayed on display means (not shown) such as a CRT.

The input means 2 is used to designate the position and coordinates on the image displayed on the display means at the time of area division described later, and is used here as a division means for performing area division. In this embodiment, as shown in FIG. 2, the display means is divided into a driving lane and an overtaking lane.
The lane road (including the traveling vehicle) is displayed, and the displayed 2
The lane road is divided into three areas A to C.
Here, area A indicates a side portion of the traveling lane, area B indicates a central portion centered on a center line that separates the traveling lane from the overtaking lane, and area C indicates a side portion of the overtaking lane. In addition, area division,
The state is not limited to the state shown in FIG. 2 and is appropriately determined according to the camera angle with respect to the road and the like. For example, when the camera is photographed around the center line, area B is an area where the part on the traveling lane side is half of the lane, and the part on the passing lane side is half of the lane. It is desirable to do.

The lane change detection processing unit 6 changes the lane of the traveling vehicle (here, the overtaking lane from the traveling lane) in the display image displayed on the display means based on the individual frames taken successively in time. Lane change and lane change from the overtaking lane to the driving lane, and also includes a case where the vehicle simply passes through a center line that separates these lanes.) 3,
It comprises an inter-frame difference unit 7, a pattern processing unit 8, and a lane change determination unit 9. Hereinafter, each part will be described in detail.

The area information storage memory 3 is a memory for storing the area information divided by the input means 2 described above. The stored area information is output to the pattern processing unit 8.

The inter-frame difference section 7 is a section for obtaining a difference between two consecutive frames held in the frame buffer 4 and extracting a vehicle as a moving body. In this embodiment, since the display means is a television screen, the effective scanning line is 480.
This is a book, and the difference result is binarized (for example, the moving area portion is set to “1” and the other portions are set to “0”) for each scanning line, and the vehicle is detected. The difference between the frames has been described in the above-mentioned conventional example, and therefore will not be described here.

The pattern processing unit 8 determines, for each scanning line,
The scanning line of the pattern in which the difference result binarized by the inter-frame difference unit 7 is “0” in the area A, “1” in the area B, and “0” in the area C is “1”. The scanning lines of the other patterns are set to “0”, and are units for performing processing (projection in the horizontal direction) for converting the pattern of each scanning line into “1” and “0”.
Specifically, the following pattern processing (horizontal projection) is performed.

FIGS. 3A and 3B are diagrams showing horizontal projections by the pattern processing section 8 on the frames 1 to 4 showing a state where the vehicle has changed lanes from the traveling lane to the passing lane. (B) is between frames 2 and 3, (c)
Is for frames 3 and 4.

The images when the traveling vehicle approaches the center lane in the traveling lane are frames 1 and 2, and the horizontal projection results in this case are as shown in FIG. Become. Difference between frames 1 and 2 (vehicle)
Is a vehicle a1 covering both the area A and the area B
And a vehicle a in an area B where a part of the vehicle overlaps the vehicle a1
The scanning line in which the pattern is "1" in this case is a scanning line of a portion of the vehicle a2 that does not overlap with the vehicle a1. Therefore, as a result of the projection in the horizontal direction, only the portion of the vehicle a2 which does not overlap the vehicle a1 out of the 480 scanning lines becomes “1”.

An image in which the traveling vehicle has moved on the center line after the frame 2 is the frame 3. In this case, the result of the lateral projection is as shown in FIG. 3B. Difference between frames 2 and 3 (vehicle)
Means that the vehicle a2 in the area B and a part of the vehicle are the vehicle a2
In this case, the scanning line in which the pattern is “1” in the area B overlapping the area B is the scanning line of the vehicle a2 and the vehicle a3. Therefore, as a result of the projection in the horizontal direction, the vehicle a2 and the vehicle a3 of the 480 scanning lines are “1”.

An image in which the traveling vehicle has moved to the overtaking lane following the frame 3 is the frame 4. In this case, the result of the lateral projection is as shown in FIG. 3C. Difference between frames 3 and 4 (vehicle)
Are the vehicle a3 in the area B, the area B and the area C
And vehicle a in which a part of the vehicle overlaps vehicle a1
The scanning line in which the pattern is "1" in this case is a scanning line of a portion of the vehicle a3 that does not overlap with the vehicle a4. Therefore, as a result of the projection in the horizontal direction, only the portion of the vehicle a3 that does not overlap with the vehicle a4 among the 480 scanning lines is “1”.

The lane change judging section 9 is a section for judging a lane change of the vehicle on the basis of the result of the horizontal projection performed by the pattern processing section 8 described above. The lane change determination unit 9 calculates the total sum of the projection results of the 480 scanning lines in the horizontal direction, and determines that the lane has been changed if the sum is equal to or greater than a predetermined threshold value. The threshold value is determined in consideration of the length of the vehicle on the screen (depending on the position on the screen). Here, in a predetermined range on the screen, a value at which the moving object can be determined as a vehicle is set as the threshold. The setting of the threshold value is performed by the input means 2 described above.

Next, detection of a lane change by the lane change detecting device will be described. FIG. 4 is a flowchart showing the procedure of the lane change detection.

A video signal from an ITV camera (not shown) is input via an A / D converter 5, and a plurality of frames of individual images (that is, frames) continuously captured in time are held. You. Individual images from an ITV camera (not shown) held in the frame buffer 4 are displayed on display means.

First, the road is divided into three areas A to C on the screen (step S101). When the area division is performed, the difference between the current frame and the previous frame is calculated, and the difference result is set to 2 for each effective scanning line.
A value is extracted to extract a vehicle as a moving body (step S10)
2). When vehicle extraction is performed, subsequently, for each scanning line, the scanning line in which the area A and the area C are “0” and the area B is “1” is “1”, and the other scanning lines are “0”. (Projection in the horizontal direction) is performed (step S103). When the pattern processing is performed, subsequently, it is determined whether or not the total sum of the pattern processing results is equal to or more than the threshold value. A determination is made (step S104).

In the above-described lane change detection procedure, step 101 may be performed after step 102.

As described above, the lane change detecting means and method of the present embodiment do not require complicated processing for detecting lane change, and can detect lane change by simple algorithm processing. The detected lane change is used as traffic information as an information processing system via a communication line,
For example, it is sent to a traffic information management system.

When a camera image is taken from the lateral direction as shown in FIG. 5, it is conceivable that a lane change is erroneously detected. However, in this embodiment, pattern processing (projection in the lateral direction) is performed. Since the processing is performed for the three areas A to C, if the area division is set according to the screen, the lane change is not detected erroneously even in the case shown in FIG.

[0034]

As described above, the present invention is configured so that lane change can be detected by simple algorithm processing, real-time processing is possible, and real-time information is used as traffic information. There is an effect that it can be provided.

In addition, since only simple algorithm processing is required, large-scale hardware is not required, and there is an effect that cost can be reduced.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a schematic configuration of a lane change detection device according to an embodiment of the present invention.

FIG. 2 is a diagram showing a state where a displayed two-lane road is divided into three areas A to C;

3A and 3B are diagrams showing pattern processing for frames 1 to 4 showing a state in which a vehicle has changed lanes from a traveling lane to a passing lane, and FIG.
(B) is for frames 2 and 3, and (c) is for frames 3 and 4.

FIG. 4 is a flowchart showing a procedure of lane change detection in the lane change detection device shown in FIG.

FIG. 5 is a diagram illustrating an example of a screen captured by a camera from a horizontal direction.

FIG. 6 is a conceptual diagram for explaining a conventional lane change detection method, in which (a) shows a frame-to-frame difference between frames 1 to 3 showing a situation where a vehicle has changed lanes from a traveling lane to an overtaking lane. (B) shows the XY coordinates of the position of the center of gravity of the moving object (vehicle) in frames 1 to 3 shown in (a).

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 CPU 2 Input means 3 Area information storage memory 4 Frame buffer 5 A / D converter 6 Lane change detection processing part 7 Inter-frame difference detection part 8 Pattern processing part 9 Lane change judgment part

Claims (2)

(57) [Claims] 1. A lane change detecting device for detecting a lane change on a road divided into a first lane and a second lane, wherein the lane change detecting device displays the lane change based on individual frames taken successively in time with respect to the road. about display image, the first lane side area a, area B the center of the first lane and the reference line for dividing the second lane, the area of the second lane side C Division means for dividing the current frame and the previous frame, and inter-frame difference detection means for binarizing the difference result; and for each scanning line forming the display image, The difference result binarized by the inter-frame difference detection means is the area A
A pattern processing unit for performing pattern processing in which “0” is set in the area C, “1” is set in the area B, and “0” is set in the other scanning lines. A lane change detecting device , comprising: a lane change determining unit that calculates a total sum of pattern processing results and determines that the lane change is performed when the value is equal to or greater than a predetermined threshold value. 2. A lane change detection method for detecting a lane change on a road divided into a first lane and a second lane, wherein the lane change is detected based on individual frames temporally continuously taken on the road. about display image, the first lane side area a, area B the center of the first lane and the reference line for dividing the second lane, the area of the second lane side C A first step of dividing the current frame and the previous frame into two areas, a second step of obtaining a difference between the current frame and the previous frame, and binarizing the difference result; and a scanning line forming the display image. The scanning line whose difference result binarized in the second step is “0” in the areas A and C and “1” in the area B is “1”, and the other scanning lines are “0”. A third step of performing a pattern process, A fourth step of taking a total of the results of the pattern processing in the third step, and determining that the lane change has been made when the value is equal to or greater than a predetermined threshold value. Method.