KR100383538B1 - Method and system for warning departure of traffic lane - Google Patents

Abstract

The present invention is to alert the driver immediately before the actual lane departure in order to prevent the safety accident that may be caused in the unconscious situation of the driver during the driving unconscious situation while driving the driver's alert to prevent unwanted lane departure A lane departure warning method and system are provided.

Meanwhile, the present invention recognizes lane markings on the road through image processing to determine the degree of departure of the vehicle from the center of the lane, predicts the possibility of lane departure, and measures vehicle driving information such as vehicle speed, steering angle, direction indication, and wiper operation. The present invention relates to a lane departure warning method and system that does not generate an alarm sound when the driver's willingness to change lanes is confirmed even if the possibility of lane departure is predicted by image processing by grasping whether there is a willingness to change lanes.

Description

Lane departure warning method and system {Method and system for warning departure of traffic lane}

The present invention relates to a lane departure warning method and system, and more particularly, to alert the driver immediately before the actual lane departure for the prevention of safety accidents that may be caused in the driver's unconscious situation during an unconscious situation. The present invention relates to a lane departure warning method and system that prevents an unwanted lane departure by inducing a driver's alert by exporting.

Meanwhile, the present invention recognizes lane markings on the road through image processing to determine the degree of departure of the vehicle from the center of the lane, predicts the possibility of lane departure, and measures vehicle driving information such as vehicle speed, steering angle, direction indication, and wiper operation. The present invention relates to a lane departure warning method and system that does not generate an alarm sound when the driver's willingness to change lanes is confirmed even if the possibility of lane departure is predicted by image processing by grasping whether there is a willingness to change lanes.

In general, the driver is required to drive while paying attention so that the vehicle does not deviate from the center of the lane while driving the vehicle, that is, to keep the driving lane without leaving the lane. The reason why the driver continues to pay attention so that the vehicle does not deviate from the center of the lane during driving is, in fact, to prevent traffic accidents that may occur during lane departure, and not to disturb the traffic flow.

However, while the driver is driving, the driver must drive while keeping the driving lane without deviating lanes, but for example, when the driver is drowsy because of physical or mental fatigue, the driver may leave the driving lane and cause an accident. Increasing cases are becoming traffic problems and social issues.

As described above, devices, systems, and methods for preventing inadvertent accidents due to lane departure due to drowsy driving are known in various ways, but they have the following problems.

That is, the conventional embodiment of the lane departure prevention system detects a situation such as drowsy driving, rather than recognizing lane departure itself, and may not recognize the lane departure situation due to reasons other than drowsy driving.

In addition, another embodiment of the conventional lane departure prevention system has a problem in that the driver's will, for example, do not grasp the lane departure situation for changing lanes, causing unnecessary alarms.

In the conventional lane departure prevention system, there is a problem that the lane recognition is not a real lane departure prevention system because it recognizes the position of the lane, that is, generates an alarm in a situation in which the lane departure has already been made.

The present invention has been made to solve the above problems, by sending an alarm to the driver immediately before the actual lane departure for the prevention of safety accidents that may be caused in the driver's unconscious situation during the driving uncomfortable situation during driving It is an object of the present invention to provide a lane departure warning method and system that can induce a driver's alert to prevent an unwanted lane departure.

Another object of the present invention is to recognize lane markers on the road through image processing, to detect the degree of vehicle departure from the center of the lane, to predict the possibility of lane departure, and to measure vehicle driving information such as vehicle speed, steering angle, direction indication, and wiper operation. By identifying whether the driver is willing to change lanes and providing a lane departure warning method and system that does not generate an alarm sound when the driver's willingness to change lanes is confirmed even if the possibility of lane departure is predicted by image processing, have.

1 is a flowchart of a lane departure warning method according to the present invention;

Figure 2 is a block diagram of a lane departure warning system according to the present invention.

Figure 3 is an application example of the lane departure warning system of the present invention.

4 is a flowchart of an image processing algorithm of the lane departure warning method of the present invention.

5 is a graph related to image processing region setting of the present invention.

6 is a processing block diagram for edge extraction in the image processing of the present invention.

7A and 7B are conceptual diagrams related to a Sobel edge operator for the present invention.

8 is a conceptual diagram of a parameter space for constructing the present invention cumulative distribution function.

Figure 9 is a graph of the cumulative distribution function for the present invention.

10 is an experimental example of the present invention.

<Description of the code | symbol about the principal part of drawing>

1: CCD camera 10: control part

11: timing generator 13: processor

15: interface unit 17: alarm determination unit

21: vehicle speed sensor 22: steering angle sensor

23: brake sensor 24: direction switch

25: Wiper Switch 26: System On / Off Switch

30: buzzer

Lane departure warning method according to the present invention to achieve the above object,

Capturing an image of a road being driven through a photographing means such as a CCD camera;

If the vehicle speed is less than a predetermined vehicle speed, the data processing of the photographed image is not performed. If the vehicle speed is more than the predetermined vehicle speed, the image is acquired through the processing of the captured image data, and an edge is detected to build a lane model of the road. Image processing step;

Determining whether the image by the image processing step is a lane-recognizable image and notifying the driver of the lane-recognized image, and determining a lane departure possibility if the image is lane-recognized image;

And a warning step of alerting the driver if there is a lane departure possibility in the determination step, and the determining of the lane departure possibility may include calculating an edge distribution function based on edge information of the lane detected in the image processing step. step; Building an average edge distribution function by accumulating the edge distribution functions calculated for each successive predetermined number of captured image data; Selecting an axis of symmetry of the constructed mean edge inclusion function; It is characterized in that it comprises the step of determining that there is a possibility of lane departure if the axis of symmetry of the selected mean edge inclusion function is more than a predetermined angle relative to 90 °.

In a preferred embodiment of the method of the present invention, whether the lane departure possibility determined in the determination step is determined by the driver's will by analyzing the vehicle signal such as a direction indicator, a steering angle, a brake signal, and whether the wiper is operated or not. The driver further includes a driver's will analysis step of not performing the alarm step when the possibility is at the driver's will.

Lane departure warning system according to the present invention for achieving the above object,

Road image photographing means for photographing an image of a road being driven;

Signal generating means for generating signals relating to vehicle speed and steering angle, brake, direction indication, wiper operation and on / off of the system;

A timing generator for generating a driving timing signal of the photographing means, a processor for predicting lane departure possibility by receiving an image captured by the photographing means, and receiving a signal from the signal generating means to interface the processor to the processor; A control means having an interface unit, and an alarm determining unit configured to determine whether an alarm is generated according to the possibility of lane departure under the control of the process unit;

And alarm means for generating an alarm in accordance with an output signal of the alarm decision unit of the control means, wherein the processor extracts an edge of a lane for each consecutive number of photographed image data photographed by the road image photographing means. The edge distribution function is calculated based on the extracted edge information of the lane, and the cumulative edge distribution function is accumulated to build an average edge distribution function, and the symmetry axis of the constructed average edge distribution function is selected to calculate the average edge distribution function. It is characterized in that it is determined that there is a possibility of lane departure when the axis of symmetry deviates by more than a predetermined angle with respect to 90 °.

In a preferred embodiment of the system of the present invention, the signal generating means is

Vehicle speed sensor, steering angle sensor, brake sensor, direction indicating switch, wiper switch, and system on / off switch.

In a preferred embodiment of the system of the present invention, the control means controls not to generate an alarm when a signal based on the driver's lane change intention is output from the signal generating means even when a possibility of lane departure of the vehicle is predicted.

Hereinafter, exemplary embodiments of a lane departure warning method and system according to the present invention will be described in detail with reference to the accompanying drawings.

1 is a flowchart of a lane departure warning method according to the present invention, Figure 2 is a block diagram of a lane departure warning system according to the present invention.

First, referring to FIGS. 1 and 2, a lane departure warning method according to the present invention includes: photographing an image of a road driving through a CCD camera 1 (S10); If the vehicle speed is less than a predetermined vehicle speed, for example, 20 km / h, data processing of the photographed image is not performed. If the vehicle speed is more than the predetermined vehicle speed, an image is acquired through processing of the captured image data, and a lane edge is detected. An image processing step (S20) (S30) of constructing a lane model of the road; Determining whether the image at step S30 is a lane-recognized image and notifying the driver of the lane-recognized image (S50), and determining a lane departure possibility if the image is lane-recognized image (S40); If there is a lane departure possibility in step S40 is configured to include an alarm step (S60) to alert the driver.

The determination step (S40) is a step of determining whether the lane can be recognized image and predict the lane departure possibility (S41 (S41a)), the step of determining whether there is a lane departure possibility (S42), and lane departure possibility driver Determining whether it is by the will of the vehicle, such as the direction indicator, steering angle, brake signal, whether the wiper operation, that is, the step of determining the direction indicator (operation) operation state (S43), and the steering angle is a certain angle (E.g., 10 DEG or more) and determining (S44), determining whether the brake is operated (S45), and determining whether the wiper is operating (S45).

The alarm step (S60) generates an alarm in two stages according to lane departure information. For example, when the lane departure degree is small, a beep sound of 1 Hz period is generated, and when the lane departure degree is large, a 2 Hz period is generated. Generates a beep sound.

2 and 3, the lane departure prevention system of the present invention includes: a CCD camera (1) for photographing an image of a driving road; A signal generator 20 for generating signals related to vehicle speed, steering angle, brake, direction instruction, wiper operation, and system ON / OFF; A timing generator 11 for generating a driving timing signal of the CCD camera 1, a processor 13 for receiving an image captured by the CCD camera 1 and predicting a lane departure possibility, and a signal generator 20 Interface unit 15 for receiving a signal from the interface 13 to interface with the process unit 13, and an alarm determination unit 17 for determining whether or not an alarm according to the possibility of lane departure under the control of the process unit 13; Control unit 10; And an alarm unit 30 for generating an alarm in accordance with the output signal of the alarm determination unit 17 of the control unit 10. The signal generator 20 includes a vehicle speed sensor 21, a steering angle sensor 22, a brake sensor 23, a direction indicating switch 24, a wiper switch 25, and a system on / off switch 26. Is made, the alarm unit 30 is provided with a buzzer or a speaker.

In the lane departure warning system according to the present invention, the control unit 10 outputs a signal from the signal generation unit 20 according to the driver's lane change intention or a signal for canceling the lane departure warning even when a possibility of lane departure of the vehicle is predicted. Control not to generate an alarm.

It will be apparent to those skilled in the art that the control unit 10 may store the flowchart algorithm of FIG. 1 and the image processing algorithm described below.

Referring to the operation and operation of the lane departure warning method and system of the present invention configured as described above are as follows.

First, according to the present invention, the generation of a beep sound is generated immediately before the lane departure of the own vehicle, even if the lane departure situation is made by the driver's will, even if it is just before the lane departure of the own vehicle. When the vehicle speed is below a certain speed, for example 20 km / h or less, the steering angle of the own vehicle is above a predetermined angle, for example 10 ° or more, the direction indicator is in operation, a brake signal is detected or the wiper is activated. There is no alarm sound.

In the present invention, when the lane recognition is difficult, for example, when the lane is not visible even with the naked eye (road wear, construction section, etc.), and the operation of the system on / off switch 26, It will not work when the system is off. Operation of the system on / off switch 26 is at the driver's choice.

Determination of lane departure possibility according to the present invention is achieved through lane mark recognition by image processing, and the CCD camera 1 for this process preferably processes approximately 15 frames or more per second. The CCD camera 1 maintains 0 ° in all of the roll, tilt, and yaw angles, and is preferably installed at the center of the vehicle 0BL, that is, in the rear of the room mirror (not shown). The determination of the driver's will and the determination of the occurrence of the alarm are achieved by the interlocking of the image processing result and the vehicle driving signal, and the processing speed for this is preferably about 10 Hz.

The CCD camera 1 captures the front road image while driving as a single monochrome image. The image processing software (content of which the flowchart algorithm of FIG. 1 is programmed) processed by the processor 13 of the controller 10 uses the 3 × 3 Sobel edge operator to draw the edge of the lane marker. Detect edges, build an edge distribution function (EDF), calculate scene analysis indexes by EDF, determine scenes that are difficult to process, and search for symmetry axes and local maxima of EDF. We predict the lane departure possibility by using the position of the symmetry axis of the EDF.

The controller 10 performs image acquisition (Raw Image Grabbing) using a sync generator and an analog-to-digital converter (ADC), operates image processing software for real-time processing, and operates the interface unit 15. Receives vehicle signals such as vehicle speed and steering angle.

The alarm determiner 17 of the controller 10 not only selects an alarm mode in two stages (Light / Heavy Warning) but also predicts a direction of heading (DOH) using vehicle information (vehicle speed, steering angle, etc.). do.

Looking at the above in more detail as follows.

Logic for the present invention consists of two main. In other words, the driver can change lanes by recognizing lane markings on the road by image processing technology to determine the extent of the vehicle's departure from the center of the lane, predicting the possibility of lane departure, and measuring vehicle driving information such as speed, steering angle, and direction indication. It consists of figuring out whether you are willing to do this. Therefore, even if the possibility of lane departure is predicted by the image processing logic, if the driver's willingness to change lanes is confirmed, no alarm sound is generated.

In the image processing algorithm according to the present invention, a lane is recognized from a road image input from the CCD camera 1 to determine whether the own vehicle may leave the lane. According to the known technology at home and abroad, lane recognition is regarded as immediately recognizing the position of the lane, but the lane recognition method of the present invention focuses on the direction recognition of the lane rather than the position recognition of the lane. An example of a flowchart of an image processing algorithm is shown in FIG. 4.

Looking at the process of judging lane departure possibility by the controller 10, since the lanes are parallel to each other, in the input image photographed while the own vehicle is traveling in the center of the lane, the direction of the left and right lanes is symmetric about 90 °, but the vehicle is lane Once out of the center the symmetry of the detected lane direction is no longer established about 90 ° and the center of symmetry deviates from 90 °. Use this fact to determine the lane departure of your vehicle.

Lane recognition is performed by the CCD camera 1, wherein the image by the CCD camera 1 is a black and white image, and the frame size is approximately 160 (horizontal) x 120 (vertical) pixels.

Looking at the setting of the image processing area, assuming that the optical axis of the CCD camera 1 installed in the vehicle coincides with 0BL and is parallel to the flat road surface, that is, the yaw angle, the tilt angle, and the roll angle If all of these are 0 °, the vanishing point of the lane visible in the image will be at the center of the image, and the lane will actually appear only below the center of the image. In this situation, the image processing area for lane recognition is limited only to the lower center of the image as shown in FIG. 5. When the camera is installed in a vehicle to slightly expand the lane identification area, it is generally installed with a slight tilt angle so that the optical axis of the camera is in contact with the road surface, so the actual vanishing point is slightly above the center point of the image. To set the processing area in the same manner as in FIG.

Looking at the basic preconditions of the image processing algorithm, the image processing algorithm of the present invention was made under the following three conditions, each of which has sufficient rationality when compared with the actual highway environment.

(1) The direction of the lane does not change rapidly and has continuity.

(2) Lane markers have a brighter brightness value than the background, and the boundary is located at the division of two areas with different brightness.

(3) Because the lanes are parallel to each other, the direction of left and right lanes in the input image is symmetric about 90 ° when the vehicle travels in the center of the lane.

As for the edge extraction of the lane, according to the second precondition of the image processing algorithm, the boundary of the lane marker has a thick edge component which is larger than the other parts of the image. If used effectively, good lane recognition results can be achieved. Accordingly, in the present invention, edge extraction is performed in the following manner, which is well known in the field of computer vision, considering the raw information of an image called an edge as an important factor. The content related to the edge extraction is shown in FIG. 6 as a block.

(1) Calculation of gradient of brightness value by differential operation

(2) Calculation of the magnitude and direction of the calculated slope

(3) If possible, remove small edge pixels with no distinct edge components to leave only the edges corresponding to the lane markers.

In order to calculate the slope, edges are extracted by differential operation on the input image. Using the brightness value of each pixel in the 3 × 3 area as shown in FIG. 7A and the Sobel edge operator given in FIG. 7B, the slope component of the edge at the center position of the area is calculated as follows. 1) (2) is shown.

In the physical quantity of the edge, the size of the edge is calculated by the formula (3), and the direction of the edge is calculated by the formula (4). Edge orientation in logic implementation This value is 0 ° to 180 °.

Threshold updating for removal of small edge pixels;

In order to highlight the edge pixels related to the lane as much as possible, small edge pixels should be removed, which is one of the dilemma in computer vision. The present invention proposes a threshold updating method in which an algorithm can adapt itself to a change in brightness of an input image without user intervention. A square of a certain size in the bottom center of the image for each input image frame Next, the mean and standard deviation of the size of the edge pixels in this area are calculated by the formula (5) (6). In formula (5) (6) Represents the area of the selected rectangular area. Predicted threshold for the first frame entered Is set to the sum of [mu] and [sigma] of the formula (5) (6), and then in the input k-th frame, [mu], [sigma], and [tau] are updated by the exponential weighted average as in the formula (7). In formula (7), λ uses a value that is experimentally set to 0.6, for example. Pixels with smaller edge sizes obtained as above are excluded from consideration.

The construction of edge distribution function (EDF) is as follows.

Edge Distribution Function (EDF) as a function of edge direction;

The edge distribution function is formed by removing edge pixels smaller than the threshold of formula (7) and accumulating the edge size of formula (3) with respect to the direction of formula (4) with the remaining edge pixels. Edge direction of formula (4) to make this function discrete Calculate in 1 ° increments from 0 ° to 180 °. Distribution function Is expressed as the formula (8). In formula (8) d denotes the angle of the edge, n (d) The number of edge pixels with

Similar to the idea of the Hough Transform, a one-dimensional look-up table with 181 cells called parameter spaces is constructed to construct the actual distribution function as shown in FIG. 8. The initial value of each cell in the space is zero. For each remaining edge pixel removed, find the cell corresponding to the edge direction value of the pixel, read the current value of the cell, and add the edge size value of the pixel to this value to add the edge distribution Is built. The EDF constructed as described above has a shape as shown in FIG. 9.

EDF shown in Figure 9 shows the following two features.

1) It has a large value near the angle corresponding to the lane direction. This is because the lanes have continuity in the direction and the boundary of the lane markers has a large edge pixel. The edges other than the lane boundaries do not guarantee continuity in any particular direction, so the contribution to the specific position of the EDF falls.

2) It has symmetry around 90 °. Because the lanes are parallel, the EDF always maintains symmetry around 90 ° when the vehicle travels in the center of the lane. This fact is later applied to the alarm determination unit 17 of the lane departure control unit 10 as it is. That is, if the vehicle is out of the lane center while driving, the symmetry axis of the EDF is also out of 90 °.

Mean equity coverage;

Since the edge distribution function of Equation (8) is obtained from a single image, the lane markers are not easily revealed due to noise in the noise-inherent image, and as a result, the value of the function is not prominent near the angle corresponding to the lane direction. Can be. Therefore, if the function of formula (8) is accumulated again according to the principle of moving average for N consecutive images, the effect of noise is almost disappeared and the direction of lane is more clearly shown. In this concept, we build the mean distribution function as shown in equation (9). In equation (9), k represents the current frame. When realizing H (d) in a real system, the purpose is to eliminate the effects of noise. Therefore, it does not take a moving average but takes the form of a moving sum as follows.

Looking at the symmetry of the edge distribution function of the present invention.

Any function Is an even function And exponential Can be expressed as the sum of and each is represented by the formula (10). In formula (10) Is a function Is the interval to be defined. In particular, the right function has axis symmetry, and the odd function has a characteristic that describes the symmetry of the origin well.

Based on these mathematical backgrounds, an assessment of how well a vehicle is running through the center of the road is made by calculating how well the EDF is dominated by the EDF near 90 °.

Symmetry axis search interval setting;

The symmetry axis of the EDF appears near 90 ° when the vehicle travels in the center of the road, but as the driving position is shifted from the center, that is, when the wheel of the vehicle is closer to either side of the left and right lanes, the position of the symmetry axis also deviates from 90 °. The symmetry axis search section β is set experimentally as follows to find out its position even in lane change.

That is, β: 90 ° ± 40 ° (50 ° ≦ β ≦ 130 °).

An evaluation interval of a function for searching the position of the axis of symmetry;

Although the range of EDF is defined between 0 ° and 180 °, the driving lane direction of the own vehicle seen in the actual highway image is less than 15 ° but does not appear above 165 °. Therefore, the evaluation interval of the EDF for searching the position of the symmetry axis Is set to 150 ° between 15 ° and 165 °. sure May be set to 180 °, but it is not efficient because it takes more calculation time.

Symmetry Measure;

First, the right function and the odd function of EDF are represented by the formula (11). Substitution variable in formula (11) And Is an arbitrary assumed axis of symmetry. Any axis of symmetry Symmetry measure to assess how much symmetry a has Is defined as in formula (12). In formula (12), Integral is the calculation of the energy of each function.

Selection of the axis of symmetry;

Symmetry measure computed within the symmetry search interval β With the highest value Is chosen as the axis of symmetry of the EDF.

The interpretation of the input image based on EDF is as follows.

The road environment may be difficult to identify lanes even by human eyes due to various noise factors. For example, when the road is badly worn, when rainfall, toll gate entry and exit, when the lane is blocked by the front vehicle, or when the roughness suddenly occurs severely at the entrance and exit of the tunnel, etc. to the CCD camera 1 and the control unit 10 This can be seen as a difficult lane identification. When the situation occurs in a system that takes action based on the lane identification information, the reliability of the actuating result based on the lane information obtained without knowing such a situation cannot be guaranteed. Therefore, if the above situation can be detected, the error of operating the system with the error information can be avoided. To this end, in the present system, as shown in FIG. And standard deviation Calculate And the ratio of these two statistics Take a look at whether or not the outline of the lane is well revealed in the input image. Is called the scene interpretation index. In the case of the lane outline image, the edge pixels obtained in the lane boundary have a large value as shown in (b) of FIG. 9. Thus the mean of the function Standard deviation Has a large value and as a result The value becomes smaller. Here, "in the image where the lane outline is well revealed, "The value is small" is always true, but not vice versa. Small does not always show the lane outline well in the image. Scene Interpretation Index Is a necessary condition for the interpretation of the input image but is not sufficient. Therefore, the algorithm of the present invention has a limitation in the interpretation of the input image, which can be overcome by a later invention.

10 is shown with reference to the experimental example of the present invention system.

Looking at the alarm algorithm achieved by the present invention alarm determination unit 17 is as follows.

As a premise of an alarm, prediction of a lane departure possibility;

Symmetry axis of EDF If is far away from 90 °, the lane deviation is calculated as follows to assume that the vehicle is out of the lane center while driving.

Deviation amount:

In the above, for example If there is a possibility of lane departure, In the present invention Although 7 degrees was set, the specific value can be determined suitably by actual vehicle tuning.

When the lane departure amount satisfies the alarm condition, as shown in FIG. 1, it is determined by detecting a direction indication, a steering angle, and the like to determine whether the driver has an intention to change lanes before the alarm is issued. That is, when the direction indicator flashes or when the steering angle is ± 10 ° or more, it is determined that the driver intends to consciously change lanes. If the vehicle is leaving the lane in the unconscious state of the driver, an alarm is issued.

Even if the alarm condition is satisfied, the alarm determining unit 17 of the control unit 10 does not emit an alarm when the running speed of the vehicle is approximately 20 km / h. However, the vehicle speed may be preferably determined through actual experiments. .

And even if the alarm condition is satisfied, the alarm determining unit 17 of the control unit 10 does not emit an alarm when there is a rainfall. It is as described above that the judgment of this situation determines whether or not the wiper is operating. The reason for not warning when the wiper is operating is because it is difficult to secure the reliability of the recognized lane information in rainy weather, and if the driver is not drunk, the driver will not drive unconsciously, such as drowsy driving.

The alarm determining unit 17 generates an alarm in two stages through the buzzer 30.

Light warning occurs when Although the period of the light alarm sound was preferably set to 1 Hz, it is of course possible to newly determine through tuning the vehicle.

When is set, a heavy warning is issued. The period of the strong alarm sound is preferably 2 Hz. Although the lane departure amount is set to 10 ° in the above, this value can of course be newly determined through actual vehicle tuning.

As described above, according to the present invention, the driver's alert is raised by sending an alarm to the driver immediately before the actual lane departure in order to prevent safety accidents that may be caused when the driver undesired from the driving lane. To prevent unwanted lane departures.

Also, the present invention recognizes lane markings on the road through image processing, detects the degree of departure of the vehicle from the center of the lane, predicts the possibility of lane departure, measures vehicle driving information such as vehicle speed, steering angle, direction indication, and wiper operation. If the driver's willingness to change lanes is determined by the image processing, and the driver's willingness to change lanes is confirmed by the image processing, the driver will not sound an alarm to prevent lane departures. It offers the advantage of improving sex.

Although preferred embodiments of the present invention have been described in detail above, those of ordinary skill in the art will appreciate that the present invention may be made without departing from the spirit and scope of the invention as defined in the appended claims. It will be appreciated that various modifications or changes can be made. Therefore, changes in the future embodiments of the present invention will not be able to escape the technology of the present invention.

Claims (5)

In the lane departure warning method, Photographing an image of a road running through the photographing means; If the vehicle speed is less than a predetermined vehicle speed, the data processing of the photographed image is not performed. If the vehicle speed is more than the predetermined vehicle speed, the image is acquired through the processing of the captured image data, and an edge is detected to build a lane model of the road. Image processing step; Determining whether the image by the image processing step is a lane-recognizable image and notifying the driver of the lane-recognized image, and determining a lane departure possibility if the image is lane-recognized image; And a warning step of alerting the driver if there is a possibility of lane departure in the determining step, The determining of the lane departure possibility, Calculating an edge distribution function based on the edge information of the lane detected in the image processing step; Building an average edge distribution function by accumulating the edge distribution functions calculated for each successive predetermined number of captured image data; Selecting an axis of symmetry of the constructed mean edge inclusion function; And determining that there is a possibility of lane departure when the axis of symmetry of the selected average edge distribution function is out of a predetermined angle based on 90 °. The method of claim 1, wherein the possibility of lane departure is determined by analyzing a vehicle signal such as a direction indicator, a steering angle, a brake signal, and a wiper operation to determine whether the lane departure possibility determined in the determination step is at the driver's will. The lane departure warning method further comprises a driver's will analysis step not to perform the alarm step when the will. In the lane departure warning system, Road image photographing means for photographing an image of a road being driven; Signal generating means for generating signals relating to vehicle speed and steering angle, brake, direction indication, wiper operation and on / off of the system; A timing generator for generating a driving timing signal of the photographing means, a processor for predicting lane departure possibility by receiving an image captured by the photographing means, and receiving a signal from the signal generating means to interface the processor to the processor; A control unit having an interface unit configured to perform an alarm, and an alarm determiner configured to determine whether an alarm is generated according to the possibility of lane departure under the control of the process unit; And Alarm means for generating an alarm in accordance with the output signal of the alarm determination unit of the control means, The process unit, The edges of the lanes are respectively extracted for a predetermined number of consecutive image data captured by the road image photographing means, and the edge distribution functions are calculated based on the extracted edge information of the lanes. Constructing the symmetry axis of the average mean inclusion function, and determining that there is a possibility of lane departure if the symmetry axis of the selected average edge inclusion function deviates from the predetermined angle by more than 90 °. Lane departure alarm system. The method of claim 3, wherein the signal generating means, And a vehicle speed sensor, a steering angle sensor, a brake sensor, a direction indicating switch, a wiper switch, and a system on / off switch. 4. The lane departure warning according to claim 3, wherein the control means controls not to generate an alarm when a signal from the signal generation means is output by the driver's lane change intention even when a possibility of lane departure of the vehicle is predicted. system.