KR20010059100A - Method for discriminating of a traffic lane vedio signal - Google Patents

Abstract

PURPOSE: A lane image signal discriminating method is provided to check if the state of a lane is good or bad by detecting lane image signals while traveling. CONSTITUTION: A system photographs forward, left, and right lanes while traveling(S20) and converts the photographed lanes to image signals(S21). Then, the system detects the size and direction of each lane edge from the image signals(S22). The system obtains an average rate and a dispersion rate between a lane detecting zone and a lane non-detecting zone(S24) after adding the edge sizes in the same direction(S23). When the average rate and the dispersion rate are negative, the system decides whether the state of the lane is impossible to be recognized or not(S25).

Description

Lane image signal discrimination method {METHOD FOR DISCRIMINATING OF A TRAFFIC LANE VEDIO SIGNAL}

The present invention relates to a method for discriminating a lane image signal, and more particularly, to detect a good and bad state of a lane according to the direction of a detected lane image signal when detecting a lane image signal on an automobile road. The present invention relates to a video signal discrimination method.

Recently, due to the development of the electronic imaging system, various multimedia and even vehicles have been introduced, and many safety devices using various electronic systems have been applied to vehicles.

For example, there is a navigation system that allows the driver to easily grasp the efficient traffic control and traffic conditions, and to prevent the collision of the vehicle by controlling the alarm and speed to prevent the collision of the vehicle, and to maintain the proper distance between the vehicles to prevent the driver's safety. Electronic system, lane departure prevention system to prevent accidents caused by lane departure due to drowsiness while driving.

In particular, the lane departure prevention system, unlike other electronic systems, is a safety system for preventing large accidents caused by lane departure (center line invasion) from the moment of inattention or tiredness of the driver while driving.

The lane departure prevention system includes a camera 10 mounted at a predetermined position in front of the vehicle and photographing front left and right lanes while driving, as shown in FIG. 1; An image signal processor 11 for processing the lane photographed by the camera 10 as an image signal; A controller 12 for detecting lane presence and determining lane departure from a lane processed as an image signal by the image signal processor 11; By configuring the alarm unit 13 to warn the departure of the lane, it is possible to prevent as much as possible in advance, large and large accidents caused by the lane departure that may occur during driving.

However, in the lane departure prevention system, there are various methods for detecting a lane from a captured lane image signal. However, when the lane is erased due to road conditions, or becomes unrecognizable due to road conditions while driving, the lane may be repackaged. In this case, the lane departure prevention system does not have a function or a method for determining a state in which lane recognition is good or a state in which lane recognition is impossible from the captured lane image signal, and thus the lane detection process is performed from the captured lane image signal. Not only did it take longer, but it also had the problem of giving a false alarm of lane departure.

Accordingly, an object of the present invention is to provide a lane image signal discrimination method capable of detecting a lane image signal when driving a vehicle road and determining whether the detected lane image signal is in a good lane or a poor lane that cannot be recognized.

1 is a block diagram of a conventional lane departure detection control device.

2 is a flowchart illustrating a method for determining a lane image signal according to the present invention.

Figure 3a is a lane edge size and direction video signal waveform diagram applied to the present invention.

3B is a lane edge size, direction average, and dispersion waveform diagram applied to the present invention.

Figure 4 is a video signal waveform diagram when the lane image unattended applied to the present invention.

In order to realize the above object, the present invention includes the steps of photographing the front left and right lanes while driving on an automobile road and processing the photographed lanes as image signals, and detecting the size and direction of each lane edge from the processed lanes; Obtaining an average ratio and a dispersion ratio between the lane detection area and the non-lane detection area by obtaining a sum of magnitudes in the same direction of each lane edge detected in the step; The lane detection area obtained in the above step is characterized in that the average ratio and the dispersion ratio of the lane undetected area is processed to be impossible to recognize the lane when detected as a negative ratio.

Therefore, according to the present invention, a lane edge is detected with respect to a lane image signal detected while driving on a road, and an average ratio and a dispersion ratio of a lane detection region and a non-lane detection region are obtained for the detected lane edge, respectively. When calculated as a ratio, it is possible to recognize that the lane is defective, thereby improving image processing speed and preventing false alarm of lane departure.

Hereinafter, described in detail with reference to the accompanying drawings, preferred embodiments of the present invention.

2 is a flowchart illustrating a method for determining a lane image signal according to the present invention, FIG. 3A is a diagram of a lane edge size and direction image signal waveform applied to the present invention, and FIG. 3B is a lane edge size and direction applied to the present invention. 4 is an image signal waveform diagram when the lane image is unrecognized in the present invention, which includes photographing a front left and right lanes when driving on an automobile road; Processing (21) the lane photographed in the step (20) as an image signal; Detecting (22) the size and direction of each lane edge from the lane processed as an image signal in the step (21); Obtaining (23) a sum of magnitudes in the same direction of each lane edge detected in the step (22); Calculating (24) the lane detection area average ratio and the lane detection area dispersion ratio with respect to the sum of the direction and the size of the same lane obtained in the step (23); Determining (25) which one of the lane detection area average ratio and the lane detection area dispersion ratio respectively obtained in the step 24 is detected as a negative ratio; Processing (26) to make the lane unrecognizable if it is detected at a negative ratio in the step (25); If it is not detected as a negative ratio in the step 25 it is made to the step 27 to recognize the lane good.

According to the present invention made as described above, the front left and right lanes are photographed from the camera 10 mounted at the front predetermined position of the vehicle and input to the image signal processor 11 (step 20).

The image signal processing unit 11 processes the captured and input lane signal into an image signal and inputs it to the control unit 12 (step 21).

The controller 12 detects the size and direction (θ) of each edge with respect to the lane edge as shown in FIG. 3A with respect to the image signal processed by the predetermined program (step 22). The sum of magnitudes in the same direction with respect to [theta] is calculated (step 23).

Subsequently, the lane edge detection area average ratio and the detection area dispersion ratio are calculated for the lane edge size, wherein the lane edge detection area average ratio and the detection area dispersion ratio are calculated as shown in FIG. 3B. Each lane is separated into a lane detection area and a lane non-detection area, and the lane detection and non-detection are separated into lane average and lane dispersion.

The average ratio and the dispersion ratio are calculated in the lane detection and the undetected lanes separated as described above, that is, the average ratio = (average of the lane detection area-average of the lane undetected area) / (average of the lane detection area) + Average of non-lane detection areas), and the dispersion ratio = (variance of lane detection areas minus dispersion of lane detection areas) / (variance of lane detection areas + dispersion of lane non-detection areas) (step 24). .

Subsequently, the controller 12 determines whether either the calculated average ratio or dispersion ratio has a negative ratio or all have negative values as shown in FIG. 4 (step 25).

At this time, when either or both of the calculated average ratio or dispersion ratio have a negative value, the controller 12 determines that the lane recognition state is poor with respect to the currently imaged and input lane image signal. If both the calculated average ratio and the dispersion ratio do not have a negative value, the controller 12 determines that the currently imaged and input lane image signal is a good state signal (step 26).

As described above, the present invention detects a lane edge with respect to a lane image signal detected when driving on a road, obtains a magnitude in the same direction of the detected lane edge, and compares the average ratio between the lane detection area and the undetected area. When the dispersion ratio is obtained and the calculated value is calculated as a negative ratio, it is recognized as a lane defect, thereby improving the image processing speed of the detected lane and providing an effect of preventing a false alarm of lane departure.

Claims (4)

Photographing the front left and right lanes while driving on an automobile road and processing the photographed lanes as image signals, and detecting the size and direction of each lane edge from the processed lanes; Obtaining an average ratio and a dispersion ratio with respect to the size of the lane edge by obtaining a sum of magnitudes in the same direction of each lane edge detected in the step; And processing the lane-recognition impossible when the average ratio and the dispersion ratio obtained in the step are detected as negative ratios. The method of claim 1, wherein the size average ratio and the dispersion ratio of the lane edge are calculated by using the lane detection area and the lane non-detection area. The size average ratio of the lane edge is set to (average of lane detection area minus lane detection area) / (average of lane detection area + average of lane non-detection area) and the size of the edge. The dispersion ratio is calculated by (variance of the lane detection area minus the dispersion of the lane non-detection area) / (variance of the lane detection area + dispersion of the lane non-detection area), respectively. The method of claim 1, wherein the determination of the lane detection impossibility is performed when the non-calculation of any one of an average ratio and a dispersion ratio with respect to the magnitude of the lane edge image signal is performed.