KR20150002039A - Real-Time Lane Detection Method Using Black Box Video - Google Patents

Abstract

The present invention relates to a method for securing real-time properties of a device for recognizing, warning, and recording a traffic lane in an image using an image terminal such as a black box installed in a vehicle. The method for securing real-time properties of a device for recognizing, warning, and recording a traffic lane in an image using an image terminal such as a black box installed in a vehicle can recognize a traffic lane in real time using a black box image to improve the accuracy of recognition of the traffic lane even when the traffic lane is not recognized well and thus not detected during the driving of a vehicle.

Description

Technical Field [0001] The present invention relates to a real-time lane detection method using black box video, and more particularly, to a real-time lane detection method using a video terminal such as a black box installed in a vehicle,

The present invention relates to a method for securing real-time performance of an apparatus for recognizing and warning lanes in an image using a video terminal such as a black box installed in a vehicle, and more particularly, The method comprising:

The lane departure warning system and the lane keeping support system detect the image of the road ahead from the camera attached to the vehicle and confirm the currently running lane and then leave the lane by carelessness of the driver It is a safety device that makes an alarm sound or automatically moves the vehicle into the lane.

However, these safety devices are often mistaken for lane information in bad conditions such as shadows, backlighting, and tunnels. As described above, there are many lane recognition errors in the bad condition, so that the reliability of the lane recognition is low, and furthermore, the reliability of the safety device based on the lane recognition is lowered.

SUMMARY OF THE INVENTION The present invention has been made in view of the above problems, and it is an object of the present invention to provide a video terminal such as a black box installed in a vehicle for improving the accuracy of lane recognition even when a lane is not recognized well during driving, And to provide a method for securing the real-time performance of a device for warning, recording, and recognizing a lane in a video using the method of the present invention.

In order to achieve the above object, there is provided a method for securing real-time performance of a device for recognizing and recording a lane in a video using a video terminal such as a black box installed in a vehicle according to an aspect of the present invention, A method of identifying a lane by using only one lane in a method of recognizing a lane and identifying information of a lane; A method of reconstructing the lane recognition rate deterioration caused by the lane or the external noise using the previously recognized lane information; A method of recognizing a lane with respect to an inferred portion without performing a lane search for all images in consideration of the predicted lane; It includes a lane identification method which is healthy even in the road environment (weather, rain, sun, etc.) and is robust to the surrounding disturbance (object, vehicle, light).

According to the present invention, it is possible to improve the accuracy of lane recognition and to operate safely in conjunction with simple devices such as navigation and mobile applications.

Fig. 1 is a diagram for explaining a lane-finding error.
FIG. 2 is a flowchart illustrating a method for securing real-time performance of a device for recognizing and recording a lane in an image using a video terminal such as a black box installed in a vehicle according to an embodiment of the present invention.
Fig. 3 is a diagram for explaining correct lane recognition. Fig.

BRIEF DESCRIPTION OF THE DRAWINGS The advantages and features of the present invention and the manner of achieving them will become apparent with reference to the embodiments described in detail below with reference to the accompanying drawings. The present invention may, however, be embodied in many different forms and should not be construed as being limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art. And is intended to enable a person skilled in the art to readily understand the scope of the invention, and the invention is defined by the claims. It is to be understood that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. In the present specification, the singular form includes plural forms unless otherwise specified in the specification. It is noted that " comprises, " or "comprising," as used herein, means the presence or absence of one or more other components, steps, operations, and / Do not exclude the addition.

Hereinafter, with reference to FIGS. 1 to 3, a method for securing real-time performance of a device for recognizing and recording a lane in an image using a video terminal such as a black box installed in a vehicle according to an embodiment of the present invention Explain. FIG. 1 is a view for explaining a lane-finding error; FIG. 2 is a diagram showing a real-time characteristic of a device for recognizing and recording a lane in a video using a video terminal such as a black box installed in the vehicle according to an embodiment of the present invention; Fig. 3 is a diagram for explaining correct lane recognition. Fig.

As shown in FIG. 1, a method for securing the real-time performance of a device for recognizing and recording a lane in a video using a video terminal such as a black box installed in a vehicle of the present invention is as follows. First, In order to increase the lane recognition rate after receiving the input, the lane recognition rate can be increased by fixing the area of interest to the lane. Conventional lane recognition methods detect lanes for each region by dividing the ROI (Region of Interest) ROI into 4 ~ 6 pixels horizontally.

However, it is very inefficient to detect the lane by dividing the black box image so that the lane is detected with the part under the fixed ROI ROI.

In the present invention, 640 * 320 is used for the black box image, and ROI is selected from a range of 120 to 220 from a height of 120 to a width of 0 to a fixed ROI, The front part of the vehicle considering the image was removed.

The reason is as follows.

In the case of the black box image, it is inefficient to divide the whole image into 4 to 6 parts because the front part of the car is caught and the area where the lane is visible is very small.

Accordingly, only one portion of the first and the end of the lane is designated as an area of interest (S101).

The second is to transform the image into a gray image before extracting the edge of the ROI region, and then use the histogram stretching to uniformly distribute the brightness values and to use the binarization threshold in all images as a fixed value. Before binarizing a histogram-stained image directly, a lane filter is applied to enhance lane recognition and lane emphasis and noise removal. The lane filter is a region having a size similar to the width of the filter, and the larger the difference from the surrounding pixels is, the higher the effect is.

This makes it possible to effectively detect the area corresponding to the lane because the characteristics of the lane that is lighter than the road area and the width of the lane are considered. Then, edges are detected using the Sober operator, which is tougher in the diagonal direction among the edge extraction operators. The Sober operator consists of two masks for detecting x and y directions. The edge is obtained using the y-direction sober mask, and binarization is performed with the binarization threshold value (48) to emphasize the lane to 0 (black) lower than the threshold value and 255 (white) to the higher region (S102).

The third is that the polar coordinate system is used instead of the Cartesian coordinate system, which approaches the infinite slope of 0, assuming that the extracted image finds a straight line in the image. And the straight lines are detected using the Hough transform (S103).

Fourthly, even if the lane part is Huff transformed using the ROI in the Huff conversion, a large number of lines are caught. In particular, even if only the lane is caught, the straight line is drawn infinitely to the beginning and the end of the screen. Here's how to filter these parts:

First, determine the height at which to draw the line.

This is the same as that of drawing only the area of interest. Next, the coordinates of the y-axis are used to obtain the coordinates of the y-axis using the coordinates of the x-axis and the coordinates of the x-axis. This can be implemented using a linear equation that passes through two points. Finally, the nearest line among many lines is selected and drawn (S104).

Fifth, groups the straight lines obtained using the Huff filter into one line. This is done by grouping the average values using the coordinates of the lines (S105).

The sixth is to find the vanishing point using the left lane and the right lane found by Hough grouping. The vanishing point refers to the intersection of two lanes and the remaining lanes can be obtained through one point (s, y) on one lane by the assumption that the two lanes are symmetrical (S106).

In the seventh, several lanes are generated even if Hough transform is performed. It is necessary to detect a correct lane, particularly as shown in Fig. 1B. In order to obtain the correct lane, the left and right areas are separated and the weight of each lane is rearranged. The distance between the left lane and the right lane, or the case where the lane is not found, Put numbers and add weights or conditions.

First, if the left lane and the right lane are found, the difference between the left lane and the right lane is used to find a real lane. If only the next lane is found, find the right lane using the position value of the left lane. If you find only the right lane, find the left lane using the position value in the right lane. Finally, when the lane is maintained, the weights are increased so that the weights are reduced each time the lane changes, reducing the error range of the lane.

As shown in Fig. 3B, an optimal lane can be obtained. However, if you want to find the best lane, you have to find the left and right lanes. However, if the line is weak or can not be recognized, if one lane or lane can not be obtained, the lane can be detected by obtaining the candidate lane using the data of the previous frame. By using this, it is possible to continuously recognize the lane without holding the lane. If only one lane is recognized as shown in FIG. 1B, the position of the other rear lane can be obtained using the distance between the end point of one lane and the vanishing point (S107).

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it is to be understood that the invention is not limited to the disclosed embodiments, but, on the contrary, is intended to cover various modifications and equivalent arrangements included within the spirit and scope of the appended claims. Therefore, the scope of the present invention should not be limited by the illustrated embodiments, but should be determined by the scope of the appended claims and equivalents thereof.

Claims (2)

A method of recognizing a lane in image and moving picture information and identifying a lane by using only one lane;
A method of reconstructing the lane recognition rate deterioration caused by the lane or the external noise using the previously recognized lane information;
A method of recognizing a lane with respect to an inferred portion without performing a lane search for all images in consideration of the predicted lane;
Lane recognition method that is healthy even in the road environment (weather, rain, sun, etc.) and is strong in the surrounding disturbance (thing, vehicle, light). The method according to claim 1,
A method that can be used as a device to track the lane by driving the lane departure information to the lane departure information, such as an alarm, MDPS or the like, so that the recognized lane can be traveled, Outputting a warning message, directly controlling it, or recording a current occurrence event on an information terminal such as a black box.

Images