KR101327348B1 - A system for ego-lane detection using defined mask in driving road image - Google Patents

Abstract

The present invention relates to a lane recognition system using a mask defined in a driving road image. The lane recognition system recognizes a lane by matching images obtained through two cameras installed in a vehicle, defining a mask line in the matched image, and sensing the color and shape of the lane according to the mask line; and recognize a lane change by defining a mask block in the matched image and sensing the change of a vehicle coordinate in the mask block. [Reference numerals] (110) First image input unit;(120) Second image input unit;(130) Image matching unit;(140) Digital image conversion unit;(150) Image filter unit;(160) Lane recognizing unit;(162) Self-lane recognition unit;(164) Lane changing recognition unit;(170) Lane information output unit

Description

A vehicle lane recognition system using a mask defined in a driving road image {A SYSTEM FOR EGO-LANE DETECTION USING DEFINED MASK IN DRIVING ROAD IMAGE}

The present invention relates to a magnetic lane recognition system using a mask defined in an image of a driving road, and more particularly, to match an image acquired through two cameras mounted on a vehicle, and to define a mask line on the matched image. A mask defined in the driving road image that recognizes a lane of a lane by detecting the color and shape of the lane along a line, and recognizes a lane change by defining a mask block in a matched image and detecting a change in lane coordinates in the mask block. The present invention relates to a magnetic lane recognition system.

In general, when a moving object moves by tracking a line, an image recognition device or the like should first be used to detect an out of line and to alert the outside.

That is, the recognition apparatus using an image acquires an image of a target to be recognized by a moving object using a camera, and then extracts a feature of the target by using a digital image processing technology, and identifies the target by using the extracted feature. In order for the image recognition apparatus to perform its original function smoothly, it is necessary to be able to accurately extract the target line from the moving object.

Such image recognition device can be widely applied to the fields of automobile, robot technology (RT), automated guided vehicle (AGV), etc. Especially in the automobile field, accurate lane recognition must be ensured even at high speeds. The technical difficulty is high. Here, the lane is a reference line for driving and is a standard for all driving behaviors such as forward, reverse, lane change, course change, forward parking, reverse parking, and row parking.

The conventional lane recognition system is a system for recognizing a lane using a radar or a single camera, and has a characteristic of informing a driver of information of a lane currently being driven by visual or audio elements.

When the driving road is recognized as an image using a camera, an angle of view, which is a sharpness according to the perspective of an object recognized as an image and a viewing angle of the image recognized by the camera, is determined according to a focal length of a lens provided in the camera. That is, if the focal length of the camera is short depending on the lens provided in the camera, the viewing angle of the image taken by the camera, that is, the angle of view, becomes wider, so that the short-distance lane can be widely recognized, while the remote object is clearly recognized. On the contrary, when the focal length of the camera is long, the far point is clearly seen in the image of the driving road, but the angle of view of the image recognized by the camera is reduced, which makes it impossible to recognize all the near lanes. There is this.

For this reason, the camera of the wide-angle lens used in the current black box can obtain images of all lanes on the road because the angle of view is wide, but the closer to the edge of the image, the more severe the distortion, so it is difficult to secure the straightness characteristic of the lane. In addition, due to the reflection of the dashboard on the front glass, it is difficult to obtain a pure road image. In addition, when a lens close to the standard lens is used to acquire an image without distortion, it is difficult to obtain a road image including all lanes except for the driving lane because the angle of view is narrow.

In addition, the existing navigation system does not consider the current lane, but only provides the driver with information about the number of lanes and lanes on the road, and does not provide information about the entire lane of the road currently being driven. There is this.

In order to solve such a problem, there are various conventional techniques for recognizing the proposed lane. For example, Korean Patent Registration No. 10-0975749 (name of the invention: lane recognition using a single lane and lane departure detection method of a vehicle) can reduce the calculation amount and improve the recognition rate even when an obstacle exists. The present invention relates to a lane recognition method using a single lane and a lane departure detection method of a vehicle.

However, the related arts including the patent still have a problem in that they do not recognize the entire lane of the driving road and one of them.

The present invention has been proposed to solve the above problems of the conventionally proposed methods, by taking the image of the front of the vehicle through the two cameras mounted on the vehicle, respectively, and attaching them to the panorama image to connect all the lanes It is an object of the present invention to provide a self-driving recognition system using a mask defined in an image of a driving road that can be recognized.

In addition, the present invention, by defining a mask line in the matched image by detecting the color and shape of the lane according to the mask line to recognize the lanes, and by defining a mask block in the matched image by changing the amount of change in the vehicle coordinates in the mask block An object of the present invention is to provide a self-driving vehicle recognition system using a mask defined in an image of a driving road, which detects a lane change by detecting a driving lane.

Magnetic lane recognition system using a mask defined in the driving road image according to the characteristics of the present invention for achieving the above object,

A photographing image matching unit for matching images of driving roads acquired from two cameras installed at both sides of the vehicle into one image; And

And a lane recognizing unit defining a mask line in the matched driving road image, and detecting a magnetic lane of the driving road based on the mask line.

Preferably, the mask line,

At least one or more spaces may be spaced apart from each other in the image of the driving road in a direction crossing the direction of the lane.

More preferably, the lane recognition unit,

The color information corresponding to the lane is recognized along the mask line, and when color is continuously detected, it is recognized as a straight line, and when broken in the middle, it may be determined as a dotted line lane.

Preferably, the lane recognition unit,

When the driving road is a 3 to 5 lane road, the color information of the pixel corresponding to the mask line is recognized, and when the yellow solid line is on the left side of the vehicle and the white dotted line is on the right side, the lane is recognized as the first lane. If there is a white dotted line on the left side of the vehicle and there is a white solid line or an orange solid line, the lane is recognized as the last lane. If there are dotted lines on both sides of the vehicle, the lane is detected by detecting the number of lanes in the orange solid lane. have.

More preferably, the lane recognition unit,

The total number of lanes of the driving road may be obtained from a global positioning system (GPS).

Preferably, the photographed video matching unit,

Images of the roads obtained from the two cameras are combined in a panorama format and matched to one image, and from the cameras mounted on the right and 100 pixels on the right of the image taken from the camera mounted on the left side of the vehicle, of the two cameras. One image may be generated by matching pixels to the left 100 pixels of the captured image through the Dimension Dynamic Warping algorithm.

More preferably, the lane recognition unit,

The apparatus may further include a lane change recognizing unit that defines a mask block in the matched driving road image, and detects a lane change by detecting a change amount of vehicle coordinates in the mask block.

Even more preferably, the mask block,

The driving road image may be defined as a block for n-dividing the driving road image (n is a natural number) based on the lane closest to the vehicle.

According to the magnetic lane recognition system using a mask defined in the driving road image proposed by the present invention, the two road cameras mounted on the vehicle to take the image of the road ahead of the vehicle, respectively, and attach the panorama image to all lanes It can obtain an image containing the.

In addition, according to the present invention, by defining a mask line in the matched image to detect the lane and the shape of the lane according to the mask line to recognize the lane, and also to define a mask block in the matched image of the lane coordinates in the mask block By detecting the amount of change, it is possible to recognize the lane change and provide the driver with accurate information about the driving lane.

1 is a block diagram schematically illustrating a configuration of a magnetic lane recognition system using a mask defined in a driving road image according to an exemplary embodiment of the present invention.
2 illustrates a portion of a vehicle operation in accordance with one embodiment of the present invention.
3 is a diagram illustrating a part of an operation of a magnetic lane recognition unit of a magnetic lane recognition system using a mask defined in a driving road image according to an exemplary embodiment of the present invention.
4 is a diagram for describing an operation of a magnetic lane recognition unit of a magnetic lane recognition system using a mask defined in a driving road image, according to an exemplary embodiment.
FIG. 5 is a diagram for describing an operation of a lane change recognition unit of a magnetic lane recognition system using a mask defined in a driving road image, according to an exemplary embodiment;
6 is a flowchart illustrating a lane detection method and a lane change recognition method using a mask defined in a driving road image according to an exemplary embodiment of the present invention.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings, in order that those skilled in the art can easily carry out the present invention. In the following detailed description of the preferred embodiments of the present invention, a detailed description of known functions and configurations incorporated herein will be omitted when it may make the subject matter of the present invention rather unclear. In the drawings, like reference numerals are used throughout the drawings.

In addition, in the entire specification, when a part is referred to as being 'connected' to another part, it may be referred to as 'indirectly connected' not only with 'directly connected' . In addition, the term 'comprising' of an element means that the element may further include other elements, not to exclude other elements unless specifically stated otherwise.

1 is a block diagram schematically illustrating a configuration of a magnetic lane recognition system using a mask defined in a driving road image according to an embodiment of the present invention, and FIG. 2 illustrates a part of a vehicle operation according to an embodiment of the present invention. One drawing.

1 and 2, the magnetic lane recognition system 10 using a mask defined in an image of a driving road according to an embodiment of the present invention may include a first image input unit 110 and a second image input unit 120. The image matching unit 130, the digital image converter 140, the image filter unit 150, the lane recognizing unit 160, and the lane information output unit 170 may be configured.

The first image input unit 110 and the second image input unit 120 may be provided in the vehicle 200 to receive an image of a driving road. In particular, the first image input unit 110 and the second image input unit 120 of the magnetic lane recognition system 10 using the mask defined in the driving road image according to an embodiment of the present invention are driving roads of the vehicle 200. For example, as shown in FIG. 2, two cameras 210 and 220 may be installed at both sides in the same direction as the room mirror mounted on the vehicle 200 to photograph a situation.

The image matching unit 130 may receive two images 310 and 320 obtained from the first image input unit 110 and the second image input unit 120 and match them into one image. That is, the first image input unit 110 and the second image input unit 120 may perform a role of collecting the two images 310 and 320 obtained by the divided images as one image.

In detail, the image matching unit 130 searches for an object or a background that is common to the divided images 310 and 320 obtained by the first image input unit 110 and the second image input unit 120, respectively. This can be completely overlapped and combined into one image. For example, the two images 310 and 320 captured by the two cameras illustrated in FIG. 2 may be matched to the image by one driving road as illustrated in FIG. 3 by the image matching unit 130.

The digital image converter 140 may serve to convert an analog image into a digital image. That is, the image matching unit 130 may convert the analog image obtained by collecting the single image into a digital image capable of numerical measurement and control of the image.

The image filter unit 150 filters the digital image acquired by the digital image converter 140 to clearly display the outlines of the driving road, the vehicle, the lane, and the environment around the road by filtering the digital image. It can play a role in distinguishing. Meanwhile, one image matched by the digital image converter 140 may be transmitted to the lane recognizing unit 160 without passing through the image filter unit 150, but is not limited thereto.

The lane recognizing unit 160 may perform a role of distinguishing and recognizing lanes by calculating a geometric feature of a lane that is distinguished from a specific object obtained by the image filter unit 150, that is, another object and a background.

In particular, the lane recognizing unit 160 of the magnetic lane recognition system 10 using the mask defined in the driving road image according to an embodiment of the present invention arbitrarily masks a line in the driving road image obtained from the image filter unit 150. Define a mask block on an image of a driving road acquired from a magnetic lane recognition unit 162 and an image filter unit 150 that recognize a magnetic lane by detecting a color and a shape of a lane according to a mask line, and a mask. It may be configured to include a lane change recognition unit 164 for detecting a lane change by detecting a change amount of the vehicle coordinates in the block. Detailed configurations and roles of the own lane recognizing unit 162 and the lane change recognizing unit 164 will be described later with reference to FIGS. 3 to 5.

The lane information output unit 170 obtains information on the own lane of the vehicle 200 from the own lane recognition unit 162 and obtains information on the movement route of the vehicle 200 from the lane change recognition unit 164. For example, by outputting through a navigation (not shown) linked to the own lane recognition system 10 using a mask defined in the driving road image, the driver can provide information on the currently moving lane and lane change information. Play a role.

Accordingly, the vehicle lane recognition system 10 using the mask defined in the driving road image according to an exemplary embodiment of the present invention enables the driver to prepare for a lane change in advance, and moves to a destination when driving a complicated downtown road or driving on a highway. It is possible to be guided by a road that is more suitable and efficient for movement.

3 is a view showing a part of the operation of the magnetic lane recognition unit 162 of the magnetic lane recognition system 10 using a mask defined in the driving road image according to an embodiment of the present invention, Figure 4 is a view of the present invention FIG. 4 is a view illustrating an operation of the magnetic lane recognition unit 162 of the lane driving lane change recognition system 10 using the mask defined in the driving road image according to an exemplary embodiment.

Referring to FIG. 3, the magnetic lane recognition unit 162 first defines a mask line in an image of a driving road where two images obtained from two cameras are matched together by the image matching unit 130, for example, a mask. The lines may be spaced apart from each other by a predetermined distance from the driving road, may be arranged in a direction crossing the lane, and may be set to any number. In an embodiment of the present invention, each of the driving roads may be defined to be spaced apart from each other by a predetermined interval, but is not limited thereto.

As mentioned above, after defining the mask line, it is determined whether the lane adjacent to the mask line is disconnected. When the color information is detected in all 10 mask lines by detecting the color information corresponding to the lane in each of the 10 mask lines, If the color detection is cut off in the middle, it can be determined as a dotted line lane. Recognized colors can be limited to white, yellow and orange.

For example, referring to FIG. 4, when the driving road is a 3 to 5 lane road, when the color information of the pixel corresponding to the mask line is recognized, the yellow solid line is on the left side of the vehicle and the white dotted line is on the right side. Can recognize its lane as one lane. Alternatively, when the white dotted line is on the left side of the vehicle and there is a white solid line or an orange solid line, the own lane may be recognized as the last lane. Alternatively, in case of dotted lines on both sides, the second lane may be detected as the own lane by detecting the number of lanes in the solid orange lane.

As another example, when a yellow dotted line / solid line or a solid white line is detected at the left side of the vehicle, and a yellow or white solid line is detected at the right side of the vehicle, it may be recognized that the driving road is a one-lane road and that the own lane is a one lane.

As another example, if a yellow solid line is detected on the left side of the vehicle, a white dotted line on the right side, a white dotted line or solid line on the left side, and a white or yellow solid line on the right side, the driving road is a two-lane road and the lane is one lane. It can be recognized.

Here, the vehicle lane recognition may use total lane number information obtained from GPS (Global Positioning System) information, but is not limited thereto.

FIG. 5 is a diagram for describing an operation of the lane change recognizing unit 164 of the magnetic lane recognition system 10 using a mask defined in a driving road image, according to an exemplary embodiment. Referring to FIG. 5, the lane change recognizing unit 164 of the vehicle lane recognition system 10 using the mask defined in the driving road image according to an embodiment of the present invention first recognizes a lane change of a vehicle. First, a mask block is defined on a driving road image in which two images obtained from two cameras are matched by the image matching unit 130. For example, n images are divided into n driving road images based on a lane adjacent to the vehicle. Natural number), but is not limited thereto.

As described above, after defining the mask block, the lane change may be recognized by detecting a change amount of the vehicle coordinates in the mask block.

For example, reference is made to FIG. 5 which shows a Birds eye view when the vehicle changes from the left lane to the right lane. In fact, if the vehicle moves from the left lane to the right lane to change lanes, the lane in the mask block moves from right to left, the right lane becomes left lane, and a new right lane appears on the right side of the image. At this time, the mask block for lane detection moves to the left in the previous frame and then to the right in the current frame. Accordingly, the lane change recognition unit 164 may recognize that the vehicle has changed lanes to the right. With this in mind, it can be recognized that a lane change has been made if the change amount of the mask block coordinates is steep.

The navigation system, which represents the current driving guidance system, uses information from a global positioning system (GPS). However, using GPS, the error rate is about 15M, and the position of the vehicle is only known on the road, and it is impossible to accurately recognize the lane. In addition, the guidance of the most suitable lane for the movement route is inaccurate, and it is difficult to prevent traffic accidents caused by the driver's sudden lane change due to the provision of late information service.

The magnetic lane recognition system 10 using the mask defined in the driving road image according to an exemplary embodiment of the present invention may provide all lane images of the road by matching real-time images acquired from two cameras, and matched images. By defining a mask line on the line, it detects the exact position of the lane by detecting the color and shape of the lane according to the mask line, and by defining the mask block in the matched image, it changes the lane by detecting the change of vehicle coordinates in the mask block Can be recognized.

Therefore, when the self-driving lane recognition system 10 using the mask defined in the driving road image according to an embodiment of the present invention is linked with the navigation, not only the driving route guidance but also information on the current moving lane and lane change information of the vehicle. By providing the driver, the driver may be prepared for a lane change in advance, and may be guided to a driving road that is more suitable and efficient for moving to a destination when driving a complicated city road or a road.

6 is a flowchart illustrating a lane detection method and a lane change recognition method using a mask defined in a driving road image according to an exemplary embodiment of the present invention. Referring to FIG. 6, the method for recognizing a lane and changing lanes using a mask defined in a driving road image according to an embodiment of the present disclosure includes: acquiring a driving road image through two cameras (S610); Defining a mask line in the driving road image (S622), Detecting the color and shape of the lane based on the mask line (S632), Recognizing the own lane based on the detection result (S642), Driving road image Defining a mask block at step S624, detecting a change in vehicle coordinates based on the mask block at step S634, recognizing a lane change based on the detection result at step S644, and regarding a magnetic lane and lane change. It may be configured to include the step of outputting the information (S650).

Step S610 serves to obtain an image of the driving road through the two cameras. The two cameras 210 and 220 mounted on the vehicle 200 capture the image of the driving road to obtain the divided images 310 and 320, and the image matching unit 130 captures the divided images 310 and 320. Corresponds to the process of matching and combining them into one image. In the collected image, the analog image may be converted into a digital image by the digital image converter 140. An image converted into a digital image may be filtered by the image filter unit 150 to clarify outlines of driving roads, vehicles, lanes, and road surroundings included in the image.

In operation S622, a mask line is defined in the driving road image. This corresponds to a process in which the own lane recognizing unit 162 arbitrarily defines a mask line in the driving road image. For example, the mask line may be spaced apart from the image by a predetermined distance, disposed in a direction crossing the direction of the lane, and may be set to any number. In an embodiment of the present invention, each of the driving roads may be defined to be spaced apart from each other by a predetermined interval, but is not limited thereto.

Step S632 serves to detect the color and shape of the lane based on the mask line. This corresponds to a process in which the self-distance recognition unit 162 determines whether the lane adjacent to the mask line is cut off. In detail, the color information corresponding to the lanes may be recognized in each of the 10 mask lines, and if the color is detected in all 10 mask lines, it may be recognized as a straight line. Recognized colors can be limited to white, yellow and orange.

In step S634, the self lane is recognized based on the detection result. This corresponds to a process in which the own lane recognizing unit 162 recognizes the own lane according to the lane detection result of step S632. For example, if the driving road is a 3 to 5 lane road, when the color information of the pixel corresponding to the mask line is recognized, a yellow solid line is on the left side of the vehicle and a white dotted line is on the right side of the road. It can be recognized as a lane. Alternatively, when the white dotted line is on the left side of the vehicle and there is a white solid line or an orange solid line, the own lane may be recognized as the last lane. Alternatively, in case of dotted lines on both sides, the second lane may be detected as the own lane by detecting the number of lanes in the solid orange lane.

As another example, when a yellow dotted line / solid line or a solid white line is detected at the left side of the vehicle, and a yellow or white solid line is detected at the right side of the vehicle, it may be recognized that the driving road is a one-lane road and that the own lane is a one lane.

As another example, if a yellow solid line is detected on the left side of the vehicle, a white dotted line on the right side, a white dotted line or solid line on the left side, and a white or yellow solid line on the right side, the driving road is a two-lane road and the lane is one lane. It can be recognized.

Here, the vehicle lane recognition may use total lane number information obtained from GPS (Global Positioning System) information, but is not limited thereto.

In operation S632, a mask block is defined in the driving road image. This corresponds to a process in which the lane change recognition unit 164 defines a mask block in the driving road image. The mask block may be defined by dividing the driving road image by n equal parts (n is a natural number) based on a lane adjacent to the vehicle, but is not limited thereto.

Step S634 serves to detect a change in vehicle coordinates based on the mask block. This corresponds to a process in which the lane change recognition unit 164 detects a change amount of vehicle coordinates in the mask block.

Step S644 serves to recognize a lane change based on the detection result of step S634. This corresponds to a process of recognizing a lane change of the vehicle based on a result of the lane change recognition unit 164 detecting a change amount of lane coordinates in the mask block. For example, referring to FIG. 5, which shows a Birds eye view when a vehicle makes a change from the left lane to the right lane, in fact, if the vehicle moves from the left lane to the right lane to change lanes, the lane within the mask block will be It will move to the left, the right lane will be the left lane, and a new right lane will appear on the right side of the image. At this time, the mask block for lane detection moves to the left in the previous frame and then to the right in the current frame. Accordingly, the lane change recognition unit 164 may recognize that the vehicle has changed lanes to the right. With this in mind, it can be recognized that a lane change has been made if the change amount of the mask block coordinates is steep.

Step S650 serves to output information on the own lane and lane change. The lane information output unit 170 obtains information on the own lane of the vehicle 200 from the own lane recognizing unit 162, and obtains information on the moving route of the vehicle 200 from the lane change recognizing unit 164. This corresponds to a process of acquiring and outputting the same through a navigation (not shown) linked to the magnetic lane recognition system 10 using the mask defined in the driving road image.

The present invention may be embodied in many other specific forms without departing from the spirit or essential characteristics of the invention.

10: magnetic lane recognition system using a mask defined in the driving road image according to an embodiment of the present invention
110: first image input unit 120: second image input unit
130: image matching unit140: digital image conversion unit
150: image filter unit 160: lane recognition unit
170: lane information output unit 200: vehicle
210: left camera 220: right camera
310: split image 320: split image

Claims (8)

In a system for recognizing a lane by using a mask defined in the driving road image,
A photographing image matching unit for matching images of driving roads acquired from two cameras installed at both sides of the vehicle into one image; And
Defining a mask line in the matched driving road image and detecting a magnetic lane of the driving road based on the mask line;
The mask line is,
At least one or more spaced apart from the driving road image by a predetermined interval, and is disposed in a direction crossing the direction of the lane,
The lane recognizing unit,
Recognize the color information corresponding to the lane along the mask line, and if the color is continuously detected, it is recognized as a straight line, and if it breaks in the middle, it is determined as the dotted line lane, and the total number of lanes of the driving road is obtained from the GPS (Global Positioning System) And a lane change recognizing unit defining a mask block in the matched driving road image, and detecting a lane change by sensing a change amount of the vehicle coordinates in the mask block, wherein the mask block includes a vehicle in the driving road image. And a block for dividing the driving road image into n equal parts (n is a natural number) based on a lane closest to the driving lane, using the mask defined in the driving road image.
delete delete The method of claim 1, wherein the lane recognition unit,
When the driving road is a 3 to 5 lane road, the color information of the pixel corresponding to the mask line is recognized, and when the yellow solid line is on the left side of the vehicle and the white dotted line is on the right side, the lane is recognized as the first lane. If a white dotted line is on the left side of a vehicle and there is a white solid line or an orange solid line, the lane is recognized as the last lane. If there are dotted lines on both sides of the vehicle, the lane is detected by detecting the number of lanes in the orange solid lane. Magnetic lane recognition system using the mask defined in the driving road image.
delete The method of claim 1, wherein the captured image matching unit,
Images of the roads obtained from the two cameras are combined in a panorama format and matched to one image, and from the cameras mounted on the right and 100 pixels on the right of the image taken from the camera mounted on the left side of the vehicle, of the two cameras. A self-driving vehicle recognition system using a mask defined in an image of a driving road, generating a single image by matching pixels to the left 100 pixels of a captured image through a dimension dynamic warping algorithm.
delete delete

Images