KR20210130688A - Apparatus for generating top-view image and method thereof - Google Patents

Abstract

Disclosed are a top-view image generating device and a method thereof. According to one aspect of the present invention, the top-view image generating method includes: a process of acquiring front, rear, left, and right images respectively from front, rear, left, and right cameras installed in a vehicle; a process of detecting a left lane and a right lane of the vehicle from the front, rear, left, and right images; a process of determining whether the color and shape of the left lane are the same as the color and shape of the right lane; a process of estimating an attitude angle of at least one of the front, rear, left, and right cameras when the color and shape are the same; and a process of generating a top-view image based on the camera attitude angle.

Description

Apparatus for generating a top-view image and a method therefor

The present invention relates to an apparatus and method for generating a top-view image, and more particularly, to an apparatus for generating a top-view image that generates a top-view image matched with an around-view image of an around-view monitoring system mounted on a vehicle. and methods thereof.

The Around View Monitoring (AVM) system of the vehicle synthesizes images obtained from four cameras installed on the front, both side mirrors, and trunk of the vehicle, respectively, and provides a bird's eye view of the vehicle's surroundings from above (Top view or Bird's Eye). It is a system that provides an around-view image that is displayed in the form of a View).

In the process of assembling cameras used in the around-view monitoring system, compensating for tolerances is essential. The vehicle manufacturer compensates for the tolerance for a vehicle equipped with an around-view monitoring system to meet the consistency standard of the around-view image, and then releases the vehicle.

However, new tolerances occur in each camera due to vibration of the vehicle that occurs while the vehicle is running after leaving the vehicle, repeated folding of side mirrors, and repeated opening/closing of the trunk, etc. Decreases the consistency of the view image. This causes deterioration of the display quality of the around-view image.

Accordingly, an object of the present invention is to provide an apparatus and method for generating a top-view image capable of compensating for camera tolerance while driving a vehicle in order to improve the consistency of around-view images.

According to one aspect of the present embodiment, the process of obtaining each of the front, rear, left, and right images from the front, rear, left, and right cameras installed in the vehicle; detecting a left lane and a right lane of the vehicle from the front, rear, left, and right images; determining whether the color and shape of the left lane and the color and shape of the right lane are the same; estimating an attitude angle of at least one of the front, rear, left, and right cameras when the color and shape are the same; and generating a top-view image based on the camera attitude angle.

According to another aspect of the present embodiment, a camera unit for generating front, rear, left and right images by photographing the front, rear, left, and right of the vehicle; a lane detection unit detecting a left lane and a right lane of the vehicle from the front, rear, left, and right images; a lane color/shape determining unit configured to determine whether the color and shape of the left lane and the color and shape of the right lane are the same; a camera posture angle estimator for estimating at least one camera posture angle among the front, rear, left and right cameras when the color and shape are the same; and an image synthesizing unit for generating a top-view image based on the camera attitude angle.

According to the present invention, by compensating for the camera tolerance in real time in a situation where the vehicle is traveling on the road, it is possible to reduce the hassle of the driver having to periodically visit a maintenance company with the vehicle to compensate for the camera tolerance.

In addition, by compensating for the camera tolerance in real time, it is possible to prevent deterioration of the display quality of the top view image due to the camera tolerance.

Effects of the present invention are not limited to those mentioned above, and other effects not mentioned will be clearly understood by those of ordinary skill in the art from the following description.

1 is a functional block diagram illustrating functions of an apparatus for generating a top-view image in units of blocks according to an embodiment of the present invention.
FIG. 2 is a diagram schematically illustrating a lane pattern detection process performed by the lane detection unit illustrated in FIG. 1 .
3 schematically shows rotation angles in a pitch direction, a roll direction, and a yaw direction defined in the camera coordinate system.
4 is a diagram schematically illustrating a method of estimating a rotation angle in a roll direction according to an embodiment of the present invention.
5 is a flowchart illustrating a method of generating a top-view image in which camera tolerance is corrected according to an embodiment of the present invention.

Advantages and/or features of the present invention, and methods of achieving them will become apparent with reference to the embodiments described below in detail in conjunction with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below, but will be implemented in various different forms, and only these embodiments allow the disclosure of the present invention to be complete, and common knowledge in the art to which the present invention pertains It is provided to fully inform the possessor of the scope of the invention, and the present invention is only defined by the scope of the claims. Like reference numerals refer to like elements throughout.

The present invention relates to an apparatus and method for estimating an AVM camera attitude angle and updating an image synthesis look-up table (LUT) of an AVM by using a lane existing on the ground detectable in an image.

In the AVM composite image, using the condition that the left and right lanes of the own vehicle are parallel, the thickness of the left and right lanes are the same, and the condition that each of the left and right lanes exist on the same line, It is possible to accurately estimate the camera attitude angle, and at the same time to update the lookup table (LUT) for synthesizing AVM images.

However, in an actual road environment, there are cases in which the condition that the thickness of the left and right lanes is the same is not satisfied. For example, when the color of the left lane and the color of the right lane are different from each other, the thickness of the left/right lane may be different from each other, and even when the shape of the left lane is a solid line and the shape of the right lane is a dotted line, the left/right The thickness of the lanes may be different.

As such, in the case of the left/right lanes having different colors and/or shapes, it is highly likely that the thickness of the left/right lanes does not satisfy the same condition, so in this case the ground-based (top-view) The camera attitude angle (relative to the coordinate system) cannot be accurately estimated.

In order to solve this problem, in the present invention, whether to perform AVM automatic tolerance correction is determined by using the detected color and shape (solid line or dotted line) of the left/right lane. That is, more accurate tolerance correction is possible by estimating the camera attitude angle for tolerance correction of each camera only when the color and shape (solid line or dotted line) of the left and right lanes are the same.

In addition, in an actual road driving environment, it is difficult to simultaneously detect lanes from the front/rear/left/right camera images. In the present invention, this problem is solved by accumulating and using lanes detected from images acquired while the vehicle is driving straight. solve it An improvement in AVM automatic tolerance correction performance can be expected through the present invention.

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

1 is a functional block diagram illustrating functions of an apparatus for generating a top-view image in units of blocks according to an embodiment of the present invention.

Referring to FIG. 1 , the top-view image generating apparatus 100 according to an embodiment of the present invention includes a camera unit 110 and a lens distortion correcting unit 120 to compensate for a tolerance of a camera for a top-view image. ), a lane detection unit 130 , a lane color/shape determination unit 140 , a camera angle estimation unit 150 , a storage unit 160 , an image rotation unit 170 , an image synthesis unit 180 , and a display unit 190 . include In addition, although not shown in FIG. 1 , a control unit for managing and controlling each operation of the components 110 to 190 may be further included, and the control unit may be implemented with a microcomputer or a central processing unit (CPU). have. In addition, the components 110 to 190 may be implemented as logical blocks in the microcomputer or central processing unit (CPU).

camera unit 110

The camera unit 110 generates a surrounding image taken around 360 degrees of the own vehicle (own vehicle) while driving the vehicle, and the surrounding image is a front image, a rear image, a left image and a right image (front/rear/ left/right image).

To this end, the camera unit 110 includes a front camera 112 , a rear camera 114 , a left camera 116 , and a right camera 118 .

The front camera may be installed on the front of the vehicle (eg, a front bumper) to generate a front image of the front of the vehicle. The rear camera may be installed on a rear side of the vehicle (eg, a rear bumper) to generate a rear image photographing the rear of the vehicle. The left camera may be installed on the left side of the vehicle to generate a left image of the left side of the vehicle. The right camera may be installed on the right side of the vehicle to generate a right image of the right side of the vehicle. These cameras 112 to 118 preferably include a wide-angle lens having a wide angle of view to photograph 360 degrees around the own vehicle.

Lens distortion correction unit 120

The lens distortion correcting unit 120 may use a lens distortion model or The peripheral image (front/rear/left/right image) is corrected based on a lens distortion correction algorithm. A lens distortion model or a lens distortion correction algorithm is a well-known technology, and since the present invention is not characterized in limiting a lens distortion model or a lens distortion correction algorithm, a detailed description thereof will be omitted.

lane detection unit 130

The lane detector 130 detects at least two lane patterns from each image (front/rear/left/right) for which lens distortion is corrected. To detect the lane pattern, various lane detection algorithms may be used. For example, a Top-Hat filter, a Hough transformation algorithm, a RANSAC-based line fitting algorithm, a Kalman filter, or the like may be used.

Although not particularly limited, an example of detecting a lane pattern performed by the lane detection unit 130 will be described with reference to FIG. 2 .

First, in step (1), a region of interest is selected in an image in which lens distortion is corrected. For example, the image may be halved into two regions ⓐ and ⓑ in the horizontal direction, and a lower region ⓑ among the two halved regions may be selected as the region of interest.

Excluding the upper area (A) from the area of interest is because the upper area (ⓐ) has a background such as the sky where no lanes exist in addition to the road surface on which the lanes exist. This is to shorten the calculation processing time by excluding an unnecessary search process for

Next, in step (2), a morphological operation is performed on the image set as the region of interest (ⓑ), and the image set as the region of interest is converted into a binary image.

Next, in step (3), connected component analysis is performed on the binary image to label objects (lane-line objects) appearing in the binary image. The connection element analysis objectifies a set of interconnected single pixels into an independent one. Using this method, objects (suboptimal objects) existing in the binary image can be efficiently separated.

Next, in step (4), since the lane pattern cannot exist in the horizontal direction and takes the form of a straight line, an object that cannot be viewed as a lane pattern among the labeled objects (lane objects) in consideration of the angle and the degree of straight line are removed, and objects on the same line that are not removed are detected as the next best candidate group.

Next, in step 5, objects included in the detected lane candidate group are connected as one line using a line fitting algorithm, and the connected lines are detected as a left lane pattern and a right lane pattern. Meanwhile, FIG. 2 exemplifies a detection process for one lane pattern for the sake of simplification of the drawing.

Lane color/shape determination unit 140

Referring back to FIG. 1 , the lane color/shape determination unit 140 determines whether the detected color and shape (solid line or dotted line) of the detected left lane pattern and the right lane pattern are the same.

In an embodiment of the present invention, only when the color and shape between the left lane pattern and the right lane pattern are the same, the camera attitude angle estimation unit 150 described below estimates the camera attitude angle. That is, the work of estimating the camera angle is not performed with respect to a corresponding camera that provides an image in which left and right lane patterns having different colors and shapes are detected.

In order to determine whether the color of the left lane pattern and the color of the right lane pattern are the same, first, the lane color/shape determination unit 140 calculates a left average color value of pixels included in the left lane pattern, A right average color value of pixels included in the right lane pattern is calculated. Here, the average color value may be an average color value for RGB color values based on an RGB color model or an average color value for Hue values based on a Hue/Saturation/Value (HSV) model.

When the left average color value and the right average color value are calculated, the lane color/shape determination unit 140 calculates a difference value between the left average color value and the right average color value, and the calculated difference value is a critical value If the value is less than or equal to the value, it is determined that the color of the left lane pattern and the color of the right lane pattern are the same.

In order to determine whether the shape of the left lane pattern and the shape of the right lane pattern are the same, first, the lane color/shape determining unit 140 determines whether the shape of the left lane pattern is a dotted line or a solid line, It is determined whether the shape of the right lane pattern is a dotted line shape or a solid line shape.

A method of determining whether each lane pattern is a dotted line or a solid line may be determined from the number of labeled objects (objects on the same line) belonging to the lane candidate group in the aforementioned lane detection process. For example, if the number of labeled objects (objects on the same line) belonging to the lane candidate group is less than or equal to a preset number, the corresponding lane pattern is determined as a solid line, and when the number exceeds the preset number, it is determined as a dotted line. .

Camera attitude angle estimation unit 150

The camera attitude angle estimation unit 150 estimates a camera attitude angle of a camera that provides an image in which a left lane pattern and a right lane pattern having the same color and shape are detected. Here, the camera attitude angle includes a rotation angle in a pitch direction, a rotation angle in a roll direction, and a rotation angle in a yaw direction, as shown in FIG. 3 .

When the camera attitude angle is estimated, a rotation matrix including the estimated camera attitude angle is generated, and the generated rotation matrix is stored in the storage unit 160 in the form of a lookup table to update the previously stored lookup table. The storage unit 160 may be a non-volatile memory.

The camera attitude angle may be estimated using geometric conditions (geometric properties, geometric properties) of the left/right lane patterns in the top-view image coordinate system.

Camera attitude estimation method1

The camera attitude angle estimation method 1 is to estimate (or calculate) the rotation angle in the pitch direction and the rotation angle in the yaw direction.

In the camera coordinate system, the left lane pattern and the right lane pattern are not parallel, and the thickness of the left lane pattern and the thickness of the right lane pattern are different from each other, but in the top view coordinate system, the left lane pattern and the right lane pattern are parallel to each other, and the left lane pattern The thickness of the right lane pattern and the thickness of the right lane pattern must be the same. Here, 'thickness' refers to the width of the lane itself, which is generally white or yellow, applied on the road.

Therefore, the camera attitude angle estimation method 1 in the pitch direction satisfies all geometric conditions including the condition that the left lane pattern and the right lane pattern are parallel and the thickness of the left lane pattern and the thickness of the right lane pattern are the same To calculate the rotation angle of and the rotation angle in the yaw direction.

Camera attitude estimation method2

The camera attitude angle estimation method 2 is another method for estimating the rotation angle in the pitch direction. The condition that the left lane pattern and the right lane pattern are parallel and the left lane pattern and the right lane pattern are the yz-axis plane of FIG. To calculate the rotation angle in the pitch direction that satisfies all geometric conditions including the condition to be parallel.

Camera attitude estimation method3

The camera attitude angle estimation method 3 is a method of estimating the rotation angle in the roll direction. Referring to FIG. 4 , first, an arbitrary image using the rotation angle estimated according to the camera attitude angle estimation methods 1 and 2 described above. (hereinafter referred to as a 'front image') is rotated to generate a front top-view image.

_{Then, based on the feature point (P1 n-1} ) included in the front top-view image of the previous frame, a movement vector 71 of the _{feature point (P1 n} ) included in the front top-view image of the current frame is calculated, A predicted motion vector 73 of _{the feature point P1 n-1} according to the vehicle motion information is calculated. Thereafter, the rotation angle in the roll direction can be estimated by calculating the rotation angle Θ at which the motion vector 71 coincides with the predicted motion vector 73 .

On the other hand, the feature point on the ground may be a crack, a scratch, an application mark line other than a lane, etc. that can be seen on the road surface, and the movement information of the vehicle is GPS information, information obtained from a gyro sensor, etc. It can be obtained from all kinds of information.

When the rotation angle in the pitch direction, the rotation angle in the roll direction, and the rotation angle in the yaw direction are estimated by the above-described camera attitude angle estimation method 1-3, the camera attitude angle estimation unit 150 ) generates a rotation matrix consisting of the estimated rotation angle in the pitch direction, the rotation angle in the roll direction, and the rotation angle in the yaw direction, and stores it in the storage unit 160 in the form of a lookup table. Save.

image rotation unit 170

The image rotation unit 170 rotates the front/rear/left/right images generated by the camera unit 110 by using the rotation matrix R stored in the storage unit 160, respectively, and performs four front/rear images. /Converts to the left/right top view image.

Image synthesizing unit 180

The image synthesizing unit 180 synthesizes four front/rear/left/right top-view images rotationally converted by the image rotating unit 170 to generate one final top-view image, which is then displayed through the display unit 190 . print out

5 is a flowchart illustrating a method of generating a top-view image in which camera tolerance is corrected according to an embodiment of the present invention. In the description of each step below, the description overlapping with the description with reference to FIGS. 1 to 4 will be briefly described. to be listed.

Referring to FIG. 5 , first, in step S551, the vehicle starts to travel.

Next, in step S513, it is determined whether the vehicle travels in a straight line. Here, the reason for determining whether to drive in a straight line is that lanes may not be simultaneously detected in front/rear/left/right images captured around the vehicle in an actual road driving environment. For example, in curve driving, a lane may not be detected in the left image or the right image. In this case, since the lane is not detected in the left or right image, the camera attitude angle estimation process for the left/right camera cannot be performed.

Accordingly, in an embodiment of the present invention, the camera attitude angle estimation process for some cameras is prevented by performing the camera attitude angle estimation process only in a straight-line driving environment in which lanes can be simultaneously detected from the front/rear/left/right images. can do. If the driving environment of the vehicle is not linear driving, the camera attitude angle estimation process may be stopped until the vehicle travels in a straight line. On the other hand, whether the vehicle is traveling in a straight line can be known from all kinds of information that can determine the traveling direction of the vehicle, such as GPS information installed in the vehicle, information obtained from a gyro sensor, and the like.

Next, if it is confirmed that the vehicle is traveling in a straight line, in step S515 , front/rear/left/right images are respectively acquired from the front/rear/left/right cameras.

Next, in step S517, the lens distortion of the front/rear/left/right image generated due to the wide angle of view of the wide-angle lens is corrected using a lens distortion algorithm.

Next, in step S519, a lane line (or lane pattern) is detected from the front/rear/left/right image for which lens distortion is corrected using a lane detection algorithm.

Next, in step S521, a process of determining whether a lane is detected from all of the front/rear/left/right images is performed. If a lane is detected in all images, the process proceeds to step S523. If no lane is detected in at least one of all images, the process returns to before step S513 and performs steps S513 to S519 again.

Next, when lanes are detected in all images, in step S523, the color/shape of lanes detected in each image is determined.

Next, in step S525, it is determined whether the color/shape of the lane detected in each image is the same. Specifically, when the left and right lanes are detected in the front image, it is determined whether the colors/shapes of the left and right lanes are the same, and similarly, whether the colors/shapes of the left and right lanes detected in the rear image are the same. judge

Meanwhile, only one left lane may be detected in the left image, and only one right lane may be detected in the right image. In this case, it is determined whether the color/shape of the left lane detected in the left image is the same as the color/shape of the left lane detected in the front image (or rear image). Similarly, it is determined whether the color/shape of the right lane detected in the right image is the same as that of the right lane detected in the front image (or rear image).

Next, if the color/shape of the left and right lanes are the same, in step S527, the camera attitude angle (or the rotation angle in the pitch direction, roll direction, and yaw direction) for camera tolerance correction is estimated using the detected left and right lanes. That is, with the camera attitude angle estimation method 1-3 described above, the camera attitude angle (or the rotation angle in the pitch direction, roll direction, and yaw direction) with respect to the ground (based on the top-view coordinate system) is estimated (calculated) )do.

On the other hand, when only one left lane is detected in the left image and only one right lane is detected in the right image, the left lane detected in the left image and the right lane detected in the right image are the aforementioned rotation angles. Camera attitude angles satisfying the conditions described in Estimation Methods 1, 2, and 3 are calculated, and the camera attitude angles of the left and right cameras can be estimated through this.

Alternatively, the camera attitude angles of the left and right cameras may be estimated using the camera attitude angles of the front camera (or rear camera) estimated using the left and right lanes detected from the front image (or the rear image). For example, under the condition that the thickness of the left lane detected in the front image (or rear image) and the thickness of the left lane detected in the left image are the same, and detection in the left lane and left image detected in the front image (or rear image) The camera attitude angle of the left camera can be estimated by estimating the camera attitude angle that satisfies the condition that the left lane is parallel. In a similar manner, the camera attitude angle of the left camera may be estimated by comparing the right lane detected from the front image (or the rear image) with the right lane detected from the right image.

When all camera attitude angles for the front/rear/left/right cameras are estimated, a rotation matrix including the estimated camera attitude angles is constructed.

Next, in step S529, the configured rotation matrix is stored in the storage unit in the form of a lookup table, and the previously stored lookup table is updated.

Then, using the rotation matrix stored in the form of a lookup table in the storage unit, the four currently input before/after/left/right images are rotationally converted into before/after/left/right top-view images, and the rotation-converted before-and-after images are rotated. By synthesizing the /rear/left/right top-view image, the final top-view image with the camera tolerance corrected can be provided to the driver.

Meanwhile, the block diagram of FIG. 1 showing an apparatus for generating a top-view image should be understood as embodied in the principle of the invention from a functional point of view. Similarly, the flow diagram of FIG. 5 is to be understood as representing various processes that may be tangibly embodied on a computer-readable medium and performed by a computer or processor, whether or not a computer or processor is explicitly shown.

Blocks 120 to 190 of FIG. 1 indicated by a processor or a similar concept may be provided by using hardware capable of executing software as well as dedicated hardware.

When blocks 120 to 190 of FIG. 1 are implemented by a processor, the functions of blocks 120 to 190 shown in FIG. 1 may be provided by a single dedicated processor, a single shared processor, or a plurality of individual processors. and some of them may be shared.

Further, clear use of terms presented as processor, control, or similar concepts should not be construed as citing exclusively hardware capable of executing software, without limitation, digital signal processor (DSP) hardware, ROM (for storing software) ROM), RAM, and non-volatile memory are to be understood as implicitly included. Of course, other hardware for general use may also be included.

The above description is merely illustrative of the technical idea of the present invention, and various modifications, changes, and substitutions are possible within the range that does not depart from the essential characteristics of the present invention by those of ordinary skill in the art to which the present invention pertains. will be. Accordingly, the embodiments disclosed in the present invention and the accompanying drawings are for explaining, not limiting, the technical spirit of the present invention, and the scope of the technical spirit of the present invention is not limited by these embodiments and the accompanying drawings . The protection scope of the present invention should be interpreted by the following claims, and all technical ideas within the scope equivalent thereto should be construed as being included in the scope of the present invention.

Claims (18)

a process of acquiring front, rear, left, and right images from front, rear, left, and right cameras installed in the vehicle;
detecting a left lane and a right lane of the vehicle from the front, rear, left, and right images;
determining whether the color and shape of the left lane and the color and shape of the right lane are the same;
estimating an attitude angle of at least one of the front, rear, left, and right cameras when the color and shape are the same; and
The process of generating a top-view image based on the camera attitude angle
A top-view image generating method comprising a. According to claim 1,
Before the process of acquiring the front and rear left and right images,
The method for generating a top-view image, characterized in that it further comprises the step of determining whether the driving state of the vehicle is a straight-line driving. According to claim 1,
The process of determining whether the color and shape are the same,
determining whether the shape of each of the left lane and the right lane is a dotted line or a solid line; and
A process of determining whether the shapes of the left lane and the right lane are the same dotted line or the same solid line
A top-view image generating method comprising a. According to claim 1,
The process of determining whether the color and shape are the same,
Comparing a difference value between the average color value of the left lane and the average color value of the right lane with a threshold value, and determining that the colors of the left lane and the right lane are the same when the difference value is less than or equal to the threshold value
A top-view image generating method comprising a. According to claim 1,
The process of estimating the camera attitude angle is
The method for generating a top-view image, characterized in that the camera attitude angle is estimated using geometric conditions of the left and right lanes required in a top view image coordinate system. 6. The method of claim 5,
The geometric condition is
and at least one of a condition in which the left lane and the right lane are parallel, and a condition in which a thickness of the left lane and a thickness of the right lane are the same. 6. The method of claim 5,
The process of estimating the camera attitude angle is
A method for generating a top-view image, characterized in that it is a process of estimating the attitude angles of a left camera and a right camera so that the left lane detected from the left image and the right lane detected from the right image satisfy the geometric condition. 6. The method of claim 5,
The process of estimating the camera attitude angle is
It is characterized in that it is a process of estimating the attitude angles of at least one of the front camera and the rear camera and the left camera so that the left lane detected from at least one of the front image and the rear image and the left lane detected from the left image satisfy the geometric condition. How to create a top-view video. 6. The method of claim 5,
The process of estimating the camera attitude angle is
It is characterized in that it is a process of estimating the attitude angle of at least one of the front camera and the rear camera and the right camera so that the right lane detected from at least one of the front image and the rear image and the right lane detected from the right image satisfy the geometric condition. How to create a top-view video. A camera unit for generating front, rear, left, and right images by photographing the front, rear, left, and right of the vehicle;
a lane detection unit detecting a left lane and a right lane of the vehicle from the front, rear, left, and right images;
a lane color/shape determining unit configured to determine whether the color and shape of the left lane and the color and shape of the right lane are the same;
a camera posture angle estimator for estimating at least one camera posture angle among the front, rear, left and right cameras when the color and shape are the same; and
An image synthesizing unit that generates a top-view image based on the camera attitude angle
A top-view image generating device comprising a. 11. The method of claim 10,
Top-view image generating apparatus, characterized in that it further comprises a driving determination unit for determining whether the driving state of the vehicle is a straight-line driving. 11. The method of claim 10,
The lane color/shape determination unit,
The apparatus for generating a top-view image, characterized in that by determining whether the shapes of the left and right lanes are dotted lines or solid lines, determining whether the shapes of the left and right lanes are the same dotted line or the same solid line. 11. The method of claim 10,
The lane color/shape determination unit,
a difference value between the average color value of the left lane and the average color value of the right lane is compared with a threshold value, and when the difference value is less than or equal to the threshold value, it is determined that the colors of the left lane and the right lane are the same A top-view image generating device with 11. The method of claim 10,
The camera attitude angle estimation unit,
The apparatus for generating a top-view image, characterized in that the camera attitude angle is estimated using geometric conditions of the left and right lanes required in a top view image coordinate system. 15. The method of claim 14,
The geometric condition is
and at least one of a condition in which the left lane and the right lane are parallel and a thickness of the left lane and a thickness of the right lane are the same. 15. The method of claim 14,
The camera attitude angle estimation unit,
A top-view image generating apparatus, characterized in that the attitude angles of the left and right cameras are estimated so that the left lane detected from the left image and the right lane detected from the right image satisfy the geometric condition. 15. The method of claim 14,
The camera attitude angle estimation unit,
A tower, characterized in that at least one of the front camera and the rear camera and the left camera are estimated so that the left lane detected from at least one of the front image and the rear image and the left lane detected from the left image satisfy the geometric condition. -View image generating device. 15. The method of claim 14,
The camera attitude angle estimation unit,
A tower, characterized in that the at least one of the front camera and the rear camera and the right camera are estimated so that the right lane detected from at least one of the front image and the rear image and the right lane detected from the right image satisfy the geometric condition. -View image generating device.

Images