KR102650906B1 - Detecting traffic lane apparatus and method - Google Patents

Abstract

This disclosure relates to a lane detection apparatus and method. Specifically, the lane detection device according to the present disclosure includes an image receiving unit that receives image data in which the surroundings of the own vehicle are captured, a first image processing unit that detects a sector determined to contain a lane based on the image data, and an image processing unit that detects a sector determined to contain a lane based on the image data. A second image processing unit that divides the image data into a predetermined number of grids and detects a grid that is determined to contain a lane in the image data, and a lane detection unit that detects a lane based on the grid containing the lane and the sector containing the lane. It should include.

Description

Lane detection device and method {DETECTING TRAFFIC LANE APPARATUS AND METHOD}

The present embodiments relate to a lane detection apparatus and method for detecting lanes.

Recently, ADAS (Advanced Driver Assistance System), which provides driving convenience functions to drivers through the use of various sensors, is being installed in vehicles. Specifically, ADAS can help prevent a vehicle from leaving its lane (LKAS, Line Keeping Assist System) or provide the driving function itself (SCC, Smart Cruise Control).

The above-described functions may be performed by detecting the distance between the vehicle and surrounding objects, or by detecting the lane, etc., by a camera mounted on the vehicle.

However, accurate lane detection is essential for the performance of the above-mentioned functions, and research to reduce the false detection rate for accurate lane detection still remains a difficult task.

Against this background, the present disclosure seeks to provide a lane detection device and method for detecting lanes around a host vehicle by utilizing a plurality of image detection methods.

In order to solve the above-described problem, in one aspect, the present disclosure includes an image receiving unit that receives image data in which the surroundings of the own vehicle are photographed, a first image processing unit that detects a sector determined to contain a lane based on the image data, and A second image processing unit that divides the image data into a predetermined number of grids and detects a grid that is determined to contain a lane in the image data, and a second image processing unit that detects a grid containing the lane and a sector containing the lane Provided is a lane detection device including a lane detection unit that detects.

In another aspect, the present disclosure includes an image receiving step of receiving image data in which the surroundings of the own vehicle are captured, a first image processing step of detecting a sector determined to contain a lane based on the image data, and a first image processing step of detecting a sector determined to contain a lane based on the image data. A second image processing step of dividing the image data into a set number of grids and detecting a grid that is determined to contain a lane, and lane detection of detecting a lane based on the grid containing the lane and the sector containing the lane. A lane detection method including the following steps is provided.

According to the present disclosure, a lane detection device can reduce errors and shorten post-processing time by comparing data through a plurality of methods for detecting lanes.

1 is a block diagram for explaining a lane detection device according to an embodiment of the present disclosure.
FIG. 2 is a diagram briefly illustrating a lane detection process according to an embodiment.
FIG. 3 is a diagram illustrating how a first image processor detects a lane by processing image data according to an embodiment.
FIG. 4 is a diagram for explaining a morphology operation on a sector detection result according to an embodiment.
FIG. 5 is a diagram illustrating how a second image processor detects a lane by processing image data according to an embodiment.
FIG. 6 is a diagram illustrating dividing image data into a plurality of y-axis coordinates with equal intervals, according to an embodiment.
FIG. 7 is a diagram illustrating how a first image processor detects a lane line from image data according to an embodiment.
Figure 8 is a diagram for explaining types of occlusion phenomenon according to one embodiment.
FIG. 9 is a diagram illustrating how a second image processor detects a lane line from image data according to an embodiment.
Figure 10 is a diagram for explaining calculating core coordinates and candidate coordinates according to an embodiment.
FIG. 11 is a diagram illustrating classification of detected lanes into three cases according to an embodiment.
Figure 12 is a diagram for explaining post-processing of candidate coordinates according to an embodiment.
FIG. 13 is a diagram illustrating performing a spline fitting algorithm on key coordinates according to an embodiment.
Figure 14 is a flowchart explaining a lane detection method according to an embodiment of the present disclosure.

Hereinafter, some embodiments of the present disclosure will be described in detail with reference to illustrative drawings. In adding reference numerals to components in each drawing, the same components may have the same reference numerals as much as possible even if they are shown in different drawings. Additionally, in describing the present embodiments, if it is determined that a detailed description of a related known configuration or function may obscure the gist of the present technical idea, the detailed description may be omitted. When “comprises,” “has,” “consists of,” etc. mentioned in the specification are used, other parts may be added unless “only” is used. When a component is expressed in the singular, it can also include the plural, unless specifically stated otherwise.

Additionally, in describing the components of the present disclosure, terms such as first, second, A, B, (a), and (b) may be used. These terms are only used to distinguish the component from other components, and the nature, sequence, order, or number of the components are not limited by the term.

In the description of the positional relationship of components, when two or more components are described as being “connected,” “coupled,” or “connected,” the two or more components are directly “connected,” “coupled,” or “connected.” ", but it should be understood that two or more components and other components may be further "interposed" and "connected," "combined," or "connected." Here, other components may be included in one or more of two or more components that are “connected,” “coupled,” or “connected” to each other.

In the description of temporal flow relationships related to components, operation methods, production methods, etc., for example, temporal precedence relationships such as “after”, “after”, “after”, “before”, etc. Or, when a sequential relationship is described, non-continuous cases may be included unless “immediately” or “directly” is used.

On the other hand, when a numerical value or corresponding information (e.g., level, etc.) for a component is mentioned, even if there is no separate explicit description, the numerical value or corresponding information is related to various factors (e.g., process factors, internal or external shocks, It can be interpreted as including the error range that may occur due to noise, etc.).

1 is a block diagram for explaining a lane detection device according to an embodiment of the present disclosure.

Referring to FIG. 1, the lane detection device according to the present disclosure may include an image receiver, a second image processor, a first image processor, and a lane detection unit.

The lane detection device according to an embodiment of the present disclosure may be an ADAS (Advance Driver Assistance Systems) that provides information to help the driver drive the vehicle or to help the driver control the vehicle.

Here, ADAS can refer to various types of advanced driver assistance systems, and driver assistance systems include, for example, Traffic Sign Recognition, Autonomous Emergency Braking, and Smart Parking Assistance System (SPAS). : Smart Parking Assistance System, Blind Spot Detection (BSD) system, Adaptive Cruise Control (ACC) system, Lane Departure Warning System (LDWS), Lane Keeping Assist System (LKAS) : Lane Keeping Assist System), Lane Change Assist System (LCAS), etc. However, it is not limited to this.

The image receiver may receive image data captured around the vehicle.

Here, image data may be received from an image sensor, and the image sensor may include a camera sensor, LiDAR sensor, etc.

A camera sensor may be a video camera module that converts and reproduces images into electrical signals. Here, if the electrical signal is processed into a standard signal (NTSC) and connected to a video tape recorder (VTR) and a monitor television, the image of the subject can be reproduced.

In one example, the image receiver may receive image data from a first camera and a second camera mounted on both ends of the vehicle and photographing the front.

Image data includes a series of unit frame images. In other words, it consists of a series of video frames. For example, if the frame rate is 60 FPS (Frames Per Second), this may mean that 60 still images, or unit frame images, are transmitted and received per second. Accordingly, video frame, frame video, or unit frame are used as almost similar concepts, but the terminology used may be slightly different depending on whether the object being referred to is a frame or data.

Hereinafter, in the present disclosure, image information captured from an image sensor may refer to image data captured from an image sensor.

Image data captured by an image sensor may be generated, for example, in one of the following formats: raw AVI, MPEG-4, H.264, DivX, or JPEG. For example, image data captured from a front image sensor or a side image sensor may be processed by the processor. Here, the processor may be included in each image sensor, and may be included in a second image processing unit, a first image processing unit, or a lane detection unit, which will be described later. Additionally, a separate module that performs the functions of a processor may be installed.

FIG. 2 is a diagram briefly illustrating a lane detection process according to an embodiment.

The lane detection device is mounted on the vehicle and can detect the lane in which the vehicle is traveling or nearby lanes. As an example, the lane detection device receives image data captured around the vehicle, divides it into a predetermined number of grids based on the image data, and detects a grid that is determined to contain a lane in the image data. , A sector determined to contain a lane may be detected based on image data, and a lane may be detected based on the grid containing the lane and the sector containing the lane.

The lane detection device can receive image data, detect each lane through two methods (Row-wise Classification, Segmentation), and detect the lanes by performing post processing.

FIG. 3 is a diagram illustrating how a first image processor detects a lane by processing image data according to an embodiment.

Referring to FIG. 3, a sector determined to contain a lane can be detected based on image data.

In one example, the first image processing unit may detect objects in an image (e.g., frame image) using a pre-specified technique (e.g., Mask R CNN technique) and classify the objects on an instance basis. Additionally, the first image processing unit may classify classes of objects.

The Mask R CNN technique described above is a technology proposed in 2017 and is one of the best algorithms for recognizing instance object technology in images, and all pixels corresponding to instance objects can be found in the image. .

Here, there are various objects that need to be detected in the above-mentioned image (e.g., frame image), and among these, objects that have the same properties are grouped and defined and are called classes. And the form of the class is variable depending on the application to which it is applied. At this time, depending on the type of class defined, it may be the same object or a different object.

For example, when there is a sky class, isolated sky areas may appear in two or more places within an image (e.g. frame image). In this case, these areas are the same object. will be. Also, in cases where objects are different, such as vegetables, but the boundaries are vague and difficult to distinguish, only classes are defined without distinguishing between objects. However, when defining classes such as traffic signs, cars, pedestrians, and bicycles, each of these objects has clear boundaries and it is important information to distinguish them depending on the time. Objects are essentially distinguished.

At this time, each object classified as described above may be referred to as an instance object.

FIG. 4 is a diagram for explaining a morphology operation on a sector detection result according to an embodiment.

Referring to FIG. 4 , in one example, the first image processor may remove background noise by performing a morphology operation on the sector detection result. Here, the morphology operation, for example, 3 erosions and 3 dilations can be processed using a 3x3 kernel. The first image processor may remove background noise through erosion and increase the reduced suboptimal pixels again through dilation.

FIG. 5 is a diagram illustrating how a second image processor detects a lane by processing image data according to an embodiment.

Referring to FIG. 5, the second image processor may divide the image data into a predetermined number of grids and detect a grid that is determined to contain a lane in the image data. Accordingly, in FIG. 3, it can be seen that the second image processing unit detects a grid including the left lane and the right lane.

Row-wise classification methods can be used to divide into such grids and detect lanes. Specifically, the row-wise classification method may refer to a grid-based classification method, and can divide the incoming image into a certain number of grids and then classify whether each grid is a lane or not.

The lane detection unit may detect the lane based on the grid containing the lane and the sector containing the lane.

FIG. 6 is a diagram illustrating dividing image data into a plurality of y-axis coordinates with equal intervals, according to an embodiment.

Referring to FIG. 6, the lane detection unit may set a plurality of y-axis coordinates at equal intervals in the image data as fixed coordinates. That is, as shown in FIG. 6, a plurality of horizontal lines that equally divide the image data can be set. Here, the y-axis coordinate can map image data to an x-y window consisting of coordinates with x and y values. Accordingly, a specific sector included in image data can be expressed as specific coordinates with x and y values.

The plurality of horizontal lines described above may be an evaluation method of TuSimple Dataset. Additionally, when only a portion of the lane is detected in the image data, the lane detection unit may not set the y-axis fixed coordinates to the non-detected area. That is, the lane detection unit may set a plurality of horizontal lines with equal intervals in areas where lanes are detected, and may not set horizontal lines in areas where lanes are not detected.

Referring to FIG. 6, in one example, the lane detection unit may set 56 y-axis fixed coordinates at intervals of 10 at the bottom of the image data.

FIG. 7 is a diagram illustrating how a first image processor detects a lane line from image data according to an embodiment.

Referring to FIG. 7, the first image processing unit can distinguish between when an occlusion phenomenon in which a lane is not detected occurs (b) and when it does not occur (a). Here, the occlusion phenomenon may mean a phenomenon in which a lane is obscured by a specific object.

Figure 8 is a diagram for explaining types of occlusion phenomenon according to one embodiment.

Since this phenomenon can occur frequently in the driving environment of the own vehicle, when a blockage phenomenon occurs in a detected lane when detecting a lane, it can be classified into a plurality of cases. Referring to FIG. 8 , for example, when a blockage phenomenon 810 occurs, the lane detection unit determines the top (a), middle (b), and bottom ( Bottom)(c), etc. Additionally, the lane detection unit may detect the lane by performing post-processing based on this classification.

In order to classify cases for the above-described occlusion phenomenon 810, the lane detection unit sets as the core coordinate a coordinate having a y-axis coordinate that overlaps the fixed coordinate among the coordinates included in the sector containing the lane and the grid containing the lane, , lanes can be detected based on key coordinates.

Specifically, the lane detection unit may calculate coordinates included in the grid included in the lane.

FIG. 9 is a diagram illustrating how a second image processor detects a lane line from image data according to an embodiment.

Referring to FIG. 9, the lane detection unit may calculate coordinates having at least one x value and one y value from the grid detected by the second image processing unit. Accordingly, when n grids are detected, the lane detection unit can calculate a total of n or more coordinates in each grid. Here, the coordinate value calculated from each grid can be calculated as the center value of each grid, and is not limited to a specific standard as long as the calculated coordinate value can be included in the detected grid.

Figure 10 is a diagram for explaining calculating core coordinates and candidate coordinates according to an embodiment.

Referring to FIG. 10, the lane detection unit determines the grid-calculated coordinates detected by the second image processor in the x-axis range a of the pixels corresponding to the preset y-axis fixed coordinates among the pixels detected by the first image processor. If you have a coordinate and it is included in the x-axis range a, that coordinate can be set as the core coordinate (910).

In addition, the first image processing unit may further extract adjacent sectors that exist within a predetermined distance from the sector containing the lane, and the lane detection unit may detect the coordinates included in the above-described grid in the x-axis range (b) of the adjacent sectors. If there is a fixed y-axis coordinate and the x-coordinate is included in the b range of FIG. 10, the corresponding coordinate can be set as a candidate coordinate.

In one example, the lane detection unit may determine that there is no lane if the core coordinates calculated as described above are less than a predetermined number. For example, if the calculated core coordinates are four or less, the lane detection unit may determine that there is no lane in the lane in which the host vehicle is traveling.

FIG. 11 is a diagram illustrating classification of detected lanes into three cases according to an embodiment.

The lane detection unit performs adaptive position comparison between core coordinates and candidate coordinates and classifies candidate coordinates into top, middle, and bottom, based on the positions of the core coordinates and candidate coordinates included in each boundary of the classified top, middle, and bottom. You can detect lanes with

Referring to FIG. 11, for example, the lane detection unit compares the y-axis value between the core coordinate a at the bottom of the y-axis and the candidate coordinate b among the core coordinates, and if the candidate coordinate is located below the core coordinate, starting from the core coordinate a. Candidate coordinates including the key coordinates at the top of the y-axis can be classified as middle c, and candidate coordinates including from the candidate coordinate b to the bottom of the y-axis can be classified as bottom d.

Likewise, the lane detection unit compares the y-axis value between the core coordinate at the top of the y-axis and the candidate coordinate among the core coordinates, and if the candidate coordinate is located above the core coordinate, the candidate coordinates included from the core coordinate to the bottom core coordinate on the y-axis are compared. As a stop, the candidate coordinates included from the corresponding sign coordinate to the top candidate coordinate of the y-axis can be classified into the top.

Here, the top and bottom y-value coordinates of the core coordinates may be selected from among the core coordinates included in the core coordinate area.

In one example, the core coordinates included in the core coordinate area may each be located within a predetermined distance from the nearest core coordinate.

In another example, when there are multiple core coordinate areas, the lane detection unit may classify the candidate coordinates detected from the image data into top, middle, and bottom based on the core coordinate area containing the most core coordinates.

Figure 12 is a diagram for explaining post-processing of candidate coordinates according to an embodiment.

If the angle between the first straight line passing through the lowest coordinate and the lowest coordinate among the coordinates included in the middle stop and the second straight line passing through the highest coordinate among the lowest coordinates and the bottom coordinate is less than or equal to the reference angle, the first straight line And the coordinates included in the second straight line can be set as key coordinates.

Referring to FIG. 12 , for example, the lane detection unit detects a first straight line 1210 through which the bottom core coordinate b and the bottom core coordinate c among the coordinates included in middle a and the coordinates included in the bottom core coordinate b and bottom d. The included angle formed by the second straight line 1220 passing through the top candidate coordinate e is calculated, and if the calculated included angle is less than or equal to the reference angle, the candidate coordinate e can be set as the core coordinate.

Likewise, the lane detection unit detects when the angle between the straight line passing through the top key coordinate and the next top core coordinate among the coordinates included in the middle and the straight line passing through the lowest candidate coordinate among the coordinates included in the top and the top core coordinate described above is less than the reference angle. , the lowest candidate coordinates can be set as the core coordinates.

And, for the candidate coordinates included in the interruption, the lane detection unit is based on a straight line passing through the core coordinate located at the shortest distance from the center of the interruption and the core coordinate located at the second closest distance, and a straight line to adjacent coordinates such as the top or bottom. Candidate coordinates can be set as key coordinates according to the included angle by comparing . Here, the center of the break may be the middle value between the top value of the y-axis and the bottom value of the y-axis. And, the closest distance from the center may be the y-axis distance.

The above-mentioned included angle can be calculated through [Equation 1] below.

[Equation 1]

FIG. 13 is a diagram illustrating performing a spline fitting algorithm on key coordinates according to an embodiment.

Referring to FIG. 13, the lane detection unit may correct the positions of key coordinates calculated by performing a spline fitting algorithm.

This spline fitting algorithm can be performed through Equation 2 below.

[Equation 2]

In one example, the lane detection device may detect a key coordinate area including key coordinates as a lane.

Accordingly, the lane detection device according to the present disclosure can compare data through a plurality of methods for detecting lanes, thereby reducing errors and shortening post-processing time.

This lane detection device can be implemented with an electronic control unit (ECU), microcomputer, etc.

In one embodiment, a computer system (not shown), such as a lane detection device, may be implemented as an electronic control unit. The electronic control unit may include at least one of one or more processors, memory, storage, user interface input, and user interface output, which may communicate with each other through a bus. Additionally, the computer system may also include a network interface for connecting to a network. A processor may be a CPU or a semiconductor device that executes processing instructions stored in memory and/or storage. Memory and storage may include various types of volatile/non-volatile storage media. For example, memory may include ROM and RAM.

In one example, below is exemplary coding data that can carry out all of the foregoing disclosure.

# Input: Segmentation S, Row-wise classification C,

# Coordinates Threshold CT, Angle Threshold AT

# Output: Coordinates of lanes

# Morphology calculation for noise removal

morph_map <- dilate(erode(S))

for x, y in points do

S_guide <- where(S==255)

if len(S_guide) == 0

Candidate <- (x,y)

Continue

# Calculate the x-coordinate range of the S result for AND operation

max_width <- max(S_guide_width) + margin

min_width <- min(S_guide_width) - margin

# Classification of core coordinates and candidate coordinates through AND operation for x-coordinates of S and C

if AND(S, C) then

Candidate <- (x, y)

continue

Core <- (x, y)

else

if y not in S

Candidate <- (x, y)

# Number of core coordinates > Coordinate Threshold

if len(core) < CT

continue

# Divide candidate coordinates into Top, Mid, and Bottom through adaptive position comparison

if len(candidate) > 0

# Create a vector between the candidate coordinates of Top, Mid, Bottom and the core coordinates and calculate the inner product between the vectors

if candidate's y-coordinate < lowest y-coordinate of core

bottom_angle <- inner_product(vector1, vector2)

Coordinates of elif candidate > Highest y coordinate of core

top_angle <- inner_product(vector1, vector2)

else

mid_angle <- inner_product(vector1, vector2)

# If the angle of the candidate coordinates calculated through the inner product is greater than AT, the candidate coordinates of Top, Mid, and Bottom are classified as core coordinates.

if top_angle < AT or bottom_angle < AT or mid_angle < AT

core <- candidate

# Apply spline fitting for smooth position correction of key coordinates -> Final result is coordinate information of each lane

Coordinates of lane <- Spline_fitting(core)

Hereinafter, a lane detection method using a lane detection device that can perform all of the above-described present disclosure will be described.

Figure 14 is a flowchart explaining a lane detection method according to an embodiment of the present disclosure.

Referring to FIG. 14, the lane detection method of the present disclosure includes an image receiving step (S1410) of receiving image data in which the surroundings of the own vehicle are photographed, and a first image detecting a sector determined to contain a lane based on the image data. Processing step (S1420), dividing the image data into a predetermined number of grids, and detecting a grid that is determined to contain a lane in the image data (S1430), and a grid containing the lane And it may include a lane detection step (S1440) in which a lane is detected based on the sector containing the lane.

In the lane detection step (S1440), a plurality of y-axis coordinates having equal intervals in the image data are set as fixed coordinates, and a y-axis coordinate that overlaps the fixed coordinate among the coordinates included in the sector containing the lane and the grid containing the lane Coordinates with are set as core coordinates, and lanes can be detected based on the core coordinates.

The first image processing step (S1420) may further extract adjacent sectors that exist within a predetermined distance from the sector containing the lane.

In the lane detection step (S1440), among the coordinates included in the grid containing adjacent sectors and lanes, coordinates with y-axis coordinates overlapping with fixed coordinates are calculated as candidate coordinates, and lanes are detected based on the core coordinates and candidate coordinates. You can.

The lane detection step (S1440) performs an adaptive position comparison between the core coordinates and the candidate coordinates to classify the candidate coordinates into top, middle, and bottom, and the core coordinates and candidate coordinates included in each boundary of the classified top, middle, and bottom. Lanes can be detected based on their locations.

The lane detection step (S1440) is performed when the angle between the first straight line passing through the lowest coordinate and the lowest coordinate among the coordinates included in the stop and the second straight line passing through the highest coordinate among the lowest coordinate and the bottom coordinate is less than or equal to the reference angle. , the coordinates included in the first straight line and the second straight line can be set as core coordinates, and the core coordinate area including the core coordinates can be detected as a lane.

As described above, according to the present disclosure, the lane detection apparatus and method can improve accuracy by correcting detection results through a plurality of detection methods, and can improve post-processing speed by considering other detection data.

The above description is merely an illustrative explanation of the technical idea of the present disclosure, and those skilled in the art will be able to make various modifications and variations without departing from the essential characteristics of the present disclosure. In addition, the present embodiments are not intended to limit the technical idea of the present disclosure, but rather to explain it, so the scope of the present technical idea is not limited by these embodiments. The scope of protection of this disclosure should be interpreted in accordance with the claims below, and all technical ideas within the equivalent scope should be interpreted as being included in the scope of rights of this disclosure.

10: lane detection device 110: image receiver
120: first image processing unit 130: second image processing unit
140: Lane detection unit

Claims (12)

an image receiver that receives image data captured around the vehicle;
a first image processing unit that detects a sector determined to contain a lane based on the image data;
a second image processor that divides the image data into a predetermined number of grids and detects a grid that is determined to contain a lane in the image data; and
A lane detection unit that detects the lane based on a sector containing the lane and a grid containing the lane,
The first image processing unit,
Further extracting adjacent sectors that exist within a predetermined distance from the sector containing the lane,
The lane detection unit,
In the image data, a plurality of y-axis coordinates having the same spacing are set as fixed coordinates, and a coordinate having a y-axis coordinate overlapping with the fixed coordinate among coordinates included in a sector including the lane and a grid including the lane is set as the core coordinate,
Among the coordinates included in the grid including the adjacent sector and the lane, a coordinate having a y-axis coordinate overlapping with the fixed coordinate is set as a candidate coordinate,
If the set key coordinates are less than a predetermined number, it is determined that there is no lane for driving,
Perform adaptive position comparison between the key coordinates and the candidate coordinates to classify the candidate coordinates into top, middle, and bottom,
Among the core coordinates, the y-axis value is compared between the y-axis bottom core coordinate and the candidate coordinate adjacent to the y-axis bottom core coordinate, and if the candidate coordinate is located below the y-axis bottom core coordinate, the y-axis starts from the bottom core coordinate of the y-axis. Candidate coordinates including the top core coordinate are classified as middle, and candidate coordinates including from the candidate coordinate adjacent to the bottom core coordinate on the y-axis to the bottom candidate coordinate on the y-axis are classified as bottom.
Among the core coordinates, the y-axis value is compared between the key coordinate at the top of the y-axis and the candidate coordinate adjacent to the core coordinate at the top of the y-axis, and if the candidate coordinate is located above the core coordinate at the top of the y-axis, the core coordinate at the top of the y-axis to the bottom of the y-axis Candidate coordinates that include the core coordinates are classified as middle, and candidate coordinates that are included from the candidate coordinate adjacent to the core coordinate at the top of the y-axis to the candidate coordinate at the top of the y-axis are classified as top.
A lane detection device that detects the lane based on the positions of the core coordinates and the candidate coordinates included in each classified boundary of the top, middle, and bottom. delete delete delete According to paragraph 1,
The lane detection unit,
If the angle formed between the first straight line passing through the lowest coordinate and the second lowest coordinate among the coordinates included in the middle stop and the second straight line passing through the highest coordinate among the lowest coordinate and the bottom coordinate is less than or equal to the reference angle, the first straight line And set the coordinates included in the second straight line as key coordinates,
A lane detection device that detects a core coordinate area including the core coordinates as the lane. According to paragraph 1,
The first image processing unit,
A lane detection device that removes background noise by performing a morphology operation on the sector detection result. An image receiving step of receiving image data captured around the vehicle;
A first image processing step of detecting a sector determined to contain a lane based on the image data;
A second image processing step of dividing the image data into a predetermined number of grids and detecting a grid that is determined to contain a lane in the image data; and
A lane detection step of detecting the lane based on a sector containing the lane and a grid containing the lane,
The first image processing step is,
Further extracting adjacent sectors that exist within a predetermined distance from the sector containing the lane,
The lane detection step is,
In the image data, a plurality of y-axis coordinates having the same spacing are set as fixed coordinates, and a coordinate having a y-axis coordinate overlapping with the fixed coordinate among coordinates included in a sector including the lane and a grid including the lane is set as the core coordinate,
Among the coordinates included in the grid including the adjacent sector and the lane, a coordinate having a y-axis coordinate overlapping with the fixed coordinate is set as a candidate coordinate,
If the set key coordinates are less than a predetermined number, it is determined that there is no lane for driving,
Perform adaptive position comparison between the key coordinates and the candidate coordinates to classify the candidate coordinates into top, middle, and bottom,
Among the core coordinates, the y-axis value is compared between the y-axis bottom core coordinate and the candidate coordinate adjacent to the y-axis bottom core coordinate, and if the candidate coordinate is located below the y-axis bottom core coordinate, the y-axis starts from the bottom core coordinate of the y-axis. Candidate coordinates including the top core coordinate are classified as middle, and candidate coordinates including from the candidate coordinate adjacent to the bottom core coordinate on the y-axis to the bottom candidate coordinate on the y-axis are classified as bottom.
Among the core coordinates, the y-axis value is compared between the key coordinate at the top of the y-axis and the candidate coordinate adjacent to the core coordinate at the top of the y-axis, and if the candidate coordinate is located above the core coordinate at the top of the y-axis, the core coordinate at the top of the y-axis to the bottom of the y-axis Candidate coordinates that include the core coordinates are classified as middle, and candidate coordinates that are included from the candidate coordinate adjacent to the core coordinate at the top of the y-axis to the candidate coordinate at the top of the y-axis are classified as top.
A lane detection method for detecting the lane based on the positions of the key coordinates and the candidate coordinates included in each classified boundary of the top, middle, and bottom. delete delete delete In clause 7,
The lane detection step is,
If the angle formed between the first straight line passing through the lowest coordinate and the second lowest coordinate among the coordinates included in the middle stop and the second straight line passing through the highest coordinate among the lowest coordinate and the bottom coordinate is less than or equal to the reference angle, the first straight line And set the coordinates included in the second straight line as key coordinates,
A lane detection method for detecting a core coordinate area including the core coordinates as the lane. In clause 7,
The first image processing step is,
A lane detection method that removes background noise by performing a morphology operation on the sector detection result.

Images