CN109190483A - A kind of method for detecting lane lines of view-based access control model - Google Patents

Abstract

The invention proposes a vision-based lane line detection method. The invention collects images through a camera and converts them into grayscale images, sets the center point of the grayscale image as a reference point to demarcate the region of interest; and extracts the rising edge point and the falling edge point in the region of interest through the line scanning gradient value method. , the rising edge point and the falling edge point are respectively obtained through inverse perspective transformation to obtain the inverse perspective rising edge point and the inverse perspective falling edge point, and the filtered rising edge point and the filtered falling edge point are obtained respectively according to the characteristic filtering of the lane width; Perform custom parameter space transformation for edge points and falling edge points, and count the number of line angles and lateral offsets of the filtered ascending edge points and the filtered descending edge points that are equal, and fit the lane curve to construct the current frame road The lane line position of the image; the lane line position of the next frame of road image is associated with the lane line position of the current frame of road image.

Description

A Vision-Based Lane Line Detection Method

technical field

The invention relates to the technical field of signal processing, in particular to a vision-based lane line detection method.

Background technique

The primary reason for the frequent occurrence of accidents is that drivers unintentionally deviate from their lanes while driving. With the increasing number of cars year by year, the road environment is becoming more and more complex. In order to allow people to travel more efficiently, autonomous driving or assisted driving has gradually become one of the research hotspots. Lane line recognition is the most basic part of it. It is not only a key unit to ensure the effective control of the road scene by the vehicle, but also provides accurate location information for the vehicle departure warning technology, thereby reducing traffic pressure and reducing traffic accidents to a certain extent. The realization of lane line recognition and departure warning under the embedded system has the characteristics of low cost, low power consumption, miniaturization and easy integration, so it has high practical value and broad application prospects.

With the development of computer technology, systems that provide timely warnings to drivers in danger have great potential to save a large number of lives. The system designed to help the driver while driving is called advanced driver assistance, and it has adaptive cruise control, collision avoidance, blind spot detection and traffic sign detection, among many other features. Lane departure systems also fall into this category. Lane detection is locating lane markings on the road and presenting this location information to intelligent systems. In an intelligent transportation system, intelligent vehicles are integrated with intelligent infrastructure to provide a safer environment and better traffic conditions.

At present, the methods of lane line detection can be mainly divided into three categories: model-based methods, feature-based methods, and sign-based methods. The feature method uses the basic features of the lane, such as color, texture, etc., to locate the lane lines in the road surface image. Model-based methods typically analyze lanes with linear or curvilinear templates, and detection is simpler once the model is defined. Based on the deep learning method, the basic principle is to mark a large number of sample sets in advance, and use the convolutional neural network method to train the sample set to obtain network parameters, so as to achieve the purpose of lane detection and classification. Compared with the deep learning method, the road information detection of the feature and model method is necessary. The model and feature method can be used as a reference, and the deep learning method can be used to identify the lane line more accurately. Due to the limited hardware conditions, the model-to-feature approach also has great potential.

For the lane line detection algorithm, there are the following problems: achieve real-time detection. The road conditions are complex, such as occlusion, lack of lane lines, ground markings, tunnels, etc., which make the detection rate low. Since there is no significant change in the position information of the lane lines in consecutive multiple frames except for lane changes, it is necessary to stably detect the lanes of the multi-frame images, so that the interference lane lines and the accurate lane lines will not be exchanged all the time. Lane line position information needs to be accurately detected before changing lanes to provide correct guidance for departure warning. Based on the above situation, how to improve efficient, real-time and stable detection of lane lines is a problem to be solved in this field.

SUMMARY OF THE INVENTION

The embedded platform used in this method has designed an efficient and real-time lane line detection algorithm, and has strong adaptability. The purpose is to solve the problem of low detection efficiency of lane lines in the prior art.

In order to achieve the above purpose, the present invention proposes a vision-based lane line detection method, which includes the following steps:

Step 1: collect an image through a camera, convert the collected image into a grayscale image, set the center point of the grayscale image as a reference point, and delineate a region of interest according to the reference point;

Step 2: Extract the rising edge point and the falling edge point in the region of interest by the line scanning gradient value method, respectively obtain the rising edge point and the falling edge point of the inverse perspective through the inverse perspective transformation. The inverse perspective rising edge point and the inverse perspective descending edge point use the lane width feature filtering to obtain the filtered ascending edge point and the filtered descending edge point respectively;

Step 3: Perform a self-defined parameter space transformation on the rising edge point after screening and the falling edge point after screening, and count the number of the same line angle and horizontal offset of the rising edge point after screening and the falling edge point after screening. Obtain candidate lane lines and fit lane curves;

Step 4: Associate the lane line position of the next frame of road image with the lane line position of the current frame of road image.

Preferably, the width of the collected image in step 1 is u, and the height is v;

The center point of the grayscale image in step 1 is Set the center point as the fiducial point of the grayscale image;

The delineation of the region of interest according to the reference point in step 1 is:

According to the reference point Delineate a rectangular box, the value range of the width of the rectangular box is The value range of the height of the rectangular block is

Among them, the value range of w is The value range of h is

Preferably, the extraction of edge points in the region of interest by the line scanning gradient value method described in step 2 is:

The calculation is based on each pixel edge intensity on the horizontal line scan line:

Among them, I(i+k,j) represents the pixel value of the i+kth row and the jth column of the region of interest, the number of image lines representing the region of interest, Represents the number of image columns in the region of interest, and L represents the filter length of each row;

Compare the pixel edge intensity with the first threshold and the second threshold respectively, and classify the pixel points in the region of interest according to the detection results: when E(i,j)>Th ₁ , I(i,j) is the rising edge point , when E(i,j)<Th ₂ , I(i,j) is the falling edge point;

Convert the rising edge points and falling edge points in the region of interest to the edge feature points in the actual road under the world coordinate system through inverse perspective transformation, that is, the inverse perspective rising edge points and inverse perspective falling edge points described in step 2;

Inverse perspective rising edge points and inverse perspective descending edge points Use lane width feature filtering to remove interference points, and calculate the Euclidean distance for the inverse perspective rising edge points and inverse perspective descending edge points of the same number of image lines in the region of interest: if |dis-D |≤dh, dis is the Euclidean distance, D is the distance threshold, and dh is the distance error, then the inverse perspective rising edge point is the rising edge point after screening described in step 2:

(x _m ,y _m )

in, M is the number of rising edge points after screening;

And the inverse perspective descending edge point is the descending edge point after screening described in step 2:

in, N is the number of falling edge points after screening;

Preferably, the custom parameter space transformation described in step 3 is:

The customized parameter space of the rising edge point after filtering is:

x _m =p _k,m +y _m *tanθ _k,m

Among them, (x _m , y _m ) are the coordinates of the rising edge point after screening described in step 2, θ _k,m is the angle of the rising edge point line after screening, and θ _k,m ∈[α,β], k ∈[1,K], K represents the angle number of the rising edge point line, p _k,m represents the lateral offset of the rising edge point line, perform the traversal calculation of θ _k,m to obtain the corresponding p _k,m ;

The customized parameter space of the descending edge point after filtering is:

in, are the coordinates of the falling edge point after screening as described in step 2, represents the angle of the falling edge point line, and L represents the angle number of the falling edge point line, Indicates the lateral offset of the falling edge point line, traversal calculation to obtain the corresponding

In step 3, count the number of the rising edge points after screening and the falling edge points after screening whose line angles and lateral offsets are equal:

In the custom parameter space of the rising edge point after screening, compare the angle of the rising edge point line and the horizontal offset of the rising edge point line of any two different rising edge points after screening. If both are equal, then :

H _r (p, θ)=H _r (p, θ)+1 r∈[1,N _r ]

Among them, H _r (p, θ) is the number of rising edge points after screening, the angle of the rth group of rising edge point lines and the horizontal offset of the rising edge point line are equal;

In the customized parameter space of the descending edge point after filtering, compare the angle of the descending edge point line and the horizontal offset of the descending edge point line of any two different descending edge points after filtering. If both are equal, then :

H _d (p, θ)=H _d (p, θ)+1 d∈[1,N _d ]

Among them, H _d (p, θ) is the number of descending edge points after screening where the angle of the d-th group of descending edge point lines and the lateral offset of the descending edge point lines are equal;

Among the upper edge points after screening where the angle of the rising edge point line and the lateral offset of the rising edge point line in the N _r groups are equal, select one of the first G groups in the order of H _r (p, θ) values from high to low. Group

(p _g ,θ _g )g∈[1,G], different (p _g ,θ _g ) are represented as different straight lines according to the angle of the rising edge point line and the lateral offset value of the rising edge point line;

Among the upper edge points after screening where the angle of the falling edge point line and the lateral offset of the falling edge point line in the N _d groups are equal, select one of the first G groups in the order of H _d (p, θ) values from high to low. Group

different It is expressed as different straight lines according to the angle of its falling edge point line and the lateral offset value of the falling edge point line;

The candidate lane lines obtained in step 3 are:

For the rising edge point, the straight line determined by the parameter value (p _g ,θ _g )g∈[1,G] is:

x _i =p _g +y _i *tanθ _g

in, The specific value of x _i is calculated by the formula of the straight line, ( _xi , y _i ) is the coordinate of the straight line, and based on ( x _i , y _i ), the obtained filtered edge points are further screened by external expansion, and only retain Rising edge point within the extended range at the same time δ and φ are the set thresholds;

The same processing is performed for the falling edge point and the rising edge point, and only the falling edge point within the extended range is retained

The fitted lane lines described in step 3 are:

The ascending edge points within the expansion range are fitted multiple times to obtain the ascending fitted lane curve, and the parameter value is

The descending edge points within the expansion range are fitted multiple times to obtain the descending fitted lane curve, and the parameter value is

Multi-term fitting can use least squares method or Bezier curve method;

The ascending fitted lane curve and the descending fitted lane curve constitute the lane line position of the current frame road image;

Preferably, the association described in step 4 is:

The rising fitted lane curve in the lane line position of the next frame of road image, the parameter value is

Descending the fitted lane curve in the lane line position of the next frame of road image, the parameter value is

like α, β are the set thresholds, γ, λ are the set thresholds,

The lane line position of the next frame of road image is valid, otherwise it is invalid;

like α, β are the set thresholds, γ, λ are the set thresholds,

The lane line position of the next frame of road image is valid, otherwise it is invalid.

The method of the invention can detect the lane lines in each scene accurately in real time, and has quite good anti-interference performance. At the same time, combined with the lane structure feature information, and looking for breakthrough points in real-time, a self-defined parameter space transformation detection algorithm was proposed, and it was applied to the embedded environment. Use vanishing point to automatically delineate the region of interest, avoid complex calculation of the whole image, eliminate redundant information to improve detection efficiency, extract edge gradient values based on line scanning, extract edge points for fast inverse perspective transformation, and then inverse The feature fusion of the edge points in the perspective transformed road map and the edge points extracted from the original image scan is performed to eliminate the interference points and retain only the effective edge feature points, which provides a guarantee for efficient parameter space transformation; after the effective edge information is obtained, the A custom parameter space transformation method suitable for edge points is used to obtain candidate lane lines, and then the lane lines are screened by feature information, and the obtained lane lines are used to achieve stable detection of subsequent frame images.

Description of drawings

Fig. 1: method flow schematic diagram of the present invention;

Figure 2: Schematic diagram of the setting of the region of interest and the scan line in the embodiment of the lane line detection algorithm of the present invention;

Fig. 3: The bird's-eye view of edge point extraction result graph and inverse perspective transformation in the embodiment of the lane line detection algorithm of the present invention;

Fig. 4: The result diagram of the width feature filtering in the embodiment of the lane line detection algorithm of the present invention;

Figure 5: Schematic diagram of the parameter space in the embodiment of the lane line detection algorithm of the present invention;

FIG. 6 is a result diagram of the inner boundary of the lane line after integration in the embodiment of the lane line detection algorithm of the present invention.

FIG. 7 is a diagram of the result of lane line detection in the embodiment of the lane line detection algorithm of the present invention.

Detailed ways

In order to facilitate the understanding and implementation of the present invention by those of ordinary skill in the art, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the embodiments described herein are only used to illustrate and explain the present invention, but not to limit it. this invention.

Embodiments of the present invention are described below in conjunction with FIGS. 1 to 7 , which specifically include the following steps:

Step 1: collect an image through a camera, convert the collected image into a grayscale image, set the center point of the grayscale image as a reference point, and delineate a region of interest according to the reference point;

In step 1, the width of the collected image is u=1280, and the height is v=720;

The center point of the grayscale image in step 1 is Set the center point as the fiducial point of the grayscale image;

The delineation of the region of interest according to the reference point in step 1 is:

According to the reference point Delineate a rectangular box, the value range of the width of the rectangular box is The value range of the height of the rectangular block is

Among them, the value range of w is The value range of h is

Step 2: Extract the rising edge point and the falling edge point in the region of interest by the line scanning gradient value method, respectively obtain the rising edge point and the falling edge point of the inverse perspective through the inverse perspective transformation. The inverse perspective rising edge point and the inverse perspective descending edge point use the lane width feature filtering to obtain the filtered ascending edge point and the filtered descending edge point respectively;

The extraction of edge points in the region of interest by the line scanning gradient value method described in step 2 is:

The calculation is based on each pixel edge intensity on the horizontal line scan line:

Among them, I(i+k,j) represents the pixel value of the i+kth row and the jth column of the region of interest, the number of image lines representing the region of interest, Represents the number of image columns in the region of interest, L represents the filter length of each row, L=8;

Compare the pixel edge intensity with the first threshold and the second threshold respectively, and classify the pixel points in the region of interest according to the detection results: when E(i,j)>Th ₁ , I(i,j) is the rising edge point , Th ₁ =16, when E(i,j)<Th ₂ , I(i,j) is the falling edge point, Th ₂ =-16;

Convert the rising edge points and falling edge points in the region of interest to the edge feature points in the actual road under the world coordinate system through inverse perspective transformation, that is, the inverse perspective rising edge points and inverse perspective falling edge points described in step 2;

Inverse perspective rising edge points and inverse perspective falling edge points Use lane width feature filtering to remove interference points, and calculate the Euclidean distance for the inverse perspective rising edge points and inverse perspective falling edge points of the same image line in the region of interest: if |dis-D |≤dh, dis is the Euclidean distance, D is the distance threshold, which is a distance of 14 pixels, and dh is the distance error, which is a distance of 4 pixels, then the inverse perspective rising edge point is the rising edge point after screening described in step 2:

(x _m ,y _m )

in, M is the number of rising edge points after screening;

And the inverse perspective descending edge point is the descending edge point after screening described in step 2:

in, N is the number of falling edge points after screening;

Step 3: Perform a self-defined parameter space transformation on the rising edge point after screening and the falling edge point after screening, and count the number of the same line angle and horizontal offset of the rising edge point after screening and the falling edge point after screening. Obtain candidate lane lines and fit lane curves;

The custom parameter space transformation described in step 3 is:

The customized parameter space of the rising edge point after filtering is:

x _m =p _k,m +y _m *tanθ _k,m

Among them, (x _m , y _m ) are the coordinates of the rising edge point after screening described in step 2, θ _k,m is the angle of the rising edge point line after screening, and θ _k,m ∈[α,β], k ∈[1,K], α=1, β=75, K represents the angle number of the rising edge point line, p _k,m represents the lateral offset of the rising edge point line, perform the traversal calculation of θ _k,m to obtain the corresponding _pk,m of ;

The customized parameter space of the descending edge point after filtering is:

in, are the coordinates of the falling edge point after screening as described in step 2, represents the angle of the falling edge point line, and α=1, β=75, L represents the angle number of the falling edge point line, Indicates the lateral offset of the falling edge point line, traversal calculation to obtain the corresponding

In step 3, count the number of the rising edge points after screening and the falling edge points after screening whose line angles and lateral offsets are equal:

In the custom parameter space of the rising edge point after screening, compare the angle of the rising edge point line and the horizontal offset of the rising edge point line of any two different rising edge points after screening. If both are equal, then :

H _r (p, θ)=H _r (p, θ)+1 r∈[1,N _r ]

Among them, H _r (p, θ) is the number of rising edge points after screening, the angle of the rth group of rising edge point lines and the horizontal offset of the rising edge point line are equal;

In the customized parameter space of the descending edge point after filtering, compare the angle of the descending edge point line and the horizontal offset of the descending edge point line of any two different descending edge points after filtering. If both are equal, then :

H _d (p, θ)=H _d (p, θ)+1 d∈[1,N _d ]

Among them, H _d (p, θ) is the number of descending edge points after screening where the angle of the d-th group of descending edge point lines and the lateral offset of the descending edge point lines are equal;

Among the upper edge points after screening where the angle of the rising edge point line and the lateral offset of the rising edge point line in the N _r groups are equal, select one of the first G groups in the order of H _r (p, θ) values from high to low. Group

(p _g ,θ _g )g∈[1,G], G=10, different (p _g ,θ _g ) are expressed as different according to the angle of the rising edge point line and the lateral offset value of the rising edge point line straight line;

Among the upper edge points after screening where the angle of the falling edge point line and the lateral offset of the falling edge point line in the N _d groups are equal, select one of the first G groups in the order of H _d (p, θ) values from high to low. Group

different It is expressed as different straight lines according to the angle of its falling edge point line and the lateral offset value of the falling edge point line;

The candidate lane lines obtained in step 3 are:

For the rising edge point, the straight line determined by the parameter value (p _g ,θ _g )g∈[1,G] is:

x _i =p _g +y _i *tanθ _g

in, The specific value of x _i is calculated by the formula of the straight line, ( _xi , y _i ) is the coordinate of the straight line, and based on ( x _i , y _i ), the obtained filtered edge points are further screened by external expansion, and only retain Rising edge point within the extended range at the same time delta and is the set threshold, δ=3,

The same processing is performed for the falling edge point and the rising edge point, and only the falling edge point within the extended range is retained

The fitted lane lines described in step 3 are:

The ascending edge points within the expansion range are fitted multiple times to obtain the ascending fitted lane curve, and the parameter value is

The descending edge points within the expansion range are fitted multiple times to obtain the descending fitted lane curve, and the parameter value is

Multi-term fitting can use least squares method or Bezier curve method;

The ascending fitted lane curve and the descending fitted lane curve constitute the lane line position of the current frame road image;

Step 4: Associate the lane line position of the next frame of road image with the lane line position of the current frame of road image;

The associations described in step 4 are:

The rising fitted lane curve in the lane line position of the next frame of road image, the parameter value is

Descending the fitted lane curve in the lane line position of the next frame of road image, the parameter value is

like α, β are the set thresholds, α=20, β=25, γ, λ are the set thresholds, γ=6, λ=7,

The lane line position of the next frame of road image is valid, otherwise it is invalid;

like α, β are the set thresholds, α=20, β=25, γ, λ are the set thresholds, γ=6, λ=7.

The lane line position of the next frame of road image is valid, otherwise it is invalid.

It should be understood that the parts not described in detail in this specification belong to the prior art.

It should be understood that the above description of the preferred embodiments is relatively detailed, and therefore should not be considered as a limitation on the protection scope of the patent of the present invention. In the case of the protection scope, substitutions or deformations can also be made, which all fall within the protection scope of the present invention, and the claimed protection scope of the present invention shall be subject to the appended claims.

Claims (5)

1. A lane line detection method based on vision is characterized by comprising the following steps:

step 1: acquiring an image through a camera, converting the acquired image into a gray image, setting a central point of the gray image as a reference point, and defining an interested area according to the reference point;

step 2: respectively extracting an ascending edge point and a descending edge point in the region of interest by a line scanning gradient value method, respectively obtaining an inverse perspective ascending edge point and an inverse perspective descending edge point by the ascending edge point and the descending edge point through inverse perspective transformation, and respectively obtaining a screened ascending edge point and a screened descending edge point by the inverse perspective ascending edge point and the inverse perspective descending edge point through lane width characteristic filtering;

and step 3: carrying out self-defined parameter space transformation on the screened rising edge points and the screened falling edge points, counting the number of the screened rising edge points and the screened falling edge points, wherein the angles and the transverse offsets of the lines of the screened rising edge points and the screened falling edge points are equal, obtaining candidate lane lines, and fitting a lane curve;

and 4, step 4: and correlating the lane line position of the next frame of road image by the lane line position of the current frame of road image.

2. The vision-based lane line detection method of claim 1, wherein: in the step 1, the width of the collected image is u, and the height of the collected image is v;

the central point of the gray level image in the step 1 isSetting the central point as a reference point of the gray image;

the step 1 of defining the region of interest according to the reference points comprises the following steps:

according to the reference pointDefining a rectangular square block with the width value range of the rectangular square block asThe height of the rectangular block is in the range of

Wherein, the value range of w ish has a value range of

3. The vision-based lane line detection method of claim 1, wherein: in the step 2, the extraction of the edge points in the region of interest by the line scanning gradient value method is as follows:

the calculation is based on the edge intensity of each pixel on the horizontal row of scan lines:

wherein I (I + k, j) represents the pixel values of the I + k th row and the j th column of the interested area,the number of image lines representing the region of interest,the number of image columns representing the region of interest, L representing the filter length of each row;

comparing the pixel edge strength with a first threshold and a second threshold respectively, and classifying the pixel points in the region of interest according to the detection result: when E (i, j) > Th₁When I (I, j) is a rising edge point, when E (I, j) < Th₂When I (I, j) is the falling edge point;

converting the rising edge point and the falling edge point in the region of interest into edge feature points in the actual road under a world coordinate system through inverse perspective transformation, namely the inverse perspective rising edge point and the inverse perspective falling edge point in the step 2;

the inverse perspective rising edge point and the inverse perspective falling edge point use lane width characteristic filtering to remove interference points, and the Euclidean distance is calculated for the inverse perspective rising edge point and the inverse perspective falling edge point of the same image line number in the region of interest: if | dis-D | is less than or equal to dh, dis is the Euclidean distance, D is the distance threshold, and dh is the distance error, the inverse perspective rising edge point is the rising edge point after the screening in the step 2:

(x_m,y_m)

wherein,m∈[1,M]m is the number of the rising edge points after screening;

and the inverse perspective descent edge points are the screened descent edge points in the step 2:

wherein,n∈[1,N]and N is the number of the edge points which are reduced after screening.

4. The vision-based lane line detection method of claim 1, wherein: the customized parameter space transformation in step 3 is:

the self-defined parameter space of the rising edge points after screening is as follows:

x_m＝p_k,m+y_m*tanθ_k,m

wherein (x)_m,y_m) Is the coordinate of the filtered rising edge point, θ in step 2_k,mIndicates the angle of the rising edge dotted line after screening, and theta_k,m∈[α,β]，k∈[1,K]K denotes the number of angles of the rising edge dot line, p_k,mRepresents the lateral offset of the rising edge point line, and is performed by_k,mGet corresponding p_k,m；

The self-defined parameter space of the screened descending edge points is as follows:

wherein,for the coordinates of the filtered falling edge points in step 2,indicates the angle of the falling edge point line, andl∈[1,L]l represents the angular number of the falling edge point line,indicating the lateral offset of the falling edge dot line, is carried outThe traversal calculation of (2) obtains corresponding

And 3, counting the number of the line angles and the transverse offsets of the screened rising edge points and the screened falling edge points which are equal to each other:

in the parameter space defined by the screened rising edge points, comparing the angles of the rising edge point lines of any two different screened rising edge points with the transverse offset of the rising edge point lines, if the angles are equal to each other, then:

H_r(p,θ)＝H_r(p,θ)+1r∈[1,N_r]

wherein H_r(p, theta) is the number of the screened rising edge points with equal angle of the r group of rising edge point lines and equal transverse offset of the rising edge point lines;

in the parameter space defined by the screened descending edge points, comparing the angles of the descending edge point lines of any two different screened descending edge points with the transverse offset of the descending edge point lines, if the angles are equal to each other, then:

H_d(p,θ)＝H_d(p,θ)+1d∈[1,N_d]

wherein H_d(p, theta) is the number of screened descending edge points with equal angles of the group d descending edge point lines and equal transverse offset of the descending edge point lines;

in N_rH is selected from the screened upper edge points with equal angles of the rising edge point lines and equal transverse offsets of the rising edge point lines_rOne of the first G groups is ordered by the (p, theta) value from high to low

(p_g,θ_g)g∈[1,G]Different (p)_g,θ_g) Different straight lines are represented according to the angle of the rising edge dotted line and the transverse offset value of the rising edge dotted line;

in N_dH is selected from the screened upper edge points with equal angles of the descending edge point lines and equal transverse offsets of the descending edge point lines_dOne of the first G groups is ordered by the (p, theta) value from high to low

Is differentDifferent straight lines are represented according to the angle of the descending edge dotted line and the transverse offset value of the descending edge dotted line;

in step 3, the obtaining of the candidate lane lines is as follows:

for rising edge points, by the parameter value (p)_g,θ_g)g∈[1,G]The straight line is determined as:

x_i＝p_g+y_i*tanθ_g

wherein,x_icalculating to obtain specific value (x) by using linear formula_i,y_i) Is the coordinate of a straight line, in (x)_i,y_i) Further screening the obtained screened edge points in an outward expansion mode for reference, and only keeping the rising edge points in an outward expansion range At the same timeδ and φ are set thresholds;

the same treatment is carried out on the falling edge point and the rising edge point, and only the falling edge point in the outward expansion range is reserved

The fitted lane line in the step 3 is as follows:

multiple fitting is carried out on the rising edge points in the outer expansion range to obtain a rising fitting lane curve, and the parameter value is

Multiple item fitting is carried out on the falling edge points in the outer expansion range to obtain a falling fitting lane curve, and the parameter value is

The multiple fitting can adopt a least square method or a Bezier curve method;

and the rising fitted lane curve and the falling fitted lane curve form the lane line position of the current frame road image.

5. The vision-based lane line detection method of claim 1, wherein: the association in step 4 is:

the rising fitting lane curve in the lane line position of the next frame of road image has the parameter value of

A lane curve is fitted in the position of the lane line of the next frame of road image in a descending way, and the parameter value is

If it isα is the threshold value set for the purpose of,gamma, lambda are set threshold values,

the lane line position of the next frame of road image is valid, otherwise, the lane line position is invalid;

if it isα is the threshold value set for the purpose of,gamma, lambda are set threshold values,

the lane line position of the next frame of road image is valid, otherwise, it is invalid.