A lane-line-based pitch angle calibration method and readable storage medium

Abstract

The present invention relates to the field of computer vision technology, and in particular to a pitch angle calibration method based on lane lines and a readable storage medium. The method comprises: setting n sampling levels, each sampling level corresponds to a different coordinate value of the lane line in the x direction; calculating the vanishing points between each pair of adjacent lane lines at each sampling level, and forming a vanishing point set; selecting the final vanishing point from the vanishing point set by median filtering; and calculating the corresponding pitch angle based on the final vanishing point. The present invention also discloses a readable storage medium. The pitch angle calibration method of the present invention can obtain a more stable vanishing point, thereby improving the accuracy of the pitch angle calibration of the vehicle-mounted camera.

Description

Technical Field

[0001] The present invention relates to the field of computer vision technology, and in particular to a lane line-based pitch angle calibration method and a readable storage medium. Background Art

[0002] In the field of autonomous driving, the change in the pitch angle of the on-board camera during driving will lead to inaccurate reconstruction of the lane line in the bird's-eye view, thereby increasing the risk of the vehicle running out of the lane. Therefore, how to quickly and effectively calibrate the pitch angle of the on-board camera has become an important research direction in this field.

[0003] At present, the prior art has conducted extensive research on the calibration of the pitch angle of the vehicle camera, among which the method of calculating the pitch angle of the vehicle based on the vanishing point is very mature. For example, the Chinese patent with the publication number CN110386142A discloses a pitch angle calibration method for an autonomous driving vehicle, which includes: determining the two-dimensional bounding box of the object and the perceived distance from the autonomous driving vehicle to the object based on the image of the object; in the image space of the object, based on the two-dimensional bounding box and the perceived distance from the autonomous driving vehicle to the object, searching for the dynamic pitch angle by at least modifying the previous external calibration value to calculate the distance value; determining the distance from the autonomous driving vehicle to the point on the ground of the autonomous driving vehicle's perspective map based on the searched dynamic pitch angle.

[0004] The pitch angle calibration method in the above-mentioned existing scheme assumes that the lane lines are parallel, and calculates the pitch angle based on the vanishing point of the parallel straight lines on the imaging screen and the line connecting the camera focus parallel to the road surface. The applicant found that in the pitch angle calibration process, the accuracy of the vanishing point directly affects the accuracy of the pitch angle calibration. However, the existing pitch angle calibration method needs to assume that the lane lines are parallel when obtaining the vanishing point. However, the lane lines perceived by the on-board camera are not always parallel, and some lane lines are curves with a certain curvature. Therefore, the vanishing point obtained by the existing pitch angle calibration method is not accurate, which leads to low accuracy of the pitch angle calibration. Therefore, how to design a pitch angle calibration method that can obtain a more stable vanishing point is a technical problem that needs to be solved urgently. Summary of the invention

[0005] In view of the above-mentioned deficiencies in the prior art, the technical problem to be solved by the present invention is: how to provide a pitch angle calibration method based on lane lines to obtain a more stable vanishing point, thereby improving the accuracy of the pitch angle calibration of the vehicle-mounted camera.

[0006] In order to solve the above technical problems, the present invention adopts the following technical solutions:

[0007] A lane line-based pitch angle calibration method comprises the following steps:

[0008] S1: Set n sampling levels, each sampling level corresponds to a different coordinate value of the lane line in the x direction;

[0009] S2: Calculate the vanishing points between each pair of adjacent lane lines at each sampling level and form a vanishing point set;

[0010] S3: Select the final vanishing point from the vanishing point set through median filtering;

[0011] S4: Calculate the corresponding pitch angle based on the final vanishing point.

[0012] Preferably, in step S1, the lane lines are first screened, and the lane lines whose lengths do not meet the requirements are deleted.

[0013] Preferably, in step S1, when the i-th lane line satisfies When , keep the i-th lane line; otherwise, delete the i-th lane line;

[0014] Among them, x imax Indicates the maximum coordinate of the i-th lane line in the x direction; Indicates the average value of the maximum x-coordinates of all lane lines; Δ x Indicates the lane line filter variable that is set.

[0015] Preferably, in step S1, based on the average value of the maximum x-coordinates of all lane lines Set the corresponding sampling level.

[0016] Preferably, in step S1, three sampling levels are set, namely, and The sampling level.

[0017] Preferably, in step S2, a vanishing point set is formed by the following steps:

[0018] S201: Initialize lane line number i=0, sampling level n=0;

[0019] S202: Take the starting point of lane number i and the coordinate value of sampling level n in the x direction to form a straight line l ni , the lane line number is i+1, and the starting point and the coordinate value of sampling level n in the x direction form a straight line l ni+1 ;

[0020] S203: Calculate straight line l ni With l ni+1 The intersection point p v As the corresponding vanishing point, and put it into the vanishing point set V s middle;

[0021] S204: If i<i max , then execute i=i+1 and return to step S202; otherwise execute i=0, n=n+1 and go to step S205;

[0022] S205: If n>n max , then stop and output the vanishing point set V s ; Otherwise return to step S202.

[0023] Preferably, in step S3, in the vanishing point set V s In the above example, sort the coordinate values ​​of the vanishing points in the x direction and put the median x in the x direction. vm The corresponding vanishing point (x vm ,y vm ) as the final vanishing point.

[0024] Preferably, in step S4, the formula fcosθ d +x vm sinθ d = 0 Calculate the pitch angle θ d ;

[0025] Where, f represents the focal length; (x vm ,y vm ) represents the coordinates of the final vanishing point.

[0026] Preferably, in step S4, if the pitch angle θ d Satisfy |θ d -θ s |<Δ θ , then the pitch angle θ d Credible; otherwise the pitch angle θ d Untrustworthy;

[0027] Among them, θ s represents the static pitch angle, Δ θ Indicates the set pitch angle fluctuation range.

[0028] The present invention also discloses a readable storage medium on which a computer management program is stored. When the computer management program is executed by a processor, the steps of the lane line-based pitch angle calibration method of the present invention are implemented.

[0029] Compared with the prior art, the pitch angle calibration method of the present invention has the following beneficial effects:

[0030] The present invention sets different sampling levels for different coordinate values ​​of lane lines in the x-direction, and traverses the vanishing points of each pair of lane lines at different sampling levels to generate a vanishing point set, then selects a final vanishing point from the vanishing point set by means of median filtering, and then calculates the pitch angle of the vehicle-mounted camera through the final vanishing point, so that a more stable vanishing point can be obtained, thereby improving the accuracy of the pitch angle calibration of the vehicle-mounted camera. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to make the purpose, technical solution and advantages of the invention more clear, the present invention will be further described in detail below with reference to the accompanying drawings, in which:

[0032] Figure 1 It is a logic block diagram of the pitch angle calibration method based on lane lines;

[0033] Figure 2 A schematic diagram for selecting lane lines;

[0034] Figure 3 A schematic diagram for generating a set of vanishing points;

[0035] Figure 4 Flowchart for generating a set of vanishing points;

[0036] Figure 5 Schematic diagram of selecting the final vanishing point for median filtering. DETAILED DESCRIPTION

[0037] The following is a further detailed description through specific implementation methods:

[0038] Embodiment 1:

[0039] This embodiment discloses a pitch angle calibration method based on lane lines.

[0040] like Figure 1 As shown, the pitch angle calibration method based on lane lines includes the following steps:

[0041] S1: Set n sampling levels, each sampling level corresponds to a different coordinate value of the lane line in the x direction;

[0042] S2: Calculate the vanishing points between each pair of adjacent lane lines at each sampling level and form a vanishing point set;

[0043] S3: Select the final vanishing point from the vanishing point set through median filtering;

[0044] S4: Calculate the corresponding pitch angle based on the final vanishing point.

[0045] It should be noted that the lane line-based pitch angle calibration method of the present invention can generate corresponding software codes or software services through program programming, and can then be run and implemented on a server and a computer.

[0046] The present invention sets different sampling levels for different coordinate values ​​of lane lines in the x-direction, and traverses the vanishing points of each pair of lane lines at different sampling levels to generate a vanishing point set, then selects a final vanishing point from the vanishing point set by means of median filtering, and then calculates the pitch angle of the vehicle-mounted camera through the final vanishing point, so that a more stable vanishing point can be obtained, thereby improving the accuracy of the pitch angle calibration of the vehicle-mounted camera.

[0047] In the specific implementation process, Figure 2 As shown, the lane lines are first screened and the lane lines that do not meet the conditions (lane lines that are too short) are deleted. , keep the i-th lane line; otherwise, delete the i-th lane line; where x imax Indicates the maximum coordinate of the i-th lane line in the x direction; Indicates the average value of the maximum x-coordinates of all lane lines; Δ x Indicates the lane line filter variable that is set.

[0048] The present invention screens lane lines and deletes lane lines whose lengths do not meet the requirements, so that unqualified lane lines can be excluded, thereby improving the quality of vanishing points in a vanishing point set.

[0049] In the specific implementation process, the average value of the maximum x-coordinates of all lane lines is calculated. Set the corresponding sampling level. Set three sampling levels, namely, the lane line in the x direction coordinate value and The sampling level.

[0050] In the specific implementation process, Figure 3 and Figure 4 As shown, the vanishing point set is formed by the following steps:

[0051] S201: Initialize lane line number i=0, sampling level n=0;

[0052] S202: Take the starting point of lane number i and the coordinate value of sampling level n in the x direction to form a straight line l ni , the lane line number is i+1, and the starting point and the coordinate value of sampling level n in the x direction form a straight line l ni+1 ;

[0053] S203: Calculate straight line l ni With l ni+1 The intersection point pv As the corresponding vanishing point, and put it into the vanishing point set (intersection point set) V s middle;

[0054] S204: If i<i max , then execute i=i+1 and return to step S202; otherwise execute i=0, n=n+1 and go to step S205;

[0055] S205: If n>n max , then stop and output the vanishing point set V s ; Otherwise return to step S202.

[0056] The present invention can generate a vanishing point set by traversing the vanishing points of each pair of lane lines at different sampling levels, and can further assist in obtaining more stable vanishing points, thereby improving the accuracy of elevation angle calibration of a vehicle-mounted camera.

[0057] In the specific implementation process, Figure 5 As shown, at the vanishing point set V s In the above example, sort the coordinate values ​​of the vanishing points in the x direction and put the median x in the x direction. vm The corresponding vanishing point (x vm ,y vm ) as the final vanishing point.

[0058] The present invention can effectively obtain more stable vanishing points from a vanishing point set by means of median filtering, thereby improving the accuracy of elevation angle calibration of a vehicle-mounted camera.

[0059] In the specific implementation process, through the formula fcosθ d +x vm sinθ d = 0 Calculate the pitch angle θ d ; Where f represents the focal length; (x vm ,y vm ) represents the coordinates of the final vanishing point. If the pitch angle θ d Satisfy |θ d -θ s |<Δ θ , then the pitch angle θ d Credible; otherwise the pitch angle θ d Unreliable (at this time, the static pitch angle θ s As output), where θ s represents the static pitch angle (known), Δ θ Indicates the set pitch angle fluctuation range.

[0060] The present invention can accurately calculate the pitch angle based on the final vanishing point, and can verify whether the pitch angle is credible, thereby improving the accuracy of the elevation angle calibration of the vehicle-mounted camera.

[0061] Embodiment 2:

[0062] This embodiment discloses a readable storage medium.

[0063] A readable storage medium stores a computer management program, which, when executed by a processor, implements the steps of the lane line-based pitch angle calibration method of the present invention. The readable storage medium may be a device with a readable storage function such as a USB flash drive or a computer.

[0064] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit the technical solution. Those skilled in the art should understand that those modifications or equivalent substitutions of the technical solution of the present invention that do not depart from the purpose and scope of the technical solution should be included in the scope of the claims of the present invention.

Claims

1. A pitch angle calibration method based on lane lines, characterized in that: The following steps are involved: S1: Set n sampling levels, each sampling level corresponds to a different coordinate value of the lane line in the x direction; In step S1, the lane lines are first screened and the lane lines whose lengths do not meet the requirements are deleted; When the i-th lane line satisfies When , keep the i-th lane line; otherwise, delete the i-th lane line; Among them, x imax Indicates the maximum coordinate of the i-th lane line in the x direction; Indicates the average value of the maximum x-coordinates of all lane lines; Δ x Indicates the lane line screening variable that is set; Based on the average value of the maximum x-coordinates of all lane lines Set the corresponding sampling level; Set three sampling levels, namely, the lane line has a coordinate value in the x direction and The sampling level; S2: Calculate the vanishing points between each pair of adjacent lane lines at each sampling level and form a vanishing point set; In step S2, a vanishing point set is formed by the following steps: S201: Initialize lane line number i=0, sampling level n=0; S202: Take the starting point of lane number i and the coordinate value of sampling level n in the x direction to form a straight line l ni , the lane line number is i+1, and the starting point and the coordinate value of sampling level n in the x direction form a straight line l ni+1 ; S203: Calculate straight line l ni With l ni+1 The intersection point p v As the corresponding vanishing point, and put it into the vanishing point set V s middle; S204: If i max , then execute i=i+1 and return to step S202; otherwise execute i=0, n=n+1 and go to step S205;​ S205: If n>n max , then stop and output the vanishing point set V s ; Otherwise, return to step S202; S3: Select the final vanishing point from the vanishing point set through median filtering; S4: Calculate the corresponding pitch angle based on the final vanishing point.

2. The lane line-based pitch angle calibration method according to claim 1, characterized in that: In step S3, in the vanishing point set V s In the above example, sort the coordinate values ​​of the vanishing points in the x direction and put the median x in the x direction. vm The corresponding vanishing point (x vm ,y vm ) as the final vanishing point.

3. The lane line-based pitch angle calibration method according to claim 1, characterized in that: In step S4, the formula fcosθ is used d +x vm sinθ d = 0 Calculate the pitch angle θ d ; Where, f represents the focal length; (x vm ,y vm ) represents the coordinates of the final vanishing point.

4. The lane line-based pitch angle calibration method according to claim 3, characterized in that: In step S4, if the pitch angle θ d Satisfy |θ d -θ s |<Δ θ , then the pitch angle θ d Credible; otherwise the pitch angle θ d Untrustworthy; Among them, θ s represents the static pitch angle, Δ θ Indicates the set pitch angle fluctuation range.

5. A readable storage medium, characterized in that: A computer management program is stored thereon, and when the computer management program is executed by a processor, the steps of the lane line-based pitch angle calibration method as described in any one of claims 1 to 4 are implemented.

Images