A lane curvature calculation implementation method, system, vehicle and storage medium

Abstract

The present application proposes a lane curvature calculation implementation method, system, vehicle and storage medium, the method comprising: if the disappearance time of the lane line on either side of the left or right lane line is within the preset time calibration value range, the fused mean of the polynomial coefficients when the lane line on that side disappears is calculated based on the longitudinal speed of the vehicle and the time difference from the disappearance time to the current moment; the variance of the fused mean is calculated based on the fused mean of the polynomial coefficients and the polynomial coefficients of the lane line on the other side at the current moment; when the calculation result is within the preset threshold range, the current polynomial coefficient is selected as the target parameter value of the lane centerline; the target lane curvature is calculated based on the target parameter value. The present application uses variance-weighted fusion calculation of lane line information on both sides, so that when the quality of single- or double-side lane line detection information is poor, the curvature information of the road is restored to the greatest extent possible, thereby improving the lane line detection quality and reducing the probability of invalid values.

Description

Technical Field

[0001] The present application belongs to the field of vehicle autonomous driving technology, and in particular, relates to a lane curvature calculation implementation method, system, vehicle and storage medium. Background Art

[0002] Lane centering and adaptive cruise control of autonomous driving require more accurate calculation of lane curvature. Current technology uses the result of perception module calculation when calculating lane curvature, that is, perception fits the cubic polynomial equation of the lane line according to the lane line image collected by the camera. Usually, the perception module outputs the coefficient of the fitted cubic polynomial equation and also outputs the standard deviation or confidence of the coefficient, such as C0_var, C1_var, C2_var, C3_var, to characterize the reliability of the fitting result of the lane line equation. However, in the case of poor detection quality of lane lines on one side, the variance is large, that is, the lane line information on the side with poor detection quality cannot be fully utilized, but is filtered by the variance judgment condition; or, in the case of poor detection quality of lane lines on both sides, the lane line information on both sides is filtered, and the effective information cannot be fully utilized to judge the lane features. In addition, when the variance of one side jumps between the upper and lower ranges of the threshold, it will cause jumps such as parameter calculation, which will affect the actual driving, resulting in the inability to calculate the accurate lane line curvature and reduce the quality of lane line detection. Summary of the invention

[0003] In response to the defects of the above-mentioned prior art, the present application provides a lane curvature calculation implementation method, system, vehicle and storage medium, which fuses and calculates the lane line information on both sides in a variance-weighted manner, so that when the quality of lane line detection information on one side or both sides is poor, the curvature information of the road can be restored to the greatest extent possible. The variance of the characteristic parameters of the calculated lane centerline will be lower than the variance of the left and right lane lines in the prior art, thereby improving the lane line detection quality and greatly reducing the probability of invalid values.

[0004] To achieve the above purpose, the present application provides a lane curvature calculation implementation method, the implementation method mainly includes:

[0005] S1: If the disappearance time of any lane line on the left or right side is within the preset time calibration value range, calculate the fused mean of the polynomial coefficients of the lane line on either side when it disappears based on the first lane line parameter of the vehicle, the longitudinal speed vl, and the time difference △T from the disappearance time to the current moment.

[0006] According to the fused mean of the polynomial coefficients of the lane line on either side when it disappears and the polynomial coefficient of the lane line on the other side at the current moment, the variance of the current fused mean is calculated to be one.

[0007] S2: Only when the calculation result of step S1 is within the preset threshold range, the current polynomial coefficient is selected as the target parameter value of the lane centerline.

[0008] S3: Calculate the target lane curvature according to the target parameter value.

[0009] In the present application, before step S1, the following steps are also included:

[0010] Get the current lane line image.

[0011] If it is determined that the left and right lane lines both exist within the preset time calibration value range, the variance 2 of the fusion mean is calculated based on the polynomial coefficients Cr and Cl of the left and right lane lines; the first lane line parameter is obtained based on the variance 2 of the fusion mean.

[0012] Wherein, the step S1 further includes:

[0013] Only when calculating the fused mean of the polynomial first-order coefficients when the lane line on either side disappears, the variance three of the current fused mean is calculated according to the vehicle's yaw angular velocity within the time difference △T and the polynomial first-order coefficients of the lane line on the other side at the current moment.

[0014] In this application, the step S1 also includes:

[0015] Only when calculating the fused mean of the cubic coefficients of the polynomial when the lane line on either side disappears, the fused mean variance of the cubic coefficients of the polynomial when the lane line on either side disappears is calculated according to the calibration value VAR3 of the vehicle within the time difference △T; wherein the calibration value VAR3 is the preset degree value of the fused mean variance of the cubic coefficients of the polynomial relative to the time diffusion.

[0016] In this application, the step S1 also includes:

[0017] Only when the fused mean of the polynomial constant term coefficients when the lane line on either side disappears is calculated, the fused mean variance of the polynomial constant term coefficients when the lane line on either side disappears is calculated according to the lateral moving speed of the vehicle within the time difference ΔT.

[0018] In this application, the step S1 also includes:

[0019] If the lane line on either side of the lane line on the left or right side does not exist within the preset time calibration value range, only when the roadside where the lane line on either side is located exists, the variance of the current fusion mean is calculated based on the polynomial coefficients of the roadside.

[0020] In this application, the step S1 also includes:

[0021] If the lane lines on the left and right sides do not exist within the preset time calibration value range, only when the left and right curbs where the lane lines on the left and right sides are respectively located exist, the variance of the current fusion mean is calculated based on the polynomial coefficients of the left and right curbs.

[0022] To achieve the above purpose, the present application also provides a lane curvature calculation implementation system, the system mainly includes:

[0023] A first processing module, a second processing module and a third processing module.

[0024] Among them, the first processing module is used to calculate the fused mean of the polynomial coefficients of the lane lines on either side when the disappearance time of the lane lines on either side is within the preset time calibration value range, based on the first lane line parameters of the vehicle, the longitudinal speed vl, and the time difference △T from the disappearance time to the current moment; calculate the variance of the current fused mean based on the fused mean of the polynomial coefficients of the lane lines on either side when they disappear and the polynomial coefficients of the lane lines on the other side at the current moment.

[0025] The second processing module is used to select the current polynomial coefficient as the target parameter value of the lane centerline only when the calculation result of the step S1 is within a preset threshold range.

[0026] And, the third processing module is used to calculate the target lane curvature according to the target parameter value.

[0027] To achieve the above objectives, the present application also provides a vehicle, which at least includes the lane curvature calculation implementation system as described above.

[0028] To achieve the above objectives, the present application also provides a storage medium, which is a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it implements any of the lane curvature calculation methods described above.

[0029] Compared with the prior art, the present invention has the following beneficial effects:

[0030] The present application proposes a lane curvature calculation implementation method, system, vehicle and storage medium, which fuses and calculates the lane line information on both sides in a variance-weighted manner, so that when the quality of lane line detection information on one side or both sides is poor, the curvature information of the road can be restored to the greatest extent possible, and the variance of the characteristic parameters of the calculated lane centerline will be lower than the variance of the lane lines on the left and right sides in the prior art, thereby improving the quality of lane line detection. The present application can also achieve that when the lane line detection quality on one side is low and the detection quality on the other side is high, the result of the side with lower detection quality will not be simply discarded, and the detection result on the other side will be fully trusted, but a similar weighted calculation will be performed. When the variance on one side is large, the weight of the parameters on that side is low, so that the final result tends to the lane line parameters on the side with smaller variance more smoothly, greatly reducing the probability of invalid values. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 This is a flow chart of a method for calculating lane curvature in one embodiment of the present application.

[0032] Figure 2 A schematic diagram of the structure of a lane curvature calculation implementation system in one embodiment of the present application. DETAILED DESCRIPTION

[0033] In order to make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme will be clearly and completely described below in conjunction with the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0034] Embodiment 1:

[0035] In order to solve the above technical problems, the present application provides a lane curvature calculation implementation method, which fully utilizes the lane line information on both sides through variance weighting, so as to restore the curvature information of the road to the greatest extent possible when the quality of the lane line detection information on one side or both sides is poor.

[0036] It should be noted that the current technology uses the result of the perception module to calculate the lane line curvature, that is, the perception module fits the lane line's cubic polynomial equation based on the lane line image collected by the camera:

[0037] y=C3*x 3 +C2*x 2 +C1*x+C0.

[0038] Among them, C0, C1, C2, and C3 are the coefficients of the fitted cubic polynomial equation respectively. Usually, the perception module outputs the standard deviation or confidence of the above four coefficients at the same time, namely C0_var, C1_var, C2_var, and C3_var, to characterize the reliability of the fitting results of the lane line equation.

[0039] The lane curvature calculation formula is as follows:

[0040]

[0041] generally Therefore

[0042] Derivative the above lane line cubic polynomial equation can be obtained:

[0043] k≈6C3x+2C2.

[0044] Since C3 itself does not change much, the above formula is simplified as follows:

[0045] k≈2C2.

[0046] Since the lane is usually obtained by combining the lane lines on the left and right sides, the lane curvature is usually calculated by the average value of the lane lines on both sides, that is:

[0047]

[0048] When the quality of lane line detection on one side is not high, the credibility of lane line information is usually judged based on the variance or confidence of the lane line. Take variance as an example. If C2_var_left<threshold and C2_var_right<threshold, then:

[0049]

[0050] If C2_var_left < threshold and C2_var_right > threshold, then:

[0051] k=k l .

[0052] If C2_var_left>threshold and C2_var_right<threshold, then:

[0053] k=k r .

[0054] If C2_var_left>threshold and C2_var_right>threshold, then k is invalid.

[0055] However, when the detection quality of lane lines on one side is poor, the variance is large, that is, the lane line information on the side with poor detection quality cannot be fully utilized, but is filtered by the variance judgment condition. Alternatively, when the detection quality of lane lines on both sides is poor, the lane line information on both sides is filtered, and the effective information cannot be fully utilized to judge the lane characteristics; and when the variance on one side jumps between the upper and lower ranges of the threshold, it will cause jumps such as parameter calculations, thus affecting actual driving.

[0056] It should be noted that when the quality of lane line detection information on one side or both sides is poor, the curvature information of the road is restored as much as possible, for example:

[0057] Poor lighting conditions: At night, in bad weather (such as rain, snow, fog, etc.) or under strong direct sunlight, the lane line information captured by the camera may become blurred or unclear due to the lighting conditions.

[0058] Blurred or worn road markings: Some road markings may become blurred or worn due to long-term use or poor construction and maintenance, making lane line detection more difficult or even impossible to accurately identify.

[0059] Missing road markings: Some roads may have missing road markings due to construction, diversion or other reasons, which will make lane line detection difficult.

[0060] Also, in some unstructured road environments, such as country roads and wild roads, there may be a lack of standard road markings or poor visibility of the markings, which makes lane line detection difficult.

[0061] The above situations may lead to poor quality of lane line detection information, which in turn affects the accuracy of road curvature information.

[0062] Compared with the above-mentioned prior art, in this embodiment, the implementation method mainly includes:

[0063] S1: If the disappearance time of any lane line on the left or right side is within the preset time calibration value range, calculate the fused mean of the polynomial coefficients of the lane line on either side when it disappears based on the first lane line parameter of the vehicle, the longitudinal speed vl, and the time difference △T from the disappearance time to the current moment.

[0064] According to the fused mean of the polynomial coefficients of the lane line on either side when it disappears and the polynomial coefficient of the lane line on the other side at the current moment, the variance of the current fused mean is calculated to be one.

[0065] In the present application, before step S1, the following steps are also included:

[0066] Get the current lane line image.

[0067] Preferably, a vehicle-mounted camera may be used to capture the current lane line image, and then the image information may be extracted and calculated through a vehicle-mounted perception module, but this is not limited to this.

[0068] If it is determined that the left and right lane lines both exist within the preset time calibration value range, the variance 2 of the fusion mean is calculated based on the polynomial coefficients Cr and Cl of the left and right lane lines; the first lane line parameter is obtained based on the variance 2 of the fusion mean.

[0069] When both left and right lane lines exist:

[0070] Get the binomial coefficient C2_left of the left lane line, the binomial coefficient C2_right of the right lane line, as well as C2_left_var and C2_right_var, where C2_left_var and C2_right_var are the variances of the binomial coefficients of the left and right lanes, respectively.

[0071] Next, calculate the sum of the squared errors:

[0072] C2_sum_var=C2_left_var+C2_right_var.

[0073] Then calculate the mean:

[0074] C2_avg=(C2_left*C2_right_var+C2_right*C2_left_var) / C2_sum_var.

[0075] Calculate the variance of the mean based on the above mean results:

[0076] C2_avg_var=C2_left_var*C2_right_var / C2_sum_var.

[0077] If C2_avg_var>threshold, the C2_avg value is marked as invalid; when C2_avg_var<threshold, C2 is used as the target parameter of the lane centerline.

[0078] By substituting the target parameter into the above lane line cubic polynomial equation, the lane curvature is calculated, that is:

[0079] k=2*C2_avg.

[0080] In other embodiments, the lane curvature when the lane line parameters C1 and C3 are calculated by iterating the above fusion calculation method is not limited thereto.

[0081] It should be noted that when calculating the polynomial constant term C0 parameter, since it is necessary to calculate the C0 parameter of the lane centerline, the following method should be used when calculating the mean:

[0082] C0_avg=(C0_left+C0_right) / 2; other steps remain unchanged.

[0083] S1: If the disappearance time of any lane line on the left or right side is within the preset time calibration value range, calculate the fused mean of the polynomial coefficients of the lane line on either side when it disappears based on the first lane line parameter of the vehicle, the longitudinal speed vl, and the time difference △T from the disappearance time to the current moment; calculate the variance of the current fused mean based on the fused mean of the polynomial coefficients of the lane line on either side when it disappears and the polynomial coefficient of the lane line on the other side at the current moment.

[0084] In this embodiment, when the left lane line exists and the right lane line does not exist for a short time:

[0085] The first lane line parameter is the lane line parameter calculated by the method when the left and right lane lines exist. For example, when the coefficient of the quadratic term of the polynomial needs to be calculated:

[0086] C2_avg_last=(C2_left*C2_right_var+C2_right*C2_left_var) / C2_sum_var;

[0087] C2_avg_var_last=C2_left_var*C2_right_var / C2_sum_va.

[0088] C3_avg_last=(C3_left*C3_right_var+C3_right*C3_left_var) / C3_sum_var;

[0089] C3_avg_var_last=C3_left_var*C3_right_var / C3_sum_var.

[0090] Next, calculate the estimated value from the previous moment to the current moment:

[0091] C2_avg_this=C2_avg_last+C3_avg_last*V_long*ΔT.

[0092] C2_avg_var_this=C2_avg_var_last+C3_avg_var_last*V_long_var*ΔT; V_long is the longitudinal speed of the vehicle, ΔT is the time difference between the last calculation and this calculation

[0093] Further, similar to the above calculation method, obtain the variance of C2_left and C2_left_var of the left lane line at the current moment. Then calculate the parameter mean at the current moment:

[0094] C2_avg=(C2_left*C2_avg_var_this+C2_avg_this*C2_left_var) / C2_sum_var.

[0095] Calculate the variance of the mean based on the parameter mean at the current moment:

[0096] C2_avg_var=C2_left_var*C2_avg_var_this / C2_sum_var.

[0097] Among them, if C2_avg_var>threshold, C2_avg is invalid. If C2_avg_var<threshold, C2 is used as the parameter of the lane centerline.

[0098] Finally, the lane curvature is calculated, that is: k = 2*C2_avg.

[0099] If the right lane line appears in the next calculation cycle, the calculation method is switched back to the scenario when the left and right lane lines exist. If the right lane line still does not exist in the next calculation cycle, the calculation steps are repeated when the left lane line exists and the right lane line does not exist for a short time.

[0100] S2: Only when the calculation result of step S1 is within the preset threshold range, the current polynomial coefficient is selected as the target parameter value of the lane centerline.

[0101] It should be noted that if the calculation result of step S1 exceeds the preset threshold range, the data information is marked as invalid. The calculation result of step S1 is the fusion mean variance under each lane line coefficient.

[0102] S3: Calculate the target lane curvature according to the target parameter value.

[0103] In this embodiment, the calculation formula may adopt the cubic polynomial equation of the lane line:

[0104] y=C3*x 3 +C2*x 2 +C1*x+C0.

[0105] Among them, C0, C1, C2, and C3 are the coefficients of the fitted cubic polynomial equation.

[0106] Wherein, the step S1 further includes:

[0107] Only when calculating the fused mean of the polynomial first-order coefficients when the lane line on either side disappears, the variance three of the current fused mean is calculated according to the vehicle's yaw angular velocity within the time difference △T and the polynomial first-order coefficients of the lane line on the other side at the current moment.

[0108] In this embodiment, when calculating the C1 parameter of the middle lane line when the left lane line exists and the right lane line does not exist for a short time, the specific calculation is as follows:

[0109] C1_avg_last=(C1_left*C1_right_var+C1_right*C1_left_var) / C1_sum_var;

[0110] C1_avg_var_last=C1_left_var*C1_right_var / C1_sum_var.

[0111] C2_avg_last=(C2_left*C2_right_var+C2_right*C2_left_var) / C2_sum_var;

[0112] C2_avg_var_last=C2_left_var*C2_right_var / C2_sum_var.

[0113] Next, calculate the estimated value from the previous moment to the current moment:

[0114] C1_avg_this=C1_avg_last+C2_avg_last*V_long*ΔT+tan(-Yawrate*ΔT);

[0115] C1_avg_var_this=C1_avg_var_last+C2_avg_var_last*V_long*ΔT+Yawrate_var*ΔT.

[0116] Among them, yawrate is the yaw rate of the vehicle at time ΔT, yawrate_var is the variance of the yaw rate observation, and the other calculation steps remain unchanged.

[0117] In this application, the step S1 also includes:

[0118] Only when calculating the fused mean of the cubic coefficients of the polynomial when the lane line on either side disappears, the fused mean variance of the cubic coefficients of the polynomial when the lane line on either side disappears is calculated according to the calibration value VAR3 of the vehicle within the time difference △T; wherein the calibration value VAR3 is the preset degree value of the fused mean variance of the cubic coefficients of the polynomial relative to the time diffusion.

[0119] In this embodiment, when calculating the C3 parameter of the middle lane line when the left lane line exists and the right lane line does not exist for a short time, the specific calculation is as follows:

[0120] C3_avg_last=(C3_left*C3_right_var+C3_right*C3_left_var) / C3_sum_var;

[0121] C3_avg_var_last=C3_left_var*C3_right_var / C3_sum_var.

[0122] Next, calculate the estimated value from the previous moment to the current moment:

[0123] C3_avg_this = C3_avg_last;

[0124] C3_avg_var_this=C3_avg_var_last+VAR3*ΔT.

[0125] Among them, VAR3 can be set to a calibration value to characterize the diffusion degree of C3 variance relative to time, preferably 0.5 to 1. The other calculation steps remain unchanged.

[0126] In this application, the step S1 also includes:

[0127] Only when the fused mean of the polynomial constant term coefficients when the lane line on either side disappears is calculated, the fused mean variance of the polynomial constant term coefficients when the lane line on either side disappears is calculated according to the lateral moving speed of the vehicle within the time difference ΔT.

[0128] In this embodiment, when calculating the C3 parameter of the middle lane line when the left lane line exists and the right lane line does not exist for a short time, since it is necessary to calculate the C0 parameter of the lane center line, the following method can be used for calculation:

[0129] The lane line C0 parameters calculated by the method when the left and right lane lines exist are further calculated:

[0130] C0_avg_last=(C0_lfet+C0_right) / 2;

[0131] C0_avg_var_last=C0_left_var*C0_right_var / C0_sum_var.

[0132] C1_avg_last=(C1_left*C1_right_var+C1_right*C1_left_var) / C1_sum_var;

[0133] C1_avg_var_last=C1_left_var*C1_right_var / C1_sum_var.

[0134] LaneWidth=C0_lfet-C0_right.

[0135] Next, calculate the estimated value from the previous moment to the current moment:

[0136] C0_avg_this=C0_avg_last+C1_avg_last*V_long*ΔT+V_latt*ΔT;

[0137] C0_avg_var_this=C0_avg_var_last+C1_avg_var_last*V_long_var*ΔT+V_latt_var*ΔT.

[0138] Where V_latt is the lateral velocity at time ΔT, and V_latt_var is the variance of the lateral velocity.

[0139] After obtaining the variance of C2_left and C2_left_var of the left lane line at the current moment, calculate the parameter mean at the current moment:

[0140] C0_avg=((C0_left-0.5LaneWidth)*C0_avg_var_this+C0_avg_this*C0_left_var) / C2_sum_var.

[0141] Furthermore, the variance of the mean is calculated:

[0142] C0_avg_var=C0_left_var*C0_avg_var_this / C0_sum_var;

[0143] If C0_avg_var>threshold, C0_avg is invalid. If C0_avg_var<threshold, C0 is used as the parameter of the lane centerline.

[0144] In this application, the step S1 also includes:

[0145] If the lane line on either side of the lane line on the left or right side does not exist within the preset time calibration value range, only when the roadside where the lane line on either side is located exists, the variance of the current fusion mean is calculated based on the polynomial coefficients of the roadside.

[0146] In this embodiment, when the left lane line exists and the right lane line does not exist for a long time, but the right curb exists, the right curb is used to replace the right lane line:

[0147] Get the variance of C2_left and C2_left_var of the left lane line.

[0148] Next, obtain the variance of C2_rightEdge and C2_rightEdge_var of the right curb.

[0149] Furthermore, the sum of the variances of the left lane line and the right curb is calculated:

[0150] C2_sum_var=C2_left_var+C2_rightEdge_var.

[0151] Then calculate the mean based on the above variance:

[0152] C2_avg=(C2_left*C2_rightEdge_var+C2_rightEdge*C2_left_var) / C2_sum_var.

[0153] Calculate the variance of the mean:

[0154] C2_avg_var=C2_left_var*C2_rightEdge_var / C2_sum_var.

[0155] If C2_avg_var>threshold, C2_avg is invalid. If C2_avg_var<threshold, C2 is used as the parameter of the lane centerline.

[0156] Finally, the lane curvature is calculated according to the cubic polynomial equation, that is: k = 2*C2_avg.

[0157] In other embodiments, the above method may also be used to calculate the C1 and C3 parameters of the lane line.

[0158] When calculating the C0 parameter, since it is necessary to calculate the C0 parameter of the lane centerline, the following method can be used to calculate the mean:

[0159] C0_avg=(C0_left+C0_right) / 2; other steps remain unchanged.

[0160] In this application, the step S1 also includes:

[0161] If the lane lines on the left and right sides do not exist within the preset time calibration value range, only when the left and right curbs where the lane lines on the left and right sides are respectively located exist, the variance of the current fusion mean is calculated based on the polynomial coefficients of the left and right curbs.

[0162] In this embodiment, when both left and right lane lines do not exist for a long time, but left and right curbs exist, the left and right curbs are used to replace the left and right lane lines:

[0163] Get the variance of C2_leftEdge C2_rightEdge and C2_leftEdge_var C2_rightEdge_var of the left and right curbs.

[0164] Calculate the sum of the squared differences between the left and right curbs:

[0165] C2_sum_var=C2_leftEdge_var+C2_rightEdge_var.

[0166] Calculate the mean:

[0167] C2_avg=(C2_leftEdge*C2_rightEdge_var+C2_rightEdge*C2_leftEdge_var) / C2_sum_var.

[0168] Next, calculate the variance of the mean:

[0169] C2_avg_var=C2_leftEdge_var*C2_rightEdge_var / C2_sum_var.

[0170] If C2_avg_var>threshold, C2_avg is invalid. If C2_avg_var<threshold, C2 is used as the parameter of the lane centerline.

[0171] Finally, the lane curvature is calculated according to the cubic polynomial equation, that is: k = 2*C2_avg.

[0172] In other embodiments, the above method may also be used to calculate the C1 and C3 parameters of the lane line.

[0173] When calculating the C0 parameter, since it is necessary to calculate the C0 parameter of the lane centerline, the following method can be used to calculate the mean:

[0174] C0_avg = (C0_leftEdge + C0_rightEdge) / 2; Other steps remain unchanged.

[0175] In this embodiment, the determination of short - time absence can use a calibrated value TimeOut of a time dimension as a threshold judgment. That is, when the disappearance time of a certain lane line < TimeOut, it is determined that the lane line is lost for a short time; otherwise, it is determined that the lane line is lost for a long time.

[0176] Preferably, the value of the TimeOut time can range from 0.5 s to 2 s, and is not limited to this.

[0177] In summary, Embodiment 1 makes a detailed analysis and example illustration of a method for calculating lane curvature proposed in this application. By fusing and calculating the information of the lane lines on both sides in a variance - weighted manner, the curvature information of the road can be restored as much as possible when the detection information quality of the lane lines on one or both sides is poor. The variance of the characteristic parameters of the calculated lane center line will be lower than the variances of the lane lines on the left and right sides in the prior art, improving the detection quality of the lane lines.

[0178] Embodiment 2:

[0179] As shown in the attached Figure 2 In order to solve the above - mentioned technical problems, this application also proposes a system for calculating lane curvature implementation. The system mainly includes:

[0180] A first processing module, a second processing module, and a third processing module.

[0181] Among them, the first processing module is used to calculate the fusion mean of the polynomial coefficients of any one - side lane line at the time of disappearance according to the first lane line parameter of the vehicle, the longitudinal speed vl, and the time difference △T from the disappearance time to the current moment when the disappearance time of any one - side lane line among the lane lines on the left and right sides is within the preset time calibration value range; and calculate the variance one of the current fusion mean according to the fusion mean of the polynomial coefficients of any one - side lane line at the time of disappearance and the polynomial coefficients of the other - side lane line at the current moment.

[0182] It should be noted that the first processing module is an important part of the system for calculating lane curvature implementation. Its main function is to estimate the fusion mean of the polynomial coefficients of the lane line at the time of disappearance and calculate the variance of the current fusion mean by calculating the disappearance time and related parameters of the lane lines on the left and right sides.

[0183] In this embodiment, the first processing module calculates the fused mean of the polynomial coefficients of the lane lines on either side when they disappear according to the disappearance time of the lane lines on the left and right sides during the vehicle's driving process, as well as the first lane line parameters of the vehicle, the longitudinal speed vl, and the time difference △T from the disappearance time to the current moment. This fused mean of the polynomial coefficients represents the fitting curve of the lane line when it disappears. Because there is still a distance after the lane line disappears, the lane line information within this distance can be used to refit the lane line, thereby more accurately estimating the position and curvature of the lane line.

[0184] Next, the first processing module also calculates the variance of the current fused mean value based on the polynomial coefficient of the lane line on the other side at the current moment and the fused mean value of the polynomial coefficient of the lane line on either side when it disappears. This variance can be used to evaluate the reliability of lane line fitting. If the variance is large, it means that the error of lane line fitting is large, and corresponding measures need to be taken to improve the fitting accuracy.

[0185] Preferably, when a lane line on one side disappears during vehicle driving, the first processing module can use the information and related parameters of the lane line on the other side, combined with factors such as disappearance time and longitudinal speed, to calculate the fused mean of the polynomial coefficients of the lane line when it disappears, and refit the lane line accordingly. At the same time, based on the polynomial coefficients of the lane line on the other side at the current moment, the variance of the current fused mean is calculated to evaluate the reliability of the lane line fitting. This process can help the vehicle identify the position and curvature of the lane line more accurately, and improve the safety and stability of vehicle driving.

[0186] The second processing module is used to select the current polynomial coefficient as the target parameter value of the lane centerline only when the calculation result of the step S1 is within a preset threshold range.

[0187] It should be noted that the second processing module is a key component in the lane curvature calculation implementation system. Its main function is to select the current polynomial coefficient as the target parameter value of the lane centerline under the premise of meeting the preset threshold range.

[0188] In this embodiment, after the first processing module calculates the fused mean of the lane line polynomial coefficients, the second processing module will judge the result. If the fused mean is within the preset threshold range, that is, the calculation result meets the preset conditions, the second processing module will select the current polynomial coefficient as the target parameter value of the lane centerline.

[0189] Preferably, it is assumed that the fused mean of the lane line polynomial coefficients calculated by the first processing module is A. At this time, the second processing module will compare this fused mean with a preset threshold. If A is within the preset threshold range, the second processing module will use A as the target parameter value of the lane centerline, that is, A is considered to be the best fitting curve parameter, indicating the centerline position of the current lane. In this way, the vehicle can adjust its own driving trajectory according to the target parameter value, keep driving in the center of the lane, and improve driving accuracy and stability.

[0190] Through the judgment and selection of the second processing module, it can be ensured that only the calculation results that meet the preset threshold range are selected as the target parameter value of the lane centerline. The purpose of this is to prevent inaccurate lane line parameters caused by calculation errors or other factors from being mistakenly selected as the target parameter value, thereby affecting the driving safety of the vehicle.

[0191] And, the third processing module is used to calculate the target lane curvature according to the target parameter value.

[0192] It should be noted that the third processing module is the last component in the lane curvature calculation implementation system, and its main function is to calculate the curvature of the target lane according to the target parameter value determined by the second processing module.

[0193] In this embodiment, after the second processing module confirms that the current polynomial coefficient is the target parameter value of the lane centerline, the third processing module will use this target parameter value to calculate the curvature of the lane. This calculation process involves the derivation and calculation of the polynomial equation, as well as the estimation of the distance and time traveled by the vehicle within a certain period of time, etc., and the curvature information of the lane, that is, the curvature value of the target lane, can be obtained.

[0194] Preferably, it is assumed that the second processing module has selected the current polynomial coefficient as the target parameter value of the lane centerline, that is, it is considered that the centerline position of the current lane has been accurately determined. At this time, the third processing module will use this target parameter value to calculate the curvature value of the target lane by means of derivation and calculation of the polynomial equation. This curvature value represents the curvature of the current lane and can be directly used in vehicle control and driving assistance.

[0195] Through the calculation of the third processing module, the vehicle can obtain the curvature information of the current lane, so as to better control the speed and steering, ensure that the vehicle is driving in the appropriate lane position and speed, and improve driving safety and reliability. At the same time, this curvature information can also be used for intelligent assisted driving and automatic driving of the vehicle, realizing more intelligent and autonomous vehicle control.

[0196] In summary, Example 2 provides a detailed analysis and example description of a lane curvature calculation implementation system proposed in this application. The system provides accurate road information for vehicle positioning and driving through lane line data processing and curvature calculation, thereby achieving a safer and more stable driving experience, and providing basic support for intelligent assisted driving and autonomous driving technology. The specific implementation process and principle have been introduced in detail in Example 1 and will not be repeated in this embodiment.

[0197] Embodiment three:

[0198] The present application also provides a vehicle, which at least includes the lane curvature calculation implementation system as described above.

[0199] It should be noted that vehicles equipped with a lane curvature calculation implementation system can play a role in a variety of application scenarios, such as:

[0200] The vehicle can be used as part of an autonomous driving system to help the vehicle with path planning, obstacle avoidance and driving control by calculating the curvature information of the lane in real time. On highways, the vehicle can autonomously adjust the speed and steering according to the curvature information to achieve automatic cruise control. On urban roads, the vehicle can perform operations such as lane keeping and turning according to the curvature information.

[0201] In other embodiments, vehicles equipped with the system can provide more powerful driving assistance functions, such as smart cruise control, lane keeping assist, steering assist, etc. By calculating the curvature of the target lane, the vehicle can make adaptive adjustments based on the curvature of the road and assist the driver in safe driving.

[0202] The vehicle can also be used to collect and analyze road information. By processing lane line data and calculating curvature information, parameters such as the curvature of the road and slope changes can be obtained, providing important data support for road design, traffic planning and road maintenance.

[0203] Assume that there is a self-driving car equipped with a lane curvature calculation implementation system. The following is an example of the specific implementation process and its role:

[0204] First, the vehicle is equipped with sensors, such as cameras or lidar, to obtain visual or laser data on the road.

[0205] Then, the acquired data is processed through image processing or point cloud processing technology to extract lane line information.

[0206] Next, the lane detection algorithm is used to determine the position and shape of the lane centerline based on the processed data, and the polynomial coefficients are extracted as the target parameter values.

[0207] Finally, the target parameter value is passed to the lane curvature calculation implementation system to perform curvature calculation and obtain the curvature information of the target lane.

[0208] In this embodiment, based on the lane line information, the vehicle can accurately determine its position on the road, thereby keeping the vehicle traveling in a suitable lane.

[0209] By calculating the curvature of the target lane, the vehicle can make corresponding steering adjustments according to the curvature of the road to achieve smoother and safer driving. Curvature information can be used in the vehicle's driving assistance system, such as smart cruise control, steering assistance and other functions, to help the driver better control the vehicle and enhance driving safety. Lane curvature information is crucial for autonomous driving systems. By calculating the curvature of the target lane in real time, the vehicle can plan and control the path according to road conditions to achieve autonomous driving functions.

[0210] In summary, vehicles equipped with a lane curvature calculation implementation system can realize multiple functions such as lane positioning, curve adaptability, driving assistance and automatic driving by processing lane line data and calculating curvature information, thereby improving driving safety, comfort and intelligence.

[0211] Embodiment 4:

[0212] The present application also provides a storage medium, which is a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it implements any of the lane curvature calculation methods described above.

[0213] The computer-readable storage medium may be configured in any computer device, which may be a computing device such as a tablet computer, a desktop computer, and a cloud server, and may also be a vehicle-mounted terminal or a control system, for example, and may also include input and output devices, network access devices, etc.

[0214] In this embodiment, visual or laser data on the road is acquired through sensors (such as cameras or lidar).

[0215] Next, the acquired data is processed using techniques such as image processing or point cloud processing to extract lane line information.

[0216] The lane detection algorithm is then applied to determine the position and shape of the lane centerline based on the processed data, and the polynomial coefficients are extracted as the target parameter values.

[0217] The target parameter value is passed to the curvature calculation module for curvature calculation.

[0218] Finally, the calculated curvature information is obtained and used for autonomous driving, driver assistance systems, or other related applications.

[0219] Curvature information can be used in driving assistance systems such as intelligent cruise control and steering assistance to help drivers better control the vehicle and enhance driving safety.

[0220] In summary, the computer program in this storage medium implements the calculation method of lane curvature, provides flexible and efficient data processing and decision support capabilities for various computing devices, and is expected to promote the widespread application and development of intelligent technology in various fields.

[0221] The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or the processor may be any conventional processor, etc. The memory is used to store an operating system, an application program, a boot loader, data, and other programs, such as the program code of the computer program, etc. The memory may also be used to temporarily store data that has been output or is to be output.

[0222] In several embodiments provided in the present application, it is understood that each box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, a program segment or a part of a code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from that marked in the accompanying drawings. For example, two consecutive boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved.

[0223] If the functions are implemented in the form of software function modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device to execute all or part of the steps of the methods described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0224] The specific embodiments described above further describe the purpose, technical solutions and beneficial effects of the present application in detail. It should be understood that the above description is only a specific embodiment of the present application and is not intended to limit the scope of protection of the present application. It is particularly pointed out that for those skilled in the art, any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of protection of the present application.

Claims

1. A method for calculating lane curvature, characterized in that: include: S1: If the disappearance time of any lane line on the left or right side is within the preset time calibration value range, calculate the fused mean value of the polynomial coefficients of the lane line on either side when it disappears according to the first lane line parameter of the vehicle, the longitudinal speed vl, and the time difference △T from the disappearance time to the current moment; Calculate the variance of the current fused mean value based on the fused mean value of the polynomial coefficients of the lane line on either side when it disappears and the polynomial coefficients of the lane line on the other side at the current moment; S2: Only when the variance of the current fusion mean is within a preset threshold, the current polynomial coefficient is selected as the target parameter value of the lane centerline; S3: Calculating a target lane curvature according to the target parameter value; Before step S1, the method further includes: Get the current lane line image; When it is determined that the lane lines on the left and right sides both exist within the preset time calibration value range, the variance of the fusion mean is calculated according to the polynomial coefficients Cr and Cl of the lane lines on the left and right sides; The first lane line parameter is obtained according to the variance 2 of the fusion mean.

2. A lane curvature calculation implementation method according to claim 1, characterized in that: The step S1 further comprises: Only when calculating the fused mean of the polynomial first-order coefficients when the lane line on either side disappears, the variance three of the current fused mean is calculated according to the vehicle's yaw angular velocity within the time difference △T and the polynomial first-order coefficients of the lane line on the other side at the current moment.

3. The method for calculating lane curvature according to claim 1, characterized in that: The step S1 further comprises: Only when calculating the fused mean of the cubic coefficients of the polynomial when the lane line on either side disappears, the fused mean variance of the cubic coefficients of the polynomial when the lane line on either side disappears is calculated according to the calibration value VAR3 of the vehicle within the time difference △T; wherein the calibration value VAR3 is the preset degree value of the fused mean variance of the cubic coefficients of the polynomial relative to the time diffusion.

4. A lane curvature calculation implementation method according to claim 3, characterized in that: The step S1 further comprises: Only when the fused mean of the polynomial constant term coefficients when the lane line on either side disappears is calculated, the fused mean variance of the polynomial constant term coefficients when the lane line on either side disappears is calculated according to the lateral moving speed of the vehicle within the time difference ΔT.

5. A lane curvature calculation implementation method according to claim 4, characterized in that: The step S1 further comprises: If the lane line on either side of the lane line on the left or right side does not exist within the preset time calibration value range, only when the roadside where the lane line on either side is located exists, the variance of the current fusion mean is calculated based on the polynomial coefficients of the roadside.

6. A lane curvature calculation implementation method according to claim 5, characterized in that: The step S1 further comprises: If the lane lines on the left and right sides do not exist within the preset time calibration value range, only when the left and right curbs where the lane lines on the left and right sides are respectively located exist, the variance of the current fusion mean is calculated based on the polynomial coefficients of the left and right curbs.

7. A lane curvature calculation implementation system, characterized in that: include: a first processing module, a second processing module and a third processing module; The first processing module is used to calculate the fused mean of the polynomial coefficients of the lane line on either side when the lane line on the left or right side disappears, according to the first lane line parameter of the vehicle, the longitudinal speed vl, and the time difference △T from the disappearance time to the current moment, when the disappearance time of the lane line on either side is within the preset time calibration value range; calculate the variance of the current fused mean value one according to the fused mean of the polynomial coefficients of the lane line on either side when the lane line on either side disappears and the polynomial coefficient of the lane line on the other side at the current moment; The second processing module is used to select the current polynomial coefficient as the target parameter value of the lane centerline only when the variance of the current fusion mean is within a preset threshold range; And, the third processing module is used to calculate the target lane curvature according to the target parameter value; Among them, obtaining the first lane line parameter is specifically as follows: obtaining the current lane line image; when it is determined that the left and right lane lines exist within the preset time calibration value range, calculating the variance two of the fusion mean according to the polynomial coefficients Cr and Cl of the left and right lane lines; obtaining the first lane line parameter according to the variance two of the fusion mean.

8. A vehicle, characterized in that: The vehicle at least includes the lane curvature calculation implementation system described in claim 7.

9. A storage medium, characterized in that: It is a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the lane curvature calculation method described in any one of claims 1 to 6 is implemented.

Images