Lane line optimization method and electronic equipment

Abstract

The invention relates to the technical field of intelligent auxiliary driving, and provides a lane line optimization method and electronic equipment, which are characterized in that an algorithm framework model based on classical Kalman filtering is set, the lane line precision is specifically optimized by using vehicle motion information, IMU information, HDmap information, INS information and lane line sensing information, the lane line detection stability and robustness are optimized, and the lane line detection accuracy is improved. Meanwhile, abnormal lane line information is detected; the main model predicts lane line information based on adaptive Kalman filtering in combination with vehicle motion information, and realizes observation updating compensation in combination with high-precision map lane lines and sensing lane line information; abnormality detection is achieved through observation and prediction of residual error information, and stable and reliable guarantee can be provided for all automatic driving vehicles driving on public roads.

Description

Technical Field

[0001] The present invention relates to the field of intelligent assisted driving technology, and in particular to a lane line optimization method and electronic equipment. Background Art

[0002] Lane line detection is a key part of autonomous driving assistance technology. With the deepening of research, the scenarios that lane line detection tasks deal with are becoming more and more diverse. For example, it is used in multiple functions such as lane keeping, lane departure warning, and adaptive cruise control. In actual traffic environments, lane line generation may fail due to objective reasons such as weather changes, lighting, road damage, or camera damage, which may cause trouble to vehicle driving and cause the vehicle's autonomous driving assistance technology to exit or downgrade.

[0003] The key parameters of the classic lane line equation (c0, c1, c2, c3) represent the lateral distance of the vehicle from the lane line, the lane line heading angle, the lane line curvature, and the curvature change rate. This traditional lane line perception solution has problems such as false detection, missed detection, unstable detection results, poor lane line curvature estimation accuracy / even errors, such as:

[0004] 1) Lane line accuracy depends on the results of the perception end. Lane line accuracy is limited by distance and perception range. The farther away from the vehicle, the lower the accuracy.

[0005] 2) Abnormal lane lines at intersections cannot be distinguished, and intersection scenarios cannot be effectively identified. Lane line perception anomalies will occur in situations such as lane change and lane line merging, so it is impossible to provide effective and accurate lane line recognition services. Summary of the invention

[0006] The present invention provides a lane line optimization method and electronic equipment, which solves the technical problem that the perception accuracy of the existing lane line perception solution is limited by the perception range and abnormal lane lines cannot be effectively identified, resulting in poor lane line recognition accuracy.

[0007] In order to solve the above technical problems, the present invention provides a lane line optimization method and electronic device, comprising the steps of:

[0008] Carry out data collection to obtain vehicle motion information, IMU information, HDmap information, INS information and perceived lane line information, and perform time synchronization processing on the vehicle motion information, IMU information, INS information and perceived lane line information;

[0009] The first observation information of the lane line is constructed by using the INS information and the HDmap information, and the second observation information of the lane line fitting is obtained by performing lane line extraction by image recognition of the sensed lane line information;

[0010] Input the sensed lane line information into a Kalman filter, perform lane line tracking prediction based on the IMU information, and optimize and smooth the lane line parameters in combination with the vehicle motion information to obtain prediction information;

[0011] During the tracking and prediction process of the lane line, the prediction information is verified based on the first observation information and the second observation information, and is fed back to the Kalman filter for iterative update.

[0012] This basic solution sets up an algorithm framework model based on the classic Kalman filter, and uses vehicle motion information, IMU information, HDmap information, INS information and perceived lane line information to optimize lane line accuracy, lane line detection stability and robustness, and detect abnormal lane line information at the same time; the main model is based on adaptive Kalman filtering combined with vehicle motion information to predict lane line information, combined with high-precision map lane lines, and perceived lane line information to achieve observation update compensation; the use of observation and prediction residual information to achieve anomaly detection can provide stable and reliable guarantees for all autonomous driving vehicles traveling on public roads.

[0013] In a further embodiment, the constructing the first lane line observation information using the INS information and the HDmap information includes:

[0014] Performing coordinate system conversion on the INS information to obtain the current positioning coordinates of the vehicle;

[0015] Determining lane information of the area where the vehicle is located according to the HDmap information;

[0016] Matching the lane information according to the positioning coordinates, and then determining the current driving lane of the vehicle based on a dichotomy method as a target lane;

[0017] The lane data of the target lane is obtained, and it is determined whether the lane data is valid. If so, lane line parameter fitting is performed to construct a vehicle driving road model, and the corresponding key parameters of the lane line equation are obtained as the first observation information.

[0018] This solution converts the INS information into a coordinate system to obtain the current positioning coordinates of the vehicle, and then determines the lane information of the vehicle's area based on the HDmap information; this is based on the combination of high-precision maps and vehicle GPS positioning technology, providing lane-level navigation information for driving, and the positioning of autonomous driving vehicles provides extremely high accuracy, so that correct observation data can be obtained as the basis for filter correction, thereby further optimizing the filter's prediction accuracy.

[0019] In a further embodiment, the coordinate system conversion of the INS information to obtain the current positioning coordinates of the vehicle includes:

[0020] Obtain the user's GPS location coordinates based on INS information;

[0021] Based on a universal coordinate conversion formula, the GPS positioning coordinates are converted into first coordinates in a geocentric coordinate system;

[0022] Based on the universal coordinate conversion formula, the first coordinate is converted from the geocentric coordinate system to the projection coordinate system to obtain the UTM coordinate as the current positioning coordinate of the vehicle.

[0023] In a further embodiment, the second observation information of lane line fitting is obtained by performing lane line extraction based on the perceived lane line information through image recognition, including:

[0024] Collect different original road images, generate image datasets and divide them into training sets and test sets;

[0025] Construct a deep learning network model, input the training set and the test set into the deep learning network model for optimization training, and obtain an optimized deep network learning model;

[0026] Inputting the sensed lane line information into the deep learning network model to obtain a feature map of the road image to be identified, fitting the feature map to obtain a recognition result map of the lane line to construct a vehicle driving road model, and outputting a confidence level;

[0027] It is determined whether the confidence level is greater than a preset threshold. If so, the corresponding lane line equation key parameters are obtained as the second observation information based on the corresponding vehicle driving road model.

[0028] This solution uses a deep learning model for image recognition to obtain the second observation information. Since the deep learning model can extract useful features from a large amount of data through learning, these features can help the model better adapt to different road environments and lighting conditions, and improve the accuracy and robustness of detection. Through its powerful learning and adaptability and efficient processing speed, it has shown significant advantages in the field of lane line detection, providing more reliable and real-time lane line detection for autonomous driving and assisted driving systems.

[0029] In a further implementation, the sensing lane line information is input into a Kalman filter, tracking and predicting the lane line is performed based on the IMU information, and the lane line parameters are optimized and smoothed in combination with the vehicle motion information to obtain prediction information, including:

[0030] Based on a universal coordinate system conversion formula, the sensed lane line information is converted into lane line data in a local coordinate system;

[0031] The lane line data is input into a Kalman filter, and key parameters of a lane line equation for constructing a vehicle driving road model are output as first prediction parameters;

[0032] Constructing a vehicle kinematics model according to the vehicle motion information and the IMU information;

[0033] updating the state transfer matrix and control matrix of the Kalman filter in combination with the vehicle kinematic model;

[0034] Based on the state transfer matrix and the control matrix, the lane line parameters are optimized and smoothed according to the first prediction parameter, and the current lane line information is predicted to obtain the prediction information.

[0035] In a further embodiment, the system model of the Kalman filter is as follows:

[0036] X t =AX t-1 +BU t-1

[0037] Where, X t , X t-1 are the state quantities at time t and time t-1 respectively, U t-1 is the control quantity at time t-1, A is the state transfer matrix, and B is the control matrix;

[0038] The state transfer matrix is ​​as follows:

[0039]

[0040] The control matrix is ​​as follows:

[0041]

[0042] Where, d x is the vehicle moving distance, w x is the vehicle yaw rate, and t is the time.

[0043] This solution uses the perceived lane line information, IMU information and vehicle motion information to construct the system equation of the filter, and realizes short-term prediction of the lane line based on the current motion trend of the vehicle, thereby effectively identifying abnormal lane lines and improving driving safety.

[0044] In a further embodiment, in the observation verification, it includes:

[0045] The observation noise is estimated based on the adaptive filtering model, where the adaptive filtering model is as follows:

[0046]

[0047] Where k is the current frame, k - 1 is the previous frame, and H k is the observation matrix, is the covariance prediction matrix of the corresponding parameters, and X k|k-1 is the predicted value of the lane line parameters, is the residual between the predicted value and the observed value, is the confidence level, and R k is the covariance matrix of the observation noise, is the i-th value in matrix R k and is the i-th value in matrix H k and Z k is the observed value at time k;

[0048] where 0 < b < 1, which is called the fading factor, and according to experience, b = 0.9 - 0.999.

[0049] In a further embodiment, in the observation verification, the observation model is as follows;

[0050] Y t = CX t-1

[0051]

[0052] where y pf represents the perceived lane line data, and y hd represents the lane line data of the high-precision map; C pf / C hd is the observation matrix, and R pf / R hd is the variance matrix of the observation state.

[0053] This solution introduces the perceived lane line information and the lane line information of the HDmap into the observation model of the filter, and adds the constrained lane line width information to realize the observation update compensation, performs the optimization and smoothing processing of the lane line parameters, and combines the actual observation data for prediction to be able to break away from the distance limitation, gives play to the over-the-horizon perception ability of the HDmap to improve the lane line accuracy, and the autonomous driving system can obtain higher positioning accuracy and stronger environmental perception ability, and further improve the lane line prediction ability and the lane line information anomaly detection performance.

[0054] In a further embodiment, the time synchronization processing is specifically as follows: Based on the Newton interpolation algorithm, data reconstruction is performed on the vehicle motion information, IMU information, INS information, and perceived lane line information, and the data times in the vehicle motion information, IMU information, INS information, and perceived lane line information are aligned on a benchmark.

[0055] The present invention also provides an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the lane line optimization method as described above when executing the computer program. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 is a flow chart of a lane line optimization method provided by an embodiment of the present invention;

[0057] Figure 2 It is a workflow diagram of a lane line optimization method provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0058] The following specifically illustrates the implementation mode of the present invention in conjunction with the accompanying drawings. The embodiments are provided for illustrative purposes only and cannot be understood as limiting the present invention. The accompanying drawings are provided for reference and illustration only and do not constitute a limitation on the scope of patent protection of the present invention, because many changes can be made to the present invention without departing from the spirit and scope of the present invention.

[0059] An embodiment of the present invention provides a lane line optimization method, such as Figure 1 , Figure 2 As shown, in this embodiment, steps S1 to S4 are included:

[0060] S1. Collect data to obtain vehicle motion information, IMU information, HDmap information, INS information and perceived lane line information, and perform time synchronization processing on the vehicle motion information, IMU information, INS information and perceived lane line information.

[0061] The vehicle motion information includes the vehicle motion distance and the vehicle yaw angular velocity.

[0062] In this embodiment, the time synchronization processing is specifically: reconstructing the vehicle motion information, IMU information, INS information and perceived lane line information based on the Newton interpolation algorithm, and aligning the data time in the vehicle motion information, IMU information, INS information and perceived lane line information to a benchmark.

[0063] In this embodiment, the vehicle driving road model is as follows:

[0064] y=c0+c1*x+c2*x 2 +c3*x 3

[0065] The key parameters of the classic lane line equation are (c0, c1, c2, c3), where c0 represents the lateral distance of the vehicle from the lane line, c1 represents the lane line heading angle, c2 represents the lane line curvature, and c3 represents the curvature change rate.

[0066] S2. Construct first lane line observation information using INS information and HDmap information, and extract lane lines by image recognition to obtain second lane line fitting observation information.

[0067] In this embodiment, the first lane line observation information is constructed using INS information and HDmap information, including:

[0068] S21A, performing coordinate system conversion on the INS information to obtain the current positioning coordinates of the vehicle, including:

[0069] S211A, obtaining the GPS positioning coordinates of the user based on the INS information;

[0070] S212A, based on a universal coordinate conversion formula, converting the GPS positioning coordinates into first coordinates in a geocentric coordinate system, for example, into a GCJ02 National Bureau of Surveying and Mapping coordinate system;

[0071] S212A: Based on a universal coordinate conversion formula, convert the first coordinate from the geocentric coordinate system to a projection coordinate system to obtain a UTM coordinate as the current positioning coordinate of the vehicle.

[0072] S22A, determining lane information of the area where the vehicle is located according to the HDmap information;

[0073] S23A, matching the lane information according to the positioning coordinates, and then determining the current driving lane of the vehicle based on a dichotomy method as a target lane;

[0074] S24A, obtaining lane data of the target lane, determining whether the lane data is valid, and if so, performing lane line parameter fitting to construct a vehicle driving road model, and obtaining corresponding key parameters of the lane line equation as first observation information.

[0075] The above general coordinate conversion formula is common knowledge in the art and will not be described in detail in this embodiment.

[0076] Among them, the method for obtaining the lane data of the target lane is specifically to provide least squares fitting data points to obtain the lane line data corresponding to the target lane as the lane data, and the lane line data includes the lane width and the lane line UTM coordinate points.

[0077] This embodiment converts the INS information into a coordinate system to obtain the current positioning coordinates of the vehicle, and then determines the lane information of the area where the vehicle is located based on the HDmap information; in this way, based on the combination of high-precision maps and vehicle GPS positioning technology, lane-level navigation information is provided for driving, and the positioning of the autonomous driving vehicle provides extremely high accuracy, so that the correct observation data can be obtained as the correction basis of the filter, thereby further optimizing the prediction accuracy of the filter.

[0078] In this embodiment, the second observation information of lane line fitting is obtained by performing lane line extraction through image recognition of the perceived lane line information, including:

[0079] S21B, collect different original road images, generate image data sets and divide them into training sets and test sets;

[0080] S22B, constructing a deep learning network model, inputting the training set and the test set into the deep learning network model for optimization training, and obtaining an optimized deep network learning model;

[0081] S23B, inputting the sensed lane line information into the deep learning network model, obtaining a feature map of the road image to be identified, fitting the feature map, obtaining a recognition result map of the lane line, constructing a vehicle driving road model, and outputting a confidence level;

[0082] S24B, determine whether the confidence level is greater than a preset threshold value. If so, obtain the corresponding lane line equation key parameters as the second observation information based on the corresponding vehicle driving road model.

[0083] This embodiment uses a deep learning model to perform image recognition to obtain the second observation information. Since the deep learning model can extract useful features from a large amount of data through learning, these features can help the model better adapt to different road environments and lighting conditions, and improve the accuracy and robustness of detection. Through its powerful learning and adaptability and efficient processing speed, it has demonstrated significant advantages in the field of lane line detection, and provides more reliable and real-time lane line detection for autonomous driving and assisted driving systems.

[0084] S3, inputting the sensed lane line information into a Kalman filter, performing lane line tracking prediction based on the IMU information, and optimizing and smoothing lane line parameters in combination with the vehicle motion information to obtain prediction information, including:

[0085] S31, based on a universal coordinate system conversion formula, converting the sensed lane line information into lane line data in a local coordinate system (e.g., a northeastern sky rectangular coordinate system, an ENU coordinate system);

[0086] S32, inputting the lane line data into a Kalman filter, and outputting key parameters of a lane line equation for constructing a vehicle driving road model as a first prediction parameter;

[0087] S33, constructing a vehicle kinematics model according to the vehicle motion information and the IMU information;

[0088] S34, updating the state transfer matrix and control matrix of the Kalman filter in combination with the vehicle kinematic model;

[0089] In this embodiment, the classic Kalman filter model is as follows:

[0090]

[0091] P k =(IK k H k ) k|k-1

[0092] In the formula, represents the previous state estimate, represents the system transfer matrix, represents the predicted value, P k|k-1 The covariance matrix of the corresponding parameters, K k represents the Kalman gain, H k is the observation matrix, I is the identity matrix, R k Covariance matrix of the observation noise.

[0093] The system model of the Kalman filter is as follows:

[0094] X t =AX t-1 +BU t-1

[0095] Where, X t , X t-1 are the state quantities at time t and time t-1 respectively, U t-1 is the control quantity at time t-1, A is the state transfer matrix, and B is the control matrix;

[0096] Among them, the state quantity X is as follows:

[0097]

[0098]

[0099] The state transfer matrix is ​​as follows:

[0100]

[0101] The control matrix is ​​as follows:

[0102]

[0103] d x =v x *dt

[0104] In the formula, v x is the vehicle speed, d x is the vehicle moving distance, w x is the vehicle yaw rate, and t is the time.

[0105] S35. Based on the state transfer matrix and the control matrix, the lane line parameters are updated according to the first prediction parameter, and the lane line information is predicted to obtain the prediction information.

[0106] This embodiment uses the perceived lane line information, IMU information and vehicle motion information to construct the system equation of the filter, and realizes short-term prediction of the lane line based on the current motion trend of the vehicle, thereby effectively identifying abnormal lane lines and improving driving safety.

[0107] S4. During the lane line tracking and prediction process, the prediction information is verified based on the first observation information and the second observation information, and fed back to the Kalman filter for iterative update.

[0108] In this embodiment, the observation model of the observation verification is as follows:

[0109] Y t =CX t-1

[0110]

[0111] In the formula, y pf Represents the perceived lane line data, y hd Represents the lane line data of the high-precision map; C pf / C hd is the observation matrix, R pf / R hd is the variance matrix of the observed state.

[0112] In this embodiment, the observation verification includes:

[0113] The observation noise is estimated based on the adaptive filtering model, where the adaptive filtering model is as follows:

[0114]

[0115] In the formula, k is the current frame, k-1 is the previous frame, and H k is the observation matrix, is the covariance prediction matrix for the corresponding parameters, X k|k-1 is the predicted value of the lane line parameters, is the residual between the predicted value and the observed value (i.e., the difference between the predicted lane line parameters and the lane line parameters obtained by fitting, which is the perceived lane line parameters), is the confidence level, R k is the covariance matrix of the observation noise, is the i-th value in matrix R k is the i-th value in matrix H k Z k is the observed value at time k;

[0116] In the formula, where 0 < b < 1, which is called the fading factor. According to experience, b = 0.9 - 0.999.

[0117] where the process noise w k and the observation noise v k covariance matrix Q k and R k respectively represent the degree of trust in the model and the observation. Their values indirectly affect the Kalman gain K k , so they have a greater impact on the accuracy of the filtering result; R k can be obtained from the accuracy parameters of the sensor, while Q k is usually set according to experience;

[0118] In this embodiment, the perception lane line information and the lane line information of the HDmap are introduced into the observation model of the filter, and the lane line width information with additional constraints is used to achieve observation update compensation, and the lane line parameter optimization and smoothing processing are carried out. Combining the actual observation data for prediction can break away from the distance limitation, give full play to the over-the-horizon perception ability of the HDmap to improve the lane line accuracy, and the autonomous driving system can obtain higher positioning accuracy and stronger environmental perception ability, further improving the lane line prediction ability and the lane line information anomaly detection performance.

[0119] In this embodiment, the steps of measurement update are as follows:

[0120] Based on the covariance matrices of the measurement noise and the process noise, a weight factor, i.e., the Kalman gain, is calculated;

[0121] Using the calculated Kalman gain and the new measurement value, the previous state estimate is corrected to obtain the updated state estimate value;

[0122] According to the updated state estimate and the new measurement information, the covariance matrix of the state estimate is adjusted to reflect the new uncertainty level.

[0123] ​This embodiment aims at detecting abnormal lane line information and proposes a robust adaptive Kalman filter algorithm to make a one-step prediction of the state based on the prior lane line information combined with the vehicle motion information, and to verify the current observation by comparing the predicted information with the current observation information; for scenes such as turning intersections and lane line disappearance, the vehicle motion information is used to realize the prediction of lane line information.

[0124] The embodiment of the present invention sets up an algorithm framework model based on classic Kalman filtering, and uses vehicle motion information, IMU information, HDmap information, INS information and perceived lane line information to optimize lane line accuracy, optimize lane line detection stability and robustness, and detect abnormal lane line information at the same time; the main model is based on adaptive Kalman filtering combined with vehicle motion information to predict lane line information, combined with high-precision map lane lines, and perceived lane line information to achieve observation update compensation; among them, the use of observation and prediction residual information to achieve anomaly detection can provide stable and reliable guarantees for all autonomous driving vehicles traveling on public roads.

[0125] Example 2

[0126] An embodiment of the present invention also provides an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of a lane line optimization method in the above-mentioned embodiment 1 when executing the computer program.

[0127] The above embodiments are preferred implementation modes of the present invention, but the implementation modes of the present invention are not limited to the above embodiments. Any other changes, modifications, substitutions, combinations, and simplifications that do not deviate from the spirit and principles of the present invention should be equivalent replacement methods and are included in the protection scope of the present invention.

Claims

1. A lane line optimization method, characterized in that: Includes steps: Carry out data collection to obtain vehicle motion information, IMU information, HDmap information, INS information and perceived lane line information, and perform time synchronization processing on the vehicle motion information, IMU information, INS information and perceived lane line information; The first observation information of the lane line is constructed by using the INS information and the HDmap information, and the second observation information of the lane line fitting is obtained by performing lane line extraction by image recognition of the perceived lane line information; Input the sensed lane line information into a Kalman filter, perform lane line tracking prediction based on the IMU information, and optimize and smooth the lane line parameters in combination with the vehicle motion information to obtain prediction information; During the tracking and prediction process of the lane line, the prediction information is verified based on the first observation information and the second observation information, and is fed back to the Kalman filter for iterative update.

2. A lane line optimization method according to claim 1, characterized in that: The method of constructing the first lane line observation information by using the INS information and the HDmap information includes: Performing coordinate system conversion on the INS information to obtain the current positioning coordinates of the vehicle; Determining lane information of the area where the vehicle is located according to the HDmap information; Matching the lane information according to the positioning coordinates, and then determining the current driving lane of the vehicle as the target lane based on the dichotomy method; The lane data of the target lane is obtained, and it is determined whether the lane data is valid. If so, lane line parameter fitting is performed to construct a vehicle driving road model, and the corresponding key parameters of the lane line equation are obtained as the first observation information.

3. A lane line optimization method as claimed in claim 2, characterized in that: The coordinate system conversion of the INS information to obtain the current positioning coordinates of the vehicle includes: Obtaining the GPS location coordinates of the user based on the INS information; Based on a universal coordinate conversion formula, the GPS positioning coordinates are converted into first coordinates in a geocentric coordinate system; Based on the universal coordinate conversion formula, the first coordinate is converted from the geocentric coordinate system to the projection coordinate system to obtain the UTM coordinate as the current positioning coordinate of the vehicle.

4. A lane line optimization method as claimed in claim 1, characterized in that: The second observation information of lane line fitting is obtained by performing lane line extraction through image recognition of the sensed lane line information, including: Collect different original road images, generate image datasets and divide them into training sets and test sets; Construct a deep learning network model, input the training set and the test set into the deep learning network model for optimization training, and obtain an optimized deep network learning model; Inputting the sensed lane line information into the deep learning network model to obtain a feature map of the road image to be identified, fitting the feature map to obtain a recognition result map of the lane line to construct a vehicle driving road model, and outputting a confidence level; It is determined whether the confidence level is greater than a preset threshold. If so, the corresponding lane line equation key parameters are obtained as the second observation information based on the corresponding vehicle driving road model.

5. A lane line optimization method as claimed in claim 4, characterized in that: The sensed lane line information is input into a Kalman filter, and the lane line is tracked and predicted based on the IMU information. The lane line parameters are optimized and smoothed in combination with the vehicle motion information to obtain prediction information, including: Based on a universal coordinate system conversion formula, the sensed lane line information is converted into lane line data in a local coordinate system; The lane line data is input into a Kalman filter, and key parameters of a lane line equation for constructing a vehicle driving road model are output as first prediction parameters; Constructing a vehicle kinematics model according to the vehicle motion information and the IMU information; updating the state transfer matrix and control matrix of the Kalman filter in combination with the vehicle kinematic model; Based on the state transfer matrix and the control matrix, the lane line parameters are optimized and smoothed according to the first prediction parameter, and the current lane line information is predicted to obtain the prediction information.

6. A lane line optimization method according to claim 5, characterized in that: The system model of the Kalman filter is as follows: X t =AX t-1 +BU t-1 Where, X t , X t-1 are the state quantities at time t and time t-1 respectively, U t-1 is the control quantity at time t-1, A is the state transfer matrix, and B is the control matrix; The state transfer matrix is ​​as follows: The control matrix is ​​as follows: Where, d x is the vehicle moving distance, w x is the vehicle yaw rate, and t is the time.

7. The lane line optimization method according to claim 1, characterized in that: The observation verification includes: The observation noise is estimated based on the adaptive filtering model, where the adaptive filtering model is as follows: In the formula, k is the current frame, k-1 is the previous frame, and H k is the observation matrix, is the covariance prediction matrix of the corresponding parameters, X k|k-1 is the predicted value of the lane line parameter, is the residual between the predicted value and the observed value, is the confidence level, R k The covariance matrix of the observation noise, is the matrix R k The i-th value in is the matrix H k The i-th value in Z k is the observed value at time k; In the formula, Where 0 <b<1,β k It is called the fading factor.

8. A lane line optimization method according to claim 7, characterized in that: In the observation verification, the observation model is as follows; Y t =CX t-1 In the formula, y pf Represents the perceived lane line data, y hd Represents the lane line data of the high-precision map; C pf / C hd is the observation matrix, R pf / R hd is the variance matrix of the observed state.

9. The lane line optimization method according to claim 1, characterized in that: The time synchronization processing is specifically as follows: based on the Newton interpolation algorithm, data of the vehicle motion information, IMU information, INS information and the perceived lane line information is reconstructed, and the data time in the vehicle motion information, IMU information, INS information and the perceived lane line information is aligned to a benchmark.

10. An electronic device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the steps of a lane optimization method as described in any one of claims 1 to 9 are implemented.

Images