Method for automatic calibration of mounting angle

Abstract

The present application relates to the technical field of millimeter wave radar, in particular to a kind of automatic calibration method of installation angle;It is suitable for the installation angle of vehicle-mounted millimeter wave angle radar is automatically calibrated;Collect the radar data of vehicle-mounted millimeter wave angle radar, obtain the standard point cloud data of each frame radar data and store in point cloud data group;When the standard point cloud data stored in the point cloud data group reaches the preset frame number, the installation angle error of the vehicle-mounted millimeter wave angle radar is obtained by processing the point cloud data group;According to the installation angle error obtained by processing multiple times, the installation angle error wave gate is obtained, and the installation angle of the vehicle-mounted millimeter wave angle radar is automatically calibrated according to the installation angle error wave gate;Advantages: real-time calibration during vehicle driving process;Angle error range can be dynamically adjusted, so that the calibration value converges;It can be judged that the position of radar may have been shifted, and large-scale shift caused by collision and jolt can be calibrated.

Description

Technical Field

[0001] This invention relates to the field of millimeter-wave radar technology, and more specifically to an automatic calibration method for installation angle. Background Technology

[0002] Millimeter-wave radar sensors have become one of the mainstream detection sensors due to their moderate cost, strong environmental adaptability, and good long-range detection capabilities. Existing calibration methods for installation angle errors mainly include: calibration using the error between the true angle of a calibration object (such as an angle reflector) and the radar measurement angle; calibration using straight objects (such as railings) near the vehicle during vehicle movement; and calibration using the relationship between the vehicle speed and the radial velocity of a stationary object during vehicle movement.

[0003] In practical methods for calibrating installation angle errors, using calibration objects cannot meet the requirements for real-time and dynamic calibration; using straight objects for calibration requires deliberately finding long railings, and straight-line fitting sometimes has errors; using the relationship between vehicle speed and the radial velocity of a stationary object for calibration requires accurate vehicle speed information, and when the vehicle speed information is inaccurate, it is necessary to fit the vehicle speed information first, thus increasing the complexity of the calculation. Summary of the Invention

[0004] The purpose of this invention is to provide an automatic calibration method for installation angle, thereby solving the above-mentioned technical problems;

[0005] The technical problem solved by this invention can be achieved by the following technical solutions:

[0006] An automatic calibration method for installation angle, applicable to the automatic calibration of the installation angle of vehicle-mounted millimeter-wave angle radar; comprising:

[0007] Step S1: Collect radar data from the vehicle-mounted millimeter-wave angle radar, obtain standard point cloud data for each frame of radar data, and store it in a point cloud data group.

[0008] Step S2: When the standard point cloud data stored in the point cloud data group reaches a preset number of frames, the installation angle error of the vehicle-mounted millimeter-wave angle radar is obtained by processing the point cloud data group.

[0009] Step S3: Based on the installation angle error obtained from multiple processing steps, an installation angle error gate is obtained. The installation angle of the vehicle-mounted millimeter-wave angle radar is automatically calibrated based on the installation angle error gate, and then the process returns to step S1.

[0010] Preferably, step S1 specifically includes:

[0011] Step S11: Collect the radar data;

[0012] Step S12: For each frame of radar data, the point cloud data that meets the preset conditions in the current frame of radar data is used as standard point cloud data and stored in the point cloud data group.

[0013] Preferably, the preset conditions include:

[0014] The radial velocity of the point cloud data relative to the vehicle is 0;

[0015] At the moment corresponding to the radar data in the current frame, the vehicle is in motion;

[0016] At the moment corresponding to the radar data in the current frame, the vehicle is traveling in a straight line;

[0017] The distance between the point cloud data and the vehicle is within a preset range; and

[0018] The angle data of the point cloud data is within a preset angle threshold range.

[0019] Preferably, in a frame of radar data, if there are multiple point cloud data that meet the preset conditions, the point cloud data closest to the vehicle is used as the standard point cloud data and stored in the point cloud data group.

[0020] Preferably, step S2 specifically includes:

[0021] Step S21: Process to obtain the mean of the point cloud data group, and process based on the mean to obtain the standard deviation of the point cloud data group;

[0022] Step S22: Remove standard point cloud data with angles greater than 3 times the standard deviation from the point cloud data group to obtain the filtered point cloud data group.

[0023] Step S23: Process the filtered point cloud data group to obtain the mean value as the installation angle error;

[0024] Step S24: Output the installation angle error and clear the point cloud data group.

[0025] Preferably, step S3 specifically includes:

[0026] Step S31: Store the received installation angle error into a calibration history frame data group;

[0027] Step S32: When the installation angle error in the calibration history frame data group reaches a first preset value, the standard deviation of the calibration history frame data group is obtained, and the angle threshold range is adjusted according to the standard deviation.

[0028] Preferably, in step S32, one of the multiple preset standard angle threshold ranges is matched according to the standard deviation to adjust the current angle threshold range.

[0029] Preferably, in step S31, if the number of consecutive frames without receiving the installation angle error reaches the second preset value, the angle threshold range is directly expanded.

[0030] Preferably, the point cloud data in the radar data includes at least:

[0031] The distance between the location of the point cloud data and the vehicle;

[0032] The location of the point cloud data is relative to the radial velocity of the vehicle;

[0033] The azimuth angle between the location of the point cloud data and the vehicle.

[0034] A vehicle-mounted millimeter-wave angle radar, in actual operation, uses the aforementioned automatic calibration method of installation angle to automatically calibrate the angle threshold range.

[0035] Beneficial effects: Due to the adoption of the above technical solutions, the present invention can perform real-time calibration during vehicle operation; it can dynamically adjust the angle error range of the selected data to make the calibration value converge; if no data meets the angle threshold range for a long time, it is assumed that the radar position may have shifted, thereby expanding the angle threshold range and calibrating large-scale shifts caused by collisions and bumps. Attached Figure Description

[0036] Figure 1 This is a flowchart of the main steps in an embodiment of the present invention;

[0037] Figure 2 This is a flowchart illustrating the specific steps of S1 in an embodiment of the present invention;

[0038] Figure 3 This is a flowchart of point cloud data filtering in an embodiment of the present invention;

[0039] Figure 4 This is a flowchart illustrating the specific steps of S2 in an embodiment of the present invention;

[0040] Figure 5 This is a flowchart illustrating the installation angle calculation in an embodiment of the present invention;

[0041] Figure 6 This is a flowchart illustrating the specific steps of S3 in an embodiment of the present invention;

[0042] Figure 7 This is a flowchart of the angle threshold update process in an embodiment of the present invention. Detailed Implementation

[0043] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0044] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other.

[0045] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, but this is not intended to limit the scope of the invention.

[0046] In a preferred embodiment of the present invention, an automatic calibration method for installation angle is provided, the main steps of which are shown in the flowchart below. Figure 1 As shown, this is suitable for automatically calibrating the installation angle of vehicle-mounted millimeter-wave angle radar; including:

[0047] Step S1: Collect radar data from the vehicle-mounted millimeter-wave angle radar, obtain standard point cloud data for each frame of radar data, and store it in the point cloud data group.

[0048] Step S2: When the standard point cloud data stored in the point cloud data group reaches the preset number of frames, the installation angle error of the vehicle millimeter-wave angle radar is obtained by processing the point cloud data group.

[0049] Step S3: Based on the installation angle error obtained from multiple processing steps, an installation angle error gate is obtained. The installation angle of the vehicle-mounted millimeter-wave angle radar is automatically calibrated based on the installation angle error gate, and then the process returns to step S1.

[0050] Furthermore, in a preferred embodiment of the present invention, the specific steps of step S1 are as follows: Figure 2 As shown, it includes:

[0051] Step S11: Collect radar data;

[0052] Step S12: For each frame of radar data, select the point cloud data that meets the preset conditions in the current frame of radar data as standard point cloud data and store it in the point cloud data group.

[0053] Specifically, in step S12, the point cloud data in the radar data includes at least the following information:

[0054] The distance between the location of the point cloud data and the vehicle;

[0055] The location of the point cloud data relative to the radial velocity of the vehicle;

[0056] The azimuth angle between the location of the point cloud data and the vehicle.

[0057] In this embodiment, the preset conditions for step S12 include:

[0058] The radial velocity of the point cloud data relative to the vehicle is 0;

[0059] At the moment corresponding to the radar data in the current frame, the vehicle is in motion;

[0060] At the moment corresponding to the radar data in the current frame, the vehicle is traveling in a straight line;

[0061] The distance between the point cloud data and the vehicle is within a preset range; and

[0062] The angle data of the point cloud data is within the preset angle threshold range.

[0063] The first of the above preset conditions is that the radial velocity of the point cloud data relative to the vehicle is 0, that is, the point cloud data may be in the following two states: 1) a moving object that is relatively stationary with respect to the vehicle (e.g., other moving objects in the outside world that are close to the vehicle's speed); 2) a stationary object that is perpendicular to the vehicle's direction of motion (e.g., a wall, tree, or fence located in front of the vehicle or behind the vehicle).

[0064] The second point in the above preset conditions: at the time corresponding to the radar data of the current frame, the vehicle is in motion. The judgment criterion can be set as the vehicle's current real-time speed being greater than a set speed threshold, which can be set as 5m / s or other feasible speed values.

[0065] The third point in the above preset conditions: at the time corresponding to the radar data of the current frame, the vehicle is in a straight-line state, which can be detected by the vehicle's yaw rate. That is, when the vehicle's yaw rate = 0, it indicates that the vehicle is in a straight-line state.

[0066] The fourth point in the above preset conditions: the distance between the point cloud data and the vehicle is within a preset range. This condition indicates that the selected point cloud data should not be too close to the vehicle (too close a point may be a reflection point of the vehicle's bumper), nor too far from the vehicle (too far a point will result in a lower signal-to-noise ratio, thus reducing detection accuracy). Preferably, the above preset range can be set to a distance of 8-10m from the vehicle.

[0067] Furthermore, if multiple point cloud data sets meet the preset conditions, the point cloud data set closest to the vehicle is selected as the standard point cloud data set and stored in the point cloud data set. Alternatively, multiple point cloud data sets meeting the preset conditions can be filtered based on information such as energy or signal-to-noise ratio to obtain a single point cloud data set as the standard point cloud data set and stored in the point cloud data set.

[0068] The fifth of the above preset conditions: the angle data of the point cloud data is within the preset angle threshold range. The initial range of this angle threshold range can be set relatively large and can be adjusted adaptively in the future.

[0069] In this embodiment, the angle threshold range can be set to one of [-1°, 1°], [-3°, 3°], [-5°, 5°] and [-15°, 15°]. Initially, a relatively large threshold range, such as [-15°, 15°], can be selected. Subsequently, the angle threshold range can be adaptively adjusted, that is, a smaller angle threshold range can be selected.

[0070] The sampling interval for the radar data can be 0.1s, that is, one frame of radar data is collected every 0.1s.

[0071] In this embodiment, the point cloud data in each frame of radar data can be judged by setting multiple pointer markers. Specifically, it can be done as follows: Figure 3 As shown, Figure 3 The process of judging all point cloud data in the radar data of the current frame includes:

[0072] Three pointer flags are pre-set: MeasureID, MeasureCnt, and minRange. The initial value of MeasureID is -1, the initial value of MeasureCnt is 0, and no initial value is set for minRange.

[0073] The first part of the above process is to determine the first point cloud data that meets the preset conditions. Specifically, it is first determined whether the value of MeasureCnt is less than the total number of point cloud data that need to be judged in the current frame of radar data (i.e., whether there are still unjudged point cloud data in the current frame of radar data).

[0074] Since the initial value of MeasureCnt is 0, the first judgment of MeasureCnt will always result in a value less than the total amount of point cloud data. Then, it will further determine whether the MeasureCnt-th point cloud data meets the preset conditions. If it does, it means the first point cloud data meeting the preset conditions has been found. At this point, the distance between this point cloud data and the vehicle is assigned to minRange, and MeasureID = MeasureCnt is set. Otherwise, MeasureCnt is incremented by 1, and the next point cloud data is judged to meet the preset conditions. This process is repeated until the first point cloud data meeting the preset conditions is found, or all point cloud data in the current frame's radar data has been judged once (i.e., there is no point cloud data meeting the preset conditions in the current frame's radar data), and then the next part of the processing begins.

[0075] Before proceeding to the next step of processing, we first determine whether the current MeasureID value is equal to the initial value, that is, whether MeasureID is equal to -1. This process is to determine whether at least one point cloud data that meets the preset conditions has been found in the radar data of the current frame.

[0076] If the value of MeasureID is equal to -1, then the value of MeasureID is output directly, that is, the result of -1 is output directly. This result is used to indicate that there is no point cloud data that meets the preset conditions in the radar data of the current frame.

[0077] If the value of MeasureID is not equal to -1, then MeasureCnt is assigned the current MeasureID + 1, and then proceeds to the next part, which is the process of determining whether there are multiple point cloud data that meet the preset conditions after the first point cloud data that meets the preset conditions has been determined, and the process of filtering the point cloud data according to distance when there are multiple point cloud data that meet the preset conditions. This part of the process is described in detail below:

[0078] First, we still check whether the current value of MeasureCnt is less than the total number of point cloud data in the current frame of radar data, and then determine whether all the point cloud data in the current frame of radar data has been evaluated:

[0079] If the point cloud data in the current frame of radar data has been completely judged, the current MeasureID value is directly output. The point cloud data corresponding to this value is the final filtered point cloud data.

[0080] If the point cloud data in the current frame of radar data has not yet been completely evaluated, the system continues to evaluate whether the current point cloud data to be evaluated meets the conditions. It's important to note that, compared to the first part, a distance-based condition is added here. Specifically, the current point cloud data to be evaluated is considered to meet the preset conditions only if it simultaneously satisfies the aforementioned preset conditions in the first part and the condition that its distance from the vehicle is less than `minRange`. Further, if the current point cloud data to be evaluated meets the preset conditions and is closer to the vehicle, then `MeasureID` is set to `MeasureCnt`, the value of `minRange` is updated to the distance between the current point cloud data and the vehicle, `MeasureCnt` is incremented by 1, and the system returns to evaluate the next point cloud data. If the current point cloud data does not meet the preset conditions, or although it meets the preset conditions, its distance from the vehicle is greater, then `MeasureCnt` is incremented by 1, and the system returns to evaluate the next point cloud data. After all point cloud data in the current frame of radar data has been evaluated, the current `MeasureID` is output as the result, meaning the `MeasureID`-th point cloud data is added to the point cloud data group as the filtered standard point cloud data.

[0081] In other words, in the second part of the above processing, after obtaining the first point cloud data that meets the preset conditions, it attempts to find in the remaining point cloud data whether there is point cloud data that meets the preset conditions and is closer to the vehicle. After iterative calculation and comparison, the final output is the point cloud data that meets the preset conditions and is closest to the vehicle as the standard point cloud data.

[0082] In a preferred embodiment of the present invention, the specific steps of step S2 are as follows: Figure 4 As shown, it includes:

[0083] Step S21: Process the point cloud data set to obtain the mean, and process the mean to obtain the standard deviation of the point cloud data set.

[0084] Step S22: Remove standard point cloud data with angles greater than 3 times the standard deviation from the point cloud data group to obtain the filtered point cloud data group.

[0085] Step S23: Process the filtered point cloud data sets to obtain the mean value as the installation angle error.

[0086] Step S24: Output the installation angle error and clear the point cloud data group.

[0087] In this embodiment, the triggering condition for step S2 is "when the standard point cloud data stored in the point cloud data group reaches a preset number of frames". The preset number of frames can be set to 50 frames, that is, point cloud data has been detected in 50 frames of radar data. At this time, the offset angle can be calculated.

[0088] In this embodiment, after step S2 is triggered, the mean and standard deviation of all standard point cloud data in the point cloud data group are first calculated. Then, based on the mean and standard deviation, data in the point cloud data group that does not meet the 3σ principle are removed, that is, standard point cloud data with an angle greater than 3 times the standard deviation are removed. Finally, the mean of the standard point cloud data retained after the removal operation is calculated to obtain the mean result. The mean result is the installation angle error corresponding to the point cloud data group, that is, the installation angle error corresponding to the sampling time period of the 50 frames of radar data.

[0089] Then, clear the point cloud data set to prepare for the calculation of the next installation angle error.

[0090] In this embodiment, the installation angle error of the point cloud data set can be calculated by setting multiple pointer markers, specifically as follows: Figure 5 As shown in the image. It is worth noting that... Figure 5 The steps mentioned refer to the acquisition time of the radar data in the current frame, and include:

[0091] Three pointer flags are pre-set: OffsetAngle_Cal, the output flag, and the flag for the number of data frames in the point cloud data group, and the initial values ​​of all of them are set to 0;

[0092] To calculate the installation angle error, multiple standard point cloud data sets must first be stored in a point cloud data group before calculation. This process is described in detail below:

[0093] First, we check if the value of MeasureID is not equal to -1. This process is to determine whether standard point cloud data has been found in the radar data of the current frame.

[0094] If the value of MeasureID is equal to -1, the current value of OffsetAngle_Cal will be output directly, that is, the result of OffsetAngle_Cal = 0 will be output. This indicates that there is no installation angle error output at the time of acquisition of the radar data in the current frame, so the installation angle cannot be automatically calibrated.

[0095] If the value of MeasureID is not equal to -1, the current standard point cloud data is stored in the point cloud data group, and the data frame number of the point cloud data group is incremented by 1.

[0096] Then determine whether the number of data frames in the point cloud data group has reached the update threshold, that is, whether the preset number of frames has been reached.

[0097] If the number of data frames in the point cloud data group does not reach the preset number of frames, the current value of OffsetAngle_Cal will be output directly, that is, the result of OffsetAngle_Cal=0 will be output. This indicates that there is no installation angle error output at the time of acquisition of the radar data in the current frame, and therefore the installation angle cannot be automatically calibrated.

[0098] If the number of data frames in the point cloud data group reaches the preset number of frames, the output flag is set to 1; then the calculation of the installation angle error begins, and this process is described in detail below:

[0099] First, calculate the mean and standard deviation of the point cloud data set;

[0100] Then, based on the mean and standard deviation, data that does not meet the 3σ principle in the point cloud data set is removed, that is, standard point cloud data with an angle greater than 3 times the standard deviation are removed; the mean of the point cloud data set remaining after the removal operation is recalculated, and OffsetAngle_Cal is assigned the final mean value, which is the installation angle error.

[0101] Finally, the number of data frames in the point cloud data group is cleared to 0, the counting starts again, and then the value of OffsetAngle_Cal is output, which is the installation angle error at the time of acquisition of the radar data in the current frame.

[0102] In a preferred embodiment of the present invention, the specific steps of step S3 are as follows: Figure 6 As shown, it includes:

[0103] Step S31: Store the received installation angle error into a calibration history frame data group;

[0104] Step S32: When the installation angle error in the calibration history frame data group reaches the first preset value, the standard deviation of the calibration history frame data group is obtained, and the angle threshold range is adjusted according to the standard deviation.

[0105] Specifically, in this embodiment, if the installation angle error in the calibration history frame data group has reached the first preset value (the first preset value can be set to 10), it indicates that automatic calibration of the installation angle can be performed. At this time, the standard deviation of all installation angle errors in the calibration history frame data group is calculated to obtain the standard deviation calculation result.

[0106] Finally, the angle threshold range is adjusted based on the standard deviation calculation result. Specifically, based on the standard deviation calculation result, one of the multiple angle threshold ranges preset above is selected as the current angle threshold range, thereby completing the automatic calibration of the installation angle. Then, the process returns to step S1 to continue collecting radar data, thus enabling the angle threshold range to be adjusted automatically in a cyclical manner.

[0107] In addition, in this embodiment, in step S31 above, if the number of consecutive frames without receiving installation angle error data reaches a second preset value, the angle threshold range is directly expanded. This second preset value can be set to 2000 frames, that is, if no installation angle error data output is received for 2000 consecutive frames, the angle threshold range is directly expanded by one level. For example, according to the preset situation above, the current angle threshold range is [-5°, 5°], then if no installation angle error data output is received for 2000 consecutive frames, it is directly expanded to [-15°, 15°].

[0108] The specific execution method of the above judgment process can be as follows: In step S3 above, if there is no angle error output installed in the current frame, the count is incremented by 1, and when the continuous count reaches 2000 frames, the angle threshold range is directly expanded by one level.

[0109] It should be noted that due to the concept of "continuous counting", if there is an installation angle error output in the current frame, the count will be reset to zero and the counting will start again.

[0110] In this embodiment, the update of the angle threshold range is determined by setting multiple pointer markers, specifically as follows: Figure 7 It is worth noting that... Figure 7 The steps mentioned refer to the acquisition time of the radar data in the current frame, and include:

[0111] Multiple pointer flags are pre-defined: AssociationGateID, OutPutAngleCycleNum, and MaintainCnt. The initial value of MaintainCnt is set to 0, the initial value of OutPutAngleCycleNum is set to 0, and an array OutPutAngleCycle is pre-defined.

[0112] AssociationGateID represents the number of the angle threshold range. The larger the angle threshold number, the larger the angle threshold range. In other words, increasing the angle threshold range number is equivalent to expanding the angle threshold range level. The initial value of AssociationGateID is the number of the currently set angle threshold range.

[0113] First, check if the current output flag is equal to 1. Specifically, the output flag is assigned the value 1 for each installation angle error output.

[0114] If the current output flag is not equal to 1, MaintainCnt is incremented by 1. The process after MaintainCnt is incremented by 1 will be described in detail below.

[0115] If the current output flag is equal to 1, the output flag is first cleared to 0, and then the previously calculated installation angle error is stored in the OutPutAngleCycle array (i.e., the calibration history frame data group). Each time an installation angle error is stored, OutPutAngleCycleNum is incremented by 1. OutPutAngleCycleNum represents the amount of data in OutPutAngleCycle.

[0116] Then, it is determined whether OutPutAngleCycleNum has reached the threshold, that is, whether the first preset value has been reached. The purpose of setting the first preset value is to obtain multiple consecutive installation angle errors and recalculate them in order to complete the purpose of automatic adjustment.

[0117] If OutPutAngleCycleNum does not reach the first preset value, the angle threshold range is updated based on the current AssociationGateID.

[0118] If OutPutAngleCycleNum reaches the first preset value, that is, multiple consecutive installation angle errors are stored, then the standard deviation is calculated based on the installation angle errors in OutPutAngleCycle. Based on the result of the standard deviation, AssociationGateID is updated, and then the angle threshold range is updated. The angle threshold range can only be one of the previously defined threshold ranges.

[0119] Finally, both OutPutAngleCycleNum and MaintainCnt are set to 0, restarting the counting process to prepare for the next angle threshold update.

[0120] Furthermore, the process of MaintainCnt incrementing by 1 as described above is as follows:

[0121] MaintainCnt is used to represent the number of consecutive frames in which the installation angle error has not been received. It must be a consecutive number of frames in which the error has not been received. If there is installation angle error data output, MaintainCnt is cleared to 0. The number of frames in which the installation angle error has not been received starts from the next frame.

[0122] Determine whether MaintainCnt has reached the threshold, which is the second preset value.

[0123] If MaintainCnt does not reach the second preset value, the angle threshold range can be updated based on the current AssociationGateID. It's important to note that if no installation angle error output data is received in the current frame, the AssociationGateID will not change. Therefore, the so-called "updating the angle threshold range based on the current AssociationGateID" actually does not update the angle threshold range; it simply maintains the current angle threshold range.

[0124] If MaintainCnt reaches the second preset value, that is, if no installation angle error data is received within 2000 consecutive frames, it indicates that standard point cloud data cannot be found within the current angle threshold range.

[0125] At this point, we continue to check whether AssociationGateID is less than the maximum angle threshold number.

[0126] If AssociationGateID is already the largest angle threshold number, it means that the current angle threshold range is already the largest and cannot be expanded further. In this case, the point cloud data in the radar data will continue to be filtered using this largest angle threshold.

[0127] If AssociationGateID is less than the maximum angle threshold number, it indicates that the current angle threshold range is not the maximum. In order to obtain standard point cloud data from radar data, the angle threshold range needs to be expanded. That is, AssociationGateID is incremented by 1, and the angle threshold range is expanded by one level according to the updated AssociationGateID. MaintainCnt is then cleared to start counting again.

[0128] The above process is executed cyclically for each frame of radar data acquisition time, in order to achieve the invention objective of cyclically and automatically adjusting the angle threshold range.

[0129] In a preferred embodiment of the present invention, a vehicle-mounted millimeter-wave angle radar is also provided. The vehicle-mounted millimeter-wave angle radar is installed on a vehicle, and during the actual driving process of the vehicle, the angle threshold range is automatically adjusted using the automatic calibration method of the installation angle described above.

[0130] The above description is merely a preferred embodiment of the present invention and does not limit the implementation and protection scope of the present invention. Those skilled in the art should realize that any equivalent substitutions and obvious changes made based on the description and illustrations of the present invention should be included within the protection scope of the present invention.

Claims

1. An automatic calibration method for installation angle, applicable to the automatic calibration of the installation angle of a vehicle-mounted millimeter-wave angle radar; characterized in that, include: Step S1: Collect radar data from the vehicle-mounted millimeter-wave angle radar, obtain standard point cloud data for each frame of radar data, and store it in a point cloud data group. Step S2: When the standard point cloud data stored in the point cloud data group reaches a preset number of frames, the installation angle error of the vehicle-mounted millimeter-wave angle radar is obtained by processing the point cloud data group. Step S3: Based on the installation angle error obtained from multiple processing steps, an installation angle error gate is obtained. The installation angle of the vehicle-mounted millimeter-wave angle radar is automatically calibrated based on the installation angle error gate. Then, the process returns to step S1. Step S1 specifically includes: Step S11: Collect the radar data; Step S12: For each frame of radar data, the point cloud data that meets the preset conditions in the current frame of radar data is used as standard point cloud data and stored in the point cloud data group. The preset conditions include: The radial velocity of the point cloud data relative to the vehicle is 0; At the moment corresponding to the radar data in the current frame, the vehicle is in motion; At the moment corresponding to the radar data in the current frame, the vehicle is traveling in a straight line; The distance between the point cloud data and the vehicle is within a preset range; and The angle data of the point cloud data is within a preset angle threshold range.

2. The automatic calibration method according to claim 1, characterized in that, If there are multiple point cloud data that meet the preset conditions in a single frame of radar data, the point cloud data closest to the vehicle is used as the standard point cloud data and stored in the point cloud data group.

3. The automatic calibration method according to claim 1, characterized in that, Step S2 specifically includes: Step S21: Process to obtain the mean of the point cloud data group, and process according to the mean to obtain the standard deviation of the point cloud data group; Step S22: Remove standard point cloud data with angles greater than 3 times the standard deviation from the point cloud data group to obtain the filtered point cloud data group. Step S23: Process the filtered point cloud data group to obtain the mean value as the installation angle error; Step S24: Output the installation angle error and clear the point cloud data group.

4. The automatic calibration method according to claim 3, characterized in that, Step S3 specifically includes: Step S31: Store the received installation angle error into a calibration history frame data group; Step S32: When the installation angle error in the calibration history frame data group reaches a first preset value, the standard deviation of the calibration history frame data group is obtained, and the angle threshold range is adjusted according to the standard deviation.

5. The automatic calibration method according to claim 4, characterized in that, In step S32, based on the standard deviation, one of the multiple preset standard angle threshold ranges is matched to adjust the current angle threshold range.

6. The automatic calibration method according to claim 4, characterized in that, In step S31, if the number of consecutive frames without receiving the installation angle error reaches the second preset value, the angle threshold range is directly expanded.

7. The automatic calibration method according to claim 1, characterized in that, The point cloud data in the radar data includes at least: The distance between the location of the point cloud data and the vehicle; The location of the point cloud data is relative to the radial velocity of the vehicle; The azimuth angle between the location of the point cloud data and the vehicle.

8. A vehicle-mounted millimeter-wave angle radar, characterized in that, In actual operation, the automatic calibration method for the installation angle as described in any one of claims 1-7 is used to automatically calibrate the angle threshold range.

Images