WO2022179566A1 - 外参标定方法、装置、电子设备及存储介质 - Google Patents

外参标定方法、装置、电子设备及存储介质 Download PDF

Info

Publication number
WO2022179566A1
WO2022179566A1 PCT/CN2022/077698 CN2022077698W WO2022179566A1 WO 2022179566 A1 WO2022179566 A1 WO 2022179566A1 CN 2022077698 W CN2022077698 W CN 2022077698W WO 2022179566 A1 WO2022179566 A1 WO 2022179566A1
Authority
WO
WIPO (PCT)
Prior art keywords
point cloud
cloud data
data
ground
point
Prior art date
Application number
PCT/CN2022/077698
Other languages
English (en)
French (fr)
Inventor
马政
闫国行
石建萍
Original Assignee
上海商汤智能科技有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 上海商汤智能科技有限公司 filed Critical 上海商汤智能科技有限公司
Publication of WO2022179566A1 publication Critical patent/WO2022179566A1/zh

Links

Images

Classifications

    • GPHYSICS
    • G01MEASURING; TESTING
    • G01SRADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S7/00Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
    • G01S7/02Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00
    • G01S7/40Means for monitoring or calibrating

Definitions

  • the present disclosure relates to the technical field of radar equipment, and in particular, to an external parameter calibration method, device, electronic device and storage medium.
  • Radar is an electronic device that uses electromagnetic waves to detect targets.
  • the radar emits electromagnetic waves to irradiate the target and receives its echoes, thereby obtaining information such as the distance, distance change rate, azimuth, and altitude from the target to the electromagnetic wave emission point. Due to the wide detection range and high detection accuracy of radar, radar has become a commonly used sensor in unmanned vehicle technology and assisted driving technology.
  • the radar can be installed at the top position of the vehicle.
  • the radar wiring harness and the installation viewing angle there is a blind spot for the vehicle only after the radar is installed.
  • multiple radars can be installed at different positions of the vehicle, so that multiple radars work together to complete the detection of the surrounding environment of the vehicle.
  • the present disclosure provides at least one external parameter calibration method, device, electronic device and storage medium.
  • the present disclosure provides a method for calibrating external parameters, including: acquiring point cloud data obtained by at least two radar devices disposed on a target vehicle and respectively collecting the same target scene; based on a first point collected by a first radar device The cloud data and the second point cloud data collected by the second radar device generate rotation angle data and displacement data; wherein, the rotation angle data is used to represent the correspondence between the first point cloud data and the second point cloud data The angle deviation of the points in different directions, the displacement data is used to represent the displacement deviation of the corresponding points in the first point cloud data and the second point cloud data in different directions, wherein the first radar The device and the second radar device are any two different radar devices among the at least two radar devices; the first radar device and the second radar device are determined based on the rotation angle data and the displacement data Coordinate system transformation matrix between devices.
  • the point cloud data obtained from the same target scene collected by at least two radar devices installed on the target vehicle are acquired, and the rotation angle data and displacement data are generated based on the point cloud data collected by different radar devices, and then the rotation angle is used.
  • Data and displacement data determine the coordinate system transformation matrix between different radar devices, and realize automatic external parameter calibration between different radar devices. Compared with the manual calibration process, the efficiency and accuracy of external parameter calibration are improved.
  • the generating rotation angle data and displacement data based on the first point cloud data collected by the first radar device and the second point cloud data collected by the second radar device includes: Extract first ground point cloud data and second ground point cloud data from the first point cloud data and the second point cloud data; based on the first ground point cloud data and the second ground point cloud data , generate the roll angle and pitch angle in the rotation angle data, and the height value in the displacement data; based on the first other point in the first point cloud data except the first ground point cloud data Cloud data and second point cloud data other than the second ground point cloud data in the second point cloud data, determine the yaw angle in the rotation angle data, and the length in the displacement data value and width value.
  • the roll angle, pitch angle and height values are related to the ground point cloud data, and the yaw angle, length value and width value are related to other point cloud data except the ground point cloud data, it can be collected from different radar devices first.
  • the ground point cloud data can be extracted from the point cloud data of the radar, and the roll angle and pitch angle in the rotation angle data, and the height value in the displacement data can be generated based on the ground point cloud data corresponding to different radar devices.
  • Other point cloud data corresponding to different radar devices respectively determine the yaw angle in the rotation angle data, and the length and width values in the displacement data.
  • extracting the first ground point cloud data and the second ground point cloud data from the first point cloud data and the second point cloud data respectively includes: for the first radar device and each radar device in the second radar device, based on the direction angle from the position of the radar device to the position of each point in the point cloud data collected by the radar device, the point collected by the radar device.
  • the points in the cloud data are divided into a plurality of groups, wherein the difference between the direction angles corresponding to the points in the same group is smaller than a set threshold; for each group, based on each point in the group and its The height difference between adjacent points is used to determine the ground points in the group; based on the ground points in each of the groups, the ground point cloud data in the point cloud data collected by the radar device is determined.
  • the points in the point cloud data can be divided into multiple groups according to the direction angle that the position of the radar device points to the position of each point in the point cloud data collected by the radar device.
  • the height difference between the ground points is small, so for each group, the ground points in the group can be more accurately determined based on the height difference between the points in the group and its adjacent points, and then the ground points in the group can be more accurately determined.
  • Ground point cloud data in point cloud data is
  • determining the ground point in the group includes: based on the difference between each point in the group and the adjacent points. Collect the distance between the radar devices of the point, and sort the points in the group; for each point after sorting, determine the height difference between the point and its adjacent points; when the height difference is less than or equal to a preset height difference threshold, and when the height of the point is less than the height threshold corresponding to the point, it is determined that the point is a ground point belonging to the ground.
  • the height value of the ground point is consistent with the installation height of the radar equipment. For each point after sorting, determine the height difference between the point and its adjacent points. When the height difference is less than or equal to the set height difference threshold, and the height of the point is less than the height threshold corresponding to the point, determine the point It belongs to the ground, and more accurately determines the ground points in the group.
  • the height threshold corresponding to each point is determined according to the following steps: based on the coordinate data of the point, determine the plane distance on the horizontal plane between the point and the radar device that collects the point; The installation height, the plane distance, and the set plane angle of the radar device at the point are collected, and the height threshold corresponding to the point is determined.
  • the ground may have a slope (ie, the ground has a corresponding plane angle)
  • the height distances between the ground at different positions and the radar device are different, that is, points at different positions may correspond to different height thresholds. Therefore, for each point, the height threshold corresponding to the point can be determined, and then it can be judged more accurately whether each point belongs to the ground point.
  • the roll angle and pitch angle in the rotation angle data and the height value in the displacement data are generated. , including: generating first ground fitting parameters corresponding to the first radar device and second ground fitting parameters corresponding to the second radar device based on the first ground point cloud data and the second ground point cloud data combining parameters; determining the roll angle and pitch angle in the rotation angle data and the height value in the displacement data based on the first ground fitting parameter and the second ground fitting parameter.
  • the roll angle and pitch angle in the rotation angle data and the height value in the displacement data are determined , comprising: determining, based on the first ground fitting parameter and the second ground fitting parameter, the normal of the first fitting plane corresponding to the first ground fitting parameter and the second ground fitting parameter the normal of the corresponding second fitting plane; based on the normal of the first fitting plane corresponding to the first ground fitting parameter and the normal of the second fitting plane corresponding to the second ground fitting parameter, generating the roll angle and pitch angle in the rotation angle data; adjusting the first ground fitting parameter based on the roll angle and pitch angle in the rotation angle data to obtain the adjusted first ground fitting parameter ; wherein, the fitting plane corresponding to the adjusted first ground fitting parameter is parallel to the fitting plane corresponding to the second ground fitting parameter; based on the second ground fitting parameter and the adjusted The first ground fitting parameter determines the height value in the displacement data.
  • the different normals corresponding to the ground should be parallel, so the roll angle and Elevation angle; further, since the ground is a fixed plane, that is, the fitting plane corresponding to the ground fitting parameters of different radar equipment is a plane at the same height, so using the roll angle and the pitch angle, the first among different radar equipment After the ground fitting parameters corresponding to the radar equipment are adjusted, the height value in the displacement data may be determined based on the unadjusted ground fitting parameters and the adjusted ground fitting parameters corresponding to different radar equipment respectively.
  • the method is based on the first point cloud data other than the first ground point cloud data in the first point cloud data and the second point cloud data except the first point cloud data. Determining the second point cloud data other than the ground point cloud data, the yaw angle in the rotation angle data, and the length and width values in the displacement data, including: based on the generated rotation angle data The roll angle and pitch angle in , and the height value in the displacement data, adjust the first point cloud data to generate adjusted first point cloud data; wherein, the adjusted first point cloud data The data is located in the same plane as the unadjusted second point cloud data; based on the second other point cloud data in the second point cloud data and all the adjusted first point cloud data The first other point cloud data is used to determine the yaw angle in the rotation angle data, and the length and width values in the displacement data.
  • the target vehicle is parked on a straight road, and the target scene includes a plurality of three-dimensional marking objects placed around the target vehicle; the first point cloud data other than the first ground point cloud data and the second point cloud data other than the second ground point cloud data in the second point cloud data, to determine the rotation angle
  • the yaw angle in the data, and the length value and width value in the displacement data include: based on the length coordinate value of each point indicated by the first other point cloud data and the second other point cloud data and the width coordinate value, respectively cluster the first other point cloud data and the second other point cloud data, and obtain the corresponding multiplicity of the first other point cloud data and the second other point cloud data respectively.
  • each of the point cloud sets corresponds to one of the three-dimensional identification objects; based on the coordinate data of each point included in each of the point cloud sets, determine the average of each of the point cloud sets coordinate data; determine the yaw angle in the rotation angle data, and The length and width values in the displacement data.
  • each point cloud set corresponds to a three-dimensional marking object. Since the position of the three-dimensional marking object is fixed, the three-dimensional marking object can be used as a reference point to determine the yaw angle in the rotation angle data and the displacement data. length and width values.
  • the method further includes: using the generated coordinate system transformation matrix to convert the coordinate system transformation matrix.
  • the first point cloud data collected by the first radar device and the second point cloud data collected by the second radar device are converted into the same coordinate system; the pose data of the target object is determined based on the converted point cloud data.
  • the generated coordinate system transformation matrix can be used to transform the point cloud data collected by different radar devices into the same coordinate system. Since the generated coordinate system transformation matrix has a high accuracy, the point cloud data after coordinate transformation is It is more accurate, thereby improving the accuracy of the determined pose data of the target object.
  • the present disclosure provides an external parameter calibration device, comprising: an acquisition module for acquiring point cloud data obtained from the same target scene collected by at least two radar devices disposed on a target vehicle; a generation module for Rotation angle data and displacement data are generated based on the first point cloud data collected by the first radar device and the second point cloud data collected by the second radar device; wherein the rotation angle data is used to represent the first point cloud data and the angular deviation of the corresponding points in the second point cloud data in different directions, the displacement data is used to represent the first point cloud data and the corresponding points in the second point cloud data in different directions displacement deviation, wherein the first radar device and the second radar device are any two different radar devices among the at least two radar devices; a determining module is configured to, based on the rotation angle data and the displacement data to determine a coordinate system transformation matrix between the first radar device and the second radar device.
  • the present disclosure provides an electronic device, comprising: a processor, a memory, and a bus, the memory stores machine-readable instructions executable by the processor, and when the electronic device runs, the processor communicates with the The memory communicates through a bus, and when the machine-readable instruction is executed by the processor, the steps of the external parameter calibration method according to the first aspect or any one of the implementation manners are executed.
  • the present disclosure provides a computer-readable storage medium, where a computer program is stored on the computer-readable storage medium, and the computer program executes the external operation described in the first aspect or any one of the embodiments when the computer program is run by a processor. Steps of the parameter calibration method.
  • FIG. 1 shows a schematic flowchart of an external parameter calibration method provided by an embodiment of the present disclosure
  • FIG. 2 shows a schematic diagram of a target scene in an external parameter calibration method provided by an embodiment of the present disclosure
  • FIG. 3 shows a schematic flowchart of a specific method for generating rotation angle data and displacement data based on point cloud data collected by different radar devices in an external parameter calibration method provided by an embodiment of the present disclosure
  • FIG. 4 shows a schematic flowchart of a specific method for extracting ground point cloud data from point cloud data collected by different radar devices in an external parameter calibration method provided by an embodiment of the present disclosure
  • FIG. 5 shows a schematic diagram of a method for determining a height threshold corresponding to a point in an external parameter calibration method provided by an embodiment of the present disclosure
  • FIG. 6 shows a schematic structural diagram of an external parameter calibration device provided by an embodiment of the present disclosure
  • FIG. 7 shows a schematic structural diagram of an electronic device provided by an embodiment of the present disclosure.
  • the radar can be installed at the top position of the vehicle.
  • the radar wiring harness and the installation viewing angle there is a blind spot for the vehicle only after the radar is installed.
  • multiple radars can be installed at different positions of the vehicle, so that multiple radars work together to complete the detection of the surrounding environment of the vehicle.
  • an embodiment of the present disclosure provides an external parameter calibration method.
  • the execution subject of the external parameter calibration method provided by the embodiments of the present disclosure is generally a computer device with a certain computing capability. , mobile devices, user terminals, terminals, cellular phones, cordless phones, personal digital assistants (Personal Digital Assistant, PDA), handheld devices, computing devices, in-vehicle devices, wearable devices, etc.
  • the external parameter calibration method may be implemented by the processor calling computer-readable instructions stored in the memory.
  • FIG. 1 is a schematic flowchart of an external parameter calibration method provided by an embodiment of the present disclosure, the method includes the following steps S101-S103, wherein:
  • the point cloud data obtained from the same target scene collected by at least two radar devices installed on the target vehicle are acquired, the rotation angle data and the displacement data are generated based on the point cloud data collected by different radar devices, and then the rotation angle data is used.
  • the coordinate system transformation matrix between different radar devices is determined by the displacement data and the automatic external parameter calibration between different radar devices is realized. In this way, compared with the manual calibration process, the above method effectively improves the efficiency and accuracy of external parameter calibration.
  • Steps S101 to S103 will be specifically described below.
  • the target vehicle may be parked on a straight road, and the target scene may include a plurality of three-dimensional identification objects placed around the target vehicle.
  • multiple placement areas may be set around the target vehicle, and at least one three-dimensional marker object is placed in each placement area to constitute a target scene where the target vehicle is located.
  • the three-dimensional identification object may be any three-dimensional object, for example, the three-dimensional identification object may be a cone barrel or the like. Among them, an area where the detection ranges of different radar devices overlap can be determined, and multiple placement areas can be determined within the determined overlapping area, so that different radar devices can detect the point cloud data of each placed cone bucket.
  • the figure shows a target vehicle 21 and eight conical barrels 22 placed around the target vehicle 21 .
  • the position range of the determined placement area is the center of the target vehicle as the origin, and is 2 meters away from the origin in the width direction of the vehicle (hereinafter also referred to as the horizontal direction or the x-axis direction).
  • the horizontal direction or the x-axis direction To 4 meters, in a range of 7 meters to 10 meters from the origin in the longitudinal direction of the vehicle (hereinafter may also be referred to as the vertical direction or the y-axis direction).
  • the positions and numbers of the plurality of radar devices set on the target vehicle can be set as required; and the models or types of the set multiple radar devices can also be set as required.
  • a main radar device can be set at the top center of the target vehicle, and the main radar device can be a 64-line lidar; an auxiliary radar device can be set on the side of the target vehicle (for example, a side of the target vehicle can be set Auxiliary radar equipment can also be set up on the two sides of the target vehicle), and the auxiliary radar equipment can be a 16-line LiDAR.
  • the point cloud data corresponding to the target scene collected by the main radar device installed on the target vehicle can be obtained, and the point cloud data corresponding to the target scene collected by the auxiliary radar device installed on the target vehicle can be obtained.
  • cloud data can be obtained separately.
  • the point cloud data may include three-dimensional coordinate data of multiple points, and may also include reflection intensity information, color information, and the like of each point.
  • the rotation angle data is used to represent the angular deviation of the point cloud data collected by different radar devices in different directions
  • the displacement data is used to represent the displacement deviation of the point cloud data collected by different radar devices in different directions.
  • the rotation angle data can be used to represent the angular deviation in different directions between the point cloud data collected by the main radar device and the point cloud data collected by the auxiliary radar device.
  • the rotation angle data may be represented by Euler angle data, and the rotation angle data may include pitch angle pitch, yaw angle yaw, and roll angle roll.
  • the displacement data can be used to characterize the displacement deviation of the point cloud data collected by the main radar device and the point cloud data collected by the auxiliary radar device in different directions. This displacement data includes the deviation in the length direction, the deviation in the width direction, and the deviation in the height direction.
  • generating rotation angle data and displacement data based on the first point cloud data collected by the first radar device and the second point cloud data collected by the second radar device includes the following steps: S301 to S302.
  • the roll angle, pitch angle and height values are related to the ground point cloud data, and the yaw angle, length value and width value are related to other point cloud data except the ground point cloud data, it can be collected from different radar equipment first.
  • the ground point cloud data is extracted from the point cloud data, and based on the ground point cloud data corresponding to different radar devices, the roll angle and pitch angle in the rotation angle data, and the height value in the displacement data can be generated; For other corresponding point cloud data, determine the yaw angle in the rotation angle data, and the length and width values in the displacement data.
  • the first ground point cloud data can be extracted from the first point cloud data collected by the first radar device such as the main radar device, and the second ground point cloud data can be extracted from the second radar device such as the auxiliary radar device.
  • the second ground point cloud data is extracted from the second point cloud data collected by the radar device.
  • the first point cloud data can be divided into the first ground point cloud data and the first other point cloud data except the first ground point cloud data
  • the second point cloud data can be divided into the second ground point cloud data , and second other point cloud data except the second ground point cloud data.
  • the roll angle roll and pitch angle pitch in the rotation angle data, and the height value in the displacement data may be generated.
  • the ground point cloud data is extracted from the point cloud data collected by the radar equipment, including:
  • S401 Divide the points in the point cloud data collected by the radar device into multiple groups based on the direction angle from the position of the radar equipment to the position of each point in the point cloud data collected by the radar equipment, wherein multiple groups within the same group are The difference between the direction angles corresponding to the points is less than the set threshold;
  • the points in the point cloud data can be divided into a plurality of groups according to the direction angle that the position of the radar device points to the position of each point in the point cloud data collected by the radar device.
  • the height difference between adjacent ground points within a group is small. Therefore, for each group, the ground points in the group can be more accurately determined based on the height difference between each point in the group and its adjacent points, and then the ground point cloud in the point cloud data can be more accurately determined. data.
  • a direction angle from the position of the radar device to the position of each point in the point cloud data collected by the radar device is determined, and the direction angle is the direction angle from the position of the radar device to the position of each point in the point cloud data collected by the radar device.
  • the points in the point cloud data collected by the radar device are divided into multiple groups, wherein the difference between the directional angles corresponding to multiple points in the same group is smaller than the set threshold.
  • the number of groups can be determined according to needs.
  • the number of groups can be 2400, so as to divide 360 degrees into 2400 intervals, the corresponding angle ranges are (0° ⁇ 0.15°], (0.15° ⁇ 0.30°], (0.30° ⁇ 0.45°], ..., (359.70° ⁇ 359.85°], (359.85° ⁇ 360°], it can be seen that the difference between the direction angles corresponding to multiple points in each group is less than 0.15° (the set threshold).
  • the group identification corresponding to the point can be determined according to the direction angle.
  • the Points are divided into groups.
  • determining the ground point in the group based on the height difference between each point in the group and its adjacent points including: based on each point in the group The distance between the point and the radar equipment that collected the point, sort the points in the group; for each point after sorting, determine the height difference between the point and its adjacent points; when the height difference is less than or equal to the preset When the height difference threshold of , and the height of the point is less than the height threshold corresponding to the point, it is determined that the point belongs to the ground.
  • the height value of the ground point is consistent with the installation height of the radar equipment. For each point after sorting, determine the height difference between the point and its adjacent points. When the height difference is less than or equal to the height difference threshold, and the height of the point is less than the height threshold corresponding to the point, determine that the point belongs to The ground points on the ground, and then more accurately determine the ground points within the group.
  • the points in the group are sorted according to the distance between each point in the group and the radar device that collected the point.
  • the distance between each point and the radar device that collects the point can be determined based on the abscissa value and the ordinate value in the three-dimensional coordinate data of each point, and the determined distance corresponding to each point can be used to determine the distance between each point.
  • the points within the group are sorted.
  • the sorted points include point p1, point p2, point p3, and point p4, then: for point p1, use the height h 1 in the three-dimensional coordinate data of point p1 and the height h 2 in the three-dimensional coordinate data of point p2 , calculate the height difference h 2 -h 1 between point p1 and point p2, if the height difference is less than or equal to the height difference threshold, and the height h 1 of point p1 is less than the height threshold corresponding to point p1, it is determined that point p1 belongs to the ground ; If the height difference is greater than the height difference threshold, or the height h1 of the point p1 is greater than or equal to the height threshold corresponding to the point p1, then confirm that the point p1 does not belong to the ground, that is, the point p1 is not a ground point.
  • the height difference between point p4 and point p3 can be calculated. If the height difference is less than or equal to the height difference threshold, and the height h 4 of point p4 is less than the height threshold corresponding to point p4 , it is determined that the point p4 belongs to the ground.
  • the height threshold corresponding to each point can be determined according to the following steps: based on the coordinate data of the point, determine the plane distance on the horizontal plane between the point and the radar equipment that collects the point; based on the installation height of the radar equipment that collects the point , the plane distance, and the set plane angle to determine the height threshold corresponding to the point.
  • the plane distance on the horizontal plane between the point and the radar device that collects the point can be calculated based on the abscissa value and the ordinate value of the point.
  • the plane angle is used to represent the degree of inclination of the plane, and can be determined according to the actual scene, for example, the plane angle can be 5°.
  • the figure shows that the installation height of the radar device 51 is h, the height threshold corresponding to the point 52 collected by the radar device 51 is H, the plane angle set relative to the ground 53 is ⁇ , and l is the plane distance . That is, the tangent value of the plane angle ⁇ and the plane distance l can be used to determine the plane height h 0 corresponding to the point, and the plane height h 0 corresponding to the point at this point can be determined based on the installation height h of the radar device 51 and the determined plane height h 0 .
  • the ground may have a slope (ie, the ground has a certain plane angle)
  • the height distances between the ground at different positions and the radar device are different, that is, points at different positions may correspond to different height thresholds. Therefore, by determining the corresponding height threshold for each point, it can be more accurately judged whether each point is a ground point belonging to the ground.
  • the ground point cloud data in the point cloud data may be determined based on the ground points in each group, that is, the point cloud data of each ground point in each group, forming the ground point cloud data in the point cloud data.
  • the roll angle and pitch angle in the rotation angle data and the height value in the displacement data are generated, including: based on the first ground point.
  • the cloud data and the second ground point cloud data are used to generate a first ground fitting parameter corresponding to the first radar device and a second ground fitting parameter corresponding to the second radar device; based on the first ground fitting parameter and the second ground fitting parameters that determine the roll and pitch angles in the rotation angle data, and the altitude value in the displacement data.
  • the ground point cloud data collected by each radar device can be fitted to generate ground fitting parameters corresponding to each radar device.
  • various methods for fitting ground point cloud data there are various methods for fitting ground point cloud data, and this is only an exemplary description.
  • the random sample consensus algorithm (RANSAC) can be used to fit the ground point cloud data to generate ground fitting parameters corresponding to the ground point cloud data.
  • RANSAC random sample consensus algorithm
  • the roll angle and pitch angle in the rotation angle data and the height value in the displacement data are determined, including:
  • the first radar device and the second radar device may be any two different radar devices.
  • the first radar device may be a primary radar device or a secondary radar device.
  • the second radar device may be a secondary radar device or a primary radar device, as long as it is different from the first radar device.
  • the first ground fitting parameter corresponding to the first radar device and the second ground fitting parameter corresponding to the second radar device can be obtained, and the first method of the first fitting plane corresponding to the first ground fitting parameter can be determined line, and a second normal of the second fitting plane corresponding to the second ground fitting parameter.
  • the first fitting plane may also be referred to as a fitting plane corresponding to the first radar device
  • the second fitting plane may also be referred to as a fitting plane corresponding to the second radar device.
  • the first radar device corresponds to the first fitting plane
  • the second radar device corresponds to the second fitting plane.
  • the first fitting parameters include A 1 , B 1 , C 1 , and D 1
  • the first fitting equation corresponding to the first fitting plane is:
  • the first fitting equation corresponding to the second fitting plane is:
  • the normal vector of the first normal is (A 1 , B 1 , C 1 ), and the normal vector of the second normal is (A 2 , B 2 , C 2 ).
  • the normal vector of the first normal of the first radar device should be consistent with the normal vector of the second normal of the second radar device, it can be based on the normal vector of the first normal and the normal vector of the second normal, Determine the roll angle and pitch angle in the rotation angle data.
  • the first ground fitting parameter may be adjusted to obtain the adjusted first ground fitting parameter.
  • the first adjusted fitting plane corresponding to the adjusted first ground fitting parameter is parallel to the second fitting plane corresponding to the unadjusted second ground fitting parameter.
  • the distance between the first adjusted fitting plane and the second fitting plane corresponding to the adjusted first ground fitting parameter may be calculated , as the height value in the displacement data.
  • the different normals corresponding to the ground should be parallel lines. Therefore, the roll angle and pitch angle in the rotation angle data can be generated based on the normals of the fitting planes corresponding to different radar devices. In addition, since the ground is a fixed plane, the fitting planes corresponding to different radar devices should be the planes at the same height. Therefore, after adjusting the first ground fitting parameter corresponding to the first radar device by using the roll angle and the pitch angle, the height in the displacement data can be determined based on the unadjusted ground fitting parameter and the adjusted ground fitting parameter. value.
  • the method further includes: adjusting the first point cloud data based on the roll angle and pitch angle in the generated rotation angle data and the height value in the displacement data, and generating the adjusted first point cloud data. point cloud data; wherein the adjusted first point cloud data and the unadjusted second point cloud data are located in the same plane, and the adjusted first point cloud data includes the adjusted first ground point cloud data and the adjusted The first other point cloud data.
  • the yaw angle in the data, and the length and width values in the displacement data including: based on the second other point cloud data in the second point cloud data and the first other point in the adjusted first point cloud data Cloud data, determine the yaw angle in the rotation angle data, and the length and width values in the displacement data.
  • the generated roll angle, pitch angle, and height value, and the set initial yaw angle, initial length value, and initial width value can be used to generate an intermediate coordinate system transformation matrix; when using the generated intermediate coordinate system transformation matrix , adjust the first point cloud data to generate the adjusted first point cloud data.
  • the generated roll angle, pitch angle, and height value can be used to adjust the first point cloud data collected by the first radar device, such as the secondary radar device or the primary radar device, to generate the adjusted first point cloud data, so that the adjusted first point cloud data and the unadjusted second point cloud data are located in the same plane.
  • the second point cloud data except the second ground point cloud data in the second point cloud data and the first point cloud data in the adjusted first point cloud data may be based on For the first other point cloud data other than the ground point cloud data, determine the yaw angle in the rotation angle data, and the length and width values in the displacement data.
  • the generated intermediate coordinate system transformation matrix can be used here to adjust the first other point cloud data except the first ground point cloud data in the first point cloud data to generate the adjusted first other point cloud data ; Then use the second other point cloud data and the adjusted first other point cloud data to determine the yaw angle in the rotation angle data, and the length and width values in the displacement data.
  • each point cloud set corresponds to a three-dimensional identification object; based on the coordinate data of each point included in each point cloud set, determine the average coordinate data corresponding to each point cloud; based on the first other point cloud data and the second other point cloud data
  • the respective average coordinate data of a plurality of point cloud sets corresponding to the point cloud data respectively determines the yaw angle in the rotation angle data
  • the first other point cloud data can be clustered based on the length coordinate value and the width coordinate value of each point indicated by the first other point cloud data collected by the first radar device, such as the main radar device, to obtain At least one first point cloud set corresponding to the first radar device.
  • the number of the first point cloud sets obtained after clustering is consistent with the number of conical buckets (three-dimensional marking objects) set in the target scene. For example, if the number of conical buckets set in the target scene is 8, the clustering can obtain 8 first point cloud sets.
  • the second other point cloud data can also be clustered based on the length coordinate value and the width coordinate value of each point indicated by the second other point cloud data collected by the second radar device, such as the secondary radar device, to obtain the second point cloud data.
  • the 8 first point cloud sets obtained by clustering the main radar device as the first radar device can be divided into two types: left and right according to the left position and the right position, and can be based on the first The point cloud set, and the 4 second point cloud sets corresponding to the left auxiliary radar device, determine the yaw angle, length value and width value of the left auxiliary radar device; and based on the first point cloud set on the right, and the right
  • the 4 second point cloud sets corresponding to the side auxiliary radar equipment determine the yaw angle, length value and width value of the right auxiliary radar equipment.
  • the RANSAC algorithm can be used to cluster other point cloud data to obtain a point cloud set.
  • the coordinate data of each point included in the first point cloud set can be averaged to determine the first average coordinate data of the first point cloud set, and then each first point cloud set can be obtained and for each second point cloud set, average the coordinate data of each point included in the second point cloud set, determine the second average coordinate data of the second point cloud set, and then can Obtain second average coordinate data of each second point cloud set.
  • the yaw angle, the length value and the width value may be determined based on the respective first average coordinate data of the at least one first point cloud set, and the respective second average coordinate data of the at least one second point cloud set.
  • the number of secondary radar devices serving as the second radar device is multiple, for each secondary radar device, from at least one first point cloud set corresponding to the primary radar device serving as the first radar device, determine whether the secondary radar device is associated with the secondary radar device from at least one first point cloud set corresponding to the primary radar device serving as the first radar device.
  • the corresponding first point cloud set can be based on the first average coordinate data of the first point cloud set corresponding to the primary radar device and the respective second average coordinates of at least one second point cloud set corresponding to the secondary radar device. data to determine the yaw angle, length value and width value corresponding to the auxiliary radar device.
  • an iterative close point algorithm (Iterative Closest Point, ICP) may be used, based on the respective first average coordinate data of at least one first point cloud set and the respective second average coordinate data of at least one second point cloud set, The yaw angle in the rotation angle data, and the length and width values in the displacement data are determined.
  • each point cloud set corresponds to a three-dimensional marking object. Since the position of the three-dimensional marking object is fixed, the three-dimensional marking object can be used as a reference point to determine the yaw angle in the rotation angle data and the displacement data. length and width values.
  • a coordinate system transformation matrix between the first radar device and the second radar device may be determined based on the rotation angle data and the displacement data.
  • the coordinate system transformation matrix can be determined according to the following formula:
  • T 3D is a coordinate system transformation matrix
  • T is a three-dimensional vector formed by displacement data
  • R is a rotation matrix determined by rotation angle data
  • the method further includes: using the generated coordinate system transformation matrix to convert the first radar device collected by the first radar device.
  • the point cloud data and the second point cloud data collected by the second radar device are converted into the same coordinate system; based on the converted point cloud data, the pose data of the target object is determined.
  • the generated coordinate system transformation matrix can be used to convert the second point cloud data collected by the secondary radar device serving as the second radar device to the primary radar device serving as the first radar device.
  • the generated coordinate system conversion matrix can also be used to convert the first point cloud data collected by the primary radar device serving as the first radar device to the location where the second point cloud data collected by the secondary radar device serving as the second radar device is located.
  • the second point cloud data and the first point cloud data after coordinate transformation are fused, and the fused point cloud data is used to determine the pose data of the target object detected by the target vehicle during the moving process.
  • the target vehicle can be controlled based on the pose data of the target object detected during the movement of the target vehicle, for example, the acceleration, deceleration, steering, braking, etc. of the target vehicle can be controlled, or voice prompt information can be played to prompt The driver controls the acceleration, deceleration, steering, braking, etc. of the target vehicle.
  • the generated coordinate system transformation matrix can be used to transform the point cloud data collected by different radar devices into the same coordinate system. Since the generated coordinate system transformation matrix has a high accuracy, the point cloud data after coordinate transformation is It is more accurate, thereby improving the accuracy of the determined pose data of the target object.
  • the writing order of each step does not mean a strict execution order but constitutes any limitation on the implementation process, and the specific execution order of each step should be based on its function and possible Internal logic is determined.
  • an embodiment of the present disclosure also provides an external parameter calibration device.
  • a schematic diagram of the architecture of the external parameter calibration device provided by the embodiment of the present disclosure includes an acquisition module 601 , a generation module 602 , a determination Module 603.
  • An acquisition module 601 configured to acquire point cloud data obtained by at least two radar devices arranged on the target vehicle respectively collecting the same target scene;
  • a generating module 602 is configured to generate rotation angle data and displacement data based on the first point cloud data collected by the first radar device and the second point cloud data collected by the second radar device; wherein the rotation angle data is used to represent all The angle deviation of the corresponding points in the first point cloud data and the second point cloud data in different directions, and the displacement data is used to represent the corresponding points in the first point cloud data and the second point cloud data. displacement deviation of the points in different directions, wherein the first radar device and the second radar device are any two different radar devices in the at least two radar devices;
  • a determination module 603, configured to determine a coordinate system transformation matrix between the first radar device and the second radar device based on the rotation angle data and the displacement data.
  • the generating module 602 generates rotation angle data and displacement based on the first point cloud data collected by the first radar device and the second point cloud data collected by the second radar device.
  • the data is used to: extract the first ground point cloud data and the second surface point cloud data from the first point cloud data and the second point cloud data respectively; and based on the first ground point cloud data and the second ground point cloud data, generate the roll angle and pitch angle in the rotation angle data, and the height value in the displacement data; based on the first ground point in the first point cloud data the first other point cloud data except the cloud data and the second other point cloud data in the second point cloud data except the second ground point cloud data, determine the yaw angle in the rotation angle data , and the length and width values in the displacement data.
  • the generating module 602 extracts the first ground point cloud data and the second ground point cloud data from the first point cloud data and the second point cloud data respectively, use In: for each radar device in the first radar device and the second radar device, based on the direction angle from the position of the radar device to the position of each point in the point cloud data collected by the radar device , the points in the point cloud data collected by the radar device are divided into multiple groups, wherein the difference between the direction angles corresponding to multiple points in the same group is smaller than the set threshold; for each group, Based on the height difference between each point in the group and its adjacent points, determine the ground points in the group; based on the ground points in each of the groups, determine the ground points in the point cloud data collected by the radar device cloud data.
  • the generating module 602 when determining the ground points in the group based on the height difference between each point in the group and its adjacent points for each group, is used to: Based on the distance between each point in the group and the radar device that collected the point, the points in the group are sorted; for each sorted point, the distance between the point and its adjacent points is determined Height difference; when the height difference is less than or equal to a preset height difference threshold, and the height of the point is less than the height threshold corresponding to the point, determine the point as a ground point belonging to the ground.
  • the generating module 602 is configured to determine the height threshold corresponding to each point according to the following steps: based on the coordinate data of the point, determine the point and the radar device that collected the point. The plane distance between them on the horizontal plane; the height threshold corresponding to the point is determined based on the installation height of the radar device that collected the point, the plane distance, and the set plane angle.
  • the generating module 602 generates, based on the first ground point cloud data and the second ground point cloud data, the roll angle and pitch angle in the rotation angle data, and all the When the height value in the displacement data is used, it is used to: generate the first ground fitting parameter and the first ground fitting parameter corresponding to the first radar device based on the first ground point cloud data and the second ground point cloud data. Second ground fitting parameters corresponding to two radar devices; based on the first ground fitting parameters and the second ground fitting parameters, determine the roll angle and pitch angle in the rotation angle data, and the displacement data The height value in .
  • the generation module 602 determines, based on the first ground fitting parameter and the second ground fitting parameter, the roll angle and pitch angle in the rotation angle data, and all When the height value in the displacement data is used, it is used to: determine the normal of the first fitting plane corresponding to the first ground fitting parameter based on the first ground fitting parameter and the second ground fitting parameter The normal line of the second fitting plane corresponding to the second ground fitting parameter; based on the normal line of the first fitting plane corresponding to the first ground fitting parameter and the second ground fitting parameter corresponding The normal line of the second fitting plane is used to generate the roll angle and pitch angle in the rotation angle data; based on the roll angle and pitch angle in the rotation angle data, the first ground fitting parameters are adjusted to obtain Adjusted first ground fitting parameters; wherein, the fitting plane corresponding to the adjusted first ground fitting parameters is parallel to the fitting plane corresponding to the second ground fitting parameters; based on the second ground fitting parameters The ground fitting parameter and the adjusted first ground fitting parameter determine the height value in the displacement data.
  • the apparatus further includes: an adjustment module 604, which is based on the first point cloud data other than the first ground point cloud data and the first point cloud data in the first point cloud data.
  • an adjustment module 604 is based on the first point cloud data other than the first ground point cloud data and the first point cloud data in the first point cloud data.
  • the second point cloud data other than the second ground point cloud data in the two point cloud data when determining the yaw angle in the rotation angle data and the length and width values in the displacement data, for: adjusting the first point cloud data based on the roll angle and pitch angle in the generated rotation angle data and the height value in the displacement data, and generating the adjusted first point cloud data;
  • the adjusted first point cloud data and the unadjusted second point cloud data are located in the same plane;
  • the generating module 602 is configured to: based on the second other point cloud data in the second point cloud data and the first other point cloud data in the adjusted first point cloud data, The yaw angle in the rotation angle data, and the length and width values in the displacement data are determined.
  • the target vehicle is parked on a straight road, and the target scene includes a plurality of three-dimensional marking objects placed around the target vehicle;
  • the generating module 602 is based on the first point cloud data other than the first ground point cloud data in the first point cloud data and the second point cloud data except the second ground point
  • the second other point cloud data other than the cloud data when determining the yaw angle in the rotation angle data, and the length and width values in the displacement data, are used for: based on the first other point cloud data and the length coordinate value and width coordinate value of each point indicated by the second other point cloud data, respectively cluster the first other point cloud data and the second other point cloud data to obtain the Multiple point cloud sets corresponding to the first other point cloud data and the second other point cloud data respectively; wherein, each of the point cloud sets corresponds to one of the three-dimensional identification objects; The coordinate data of each point included is determined, and the average coordinate data of each of the point cloud sets is determined; The average coordinate data, the yaw angle in the rotation angle data, and the length and width values in the displacement data are determined.
  • the apparatus further includes: a transformation module 605, configured to: use the generated A coordinate system conversion matrix, which converts the first point cloud data collected by the first radar device and the second point cloud data collected by the second radar device into the same coordinate system; and determines the target based on the converted point cloud data The pose data of the object.
  • a transformation module 605 configured to: use the generated A coordinate system conversion matrix, which converts the first point cloud data collected by the first radar device and the second point cloud data collected by the second radar device into the same coordinate system; and determines the target based on the converted point cloud data The pose data of the object.
  • the functions or templates included in the apparatus provided by the embodiments of the present disclosure may be used to execute the methods described in the above method embodiments.
  • the functions or templates included in the apparatus provided by the embodiments of the present disclosure may be used to execute the methods described in the above method embodiments.
  • an embodiment of the present disclosure also provides an electronic device.
  • a schematic structural diagram of an electronic device provided by an embodiment of the present disclosure includes a processor 701 , a memory 702 , and a bus 703 .
  • the memory 702 is used to store the execution instructions, including the memory 7021 and the external memory 7022; the memory 7021 here is also called the internal memory, which is used to temporarily store the operation data in the processor 701 and the data exchanged with the external memory 7022 such as the hard disk,
  • the processor 701 exchanges data with the external memory 7022 through the memory 7021.
  • the processor 701 communicates with the memory 702 through the bus 703, so that the processor 701 is executing the following instructions: obtain the data set on the target vehicle.
  • At least two radar devices respectively collect point cloud data obtained from the same target scene; based on the first point cloud data collected by the first radar device and the second point cloud data collected by the second radar device, the rotation angle data and the displacement data are generated; wherein , the rotation angle data is used to represent the angular deviation of the corresponding points in the first point cloud data and the second point cloud data in different directions, and the displacement data is used to represent the first point cloud data The displacement deviation from the corresponding point in the second point cloud data in different directions, wherein the first radar device and the second radar device are any two different radar devices in the at least two radar devices ; determining a coordinate system transformation matrix between the first radar device and the second radar device based on the rotation angle data and the displacement data.
  • an embodiment of the present disclosure further provides a computer-readable storage medium, where a computer program is stored on the computer-readable storage medium, and when the computer program is run by a processor, the external parameter calibration method described in the above method embodiment is executed. step.
  • the storage medium may be a volatile or non-volatile computer-readable storage medium.
  • the embodiments of the present disclosure further provide a computer program product, the computer program product carries program codes, and the instructions included in the program codes can be used to execute the steps of the external parameter calibration method described in the above method embodiments.
  • the computer program product carries program codes
  • the instructions included in the program codes can be used to execute the steps of the external parameter calibration method described in the above method embodiments.
  • please refer to the above The method embodiments are not repeated here.
  • the above-mentioned computer program product can be specifically implemented by means of hardware, software or a combination thereof.
  • the computer program product is embodied as a computer storage medium, and in another optional embodiment, the computer program product is embodied as a software product, such as a software development kit (Software Development Kit, SDK), etc. Wait.
  • the units described as separate components may or may not be physically separated, and components displayed as units may or may not be physical units, that is, may be located in one place, or may be distributed to multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution in this embodiment.
  • each functional unit in each embodiment of the present disclosure may be integrated into one processing unit, or each unit may exist physically alone, or two or more units may be integrated into one unit.
  • the functions, if implemented in the form of software functional units and sold or used as stand-alone products, may be stored in a processor-executable non-volatile computer-readable storage medium.
  • the computer software products are stored in a storage medium, including Several instructions are used to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present disclosure.
  • the aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (Read-Only Memory, ROM), random access memory (Random Access Memory, RAM), magnetic disk or optical disk and other media that can store program codes .

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Remote Sensing (AREA)
  • Radar Systems Or Details Thereof (AREA)

Abstract

一种外参标定方法、装置、电子设备及存储介质,该外参标定方法包括:获取设置在目标车辆上的至少两个雷达设备分别采集同一目标场景得到的点云数据(S101);基于第一雷达设备采集的第一点云数据和第二雷达设备采集的第二点云数据,生成旋转角数据和位移数据,其中,旋转角数据用于表征第一点云数据和第二点云数据中对应的点在不同方向上的角度偏差,位移数据用于表征第一点云数据和第二点云数据中对应的点在不同方向上位移偏差(S102);基于旋转角数据和位移数据,确定第一雷达设备和第二雷达设备之间的坐标系转换矩阵(S103)。

Description

外参标定方法、装置、电子设备及存储介质
相关申请的交叉引用
本公开要求于2021年2月26日提交的、申请号为202110219865.7的中国专利申请的优先权,该申请以引用的方式并入本文中。
技术领域
本公开涉及雷达设备技术领域,具体而言,涉及一种外参标定方法、装置、电子设备及存储介质。
背景技术
雷达是一种利用电磁波探测目标的电子设备。雷达发射电磁波对目标进行照射并接收其回波,由此获得目标至电磁波发射点的距离、距离变化率、方位、高度等信息。由于雷达的探测范围较广、探测精度较高,使得雷达成为无人车技术、辅助驾驶技术中常用的传感器。
一般的,可以在车辆的顶部位置安装雷达。但是受雷达线束和安装视角的限制,使得对于仅安装该雷达后的车辆而言存在盲区。为了消除盲区,可以在车辆的不同位置处安装多个雷达,使得多个雷达协同工作,完成对车辆周围环境的探测。为了实现多个雷达的协同工作,需要将多个雷达检测得到的点云数据统一在同一坐标系下,也即实现雷达之间的外参标定。
发明内容
有鉴于此,本公开至少提供一种外参标定方法、装置、电子设备及存储介质。
第一方面,本公开提供了一种外参标定方法,包括:获取设置在目标车辆上的至少两个雷达设备分别采集同一目标场景得到的点云数据;基于第一雷达设备采集的第一点云数据和第二雷达设备采集的第二点云数据,生成旋转角数据和位移数据;其中,所述旋转角数据用于表征所述第一点云数据和所述第二点云数据中对应的点在不同方向上的角度偏差,所述位移数据用于表征所述第一点云数据和所述第二点云数据中对应的点在不同方向上位移偏差,其中,所述第一雷达设备和所述第二雷达设备为所述至少两个雷达设备中任意两个不同的雷达设备;基于所述旋转角数据和所述位移数据,确定所述第一雷达设备和所述第二雷达设备之间的坐标系转换矩阵。
采用上述方法,通过获取设置在目标车辆上的至少两个雷达设备分别采集同一目标场景得到的点云数据,基于不同雷达设备采集的点云数据,生成旋转角数据和位移数据,再利用旋转角数据和位移数据,确定不同雷达设备之间的坐标系转换矩阵,实现了不同雷达设备之间的自动化外参标定,相比人工标定过程,提高了外参标定的效率和准确度。
一种可能的实施方式中,所述基于所述第一雷达设备采集的第一点云数据和所述第二雷达设备采集的第二点云数据,生成旋转角数据和位移数据,包括:分别从所述第 一点云数据和所述第二点云数据中提取第一地面点云数据和第二地面点云数据;基于所述第一地面点云数据和所述第二地面点云数据,生成所述旋转角数据中的翻滚角和俯仰角,以及所述位移数据中的高度值;基于所述第一点云数据中除所述第一地面点云数据之外的第一其他点云数据和所述第二点云数据中除所述第二地面点云数据之外的第二其他点云数据,确定所述旋转角数据中的偏航角,以及所述位移数据中的长度值和宽度值。
由于翻滚角、俯仰角和高度值与地面点云数据有关,而偏航角、长度值和宽度值与除地面点云数据之外的其他点云数据有关,故这里可以先从不同雷达设备采集的点云数据中提取地面点云数据,并可以基于不同雷达设备分别对应的提取到的地面点云数据,生成旋转角数据中的翻滚角和俯仰角,以及位移数据中的高度值;再利用不同雷达设备分别对应的其他点云数据,确定旋转角数据中的偏航角,以及位移数据中的长度值和宽度值。
一种可能的实施方式中,分别从所述第一点云数据和所述第二点云数据中提取第一地面点云数据和第二地面点云数据,包括:针对所述第一雷达设备和所述第二雷达设备中的每个雷达设备,基于从所述雷达设备的位置指向所述雷达设备采集的点云数据中每个点的位置的方向角,将所述雷达设备采集的点云数据中的点划分为多个小组,其中同一小组内的多个点分别对应的所述方向角之间的差值小于设定阈值;针对每个小组,基于该小组内的每个点与其相邻点之间的高度差,确定该小组内的地面点;基于各个所述小组内的地面点,确定所述雷达设备采集的点云数据中地面点云数据。
这里,可以按照雷达设备的位置指向该雷达设备采集的点云数据中每个点的位置的方向角,将该点云数据中的点划分为多个小组,一般的,一个小组内相邻的地面点之间的高度差较小,故针对每个小组,可以基于该小组内的点与其相邻点之间的高度差,较准确的确定该小组内的地面点,进而可以较准确的确定点云数据中地面点云数据。
一种可能的实施方式中,针对每个小组,基于该小组内的每个点与其相邻点之间的高度差,确定该小组内的地面点,包括:基于所述小组内每个点与采集该点的所述雷达设备之间的距离,将所述小组内的点进行排序;针对排序后的每个点,确定该点与其相邻点之间的高度差;在所述高度差小于或等于预先设置的高度差阈值、且该点的高度小于该点对应的高度阈值的情况下,确定所述点为属于地面的地面点。
考虑到相邻的地面点之间的高度差较小,以及在雷达设备的安装高度确定之后,地面点的高度值与雷达设备的安装高度一致,故这里在将小组内的点排序之后,针对排序后的每个点,确定该点与其相邻点之间的高度差,在该高度差小于或等于设置的高度差阈值,以及该点的高度小于该点对应的高度阈值时,确定该点属于地面,进而较准确的确定了该小组内的地面点。
一种可能的实施方式中,根据下述步骤确定每个点对应的高度阈值:基于所述点的坐标数据,确定所述点与采集该点的雷达设备之间在水平面上的平面距离;基于采集该点的所述雷达设备的安装高度、所述平面距离、和设置的平面角度,确定所述点对应的高度阈值。
由于地面可能存在坡度(即地面存在对应的平面角度),使得不同位置处的地面与雷达设备之间的高度距离不同,即不同位置处的点可能对应不同的高度阈值。因此,针对每个点,可以确定该点对应的高度阈值,进而可以较准确的判断每个点是否属于地面点。
一种可能的实施方式中,基于所述第一地面点云数据和所述第二地面点云数据,生成所述旋转角数据中的翻滚角和俯仰角、以及所述位移数据中的高度值,包括:基于 所述第一地面点云数据和所述第二地面点云数据,生成所述第一雷达设备对应的第一地面拟合参数和所述第二雷达设备对应的第二地面拟合参数;基于所述第一地面拟合参数和所述第二地面拟合参数,确定所述旋转角数据中的翻滚角和俯仰角、以及所述位移数据中的高度值。
一种可能的实施方式中,基于所述第一地面拟合参数和所述第二地面拟合参数,确定所述旋转角数据中的翻滚角和俯仰角、以及所述位移数据中的高度值,包括:基于所述第一地面拟合参数和所述第二地面拟合参数,确定所述第一地面拟合参数对应的第一拟合平面的法线和所述第二地面拟合参数对应的第二拟合平面的法线;基于所述第一地面拟合参数对应的第一拟合平面的法线和所述第二地面拟合参数对应的第二拟合平面的法线,生成所述旋转角数据中的翻滚角和俯仰角;基于所述旋转角数据中的翻滚角和俯仰角,对所述第一地面拟合参数进行调整,得到调整后的第一地面拟合参数;其中,所述调整后的第一地面拟合参数对应的拟合平面、与所述第二地面拟合参数对应的拟合平面平行;基于所述第二地面拟合参数和所述调整后的第一地面拟合参数,确定所述位移数据中的高度值。
这里,由于地面是固定的,地面对应的不同法线之间应该是平行的,故可以基于不同雷达设备的地面拟合参数对应的拟合平面的法线,生成旋转角数据中的翻滚角和俯仰角;进一步地,由于地面是固定的平面,即不同雷达设备的地面拟合参数对应的拟合平面为处于同一高度的平面,故在利用翻滚角和俯仰角,对不同雷达设备中第一雷达设备对应的地面拟合参数进行调整后,可以基于不同雷达设备分别对应的未调整的地面拟合参数和调整后的地面拟合参数,确定位移数据中的高度值。
一种可能的实施方式中,所述基于所述第一点云数据中除所述第一地面点云数据之外的第一其他点云数据和所述第二点云数据中除所述第二地面点云数据之外的第二其他点云数据,确定所述旋转角数据中的偏航角、和所述位移数据中的长度值和宽度值,包括:基于生成的所述旋转角数据中的翻滚角和俯仰角、以及所述位移数据中的高度值,对所述第一点云数据进行调整,生成调整后的第一点云数据;其中,所述调整后的第一点云数据与未调整的所述第二点云数据位于同一平面内;基于所述第二点云数据中的所述第二其他点云数据、和所述调整后的第一点云数据中的所述第一其他点云数据,确定所述旋转角数据中的偏航角、和所述位移数据中的长度值和宽度值。
一种可能的实施方式中,所述目标车辆停放在平直的道路上,所述目标场景包括放置在所述目标车辆的四周的多个立体标识物体;基于所述第一点云数据中除所述第一地面点云数据之外的第一其他点云数据和所述第二点云数据中除所述第二地面点云数据之外的第二其他点云数据,确定所述旋转角数据中的偏航角、和所述位移数据中的长度值和宽度值,包括:基于所述第一其他点云数据和所述第二其他点云数据所指示的每个点的长度坐标值和宽度坐标值,分别对所述第一其他点云数据和所述第二其他点云数据进行聚类,得到所述第一其他点云数据和所述第二其他点云数据分别对应的多个点云集合;其中,每个所述点云集合对应一个所述立体标识物体;基于每个所述点云集合中包括的每个点的坐标数据,确定每个所述点云集合的平均坐标数据;基于所述第一其他点云数据和所述第二其他点云数据分别对应的多个点云集合各自的所述平均坐标数据,确定所述旋转角数据中的偏航角、和所述位移数据中的长度值和宽度值。
上述实施方式中,每个点云集合对应一个立体标识物体,由于立体标识物体的位置是固定的,故可以将立体标识物体作为基准点,确定旋转角数据中的偏航角、和位移数据中的长度值和宽度值。
一种可能的实施方式中,在确定所述第一雷达设备和所述第二雷达设备之间的坐标 系转换矩阵之后,所述方法还包括:利用生成的所述坐标系转换矩阵,将所述第一雷达设备采集的第一点云数据和所述第二雷达设备采集的第二点云数据转换至同一坐标系下;基于转换后的点云数据,确定目标对象的位姿数据。
采用上述方法,可以利用生成的坐标系转换矩阵,将不同雷达设备采集的点云数据转换至同一坐标系下,由于生成的坐标系转换矩阵的准确率较高,使得坐标转换后的点云数据较为准确,进而提高了确定的目标对象的位姿数据的准确度。
以下装置、电子设备等的效果描述参见上述方法的说明,这里不再赘述。
第二方面,本公开提供了一种外参标定装置,包括:获取模块,用于获取设置在目标车辆上的至少两个雷达设备分别采集同一目标场景得到的点云数据;生成模块,用于基于第一雷达设备采集的第一点云数据和第二雷达设备采集的第二点云数据,生成旋转角数据和位移数据;其中,所述旋转角数据用于表征所述第一点云数据和所述第二点云数据中对应的点在不同方向上的角度偏差,所述位移数据用于表征所述第一点云数据和所述第二点云数据中对应的点在不同方向上位移偏差,其中,所述第一雷达设备和所述第二雷达设备为所述至少两个雷达设备中任意两个不同的雷达设备;确定模块,用于基于所述旋转角数据和所述位移数据,确定所述第一雷达设备和所述第二雷达设备之间的坐标系转换矩阵。
第三方面,本公开提供一种电子设备,包括:处理器、存储器和总线,所述存储器存储有所述处理器可执行的机器可读指令,当电子设备运行时,所述处理器与所述存储器之间通过总线通信,所述机器可读指令被所述处理器执行时执行如上述第一方面或任一实施方式所述的外参标定方法的步骤。
第四方面,本公开提供一种计算机可读存储介质,该计算机可读存储介质上存储有计算机程序,该计算机程序被处理器运行时执行如上述第一方面或任一实施方式所述的外参标定方法的步骤。
为使本公开的上述目的、特征和优点能更明显易懂,下文特举较佳实施例,并配合所附附图,作详细说明如下。
附图说明
为了更清楚地说明本公开实施例的技术方案,下面将对实施例中所需要使用的附图作简单地介绍。这些附图示出了符合本公开的实施例,并与说明书一起用于说明本公开的技术方案。应当理解,以下附图仅示出了本公开的某些实施例,因此不应被看作是对范围的限定,对于本领域普通技术人员来讲,在不付出创造性劳动的前提下,还可以根据这些附图获得其他相关的附图。
图1示出了本公开实施例所提供的一种外参标定方法的流程示意图;
图2示出了本公开实施例所提供的一种外参标定方法中,目标场景的示意图;
图3示出了本公开实施例所提供的一种外参标定方法中,基于不同雷达设备采集的点云数据,生成旋转角数据和位移数据的具体方法的流程示意图;
图4示出了本公开实施例所提供的一种外参标定方法中,分别从不同雷达设备采集的点云数据中提取地面点云数据的具体方法的流程示意图;
图5示出了本公开实施例所提供的一种外参标定方法中,确定点对应的高度阈值的方法的示意图;
图6示出了本公开实施例所提供的一种外参标定装置的架构示意图;
图7示出了本公开实施例所提供的一种电子设备的结构示意图。
具体实施方式
为使本公开实施例的目的、技术方案和优点更加清楚,下面将结合本公开实施例中的附图,对本公开实施例中的技术方案进行清楚、完整地描述,显然,所描述的实施例仅仅是本公开一部分实施例,而不是全部的实施例。通常在此处附图中描述和示出的本公开实施例的组件可以以各种不同的配置来布置和设计。因此,以下对在附图中提供的本公开的实施例的详细描述并非旨在限制要求保护的本公开的范围,而是仅仅表示本公开的选定实施例。基于本公开的实施例,本领域技术人员在没有做出创造性劳动的前提下所获得的所有其他实施例,都属于本公开保护的范围。
一般的,可以在车辆的顶部位置安装雷达。但是受雷达线束和安装视角的限制,使得对于仅安装该雷达后的车辆而言存在盲区。为了消除盲区,可以在车辆的不同位置处安装多个雷达,使得多个雷达协同工作,完成对车辆周围环境的探测。为了实现多个雷达的协同工作,需要将多个雷达检测得到的点云数据统一在同一坐标系下,也即实现雷达之间的外参标定。为了完成雷达之间的外参标定,本公开实施例提供了一种外参标定方法。
以上均是发明人在经过实践并仔细研究后得出的结果,因此,上述问题的发现过程以及下文中本公开针对上述问题所提出的解决方案,都应该是发明人在本公开过程中对本公开做出的贡献。
下面将结合本公开中附图,对本公开中的技术方案进行清楚、完整地描述,显然,所描述的实施例仅仅是本公开一部分实施例,而不是全部的实施例。通常在此处附图中描述和示出的本公开的组件可以以各种不同的配置来布置和设计。因此,以下对在附图中提供的本公开的实施例的详细描述并非旨在限制要求保护的本公开的范围,而是仅仅表示本公开的选定实施例。基于本公开的实施例,本领域技术人员在没有做出创造性劳动的前提下所获得的所有其他实施例,都属于本公开保护的范围。
应注意到:相似的标号和字母在下面的附图中表示类似项,因此,一旦某一项在一个附图中被定义,则在随后的附图中不需要对其进行进一步定义和解释。
首先对本公开实施例所公开的一种外参标定方法进行详细介绍。本公开实施例所提供的外参标定方法的执行主体一般为具有一定计算能力的计算机设备,该计算机设备例如包括终端设备或服务器或其它处理设备,终端设备可以为用户设备(User Equipment,UE)、移动设备、用户终端、终端、蜂窝电话、无绳电话、个人数字助理(Personal Digital Assistant,PDA)、手持设备、计算设备、车载设备、可穿戴设备等。在一些可能的实现方式中,该外参标定方法可以通过处理器调用存储器中存储的计算机可读指令的方式来实现。
参见图1所示,为本公开实施例所提供的外参标定方法的流程示意图,该方法包括以下步骤S101-S103,其中:
S101,获取设置在目标车辆上的至少两个雷达设备分别采集同一目标场景得到的点云数据;
S102,基于第一雷达设备和第二雷达设备采集的点云数据,生成旋转角数据和位移数据;其中,旋转角数据用于表征第一雷达设备和第二雷达设备采集的点云数据中对应点在不同方向上的角度偏差,位移数据用于表征第一雷达设备和第二雷达设备采集的点云数据中对应点在不同方向上的位移偏差,其中,第一雷达设备和第二雷达设备为所述至少两个雷达设备中不同的任意两个雷达设备;
S103,基于旋转角数据和位移数据,确定第一雷达设备和第二雷达设备之间的坐标系转换矩阵。
采用上述方法,通过获取设置在目标车辆上的至少两个雷达设备分别采集同一目标场景得到的点云数据,基于不同雷达设备采集的点云数据生成旋转角数据和位移数据,再利用旋转角数据和位移数据确定不同雷达设备之间的坐标系转换矩阵,实现了不同雷达设备之间的自动化外参标定。这样,相比人工标定过程,上述方法有效提高了外参标定的效率和准确度。
下述对步骤S101至S103进行具体说明。
针对S101,目标车辆可以停放在平直的道路上,以及目标场景可以包括放置在目标车辆的四周的多个立体标识物体。示例性的,可以在目标车辆的周围设置多个放置区域,在每个放置区域内放置至少一个立体标识物体,构成了目标车辆所处的目标场景。该立体标识物体可以为任一立体的物体,比如,该立体标识物体可以为锥形桶等。其中,可以确定不同雷达设备之间检测范围重叠的区域,在确定的重合区域内确定多个放置区域,使得不同雷达设备均可以检测到放置的每个锥形桶的点云数据。
参见图2所示的一种外参标定方法中目标场景的示意图,该图示出了目标车辆21、以及放置在目标车辆21周围的八个锥形桶22。其中,锥形桶放置后,需要避免锥形桶之间彼此遮挡。示例性的,如图2所示,确定的放置区域的位置范围为以目标车辆的中心为原点,在车辆的宽度方向(以下也可称为水平方向或x轴方向)上距离该原点2米至4米、在车辆的长度方向(以下也可称为垂直方向或y轴方向)上距离该原点7米至10米的范围内。
目标车辆上设置的多个雷达设备的位置和数量可以根据需要进行设置;以及设置的多个雷达设备的型号、或类型也可以根据需要进行设置。在一应用场景中,可以在目标车辆的顶部中央设置一个主雷达设备,主雷达设备可以为64线激光雷达;可以在目标车辆的侧面设置一个辅雷达设备(比如,可以在目标车辆的一个侧面设置辅雷达设备,也可以在目标车辆的两个侧面分别设置辅雷达设备),辅雷达设备可以为16线激光雷达。
下述以目标车辆上设置主雷达设备和辅雷达设备的应用场景进行具体说明。具体实施时,在构建了目标场景之后,可以获取设置在目标车辆上的主雷达设备采集的目标场景对应的点云数据,和获取设置在目标车辆上的辅雷达设备采集的目标场景对应的点云数据。在辅雷达设备为多个时,可以分别获取每个辅雷达设备采集的点云数据。其中,该点云数据中可以包括多个点的三维坐标数据,还可以包括每个点的反射强度信息、颜色信息等。
针对S102,承接上述S101中应用场景的示例继续说明,在获取了主雷达设备采集的点云数据和辅雷达设备采集的点云数据之后,可以基于所获取的点云数据生成辅雷达设备对应的旋转角数据和位移数据。在辅雷达设备的数量为多个时,可以针对每个辅雷达设备,基于主雷达设备采集的点云数据、和该辅雷达设备采集的点云数据,确定该辅雷达设备对应的旋转角数据和位移数据。
其中,旋转角数据用于表征不同雷达设备采集的点云数据在不同方向上的角度偏差,位移数据用于表征不同雷达设备采集的点云数据在不同方向上位移偏差。比如,在上述提到的应用场景中,旋转角数据可以用于表征主雷达设备采集的点云数据和该辅雷达设备采集的点云数据在不同方向上的角度偏差。该旋转角数据可以用欧拉角数据进行表征,该旋转角数据中可以包括俯仰角pitch、偏航角yaw、翻滚角roll。位移数据可以用于表征主雷达设备采集的点云数据和该辅雷达设备采集的点云数据在不同方向 上的位移偏差。该位移数据中包括在长度方向上的偏差、宽度方向上的偏差、和高度方向上的偏差。
一种可选实施方式中,参见图3所示,基于第一雷达设备采集的第一点云数据和第二雷达设备采集的第二点云数据,生成旋转角数据和位移数据,包括以下步骤S301至S302。
S301,分别从第一点云数据和第二点云数据中提取第一地面点云数据和第二地面点云数据;并基于第一地面点云数据和第二地面点云数据,生成旋转角数据中的翻滚角和俯仰角,以及位移数据中的高度值。
S302,基于第一点云数据中除第一地面点云数据之外的第一其他点云数据和第二点云数据中除第二地面点云数据之外的第二其他点云数据,确定旋转角数据中的偏航角,以及位移数据中的长度值和宽度值。
由于翻滚角、俯仰角和高度值与地面点云数据有关,而偏航角、长度值和宽度值与除地面点云数据之外的其他点云数据有关,故可以先从不同雷达设备采集的点云数据中提取地面点云数据,并可以基于不同雷达设备分别对应的地面点云数据,生成旋转角数据中的翻滚角和俯仰角、以及位移数据中的高度值;再利用不同雷达设备分别对应的其他点云数据,确定旋转角数据中的偏航角、以及位移数据中的长度值和宽度值。
针对S301,以上述描述的应用场景为例进行说明,可以从例如主雷达设备的第一雷达设备采集的第一点云数据中提取第一地面点云数据,以及从例如辅雷达设备的第二雷达设备采集的第二点云数据中提取第二地面点云数据。这样,可将第一点云数据划分为第一地面点云数据、和除第一地面点云数据之外的第一其他点云数据,将第二点云数据划分为第二地面点云数据、和除第二地面点云数据之外的第二其他点云数据。进而可以基于第一地面点云数据和第二地面点云数据,生成旋转角数据中的翻滚角roll和俯仰角pitch,以及位移数据中的高度值。
下述对从雷达设备采集的点云数据中提取地面点云数据的过程进行具体说明。
一种可选实施方式中,参见图4所示,从雷达设备采集的点云数据中提取地面点云数据,包括:
S401,基于从雷达设备的位置指向雷达设备采集的点云数据中每个点的位置的方向角,将雷达设备采集的点云数据中的点划分为多个小组,其中同一小组内的多个点分别对应的方向角之间的差值小于设定阈值;
S402,针对每个小组,基于该小组内的每个点与其相邻点之间的高度差,确定该小组内的地面点;
S403,基于各个小组内的地面点,确定雷达设备采集的点云数据中的地面点云数据。
可以按照雷达设备的位置指向该雷达设备采集的点云数据中每个点的位置的方向角,将该点云数据中的点划分为多个小组。一般的,一个小组内相邻的地面点之间的高度差较小。故针对每个小组,可以基于该小组内的每个点与其相邻点之间的高度差,较准确的确定该小组内的地面点,进而可以较准确的确定点云数据中的地面点云数据。
在S401中,确定从雷达设备的位置指向该雷达设备采集的点云数据中每个点的位置的方向角,该方向角为从该雷达设备的位置指向该雷达设备采集的点云数据中每个点的位置的线条与预设基准线之间的角度。基于每个点对应的方向角,将该雷达设备采集的点云数据中的点划分为多个小组,其中,同一小组内的多个点分别对应的方向角之间的差值小于设置阈值。
其中,小组的数量可以根据需要进行确定。比如,小组的数量可以为2400个,以将360度划分为2400个区间,对应的角度范围分别为(0°~0.15°]、(0.15°~0.30°]、(0.30°~0.45°]、……、(359.70°~359.85°]、(359.85°~360°],可知每个小组内的多个点对应的方向角之间的差值小于0.15°(设定阈值)。在确定了每个点对应的方向角之后,可以根据方向角确定该点对应的小组标识。例如,若点pA对应的方向角为0.20°,确定该点pA属于角度范围为(0.15°~0.30°]的第2组;若点pB对应的方向角为360°,确定该点pA属于角度范围为(359.85°~360°]的第2400组。进而,通过各个点对应的方向角,将点云数据中的点划分为多个小组。
在S402中,一种可选实施方式中,针对每个小组,基于该小组内的每个点与其相邻点之间的高度差,确定该小组内的地面点,包括:基于小组内每个点与采集该点的雷达设备之间的距离,将小组内的点进行排序;针对排序后的每个点,确定该点与其相邻点之间的高度差;在高度差小于或等于预先设置的高度差阈值、且该点的高度小于该点对应的高度阈值的情况下,确定该点属于地面。
考虑到相邻的地面点之间的高度差较小,以及在雷达设备的安装高度确定之后,地面点的高度值与雷达设备的安装高度一致,故这里在将小组内的点排序之后,针对排序后的每个点,确定该点与其相邻点之间的高度差,在该高度差小于或等于高度差阈值,以及该点的高度小于该点对应的高度阈值时,确定该点为属于地面的地面点,进而较准确的确定了该小组内的地面点。
本公开实施方式中,针对每个小组,按该小组内每个点与采集该点的雷达设备之间的距离对小组内的点进行排序。具体实施时,可以基于每个点的三维坐标数据中的横坐标值和纵坐标值,确定每个点与采集该点的雷达设备之间的距离,并利用确定的每个点对应的距离对小组内的点进行排序。
针对排序后的每个点,确定该点与其相邻点之间的高度差。若该高度差小于或等于高度差阈值、且该点的高度小于该点对应的高度阈值时,确定该点属于地面;若该高度差大于高度差阈值、和/或该点的高度大于或等于该点对应的高度阈值时,确定该点不属于地面。
比如,若排序后的点包括点p1、点p2、点p3、点p4,则:针对点p1,利用点p1的三维坐标数据中的高度h 1和点p2的三维坐标数据中的高度h 2,计算点p1与点p2之间的高度差h 2-h 1,若该高度差小于或等于高度差阈值、且点p1的高度h 1小于点p1对应的高度阈值时,确定点p1属于地面;若该高度差大于高度差阈值,或者,点p1的高度h 1大于或等于点p1对应的高度阈值,则确认点p1不属于地面,即点p1不是地面点。由于点p4为排序后的最后一个点,则可以计算点p4与点p3之间的高度差,若该高度差小于或等于高度差阈值、且点p4的高度h 4小于点p4对应的高度阈值时,确定点p4属于地面。
可以根据下述步骤确定每个点对应的高度阈值:基于该点的坐标数据,确定该点与采集该点的雷达设备之间在水平面上的平面距离;基于采集该点的雷达设备的安装高度、所述平面距离、和设置的平面角度,确定该点对应的高度阈值。
具体实施时,可基于该点的横坐标值和纵坐标值,计算该点与采集该点的雷达设备之间在水平面上的平面距离。利用设置的平面角度的正切值和平面距离,确定该点处对应的平面高度;再基于采集该点的雷达设备的安装高度和确定的该点处对应的平面高度,确定该点对应的高度阈值。其中,平面角度用于表征该平面的倾斜程度,可以根据实际场景进行确定,比如,该平面角度可以为5°。
参见图5所示,图中示出了雷达设备51的安装高度为h,该雷达设备51采集的 点52对应的高度阈值为H,相对于地面53设置的平面角度为θ,l为平面距离。即可以利用平面角度θ的正切值、和平面距离l,确定该点处对应的平面高度h 0,在基于雷达设备51的安装高度h和确定的该点处对应的平面高度h 0,确定该点对应的高度阈值H,比如,高度阈值可以为H=h-h 0+a,a的值可以根据需要进行设置,比如a可以为0.08。或者,高度阈值可以为H=h-h 0
由于地面可能存在坡度(即地面存在一定的平面角度),使得不同位置处的地面与雷达设备之间的高度距离不同,即不同位置处的点可能对应不同的高度阈值。因此,通过针对每个点确定对应的高度阈值,可以较准确的判断每个点是否为属于地面的地面点。
在S403中,可以基于各个小组内的地面点,确定了点云数据中的地面点云数据,即各个小组内的各个地面点的点云数据,构成了点云数据中的地面点云数据。
下述对确定旋转角数据中的翻滚角和俯仰角,以及位移数据中的高度值的过程进行具体说明。
一种可选实施方式中,基于第一地面点云数据和第二地面点云数据,生成旋转角数据中的翻滚角和俯仰角、以及位移数据中的高度值,包括:基于第一地面点云数据和第二地面点云数据,生成第一雷达设备对应的第一地面拟合参数和第二雷达设备对应的第二地面拟合参数;基于第一地面拟合参数和第二地面拟合参数,确定旋转角数据中的翻滚角和俯仰角、以及位移数据中的高度值。
可以对每个雷达设备采集的地面点云数据进行拟合,生成每个雷达设备对应的地面拟合参数。其中,对地面点云数据进行拟合的方法有多种,此处仅为示例性说明。比如,可以利用随机抽样一致性算法(Random sample consensus,RANSAC),对地面点云数据进行拟合,生成地面点云数据对应的地面拟合参数。
一种可选实施方式中,基于第一地面拟合参数和第二地面拟合参数,确定旋转角数据中的翻滚角和俯仰角、以及位移数据中的高度值,包括:
一、基于第一地面拟合参数和第二地面拟合参数,确定第一地面拟合参数对应的第一拟合平面的法线和第二地面拟合参数对应的第二拟合平面的法线;
二、基于第一地面拟合参数对应的第一拟合平面的法线和第二地面拟合参数对应的第二拟合平面的法线,生成旋转角数据中的翻滚角和俯仰角;
三、基于旋转角数据中的翻滚角和俯仰角,对第一地面拟合参数进行调整,得到调整后的第一地面拟合参数;其中,调整后的第一地面拟合参数对应的拟合平面与第二地面拟合参数对应的拟合平面平行;
四、基于未调整的第二地面拟合参数和调整后的第一地面拟合参数,确定位移数据中的高度值。
第一雷达设备和第二雷达设备可以为不同的任意两个雷达设备。比如,在上述描述的应用场景中,第一雷达设备可以为主雷达设备,也可以为辅雷达设备。相应地,第二雷达设备可以为辅雷达设备,也可以为主雷达设备,只要与第一雷达设备不同。
这里,可以得到第一雷达设备对应的第一地面拟合参数和第二雷达设备对应的第二地面拟合参数,并可以确定第一地面拟合参数对应的第一拟合平面的第一法线、和第二地面拟合参数对应的第二拟合平面的第二法线。其中,第一拟合平面又可称为第一雷达设备对应的拟合平面,第二拟合平面又可称为第二雷达设备对应的拟合平面。换言之,第一雷达设备对应第一拟合平面,第二雷达设备对应第二拟合平面。比如,在第一拟合参数包括A 1、B 1、C 1、D 1时,第一拟合平面对应的第一拟合方程为:
A 1x+B 1y+C 1z+D 1=0;
在第二拟合参数包括A 2、B 2、C 2、D 2时,第二拟合平面对应的第一拟合方程为:
A 2x+B 2y+C 2z+D 2=0;
则第一法线的法向量为(A 1,B 1,C 1)、第二法线的法向量为(A 2,B 2,C 2)。
考虑到第一雷达设备的第一法线的法向量、与第二雷达设备的第二法线的法向量应该一致,故可以基于第一法线的法向量和第二法线的法向量,确定旋转角数据中的翻滚角、俯仰角。
进而可以基于确定的翻滚角、和俯仰角,对第一地面拟合参数进行调整,得到调整后的第一地面拟合参数。其中,调整后的第一地面拟合参数对应的第一调整拟合平面与未调整的第二地面拟合参数对应的第二拟合平面平行。然后,可以基于第二地面拟合参数、和调整后的第一地面拟合参数,计算调整后的第一地面拟合参数对应的第一调整拟合平面与第二拟合平面之间的距离,作为位移数据中的高度值。
由于地面是固定的,地面对应的不同法线应该为平行线。故可以基于不同雷达设备对应的拟合平面的法线,生成旋转角数据中的翻滚角和俯仰角。另外由于地面是固定的平面,不同雷达设备对应的拟合平面应该为处于同一高度的平面。故在利用翻滚角和俯仰角,对第一雷达设备对应的第一地面拟合参数进行调整后,可以基于未调整的地面拟合参数和调整后的地面拟合参数,确定位移数据中的高度值。
针对S302,示例性的,可以基于第一雷达设备对应的第一其他点云数据和第二雷达设备对应的第二其他点云数据,确定旋转角数据中的偏航角,以及位移数据中的长度值和宽度值。
一种可选实施方式中,该方法还包括:基于生成的旋转角数据中的翻滚角和俯仰角、以及位移数据中的高度值,对第一点云数据进行调整,生成调整后的第一点云数据;其中,调整后的第一点云数据与未调整的第二点云数据位于同一平面内,并且调整后的第一点云数据包括调整后的第一地面点云数据和调整后的第一其他点云数据。
基于第一点云数据中除第一地面点云数据之外的第一其他点云数据和第二点云数据中除第二地面点云数据之外的第二其他点云数据,确定旋转角数据中的偏航角、和位移数据中的长度值和宽度值,包括:基于第二点云数据中的第二其他点云数据、和调整后的第一点云数据中的第一其他点云数据,确定旋转角数据中的偏航角、和位移数据中的长度值和宽度值。
具体实施时,可以利用生成的翻滚角、俯仰角、以及高度值,和设置的初始偏航角、初始长度值、初始宽度值,生成中间坐标系转换矩阵;在利用生成的中间坐标系转换矩阵,对第一点云数据进行调整,生成调整后的第一点云数据。
在应用场景中,可以利用生成的翻滚角和俯仰角、以及高度值,对例如辅雷达设备或者主雷达设备的第一雷达设备采集的第一点云数据进行调整,生成调整后的第一点云数据,使得调整后的第一点云数据与未调整的第二点云数据位于同一平面内。
在生成调整后的第一点云数据之后,可以基于第二点云数据中除第二地面点云数据之外的第二其他点云数据、和调整后的第一点云数据中除第一地面点云数据之外的第一其他点云数据,确定旋转角数据中的偏航角、和位移数据中的长度值和宽度值。
或者,这里可以利用生成的中间坐标系转换矩阵,对第一点云数据中的除第一地面点云数据之外的第一其他点云数据进行调整,生成调整后的第一其他点云数据;再利用第二其他点云数据、和调整后的第一其他点云数据,确定旋转角数据中的偏航角、 和位移数据中的长度值和宽度值。
一种可选实施方式中,基于第一点云数据中除第一地面点云数据之外的第一其他点云数据和第二点云数据中除第二地面点云数据之外的第二其他点云数据,确定旋转角数据中的偏航角、和位移数据中的长度值和宽度值,包括:基于第一其他点云数据和第二其他点云数据所指示的每个点的长度坐标值和宽度坐标值,分别对第一其他点云数据和第二其他点云数据进行聚类,得到第一其他点云数据和第二其他点云数据分别对应的多个点云集合;其中,每个点云集合对应一个立体标识物体;基于每个点云集合中包括的每个点的坐标数据,确定每个点云对应的平均坐标数据;基于第一其他点云数据和第二其他点云数据分别对应的多个点云集合各自的平均坐标数据,确定旋转角数据中的偏航角、和位移数据中的长度值和宽度值。
在应用场景中,可以基于例如主雷达设备的第一雷达设备采集的第一其他点云数据指示的每个点的长度坐标值和宽度坐标值,对第一其他点云数据进行聚类,得到第一雷达设备对应的至少一个第一点云集合。具体实施时,聚类后得到的第一点云集合的数量与目标场景中设置的锥形桶(立体标识物体)的数量一致。比如,若目标场景中设置的锥形桶的数量为8个,则聚类可以得到8个第一点云集合。
同时,还可以基于例如辅雷达设备的第二雷达设备采集的第二其他点云数据指示的每个点的长度坐标值和宽度坐标值,对第二其他点云数据进行聚类,得到第二雷达设备对应的至少一个第二点云集合。由于辅雷达设备设置在目标车辆的侧面,故辅雷达设备仅可以对目标车辆的一侧进行检测,若目标场景中在目标车辆的左侧设置有4个锥形桶,在目标车辆的右侧设置有4个锥形桶,则对左侧辅雷达设备采集的第二其他点云数据聚类后,可以得到对应的4个第二点云集合;以及对右侧辅雷达设备采集的第二其他点云数据聚类后,也可以得到对应的4个第二点云集合。
进而,可以将作为第一雷达设备的主雷达设备聚类得到的8个第一点云集合按照左侧位置和右侧位置划分为左侧和右侧两类,并可以基于左侧的第一点云集合、和左侧辅雷达设备对应的4个第二点云集合,确定左侧辅雷达设备的偏航角、长度值和宽度值;以及基于右侧的第一点云集合、和右侧辅雷达设备对应的4个第二点云集合,确定右侧辅雷达设备的偏航角、长度值和宽度值。
具体实施时,对每个雷达设备采集的其他点云数据进行聚类的方法有多种,此处不进行具体限定。比如,可以使用RANSAC算法对其他点云数据进行聚类,得到点云集合。
针对每个第一点云集合,可对该第一点云集合中包括的各个点的坐标数据求平均,确定第一点云集合的第一平均坐标数据,进而可以得到各个第一点云集合的第一平均坐标数据;以及针对每个第二点云集合,对该第二点云集合中包括的各个点的坐标数据求平均,确定第二点云集合的第二平均坐标数据,进而可以得到各个第二点云集合的第二平均坐标数据。
可以基于至少一个第一点云集合各自的第一平均坐标数据、和至少一个第二点云集合各自的第二平均坐标数据,确定偏航角、长度值和宽度值。
在作为第二雷达设备的辅雷达设备的数量为多个时,针对每个辅雷达设备,从作为第一雷达设备的主雷达设备对应的至少一个第一点云集合中确定与该辅雷达设备对应的第一点云集合,再可以基于与该主雷达设备对应的第一点云集合的第一平均坐标数据、和该辅雷达设备对应的至少一个第二点云集合各自的第二平均坐标数据,确定该辅雷达设备对应的偏航角、长度值和宽度值。
具体实施时,可以使用迭代最近点算法(Iterative Closest Point,ICP),基于至少一 个第一点云集合各自的第一平均坐标数据、和至少一个第二点云集合各自的第二平均坐标数据,确定旋转角数据中的偏航角、和位移数据中的长度值和宽度值。
上述实施方式中,每个点云集合对应一个立体标识物体,由于立体标识物体的位置是固定的,故可以将立体标识物体作为基准点,确定旋转角数据中的偏航角、和位移数据中的长度值和宽度值。
针对S103,在确定了旋转角数据和位移数据之后,可以基于旋转角数据和位移数据,确定第一雷达设备和第二雷达设备之间的坐标系转换矩阵。可以根据下述公式确定坐标系转换矩阵:
Figure PCTCN2022077698-appb-000001
其中,T 3D为坐标系转换矩阵,T为由位移数据构成的三维向量,R为由旋转角数据确定的旋转矩阵。
一种可选实施方式中,在确定第一雷达设备和第二雷达设备之间的坐标系转换矩阵之后,该方法还包括:利用生成的坐标系转换矩阵,将第一雷达设备采集的第一点云数据和第二雷达设备采集的第二点云数据转换至同一坐标系下;基于转换后的点云数据,确定目标对象的位姿数据。
在应用场景中,在生成坐标系转换矩阵之后,可以利用生成的坐标系转换矩阵,将作为第二雷达设备的辅雷达设备采集的第二点云数据转换至作为第一雷达设备的主雷达设备采集的第一点云数据所处的坐标系下;再将第一点云数据和坐标转换后的第二点云数据融合,利用融合后的点云数据确定目标车辆在移动过程中检测到的目标对象的位姿数据。或者,也可以利用生成的坐标系转换矩阵,将作为第一雷达设备的主雷达设备采集的第一点云数据转换至作为第二雷达设备的辅雷达设备采集的第二点云数据所处的坐标系下;再将第二点云数据和坐标转换后的第一点云数据融合,利用融合后的点云数据确定目标车辆在移动过程中检测到的目标对象的位姿数据。
进而,可以基于目标车辆在移动过程中检测到的目标对象的位姿数据,控制目标车辆,比如,可以控制目标车辆的加速、减速、转向、制动等,或者可以播放语音提示信息,以提示驾驶员控制目标车辆加速、减速、转向、制动等。
采用上述方法,可以利用生成的坐标系转换矩阵,将不同雷达设备采集的点云数据转换至同一坐标系下,由于生成的坐标系转换矩阵的准确率较高,使得坐标转换后的点云数据较为准确,进而提高了确定的目标对象的位姿数据的准确度。
本领域技术人员可以理解,在具体实施方式的上述方法中,各步骤的撰写顺序并不意味着严格的执行顺序而对实施过程构成任何限定,各步骤的具体执行顺序应当以其功能和可能的内在逻辑确定。
基于相同的构思,本公开实施例还提供了一种外参标定装置,参见图6所示,为本公开实施例提供的外参标定装置的架构示意图,包括获取模块601、生成模块602、确定模块603。
获取模块601,用于获取设置在目标车辆上的至少两个雷达设备分别采集同一目标场景得到的点云数据;
生成模块602,用于基于第一雷达设备采集的第一点云数据和第二雷达设备采集的第二点云数据,生成旋转角数据和位移数据;其中,所述旋转角数据用于表征所述第一点云数据和所述第二点云数据中对应的点在不同方向上的角度偏差,所述位移数据用于表征所述第一点云数据和所述第二点云数据中对应的点在不同方向上位移偏差, 其中,所述第一雷达设备和所述第二雷达设备为所述至少两个雷达设备中任意两个不同的雷达设备;
确定模块603,用于基于所述旋转角数据和所述位移数据,确定所述第一雷达设备和所述第二雷达设备之间的坐标系转换矩阵。
一种可能的实施方式中,所述生成模块602,在基于所述第一雷达设备采集的第一点云数据和所述第二雷达设备采集的第二点云数据,生成旋转角数据和位移数据时,用于:分别从所述第一点云数据和所述第二点云数据中提取地第一地面点云数据和第二面点云数据;并基于所述第一地面点云数据和所述第二地面点云数据,生成所述旋转角数据中的翻滚角和俯仰角,以及所述位移数据中的高度值;基于所述第一点云数据中除所述第一地面点云数据之外的第一其他点云数据和所述第二点云数据中除所述第二地面点云数据之外的第二其他点云数据,确定所述旋转角数据中的偏航角,以及所述位移数据中的长度值和宽度值。
一种可能的实施方式中,所述生成模块602,在分别从所述第一点云数据和所述第二点云数据中提取第一地面点云数据和第二地面点云数据时,用于:针对所述第一雷达设备和所述第二雷达设备中的每个雷达设备,基于从所述雷达设备的位置指向所述雷达设备采集的点云数据中每个点的位置的方向角,将所述雷达设备采集的点云数据中的点划分为多个小组,其中同一小组内的多个点分别对应的所述方向角之间的差值小于设定阈值;针对每个小组,基于该小组内的每个点与其相邻点之间的高度差,确定该小组内的地面点;基于各个所述小组内的地面点,确定所述雷达设备采集的点云数据中的地面点云数据。
一种可能的实施方式中,所述生成模块602,在针对每个小组,基于该小组内的每个点与其相邻点之间的高度差,确定该小组内的地面点时,用于:基于所述小组内每个点与采集该点的所述雷达设备之间的距离,将所述小组内的点进行排序;针对排序后的每个点,确定该点与其相邻点之间的高度差;在所述高度差小于或等于预先设置的高度差阈值、且该点的高度小于该点对应的高度阈值的情况下,确定所述点为属于地面的地面点。
一种可能的实施方式中,所述生成模块602,用于根据下述步骤确定每个点对应的高度阈值:基于所述点的坐标数据,确定所述点与采集该点的所述雷达设备之间在水平面上的平面距离;基于采集该点的所述雷达设备的安装高度、所述平面距离、和设置的平面角度,确定所述点对应的所述高度阈值。
一种可能的实施方式中,所述生成模块602,在基于所述第一地面点云数据和所述第二地面点云数据,生成所述旋转角数据中的翻滚角和俯仰角、以及所述位移数据中的高度值时,用于:基于所述第一地面点云数据和所述第二地面点云数据,生成所述第一雷达设备对应的第一地面拟合参数和所述第二雷达设备对应的第二地面拟合参数;基于所述第一地面拟合参数和所述第二地面拟合参数,确定所述旋转角数据中的翻滚角和俯仰角、以及所述位移数据中的高度值。
一种可能的实施方式中,所述生成模块602,在基于所述第一地面拟合参数和所述第二地面拟合参数,确定所述旋转角数据中的翻滚角和俯仰角、以及所述位移数据中的高度值时,用于:基于所述第一地面拟合参数和所述第二地面拟合参数,确定所述第一地面拟合参数对应的第一拟合平面的法线和所述第二地面拟合参数对应的第二拟合平面的法线;基于所述第一地面拟合参数对应的第一拟合平面的法线和所述第二地面拟合参数对应的第二拟合平面的法线,生成所述旋转角数据中的翻滚角和俯仰角;基于所述旋转角数据中的翻滚角和俯仰角,对所述第一地面拟合参数进行调整,得到调整后的第一地面拟合参数;其中,所述调整后的第一地面拟合参数对应的拟合平面、 与所述第二地面拟合参数对应的拟合平面平行;基于所述第二地面拟合参数和所述调整后的第一地面拟合参数,确定所述位移数据中的高度值。
一种可能的实施方式中,所述装置还包括:调整模块604,在基于所述第一点云数据中除所述第一地面点云数据之外的第一其他点云数据和所述第二点云数据中除所述第二地面点云数据之外的第二其他点云数据,确定所述旋转角数据中的偏航角、和所述位移数据中的长度值和宽度值时,用于:基于生成的所述旋转角数据中的翻滚角和俯仰角、以及所述位移数据中的高度值,对所述第一点云数据进行调整,生成调整后的第一点云数据;其中,所述调整后的第一点云数据与未调整的所述第二点云数据位于同一平面内;
所述生成模块602,用于:基于所述第二点云数据中的所述第二其他点云数据、和所述调整后的第一点云数据中的所述第一其他点云数据,确定所述旋转角数据中的偏航角、和所述位移数据中的长度值和宽度值。
一种可能的实施方式中,所述目标车辆停放在平直的道路上,所述目标场景包括放置在所述目标车辆的四周的多个立体标识物体;
所述生成模块602,在基于所述第一点云数据中除所述第一地面点云数据之外的第一其他点云数据和所述第二点云数据中除所述第二地面点云数据之外的第二其他点云数据,确定所述旋转角数据中的偏航角、和所述位移数据中的长度值和宽度值时,用于:基于所述第一其他点云数据和所述第二其他点云数据所指示的每个点的长度坐标值和宽度坐标值,分别对所述第一其他点云数据和所述第二其他点云数据进行聚类,得到所述第一其他点云数据和所述第二其他点云数据分别对应的多个点云集合;其中,每个所述点云集合对应一个所述立体标识物体;基于每个所述点云集合中包括的每个点的坐标数据,确定每个所述点云集合的平均坐标数据;基于所述第一其他点云数据和所述第二其他点云数据分别对应的多个点云集合各自的所述平均坐标数据,确定所述旋转角数据中的偏航角、和所述位移数据中的长度值和宽度值。
一种可能的实施方式中,在确定所述第一雷达设备和所述第二雷达设备之间的坐标系转换矩阵之后,所述装置还包括:转换模块605,用于:利用生成的所述坐标系转换矩阵,将所述第一雷达设备采集的第一点云数据和所述第二雷达设备采集的第二点云数据转换至同一坐标系下;基于转换后的点云数据,确定目标对象的位姿数据。
在一些实施例中,本公开实施例提供的装置具有的功能或包含的模板可以用于执行上文方法实施例描述的方法,其具体实现可以参照上文方法实施例的描述,为了简洁,这里不再赘述。
基于同一技术构思,本公开实施例还提供了一种电子设备。参照图7所示,为本公开实施例提供的电子设备的结构示意图,包括处理器701、存储器702、和总线703。其中,存储器702用于存储执行指令,包括内存7021和外部存储器7022;这里的内存7021也称内存储器,用于暂时存放处理器701中的运算数据,以及与硬盘等外部存储器7022交换的数据,处理器701通过内存7021与外部存储器7022进行数据交换,当电子设备700运行时,处理器701与存储器702之间通过总线703通信,使得处理器701在执行以下指令:获取设置在目标车辆上的至少两个雷达设备分别采集同一目标场景得到的点云数据;基于第一雷达设备采集的第一点云数据和第二雷达设备采集的第二点云数据,生成旋转角数据和位移数据;其中,所述旋转角数据用于表征所述第一点云数据和所述第二点云数据中对应的点在不同方向上的角度偏差,所述位移数据用于表征所述第一点云数据和所述第二点云数据中对应的点在不同方向上位移偏差,其中,所述第一雷达设备和所述第二雷达设备为所述至少两个雷达设备中任意两个不同的雷达设备;基于所述旋转角数据和所述位移数据,确定所述第一雷达设备和所述第 二雷达设备之间的坐标系转换矩阵。
此外,本公开实施例还提供一种计算机可读存储介质,该计算机可读存储介质上存储有计算机程序,该计算机程序被处理器运行时执行上述方法实施例中所述的外参标定方法的步骤。其中,该存储介质可以是易失性或非易失的计算机可读取存储介质。
本公开实施例还提供一种计算机程序产品,该计算机程序产品承载有程序代码,所述程序代码包括的指令可用于执行上述方法实施例中所述的外参标定方法的步骤,具体可参见上述方法实施例,在此不再赘述。
其中,上述计算机程序产品可以具体通过硬件、软件或其结合的方式实现。在一个可选实施例中,所述计算机程序产品具体体现为计算机存储介质,在另一个可选实施例中,计算机程序产品具体体现为软件产品,例如软件开发包(Software Development Kit,SDK)等等。
所属领域的技术人员可以清楚地了解到,为描述的方便和简洁,上述描述的系统和装置的具体工作过程,可以参考前述方法实施例中的对应过程,在此不再赘述。在本公开所提供的几个实施例中,应该理解到,所揭露的系统、装置和方法,可以通过其它的方式实现。以上所描述的装置实施例仅仅是示意性的,例如,所述单元的划分,仅仅为一种逻辑功能划分,实际实现时可以有另外的划分方式,又例如,多个单元或组件可以结合或者可以集成到另一个系统,或一些特征可以忽略,或不执行。另一点,所显示或讨论的相互之间的耦合或直接耦合或通信连接可以是通过一些通信接口,装置或单元的间接耦合或通信连接,可以是电性,机械或其它的形式。
所述作为分离部件说明的单元可以是或者也可以不是物理上分开的,作为单元显示的部件可以是或者也可以不是物理单元,即可以位于一个地方,或者也可以分布到多个网络单元上。可以根据实际的需要选择其中的部分或者全部单元来实现本实施例方案的目的。
另外,在本公开各个实施例中的各功能单元可以集成在一个处理单元中,也可以是各个单元单独物理存在,也可以两个或两个以上单元集成在一个单元中。
所述功能如果以软件功能单元的形式实现并作为独立的产品销售或使用时,可以存储在一个处理器可执行的非易失的计算机可读取存储介质中。基于这样的理解,本公开的技术方案本质上或者说对现有技术做出贡献的部分或者该技术方案的部分可以以软件产品的形式体现出来,该计算机软件产品存储在一个存储介质中,包括若干指令用以使得一台计算机设备(可以是个人计算机,服务器,或者网络设备等)执行本公开各个实施例所述方法的全部或部分步骤。而前述的存储介质包括:U盘、移动硬盘、只读存储器(Read-Only Memory,ROM)、随机存取存储器(Random Access Memory,RAM)、磁碟或者光盘等各种可以存储程序代码的介质。
以上仅为本公开的具体实施方式,但本公开的保护范围并不局限于此,任何熟悉本技术领域的技术人员在本公开揭露的技术范围内,可轻易想到变化或替换,都应涵盖在本公开的保护范围之内。因此,本公开的保护范围应以权利要求的保护范围为准。

Claims (13)

  1. 一种外参标定方法,包括:
    获取设置在目标车辆上的至少两个雷达设备分别采集同一目标场景得到的点云数据;
    基于第一雷达设备采集的第一点云数据和第二雷达设备采集的第二点云数据,生成旋转角数据和位移数据;其中,所述旋转角数据用于表征所述第一点云数据和所述第二点云数据中对应的点在不同方向上的角度偏差,所述位移数据用于表征所述第一点云数据和所述第二点云数据中对应的点在不同方向上位移偏差,其中,所述第一雷达设备和所述第二雷达设备为所述至少两个雷达设备中任意两个不同的雷达设备;
    基于所述旋转角数据和所述位移数据,确定所述第一雷达设备和所述第二雷达设备之间的坐标系转换矩阵。
  2. 根据权利要求1所述的方法,其特征在于,所述基于所述第一雷达设备采集的第一点云数据和所述第二雷达设备采集的第二点云数据,生成旋转角数据和位移数据,包括:
    分别从所述第一点云数据和所述第二点云数据中提取第一地面点云数据和第二地面点云数据;
    基于所述第一地面点云数据和所述第二地面点云数据,生成所述旋转角数据中的翻滚角和俯仰角,以及所述位移数据中的高度值;
    基于所述第一点云数据中除所述第一地面点云数据之外的第一其他点云数据和所述第二点云数据中除所述第二地面点云数据之外的第二其他点云数据,确定所述旋转角数据中的偏航角,以及所述位移数据中的长度值和宽度值。
  3. 根据权利要求2所述的方法,其特征在于,所述分别从所述第一点云数据和所述第二点云数据中提取第一地面点云数据和第二地面点云数据,包括:
    针对所述第一雷达设备和所述第二雷达设备中的每个雷达设备,
    基于从所述雷达设备的位置指向所述雷达设备采集的点云数据中每个点的位置的方向角,将所述雷达设备采集的点云数据中的点划分为多个小组,其中同一小组内的多个点分别对应的所述方向角之间的差值小于设定阈值;
    针对每个小组,基于该小组内的每个点与其相邻点之间的高度差,确定该小组内的地面点;
    基于各个所述小组内的地面点,确定所述雷达设备采集的点云数据中的地面点云数据。
  4. 根据权利要求3所述的方法,其特征在于,针对每个小组,基于该小组内的每个点与其相邻点之间的高度差,确定该小组内的地面点,包括:
    基于所述小组内每个点与采集该点的所述雷达设备之间的距离,将所述小组内的点进行排序;
    针对排序后的每个点,
    确定该点与其相邻点之间的高度差;
    在所述高度差小于或等于预先设置的高度差阈值、且该点的高度小于该点对应的高度阈值的情况下,确定所述点为属于地面的地面点。
  5. 根据权利要求4所述的方法,其特征在于,根据下述步骤确定每个点对应的高度阈值:
    基于所述点的坐标数据,确定所述点与采集该点的雷达设备之间在水平面上的平面距离;
    基于采集该点的雷达设备的安装高度、所述平面距离、和设置的平面角度,确定所述点对应的所述高度阈值。
  6. 根据权利要求2至5任一所述的方法,其特征在于,所述基于所述第一地面点云数据和所述第二地面点云数据,生成所述旋转角数据中的翻滚角和俯仰角、以及所述位移数据中的高度值,包括:
    基于所述第一地面点云数据和所述第二地面点云数据,生成所述第一雷达设备对应的第一地面拟合参数和所述第二雷达设备对应的第二地面拟合参数;
    基于所述第一地面拟合参数和所述第二地面拟合参数,确定所述旋转角数据中的翻滚角和俯仰角、以及所述位移数据中的高度值。
  7. 根据权利要求6所述的方法,其特征在于,基于所述第一地面拟合参数和所述第二地面拟合参数,确定所述旋转角数据中的翻滚角和俯仰角、以及所述位移数据中的高度值,包括:
    基于所述第一地面拟合参数和所述第二地面拟合参数,确定所述第一地面拟合参数对应的第一拟合平面的法线和所述第二地面拟合参数对应的第二拟合平面的法线;
    基于所述第一地面拟合参数对应的第一拟合平面的法线和所述第二地面拟合参数对应的第二拟合平面的法线,生成所述旋转角数据中的翻滚角和俯仰角;
    基于所述旋转角数据中的翻滚角和俯仰角,对所述第一地面拟合参数进行调整,得到调整后的第一地面拟合参数;其中,所述调整后的第一地面拟合参数对应的拟合平面、与所述第二地面拟合参数对应的拟合平面平行;
    基于所述第二地面拟合参数和所述调整后的第一地面拟合参数,确定所述位移数据中的高度值。
  8. 根据权利要求2至7任一所述的方法,其特征在于,所述基于所述第一点云数据中除所述第一地面点云数据之外的第一其他点云数据和所述第二点云数据中除所述第二地面点云数据之外的第二其他点云数据,确定所述旋转角数据中的偏航角、和所述位移数据中的长度值和宽度值,包括:
    基于生成的所述旋转角数据中的翻滚角和俯仰角、以及所述位移数据中的高度值,对所述第一点云数据进行调整,生成调整后的第一点云数据;其中,所述调整后的第一点云数据与未调整的所述第二点云数据位于同一平面内;
    基于所述第二点云数据中的所述第二其他点云数据、和所述调整后的第一点云数据中的所述第一其他点云数据,确定所述旋转角数据中的偏航角、和所述位移数据中的长度值和宽度值。
  9. 根据权利要求2至8任一所述的方法,其特征在于,所述目标车辆停放在平直的道路上,所述目标场景包括放置在所述目标车辆的四周的多个立体标识物体;基于所述第一点云数据中除所述第一地面点云数据之外的第一其他点云数据和所述第二点云数据中除所述第二地面点云数据之外的第二其他点云数据,确定所述旋转角数据中的偏航角、和所述位移数据中的长度值和宽度值,包括:
    基于所述第一其他点云数据和所述第二其他点云数据所指示的每个点的长度坐标 值和宽度坐标值,分别对所述第一其他点云数据和所述第二其他点云数据进行聚类,得到所述第一其他点云数据和所述第二其他点云数据分别对应的多个点云集合;其中,每个所述点云集合对应一个所述立体标识物体;
    基于每个所述点云集合中包括的每个点的坐标数据,确定每个所述点云集合的平均坐标数据;
    基于所述第一其他点云数据和所述第二其他点云数据分别对应的多个点云集合各自的所述平均坐标数据,确定所述旋转角数据中的偏航角、和所述位移数据中的长度值和宽度值。
  10. 根据权利要求1至9任一所述的方法,其特征在于,在确定所述第一雷达设备和所述第二雷达设备之间的坐标系转换矩阵之后,所述方法还包括:
    利用生成的所述坐标系转换矩阵,将所述第一雷达设备采集的第一点云数据和所述第二雷达设备采集的第二点云数据转换至同一坐标系下;
    基于转换后的点云数据,确定目标对象的位姿数据。
  11. 一种外参标定装置,其特征在于,包括:
    获取模块,用于获取设置在目标车辆上的至少两个雷达设备分别采集同一目标场景得到的点云数据;
    生成模块,用于基于第一雷达设备采集的第一点云数据和第二雷达设备采集的第二点云数据,生成旋转角数据和位移数据;其中,所述旋转角数据用于表征所述第一点云数据和所述第二点云数据中对应的点在不同方向上的角度偏差,所述位移数据用于表征所述第一点云数据和所述第二点云数据中对应的点在不同方向上位移偏差,其中,所述第一雷达设备和所述第二雷达设备为所述至少两个雷达设备中任意两个不同的雷达设备;
    确定模块,用于基于所述旋转角数据和所述位移数据,确定所述第一雷达设备和所述第二雷达设备之间的坐标系转换矩阵。
  12. 一种电子设备,包括:处理器、存储器和总线,所述存储器存储有所述处理器可执行的机器可读指令,当电子设备运行时,所述处理器与所述存储器之间通过总线通信,所述机器可读指令被所述处理器执行时执行如权利要求1至10任一所述的外参标定方法的步骤。
  13. 一种计算机可读存储介质,其上存储有计算机程序,该计算机程序被处理器运行时执行如权利要求1至10任一所述的外参标定方法的步骤。
PCT/CN2022/077698 2021-02-26 2022-02-24 外参标定方法、装置、电子设备及存储介质 WO2022179566A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202110219865.7A CN112946591A (zh) 2021-02-26 2021-02-26 外参标定方法、装置、电子设备及存储介质
CN202110219865.7 2021-02-26

Publications (1)

Publication Number Publication Date
WO2022179566A1 true WO2022179566A1 (zh) 2022-09-01

Family

ID=76246630

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2022/077698 WO2022179566A1 (zh) 2021-02-26 2022-02-24 外参标定方法、装置、电子设备及存储介质

Country Status (2)

Country Link
CN (1) CN112946591A (zh)
WO (1) WO2022179566A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117590362A (zh) * 2024-01-19 2024-02-23 深圳海星智驾科技有限公司 一种多激光雷达外参标定方法、装置和设备

Families Citing this family (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112946591A (zh) * 2021-02-26 2021-06-11 商汤集团有限公司 外参标定方法、装置、电子设备及存储介质
CN113284197B (zh) * 2021-07-22 2021-11-23 浙江华睿科技股份有限公司 Agv的tof相机外参标定方法及装置、电子设备
CN113640756B (zh) * 2021-08-11 2024-05-17 北京航迹科技有限公司 一种数据标定方法、系统、装置、计算机程序以及存储介质
CN113655453B (zh) * 2021-08-27 2023-11-21 阿波罗智能技术(北京)有限公司 用于传感器标定的数据处理方法、装置及自动驾驶车辆
CN114829971A (zh) * 2021-08-30 2022-07-29 华为技术有限公司 一种激光雷达的标定方法、装置及存储介质
CN116381632B (zh) * 2023-06-05 2023-08-18 南京隼眼电子科技有限公司 雷达横滚角的自标定方法、装置及存储介质

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109839624A (zh) * 2017-11-27 2019-06-04 北京万集科技股份有限公司 一种多激光雷达位置标定方法及装置
CN110031824A (zh) * 2019-04-12 2019-07-19 杭州飞步科技有限公司 激光雷达联合标定方法及装置
US20200081119A1 (en) * 2018-09-07 2020-03-12 Baidu Online Network Technology (Beijing) Co., Ltd. Method and apparatus for determining relative pose, device and medium
CN111090084A (zh) * 2018-10-24 2020-05-01 舜宇光学(浙江)研究院有限公司 多激光雷达外参标定方法、标定装置、标定系统和电子设备
CN111965627A (zh) * 2020-08-18 2020-11-20 湖北亿咖通科技有限公司 一种车辆的多激光雷达标定方法
CN112946612A (zh) * 2021-03-29 2021-06-11 上海商汤临港智能科技有限公司 外参标定方法、装置、电子设备及存储介质
CN112946591A (zh) * 2021-02-26 2021-06-11 商汤集团有限公司 外参标定方法、装置、电子设备及存储介质

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109521403B (zh) * 2017-09-19 2020-11-20 百度在线网络技术(北京)有限公司 多线激光雷达的参数标定方法及装置、设备及可读介质
CN109696663B (zh) * 2019-02-21 2021-02-09 北京大学 一种车载三维激光雷达标定方法和系统
CN111724429B (zh) * 2019-03-21 2024-03-01 北京京东乾石科技有限公司 地面特征提取方法及装置
CN110320506B (zh) * 2019-08-06 2023-05-02 深圳市海梁科技有限公司 一种车用激光雷达自动标定装置及方法
CN111427026B (zh) * 2020-02-21 2023-03-21 深圳市镭神智能系统有限公司 激光雷达的标定方法、装置、存储介质及自移动设备
CN112164063A (zh) * 2020-10-30 2021-01-01 上海商汤临港智能科技有限公司 一种数据处理方法及装置

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109839624A (zh) * 2017-11-27 2019-06-04 北京万集科技股份有限公司 一种多激光雷达位置标定方法及装置
US20200081119A1 (en) * 2018-09-07 2020-03-12 Baidu Online Network Technology (Beijing) Co., Ltd. Method and apparatus for determining relative pose, device and medium
CN111090084A (zh) * 2018-10-24 2020-05-01 舜宇光学(浙江)研究院有限公司 多激光雷达外参标定方法、标定装置、标定系统和电子设备
CN110031824A (zh) * 2019-04-12 2019-07-19 杭州飞步科技有限公司 激光雷达联合标定方法及装置
CN111965627A (zh) * 2020-08-18 2020-11-20 湖北亿咖通科技有限公司 一种车辆的多激光雷达标定方法
CN112946591A (zh) * 2021-02-26 2021-06-11 商汤集团有限公司 外参标定方法、装置、电子设备及存储介质
CN112946612A (zh) * 2021-03-29 2021-06-11 上海商汤临港智能科技有限公司 外参标定方法、装置、电子设备及存储介质

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117590362A (zh) * 2024-01-19 2024-02-23 深圳海星智驾科技有限公司 一种多激光雷达外参标定方法、装置和设备
CN117590362B (zh) * 2024-01-19 2024-04-16 深圳海星智驾科技有限公司 一种多激光雷达外参标定方法、装置和设备

Also Published As

Publication number Publication date
CN112946591A (zh) 2021-06-11

Similar Documents

Publication Publication Date Title
WO2022179566A1 (zh) 外参标定方法、装置、电子设备及存储介质
EP3407294A1 (en) Information processing method, device, and terminal
CN110782465B (zh) 一种基于激光雷达的地面分割方法、装置及存储介质
US11783507B2 (en) Camera calibration apparatus and operating method
JP2002352225A (ja) 障害物検出装置及びその方法
CN112051575B (zh) 一种毫米波雷达与激光雷达的调整方法及相关装置
CN111383279A (zh) 外参标定方法、装置及电子设备
CN108362205B (zh) 基于条纹投影的空间测距方法
JP2014523572A (ja) 地図データの生成
EP4016457A1 (en) Positioning method and apparatus
WO2020259506A1 (zh) 一种确定摄像头的畸变参数的方法及装置
CN114047487B (zh) 雷达和车体的外参标定方法、装置、电子设备和存储介质
WO2022217988A1 (zh) 传感器配置方案确定方法、装置、计算机设备、存储介质及程序
JP2017181476A (ja) 車両位置検出装置、車両位置検出方法及び車両位置検出用コンピュータプログラム
CN113569958A (zh) 激光点云数据聚类方法、装置、设备及介质
CN110160503B (zh) 一种顾及高程的无人机景观匹配定位方法
CN107765257A (zh) 一种基于反射强度辅助外部校准的激光探测与测量方法
CN114241057A (zh) 相机和激光雷达的外参标定方法、系统及可读存储介质
CN111862208A (zh) 一种基于屏幕光通信的车辆定位方法、装置及服务器
CN113033248A (zh) 一种图像识别方法及装置、计算机可读存储介质
CN110542422A (zh) 机器人定位方法、装置、机器人及存储介质
CN113763457B (zh) 落差地形的标定方法、装置、电子设备和存储介质
WO2021253333A1 (zh) 一种基于屏幕光通信的车辆定位方法、装置及服务器
WO2019150483A1 (ja) 速度算出装置、速度算出方法、及び、プログラム
TWI819613B (zh) 物件的雙感測方法及用於物件感測的運算裝置

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 22758930

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE