CN113933820A - Laser radar external reference calibration method without calibration object - Google Patents
Laser radar external reference calibration method without calibration object Download PDFInfo
- Publication number
- CN113933820A CN113933820A CN202111538615.6A CN202111538615A CN113933820A CN 113933820 A CN113933820 A CN 113933820A CN 202111538615 A CN202111538615 A CN 202111538615A CN 113933820 A CN113933820 A CN 113933820A
- Authority
- CN
- China
- Prior art keywords
- point
- point cloud
- laser radar
- points
- current point
- Prior art date
- Legal status (The legal status 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 status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO 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
- G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
- G01S7/48—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S17/00
- G01S7/497—Means for monitoring or calibrating
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO 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
- G01S17/00—Systems using the reflection or reradiation of electromagnetic waves other than radio waves, e.g. lidar systems
- G01S17/88—Lidar systems specially adapted for specific applications
- G01S17/89—Lidar systems specially adapted for specific applications for mapping or imaging
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Computer Networks & Wireless Communication (AREA)
- General Physics & Mathematics (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Electromagnetism (AREA)
- Optical Radar Systems And Details Thereof (AREA)
Abstract
The invention discloses a calibration object-free laser radar external reference calibration method, which relates to the technical field of laser radars and comprises the steps of obtaining a point cloud map, and enabling a laser radar to be calibrated to be located in the range of the point cloud map; scanning a laser radar to be calibrated to obtain a frame of original point cloud data, and extracting line features and surface features of the original point cloud data; positioning the laser radar to be calibrated through a positioning device to obtain a positioning point of the laser radar to be calibrated, and then loading a local point cloud map with the positioning point as a circle center; and roughly matching the line features and the surface features extracted from the original point cloud data with the line features and the surface features of the local point cloud map to obtain corresponding yaw angles and displacements, and using the corresponding yaw angles and displacements as initial external parameters of the laser radar to be calibrated. The method and the device directly obtain the pose, namely the external reference, of the laser radar to be calibrated in the map coordinate system by fusing the laser radar to be calibrated and the point cloud map established in advance, and have high calibration precision and high efficiency.
Description
Technical Field
The invention relates to the technical field of laser radars, in particular to a laser radar external reference calibration method without a calibration object.
Background
The purpose of sensor calibration is to obtain the attitude of the sensor in a certain coordinate system, i.e. the position and orientation of 2D/3D, which is the external reference of the sensor relative to the coordinate system. So if one wants to calibrate one sensor, it is clear which coordinate system the external reference needs to be acquired with respect to. In the case of driverless applications, this target coordinate system is usually the own vehicle coordinate system or the map coordinate system.
In the automatic driving technology, the position of an object is detected by using a laser radar arranged at the end of a vehicle and a laser radar arranged at the roadside, and the laser radar arranged at the end of the vehicle and other various sensors are fused to position the self-vehicle. In the application scenes needing to fuse laser radar data, external parameters of the laser radar need to be calibrated firstly.
At present, the common method for calibrating external parameters of laser radars in the industry is as follows: and fusion calibration among the multiple laser radars, namely scanning point clouds of the same scene by using a laser radar to be calibrated and a reference laser radar respectively, and then registering the two pairs of acquired point cloud data by icp or ndt, so that the pose change between the two laser radars is calculated, and the external reference of the laser radar to be calibrated in a reference laser radar coordinate system can be obtained. However, this calibration method brings about problems: one is how to obtain external parameters of a reference laser radar, and only knowing the pose relation of the laser radar to be calibrated relative to the reference laser radar may not meet the requirement of data fusion in unmanned driving; the other is that the calibration method has higher requirements on the installation positions of the two laser radars, and requires that the scenes scanned by the two laser radars have more overlapped areas enough for point cloud registration; furthermore, if there is only one lidar to be calibrated, no reference lidar can be provided, and this approach is not suitable.
Disclosure of Invention
The invention aims to solve the technical problem of overcoming the defects of the prior art and provides a laser radar external reference calibration method without a calibration object.
In order to solve the technical problems, the technical scheme of the invention is as follows:
a calibration method of external reference of laser radar without calibration object includes,
acquiring a pre-established point cloud map, and enabling the laser radar to be calibrated to be located in the range of the point cloud map;
scanning a laser radar to be calibrated to obtain a frame of original point cloud data, and then extracting line features and surface features of the original point cloud data;
positioning the laser radar to be calibrated through a positioning device to obtain a positioning point of the laser radar to be calibrated, and then loading a local point cloud map with the positioning point as a circle center;
and matching the line features and the surface features extracted from the original point cloud data with the line features and the surface features of the local point cloud map to obtain corresponding yaw angles and displacements, and taking the corresponding yaw angles and the corresponding displacements as initial external parameters of the laser radar to be calibrated.
As a preferred scheme of the calibration method of the laser radar external reference without the calibration object, the method comprises the following steps: after the line features and the surface features extracted from the original point cloud data are matched with the line features and the surface features of the local point cloud map to obtain corresponding yaw angle and displacement which are used as initial external parameters of the laser radar to be calibrated, the method also comprises the following steps of,
loading a local point cloud map by taking the initial external parameters as a center;
searching matching points and matching data corresponding to line features and surface features extracted from original point cloud data in a local point cloud map, then projecting point cloud points onto the point cloud map, and adjusting initial external parameters to obtain optimized external parameters;
scanning a laser radar to be calibrated to obtain a new frame of original point cloud data, and extracting line features and surface features from the new frame of original point cloud data;
loading a local point cloud map by taking the optimized external parameters as a center;
searching matching points and matching data corresponding to line features and surface features extracted from a new frame of original point cloud data in a local point cloud map, projecting point cloud points to the point cloud map, and adjusting the optimized external parameters again to obtain the optimized external parameters;
and repeating the steps for a plurality of times until the error of the result obtained by the optimization calculation of the point cloud data of the front frame and the back frame is smaller than the set threshold value, and obtaining the external parameters of the laser radar to be calibrated.
As a preferred scheme of the calibration method of the laser radar external reference without the calibration object, the method comprises the following steps: the extracting line features and surface features of the raw point cloud data includes,
splitting the original point cloud data;
calculating the curvature of a point on each wire harness;
judging whether an included angle between a current point and a next point in the wire harness is larger than 1 degree, if so, rejecting the current point, ten continuous points in front of the current point and ten continuous points behind the current point;
dividing the rest points on each wire harness into a plurality of sections;
screening line characteristics of points in each section according to the sequence of curvature from large to small, sequentially judging whether a current point is a shielding point or not, if so, setting the current point, ten continuous points in front of the current point and ten continuous points behind the current point as dead points, then sequentially judging whether the curvature of the current point is larger than a preset threshold value or not, if the curvature of the current point is larger than the preset threshold value and the current point is not the dead point, determining that the current point is the line characteristics, and after the current point is determined as the line characteristics, setting the current point, ten continuous points in front of the current point and ten continuous points behind the current point as the dead points;
screening surface characteristics of points in each section according to the sequence of curvature from small to large, sequentially judging whether a current point is parallel to a laser beam, if so, setting the current point, ten continuous points in front of the current point and ten continuous points behind the current point as dead points, then sequentially judging whether the curvature of the current point is smaller than a preset threshold value, if the curvature of the current point is smaller than the preset threshold value and the current point is not the dead point, determining the current point as the surface characteristics, and after the current point is determined as the surface characteristics, setting the current point, ten continuous points in front of the current point and ten continuous points behind the current point as the dead points.
As a preferred scheme of the calibration method of the laser radar external reference without the calibration object, the method comprises the following steps: the calculating of the curvature of the point on each line bundle includes,
calculating the curvature of the ith point on the kth wire harness according to the formula IComprises the following steps:
，
wherein the content of the first and second substances,
i point, j =1,2,3, …,10, representing the point cloud on the k-th line bundle;
calculated according to the formula I
,
,…, 
Then, sorting the ten curvatures, and taking a median value as the curvature of the ith point on the kth line beam;
the curvature of the point on each line bundle is calculated as above.
As a preferred scheme of the calibration method of the laser radar external reference without the calibration object, the method comprises the following steps: the points for which curvature needs to be calculated are the tenth to the n-10 th points on each beam, n being the total number of points on the beam.
As a preferred scheme of the calibration method of the laser radar external reference without the calibration object, the method comprises the following steps: the radius of the local point cloud map is 50-200 m.
As a preferred scheme of the calibration method of the laser radar external reference without the calibration object, the method comprises the following steps: the positioning device comprises a GPS positioning device and an inertial sensor.
The invention has the beneficial effects that:
(1) the method and the device directly obtain the pose, namely the external reference, of the laser radar to be calibrated in the map coordinate system by fusing the laser radar to be calibrated and the point cloud map established in advance, and have high calibration precision and high efficiency.
(2) The method comprises the steps of obtaining the position of the laser radar to be calibrated through the positioner, loading a corresponding point cloud map by taking the position as a center, matching point cloud data scanned by the laser radar to be calibrated with the map to obtain external reference information of the laser radar to be calibrated, and calibrating at any time without any calibration object and any requirement on a calibrated environment by only placing the laser radar to be calibrated in a point cloud map area.
(3) According to the method, a pre-established point cloud map is utilized, and after the point cloud map is subjected to image building steps such as dynamic object filtering, loop detection and the like, the image building precision of the point cloud map is very high, so that the calibration precision of the laser radar is higher.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings needed to be used in the description of the embodiments will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without inventive exercise.
FIG. 1 is a schematic flow chart of a calibration method for external reference of a laser radar without a calibration object according to the present invention;
FIG. 2 is a schematic flow chart of extracting line features and surface features of original point cloud data in the calibration method of the external reference of the laser radar without the calibration object provided by the invention;
FIG. 3 is a schematic flow chart illustrating the optimization of the initial external parameter of the lidar to be calibrated by using the point cloud data frame after the point cloud data frame is utilized in the calibration method for the external parameter of the lidar without a calibration object provided by the invention;
fig. 4 is a schematic diagram of a branch-and-bound method search algorithm according to an embodiment.
Detailed Description
In order that the present invention may be more readily and clearly understood, a more particular description of the invention briefly described above will be rendered by reference to specific embodiments that are illustrated in the appended drawings.
The embodiment provides a calibration object-free laser radar external reference calibration method, which comprises the following steps of S101-S105, and the specific steps are as follows:
step S101: and acquiring a pre-established point cloud map, and enabling the laser radar to be calibrated to be located in the range of the point cloud map.
Specifically, a pre-established point cloud map is obtained, which is formed after the map building steps of dynamic object filtering, loop detection and the like are performed, and the map is high in self precision and stored in the form of line features and surface features. And then, placing the laser radar to be calibrated in the area covered by the point cloud map, and keeping the laser radar to be calibrated still in the whole external reference calibration process.
Step S102: scanning a laser radar to be calibrated to obtain a frame of original point cloud data, and then extracting line features and surface features of the original point cloud data.
Specifically, the method for extracting the line features and the surface features of the original point cloud data comprises the following steps:
step S102 a: and splitting the original point cloud data. The line beam models of the laser radars of different models and different line beams are different, and the line beam models corresponding to the used laser radars can be obtained by referring to the corresponding user manual. Taking the mechanical lidar as an example, the mechanical lidar operates in a rotating scanning mode, a light source scans a circle, and point cloud data of the circle is a point of a beam. The following description will be made of the operation of the sub-beam processing, taking a velodyne 16 line lidar as an example:
tables 1 to 1: corresponding relation table of number of beam of velodyne 16 line laser radar and vertical angle
| Laser ID | Vertical Angle |
| 0 | -15° |
| 1 | 1° |
| 2 | -13° |
| 3 | 3° |
| 4 | -11° |
| 5 | 5° |
| 6 | -9° |
| 7 | 7° |
| 8 | -7° |
| 9 | 9° |
| 10 | -5° |
| 11 | 11° |
| 12 | -3° |
| 13 | 13° |
| 14 | -1° |
| 15 | 15° |
Table 1-1 shows the corresponding relation between the number of the beam of the velodyne 16 line laser radar and the vertical angle. Assuming that point p (x, y, z) is a point scanned by the lidar, its vertical angle may be calculated as:
。
the wire harness id corresponding to a certain vertical angle can be calculated according to the corresponding relationship between the wire harness id and the vertical angle shown in the above table 1-1:
therefore, the corresponding wire harness number of each three-dimensional laser point is calculated, and the branch wire harness processing of the original point cloud data is realized.
Step S102 b: the curvature of the point on each line bundle is calculated. It should be noted that, according to the arrangement order of the points on the harnesses, the tenth point to the tenth last point on each harness are calculated. The curvature of the points is calculated because the first ten points and the last ten points on each line bundle are not stable enough, and when the current point is calculated, the curvature of the first ten points and the curvature of the last ten points are not calculated because the current point needs to be operated with the first ten points and the last ten points. The curvature calculation of the point on the wire harness specifically comprises the following steps:
step S102 b-1: calculating the curvature of the ith point on the kth beam according to a first formula:
wherein, in the step (A),
data representing the ith point of the point cloud on the kth line, i.e., three-dimensional coordinate data, j =1,2,3, …, 10.
Step S102 b-2: calculated according to the formula I
,
,…, 
The ten curvatures are then sorted, with the median taken as the curvature of the ith point on the kth line.
The curvature of the point on each line bundle is then calculated as above.
Step S102 c: and judging whether the included angle between the current point and the next point in the wire harness is larger than 1 degree or not, and if so, rejecting the current point, ten continuous points in front of the current point and ten continuous points behind the current point.
Step S102 d: the remaining points on each bundle are equally divided into segments. The purpose of the segmentation is to evenly distribute the subsequently selected line and surface features. In this embodiment, the remaining points on each wire harness are equally divided into six segments.
Step S102 e: screening line characteristics of points in each section according to the sequence of curvature from large to small, specifically comprising the following steps: judging in proper order earlier whether current point is for sheltering from the point, if, then with current point, ten continuous points that lie in front of current point and ten continuous points that lie in behind the current point all are bad points, then judge whether the camber of current point is greater than preset threshold value in proper order, if the camber of current point is greater than preset threshold value and current point is not bad point, then think current point and be the line characteristic, and after the current point is thought to be the line characteristic, with current point, ten continuous points that lie in front of current point and ten continuous points that lie in behind the current point all are bad points. And repeating the steps until the line characteristics on all the wire harnesses are screened.
Step S102 f: screening surface characteristics of points in each section according to the sequence of curvature from small to large, and specifically comprising the following steps: judging in proper order earlier whether current point is parallel with the laser line bundle, if parallel, then with the current point, ten continuous points that lie in front of the current point and ten continuous points that lie in the back of the current point all are bad points, then judge whether the camber of current point is less than preset threshold value in proper order, if the camber of current point is less than preset threshold value and current point is not bad point, then think the current point is the face characteristic, and after the current point is regarded as the face characteristic, with the current point, ten continuous points that lie in front of the current point and ten continuous points that lie in the back of the current point all are bad points. And repeating the steps until the surface characteristics on all the wire harnesses are screened.
Step S103: and positioning the laser radar to be calibrated through a positioning device to obtain a positioning point of the laser radar to be calibrated, and then loading a local point cloud map with the positioning point as a circle center.
Specifically, a positioning device is installed to position the position of the laser radar to be calibrated to obtain a two-dimensional or three-dimensional coordinate point, and then a local point cloud map with the coordinate point as the center of a circle is loaded. Wherein, the radius of the local point cloud map is 50-200 m.
In this embodiment, the laser radar to be calibrated is installed at the vehicle end, and the positioning device may be a GPS positioning device or a vehicle-end inertial sensor. And roughly positioning the laser radar to be calibrated through a GPS positioning device or a vehicle-end inertial sensor. It can be understood that other positioning devices can be adopted, and only the position of the laser radar to be calibrated can be positioned to obtain the positioning point. In addition, the laser radar to be calibrated can also be fixedly arranged at the road end and used for vehicle-road cooperation.
Step S104: and matching the line features and the surface features extracted from the original point cloud data with the line features and the surface features of the local point cloud map to obtain corresponding yaw angles and displacements, and taking the corresponding yaw angles and the corresponding displacements as initial external parameters of the laser radar to be calibrated.
Specifically, the method is a fast search algorithm, and the graphical description of the algorithm is shown in fig. 4, which can be understood as follows:
step S104 a: first, the head node is divided into a plurality of sub-nodes according to possible values of yaw angle and displacement, and for convenience of description in this embodiment, the head node is divided into two sub-nodes, and then the nodes are sorted in order of scores from large to small. In fig. 4, a score of 2 is greater than a score of 1.
It should be noted that a node is a data structure that includes displacement, rotation, number of levels, and score of the node. The value of the displacement is all possible combinations searched currently, including: t1= (0,0, 0), (0, search _ step, 0), (search _ step, 0, 0), (search _ step, search _ step, 0), (0,0, search _ step), (0, search _ step, search _ step), (search _ step, 0, search _ step), (search _ step, search _ step, and search _ step). search step represents the step size of one search.
The value range of the yaw angle is
Step size is
So taking the value of
The number of layers is the number of layers of the tree, and from bottom to top, the number of layers where the leaf nodes are located is the 0 th layer.
And (3) calculating a score: for the point cloud of the current frame, calculating the matching condition of each point and the point cloud lower, namely, the point needs to be converted into a map coordinate system from a laser radar coordinate system, namely, the displacement and the yaw angle need to be used for conversion. Taking one displacement and one yaw angle at a time, there will be 8 x 180 value cases, and the score of each combination, i.e. the probability that each point falls into the grid point of the map, is distributed, and if it falls, the current score is increased by one. The map grid point is a small grid formed by rasterizing the point cloud map, and can be understood as a small grid formed by dividing one map into several parts.
Step S104 b: node number 2 is then divided up to leaf nodes 5 and 6. According to the above rule, a score of 5 is greater than a score of 5. The current maximum score is a score of 5.
Step S104 c: returning to the previous layer, comparing the score of the other sub-node 4 of the 2 with the maximum score, if the score of 4 is greater than the maximum score, further dividing 4 until the leaf nodes, comparing the maximum score,
step S104 d: and repeating the steps until the maximum score is found, namely considering that the yaw angle and the displacement corresponding to the maximum score are initial external parameters of the laser radar to be calibrated.
It is understood that the branch-and-bound method describes a matching relationship between points and a map, i.e. points in the point cloud data are transformed from the lidar coordinate system to the point cloud map coordinate system, i.e. the displacement and yaw angle are used for transformation. The points are line features and surface features extracted from the point cloud data.
Step S105: and after the initial external parameters of the laser radar to be calibrated are obtained, optimizing the initial external parameters by using the subsequent point cloud data frame. The method specifically comprises the following steps:
step S105 a: and loading a local point cloud map by taking the initial external reference as a center.
Specifically, in the initial external reference obtained in step S104, the displacement is a 3 × 1 vector, which can be understood as a three-dimensional coordinate, so that the local point cloud map is loaded with the pose of the initial external reference. Wherein, the radius of the local point cloud map is also 50-200 m.
Step S105 b: and searching matching points and matching numbers corresponding to line features and surface features extracted from the original point cloud data in the local point cloud map, then projecting the point cloud points to the local point cloud map, and adjusting the initial external parameters to obtain optimized external parameters.
Specifically, the projection is to transform the three-dimensional point cloud to point cloud coordinates using calibrated external parametersLet us assumeTIs a calibration parameter that is a function of,Pis a three-dimensional point cloud of points,Pis the corresponding map point, thenT p =POne such process is projection. Optimizing the external parameters by using the projection relation can adjust the corresponding external parameters. The adjustment here includes displacement and three angles of yaw, pitch and roll.
It should be noted that the external reference is composed of translation and rotation, where translation, i.e. displacement, is a vector of 3 x 1, and there are many representations of rotation. In the present embodiment, the euler angle is adopted, and includes roll, pitch, and yaw.
The specific optimization process can be described as follows:
suppose thatPA 1 is the extracted line feature point,T x _MP_Lis the currently estimated external parameter, p_MPis a map point, thenp_MP=T x _MP_L·P1, the corresponding objective function can be written:
wherein, in the step (A),
. Minimizing this objective function, estimatingT x _MP_LThe external reference is made.
Scanning a laser radar to be calibrated to obtain a new frame of original point cloud data, and extracting line features and surface features from the new frame of original point cloud data;
step S105 c: and taking the optimized external parameters as a center, and loading a local point cloud map.
Step S105 d: and searching matching points and matching data corresponding to line features and surface features extracted from a new frame of original point cloud data in the local point cloud map, projecting the point cloud points to the point cloud map, and adjusting the optimized external parameters again to obtain the optimized external parameters.
Step S105 e: and repeating the steps for a plurality of times until the error of the result obtained by the optimization calculation of the point cloud data of the front frame and the back frame is smaller than the set threshold value, and obtaining the external parameters of the laser radar to be calibrated.
Specifically, the step S105a to the step S105d are repeated three times, and it is considered that the extrinsic optimization is completed once. And repeating the above external parameter optimization for a plurality of times until the error of the result obtained by the optimization calculation of the front and back two frames of point cloud data is less than the set threshold value, and determining that the optimization stage is converged. And at the moment, the external parameter obtained by optimizing the point cloud data of the next frame is used as the final external parameter result of the laser radar to be calibrated.
According to the specific process steps of the calibration-object-free laser radar external reference calibration method, the laser radar to be calibrated is fused with the point cloud map established in advance, so that the pose of the laser radar to be calibrated in a map coordinate system is directly obtained, and calibration objects are not needed. The method has more advantages than the prior technical scheme in terms of calibration precision and calibration efficiency. In addition, if the coordinate system needs to be converted into the coordinate system of the self-vehicle, only one pose transformation relation is needed to be calculated.
After the laser radar is calibrated by adopting the calibration object-free laser radar external reference calibration method provided by the invention, scanned point cloud data and a point cloud map can be perfectly fused. The adopted laser radar is a hesai 64-line laser radar, the whole calibration time takes 72 seconds, and the calibration efficiency is very high.
In addition to the above embodiments, the present invention may have other embodiments; all technical solutions formed by adopting equivalent substitutions or equivalent transformations fall within the protection scope of the claims of the present invention.
Claims (7)
1. A calibration object-free laser radar external reference calibration method is characterized by comprising the following steps: comprises the steps of (a) preparing a mixture of a plurality of raw materials,
acquiring a pre-established point cloud map, and enabling the laser radar to be calibrated to be located in the range of the point cloud map;
scanning a laser radar to be calibrated to obtain a frame of original point cloud data, and then extracting line features and surface features of the original point cloud data;
positioning the laser radar to be calibrated through a positioning device to obtain a positioning point of the laser radar to be calibrated, and then loading a local point cloud map with the positioning point as a circle center;
and matching the line features and the surface features extracted from the original point cloud data with the line features and the surface features of the local point cloud map to obtain corresponding yaw angles and displacements, and taking the corresponding yaw angles and the corresponding displacements as initial external parameters of the laser radar to be calibrated.
2. The calibration method for external reference of lidar without calibration object according to claim 1, wherein: after the line features and the surface features extracted from the original point cloud data are matched with the line features and the surface features of the local point cloud map to obtain corresponding yaw angle and displacement which are used as initial external parameters of the laser radar to be calibrated, the method also comprises the following steps of,
loading a local point cloud map by taking the initial external parameters as a center;
searching matching points and matching data corresponding to line features and surface features extracted from original point cloud data in a local point cloud map, then projecting point cloud points onto the point cloud map, and adjusting initial external parameters to obtain optimized external parameters;
scanning a laser radar to be calibrated to obtain a new frame of original point cloud data, and extracting line features and surface features from the new frame of original point cloud data;
loading a local point cloud map by taking the optimized external parameters as a center;
searching matching points and matching data corresponding to line features and surface features extracted from a new frame of original point cloud data in a local point cloud map, projecting point cloud points to the point cloud map, and adjusting the optimized external parameters again to obtain the optimized external parameters;
and repeating the steps for a plurality of times until the error of the result obtained by the optimization calculation of the point cloud data of the front frame and the back frame is smaller than the set threshold value, and obtaining the external parameters of the laser radar to be calibrated.
3. The calibration method for external reference of lidar without calibration object according to claim 1, wherein: the extracting line features and surface features of the raw point cloud data includes,
splitting the original point cloud data;
calculating the curvature of a point on each wire harness;
judging whether an included angle between a current point and a next point in the wire harness is larger than 1 degree, if so, rejecting the current point, ten continuous points in front of the current point and ten continuous points behind the current point;
dividing the rest points on each wire harness into a plurality of sections;
screening line characteristics of points in each section according to the sequence of curvature from large to small, sequentially judging whether a current point is a shielding point or not, if so, setting the current point, ten continuous points in front of the current point and ten continuous points behind the current point as dead points, then sequentially judging whether the curvature of the current point is larger than a preset threshold value or not, if the curvature of the current point is larger than the preset threshold value and the current point is not the dead point, determining that the current point is the line characteristics, and after the current point is determined as the line characteristics, setting the current point, ten continuous points in front of the current point and ten continuous points behind the current point as the dead points;
screening surface characteristics of points in each section according to the sequence of curvature from small to large, sequentially judging whether a current point is parallel to a laser beam, if so, setting the current point, ten continuous points in front of the current point and ten continuous points behind the current point as dead points, then sequentially judging whether the curvature of the current point is smaller than a preset threshold value, if the curvature of the current point is smaller than the preset threshold value and the current point is not the dead point, determining the current point as the surface characteristics, and after the current point is determined as the surface characteristics, setting the current point, ten continuous points in front of the current point and ten continuous points behind the current point as the dead points.
4. The calibration method for external reference of lidar without a calibration object according to claim 3, wherein: the calculating of the curvature of the point on each line bundle includes,
calculating the curvature of the ith point on the kth beam according to a first formula:
，
wherein the content of the first and second substances,
i point, j =1,2,3, …,10, representing the point cloud on the k-th line bundle;
calculated according to the formula I
,
,…, 
Then, sorting the ten curvatures, and taking a median value as the curvature of the ith point on the kth line beam;
the curvature of the point on each line bundle is calculated as above.
5. The calibration method for external reference of lidar without calibration object according to claim 4, wherein: the points for which curvature needs to be calculated are the tenth to the n-10 th points on each beam, n being the total number of points on the beam.
6. The external reference calibration method for lidar without a calibration object according to claim 1 or 2, wherein: the radius of the local point cloud map is 50-200 m.
7. The calibration method for external reference of lidar without calibration object according to claim 1, wherein: the positioning device comprises a GPS positioning device and an inertial sensor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111538615.6A CN113933820B (en) | 2021-12-16 | 2021-12-16 | Laser radar external reference calibration method without calibration object |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111538615.6A CN113933820B (en) | 2021-12-16 | 2021-12-16 | Laser radar external reference calibration method without calibration object |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113933820A true CN113933820A (en) | 2022-01-14 |
| CN113933820B CN113933820B (en) | 2022-03-08 |
Family
ID=79289058
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111538615.6A Active CN113933820B (en) | 2021-12-16 | 2021-12-16 | Laser radar external reference calibration method without calibration object |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113933820B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115236645A (en) * | 2022-09-23 | 2022-10-25 | 北京小马易行科技有限公司 | Laser radar attitude determination method and attitude determination device |
| CN115236644A (en) * | 2022-07-26 | 2022-10-25 | 广州文远知行科技有限公司 | Laser radar external parameter calibration method, device, equipment and storage medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109901139A (en) * | 2018-12-28 | 2019-06-18 | 文远知行有限公司 | Laser radar scaling method, device, equipment and storage medium |
| US20200142027A1 (en) * | 2018-11-02 | 2020-05-07 | Hyundai Motor Company | Apparatus and method for calibrating a zero point of a radar for a vehicle |
| CN112014857A (en) * | 2020-08-31 | 2020-12-01 | 上海宇航系统工程研究所 | Three-dimensional laser radar positioning and navigation method for intelligent inspection and inspection robot |
| CN112462350A (en) * | 2020-12-10 | 2021-03-09 | 苏州一径科技有限公司 | Radar calibration method and device, electronic equipment and storage medium |
| CN112763993A (en) * | 2020-12-22 | 2021-05-07 | 北京百度网讯科技有限公司 | Method and device for calibrating radar parameters, electronic equipment and storage medium |
| CN113687336A (en) * | 2021-09-09 | 2021-11-23 | 北京斯年智驾科技有限公司 | Radar calibration method and device, electronic equipment and medium |
-
2021
- 2021-12-16 CN CN202111538615.6A patent/CN113933820B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20200142027A1 (en) * | 2018-11-02 | 2020-05-07 | Hyundai Motor Company | Apparatus and method for calibrating a zero point of a radar for a vehicle |
| CN109901139A (en) * | 2018-12-28 | 2019-06-18 | 文远知行有限公司 | Laser radar scaling method, device, equipment and storage medium |
| CN112014857A (en) * | 2020-08-31 | 2020-12-01 | 上海宇航系统工程研究所 | Three-dimensional laser radar positioning and navigation method for intelligent inspection and inspection robot |
| CN112462350A (en) * | 2020-12-10 | 2021-03-09 | 苏州一径科技有限公司 | Radar calibration method and device, electronic equipment and storage medium |
| CN112763993A (en) * | 2020-12-22 | 2021-05-07 | 北京百度网讯科技有限公司 | Method and device for calibrating radar parameters, electronic equipment and storage medium |
| CN113687336A (en) * | 2021-09-09 | 2021-11-23 | 北京斯年智驾科技有限公司 | Radar calibration method and device, electronic equipment and medium |
Non-Patent Citations (1)
| Title |
|---|
| 孔志飞: "应急救援车辆的高精度、高稳定性定位技术研究", 《中国优秀博硕士学位论文全文数据库(硕士)工程科技Ⅱ辑》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115236644A (en) * | 2022-07-26 | 2022-10-25 | 广州文远知行科技有限公司 | Laser radar external parameter calibration method, device, equipment and storage medium |
| CN115236645A (en) * | 2022-09-23 | 2022-10-25 | 北京小马易行科技有限公司 | Laser radar attitude determination method and attitude determination device |
| CN115236645B (en) * | 2022-09-23 | 2023-01-24 | 北京小马易行科技有限公司 | Laser radar attitude determination method and attitude determination device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113933820B (en) | 2022-03-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110298298B (en) | Target detection and target detection network training method, device and equipment | |
| CN113933820B (en) | Laser radar external reference calibration method without calibration object | |
| CN108921947B (en) | Method, device, equipment, storage medium and acquisition entity for generating electronic map | |
| CN111882612B (en) | Vehicle multi-scale positioning method based on three-dimensional laser detection lane line | |
| CN112347550B (en) | Coupling type indoor three-dimensional semantic graph building and modeling method | |
| Kang et al. | Automatic targetless camera–lidar calibration by aligning edge with gaussian mixture model | |
| CN112084869B (en) | Compact quadrilateral representation-based building target detection method | |
| CN105160702A (en) | Stereoscopic image dense matching method and system based on LiDAR point cloud assistance | |
| CN110223380B (en) | Scene modeling method, system and device fusing aerial photography and ground visual angle images | |
| CN111257909B (en) | Multi-2D laser radar fusion mapping and positioning method and system | |
| CN114972968A (en) | Tray identification and pose estimation method based on multiple neural networks | |
| CN112346463A (en) | Unmanned vehicle path planning method based on speed sampling | |
| CN114155265A (en) | Three-dimensional laser radar road point cloud segmentation method based on YOLACT | |
| CN112669458A (en) | Method, device and program carrier for ground filtering based on laser point cloud | |
| CN113589306B (en) | Positioning method, positioning device, electronic equipment and storage medium | |
| CN112070800A (en) | Intelligent vehicle positioning method and system based on three-dimensional point cloud polarization map representation | |
| KR102490521B1 (en) | Automatic calibration through vector matching of the LiDAR coordinate system and the camera coordinate system | |
| CN117590362A (en) | Multi-laser radar external parameter calibration method, device and equipment | |
| CN117392237A (en) | Robust laser radar-camera self-calibration method | |
| CN114782556B (en) | Camera and laser radar registration method and system and storage medium | |
| WO2022021209A1 (en) | Electronic map generation method and apparatus, computer device, and storage medium | |
| CN111504317B (en) | Indoor positioning method based on single-line laser radar | |
| CN112433193A (en) | Multi-sensor-based mold position positioning method and system | |
| Huang et al. | Ground filtering algorithm for mobile LIDAR using order and neighborhood point information | |
| US20210405197A1 (en) | GLOBAL LOCALIZATION APPARATUS AND METHOD IN DYNAMIC ENVIRONMENTS USING 3D LiDAR SCANNER |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20230329 Address after: 2nd floor, building A3, Hongfeng science and Technology Park, Nanjing Economic and Technological Development Zone, Nanjing, Jiangsu Province 210033 Patentee after: CIIC Technology Co.,Ltd. Patentee after: Tianyi Transportation Technology Co.,Ltd. Address before: 210033 floor 2, building A3, Hongfeng Science Park, Nanjing Economic and Technological Development Zone, Jiangsu Province Patentee before: CIIC Technology Co.,Ltd. |