CN115079143A - Multi-radar external parameter rapid calibration method and device for double-axle steering mine card - Google Patents
Multi-radar external parameter rapid calibration method and device for double-axle steering mine card Download PDFInfo
- Publication number
- CN115079143A CN115079143A CN202210670457.8A CN202210670457A CN115079143A CN 115079143 A CN115079143 A CN 115079143A CN 202210670457 A CN202210670457 A CN 202210670457A CN 115079143 A CN115079143 A CN 115079143A
- Authority
- CN
- China
- Prior art keywords
- point cloud
- cloud data
- laser radar
- coordinate system
- laser
- 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
- 238000000034 method Methods 0.000 title claims abstract description 51
- 238000013507 mapping Methods 0.000 claims abstract description 15
- 238000009434 installation Methods 0.000 claims description 12
- 230000011218 segmentation Effects 0.000 claims description 6
- 238000005457 optimization Methods 0.000 claims description 5
- 238000003860 storage Methods 0.000 claims description 5
- 238000011065 in-situ storage Methods 0.000 claims description 3
- 230000036544 posture Effects 0.000 description 6
- 230000009466 transformation Effects 0.000 description 5
- 238000011156 evaluation Methods 0.000 description 3
- 230000008447 perception Effects 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 238000012549 training Methods 0.000 description 3
- 238000013519 translation Methods 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 238000012937 correction Methods 0.000 description 2
- 238000000691 measurement method Methods 0.000 description 2
- 238000005065 mining Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000002310 reflectometry Methods 0.000 description 2
- 230000000007 visual effect Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 239000003086 colorant Substances 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000009977 dual effect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000012805 post-processing Methods 0.000 description 1
- 230000001131 transforming effect Effects 0.000 description 1
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
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Optical Radar Systems And Details Thereof (AREA)
Abstract
The invention belongs to the field of sensor and laser radar mapping and positioning, and relates to a multi-radar external reference rapid calibration method and device for a double-axle steering mine card. The invention discloses a multi-radar external reference calibration method for a double-axle steering mine card, which comprises the following steps: collecting point cloud data of a plurality of laser radars; dividing the point cloud data into ground point cloud data and non-ground point cloud data, and adding semantic category information to the non-ground point cloud data; performing plane parameter fitting on the ground point cloud data to construct a plurality of freedom degree constraint relations among a plurality of laser radars; building a cloud map; carrying out point cloud map registration on the point cloud data to obtain a 6-degree-of-freedom pose relationship among the plurality of laser radars; and (5) solving external parameters of each laser radar relative to the mine truck body coordinate system. The calibration method can quickly calibrate a plurality of laser radars on the double-axle steering mine car without the cab, ensure the coordinate uniformity of the fused point cloud and provide accurate position information for subsequent sensing tasks.
Description
Technical Field
The invention belongs to the field of sensor and laser radar mapping and positioning, and relates to a multi-radar external reference rapid calibration method and device for a double-axle steering mine card, in particular to a multi-radar external reference rapid calibration method, device and storage medium suitable for a cab-free double-axle steering mine card.
Background
The mining truck is huge in size, and a sensing system suitable for the mining truck needs to consider the problems that the view field in the overall range of a truck body is as large as possible and the dead angle is as small as possible, so that the safety in the automatic driving process is guaranteed. The laser radar has the characteristics of high precision and all-weather work, and a plurality of laser radars are usually arranged around a vehicle body to realize the full-coverage perception of the surrounding environment in consideration of the scanning range of the laser radar. However, due to different installation positions of the laser radar devices, the data acquired by the laser radars have certain offset. Therefore, the accurate calibration of the external installation parameters (external parameters) of each lidar is an essential step for simultaneously using a plurality of lidars.
At present, the external reference calibration of the laser radar is mainly carried out by adopting a traditional manual calibration method or a calibration object measurement method. The method is simple in principle, but has high requirements on the operation precision of an operator and the precision of a measuring instrument, so that the popularization is not high, the calibration precision is unstable, and the calibration precision can change along with the level of the operator. The calibration object measurement method determines external parameters of the laser radar device through matching of a specific calibration object, and the method has certain disadvantages: because the reflectivity of the laser radar is different for objects with different colors and materials, the search of the calibration object is difficult, a large amount of manpower is needed in the process of moving the calibration object, and the calibration precision depends on parameters between other sensors and the calibration object, so the calibration method has low calibration efficiency and low precision.
Chinese patent CN114047487A, entitled "method and apparatus for calibrating external reference of radar and vehicle body", provides a method for calibrating external reference of vehicle-mounted laser radar, which converts the coordinates of a reference point under an auxiliary radar coordinate system into the coordinates under a vehicle-mounted radar coordinate system through an auxiliary radar and a calibration plate independent of the vehicle body, and calibrates external parameters from the vehicle-mounted radar coordinate system to the vehicle body coordinate system according to the position coordinates of the reference point under the vehicle body coordinate system. The method only reduces the requirement on the calibration site, and the operation simplicity is not improved when the number of the radars to be calibrated is large.
Chinese patent CN113702931A, entitled "method, apparatus and storage medium for calibrating external parameters of vehicle-mounted radar", provides a method for automatically calibrating external parameters between radars on line in real time by using characteristics of multiple radars and using first object set information and second object set information. The method requires sufficient repeated views among multiple lidar and can calibrate fewer radars simultaneously.
For the double-axle steering mine card without the cab, the front driving and the rear driving can be switched during actual operation, and laser radar sensing equipment with the same capacity is needed before and after the comprehensive sensing visual field is obtained. Obviously, the laser radars cannot have repeated perception fields at the same time, if a pairwise calibration method is adopted, the efficiency is not high, errors are accumulated due to multiple times of calibration, and ideal precision cannot be obtained.
Disclosure of Invention
The double-axle steering mine card without the cab needs a plurality of laser radars to sense the surrounding environment, and the existing calibration method is difficult to quickly and accurately obtain the installation parameters among the plurality of radars. Therefore, the invention provides a multi-laser-radar external-parameter quick calibration method, which can quickly calibrate a plurality of laser radars on a cab-free double-axle steering mine car, ensure the coordinate uniformity of fusion point cloud and provide accurate position information for subsequent sensing tasks.
In order to achieve the purpose, the invention provides a multi-radar external reference quick calibration method for a double-axle steering mine card, which comprises the following steps:
s1: when the mine card is in a calibration scene, collecting point cloud data of a plurality of laser radars on a mine card vehicle body, measuring the ground-off installation height of each laser radar, and taking the laser radar positioned in the middle of the mine card vehicle body as a main laser radar;
s2: dividing the point cloud data into ground point cloud data and non-ground point cloud data based on the ground-above-ground installation height of each laser radar, and adding semantic category information to the non-ground point cloud data;
s3: performing plane parameter fitting on the ground point cloud data to obtain ground point cloud plane parameters, and constructing multiple degree-of-freedom constraint relations among the laser radars based on the ground point cloud plane parameters;
s4: constructing a point cloud map based on the non-ground point cloud data with semantic category information of the main laser radar;
s5: based on the constructed multiple degree-of-freedom constraint relations among the laser radars and the constructed point cloud map, performing point cloud map registration on the point cloud data of the laser radars acquired in the step S1 to obtain the positions of the coordinate systems of the laser radars relative to the coordinate system of the point cloud map, and further obtain the position relation of 6 degrees-of-freedom of the coordinate systems of the laser radars;
s6: and calculating the origin position of the mine truck body coordinate system based on the 6-degree-of-freedom pose relation of the coordinate system of each laser radar, and solving the external reference of each laser radar relative to the mine truck body coordinate system.
Further, in step S1, point cloud data of the laser radar in the front direction and the back direction of the mine truck body are respectively collected by turning the mine truck in situ.
Further, the step S2 specifically includes:
according to the ground-to-ground installation height of each laser radar, point cloud data of each laser radar is divided into ground point cloud data and non-ground point cloud data, and the non-ground point cloud data is subjected to semantic division, so that each point in the non-ground point cloud data has semantic category information except space geometric information, and a plurality of analog point clouds are generated.
Further, the step S3 specifically includes:
and performing plane parameter fitting on the ground point cloud data of each laser radar by using a RANSAC method to obtain the ground point cloud plane parameters of each laser radar, and constructing a plurality of freedom degree constraint relations of each laser radar in a space position.
Further, the step S4 specifically includes:
and extracting line and surface features by using non-ground point cloud data of the main laser radar, adding semantic category information constraint in a feature matching stage, calculating the pose relationship of adjacent frame point clouds, and overlapping to construct a point cloud map.
Further, the step S5 specifically includes:
based on a plurality of degree-of-freedom constraint relations among the constructed laser radars, performing point cloud map matching on point cloud data of the laser radars on the constructed point cloud map, continuously iterating to obtain positioning parameters of the laser radars relative to the point cloud map, and finally obtaining x, y, z, roll, pitch and yaw information of a coordinate system of the laser radars relative to a coordinate system of the point cloud map, so as to obtain a 6-degree-of-freedom pose relation of the coordinate system of the laser radars; the offset of the coordinate system of each laser radar relative to the coordinate system of the point cloud map in the X-axis, Y-axis and Z-axis directions is represented by X, Y and Z, the rotation angle of the coordinate system of each other laser radar relative to the coordinate system of the point cloud map around the X-axis is represented by roll, the rotation angle of the coordinate system of each laser radar relative to the coordinate system of the point cloud map around the Y-axis is represented by pitch, and the rotation angle of the coordinate system of each laser radar relative to the coordinate system of the point cloud map around the Z-axis is represented by yaw.
Further, the step S6 specifically includes:
and calculating the origin position of the ore removal truck body coordinate system based on the 6-degree-of-freedom position and posture relation of the coordinate system of each laser radar, and converting the 6-degree-of-freedom position and posture relation of the coordinate system of each laser radar into parameters under the ore removal truck body coordinate system to obtain the external parameters of each laser radar relative to the ore removal truck body coordinate system.
The invention also provides a multi-radar external reference quick calibration device for the double-axle steering mine card, which comprises the following components:
the acquisition module is used for acquiring point cloud data of a plurality of laser radars on the mine truck body;
the point cloud segmentation module is used for segmenting the point cloud data acquired by the acquisition module into ground point cloud data and non-ground point cloud data; meanwhile, the method is used for adding semantic category information to the non-ground point cloud data;
the plane fitting module is used for performing plane parameter fitting on the ground point cloud data of each laser radar to obtain ground point cloud plane parameters, constructing multiple freedom degree constraint relations among the laser radars and transmitting the multiple freedom degree constraint relations to the repositioning module;
the mapping module is used for constructing a point cloud map based on non-ground point cloud data with semantic category information of a main laser radar, wherein the main laser radar is a laser radar positioned in the middle of the mine truck body;
the repositioning module is used for receiving a plurality of freedom degree constraints among the laser radars constructed by the plane fitting module, and performing point cloud map matching on the point cloud data of the laser radars from the acquisition module to obtain a 6-freedom degree pose relation of a coordinate system of each laser radar;
and the external reference optimization module is used for calculating the origin position of the ore removal truck body coordinate system based on the 6-degree-of-freedom pose relationship of the coordinate system of each laser radar and solving the external reference of each laser radar relative to the ore removal truck body coordinate system.
The present invention also provides an electronic device, comprising:
a memory, a processor and a program stored on the memory and executable on the processor, the processor executing the program to implement the above method.
The present invention also provides a computer-readable storage medium having a program stored thereon, the program being executed by a processor to implement the above-mentioned method.
The invention has the beneficial effects that:
1) the invention can realize multi-laser radar external reference calibration of indoor and outdoor environment only depending on the self characteristics of the laser radar without other calibration equipment and mutual cooperation with other sensors;
2) the method is combined with the Slam mapping principle, the problem that a plurality of laser radars have no repeated visual field can be well solved, the plurality of laser radars can be calibrated quickly and accurately at the same time, and the calibration efficiency is greatly improved;
3) according to the invention, ground point cloud data information is utilized to provide multiple freedom degree constraints for positioning, so that the positioning precision and the positioning speed are improved to a certain extent;
4) according to the invention, the external parameters of a plurality of laser radars are finally calibrated to the middle of the mine truck body, which is equivalent to that a virtual radar capable of sensing in a full view angle range is installed in the middle of the truck body, and the method has no forward and backward sensing switching step, and is more in line with the characteristic that double-axle steering has no obvious forward and backward difference.
Drawings
Fig. 1 is a flowchart of a multi-radar external reference calibration method for a double-axle bogie mine card according to embodiment 1 of the invention;
fig. 2 is a block diagram of a multi-radar external reference calibration apparatus of a dual axle bogie mine card according to embodiment 2 of the present invention.
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings and examples.
Example 1
Referring to fig. 1, the multi-radar external reference rapid calibration method for the double-axle bogie mine card in the embodiment includes the following steps:
step 1: 3 laser radars are respectively installed in the front direction and the rear direction of the cab-free double-axle bogie mine card, and the ground clearance installation height of 6 laser radars is measured.
In the embodiment, 2 laser radars installed at the middle position between the front direction and the rear direction of the mine truck body are used as main laser radars, and finally point cloud data of 6 laser radars are converted to a point cloud data which takes the central position of the 2 main laser radars as an origin and the direction of the truck head as X Vehicle with wheels The right left direction of the axle and the vehicle body is Y Vehicle with wheels The shaft is arranged in the direction Z directly above the vehicle body Vehicle with a detachable front cover In the body coordinate system of the axle.
Step 2: and (3) placing the double-axle steering mine card in a proper calibration scene, and collecting point cloud data of 6 laser radars. Because the front and back sensing capability of the double-axle steering is the same, particularly, in order to ensure that the difference of the front and back environment is not large when data are collected, the point cloud data of the laser radar on the front and back of the mine truck body can be collected respectively in an in-situ steering mode, and the consistency of the data collected by the front and back laser radar is ensured.
And step 3: dividing point cloud data of the laser radar into ground point cloud data and non-ground point cloud data based on the ground-above-ground installation height of each laser radar; performing ball mapping on non-ground point cloud data by using a randnet + + deep learning frame, and mapping a 3D point cloud into a 2D depth image, wherein a mapping formula is as follows:
in the formula: u, v represent pixel coordinate system coordinates; x ', y ', z ' are coordinate values of 3D point cloud in the laser radar; r is the Euclidean distance from the point to the laser radar; w and h are the width and height of the depth image; f is the total view angle range of the laser radar in the vertical direction, f up Is the upward viewing angle range in the vertical direction.
All points in the point cloud data of the laser radar can be mapped to each pixel point in the depth image through the mapping formula (1). Because the width, the height w and the height h of the depth image cannot be set too large in order to improve the convolution speed, any pixel point on the depth image corresponds to a plurality of points in the original non-ground point cloud data. And 5 channels are arranged at each pixel position, and each channel value corresponds to the 3D coordinate (x ', y ', z '), the reflectivity and the distance of the 3D point respectively, so that a depth image of [5 x h x w ] is finally formed. And then carrying out range 53 network training on the mapped depth image, obtaining the semantic label of each pixel point according to the training result, endowing the same semantic label to a plurality of 3D points corresponding to each pixel point, and reconstructing the 2D pixel points into 3D point clouds according to the stored channel values. And finally, carrying out post-processing correction on the edge position of a blocked object (obstacle) in the 3D point cloud, namely determining the category of each 3D point according to the neighbor point category of each 3D point, and finally, dividing each point of non-ground point cloud data into different semantic category information to generate a plurality of analog point clouds, such as a pedestrian point cloud, a vehicle point cloud, a traffic sign point cloud and the like.
And 4, step 4: and performing plane parameter fitting on the ground point cloud data of each laser radar by using a RANSAC plane fitting method to obtain the ground point cloud parameters of each laser radar, and accordingly obtaining a coordinate transformation relation equation between each laser radar in a space position to construct a plurality of freedom degree constraint relations between each laser radar.
In this embodiment, the method for performing plane parameter fitting on the ground point cloud data of each laser radar by using the RANSAC plane fitting method specifically includes the following steps:
1) randomly selecting 3 points in the ground point cloud data in a centralized manner as initial points, and calculating a ground plane model equation of the ground point cloud:
Ax'+By'+Cz'+D＝0 (2)
where A, B, C, D represents the coefficients in the ground plane model.
2) Calculating Euclidean distance d from all points in ground point cloud data to the ground plane model i ＝|Ax′ i +By′ i +Cz′ i + D |, where x' i ,y′ i ,z′ i Indicating the coordinates of the ith point. Setting a distance threshold d t If the Euclidean distance d of a certain point is calculated i ≤d t If the point is not within the current ground plane model, the point is considered as a point within the ground plane model, otherwise the point is considered as a point outside the ground plane model, i.e. the point is not within the current ground plane model.
3) And repeating the step 2), recording the number of the points in the current ground plane model, selecting the ground plane model with the largest number of the points as the optimal ground plane model, and outputting the optimal model parameters as the final ground plane model parameter estimation value.
4) Constructing ground constraint:
firstly, solving a final ground plane model parameter estimation value of point cloud of each laser radar, and determining a constraint equation according to the model parameter estimation value as follows:
writing the ground plane model equation (2) as a matrix form, i.e.:
for any two coordinate systems (X, Y, Z) of the laser radar, the transformation parameters between the coordinate systems are assumed to be tx, ty, tz and theta respectively r ,θ y ,θ p Where tx is the translation in the X-axis direction, ty is the translation in the Y-axis direction, tz is the translation in the Z-axis direction, and θ r Is the difference between the roll angles, theta, of the two lidar y For the difference in yaw angle, theta, of two lidar z The pitch angle difference of the two laser radars.
According to the rigid body surface transformation characteristics, the corresponding relation between any two corresponding points (x '1, y' 1, z '1) and (x' 2, y '2, z' 2) in the two laser radar coordinate systems is
From the rectangular form (3) of the ground plane model equation:
substituting equation (4) into the rectangular form (3) of the ground plane model equation yields:
therefore, it is not only easy to use
In the formula (A) 1 B 1 C 1 D 1 ) And (A) 2 B 2 C 2 D 2 ) Known as tx, ty, tz, theta r ,θ y ,θ p Is an unknown quantity.
From equation (6) above, tx, ty, tz, θ can be determined r ,θ y ,θ p The constraint relation among the unknown quantities can determine the multi-degree-of-freedom constraint relation of two laser radar coordinate systems, and further can determine the multi-degree-of-freedom constraint relation of a plurality of laser radars.
And 5: the method comprises the steps of utilizing 2 pieces of non-ground point cloud data with semantic information of a main laser radar, taking a large curvature point as an angular point and a small curvature point as a plane point in the non-ground point cloud data to conduct line-surface feature extraction, conducting upper and lower frame point cloud registration according to the semantic information, namely when matching of corresponding feature points of adjacent frames is conducted, judging whether the corresponding feature points of the adjacent frames have the same semantic information or not to conduct matching, if matching of the semantic information fails, conducting semantic information matching on feature points of the frame adjacent to the feature points of the previous frame, which are failed to match, if the same semantic feature points are found, conducting matching successfully, calculating the pose relation of the adjacent frame point clouds, and building a point cloud map in an overlapping mode.
The curvature calculation formula of the midpoint of the non-ground point cloud data is as follows:
in the formula: rho i Is the curvature of the ith point, X i Is the coordinate (x ') of the ith point' i ,y' i ,z' i )，X j And m is the number of adjacent points and is a set value.
Step 6: and performing point cloud map registration on the real-time point cloud data of each laser radar in the constructed point cloud map based on the constructed multiple-freedom-degree constraint relation among the laser radars, and obtaining positioning parameters meeting the precision requirement after continuous iteration to obtain the 6-freedom-degree pose of each laser radar coordinate system. The accuracy evaluation can be performed using the following accuracy evaluation method:
suppose that the coordinate of any point A in the real-time point cloud after primary registration is (x' A ,y′ A ,z′ A ) And the coordinate of the corresponding feature point B found in the point cloud map is (x' B ,y′ B ,z′ B ) Euclidean distance d between AB AB Is composed of
Then calculating the distance between all corresponding points, and solving the Euclidean distance sum
And n is the total point cloud number of the laser radar as an evaluation index. And if d meets the set threshold, the precision requirement is met, otherwise, iterative registration is continuously carried out on the real-time point cloud after registration.
And 7: and calculating the origin position of the ore removal truck body coordinate system based on the position and posture relation of each laser radar coordinate system, converting the position and posture relation of each laser radar into parameters under the ore removal truck body coordinate system, and obtaining the external parameters of each laser radar relative to the ore removal truck body coordinate system. The specific process is as follows:
calculating the position and pose of 6 degrees of freedom among all the laser radars according to the position and pose (x) of two main laser radars in step 6 1 ,y 1 ,z 1 ,roll 1 ,pitch 1 ,yaw 1 ) And (x) 2 ,y 2 ,z 2 ,roll 2 ,pitch 2 ,yaw 2 ) Calculating the origin of the coordinate system of the vehicle body
And transforming the other positions and postures of the laser radars into the vehicle body coordinate system to obtain the external parameters of the laser radars relative to the vehicle body coordinate system, and finally successfully calibrating.
Example 2:
as shown in fig. 2, the multi-radar external reference rapid calibration device for the dual-axle bogie mine card of the embodiment includes an acquisition module, a segmentation module, a plane fitting module, a mapping module, a repositioning module and an external reference optimization module. Wherein:
and the acquisition module is used for acquiring point cloud data of a plurality of laser radars on the mine truck body. In this embodiment 2, 3 lidar are respectively installed in the front and back of the double-axle bogie mine card without a cab, and the acquisition module is used for subsequent off-line drawing construction and marking by acquiring off-line data of 6 lidar.
The point cloud segmentation module is used for segmenting the point cloud data acquired by the acquisition module into ground point cloud data and non-ground point cloud data; and meanwhile, the method is used for adding semantic information to the non-ground point cloud data. Specifically, the point cloud segmentation module segments the original point cloud data acquired by the acquisition module into ground point cloud data and non-ground point cloud data by using the ground clearance installation height information of 6 laser radars, and performs spherical coordinate projection, network training, remapping and semantic segmentation on the non-ground point cloud data by using a randent + + network structure. The non-ground point cloud data not only has space geometric information such as x, y, z and the like, but also has semantic information capable of expressing the specific meaning represented by each point, namely whether each point belongs to a pedestrian point cloud or a vehicle point cloud and the like can be obtained, and the method plays an important role in frame registration before and after a subsequent mapping stage and positioning in a repositioning stage.
And the plane fitting module is used for performing plane parameter fitting on the ground point cloud data of each laser radar to obtain ground point cloud plane parameters, constructing multiple freedom degree constraint relations among the plurality of laser radars and transmitting the multiple freedom degree constraint relations to the repositioning module. Specifically, the plane fitting module performs plane fitting on the ground point cloud by using a ransac method to obtain ground plane model parameters of the ground point cloud in each laser radar, and because coordinate transformation between each laser radar is rigid body transformation, under the condition of a known ground plane model equation, multiple degree-of-freedom constraint relations between each laser radar coordinate system can be determined according to different ground plane model parameters of the ground, so that a certain correction effect is achieved on subsequent positioning.
And the mapping module is used for constructing a point cloud map based on the non-ground point cloud data with the semantic information of the main laser radar. Specifically, the mapping module extracts angular points and plane points of the non-ground point cloud data of the main laser radar as feature points based on the leam mapping principle, adds semantic information as prior information during odometer calculation, and the matching of the feature points of the front frame and the back frame must have the same semantic information, so that the corresponding feature points can be quickly and accurately found out, and the mapping module plays an important role in reducing odometer errors.
And the repositioning module is used for receiving multiple degree-of-freedom constraints among the 6 laser radars constructed by the plane fitting module, and performing point cloud map registration on the point cloud data of the laser radars from the acquisition module to obtain 6 degree-of-freedom pose relations of 6 laser radar coordinate systems. Specifically, the repositioning module performs icp matching on a map according to instant point cloud data among all the laser radars, and meanwhile, semantic information is added as assistance, so that the relative position of each laser radar on a point cloud map can be obtained, that is, the x, y, z, roll, pitch and yaw information of each laser radar coordinate system relative to a point cloud map coordinate system can be determined, and finally, the 6-degree-of-freedom position and posture relation of each laser radar coordinate system is obtained; the method comprises the following steps of calculating the coordinate system of each laser radar coordinate system, wherein X, Y and Z represent the offset of each laser radar coordinate system relative to the X-axis, Y-axis and Z-axis directions of a point cloud map coordinate system, roll represents the rotation angle of each laser radar coordinate system relative to the point cloud map coordinate system around the X-axis, pitch represents the rotation angle of each laser radar coordinate system relative to the point cloud map coordinate system around the Y-axis, and yaw represents the rotation angle of each laser radar coordinate system relative to the point cloud map coordinate system around the Z-axis.
And the external reference optimization module is used for calculating the origin position of the ore removal truck body coordinate system based on the 6-degree-of-freedom pose relationship among the 6 laser radars and solving the external reference of each laser radar relative to the ore removal truck body coordinate system. The external reference optimization module mainly transforms the relative pose of each laser radar to the pose of a relative vehicle body coordinate system. Because the double-axle steering mine car without the cab does not have obvious front-end and rear-end distinction during driving, the laser radar data is calibrated in a car body coordinate system taking the center of a car body as the origin of coordinates, and the perception data at the moment does not have obvious front-back distinction any more, so that the distance between obstacles can be perceived to be rapidly switched when the driving direction of the double-axle steering mine car is switched, and the driving characteristics of the double-axle steering mine car without the cab are met.
It will be apparent to those skilled in the art that various modifications and improvements can be made to the embodiments of the present invention without departing from the inventive concept thereof, and these modifications and improvements are intended to be within the scope of the invention.
Claims (10)
1. A multi-radar external reference rapid calibration method for a double-axle steering mine card is characterized by comprising the following steps:
s1: when the mine truck is in a calibration scene, acquiring point cloud data of a plurality of laser radars on a mine truck body, measuring the ground clearance installation height of each laser radar, and taking the laser radar positioned in the middle of the mine truck body as a main laser radar;
s2: dividing the point cloud data into ground point cloud data and non-ground point cloud data based on the ground-above-ground installation height of each laser radar, and adding semantic category information to the non-ground point cloud data;
s3: performing plane parameter fitting on the ground point cloud data to obtain ground point cloud plane parameters, and constructing multiple degree-of-freedom constraint relations among the laser radars based on the ground point cloud plane parameters;
s4: constructing a point cloud map based on the non-ground point cloud data with semantic category information of the main laser radar;
s5: based on the constructed multiple degree-of-freedom constraint relations among the laser radars and the constructed point cloud map, performing point cloud map registration on the point cloud data of the laser radars acquired in the step S1 to obtain the positions of the coordinate systems of the laser radars relative to the coordinate system of the point cloud map, and further obtain the position relation of 6 degrees-of-freedom of the coordinate systems of the laser radars;
s6: and calculating the origin position of the mine truck body coordinate system based on the 6-degree-of-freedom pose relation of the coordinate system of each laser radar, and solving the external reference of each laser radar relative to the mine truck body coordinate system.
2. The method as claimed in claim 1, wherein in step S1, the point cloud data of the lidar is collected in the forward and backward directions of the truck body of the mine truck, respectively, in a manner of in-situ steering of the mine truck.
3. The method according to claim 1, wherein the step S2 is specifically performed by:
according to the ground-to-ground installation height of each laser radar, point cloud data of each laser radar is divided into ground point cloud data and non-ground point cloud data, and the non-ground point cloud data is subjected to semantic division, so that each point in the non-ground point cloud data has semantic category information except space geometric information, and a plurality of analog point clouds are generated.
4. The method according to claim 1, wherein the step S3 is specifically performed by:
and performing plane parameter fitting on the ground point cloud data of each laser radar by using a RANSAC method to obtain the ground point cloud plane parameters of each laser radar, and constructing a plurality of freedom degree constraint relations of each laser radar in a spatial position.
5. The method according to claim 1, wherein the step S4 is specifically performed by:
and extracting line and surface features by using non-ground point cloud data of the main laser radar, adding semantic category information constraint in a feature matching stage, calculating the pose relationship of adjacent frame point clouds, and overlapping to construct a point cloud map.
6. The method according to claim 1, wherein the step S5 is specifically performed by:
based on a plurality of degree-of-freedom constraint relations among the constructed laser radars, performing point cloud map matching on point cloud data of the laser radars on the constructed point cloud map, continuously iterating to obtain positioning parameters of the laser radars relative to the point cloud map, and finally obtaining x, y, z, roll, pitch and yaw information of a coordinate system of the laser radars relative to a coordinate system of the point cloud map, so as to obtain a 6-degree-of-freedom pose relation of the coordinate system of the laser radars; the offset of the coordinate system of each laser radar relative to the coordinate system of the point cloud map in the X-axis, Y-axis and Z-axis directions is represented by X, Y and Z, the rotation angle of the coordinate system of each other laser radar relative to the coordinate system of the point cloud map around the X-axis is represented by roll, the rotation angle of the coordinate system of each laser radar relative to the coordinate system of the point cloud map around the Y-axis is represented by pitch, and the rotation angle of the coordinate system of each laser radar relative to the coordinate system of the point cloud map around the Z-axis is represented by yaw.
7. The method according to claim 1, wherein step S6 specifically comprises the following steps:
and calculating the origin position of the ore removal truck body coordinate system based on the 6-degree-of-freedom position and posture relation of the coordinate system of each laser radar, converting the 6-degree-of-freedom position and posture relation of the coordinate system of each laser radar into parameters under the ore removal truck body coordinate system, and obtaining external parameters of each laser radar relative to the ore removal truck body coordinate system.
8. A quick calibration device of many radars external reference for two bridge steering mine cards which characterized in that includes:
the acquisition module is used for acquiring point cloud data of a plurality of laser radars on the mine truck body;
the point cloud segmentation module is used for segmenting the point cloud data acquired by the acquisition module into ground point cloud data and non-ground point cloud data; meanwhile, the method is used for adding semantic category information to the non-ground point cloud data;
the plane fitting module is used for performing plane parameter fitting on the ground point cloud data of each laser radar to obtain ground point cloud plane parameters, constructing multiple freedom degree constraint relations among the laser radars and transmitting the multiple freedom degree constraint relations to the repositioning module;
the mapping module is used for constructing a point cloud map based on non-ground point cloud data with semantic category information of a main laser radar, wherein the main laser radar is a laser radar positioned in the middle of the mine truck body;
the repositioning module is used for receiving a plurality of freedom degree constraints among the laser radars constructed by the plane fitting module, and performing point cloud map matching on the point cloud data of the laser radars from the acquisition module to obtain a 6-freedom degree pose relation of a coordinate system of each laser radar;
and the external reference optimization module is used for calculating the origin position of the ore removal truck body coordinate system based on the 6-degree-of-freedom pose relationship of the coordinate system of each laser radar and solving the external reference of each laser radar relative to the ore removal truck body coordinate system.
9. An electronic device, comprising:
memory, processor and program stored on the memory and executable on the processor, the processor executing the program to implement the method according to one of claims 1 to 7.
10. A computer-readable storage medium, characterized in that the computer-readable storage medium has stored thereon a program which is executed by a processor to implement the method according to one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210670457.8A CN115079143B (en) | 2022-06-14 | 2022-06-14 | Multi-radar external parameter quick calibration method and device for double-bridge steering mine card |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210670457.8A CN115079143B (en) | 2022-06-14 | 2022-06-14 | Multi-radar external parameter quick calibration method and device for double-bridge steering mine card |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115079143A true CN115079143A (en) | 2022-09-20 |
| CN115079143B CN115079143B (en) | 2024-04-26 |
Family
ID=83252093
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210670457.8A Active CN115079143B (en) | 2022-06-14 | 2022-06-14 | Multi-radar external parameter quick calibration method and device for double-bridge steering mine card |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115079143B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115797422A (en) * | 2022-12-01 | 2023-03-14 | 西南交通大学 | Semantic map-based cross-view repositioning method from ground to unmanned aerial vehicle laser point cloud |
| CN117269939A (en) * | 2023-10-25 | 2023-12-22 | 北京路凯智行科技有限公司 | Parameter calibration system, method and storage medium for sensor |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20130226344A1 (en) * | 2012-02-29 | 2013-08-29 | Irobot Corporation | Mobile Robot |
| CN110596683A (en) * | 2019-10-25 | 2019-12-20 | 中山大学 | Multi-group laser radar external parameter calibration system and method thereof |
| CN113552559A (en) * | 2021-07-23 | 2021-10-26 | 北京轻舟智航科技有限公司 | Laser radar external parameter correction method in unmanned driving |
| CN113640778A (en) * | 2021-08-12 | 2021-11-12 | 东风悦享科技有限公司 | Multi-laser radar combined calibration method based on non-overlapping view field |
| WO2021253193A1 (en) * | 2020-06-15 | 2021-12-23 | 深圳市大疆创新科技有限公司 | Calibration method and calibration apparatus for external parameters of multiple groups of laser radars, and computer storage medium |
-
2022
- 2022-06-14 CN CN202210670457.8A patent/CN115079143B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20130226344A1 (en) * | 2012-02-29 | 2013-08-29 | Irobot Corporation | Mobile Robot |
| CN110596683A (en) * | 2019-10-25 | 2019-12-20 | 中山大学 | Multi-group laser radar external parameter calibration system and method thereof |
| WO2021253193A1 (en) * | 2020-06-15 | 2021-12-23 | 深圳市大疆创新科技有限公司 | Calibration method and calibration apparatus for external parameters of multiple groups of laser radars, and computer storage medium |
| CN113552559A (en) * | 2021-07-23 | 2021-10-26 | 北京轻舟智航科技有限公司 | Laser radar external parameter correction method in unmanned driving |
| CN113640778A (en) * | 2021-08-12 | 2021-11-12 | 东风悦享科技有限公司 | Multi-laser radar combined calibration method based on non-overlapping view field |
Non-Patent Citations (2)
| Title |
|---|
| 康国华;张琪;张晗;徐伟证;张文豪;: "基于点云中心的激光雷达与相机联合标定方法研究", 仪器仪表学报, no. 12, 15 December 2019 (2019-12-15) * |
| 韩栋斌;徐友春;李华;谢德胜;陈文;: "基于手眼模型的三维激光雷达外参数标定", 光电工程, no. 08, 15 August 2017 (2017-08-15) * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115797422A (en) * | 2022-12-01 | 2023-03-14 | 西南交通大学 | Semantic map-based cross-view repositioning method from ground to unmanned aerial vehicle laser point cloud |
| CN117269939A (en) * | 2023-10-25 | 2023-12-22 | 北京路凯智行科技有限公司 | Parameter calibration system, method and storage medium for sensor |
| CN117269939B (en) * | 2023-10-25 | 2024-03-26 | 北京路凯智行科技有限公司 | Parameter calibration system, method and storage medium for sensor |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115079143B (en) | 2024-04-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP7073315B2 (en) | Vehicles, vehicle positioning systems, and vehicle positioning methods | |
| CN112396650B (en) | Target ranging system and method based on fusion of image and laser radar | |
| CN110859044B (en) | Integrated sensor calibration in natural scenes | |
| CN113870343B (en) | Relative pose calibration method, device, computer equipment and storage medium | |
| CN110119698B (en) | Method, apparatus, device and storage medium for determining object state | |
| CN103487034B (en) | Method for measuring distance and height by vehicle-mounted monocular camera based on vertical type target | |
| CN109752701A (en) | A kind of road edge detection method based on laser point cloud | |
| CN110031829B (en) | Target accurate distance measurement method based on monocular vision | |
| CN115079143B (en) | Multi-radar external parameter quick calibration method and device for double-bridge steering mine card | |
| CN104766058A (en) | Method and device for obtaining lane line | |
| CN111680611B (en) | Road trafficability detection method, system and equipment | |
| CN107796373B (en) | Distance measurement method based on monocular vision of front vehicle driven by lane plane geometric model | |
| CN112346463B (en) | Unmanned vehicle path planning method based on speed sampling | |
| CN113327296B (en) | Laser radar and camera online combined calibration method based on depth weighting | |
| Kellner et al. | Road curb detection based on different elevation mapping techniques | |
| Heng | Automatic targetless extrinsic calibration of multiple 3D LiDARs and radars | |
| CN112232275A (en) | Obstacle detection method, system, equipment and storage medium based on binocular recognition | |
| CN113989766A (en) | Road edge detection method and road edge detection equipment applied to vehicle | |
| CN114413958A (en) | Monocular vision distance and speed measurement method of unmanned logistics vehicle | |
| CN114325634A (en) | Method for extracting passable area in high-robustness field environment based on laser radar | |
| CN113985405A (en) | Obstacle detection method and obstacle detection equipment applied to vehicle | |
| CN111506069A (en) | All-weather all-ground crane obstacle identification system and method | |
| CN116129669A (en) | Parking space evaluation method, system, equipment and medium based on laser radar | |
| CN115267756A (en) | Monocular real-time distance measurement method based on deep learning target detection | |
| CN113624223B (en) | Indoor parking lot map construction method and device |
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 |