CN114035167A - Target high-precision sensing method based on roadside multi-sensors - Google Patents
Target high-precision sensing method based on roadside multi-sensors Download PDFInfo
- Publication number
- CN114035167A CN114035167A CN202111293177.1A CN202111293177A CN114035167A CN 114035167 A CN114035167 A CN 114035167A CN 202111293177 A CN202111293177 A CN 202111293177A CN 114035167 A CN114035167 A CN 114035167A
- Authority
- CN
- China
- Prior art keywords
- coordinates
- radar
- target
- sensor
- gps
- 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.)
- Pending
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
- G01S7/02—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00
- G01S7/40—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
- G01S19/00—Satellite radio beacon positioning systems; Determining position, velocity or attitude using signals transmitted by such systems
- G01S19/01—Satellite radio beacon positioning systems transmitting time-stamped messages, e.g. GPS [Global Positioning System], GLONASS [Global Orbiting Navigation Satellite System] or GALILEO
- G01S19/13—Receivers
- G01S19/23—Testing, monitoring, correcting or calibrating of receiver elements
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Computer Networks & Wireless Communication (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Position Fixing By Use Of Radio Waves (AREA)
Abstract
A high-precision target sensing method based on roadside multisensors comprises the following steps: installing a sensor and acquiring a GPS coordinate of the sensor; detecting a target, namely detecting the target by using each sensor to obtain the coordinates of the target under each sensor; converting coordinates of the target under each sensor into GPS coordinates, wherein the GPS coordinates comprise pixel coordinates converted into GPS coordinates, radar coordinates converted into GPS coordinates and target GPS coordinates detected by a laser radar; and secondary coordinate conversion is carried out, and the GPS coordinates are converted into other coordinates. The invention integrates a plurality of methods; the GPS positioning information of the target can be acquired only by utilizing the road side sensor, the problem of inaccurate GPS positioning when a GPS signal is weak is avoided, and meanwhile, the sensor does not need to be installed on the vehicle, so that the requirement and the cost on the vehicle are reduced.
Description
Technical Field
The invention relates to the field of roadside sensing, in particular to a high-precision target sensing method based on roadside multi-sensors.
Background
In automatic driving, it is a crucial part to perform environmental perception. The use of vehicle sensors for environmental sensing requires one or more sensors for each vehicle, requires high computational power on the vehicle sensors and vehicles, and increases vehicle costs. And the roadside sensor is used for sensing the environment, so that the problems of computing power and cost can be solved, and the problem of the environmental sensing performance of the vehicle sensor under the condition of target shielding and long distance can be solved. Detection principles and coordinate representations of various roadside sensors are different, and how to use the roadside sensors to position and fuse targets is important.
In the prior art, for example, patent CN113433542A discloses a vehicle positioning scheme for determining the GPS position of a vehicle by acquiring the GPS position of a roadside device and the position of the vehicle relative to the roadside device, but this step is not described in detail. Patent CN111787481A proposes a high-precision sensing method for road-vehicle coordination. Respectively using a visual sensor and a millimeter wave radar to detect a target and perform visual ranging; projecting a millimeter wave radar detection target onto a pixel coordinate system, performing matching association and tracking with a vision sensor detection target, and obtaining a target coordinate xy after weighted average; converting the coordinates into site coordinates ENU and then converting the coordinates into GPS; the method has the following defects: 1. the monocular distance measurement accuracy is low, and the positioning precision and the fusion performance are directly influenced; 2. fusion is carried out under a pixel coordinate system, and data association logic and parameters are complex; and 3, the ENU coordinate system is complex to use, and the rotation angle of the U shaft of the E shaft and the N shaft of the roadside fusion equipment needs to be acquired respectively.
Disclosure of Invention
According to the problems provided by the background art, the patent provides a high-precision target sensing method based on roadside multisensors, which is used for solving the problems of positioning and fusing targets by the roadside multisensors, acquiring GPS positioning information of the targets by only using the roadside multisensors, and fusing the multisensors based on the GPS positioning. The invention will be further elucidated below.
A high-precision target sensing method based on roadside multisensors comprises the following steps:
s1, installing a sensor and acquiring a GPS coordinate of the sensor;
s2, detecting the target by using each sensor, and acquiring the coordinates of the target under each sensor;
s3, converting coordinatesTarget under each sensorConverting coordinates to GPS coordinates, including pixel coordinates to GPS coordinates, radar coordinates to GPS coordinates, laser radarConverting the coordinate to a GPS coordinate;
and S4, converting the GPS coordinates into other coordinates by secondary coordinate conversion.
Further, for the target detection in step S2, the target detection is performed using each sensor, coordinates of the target under each sensor are acquired, pixel coordinates (u, v) of the target are acquired using the camera, coordinates (x, y) of the target with respect to the millimeter wave radar are acquired using the millimeter wave radar, and coordinates (x, y, z) of the target with respect to the laser radar are acquired using the laser radar.
Further, for the pixel coordinate to GPS coordinate conversion in the coordinate conversion of step S3:
the sensor uses a camera, converts pixel coordinates to GPS coordinates using a homography matrix method, and converts pixel coordinates to GPS coordinates by acquiring a homography matrix of pixel coordinates (u, v) and GPS coordinates (L, B) regardless of altitude, and vice versa.
Further, for the conversion from the millimeter wave radar coordinate to the GPS coordinate in the coordinate conversion in step S3, a method based on the north-offset angle of the radar is used, or a single-point calibration method is used to obtain the radar coordinate and the GPS coordinate point pair of a single point for calibration.
Further, as for the lidar coordinates to GPS coordinates in the coordinate conversion in step S3, the lidar coordinates are converted to GPS coordinates using a rigid transformation matrix method, and the lidar coordinates are converted to GPS coordinates by acquiring a rigid transformation matrix of the lidar coordinates (x, y, z) and the GPS coordinates (L, B, H), and vice versa.
Further, for the secondary coordinate transformation of step S4, the method includes transforming GPS to radar coordinates based on a radar north-offset angle method and/or a single-point calibration method;
the method for converting the GPS to the radar coordinate based on the radar north-offset angle method comprises the following steps: :
s41, acquiring GPS coordinates of the radar installation position and a north-bias angle theta of the radar installation;
s42, calculating UTM coordinates (Xw0, Yw0) of the radar installation position;
s43, acquiring the GPS coordinates of the radar target;
s44, calculating UTM coordinates (Xw, Yw) of the radar target;
s45, calculating coordinates (x ', y') after rotation;
calculating (x, y) coordinates of the radar target;
the method for converting the GPS to the radar coordinate based on the single-point calibration method comprises the following steps:
sa41, acquiring GPS coordinates of the radar mounting position and a rotation angle θ of the radar coordinate system and the UTM coordinate system;
sa42, calculating UTM coordinates (Xw0, Yw0) of the radar mounting position;
sa43, calculating corresponding UTM coordinates (Xw, Yw) for the GPS coordinates of any radar target;
sa44, calculating the coordinates (x ', y') after rotation,
sa44, calculating the (x, y) coordinates of the radar target,
further, in the secondary coordinate conversion process of step S4, the coordinate conversion is performed to calculate the distance, and the distance between the target and the main sensor is obtained for target fusion, including the following steps;
erecting a sensor i, acquiring a GPS coordinate of an installation position of the sensor i, and calculating a UTM coordinate (Xwi, Ywi) of the point, wherein i is the number of sensors, and is 1, 2, 3;
calibrating a sensor and a GPS, acquiring a sensor coordinate and a GPS coordinate of a target, and calculating a UTM coordinate (Xwj, Ywj) of the target, wherein j is the number of the targets, and j is 1, 2, 3;
based on the above steps, the distances in the horizontal direction and the vertical direction between an arbitrary target and an arbitrary sensor are calculated:
wherein i is the number of sensors, i is 1, 2, 3; j is the target number, j 1, 2, 3.;
calculating the distance between any target and any target in the horizontal direction and the vertical direction:
j is the target number, j is 1, 2, 3., j is not j';
calculating the distance between any sensor and any sensor in the horizontal direction and the vertical direction:
i is the number of sensors, i ≠ i', 1, 2, 3.
Further, the distance between a target and any sensor is acquired for sensor relay, and any plurality of sensors are calibrated simultaneously in a relay mode, and the method comprises the following specific steps:
erecting a plurality of sensors, establishing a certain main sensor A, and ensuring that any sensor is positioned in the detection range of another sensor adjacent to the sensor A when the other sensors are erected to form a sensor chain;
calibrating a main sensor A, setting another sensor in the detection range of the main sensor A as a sensor B, driving a target to be right below the sensor B, and acquiring a GPS coordinate of the target through the main sensor A, namely the GPS coordinate of the installation position of the sensor B;
the target runs into the common detection range of the main sensor A and the sensor B, the GPS coordinate of the target is obtained through the main sensor A, and meanwhile, the sensor coordinate of the target under the sensor B, such as pixel coordinate (u, v) and radar coordinate (x, y), is obtained;
calibrating the sensor B according to the GPS coordinates of the target under the main sensor A and the sensor coordinates under the sensor B; if the sensor B needs to be calibrated in multiple points, repeating the previous step until the number of the point pairs meets the requirement;
in the above steps, the sensor B in the detection range of the sensor A is calibrated through the calibrated sensor A, and similarly, the sensor C in the detection range of the sensor B is calibrated through the above steps, and the calibration of any plurality of sensors can be realized by circulating and reciprocating along the sensor chain.
Furthermore, the sensor relay takes GPS coordinates as intermediate quantity to realize coordinate conversion between any sensors, including multi-source sensor calibration and homologous sensor calibration; the multi-source sensor calibration, for example, converting the pixel coordinates (u, v) and the radar coordinates (x, y), firstly converting the pixel coordinates (u, v) into GPS coordinates, and then converting the GPS coordinates into the radar coordinates, so as to realize the conversion of the pixel coordinates into the radar coordinates; and (3) calibrating the homologous sensor, for example, converting the radar A coordinate (x1, y1) and the radar B coordinate (x2, y2), firstly converting the radar A coordinate (x1, y1) into a GPS coordinate, and then converting the GPS coordinate into the radar B coordinate (x2, y2), so as to realize the conversion of the radar coordinate into the radar coordinate.
Has the advantages that: compared with the prior art, the method integrates various methods such as an image and GPS calibration method based on a homography matrix, a radar and GPS calibration method based on a radar north-bias angle, a radar and GPS calibration method based on a single-point calibration method, a laser radar and GPS calibration method based on a rigid transformation matrix, a calculation method of the distance between any target and any sensor, a coordinate conversion method between any sensors and the like; the GPS positioning information of the target can be acquired only by utilizing the road side sensor, the problem of inaccurate GPS positioning when a GPS signal is weak is avoided, and meanwhile, the sensor does not need to be installed on the vehicle, so that the requirement and the cost on the vehicle are reduced.
Drawings
FIG. 1: a diagram of the method steps of the present invention;
FIG. 2: the relationship between the UTM coordinate system and the radar coordinate system;
FIG. 3: setting a main sensor, and acquiring schematic distance diagrams between all targets and the main sensor;
FIG. 4: obtaining a distance schematic diagram between a target and any sensor;
FIG. 5: sensor relay calibration schematic diagram.
Detailed Description
A specific embodiment of the present invention will be described in detail with reference to fig. 1-5.
A high-precision target sensing method based on roadside multi-sensors,
the method mainly comprises the following steps:
s1, installing a sensor and acquiring a GPS coordinate of the sensor;
s2, detecting the target by using each sensor, and acquiring the coordinates of the target under each sensor;
s3, converting coordinatesTarget under each sensorConverting the coordinates into GPS coordinates;
and S4, converting the GPS coordinates into other coordinates by secondary coordinate conversion.
The above steps are explained in detail:
for the sensor installation of S1, the GPS of the sensor is obtained, the sensor is installed on the roadside, such as a gantry, a street lamp, an electric police pole, and the like, the sensor may be a camera, a millimeter wave radar, a laser radar, and the like, and the GPS coordinates installed on each sensor are obtained using RTK and other equipment.
For the target detection in S2, each sensor is used to perform target detection, and coordinates of the target under each sensor are obtained, and if the sensor uses a camera, pixel coordinates (u, v) of the target can be obtained; if the sensor uses a millimeter wave radar, the coordinates (x, y) of the target relative to the millimeter wave radar can be obtained, and if the sensor uses a laser radar, the GPS coordinates of the target can be obtained;it should be noted that the millimeter wave radar is used to acquire the target relative to the millimeter waveCoordinates (x, y) of the radar as phase For a target, the GPS coordinates of the lidar used to acquire the target are absolute targets, which are collectively called radar in practice, and the technology in the art The operator can automatically distinguish the two kinds of information when in use, and no ambiguity is generated。
For the coordinate transformation of S3, converting the sensor coordinates to GPS coordinates, including pixel coordinates to GPS coordinates, radar coordinates to GPS coordinates, GPS coordinates of a target detected by a lidar, wherein:
converting the pixel coordinate into a GPS coordinate;
the sensor uses a camera, and then a homography matrix method can be used for converting pixel coordinates into GPS coordinates, wherein the homography matrix expresses the conversion relation of two planes, namely the coordinates (x ', y') of a plane point (x, y) corresponding to the other plane can be obtained by obtaining the homography matrix; the GPS is a three-dimensional coordinate (L, B, H), the longitude and latitude height of the target is expressed, when the detected target is on the ground, the height information of the target can be ignored, and the GPS can be similar to a two-dimensional coordinate (L, B), so that the pixel coordinate can be converted into the GPS coordinate by acquiring a homography matrix of the pixel coordinate (u, v) and the GPS coordinate (L, B), and vice versa; the method for acquiring the homography matrix is a published and mature algorithm, and at least 4 groups of corresponding pixel coordinate and GPS coordinate point pairs are acquired, so that the homography matrix can be acquired.
For the conversion from radar coordinates to GPS coordinates, the present embodiment provides three methods, namely a radar north-offset angle method, a homography matrix method, and a single-point calibration method.
Based on the method of the north-offset angle of the radar, the millimeter wave radar coordinate is converted into a GPS coordinate, the UTM coordinate System is introduced as an intermediate coordinate System in this embodiment, and the UTM (Universal transform Mercator Grid System) coordinate System is a planar rectangular coordinate, and the representation format thereof is as follows: longitude/latitude/east/north, where east represents the projected distance from the center meridian of the longitude and north represents the projected distance (in meters) from the equator, the GPS coordinates and UTM coordinates can be converted to each other, and the conversion method is a well-known and well-established algorithm.
The radar coordinate system is a plane rectangular coordinate system, the Y axis points to the normal direction of the radar, the X axis is vertical to the Y axis and accords with the rule of a right-hand coordinate system, a target detected by the radar has unique (X, Y) coordinates, and the expression is that the distance (unit: meter) between the target and the X axis and the Y axis of the radar.
As can be seen from the above, the UTM coordinate system and the radar coordinate system are both rectangular coordinate systems, the radar installation position is used as the origin, the UTM coordinate system and the radar coordinate system are simultaneously established, and the UTM coordinate system is set to use the east direction as the X axis and the north direction as the Y axis. The two coordinate systems are in a rotation translation relation, and the rotation angle is an included angle between the two Y axes, namely an included angle between the radar and the north direction, namely a radar north-bias angle. Since the origins coincide, the amount of translation is 0.
Referring to fig. 2, the UTM coordinate system and the radar coordinate system perform conversion between radar coordinates and GPS coordinates based on a method of radar north-offset angle, wherein the method of converting millimeter-wave radar coordinates into GPS coordinates is as follows:
a1, acquiring GPS coordinates of a radar installation position and a north-bias angle theta of the radar installation;
a2, calculating UTM coordinates (Xw0, Yw0) of the radar installation position;
a3, acquiring (x, y) coordinates of the radar target;
a4, calculating coordinates (x ', y') after rotation;
a5, calculating UTM coordinates (Xw, Yw) of the radar target;
a6, calculating the GPS coordinates of the radar target.
The radar north-offset angle θ is obtained as follows:
a11, setting an initial north-bias angle theta according to the installation position and orientation of the radar0E.g. due to north00, east-ward direction θ0＝90；
a12, driving the target along the lane line direction, and acquiring the GPS coordinates of the target at each moment;
a13, displaying the target GPS coordinate track on a high-precision map, and checking whether the track runs in a lane line; if the vehicle just runs in the lane line, the north deflection angle of the radar is theta0(ii) a Otherwise, readjusting theta, theta being theta0And +/-delta theta, and circularly executing the operations a12 and a 13.
Based on the homography matrix method, the aforementioned pixel coordinates may be referred to as GPS coordinates, and details thereof are not repeated herein.
Based on a single-point calibration method, acquiring radar coordinates and GPS coordinate point pairs of a single point for calibration, and specifically comprising the following steps:
b1, calculating the rotation angle theta according to the radar coordinates and the GPS coordinates of the single point, and the steps are as follows;
b11, acquiring the GPS coordinates of the radar installation position;
b12, calculating UTM coordinates (Xw0, Yw0) of the radar installation position;
b13, acquiring (x0, y0) coordinates and GPS coordinates of a certain target; (when the target travels in the radar detection area, the radar detection coordinates (x0, y0) can be read, and the GPS coordinates can be acquired by using RTK or the like)
b14, calculating UTM coordinates (Xw, Yw) of the radar target;
calculating the rotated coordinates (x ', y');
calculating a rotation angle theta:
b2, converting the radar coordinate into the GPS coordinate according to the rotation angle theta
b21, calculating the coordinates (x ', y') after rotation for any radar target (x, y) coordinates;
b22, calculating UTM coordinates (Xw, Yw) of the radar target;
b23, calculating the GPS coordinates of the radar target.
For the conversion from the laser radar coordinate (x, y, z) to the GPS coordinate, the sensor uses the laser radar, the sensor uses the rigid transformation matrix method to convert the laser radar coordinate to the GPS coordinate, and the sensor converts the laser radar coordinate to the GPS coordinate by obtaining the rigid transformation matrix between the laser radar coordinate (x, y, z) and the GPS coordinate (L, B, H), and vice versa.
For the secondary coordinate conversion in step S4, because the GPS coordinates describe longitude and latitude, the coordinates are not intuitive in actual use, and the coordinates can be converted to have an actual distance meaning, which specifically includes:
s4.1, calculating the distance;
the GPS coordinates of the sensor may be converted to UTM coordinates (x0, y0), the GPS coordinates of the target may be converted to UTM coordinates (x1, y1), and then the coordinates (x2, y2) are expressed as the distance of the arbitrary point from the sensor:
s4.2, calibrating the sensor;
the GPS coordinates are used as intermediate quantity, coordinate conversion between any sensors can be realized, and pairwise calibration between the sensors, including multi-source sensor calibration and homologous sensor calibration, is avoided.
And (3) calibrating the multi-source sensor, namely converting the pixel coordinates (u, v) into GPS coordinates and converting the GPS coordinates into radar coordinates if the pixel coordinates (u, v) and the radar coordinates are converted, so that the conversion of the pixel coordinates into the radar coordinates can be realized, and vice versa.
And (3) calibrating the homologous sensor, if radar coordinates (x1, y1) and radar coordinates (x2, y2) are converted, firstly converting the radar coordinates (x1, y1) into GPS coordinates, and then converting the GPS coordinates into the radar coordinates (x2, y2), so that the conversion of the radar coordinates into the radar coordinates can be realized.
In environmental sensing, a sensor is generally used for detecting a target and acquiring a coordinate position of the target under the sensor, but due to the fact that the sensor is insufficient, fusion detection can be performed through multiple sensors, detection accuracy is improved, and a detection range is expanded. The calibration between the sensors is carried out on the premise of multi-sensor fusion, the coordinate conversion relation can be obtained by calibrating every two sensors, or any sensor and a GPS are calibrated and fused under a GPS coordinate system.
The multi-sensor and GPS calibration has the following defects: the GPS coordinates are not visual, the longitude and the latitude and the height are represented, and the distance cannot be directly represented; the information between the sensors is independent, the information is not fully utilized, if a certain target is in the detection range of the sensor A and out of the detection range of the sensor B, the sensor A detects the target and then converts the target into a GPS coordinate, and the GPS coordinate has no any significance to the sensor B. Based on this, the present embodiment gives GPS coordinates to be converted into local coordinates having a distance meaning, specifically as follows:
s4.1.1, erecting a sensor i, acquiring a GPS coordinate of an installation position of the sensor i, and calculating a UTM coordinate (Xwi, Ywi) of the point, wherein i is the number of sensors, and i is 1, 2 and 3;
s4.1.2, calibrating the sensor and the GPS, such as calibrating the camera and the GPS, and calibrating the millimeter wave radar and the GPS;
acquiring sensor coordinates and GPS coordinates of a target, and calculating UTM coordinates (Xwj, Ywj) of the target, wherein j is the number of targets, and j is 1, 2 and 3;
s4.1.3, based on the above steps, calculating the distance between any target and any sensor in the horizontal direction and the vertical direction:
wherein i is the number of sensors, i is 1, 2, 3; j is the target number, j 1, 2, 3.;
calculating the distance between any target and any target in the horizontal direction and the vertical direction:
j is the target number, j is 1, 2, 3., j is not j';
calculating the distance between any sensor and any sensor in the horizontal direction and the vertical direction:
i is the number of sensors, i ≠ i', 1, 2, 3.
Therefore, the present embodiment can convert the GPS coordinates into local coordinates having a distance meaning, and can perform information intercommunication between sensors to obtain a distance between any target and any sensor, a distance between any target and any target, and a distance between any sensor and any sensor.
It should be noted that when the sensor and the GPS are calibrated, at least 4 sets of pairs of sensor coordinates, such as pixel coordinates (u, v) and radar coordinates (x, y), need to be acquired, if the homography matrix method is adopted. GPS coordinates are inconvenient to acquire in certain specific scenes, such as high-speed scenes, and the GPS coordinates in the lane are abnormal and dangerous by using RTK; such as a tunnel scene, the signal is weak, and the GPS coordinates are not obtained or are inaccurate.
Referring to fig. 5, in order to reduce the calibration amount, a plurality of sensors are arranged in the driving direction to form a sensor chain, the field of view of the sensors is in a fan shape, one sensor is firstly calibrated, then other sensors are calibrated in a relay manner, and the calibration is divided into homodromous calibration and reverse calibration according to whether the superposition direction of the field of view of the sensor chain is the same as the driving direction, wherein the calibration steps are as follows:
erecting a plurality of sensors, wherein the visual fields of adjacent sensors are overlapped;
setting a main sensor A, and calibrating the main sensor A;
setting a sensor adjacent to the main sensor A as a sensor B, driving a target into a common detection range of the main sensor A and the sensor B, acquiring a GPS coordinate of the target through the main sensor A, and acquiring a sensor coordinate of the target under the sensor B through the main sensor B;
calibrating the sensor B according to the GPS coordinate of the target under the main sensor A and the sensor coordinate under the sensor B; and repeating the previous step and the current step in a circulating way until the calibration of all the sensors is realized.
It should be noted that, if the GPS coordinates of any sensor installation position need to be acquired, preferably, the same-direction calibration is performed, that is, the calibrated sensor is used to calibrate the adjacent sensor in the detection range, taking the calibrated main sensor a and the adjacent sensor B as an example, when a target runs to a position right below the sensor B, and the target is in the detection range of the calibrated main sensor a, the GPS coordinates of the target can be acquired through the main sensor a, the GPS coordinates are combined by the acquired latitude coordinates and the height coordinates acquired by combining the installation height, the GPS coordinates are the GPS coordinates of the sensor B installation position, and the GPS coordinates of all sensors in the sensor chain can be sequentially acquired.
Based on the arrangement of the sensor chain, any plurality of sensors can be calibrated simultaneously in a relay mode only by calibrating one sensor, so that the calibration requirement is reduced, and the calibration safety can be improved if the sensor chain is used in a high-speed scene; if the sensor calibration device is used in a tunnel scene, the sensor is calibrated outside the tunnel, and the sensor in the tunnel is calibrated in a relay mode, so that the calibration precision can be improved.
The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention, and various modifications and changes may be made by those skilled in the art. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (10)
1. A high-precision target sensing method based on roadside multi-sensors is characterized by comprising the following steps:
s1, installing a sensor and acquiring a GPS coordinate of the sensor;
s2, detecting the target by using each sensor, and acquiring the coordinates of the target under each sensor;
s3, converting coordinatesTarget under each sensorConverting the coordinates into GPS coordinates, wherein the GPS coordinates are converted from pixel coordinates, millimeter wave radar coordinates and laser radar coordinates;
and S4, converting the GPS coordinates into other coordinates by secondary coordinate conversion.
2. The method for sensing the target with high precision as claimed in claim 1, wherein for the target detection in step S2, the target detection is performed by using each sensor, the coordinates of the target under each sensor are obtained, the pixel coordinates (u, v) of the target are obtained by using the camera, the coordinates (x, y) of the target relative to the millimeter wave radar are obtained by using the millimeter wave radar, and the coordinates (x, y, z) of the target relative to the laser radar are obtained by using the laser radar.
3. The method for sensing a target with high accuracy according to claim 2, wherein for the pixel coordinate-to-GPS coordinate conversion in the step S3 coordinate conversion:
the sensor uses a camera, converts pixel coordinates to GPS coordinates using a homography matrix method, and converts pixel coordinates to GPS coordinates by acquiring a homography matrix of pixel coordinates (u, v) and GPS coordinates (L, B) regardless of altitude, and vice versa.
4. The method for sensing a target with high accuracy as claimed in claim 2, wherein for the conversion of radar coordinates to GPS coordinates in the coordinate conversion of step S3, a method based on a north-offset angle of radar is used to convert millimeter wave radar coordinates to GPS coordinates, with a radar installation position as an origin, and simultaneously establish a UTM coordinate system and a radar coordinate system, and the UTM coordinate system is set to have an east-oriented direction as an X-axis and a north-oriented direction as a Y-axis; setting a normal line of the radar as a Y axis, and establishing an X axis according to a right-hand coordinate system; the UTM coordinate system and the radar coordinate system are in a rotating translation relation, the rotating angle is an included angle of two Y axes, due to the coincidence of the original points, the translation amount is 0, and the method for converting the millimeter wave radar coordinate into the GPS coordinate based on the radar north-offset angle comprises the following steps:
sa1, acquiring GPS coordinates of a radar installation position and a north-bias angle theta of radar installation;
sa2, calculating UTM coordinates (Xw0, Yw0) of the radar mounting position;
sa3, acquiring (x, y) coordinates of the radar target;
sa4, calculating coordinates (x ', y') after rotation;
sa5, calculating UTM coordinates (Xw, Yw) of the radar target;
sa6, calculates GPS coordinates of the radar target.
5. The method for sensing the target with high precision according to claim 4, wherein the north angle θ of the radar is obtained as follows:
sa11 sets an initial north-bias angle θ ═ θ according to the radar attachment position orientation0E.g. due to north00, east-ward direction θ0＝90；
Sa12, the target drives along the lane line direction, and the GPS coordinates of the target at each moment are acquired;
sa13, displaying the target GPS coordinate track on a high-precision map, and checking whether the track runs in a lane line; if the vehicle just runs in the lane line, the north deflection angle of the radar is theta0(ii) a Otherwise, readjustθ，θ＝θ0And +/-delta theta, and circularly executing the operations a12 and a 13.
6. The method for sensing the target with high precision according to claim 2, wherein for the conversion from radar coordinates to GPS coordinates in the coordinate conversion in step S3, the radar coordinates and the GPS coordinate point pairs of a single point are obtained based on a single point calibration method for calibration, which is as follows:
sb1, calculating a rotation angle theta according to the radar coordinates and the GPS coordinates of the single point, and adopting the following steps;
sb11, acquiring a GPS coordinate of a radar mounting position;
sb12, calculating UTM coordinates (Xw0, Yw0) of the radar mounting position;
sb13, acquiring (x0, y0) coordinates and GPS coordinates of a certain target; (when the target travels in the radar detection area, the radar detection coordinates (x0, y0) can be read, and the GPS coordinates can be acquired by using RTK or the like)
Sb14, calculating UTM coordinates (Xw, Yw) of the radar target;
calculating the coordinates after rotation (x ', y'):
calculating a rotation angle theta:
sb2, converting the radar coordinates into GPS coordinates according to the rotation angle theta in the following steps;
sb21, calculating coordinates after rotation (x ', y') for arbitrary radar target (x, y) coordinates;
sb22, calculating UTM coordinates (Xw, Yw) of the radar target;
sb23, calculates GPS coordinates of the radar target.
7. The method for sensing the target with high precision according to claim 4 or 6, wherein for step S4 quadratic coordinate conversion, the method comprises converting GPS into radar coordinate method based on the radar partial north angle method and/or the single point calibration method;
the method for converting the GPS to the radar coordinate based on the radar north-offset angle method comprises the following steps:
sa41, acquiring GPS coordinates of a radar installation position and a north-bias angle theta of radar installation;
sa42, calculating UTM coordinates (Xw0, Yw0) of the radar mounting position;
sa43, acquiring GPS coordinates of the radar target;
sa44, calculating UTM coordinates (Xw, Yw) of the radar target;
sa45, calculating coordinates (x ', y') after rotation;
calculating (x, y) coordinates of the radar target;
the method for converting the GPS to the radar coordinate based on the single-point calibration method comprises the following steps:
sa41, acquiring GPS coordinates of the radar mounting position and a rotation angle θ of the radar coordinate system and the UTM coordinate system;
sa42, calculating UTM coordinates (Xw0, Yw0) of the radar mounting position;
sa43, calculating corresponding UTM coordinates (Xw, Yw) for the GPS coordinates of any radar target;
sa44, calculating the coordinates (x ', y') after rotation,
sa44, calculating the (x, y) coordinates of the radar target,
8. the method for sensing the high precision of the target according to claim 7, wherein in the secondary coordinate transformation process of the step S4, the coordinate transformation is used for distance calculation, and the distance between the target and the main sensor is obtained for target fusion, and the method comprises the following steps;
erecting a sensor i, acquiring a GPS coordinate of an installation position of the sensor i, and calculating a UTM coordinate (Xwi, Ywi) of the point, wherein i is the number of sensors, and is 1, 2, 3;
calibrating a sensor and a GPS, acquiring a sensor coordinate and a GPS coordinate of a target, and calculating a UTM coordinate (Xwj, Ywj) of the target, wherein j is the number of the targets, and j is 1, 2, 3;
based on the above steps, the distances in the horizontal direction and the vertical direction between an arbitrary target and an arbitrary sensor are calculated:
wherein i is the number of sensors, i is 1, 2, 3; j is the target number, j 1, 2, 3.;
calculating the distance between any target and any target in the horizontal direction and the vertical direction:
j is the target number, j is 1, 2, 3., j is not j';
calculating the distance between any sensor and any sensor in the horizontal direction and the vertical direction:
i is the number of sensors, i ≠ i', 1, 2, 3.
9. The method for sensing the target with high precision according to claim 7, wherein the distance between the target and any sensor is acquired for sensor relay, and any plurality of sensors are calibrated simultaneously in a relay manner, and the method comprises the following specific steps:
erecting a plurality of sensors, wherein the visual fields of adjacent sensors are overlapped;
setting a main sensor A, and calibrating the main sensor A;
setting a sensor adjacent to the main sensor A as a sensor B, driving a target into a common detection range of the main sensor A and the sensor B, acquiring a GPS coordinate of the target through the main sensor A, and acquiring a sensor coordinate of the target under the sensor B through the main sensor B;
and calibrating the sensor B according to the GPS coordinate of the target under the main sensor A and the sensor coordinate under the sensor B, and circularly repeating the previous step and the current step until the calibration of all the sensors is realized.
10. The method for sensing the target with high precision according to claim 9, wherein the sensor relay takes GPS coordinates as a middle quantity to realize coordinate conversion between any sensors, including multi-source sensor calibration and homologous sensor calibration;
the multi-source sensor calibration, for example, converting the pixel coordinates (u, v) and the radar coordinates (x, y), firstly converting the pixel coordinates (u, v) into GPS coordinates, and then converting the GPS coordinates into the radar coordinates, so as to realize the conversion of the pixel coordinates into the radar coordinates;
and (3) calibrating the homologous sensor, for example, converting the radar A coordinate (x1, y1) and the radar B coordinate (x2, y2), firstly converting the radar A coordinate (x1, y1) into a GPS coordinate, and then converting the GPS coordinate into the radar B coordinate (x2, y2), so as to realize the conversion of the radar coordinate into the radar coordinate.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111293177.1A CN114035167A (en) | 2021-11-03 | 2021-11-03 | Target high-precision sensing method based on roadside multi-sensors |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111293177.1A CN114035167A (en) | 2021-11-03 | 2021-11-03 | Target high-precision sensing method based on roadside multi-sensors |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114035167A true CN114035167A (en) | 2022-02-11 |
Family
ID=80136093
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111293177.1A Pending CN114035167A (en) | 2021-11-03 | 2021-11-03 | Target high-precision sensing method based on roadside multi-sensors |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114035167A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114910928A (en) * | 2022-05-20 | 2022-08-16 | 山东高速建设管理集团有限公司 | Method and device for positioning vehicles in tunnel and storage medium |
-
2021
- 2021-11-03 CN CN202111293177.1A patent/CN114035167A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114910928A (en) * | 2022-05-20 | 2022-08-16 | 山东高速建设管理集团有限公司 | Method and device for positioning vehicles in tunnel and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110057373B (en) | Method, apparatus and computer storage medium for generating high-definition semantic map | |
| Brenner | Extraction of features from mobile laser scanning data for future driver assistance systems | |
| EP3637371A1 (en) | Map data correcting method and device | |
| US20180273031A1 (en) | Travel Control Method and Travel Control Apparatus | |
| CA2526105C (en) | Image display method and image display apparatus | |
| CN112189225B (en) | Lane line information detection apparatus, method, and computer-readable recording medium storing computer program programmed to execute the method | |
| CN110530372B (en) | Positioning method, path determining device, robot and storage medium | |
| WO2017208503A1 (en) | Map data update device, map data update method and map data update program | |
| EP3876189A1 (en) | Geographic object detection device, geographic object detection method, and geographic object detection program | |
| JP4596566B2 (en) | Self-vehicle information recognition device and self-vehicle information recognition method | |
| CN104428686A (en) | Method for determining a position of a vehicle, and a vehicle | |
| CN112683260A (en) | High-precision map and V2X-based integrated navigation positioning precision improving system and method | |
| CN112819711B (en) | Monocular vision-based vehicle reverse positioning method utilizing road lane line | |
| CN114323050A (en) | Vehicle positioning method and device and electronic equipment | |
| CN113296133B (en) | Device and method for realizing position calibration based on binocular vision measurement and high-precision positioning fusion technology | |
| CN113989766A (en) | Road edge detection method and road edge detection equipment applied to vehicle | |
| CN115027482A (en) | Fusion positioning method in intelligent driving | |
| CN116997771A (en) | Vehicle, positioning method, device, equipment and computer readable storage medium thereof | |
| CN116704458A (en) | Transverse positioning method for automatic driving commercial vehicle | |
| CN114035167A (en) | Target high-precision sensing method based on roadside multi-sensors | |
| CN210038170U (en) | Tightly-coupled automatic driving sensing system | |
| JP2010190647A (en) | Vehicle position measuring instrument and vehicle position measuring program | |
| CN216116064U (en) | Pose calibration system of heading machine | |
| CN111141252B (en) | Monocular calibration ranging method and system | |
| CN113960548A (en) | Calibration method for millimeter wave radar and GPS calibration |
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 |