Lane positioning method, device, vehicle and storage medium
By obtaining and matching lane lines in maps and perceived information and determining the positioning posture transformation relationship, the problem of difficult lane-level positioning accuracy in autonomous driving is solved, and the positioning accuracy and accuracy of autonomous driving are improved.
Patent Information
- Application Number
- CN202211371702.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-11-03
- Publication Date
- 2024-11-26
- Estimated Expiration
- 2042-11-03
AI Technical Summary
In the prior art, lane-level positioning accuracy is difficult to achieve in autonomous driving and assisted driving, especially in finding a balance between cost and accuracy.
By acquiring the first fork lane line in the map information and the second fork lane line in the perception information, the correspondence relationship and pose transformation relationship between the two are determined, and the current positioning information is corrected using accurate map information. Achieve lane-level positioning accuracy.
Improve lane-level positioning accuracy, enable lane-level positioning accuracy within an acceptable cost, and enhance the accuracy of autonomous driving and assisted driving.
Smart Images

Figure CN115655289B_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of artificial intelligence technology, and in particular to a method, device, vehicle and storage medium for locating lanes. Background Art
[0002] With the continuous development of electronic maps and vehicle navigation technology, autonomous driving technology has begun to enter people's daily lives. Autonomous driving technology requires higher positioning accuracy.
[0003] In the related technologies, the positioning of mobile platforms is mainly carried out by satellite positioning technology and / or vehicle-mounted sensor positioning technology. However, the positioning accuracy needs to be improved. Summary of the invention
[0004] In order to solve or partially solve the problems existing in the related art, the present application provides a method, device, vehicle and storage medium for locating a lane, which can improve the positioning accuracy.
[0005] On the one hand, the present application provides a method for locating a lane, including: obtaining at least one first forked lane line in map information, and obtaining at least one second forked lane line in perception information; determining the correspondence between at least one first forked lane line and at least one second forked lane line; determining a posture transformation relationship between the first forked lane line and the second forked lane line having the corresponding relationship; and processing current positioning information based on the posture transformation relationship to obtain lane positioning information.
[0006] In one embodiment, obtaining at least one first forked lane line in map information includes: for a current lane in the map information, obtaining a preceding lane and a succeeding lane of the current lane; if the number of preceding lanes and the number of succeeding lanes of the current lane are different, determining that the current lane line is a first forked lane line, and the first forked lane line has a fork point.
[0007] In one embodiment, if the number of preceding lanes and the number of succeeding lanes of the current lane are different, the current lane line is determined to be the first bifurcated lane line, including: if the number of preceding lanes is greater than 1, the bifurcation point between the lane lines of the current lane and the preceding lane is a merging point; if the number of succeeding lanes is greater than 1, the bifurcation point between the lane lines of the current lane and the succeeding lane is a divergence point.
[0008] In one embodiment, obtaining at least one second forked lane line in the perception information includes: determining at least one second forked lane line from the perception information based on constraints, where the constraints include: at least one of position information and posture information.
[0009] In one embodiment, at least one second diverging lane line is determined from the perception information based on constraints, including: determining at least two lane lines from the perception information; determining the distance between the at least two lane lines; if the distance is less than or equal to a preset distance threshold and there is a monotonic change trend, then determining that a second diverging lane line exists in the perception information.
[0010] In one embodiment, the method further includes: if the distance is greater than a preset distance threshold, pruning one lane line of the at least two lane lines.
[0011] In one embodiment, determining the correspondence between at least one first forked lane line and at least one second forked lane line includes: pairing the fork points of the first forked lane line and the second forked lane line based on the Euclidean distance between the fork points of the first forked lane line and the fork points of the second forked lane line to obtain paired fork points.
[0012] In one embodiment, the bifurcation points include diverging points and converging points. Accordingly, the above method further includes: if the paired bifurcation points include bifurcation points and converging points, then the bifurcation point that is next closest to the bifurcation point or converging point is selected for verification until a stop matching condition is reached, and the stop matching condition includes: the paired bifurcation points only include bifurcation points or converging points, and the distance between the paired bifurcation points of the same type exceeds a distance threshold.
[0013] In one embodiment, the first forked lane line includes at least three first lane lines, the second forked lane line includes at least three second lane lines, and the first lane line and the second lane line each include a plurality of points. Accordingly, determining the posture transformation relationship between the first forked lane line and the second forked lane line having a corresponding relationship includes: matching the at least three first lane lines one by one in the three second lane lines in a preset order to obtain at least three paired point sets of paired lane lines; adjusting the rotation parameter R and the translation parameter t between the first forked lane line and the second forked lane line so that the Euclidean distance and error between at least some paired points in the paired point set are minimized; if the Euclidean distance and error are less than or equal to a preset Euclidean distance threshold, determining the posture transformation relationship based on the rotation parameter R and the translation parameter t; if the Euclidean distance and error are greater than the preset Euclidean distance threshold, replacing the first forked lane line or the second forked lane line, and calculating the Euclidean distance and error between the updated forked lane lines until the stop condition is met.
[0014] In one embodiment, the Euclidean distance and error between the updated bifurcated lane lines are calculated until a stop condition is met, including: transforming the first bifurcated lane line or the second bifurcated lane line to an updated position according to a rotation parameter R and a translation parameter t; calculating the Euclidean distance and error between the updated bifurcated lane lines until the Euclidean distance and error are less than or equal to a preset Euclidean distance threshold, or all bifurcated lane lines are traversed.
[0015] In one embodiment, the method further includes: determining the lengths of at least three first lane lines and at least three second lane lines before matching the at least three first lane lines one by one in the three second lane lines in a preset order; obtaining at least three first lane lines and at least three second lane lines of consistent length; and performing equal-density interpolation on the at least three first lane lines and at least three second lane lines of consistent length to obtain a point set of lane lines.
[0016] In one embodiment, the above method also includes: if there is no second fork lane line in the perception information, processing the current positioning information based on the correspondence between at least part of the lane lines in the perception information and at least part of the lane lines in the map information to obtain lane positioning information.
[0017] On the other hand, the present application provides a device for locating lanes, including: a bifurcated line acquisition module, a correspondence determination module, a transformation relationship determination module, and a positioning update module. The bifurcated line acquisition module is used to obtain at least one first bifurcated lane line in the map information, and obtain at least one second bifurcated lane line in the perception information; the correspondence determination module is used to determine the correspondence between at least one first bifurcated lane line and at least one second bifurcated lane line; the transformation relationship determination module is used to determine the posture transformation relationship between the first bifurcated lane line and the second bifurcated lane line with the corresponding relationship; the positioning update module is used to process the current positioning information based on the posture transformation relationship to obtain lane positioning information.
[0018] On the other hand, the present application provides a vehicle, including: a vehicle body; a sensing device, arranged on the vehicle body, for obtaining sensing information; a processor, coupled to the sensing device, for receiving sensing information; a memory, on which executable code is stored, and when the executable code is executed by the processor, the processor executes the method as described above.
[0019] On the other hand, the present application provides a computer-readable storage medium having executable codes stored thereon. When the executable codes are executed by a processor of an electronic device, the processor executes the above method.
[0020] On the other hand, the present application provides a computer program product, including executable codes, which implement the above method when the executable codes are executed.
[0021] The technical solution provided by this application may have the following beneficial effects:
[0022] The technical solution of the embodiment of the present application determines the posture transformation relationship between the first forked lane line and the second forked lane line with a corresponding relationship after obtaining the first forked lane line in the corresponding map information and the second forked lane line in the perception information. Since the map information includes accurate position information and lane information, the current positioning information can be corrected based on the posture transformation relationship and the map position information to obtain lane-level positioning accuracy.
[0023] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS
[0024] The above and other objects, features and advantages of the present application will become more apparent by describing in more detail exemplary embodiments of the present application in conjunction with the accompanying drawings, wherein the same reference numerals generally represent the same components in the exemplary embodiments of the present application.
[0025] Figure 1 It is an exemplary system architecture of a method, device, vehicle and storage medium for locating lanes shown in an embodiment of the present application;
[0026] Figure 2 It is a schematic diagram of the structure of a vehicle and a vehicle-mounted sensor shown in an embodiment of the present application;
[0027] Figure 3 is a flow chart of a method for locating a lane shown in an embodiment of the present application;
[0028] Figure 4 is a schematic diagram of a bifurcated lane line shown in an embodiment of the present application;
[0029] Figure 5 is a schematic diagram of a leading lane and a following lane shown in an embodiment of the present application;
[0030] Figure 6 is a schematic diagram of a confluence point and a divergence point shown in an embodiment of the present application;
[0031] Figure 7 is a schematic diagram of the corresponding relationship between the first fork lane line and the second fork lane line shown in an embodiment of the present application;
[0032] Figure 8 is a schematic diagram of the position transformation relationship between the first forked lane line and the second forked lane line shown in an embodiment of the present application;
[0033] Fig. 9 is a schematic diagram of lane positioning information shown in an embodiment of the present application;
[0034] Fig.10 is a block diagram of a lane positioning device according to an embodiment of the present application; and
[0035] Fig.11 It is a block diagram of a vehicle according to an embodiment of the present application. DETAILED DESCRIPTION
[0036] The embodiments of the present application will be described in more detail below with reference to the accompanying drawings. Although the embodiments of the present application are shown in the accompanying drawings, it should be understood that the present application can be implemented in various forms and should not be limited by the embodiments described herein. On the contrary, these embodiments are provided to make the present application more thorough and complete, and to fully convey the scope of the present application to those skilled in the art.
[0037] The terms used in this application are for the purpose of describing specific embodiments only and are not intended to limit this application. The singular forms of "a", "an" and "the" used in this application and the appended claims are also intended to include plural forms unless the context clearly indicates other meanings. It should also be understood that the term "and / or" used herein refers to and includes any or all possible combinations of one or more associated listed items.
[0038] It should be understood that although the terms "first", "second", "third", etc. may be used in this application to describe various information, this information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of this application, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Thus, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of this application, the meaning of "multiple" is two or more, unless otherwise clearly and specifically defined.
[0039] With the rapid development of Advanced Driving Assistance System (ADAS) and unmanned driving technology, a variety of sensors are loaded on cars to perceive the surrounding environment of the car, such as lidar, millimeter wave radar, monocular / binocular cameras and satellite navigation.
[0040] For example, Laser Detecting and Ranging (Lidar) determines the distance by emitting and receiving laser beams, measuring the time difference and phase difference of laser signals, measuring the angle by horizontal rotation scanning, and establishing a two-dimensional polar coordinate system based on these two parameters, and then obtaining three-dimensional height information by obtaining signals at different pitch angles. High-frequency lasers can obtain a large amount of (e.g., 1.5 million) position point information (called point clouds) within one second, and perform three-dimensional modeling based on this information. Lidar has the advantages of high resolution, strong anti-active interference ability, good low-altitude detection performance, and light weight. As a large-scale environmental perception module, Lidar is now widely used in map creation, lane perception, etc. for autonomous driving. This makes it possible to obtain information that can be used for positioning from the perception information, such as lane line information.
[0041] The electronic map includes information such as lane width and actual geographical location of the lane. The location information of the mobile platform is mapped on the electronic map so that the user can clearly know the location information of the mobile platform. The location information of the mobile platform can be obtained through satellite positioning, inertial navigation positioning, etc.
[0042] For example, the Global Positioning System (GPS) system is used in traffic management, which can accurately reflect the real-time position and running trajectory of vehicles on the road network on the electronic map of the control center, so as to monitor various traffic information such as the flow direction, flow rate, flow rate, density, and road network load degree of the vehicle on the road network in real time. The GPS system is connected with the urban traffic signal control system, traffic geographic information system, and traffic intelligence information system to carry out real-time traffic signal control, traffic induction, and traffic flow organization optimization, so as to achieve the purpose of making full use of the road network, shortening vehicle travel time, reducing driving delays, reducing vehicle empty driving, ensuring driving safety, and improving road traffic capacity. The GPS system can also be used for vehicle positioning, command, dispatching, rescue and management for emergency tasks. The civilian positioning accuracy of the GPS system is 10 meters.
[0043] Inertial navigation positioning is mainly achieved by inertial measurement unit (IMU). IMU is a sensor mainly used to measure the movement of objects. The spatial movement of an object has 6 degrees of freedom, translation along three axes and rotation around three axes. IMU relies on the law of inertia for measurement. The measurement of translation is usually done by accelerometer, while the measurement of rotation is done by gyroscope. Inertial navigation positioning is mainly limited by the high price of high-precision IMU, and the accuracy and price of the parts show an exponential relationship. Low-precision IMU can only accurately position the mobile platform at a short distance. Inertial navigation is usually used to support other navigation systems, providing higher accuracy than using any single system.
[0044] For example, related technologies can combine GPS positioning and inertial navigation positioning to improve positioning accuracy. However, it is not easy to achieve lane-level positioning accuracy within a reasonable cost range. For example, a lane is about 2.5 meters wide, and lane-level positioning accuracy means that the deviation does not exceed 1.2 meters. Lane-level positioning helps to improve the level of assisted driving or autonomous driving.
[0045] Therefore, how to achieve lane-level positioning accuracy within an acceptable cost is a key issue that needs to be urgently addressed in the field of assisted driving and autonomous driving technology.
[0046] In the embodiment of the present application, at least one first forked lane line in the map information has significant features, and at least one second forked lane line in the perception information also has significant features. In this way, the first forked lane line and the second forked lane line can be accurately paired, and the current positioning information can be corrected using the accurate position information in the map to achieve lane-level positioning accuracy.
[0047] In addition, in certain embodiments of the present application, it is possible to conveniently determine whether there is a first fork lane line in the map based on the number of leading lanes and the number of succeeding lanes, and determine whether there is a second fork lane line in the perception information based on constraint conditions, so as to quickly pair the lane lines and calculate the posture transformation relationship, thereby achieving online improvement of positioning accuracy.
[0048] The following will be passed Figures 1 to 11 A method, device and electronic device for storing video in an embodiment of the present application are described in detail.
[0049] Figure 1 This is an exemplary system architecture of a method, device, vehicle, and storage medium that can be used to locate lanes in an embodiment of the present application. It should be noted that: Figure 1What is shown is merely an example of a system architecture to which the embodiments of the present application can be applied, in order to help those skilled in the art understand the technical content of the present application, but it does not mean that the embodiments of the present application cannot be used in other devices, systems, environments or scenarios.
[0050] See also Figure 1 According to the embodiment, the system architecture 100 may include mobile platforms 101, 102, 103, a network 104 and a cloud 105. The network 104 is used to provide a medium for communication links between the mobile platforms 101, 102, 103 and the cloud 105. The network 104 may include various connection types, such as wired or wireless communication links or optical fiber cables, etc. The mobile platforms 101 and 102 may be equipped with sensors, such as laser radars, millimeter wave radars, cameras, etc., to sense the surrounding environment information of the mobile platforms 101 and 102, such as lane line information.
[0051] Users can use mobile platforms 101, 102, and 103 to interact with other mobile platforms and cloud 105 through network 104 to receive or send information, etc. Such as sending model training requests, model parameter download requests, and receiving trained model parameters. The trained model can classify objects in image data or point cloud data, such as lane lines, obstacles, etc. Mobile platforms 101, 102, and 103 can be installed with various communication client applications, such as assisted driving applications, autonomous driving applications, car applications, web browser applications, database applications, search applications, instant messaging tools, email clients, social platform software, and other applications.
[0052] Mobile platforms 101 , 102 , 103 include, but are not limited to, electronic devices such as automobiles, robots, aircraft, tablet computers, and laptop computers.
[0053] The cloud 105 can receive model training requests, model parameter download requests, etc., adjust model parameters for model training, send model topology structures, send trained model parameters, etc., and can also send weather information, real-time traffic information, etc. to the mobile platforms 101, 102, 103. For example, the cloud 105 can be a background management server, a server cluster, a car network, etc.
[0054] It should be noted that the number of mobile platforms, networks and cloud servers is only illustrative, and any number of mobile platforms, networks and cloud servers may be provided as required.
[0055] Figure 2 It is a schematic diagram of the structure of a vehicle and a vehicle-mounted sensor shown in an embodiment of the present application.
[0056] See also Figure 2The vehicle 100 may include one or more of the following components: a vehicle body 110, a laser radar 120, and a camera 140. The laser radar 120 is disposed in the vehicle body 110. In addition, in order to improve the performance of assisted driving or automatic driving, the vehicle 100 may also be provided with an inertial measurement unit (IMU) 130, which is disposed in the vehicle body 110.
[0057] For example, the laser radar 120 can be installed on both sides of the front bumper. In addition, the laser radar can also be installed on the roof or above the front window of the vehicle, etc., which is not limited here. The laser radar 120 can include a pulse laser radar and a continuous wave laser radar, and can use ultraviolet light, visible light or near infrared light as a light source.
[0058] For example, the detection distance of the laser radar 120 set up in the front and side dual stations can reach 150 meters (such as for a target with a reflectivity of 10%), with a horizontal 120° field of view (FOV), and the field of view is increased to 150°. If calculated according to the point cloud density of 144 lines, the angular resolution can reach 0.16°, and the central area refresh rate is 20Hz, which can realize the detection of distant obstacles in scenes such as highways and urban roads, and improve the driving safety of assisted driving.
[0059] In addition, one or more ultrasonic sensors, millimeter wave radars, etc. may be provided on the vehicle 100 to improve the comprehensiveness and accuracy of the perceived vehicle environment information. For example, ultrasonic sensors may be provided inside the front and rear bumpers. For example, millimeter wave radars may be provided inside the front and rear bumpers.
[0060] For example, the shooting device 140 includes a camera, which images the light signal collected by the camera on the photosensitive device and transmits it to the image processor. In this way, the lane line can be obtained by image recognition. In order to ensure that a clear picture can be taken during driving, the camera pixels include but are not limited to: 300,000, 1 million, 2 million, 5 million or 8 million. The presented image quality resolution includes 360P, 720P, 1080P, 2K and even 4K. For different lighting environments, clear shooting in dim light, weak light or even backlight conditions can be achieved by mode adjustment or fill light. At the same time, the high frame rate function also makes the pictures taken under high-speed driving conditions still reliable.
[0061] Figure 3 It is a flow chart of a method for locating a lane shown in one embodiment of the present application.
[0062] See also Figure 3 The method for locating a lane may include operations S310 to S340.
[0063] In operation S310, at least one first forked lane line in the map information is obtained, and at least one second forked lane line in the perception information is obtained.
[0064] In some embodiments, at least one first bifurcated lane line can be obtained from the map information by image recognition, rule analysis, attribute information reading, etc. For example, a trained lane line recognition model can be used to process a map image to obtain a bifurcated lane line. For example, a bifurcated lane line can be determined based on a change in the number of lanes. For example, if there is attribute information of a bifurcated lane line or a change in the number of lanes in the map, the relevant attribute information can be directly read to obtain the first bifurcated lane line.
[0065] In addition, it can be obtained in a preset period, preset range, etc. For example, the preset period includes but is not limited to: 10 milliseconds, 100 milliseconds, 1 second, 3 seconds, 5 seconds, 10 seconds or a long period, etc. The preset range includes but is not limited to: 1 meter, 3 meters, 5 meters, 10 meters, 50 meters or a larger range, etc. In this way, when the mobile platform travels for a certain length of time or a certain distance, a calculation is performed, which reduces the amount of calculation while meeting the positioning accuracy requirements.
[0066] In some embodiments, at least one second bifurcated lane line can be obtained from the perception information by image recognition, point cloud recognition, feature analysis, etc. For example, the perception information is collected by a camera, and the second bifurcated lane line can be obtained from the image by image recognition. For example, the perception information is collected by a laser radar, and the second bifurcated lane line can be obtained from the point cloud by point cloud recognition. Among them, image recognition and point cloud recognition can be based on an artificial intelligence network to perform bifurcated lane line recognition.
[0067] Taking a vehicle with assisted driving function as an example, the first laser radar and the second laser radar can be respectively set on both sides of the front of the mobile platform. By splicing the point cloud collected by the first laser radar and the point cloud collected by the second laser radar, the area scanned by the laser radar at one time can be increased and the blind spot can be reduced. Among them, the mobile platform can be a car, a robot, an aircraft or a ship. The mobile platform can also be configured with IMU, camera, millimeter wave radar, ultrasonic radar, etc. The mobile platform can analyze at least one of the collected point cloud data, image data and inertial navigation data to obtain the second cross lane line.
[0068] Figure 4 is a schematic diagram of a bifurcated lane line shown in an embodiment of the present application. Figure 4, the bifurcated lane line has at least the following features: the number of lanes before the bifurcation and the number of lanes after the bifurcation change, one lane line changes from the bifurcation point to two lane lines, or two lane lines change from the bifurcation point to one lane line, etc. In this way, the bifurcated lane line can be identified based on these features.
[0069] In operation S320, a correspondence between at least one first forked lane line and at least one second forked lane line is determined.
[0070] There should be a one-to-one correspondence between the first forked lane line in the map and the second forked lane line in the perception information of the real world. After the first forked lane line and the second forked lane line are collected in operation S310, the correspondence between the two can be determined by matching.
[0071] In operation S330, a posture transformation relationship between a first forked lane line and a second forked lane line having a corresponding relationship is determined.
[0072] In this embodiment, the first forked lane line in the map information is in the first posture, and the second forked lane line in the perception information of the mobile platform is in the second posture. Since the first forked lane line and the second forked lane line with a corresponding relationship should be the same forked lane line and have the same posture, the first posture and the second posture should be the same. At this time, the posture transformation relationship between the first forked lane line and the second forked lane line with a corresponding relationship can be calculated so that the posture related information of the perception information can be corrected based on the posture transformation relationship. Among them, the posture related information can be characterized by posture parameters, and the posture parameters include but are not limited to: at least one of position parameters and posture parameters.
[0073] For example, posture parameters include but are not limited to: roll angle, pitch angle, yaw angle, at least one of X-axis displacement, Y-axis displacement or Z-axis displacement. For example, the yaw angle of the vehicle in the world coordinate system can be defined as follows: the angle between the vehicle's driving direction in the horizontal plane and the straight line where the preset direction (X-axis) in the world coordinate system is located, that is, the angle between the X-axis in the vehicle coordinate system and the XOZ plane in the world coordinate system (the Z-axis is the vertical upward direction, and O is the origin of the coordinate system). The pitch angle can be the angle between the straight line where the X-axis in the vehicle coordinate system is located and the XOY plane in the world coordinate system (the X-axis, Y-axis and Z-axis are orthogonal). The roll angle can be the angle between the Y-axis in the vehicle coordinate system and the XOY plane in the world coordinate system. If there is no angular deviation of the vehicle, that is, the vehicle coordinate system OX ′ Y ′ Z ′If the vehicle's yaw angle, pitch angle, and roll angle coincide with the world coordinate system OXYZ, the vehicle's corresponding yaw angle, pitch angle, and roll angle should all be zero. It should be noted that the above description of the X-axis, Y-axis, Z-axis, yaw angle, pitch angle, and roll angle is only for illustrative purposes and cannot be construed as a limitation on the present application.
[0074] In operation S340, the current positioning information is processed based on the posture transformation relationship to obtain lane positioning information.
[0075] In this embodiment, the position-related information of the perception information may not be accurate, while the position-related information in the map information is accurate. If the forked lane line is matched successfully, the position-related information of the perception information can be corrected based on the position-related information of the forked lane line in the matched map information to achieve lane-level positioning accuracy.
[0076] In some embodiments, obtaining at least one first bifurcated lane line in the map information may include the following operations.
[0077] First, for a current lane in the map information, a preceding lane and a succeeding lane of the current lane are obtained.
[0078] Figure 5 is a schematic diagram of a leading lane and a following lane shown in an embodiment of the present application. Figure 5 , along the direction of the car's travel, the road section that has been traveled includes the front driving lane, and the road section that may be traveled in the future includes the subsequent lane.
[0079] Then, if the number of preceding lanes and the number of succeeding lanes of the current lane in the map information are different, the current lane line is determined to be a first bifurcated lane line, and the first bifurcated lane line has a bifurcation point.
[0080] In some embodiments, if the number of preceding lanes and the number of succeeding lanes of the current lane are different, determining that the current lane marking is the first bifurcated lane marking may include the following operations.
[0081] For example, if the number of the preceding lane is greater than 1, the divergence point between the lane lines of the current lane and the preceding lane is the merging point.
[0082] For example, if the number of the subsequent lanes is greater than 1, the divergence point between the lane lines of the current lane and the subsequent lane is the divergence point.
[0083] See also Figure 5 , the leading lane includes one lane and the succeeding lane includes two lanes, which indicates that there must be a forked lane line, and the intersection of the forked lane lines is the fork point. Figure 5The number of subsequent lanes corresponding to the bifurcation point is shown to be 2, including lane 2 and lane 3. The number of leading lanes corresponding to the bifurcation point is 1, including lane 1. The bifurcation point between the lane lines of the current lane and the subsequent lane is a divergence point.
[0084] In a specific embodiment, the road information through which the driving trajectory passes is searched in the map information, and the lane lines participating in the diversion function and / or merging function are determined through the road diversion information and / or road merging information in the map information. For example, for each road in the map information, its predecessor road and successor road can be obtained. If the number of predecessor roads is greater than 1, the road belongs to a merging road, and the Y-shaped intersection of its lane line and the lane line of the predecessor road is the merging point. If the number of successor roads is greater than 1, the road belongs to a diversion road, including the first fork lane line. Specifically, the Y-shaped intersection of the lane line of the diversion road and the lane line of the successor road is the diversion point. In this way, for the roads passed by the vehicle in the map information, all relevant diversion points and / or merging points can be found. In addition, for each "Y"-shaped bifurcation point, its associated left fork, right fork and middle fork three map lane lines can be further recorded. In the above manner, the first fork lane line can be determined online in real time on the basis of consuming less computing resources.
[0085] In some embodiments, obtaining at least one second fork lane line in the perception information may include the following operations: determining at least one second fork lane line from the perception information based on constraints, where the constraints include: at least one of position information and posture information.
[0086] The sensing information is as above, including but not limited to: at least one of point cloud data, image data, and inertial navigation data. The second bifurcated lane line can be obtained by processing the sensing information. For example, it can be determined whether the current lane line is a bifurcated lane line based on the shape features, angle features, position features, etc. of the bifurcated lane line.
[0087] In some embodiments, determining at least one second bifurcated lane line from the perception information based on the constraint condition may include the following operations.
[0088] First, at least two lane lines are determined from the perception information, for example, by using a trained lane line recognition model to recognize lane lines from point cloud data and / or image data.
[0089] For example, the classification model used to determine whether the target object is a lane line can be a pre-trained model that can determine whether the target object in the input point cloud data is a lane line type. Classification models include but are not limited to: Linear Regression, Logistic Regression, Decision Trees, Naive Bayes, K-Nearest Neighbors, Support Vector Machines, Random Decision Forests or Bagging, Neural Networks, etc.
[0090] The classification model may use a deep neural network, etc. For example, the classification model may perform feature extraction on the input point cloud data to obtain a feature map. Then, based on the feature map or the feature map and other features, the point cloud data is classified. For another example, the classification model may perform feature extraction on each frame in the input point cloud data to obtain a feature map, then determine a candidate box corresponding to the target object from the feature map, then obtain feature data of the target object from the feature map based on the candidate box, and then determine the category of the candidate box based on the feature data of the target object.
[0091] Then, the distance between at least two lane lines is determined. The distance between lane lines can be calculated by a variety of related technologies. For example, based on the distance of a point in one lane line relative to another lane line. In addition, the weighted sum of multiple distances can also be calculated.
[0092] Then, if the distance is less than or equal to the preset distance threshold and there is a monotonic change trend, it is determined that there is a second bifurcated lane line in the perception information. Figure 5 , the distance between the bifurcated lane lines that are closer to the bifurcation point is smaller, and there is a trend of monotonic change, such as monotonic increase or monotonic decrease.
[0093] In addition, in order to reduce the amount of calculation, the identified lane lines can be processed to delete lane lines that cannot be bifurcated lane lines. Specifically, the above method can also include the following operation: if the distance is greater than a preset distance threshold, pruning one of the at least two lane lines.
[0094] In a specific embodiment, in the perception information, the lane lines participating in the divergence function and / or merging function are determined by constraints such as position and angle. The distance from each point on the perceived lane line to other perceived lane lines is calculated respectively. If the distance is less than a certain threshold and has a trend from small to large or from large to small, then these two lane lines may be bifurcated lane lines and are recorded. If the distance is greater than a certain threshold, then these two lane lines are non-adjacent lane lines and will not have any intersections, so they are pruned. In addition, for each "Y"-shaped bifurcation point, the three perceived lane lines of the associated left fork, right fork, and middle fork are recorded. In the above manner, the second bifurcated lane line can be determined online in real time on the basis of consuming less computing resources.
[0095] In some embodiments, the above-mentioned determination of the correspondence between at least one first forked lane line and at least one second forked lane line may include the following operations: based on the Euclidean distance between the fork point of the first forked lane line and the fork point of the second forked lane line, pairing the fork points of the first forked lane line and the second forked lane line to obtain paired fork points.
[0096] In this embodiment, since the first forked lane line in the paired map information and the second forked lane line in the perception information are the same lane line, the Euclidean distance between the two should be small. Therefore, the first forked lane line in the map information and the second forked lane line in the perception information can be paired based on the Euclidean distance between the two lane lines. The Euclidean distance calculation formula can be expressed as: sqrt((x1-x2)^2+(y1-y2)^2).
[0097] In some special scenarios, when there are multiple bifurcation points and the distances between the bifurcation points are close, there is a risk of pairing errors. In order to reduce this risk, the accuracy of pairing can be further improved with the help of other attribute information of the bifurcation points. For example, bifurcation points are divided into two types: diverging points and converging points. A bifurcation point can only be paired with another bifurcation point, but not with a converging point.
[0098] Figure 6 Schematic diagram of a junction and a divergence point shown in an embodiment of the present application. Figure 6 , including at least three bifurcation points 61, 62, and 63. Among them, the three bifurcation points 61, 62, and 63 can be divided into one confluence point 61 and two divergence points 62 and 63. The confluence point 61 and the divergence point 62 are close to each other, and there is a possibility that the distance between the two changes monotonically, resulting in the risk of misjudgment. If the attributes of the bifurcation points are combined: the confluence point 61 and the divergence point 62 are different, the possibility of incorrect pairing of the confluence point 61 and the divergence point 62 can be ruled out.
[0099] Specifically, the bifurcation point includes a divergence point and a confluence point. Accordingly, the above method may further include the following operations.
[0100] If the paired bifurcation points include diverging points and converging points, the bifurcation point that is next closest to the bifurcation point or converging point is selected for verification until the stop matching condition is met. The stop matching condition includes: the paired bifurcation points only include diverging points or converging points, and the distance between the paired bifurcation points of the same type exceeds the distance threshold.
[0101] In a specific embodiment, the map and perceived diverging points and / or converging points that are close in distance are combined through a nearest neighbor search, and a one-to-one correspondence is obtained through conditions such as type. Specifically, the diverging points / converging points in the searched map information and the diverging points / converging points in the perceived information can be matched. The matching principle includes calculating the Euclidean distance between the points of the lane lines near the diverging points / converging points in the perceived information and the diverging points / converging points in the map information, and pairing the two groups of diverging / converging points that are close in distance.
[0102] For example, in operation S310, the attributes of the bifurcation points in the first bifurcated lane line and the second bifurcated lane line, that is, whether they are "divergence points" or "merging points", are labeled. In the matching stage, if the points in the perception and map that are closer have the "divergence" attribute and the "merging" attribute, the match is rejected and the next closest intersection is selected to check whether its attributes correspond.
[0103] Figure 7 It is a schematic diagram of the correspondence between the first fork lane line and the second fork lane line shown in an embodiment of the present application.
[0104] See also Figure 7 , the solid lines show two lane lines 71 and 73 corresponding to a lane in the map information. The dotted lines show two lane lines 72 and 74 corresponding to a lane in the perception information. Among them, there is a correspondence between lane line 71 in the map information and lane line 72 in the perception information, and there is a correspondence between lane line 73 in the map information and lane line 74 in the perception information. However, the distance between lane line 72 and lane line 73 is close, and lane line 72 and lane line 73 may be mistakenly paired. If the attribute of the bifurcation point of lane line 72 is a divergence point, and the attribute of the bifurcation point of lane line 73 is a confluence point, the possibility of pairing between lane line 72 and lane line 73 can be ruled out, effectively improving the accuracy of pairing.
[0105] In some embodiments, the first bifurcated lane line includes at least three first lane lines, the second bifurcated lane line includes at least three second lane lines, and the first lane line and the second lane line each include a plurality of points. Figure 7 , each bifurcated lane line consists of three lane lines.
[0106] Accordingly, determining the posture transformation relationship between the first forked lane line and the second forked lane line having a corresponding relationship may include the following operations.
[0107] First, at least three first lane lines are matched one by one with three second lane lines in a preset order to obtain at least three paired point sets of paired lane lines.
[0108] Then, the rotation parameter R and the translation parameter t between the first bifurcated lane line and the second bifurcated lane line are adjusted so that the Euclidean distance and the error between at least some of the paired points in the paired point set are minimized.
[0109] If the Euclidean distance and error are less than or equal to the preset Euclidean distance threshold, the posture transformation relationship is determined based on the rotation parameter R and the translation parameter t.
[0110] If the Euclidean distance and error are greater than the preset Euclidean distance threshold, the first fork lane line or the second fork lane line is replaced, and the Euclidean distance and error between the updated fork lane lines are calculated until the stop condition is met.
[0111] Figure 8 It is a schematic diagram of the posture transformation relationship between the first fork lane line and the second fork lane line shown in an embodiment of the present application.
[0112] See also Figure 8 , the left side shows the perception information, which is the information in the perception coordinate system. The right side shows the map information, which is the information in the map coordinate system. The two coordinate systems represent the map information coordinate system and the perception information coordinate system respectively. Figure 8 It can be seen that the position conversion relationship between the map information coordinate system and the perception information coordinate system can be expressed as a corresponding relationship (R, t), where R can represent a rotation matrix and t can represent a translation matrix.
[0113] Specifically, the pose transformation relationship can be calculated in the following way. First, the two point clouds are decentralized, such as first finding their respective center coordinates, and using each point in the two point clouds to subtract the respective point cloud geometric center coordinates to form two new point clouds. Then, the two new point clouds are multiplied (one of them is transposed first) to obtain the matrix W. Perform singular value decomposition (SVD decomposition) on the matrix W to obtain the matrix U and the matrix V. In this way, the rotation matrix R can be obtained based on the matrix U and the matrix V, and the translation matrix t can be obtained by calculating the center coordinates of the two point clouds. The rotation matrix R and the translation matrix t can be used to represent the pose transformation relationship.
[0114] In addition, other point cloud registration methods can be used to determine the pose conversion relationship between the two coordinate systems. For example, in addition to using the Iterative Closest Point (ICP) method for registration, point cloud registration can also be performed using generalized ICP (GICP) and normal distributions transform (NDT) algorithms.
[0115] The pose of the mobile platform in the perception information can be obtained through point cloud computing. For example, each laser radar can continuously collect multiple frames of point clouds. Taking a laser radar as an example, in the coordinate system of the laser radar, the coordinates of the point clouds for the same object in the multiple frames of point clouds collected by the laser radar in a stationary state should be the same. When the pose of the laser radar changes, the coordinates of the point clouds for the same object in different point cloud frames will change. Based on this, the current posture of the mobile platform can be determined. Specifically, the pose between the mobile platform (laser radar) and an object can be determined as follows: for at least some frames in the multi-frame point cloud collected by the laser radar, the pose differences of adjacent frames between the start frame and the end frame are accumulated to obtain pose information.
[0116] In some embodiments, when the vehicle turns, etc., the accuracy of the point cloud data deteriorates, and motion compensation can be performed on the point cloud to improve the accuracy of the point cloud and increase the probability of successful pairing.
[0117] In some embodiments, when two bifurcation points fail to pair, it is necessary to pair the other two bifurcation points. The related technology can retry the pairing with a preset initial value, but this will cause the previous calculation results to be completely abandoned, which may lead to a waste of computing resources. Considering that the bifurcation points can be tried to pair separately in a way such as from far to near, the previous calculation results have a certain reference value. In this embodiment, the initial value can be determined based on the previous calculation results.
[0118] Specifically, the above calculation of the updated Euclidean distance and error between the bifurcated lane lines until the stop condition is met may include the following operations. First, the first bifurcated lane line or the second bifurcated lane line is transformed to the updated position according to the rotation parameter R and the translation parameter t. Then, the Euclidean distance and error between the updated bifurcated lane lines are calculated until the Euclidean distance and error are less than or equal to the preset Euclidean distance threshold, or all bifurcated lane lines are traversed.
[0119] In some embodiments, in order to facilitate pairing and further reduce the amount of calculation, the above method may further include the following operations before matching the at least three first lane lines one by one in the three second lane lines according to a preset order.
[0120] First, lengths of at least three first lane lines and at least three second lane lines are determined.
[0121] Then, at least three first lane lines and at least three second lane lines of the same length are obtained.
[0122] Next, at least three first lane lines of the same length and at least three second lane lines are interpolated with equal density to obtain a point set of the lane lines. In the above manner, a first point set corresponding to the first lane line and a second point set corresponding to the second lane line can be obtained, and then paired.
[0123] For example, determining the posture conversion parameters between the first fork lane line and the second fork lane line based on the first point set and the second point set may include the following operations.
[0124] First, a K-dimensional tree (KD-Tree or KD tree for short) is constructed for the first point set. Among them, the KD tree is a tree data structure that stores instance points in a K-dimensional space for rapid retrieval. The KD tree is also called a binary tree, which represents a partition of the K-dimensional space. Constructing a KD tree is equivalent to continuously dividing the K-dimensional space with hyperplanes perpendicular to the coordinate axis to form a series of K-dimensional hyperrectangular regions. Each node of the KD tree corresponds to a K-dimensional hyperrectangular region. It should be noted that the K in the KD tree refers to the K-dimensional space.
[0125] Then, use the second point set to find multiple adjacent point pairs on the K-nearest tree, where the point distance between each pair of multiple adjacent point pairs is less than the preset distance threshold. After the KD tree is generated, you can find adjacent point pairs. For example, for a target point, first find the leaf node containing the target point in the KD tree. With the target point as the center and the distance from the target point to the leaf node sample instance as the radius, a hypersphere is obtained, and the nearest neighbor point must be inside this hypersphere. Then return to the parent node of the leaf node and check whether the hyperrectangle contained in another child node intersects with the hypersphere. If it intersects, go to this child node to find out if there is a closer neighbor. If so, update the nearest neighbor. If it does not intersect, it is simple. Just return to the parent node of the parent node and continue searching for the nearest neighbor in another subtree. When backtracking to the root node, the algorithm ends, and the nearest neighbor node saved at this time is the final nearest neighbor.
[0126] Next, the posture conversion parameters between the first forked lane line and the second forked lane line are determined based on multiple adjacent point pairs. For details, please refer to the following related content.
[0127] In a specific embodiment, ICP calculation may be performed on a set of perception information and lane lines of diverging / merging points in map information to obtain a set of posture transformation relationships, for example, including the following five operations.
[0128] 1. Calculate the lengths of the three lines at the "Y"-shaped intersections in the perception information and map information, cut off the long line based on the shorter line, and perform equal-density interpolation on the six cut lines.
[0129] 2. Find the corresponding lines on the map according to the order of left, middle and right for the three perceived lines, and pair the points on the lines according to the principle of nearest neighbors.
[0130] 3. Take the sum of the Euclidean distances of the matching point pairs as the error, and use the SVD method to solve the rotation R and translation t to minimize the error. If the error is less than a certain threshold at this time, execute step 5, otherwise execute step 4.
[0131] 4. Transform the sensing point to a new position according to R, t, and return to step 2 for a new round of calculation.
[0132] 5. The relative relationship between the perceived final posture and the initial posture is the posture transformation obtained by this group of divergence / convergence points.
[0133] In some embodiments, if a set of perception information includes multiple bifurcation points, they can be matched one by one according to the above method to obtain a set of posture transformation relationships, and the final adopted posture transformation relationship can be determined based on weighted average or other methods.
[0134] In addition, if a set of perception information includes multiple bifurcation points, the mobile platform can be accurately positioned in the correct lane. If the bifurcation point is far away from the mobile platform and more accurate positioning in the lane is not possible, other characteristic information of the lane can be used to assist positioning to improve positioning accuracy. Among them, other characteristic information of the lane includes but is not limited to: the number of lanes, the turning point of the lane, the type of lane line (single line, double line, dotted line or solid line, etc.), the arrangement of the lane line, lane markings (such as turn signs, stop lines, zebra crossings, etc.), etc.
[0135] Specifically, the above method may also include the following operation: if there is no second bifurcated lane line in the perception information, the current positioning information is processed based on the correspondence between at least part of the lane lines in the perception information and at least part of the lane lines in the map information to obtain lane positioning information.
[0136] In a specific embodiment, for a set of perception information, if there are multiple diverging / merging points, they are matched to obtain a set of posture transformation relationships. At this time, the vehicle will be accurately positioned in the correct lane, but because the bifurcation point may be far away from the vehicle, all lane lines can be used for ICP calculation to achieve higher precision positioning.
[0137] Fig. 9 It is a schematic diagram of lane positioning information shown in an embodiment of the present application.
[0138] See also Fig. 9 The positioning accuracy that can be obtained by using GPS positioning and other methods is as follows: Fig. 9 As shown in the dotted circle, it is impossible to determine which lane the vehicle is in, among lane 1, lane 2, and lane 3. In the embodiment of the present application, by matching and calculating the bifurcated lane line or the bifurcated lane line and other lane lines in the vehicle perception information with the bifurcated lane line or the bifurcated lane line and other lane lines in the map information, a posture transformation relationship can be obtained. For example, it can be determined that the second bifurcated lane line has an angle of about 30° with respect to the first bifurcated lane line and is located at the bifurcation point. Then, it can be determined that the vehicle is located at the illustrated position of lane 3, rather than the illustrated position of lane 2, thereby achieving lane-level positioning accuracy.
[0139] Another aspect of the present application also provides a block diagram of a laser radar calibration device.
[0140] Fig.10 It is a block diagram of a lane positioning device according to an embodiment of the present application.
[0141] See also Fig.10 The lane positioning device 1000 may include: a bifurcation line obtaining module 1010, a correspondence relationship determining module 1020, a transformation relationship determining module 1030 and a positioning updating module 1040.
[0142] The bifurcated lane obtaining module 1010 is used to obtain at least one first bifurcated lane line in the map information and to obtain at least one second bifurcated lane line in the perception information.
[0143] The correspondence determination module 1020 is used to determine the correspondence between at least one first forked lane line and at least one second forked lane line.
[0144] The transformation relationship determination module 1030 is used to determine the posture transformation relationship between the first fork lane line and the second fork lane line that have a corresponding relationship.
[0145] The positioning update module 1040 is used to process the current positioning information based on the posture transformation relationship to obtain lane positioning information.
[0146] Regarding the device in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated again here.
[0147] Another aspect of the present application also provides a vehicle.
[0148] Fig.11 It is a block diagram of a vehicle according to an embodiment of the present application.
[0149] See also Fig.11 The vehicle 1100 includes a vehicle body, a memory 1110 and a processor 1120. In addition, the vehicle 1100 may further include at least one of a laser radar 1130 and a camera 1140.
[0150] One or more laser radars 1130 may be provided. For example, at least two laser radars 1130 are respectively provided at at least two positions of the vehicle body, and are respectively used to collect point clouds. For example, one laser radar 1130 is provided on the left side of the front end of the vehicle body, and one laser radar 1130 is provided on the right side of the front end of the vehicle body. For example, one laser radar 1130 is provided on the left side of the front end of the vehicle body, one laser radar 1130 is provided on the right side of the front end of the vehicle body, and one laser radar 1130 is provided above the front windshield of the vehicle body, or provided above the roof.
[0151] The camera 1140 is disposed on the top of the vehicle body or the front end of the vehicle body. For example, the camera 1140 may include a body, a lens, and a photosensitive material. In addition, it may also include an aperture, a shutter, a distance measurement, a framing, a light metering, a film transport, a counting, a self-timer, a focus, a zoom, and other systems.
[0152] The memory 1110 may include various types of storage units, such as system memory, read-only memory (ROM), and permanent storage. Among them, ROM may store static data or instructions required by the processor 1120 or other modules of the computer. The permanent storage may be a readable and writable storage device. The permanent storage may be a non-volatile storage device that does not lose the stored instructions and data even after the computer is powered off. In some embodiments, the permanent storage device uses a large-capacity storage device (such as a magnetic or optical disk, flash memory) as a permanent storage device. In other embodiments, the permanent storage device may be a removable storage device (such as a floppy disk, optical drive). The system memory may be a readable and writable storage device or a volatile readable and writable storage device, such as a dynamic random access memory. The system memory may store some or all instructions and data required by the processor at run time. In addition, the memory 1110 may include any combination of computer-readable storage media, including various types of semiconductor memory chips (such as DRAM, SRAM, SDRAM, flash memory, programmable read-only memory), and disks and / or optical disks may also be used. In some embodiments, the memory 1110 may include a readable and / or writable removable storage device, such as a laser disc (CD), a read-only digital versatile disc (such as a DVD-ROM, a double-layer DVD-ROM), a read-only Blu-ray disc, an ultra-density optical disc, a flash memory card (such as an SD card, a mini SD card, a Micro-SD card, etc.), a magnetic floppy disk, etc. The computer-readable storage medium does not include carrier waves and transient electronic signals transmitted wirelessly or wired.
[0153] The processor 1120 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc. The processor is coupled to at least two laser radars, and is at least used to receive point clouds from the laser radars.
[0154] The memory 1110 stores executable codes, and when the executable codes are processed by the processor 1120 , the processor 1120 can execute part or all of the methods described above.
[0155] It is understandable that, unless otherwise specified, each functional unit / module in each embodiment of the present application may be integrated into one unit / module, or each unit / module may exist physically separately, or two or more units / modules may be integrated together. The above-mentioned integrated unit / module may be implemented in the form of hardware or in the form of a software program module.
[0156] If the integrated unit / module is implemented in the form of hardware, the hardware may be a digital circuit, an analog circuit, etc. The physical implementation of the hardware structure includes but is not limited to transistors, memristors, etc. If not otherwise specified, the artificial intelligence processor may be any appropriate hardware processor, such as a CPU, a GPU, an FPGA, a DSP, and an ASIC, etc. If not otherwise specified, the storage module may be any appropriate magnetic storage medium or magneto-optical storage medium, such as Resistive Random Access Memory (RRAM), Dynamic Random Access Memory (DRAM), Static Random-Access Memory (SRAM), Enhanced Dynamic Random Access Memory (EDRAM), High-Bandwidth Memory (HBM), or Hybrid Memory Cube (HMC), etc.
[0157] If the integrated unit / module is implemented in the form of a software program module and sold or used as an independent product, it can be stored in a computer-readable memory. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a memory, including instructions for enabling a computer device (which can be a personal computer, server or network device, etc.) to perform all or part of the steps of the methods of various embodiments of the present application. The aforementioned memory includes: U disk, read-only memory (Read-Only Memory, referred to as ROM), random access memory (Random Access Memory, referred to as RAM), mobile hard disk, disk or optical disk and other media that can store program codes.
[0158] In a possible implementation, a board is also disclosed, which includes a storage device, an interface device, a control device and the above-mentioned processor; wherein the processor is connected to the storage device, the control device and the interface device respectively; the storage device is used to store data; the interface device is used to realize data transmission between the processor and an external device; and the control device is used to monitor the status of the processor.
[0159] In addition, the method according to the present application may also be implemented as a computer program or a computer program product, which includes computer program code instructions for executing some or all of the steps in the above method of the present application.
[0160] Alternatively, the present application can also be implemented as a computer-readable storage medium (or non-transitory machine-readable storage medium or machine-readable storage medium) on which executable code (or computer program or computer instruction code) is stored. When the executable code (or computer program or computer instruction code) is executed by a processor of an electronic device (or server, etc.), the processor executes part or all of the steps of the above-mentioned method according to the present application.
[0161] The embodiments of the present application have been described above, and the above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and changes will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The selection of terms used herein is intended to best explain the principles of the embodiments, practical applications, or improvements to the technology in the market, or to enable other persons of ordinary skill in the art to understand the embodiments disclosed herein.
Claims
1. A method for locating a lane, characterized in that: include: Obtain at least one first forked lane line in the map information, and obtain at least one second forked lane line in the perception information; Determine a correspondence between at least one of the first forked lane lines and at least one of the second forked lane lines; Determine a position transformation relationship between a first forked lane line and a second forked lane line having a corresponding relationship; Processing the current positioning information based on the posture transformation relationship to obtain lane positioning information; The step of obtaining at least one first bifurcated lane line in the map information includes: For the current lane in the map information, a preceding lane and a succeeding lane of the current lane are obtained; the preceding lane is a lane included in a road section that the vehicle has traveled, and the succeeding lane is a lane included in a road section that the vehicle will travel in the future; If the number of preceding lanes and the number of succeeding lanes of the current lane are different, the current lane marking is determined to be the first bifurcated lane marking.
2. The method according to claim 1, characterized in that The first bifurcated lane line has a bifurcation point; if the number of preceding lanes and the number of succeeding lanes of the current lane are different, determining that the current lane line is the first bifurcated lane line comprises: If the number of the preceding lanes is greater than 1, the bifurcation point between the lane lines of the current lane and the preceding lane is the merging point; If the number of the subsequent lanes is greater than 1, the divergence point between the lane lines of the current lane and the subsequent lane is the divergence point.
3. The method according to claim 1, characterized in that The obtaining of at least one second bifurcated lane line in the perception information includes: At least one second bifurcated lane line is determined from the perception information based on a constraint condition, wherein the constraint condition includes at least one of position information and posture information.
4. The method according to claim 3, characterized in that The determining at least one second bifurcated lane line from the perception information based on the constraint condition comprises: Determining at least two lane lines from the perception information; determining a distance between at least two of the lane markings; If the distance is less than or equal to a preset distance threshold and there is a monotonic change trend, it is determined that the second bifurcated lane line exists in the perception information.
5. The method according to claim 4, characterized in that Also includes: If the distance is greater than the preset distance threshold, one lane line of the at least two lane lines is pruned.
6. The method according to claim 1, characterized in that The determining of the correspondence between at least one of the first forked lane lines and at least one of the second forked lane lines comprises: Based on the Euclidean distance between the bifurcation point of the first bifurcated lane line and the bifurcation point of the second bifurcated lane line, the bifurcation points of the first bifurcated lane line and the second bifurcated lane line are paired to obtain paired bifurcation points.
7. The method according to claim 6, characterized in that The bifurcation point includes a divergence point and a confluence point; The method further comprises: If the paired bifurcation points include diverging points and converging points, a bifurcation point that is next closest to the bifurcation point or the converging point is selected for verification until a stop matching condition is reached. The stop matching condition includes: the paired bifurcation points only include bifurcation points or converging points, and the distance between the paired bifurcation points of the same type exceeds a distance threshold.
8. The method according to claim 1, characterized in that The first forked lane line includes at least three first lane lines, the second forked lane line includes at least three second lane lines, and the first lane line and the second lane line each include a plurality of points; The step of determining the position transformation relationship between the first forked lane line and the second forked lane line having a corresponding relationship comprises: Matching at least three of the first lane lines one by one with the three second lane lines in a preset order to obtain at least three sets of paired lane line pairing point sets; Adjusting a rotation parameter R and a translation parameter t between the first bifurcated lane line and the second bifurcated lane line so that the Euclidean distance and the error between at least some of the paired points in the paired point set are minimized; If the Euclidean distance and error are less than or equal to a preset Euclidean distance threshold, determining the posture transformation relationship based on the rotation parameter R and the translation parameter t; If the Euclidean distance and error are greater than the preset Euclidean distance threshold, the first fork lane line or the second fork lane line is replaced, and the Euclidean distance and error between the updated fork lane lines are calculated until a stop condition is met.
9. The method according to claim 8, characterized in that The calculation of the updated Euclidean distance and error between the bifurcated lane lines until the stopping condition is met includes: Transforming the first forked lane line or the second forked lane line to an updated position according to the rotation parameter R and the translation parameter t; The updated Euclidean distance and error between the bifurcated lane lines are calculated until the Euclidean distance and error are less than or equal to the preset Euclidean distance threshold, or all bifurcated lane lines are traversed.
10. The method according to claim 8, characterized in that Before matching the at least three first lane lines one by one in a preset order among the three second lane lines, the method further includes: Determining the lengths of at least three of the first lane lines and at least three of the second lane lines; Obtain at least three first lane lines and at least three second lane lines of the same length; Perform equal-density interpolation on at least three of the first lane lines and at least three of the second lane lines of the same length to obtain a point set of the lane lines.
11. The method according to any one of claims 1 to 9, characterized in that: Also includes: If there is no second bifurcated lane line in the perception information, the current positioning information is processed based on the correspondence between at least part of the lane lines in the perception information and at least part of the lane lines in the map information to obtain the lane positioning information.
12. A device for locating a lane, characterized in that: include: A bifurcated lane obtaining module, used to obtain at least one first bifurcated lane line in the map information, and to obtain at least one second bifurcated lane line in the perception information; A correspondence determination module, used to determine the correspondence between at least one of the first forked lane lines and at least one of the second forked lane lines; A transformation relationship determination module, used to determine a posture transformation relationship between a first forked lane line and a second forked lane line having a corresponding relationship; A positioning update module, used for processing the current positioning information based on the posture transformation relationship to obtain lane positioning information; Among them, the bifurcated line acquisition module is used to obtain the preceding lane and the succeeding lane of the current lane for the current lane in the map information; the preceding lane is a lane included in the road section that the vehicle has traveled, and the succeeding lane is a lane included in the road section that the vehicle will travel in the future; if the number of preceding lanes and the number of succeeding lanes of the current lane are different, the current lane line is determined to be the first bifurcated lane line.
13. A vehicle, characterized in that: include: The main body of the vehicle; A sensing device, disposed on the vehicle body, for obtaining sensing information; a processor, coupled to the sensing device, and configured to receive the sensing information; A memory having executable codes stored thereon, which, when executed by the processor, causes the processor to execute the method according to any one of claims 1 to 11.
14. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores program code, which can be called by a processor to execute the method according to any one of claims 1 to 11.
Citation Information
Patent Citations
Vehicle positioning method, vehicle positioning apparatus and vehicle positioning system
CN107643086A
Road feature point extraction method and system
CN110361015A
Method for recognizing lane-level vehicle positioning information based on lidar map matching, recording medium and device for performing the method
KR1020180080828A
Lane boundary line recognition device and computer-readable storage medium storing program of recognizing lane boundary lines on roadway
US20160012298A1
KR20200120402A
