A positioning method, apparatus, electronic device, and storage medium

Abstract

This invention discloses a positioning method, apparatus, electronic device, and storage medium. The positioning method includes: acquiring environmental point cloud data and vehicle pose increments; correcting point cloud distortion in the environmental point cloud data based on the vehicle pose increments; determining environmental feature data from the environmental point cloud data; and determining the vehicle position within a preset high-precision map based on the environmental feature data. In this embodiment, by acquiring environmental point cloud data and vehicle pose increments, correcting point cloud distortion in the environmental point cloud data based on the vehicle pose increments, and determining the vehicle position within a preset high-precision map based on the environmental feature data of the point cloud data, accurate vehicle positioning is achieved, improving the user experience.

Description

Technical Field

[0001] This invention relates to the field of computer application technology, and in particular to a positioning method, device, electronic device, and storage medium. Background Technology

[0002] Localization technology is one of the foundational and core technologies for robotic applications such as autonomous driving, providing robots with position and orientation. Currently, most autonomous vehicles use cameras as their primary localization sensors, extracting information such as lampposts from visual images through convolutional neural networks to obtain visual features of the environment, and then reconstructing the depth information of visual feature points using 3D photogrammetry.

[0003] However, visual sensors are easily affected by lighting conditions, and monocular vision has a large depth recovery error, making it difficult to significantly improve the robustness and accuracy of positioning. Therefore, a method that can achieve vehicle positioning on high-precision maps without being affected by visual factors has become an urgent problem to be solved. Summary of the Invention

[0004] This invention provides a positioning method, device, electronic device, and storage medium to achieve accurate vehicle positioning on high-precision maps and improve the user experience.

[0005] According to one aspect of the present invention, a positioning method is provided, wherein the method includes:

[0006] Acquire environmental point cloud data and vehicle pose increments;

[0007] Point cloud distortion correction is performed on environmental point cloud data based on vehicle pose increments;

[0008] Determine environmental feature data from environmental point cloud data;

[0009] The vehicle's location is determined within a pre-set high-precision map based on environmental feature data.

[0010] According to another aspect of the present invention, a positioning device is provided, wherein the device comprises:

[0011] The data acquisition module is used to acquire environmental point cloud data and vehicle pose increments;

[0012] The distortion correction module is used to correct point cloud distortion in environmental point cloud data based on vehicle pose increments.

[0013] The feature extraction module is used to determine environmental feature data from environmental point cloud data;

[0014] The vehicle positioning module is used to determine the vehicle's location within a preset high-precision map based on environmental feature data.

[0015] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:

[0016] At least one processor; and

[0017] A memory that is communicatively connected to at least one processor; wherein,

[0018] The memory stores a computer program that can be executed by at least one processor, such that the at least one processor is able to perform a positioning method according to any embodiment of the present invention.

[0019] According to another aspect of the present invention, a computer-readable storage medium is provided, which stores computer instructions for causing a processor to execute a positioning method according to any embodiment of the present invention.

[0020] The technical solution of this invention acquires environmental point cloud data and vehicle pose increments, corrects point cloud distortion in the environmental point cloud data based on the vehicle pose increments to achieve accuracy of the point cloud image, determines environmental feature data in the point cloud data, and determines the vehicle position in a preset high-precision map according to the environmental feature data, thereby achieving precise vehicle positioning and improving the user experience.

[0021] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0022] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0023] Figure 1 This is a flowchart of a positioning method provided according to Embodiment 1 of the present invention;

[0024] Figure 2 This is a flowchart of a positioning method provided according to Embodiment 2 of the present invention;

[0025] Figure 3 This is a flowchart of determining environmental feature data from environmental point cloud data according to Embodiment 3 of the present invention;

[0026] Figure 4 This is a flowchart of a positioning method provided according to Embodiment 4 of the present invention;

[0027] Figure 5 This is a schematic diagram of a ground point cloud extraction principle provided in Embodiment 4 of the present invention;

[0028] Figure 6 This is an example diagram of the incident angle for point cloud data acquisition provided in Embodiment 4 of the present invention;

[0029] Figure 7 This is a schematic diagram of a positioning device according to Embodiment 5 of the present invention;

[0030] Figure 8 This is a schematic diagram of the structure of an electronic device that implements a positioning method according to an embodiment of the present invention. Detailed Implementation

[0031] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0032] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0033] Example 1

[0034] Figure 1 This is a flowchart of a positioning method according to Embodiment 1 of the present invention. This embodiment is applicable to vehicle positioning. The method can be executed by a positioning device, which can be implemented in hardware and / or software and can be configured in a vehicle. Figure 1 As shown, the method includes:

[0035] S110: Acquire environmental point cloud data and vehicle pose increments.

[0036] Point cloud data can refer to a set of vectors in a three-dimensional coordinate system. Point cloud data can be acquired by sensors such as three-dimensional imaging sensors and lidar sensors. Vehicle pose increment can refer to the displacement increment and angle increment of the vehicle over the previous moment. It is the change in vehicle displacement and attitude.

[0037] In this embodiment of the invention, sensors can be used to detect environmental point cloud data and vehicle pose increments. Sensors for detecting point cloud data can include lidar sensors or 3D laser scanners. Vehicle pose increments can be detected directly by continuously acquiring vehicle pose data using a pose sensor and determining the pose increment based on the change in pose; alternatively, vehicle displacement data and driving angle data can be obtained by sensors that collect displacement and angle changes, thereby determining the pose increment. In one embodiment, the detection of environmental point cloud data can include using a lidar sensor. The lidar sensor sends laser signals to objects in the environment, and the objects can reflect the laser information. The reflected laser can carry information such as the orientation and distance of the objects in the environment. The lidar sensor can scan the environment according to a certain trajectory to acquire environmental point cloud data. When acquiring environmental point cloud data using a 3D laser scanner, selected target objects in the environment can be scanned. The 3D laser scanner can quickly acquire the 3D coordinate data of the surface of the measured object over a large area with high resolution, thereby acquiring point cloud data.

[0038] S120. Correct point cloud distortion in environmental point cloud data based on vehicle pose increment.

[0039] Point cloud distortion correction refers to the process of correcting distorted point cloud data. Since the points in a single frame of point cloud are not collected simultaneously, and the sensor moves with the vehicle during acquisition, and because the sensor measures the distance between the object and the sensor, the coordinate systems of different laser points may be different. Therefore, it is necessary to unify the environmental point cloud data to the same coordinate system to align the point cloud data and improve positioning accuracy.

[0040] In this embodiment of the invention, the changes in environmental point cloud data within the coordinate system within the time corresponding to the pose increment are determined by the pose increment, thereby unifying the environmental point cloud data to the same coordinate system and achieving point cloud distortion correction. Point cloud distortion correction can be performed in various ways, including correction based on a neural network model using the pose increment, or correction using a transformation matrix. In one embodiment, when using a neural network model for point cloud distortion correction, a neural network model corresponding to the pose increment and point cloud distortion correction can be pre-constructed, and the environmental point cloud data can be corrected by inputting the vehicle pose increment. When using a transformation matrix for point cloud distortion correction, the transformation matrix can be determined by the pose increment, and the transformation matrix can be used to assist in point cloud distortion correction. For example, point cloud distortion correction can be achieved by calculating the product of the environmental point cloud data and the transformation matrix.

[0041] S130. Determine environmental feature data from environmental point cloud data.

[0042] Among them, environmental feature data can refer to special data that can be identified in the environment, including ground feature data, object feature data, etc.

[0043] In this embodiment of the invention, point cloud data may contain multiple location points, and the positional relationship between any two location points may include horizontal and vertical relationships. Different positional relationships can determine different environmental feature data. In one embodiment, any two location points in the point cloud data can be extracted by means of adjacent or random filtering, and the positional relationship between the two location points is determined by their coordinates in the coordinate system. Environmental data features are then determined based on the positional relationship. In one embodiment, when the positional relationship is horizontal, the environmental feature data can be determined as ground feature data; when the positional relationship is vertical, the environmental feature data can be determined as object feature data.

[0044] S140. Determine the vehicle's location within a preset high-precision map based on environmental feature data.

[0045] The pre-set high-precision map can refer to a map used by the autonomous driving system, and can be pre-built and stored in the vehicle. The high-precision map can store road semantic information in the form of vector information, which may include, but is not limited to, road information such as light poles and road signs, as well as road marking information such as lane lines.

[0046] In this embodiment of the invention, the high-precision map can store attribute information such as light poles, road signs, and lane lines. Based on environmental feature data, corresponding locations can be matched in the high-precision map, and these matched locations can be determined as the vehicle's position. The methods for matching environmental feature data to corresponding locations in the high-precision map can include one-to-one matching of environmental feature data within the high-precision map, or pre-setting particle filtering rules to determine which locations the environmental feature data can match in the high-precision map. In one embodiment, environmental feature data can be mapped one-to-one to environmental features in the high-precision map. By matching the environmental feature data, the environment surrounding the vehicle's location can be determined, thus realizing the determination of the vehicle's position in the preset high-precision map. Alternatively, particle filtering rules can be pre-set. By inputting the environmental feature data into the particle filter and matching it to the environment in the high-precision map, the vehicle's position can be determined in the high-precision map using the preset particle filtering rules.

[0047] In this embodiment of the invention, by acquiring environmental point cloud data and vehicle pose increments, point cloud distortion correction is performed on the environmental point cloud data based on the vehicle pose increments, environmental feature data is determined in the point cloud data, and the vehicle position is determined in a preset high-precision map according to the environmental feature data. This achieves accurate vehicle positioning by determining the vehicle position based on the environmental feature data, thereby improving the user experience.

[0048] Example 2

[0049] Figure 2 This is a flowchart of a positioning method according to Embodiment 2 of the present invention. This embodiment is a further refinement of the positioning method based on the above embodiments. Figure 2 As shown, the method includes:

[0050] S210. Start the lidar sensor to measure environmental point cloud data, and control the pose sensor to measure at least two pose data according to the threshold time interval.

[0051] In this context, a lidar sensor can be a sensor that emits laser signals and determines point cloud data based on the reflected laser signals; a pose sensor can be a sensor that measures the vehicle's position and driving angle. The threshold time interval refers to the interval at which the pose sensor collects pose data; the pose sensor can collect pose data once every threshold time interval. The pose data can refer to the vehicle's position and driving angle data at the current acquisition moment.

[0052] In this embodiment of the invention, the vehicle can activate a lidar sensor to detect point cloud data in the environment. The lidar sensor can emit laser signals into the environment and receive laser signals reflected back from the environment. The reflected laser signals can carry information such as object location and distance. By acquiring a large amount of data, environmental point cloud data can be obtained. The vehicle can also control a pose sensor to collect pose data, which can be measured at threshold time intervals. The threshold time interval can be a time set by the vehicle manufacturer based on experience. For example, the threshold time interval stored in the vehicle can be 10 milliseconds, 100 milliseconds, 1 second, etc., and the threshold time interval can also be a specific time range. Since the pose sensor acquires pose data according to the threshold time interval, multiple pose data sets can be acquired.

[0053] S220. Determine the displacement increment and angle increment of each pose data relative to the reference pose data as the vehicle pose increment, wherein the reference pose data is the first pose data collected within each pose data.

[0054] The reference position data can be used as the starting position in measuring displacement and angle increments. The reference position data can be pre-set; since the pose sensor collects multiple pose data sets, the first pose data set collected from each pose data set can be used as the reference pose data. The displacement increment refers to the increase in displacement relative to the reference pose data in each pose data set, representing the amount of movement of the vehicle in a straight line. The angle increment refers to the change in angle relative to the reference pose data in each pose data set, representing the change in the vehicle's direction of motion.

[0055] In this embodiment of the invention, each pose data can be compared and calculated with the reference pose data to determine the displacement increment and angle increment of each pose data relative to the reference pose data. The reference pose data and each pose data can be extracted and separated into displacement data and angle data respectively. The displacement data and angle data in each pose data are then compared and calculated with the displacement data and angle data in the reference data to obtain the displacement increment and angle increment, which are then used as the vehicle's displacement increment.

[0056] S230. Determine the correction transformation matrix according to the vehicle pose increment, and determine the product of the correction transformation matrix and the environmental point cloud data.

[0057] The correction transformation matrix can be a transformation matrix determined based on the vehicle pose increment, or it can be a transformation matrix used to correct auxiliary point cloud data. The correction transformation matrix can be composed of the relationship between pose increments.

[0058] In this embodiment of the invention, since the vehicle's pose increment can include displacement increment and angle increment, the correction transformation matrix can be determined based on the displacement increment and angle enhancement. Coordinate information from the environmental point cloud data can be extracted, and this coordinate information can also be treated as a matrix; the product of this environmental point cloud coordinate information matrix and the correction transformation matrix can be calculated.

[0059] S240. Use the product as the environmental point cloud data after point cloud distortion correction.

[0060] In this embodiment of the invention, the transformations for point cloud distortion correction in the graphic can include translation, rotation, scaling, and combinations thereof. If each location point is extracted from the point cloud data, the correction transformation for each location point is essentially multiplying that location point by the correction transformation matrix. In other words, the product of the correction transformation matrix and the environmental point cloud data can be used as the environmental point cloud data after point cloud distortion correction.

[0061] S250. Determine the positional relationship between two location points within the environmental point cloud data, wherein the two location points are two point cloud data points on adjacent laser beams with the same horizontal measurement angle.

[0062] The two location points can be two point cloud data points on adjacent laser beams measured by a lidar sensor at the same horizontal measurement angle, or they can be two points with different horizontal angles. The positional relationship between the two location points can include horizontal and vertical relationships.

[0063] In this embodiment of the invention, the positional relationship between two points within environmental point cloud data can be determined using the coordinates of the points. A lidar sensor can acquire data from multiple laser beams by measuring the same horizontal angle. Two point cloud data points from adjacent laser beams can be extracted, and their coordinates can be used to determine their positional relationship. In one embodiment, the coordinates of the two extracted points can be confirmed in a three-dimensional coordinate system. Connecting the two points, when the angle between the line connecting the two points and the horizontal plane is close to or equal to 0, the positional relationship between the two points can be determined to be horizontal. When the angle between the line connecting the two points and the horizontal plane is greater than a threshold angle, the positional relationship between the two points can be considered vertical. The threshold angle can be set by the manufacturer based on experience; for example, the threshold angles stored in the vehicle can include 88 degrees, 89 degrees, 90 degrees, etc.

[0064] S260. If the positional relationship is horizontal, then the two position points are determined as ground feature data in the environmental feature data.

[0065] Ground feature data can refer to data with identifiable features on the ground surface, and can include lane line data, etc.

[0066] In this embodiment of the invention, when the positional relationship between two location points is determined to be horizontal, it can be assumed that the two location points are located on the same horizontal plane. At this time, it can be determined that the two location points correspond to ground feature data in the environmental feature data.

[0067] S270. If the positional relationship is vertical, then the two position points are determined as object feature data in the environmental feature data.

[0068] Among them, object feature data can refer to object data with identifiable features. Object feature data can include feature data of any object that may obstruct vehicle movement, such as pillars, railings, and road signs.

[0069] In this embodiment of the invention, when the positional relationship between two location points is determined to be vertical, it can be assumed that the two location points are not located on the same horizontal plane, but have a certain height difference in the vertical direction. It can be determined that the two location points correspond to object feature data in environmental feature data.

[0070] S280. Collect the vehicle's prior pose at the current moment, and determine the initial local map within the preset high-precision map according to the vehicle's prior pose.

[0071] Among them, the vehicle prior pose can refer to the vehicle pose predicted by a high-frequency sensor; the initial local map can refer to the local map of the area where the vehicle is located in the preset high-precision map, and the local map can be matched by the vehicle prior pose.

[0072] In this embodiment of the invention, the vehicle's prior pose at the current moment can be acquired using a high-frequency sensor. The sensor acquiring the prior pose may include an inertial measurement unit (IMU) or a wheel speed sensor, etc. Since the result detected by the high-frequency sensor is ahead of the vehicle pose acquired by the lidar sensor, the vehicle's prior pose can be understood as a predicted value of the vehicle's pose. Based on the vehicle's prior pose, an initial local map of the vehicle's location can be determined within a preset high-precision map.

[0073] S290. Process environmental feature data and initial local map according to preset particle filtering processing rules to determine vehicle position.

[0074] The preset particle filtering processing rule can be to approximate the probability density function by using random samples propagating in the state space through particle filtering, and use the sample mean to replace the integral operation, thereby obtaining the minimum variance estimate of the system state, which can then be used to obtain the optimal estimate of the current state.

[0075] In this embodiment of the invention, point cloud data of environmental feature data can be initialized in the initial local map. By bringing the environmental feature data into the particle filter, corresponding to the environment in the initial local map, the vehicle position can be determined in the initial local map through preset particle filter processing rules.

[0076] In this embodiment of the invention, environmental point cloud data is acquired by a lidar sensor, and vehicle pose data is acquired by a pose sensor. The displacement increment of the vehicle is determined based on the displacement increment and angle increment of the pose data compared with the reference pose data. A correction transformation matrix is ​​determined based on the vehicle displacement increment, and the product of the correction transformation matrix and the environmental point cloud is used as the environmental point cloud data after point cloud distortion correction, thus ensuring the accuracy of point cloud distortion correction. Ground feature data and object feature data are determined by determining the positional relationship between any two points in the point cloud data, thus distinguishing the types of environmental feature data. The vehicle's prior pose at the current moment is collected, and an initial local map is determined. The environmental feature data and the initial local map are processed according to a preset particle filter processing rule to determine the vehicle's position, achieving accurate vehicle positioning and improving the user experience.

[0077] Example 3

[0078] Figure 3 This is a flowchart for determining environmental feature data from environmental point cloud data according to Embodiment 3 of the present invention. This embodiment is a further refinement of S130 based on the above embodiment.

[0079] like Figure 3 As shown, the method includes:

[0080] S310. Extract the coordinate information of two location points within the environmental point cloud data.

[0081] In this embodiment of the invention, since the environmental point cloud data is a set of vectors in a three-dimensional coordinate system, the coordinate information of any point in the environmental point cloud data can be read. It is possible to acquire two point cloud data points on adjacent laser beams at the same horizontal measurement angle measured by a lidar sensor, and then read the coordinate information of the point using a coordinate system.

[0082] S320. Determine the angle of inclination between the line connecting the two location points and the horizontal plane based on the coordinate information.

[0083] In this embodiment of the invention, any two position points can be connected, and the angle of inclination between the line connecting the two position points and the horizontal plane can be determined based on the coordinate information. In actual operation, the line connecting the two position points can intersect the horizontal plane. If the two position points are not on the same horizontal plane, there can be a position point that does not intersect the horizontal plane. A projection can be made onto the horizontal plane from this position point. Since the coordinate information of the two position points is already determined, the coordinate information of the projection point on the horizontal plane can be determined based on the two position points. The two position points and the projection point on the horizontal plane can form a right triangle, and the angle of inclination between the line connecting the two position points and the horizontal plane can be determined using trigonometric functions.

[0084] S330. When the angle of inclination is less than the first threshold angle, the positional relationship is determined to be horizontal.

[0085] The first threshold angle can refer to the maximum angle at which the positional relationship is determined to be horizontal. In other words, when the angle between the line connecting the two positional points and the horizontal plane is less than the first threshold angle, the positional relationship between the two positional points can be considered to be horizontal.

[0086] In this embodiment of the invention, the angle of inclination of the line connecting two location points to the horizontal plane can be compared with a first threshold angle. When the angle of inclination is less than the first threshold angle, the two location points can be considered to be essentially on the same horizontal plane, and thus the positional relationship between the two location points can be confirmed as horizontal. The first threshold angle can be an angle set by the vehicle manufacturer based on experience. For example, the first threshold angle stored in the vehicle can be 0 degrees, 1 degree, 2 degrees, etc., and the threshold time interval can also be a specific angular range. In one embodiment, when the angle of inclination of the line connecting two location points to the horizontal plane is less than 1 degree, the positional relationship between the two location points can be considered horizontal.

[0087] S340. When the angle of inclination is greater than the second threshold angle, the positional relationship is determined to be vertical.

[0088] The second threshold angle can refer to the minimum angle required to determine if the positional relationship is vertical. In other words, when the angle between the line connecting two positional points and the horizontal plane is greater than the second threshold angle, the positional relationship between the two positional points can be considered vertical.

[0089] In this embodiment of the invention, the tilt angle and a second threshold angle can be compared. When the tilt angle is greater than the second threshold angle, it can be considered that the two position points are not on the same horizontal plane, and thus the positional relationship between the two position points can be confirmed as vertical. The second threshold angle can be an angle set by the vehicle manufacturer based on experience. For example, the second threshold angle stored in the vehicle can be 88 degrees, 89 degrees, 90 degrees, etc., and the threshold time interval can also be a specific angular range. In one embodiment, when the tilt angle is greater than 89 degrees, the positional relationship between the two position points can be considered vertical.

[0090] S350. If the positional relationship is horizontal, then the two position points are determined as ground feature data in the environmental feature data.

[0091] S360. If the positional relationship is vertical, then the two position points are determined as object feature data in the environmental feature data.

[0092] In one embodiment, determining environmental feature data in environmental point cloud data further includes: determining the point cloud intensity value of each location point in the ground feature data; if the point cloud intensity value is less than an intensity threshold, then the location point is determined to be road surface feature data; if the point cloud intensity value is greater than or equal to the intensity threshold, then the location point is determined to be lane line feature data.

[0093] Among them, the point cloud intensity value can refer to the intensity of the returned signal recorded based on the reflectivity of the surface of an object in the environment; the intensity threshold can be the critical value that distinguishes and judges the point cloud intensity values ​​of road feature data and lane line feature data. The environmental feature data corresponding to the location point can be determined through the intensity threshold, and the intensity threshold can be a value set by the manufacturer based on experience.

[0094] In this embodiment, since the point cloud intensity values ​​of the road surface and lane lines are clearly distinguishable, road surface feature data and lane line feature data can be identified by setting intensity thresholds. The point cloud intensity value of each location point can be obtained by extracting point cloud information, which can be collected by sensors when acquiring point cloud information. The vehicle can extract the intensity threshold stored locally and compare the point cloud intensity value of each location point with the intensity threshold. When the point cloud intensity value is less than the intensity threshold, the location point can be considered as road surface feature data; when the point cloud intensity value is greater than or equal to the intensity threshold, the location point can be considered as lane line feature data.

[0095] In this embodiment of the invention, by extracting the coordinate information of any two location points from environmental point cloud data, determining the angle of inclination between the line connecting the two location points and the horizontal plane according to the coordinate information, determining the positional relationship between the two location points according to the angle of inclination between the line connecting the two location points and the horizontal plane, and then determining the ground feature data and object feature data in the environmental feature data, the invention achieves accurate determination of environmental feature data in environmental point cloud data.

[0096] In one embodiment, determining the intensity threshold includes: determining the intensity threshold of the ground feature data according to a preset maximum inter-class variance rule.

[0097] The preset maximum inter-class variance rule is an algorithm for determining the image binarization segmentation threshold, and it is an adaptive threshold determination method. In this embodiment of the invention, according to the preset maximum inter-class variance rule, the image can be divided into two parts: road surface and lane lines, based on the image's grayscale characteristics. When the optimal intensity threshold is taken, the difference between the two parts should be the largest. The optimal intensity threshold can be calculated based on the image's grayscale characteristics and the preset maximum inter-class variance rule.

[0098] In one embodiment, lane line extraction is based on differences in point cloud reflection intensity. Point cloud intensity is affected not only by different materials but also by the incident angle and distance. Assuming the incident angle between the object surface and the laser beam is α, the laser measurement of the lidar is modeled according to the following formula, which quantifies the values ​​affecting point cloud intensity.

[0099]

[0100] Where: I is the intensity value of the lidar; P e P represents the power of the emitted laser beam. r D represents the energy value of radar intensity. r η is the aperture of the receiver; sys η is the system emission factor; atm ρ is the atmospheric emissivity; α is the incident angle between the object surface and the laser beam; ρ is the reflectivity of the object material; and R is the measured distance.

[0101] In one embodiment, determining environmental feature data from environmental point cloud data further includes: performing clustering operations on object feature data and obtaining at least one clustered point cloud cluster; generating geometric constraint verification results for each clustered point cloud cluster; and determining the clustered point cloud cluster as at least one of column feature data, railing feature data, and road sign feature data according to the geometric constraint verification results.

[0102] Clustering operations refer to dividing object feature data into different classes or clusters according to specific criteria (such as distance), making data objects within the same cluster more similar, while identifying data objects in different clusters that are highly dissimilar. Clustering point cloud clusters can be obtained by performing clustering operations on point cloud data. Geometric constraint verification results refer to the verification results after applying geometric constraints to the positional relationships of different parts of the object feature data.

[0103] In this embodiment of the invention, object feature data can be segmented into different clusters of point clouds, and geometric constraint verification results for each cluster of point clouds can be generated. In one embodiment, the geometric constraint verification results can be obtained by analyzing the eigenvalues ​​of each cluster of point clouds using Principal Component Analysis (PCA). Based on the eigenvalues, the object can be identified as a pillar, road sign, railing, etc. In one embodiment, point cloud clusters extracted using geometric constraint verification for pillars can be identified as pillar feature data; point cloud clusters extracted using geometric constraint verification for road signs can be identified as road sign feature data; and point cloud clusters extracted using geometric constraint verification for railings can be identified as railing feature data. In one embodiment, the extraction of road signs can also involve point cloud intensity verification and plane normal vector calculation for planar point cloud clusters. Point cloud clusters whose plane normal vectors are consistent with the vehicle's direction of travel and whose point cloud intensity values ​​are greater than an intensity threshold are considered as road signs.

[0104] Example 4

[0105] Figure 4 This is a flowchart of a positioning method provided according to Embodiment 4 of the present invention. This embodiment is a specific embodiment based on the above embodiments. Figure 4 As shown, the method includes:

[0106] S4010: Obtain point cloud data from the environment.

[0107] S4020, Obtain vehicle pose increment.

[0108] In this embodiment, the vehicle's pose increment at each moment can be calculated using sensors such as those carrying the vehicle, which can detect speed and acceleration. The pose increment can include increments in displacement and angle.

[0109] S4030: Correct point cloud distortion in the point cloud data in the environment based on the vehicle pose increment.

[0110] In this embodiment, the lidar sensor, as a time-of-flight ranging sensor, assumes that the vehicle is stationary while in motion. However, when the vehicle is moving at high speed, this static assumption no longer holds, causing the lidar sensor itself to change its pose during the measurement process. This pose change will be clearly reflected in the measured data. To reduce the measurement distortion of the point cloud caused by vehicle motion, point cloud distortion correction can be performed based on the pose increment obtained by the lidar during a single measurement.

[0111] The principle is shown in the following formula: when the lidar starts measuring in the current k-th frame, the time t is obtained. k and the vehicle's position T k Then, based on the time t of each radar measurement pointk+i Obtain the vehicle relative to t k The pose increment ΔT at time step P is then used to measure the current measurement point P. i Based on the pose increment, convert to t k time.

[0112] P′ i =ΔT×P i

[0113] Where: P′ i P represents the measurement point after distortion correction; ΔT represents the change in lidar pose of the current measurement point relative to the start of this measurement, and can be a 4x4 transformation matrix. i Mark the measurement points before correction of distortion.

[0114] S4040, Extract ground point cloud data.

[0115] Figure 5 This is a schematic diagram of a ground point cloud extraction principle provided in Embodiment 4 of the present invention, as follows: Figure 5 As shown, points A and B are measured by the lidar sensor at the same rotation angle; they are adjacent laser beams. Point C is the projection of point B onto the horizontal plane. The tilt angle β is calculated using trigonometric functions, and adjacent points close to 0 degrees are considered ground points.

[0116] S4050 Correct the point cloud intensity value of the ground point cloud data.

[0117] Figure 6 This is an example diagram of the incident angle for point cloud data acquisition according to Embodiment 4 of the present invention. Since lane lines are extracted based on the different reflection intensities of the point cloud, and the point cloud intensity is affected not only by different materials but also by the incident angle and distance. Figure 6 As shown, α is the incident angle of the point cloud, and n is the normal vector of the plane.

[0118] The laser measurements of lidar can be modeled according to the following formula, which can quantify the values ​​that affect the intensity of point clouds.

[0119]

[0120] Where: I is the intensity value of the lidar; P e P represents the power of the emitted laser beam. r D represents the energy value of radar intensity. r η is the aperture of the receiver; sys η is the system emission factor; atm ρ is the atmospheric emissivity; α is the incident angle between the object surface and the laser beam; ρ is the reflectivity of the object material; and R is the measured distance.

[0121] S4060, Extract ground lane lines.

[0122] In this embodiment, the ground point cloud after intensity correction includes lane lines and asphalt pavement. Experiments have shown that there is a clear distinction between pavement and lane lines in the point cloud intensity values. Lane lines can be extracted by setting an intensity threshold; points below the threshold are identified as ground, while those above the threshold are identified as lane lines. Due to differences in wear and materials among different urban pavements, setting a fixed segmentation threshold sometimes lacks universality. To achieve adaptive threshold setting, the Otsu's algorithm can be used to automatically calculate the threshold. The Otsu's algorithm determines the threshold that best distinguishes between the two classes of data by calculating the variance of the lane line point cloud data and the pavement point cloud data.

[0123] S4070 Extract point cloud feature data of objects such as pillars, railings, and road signs.

[0124] In this embodiment, features such as pillars, railings, and road signs are calculated by measuring the angle between the point clouds of adjacent frames and the horizontal plane. Since these features are all perpendicular to the ground, point cloud data with angles close to 90 degrees can be retained. Because urban roads also have interference from buildings, the retained potential features can be further post-processed. Specifically, clustering can be performed first, followed by geometric constraint verification for each cluster. The main steps of geometric constraint verification can include performing PCA analysis on each clustered point cloud to obtain its feature values, extracting planes and pillars, performing point cloud intensity checks and calculating plane normal vectors on the planar point cloud clusters, and identifying point clouds whose plane normal vectors are consistent with the vehicle's direction of travel and whose point cloud intensity values ​​are greater than a threshold as road signs.

[0125] S4080. Create a high-precision map (HD Map).

[0126] In this embodiment, road information can be collected using high-precision positioning devices and sensors to create a high-precision map. The high-precision map stores road semantic information in the form of vector information, including but not limited to vector information such as points, lines, and areas, and stores road surface information such as light poles, road signs, and curbs, as well as road marking information such as solid lines, dashed lines, and text.

[0127] S4090: Perform particle filtering on the ground lane lines and environmental feature data such as pillars, railings, and road signs, and high-precision maps.

[0128] In this embodiment, the vehicle's pose prior can be obtained through its own pose sensor. A local map is then loaded into a high-precision map, and the point cloud features obtained from the aforementioned observations are used to input into a particle filter to achieve continuous localization in the high-precision map. The probability density function is approximated by random samples propagating in the state space through the particle filter. The sample mean is used to replace the integral operation, thereby obtaining the minimum variance estimate of the system state, which yields the optimal estimate of the current state.

[0129] S4100, output vehicle position.

[0130] Example 5

[0131] Figure 7 This is a schematic diagram of a positioning device according to Embodiment 5 of the present invention. Figure 7 As shown, the device includes: a data acquisition module 51, a distortion correction module 52, a feature extraction module 53, and a vehicle positioning module 54.

[0132] Among them, the data acquisition module 51 is used to acquire environmental point cloud data and vehicle pose increments.

[0133] The distortion correction module 52 is used to correct the point cloud distortion of the environmental point cloud data based on the vehicle pose increment.

[0134] The feature extraction module 53 is used to determine environmental feature data from environmental point cloud data.

[0135] The vehicle positioning module 54 is used to determine the vehicle's location within a preset high-precision map based on environmental feature data.

[0136] In this embodiment of the invention, an environmental point cloud data and vehicle pose increment are acquired through a data acquisition module, a distortion correction module corrects the point cloud distortion of the environmental point cloud data based on the vehicle pose increment, a feature extraction module determines environmental feature data from the point cloud data, and a vehicle positioning module determines the vehicle position within a preset high-precision map based on the environmental feature data. This achieves accurate vehicle positioning by determining the vehicle position based on environmental feature data, thereby improving the user experience.

[0137] In one embodiment, the data acquisition module 51 includes:

[0138] The pose data acquisition unit is used to activate the lidar sensor to measure environmental point cloud data and control the pose sensor to measure at least two pose data points at threshold time intervals.

[0139] The pose increment acquisition unit is used to determine the displacement increment and angle increment between each pose data and the reference pose data as the vehicle pose increment, wherein the reference pose data is the first pose data collected within each pose data.

[0140] In one embodiment, the distortion correction module 52 includes:

[0141] The product confirmation unit is used to determine the correction transformation matrix according to the vehicle pose increment, and to determine the product of the correction transformation matrix and the environmental point cloud data.

[0142] The distortion correction unit is used to take the product as the distortion-corrected environmental point cloud data.

[0143] In one embodiment, the feature extraction module 53 includes:

[0144] The position determination unit is used to determine the positional relationship between two position points in the environmental point cloud data, wherein the two position points are two point cloud data on adjacent laser beams with the same horizontal measurement angle.

[0145] The first data determination unit is used to determine two location points as ground feature data in the environmental feature data if the positional relationship is horizontal.

[0146] The second data determination unit is used to determine two location points as object feature data in the environmental feature data if the positional relationship is vertical.

[0147] In one embodiment, the position determination unit includes:

[0148] The coordinate information extraction unit is used to extract the coordinate information of two location points within the environmental point cloud data.

[0149] Angle confirmation unit is used to determine the angle of inclination between the line connecting two location points and the horizontal plane based on coordinate information.

[0150] The first positional relationship determination unit is used to determine the positional relationship as horizontal when the included tilt angle is less than the first threshold angle.

[0151] The second position relationship determination unit is used to determine the position relationship as vertical when the included angle of inclination is greater than the second threshold angle.

[0152] In one embodiment, the feature extraction module 53 further includes:

[0153] The intensity value unit is used to determine the point cloud intensity value at each location point within the ground feature data.

[0154] The first intensity value comparison unit is used to determine the location point as road surface feature data if the point cloud intensity value is less than the intensity threshold.

[0155] The second intensity value comparison unit is used to determine the location point as lane line feature data if the point cloud intensity value is greater than or equal to the intensity threshold.

[0156] In one embodiment, determining the point cloud intensity value includes: determining the point cloud intensity value of the ground feature data according to a preset maximum inter-class variance rule.

[0157] In one embodiment, the feature extraction module 53 further includes:

[0158] The clustering point cloud acquisition unit is used to perform clustering operations on object feature data and acquire at least one clustered point cloud cluster.

[0159] The result verification unit is used to generate the geometric constraint verification results for each clustered point cloud.

[0160] The feature data determination unit is used to determine the clustered point cloud as at least one of the column feature data, railing feature data, and road sign feature data according to the geometric constraint verification results.

[0161] In one embodiment, the vehicle positioning module 54 includes:

[0162] The initial map confirmation unit is used to collect the vehicle's prior pose at the current moment and determine the initial local map within a preset high-precision map according to the vehicle's prior pose.

[0163] The vehicle location determination unit is used to process environmental feature data and initial local map according to preset particle filtering processing rules to determine the vehicle location.

[0164] The positioning device provided in this embodiment of the invention can execute a positioning method provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the method.

[0165] Example 6

[0166] Figure 8 This is a schematic diagram of the structure of an electronic device 10 implementing a positioning method according to an embodiment of the present invention. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices (such as helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0167] like Figure 8As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12 or a random access memory (RAM) 13, communicatively connected to the at least one processor 11. The memory stores computer programs executable by the at least one processor. The processor 11 can perform various appropriate actions and processes based on the computer program stored in the ROM 12 or loaded from storage unit 18 into the RAM 13. The RAM 13 may also store various programs and data required for the operation of the electronic device 10. The processor 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0168] Multiple components in electronic device 10 are connected to I / O interface 15, including: input unit 16, such as keyboard, mouse, etc.; output unit 17, such as various types of displays, speakers, etc.; storage unit 18, such as disk, optical disk, etc.; and communication unit 19, such as network card, modem, wireless transceiver, etc. Communication unit 19 allows electronic device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0169] Processor 11 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. Processor 11 performs the various methods and processes described above, such as a positioning method.

[0170] In some embodiments, a positioning method may be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or mounted on electronic device 10 via ROM 12 and / or communication unit 19. When the computer program is loaded into RAM 13 and executed by processor 11, one or more steps of a positioning method described above may be performed. Alternatively, in other embodiments, processor 11 may be configured to perform a positioning method by any other suitable means (e.g., by means of firmware).

[0171] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0172] Computer programs used to implement the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0173] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0174] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0175] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or computing systems that include middleware components (e.g., application servers), or computing systems that include frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0176] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.

[0177] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0178] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A positioning method, characterized by, The method comprises: acquiring environment point cloud data and vehicle pose increment; performing point cloud distortion correction on the environment point cloud data according to the vehicle pose increment; determining environment feature data in the environment point cloud data, wherein the environment feature data refers to special data available for identification in the environment, including ground feature data and object feature data; determining the vehicle position in a preset high-precision map according to the environment feature data, including: corresponding the environment feature data to the environment feature in the preset high-precision map, determining the environment around the position where the vehicle is located according to the corresponding environment feature data, and determining the vehicle position in the preset high-precision map.

2. The method of claim 1, wherein, The acquisition of the environment point cloud data and the vehicle pose increment comprises: starting a laser radar sensor to measure the environment point cloud data, and controlling a pose sensor to measure at least two pose data at a threshold time interval; determining the displacement increment and the angle increment of each pose data and the reference pose data as the vehicle pose increment, wherein the reference pose data is the first collected pose data in each pose data.

3. The method of claim 1, wherein, The point cloud distortion correction on the environment point cloud data according to the vehicle pose increment comprises: determining a correction transformation matrix according to the vehicle pose increment, determining the product of the correction transformation matrix and the environment point cloud data; taking the product as the environment point cloud data after the point cloud distortion correction.

4. The method of claim 1, wherein, The determination of the environment feature data in the environment point cloud data comprises: determining the positional relationship of two position points in the environment point cloud data, wherein the two position points are two point cloud data on adjacent laser beam lines at the same horizontal measurement angle; if the positional relationship is a horizontal relationship, determining that the two position points are ground feature data in the environment feature data; if the positional relationship is a vertical relationship, determining that the two position points are object feature data in the environment feature data.

5. The method of claim 4, wherein, The determination of the positional relationship of the two position points in the environment point cloud data comprises: extracting the coordinate information of the two position points in the environment point cloud data; determining the inclination angle between the connecting line of the two position points and the horizontal plane according to the coordinate information; in the case that the inclination angle is less than a first threshold angle, determining that the positional relationship is the horizontal relationship; in the case that the inclination angle is greater than a second threshold angle, determining that the positional relationship is the vertical relationship.

6. The method of claim 4, wherein, Further comprising: determining the point cloud intensity value of each position point in the ground feature data; if the point cloud intensity value is less than an intensity threshold value, determining that the position point is road surface feature data; if the point cloud intensity value is greater than or equal to the intensity threshold value, determining that the position point is lane line feature data.

7. The method of claim 6, wherein, The determination of the intensity threshold value comprises: determining the intensity threshold value of the ground feature data according to a preset maximum inter-class variance rule.

8. The method of claim 4, wherein, Further comprising: performing clustering operation on the object feature data, and acquiring at least one clustered point cloud cluster; generating a geometric constraint verification result of each clustered point cloud cluster; The clustering point cloud clusters are determined as at least one of column feature data, railing feature data and road sign feature data according to the geometric constraint checking result.

9. The method of claim 1, wherein, The vehicle position is determined in the preset high-definition map according to the environment feature data, including: The vehicle prior pose at the current time is collected, and an initial local map is determined in the preset high-definition map according to the vehicle prior pose; The environment feature data and the initial local map are processed according to a preset particle filtering processing rule to determine the vehicle position.

10. A positioning device, characterized by The device comprises: A data acquisition module is configured to acquire environment point cloud data and vehicle pose increments; A distortion correction module is configured to perform point cloud distortion correction on the environment point cloud data according to the vehicle pose increments; A feature extraction module is configured to determine environment feature data in the environment point cloud data, the environment feature data referring to special data available for identification in the environment, including ground feature data and object feature data; A vehicle positioning module is configured to determine a vehicle position in a preset high-definition map according to the environment feature data, including: corresponding the environment feature data to environment features in the preset high-definition map, determining an environment around the vehicle position through the corresponding environment feature data, and determining the vehicle position in the preset high-definition map.

11. An electronic device, comprising: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor to enable the at least one processor to execute the positioning method of any one of claims 1-9.

12. A computer-readable storage medium, characterized in that, The computer readable storage medium stores computer instructions for enabling the processor to execute the positioning method of any one of claims 1-9 when executed. The computer readable storage medium stores computer instructions for enabling the processor to execute the positioning method of any one of claims 1-9 when executed.

Images