WO2024051344A1 - 一种地图创建方法及装置 - Google Patents

一种地图创建方法及装置 Download PDF

Info

Publication number
WO2024051344A1
WO2024051344A1 PCT/CN2023/107169 CN2023107169W WO2024051344A1 WO 2024051344 A1 WO2024051344 A1 WO 2024051344A1 CN 2023107169 W CN2023107169 W CN 2023107169W WO 2024051344 A1 WO2024051344 A1 WO 2024051344A1
Authority
WO
WIPO (PCT)
Prior art keywords
data
data set
environmental data
data sets
environmental
Prior art date
Application number
PCT/CN2023/107169
Other languages
English (en)
French (fr)
Inventor
薛宇飞
Original Assignee
北京地平线机器人技术研发有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 北京地平线机器人技术研发有限公司 filed Critical 北京地平线机器人技术研发有限公司
Publication of WO2024051344A1 publication Critical patent/WO2024051344A1/zh

Links

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
    • G01C21/26—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 specially adapted for navigation in a road network
    • G01C21/28—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 specially adapted for navigation in a road network with correlation of data from several navigational instruments
    • G01C21/30—Map- or contour-matching
    • G01C21/32—Structuring or formatting of map data

Definitions

  • the present disclosure relates to the technical field of map creation, and in particular to a map creation method and device.
  • vehicles In application scenarios such as assisted driving or autonomous driving, vehicles usually need to determine the type and location of road elements (such as lane lines, road edges, road signs, etc.) contained in the surrounding environment in order to adjust driving behavior accordingly.
  • road elements such as lane lines, road edges, road signs, etc.
  • Embodiments of the present disclosure provide a map creation method and device.
  • a map creation method including:
  • At least one data set is determined, wherein the environmental data contained in each data set corresponds to the same type of pavement elements;
  • a map creation device including:
  • the data acquisition module is used to obtain the environmental data collected by the mobile device for the target area
  • the set determination module is used to determine at least one data set by classifying the environmental data obtained by the data acquisition module, wherein the environmental data contained in each data set corresponds to the same type of pavement elements;
  • the confidence determination module is used to determine the confidence corresponding to each data set based on the three-dimensional spatial coordinates corresponding to each environmental data contained in each data set determined by the set determination module;
  • the map creation module is used to create a high-precision map based on the confidence levels corresponding to each data set determined by the confidence level determination module.
  • a computer-readable storage medium stores a computer program, and the computer program is stored in the storage medium.
  • a computer program is used to execute the map creation method of any of the above embodiments of the present disclosure.
  • an electronic device includes:
  • Memory used to store instructions executable by the processor
  • a processor configured to read executable instructions from the memory and execute the instructions to implement the map creation method of any of the above embodiments of the present disclosure.
  • the environmental data after obtaining the environmental data of the target area, the environmental data can be divided into at least one data set.
  • the environmental data contained in each data set corresponds to the same type of pavement elements, and then it is determined that the environmental data contained in each data set contains The three-dimensional spatial coordinates corresponding to the environmental data are determined, and the confidence levels corresponding to each data set are determined, and then a high-precision map is created based on the confidence levels.
  • the solution provided by the embodiments of the present disclosure can create a higher-precision map and solve the problem of low accuracy of maps created by the prior art.
  • the solution provided by the embodiments of the present disclosure has lower performance requirements for mobile devices that collect environmental data, and does not require the use of high-performance mobile devices, so that high-precision maps can be obtained at reduced costs and has a wide range of applications.
  • Figure 1 is a scene diagram to which this disclosure is applicable.
  • Figure 2 is a schematic flowchart of a map creation method provided by an exemplary embodiment of the present disclosure.
  • Figure 3 is a schematic flowchart of a map creation method provided by another exemplary embodiment of the present disclosure.
  • Figure 4 is an example diagram of environmental data distribution provided by an exemplary embodiment of the present disclosure.
  • Figure 5(a) is an example diagram of environmental data distribution provided by another exemplary embodiment of the present disclosure.
  • FIG. 5(b) is an example diagram of environmental data distribution provided by another exemplary embodiment of the present disclosure.
  • Figure 6 is a schematic flowchart of a map creation method provided by another exemplary embodiment of the present disclosure.
  • Figure 7 is a schematic flowchart of a map creation method provided by another exemplary embodiment of the present disclosure.
  • Figure 8 is a schematic flowchart of a map creation method provided by another exemplary embodiment of the present disclosure.
  • Figure 9 is a schematic flowchart of a map creation method provided by another exemplary embodiment of the present disclosure.
  • Figure 10 is a schematic structural diagram of a map creation device provided by an exemplary embodiment of the present disclosure.
  • Figure 11 is a schematic structural diagram of a map creation device provided by another exemplary embodiment of the present disclosure.
  • FIG. 12 is a structural diagram of an electronic device provided by an exemplary embodiment of the present disclosure.
  • vehicles In application scenarios such as assisted driving or autonomous driving, vehicles often need to determine the type and location of road elements (such as lane lines, road edges, road signs, etc.) existing in the surrounding environment in order to adjust the vehicle's driving behavior accordingly. For example, if it is determined that there is a road sign in front of the vehicle, the driving direction of the vehicle can be appropriately adjusted according to the instructions of the road sign. In order to meet this demand of vehicles, it is common to build maps containing the surrounding environment for vehicles.
  • road elements such as lane lines, road edges, road signs, etc.
  • the environmental data of each pavement element in the surrounding environment is usually collected by a collection device.
  • the environmental data includes data indicating the type and location of the pavement element.
  • the environmental data may include the data corresponding to the pavement element. Point cloud data sets and/or depth images, etc.; then through clustering processing of environmental data, the type and location of each pavement element is determined, and a map indicating the surrounding environment is further constructed based on this.
  • the point cloud data set includes multiple points, and each point refers to the corresponding point cloud data.
  • the point cloud data set can be determined through lidar measurements (i.e., the acquisition device includes a lidar). The laser emitted by the lidar is reflected after encountering a certain position of the pavement element, and the reflected light is received by the lidar. Light, then each point in the point cloud data set (i.e., point cloud data) usually includes the three-dimensional coordinates of a certain position of the pavement element and the laser reflection intensity.
  • the environmental data collected by the collection equipment is sometimes less accurate.
  • the environmental data includes depth images of pavement elements, in rainy weather, the depth images captured by the collection equipment are sometimes blurry, resulting in lower clarity of the pavement elements in the depth images and a decrease in the accuracy of the corresponding environmental data.
  • the laser reflected by the road elements may be affected by the obstruction and cannot be received by the lidar, resulting in a sparse distribution of point cloud data.
  • the point cloud data set determined by the lidar will be accuracy decreases.
  • one current improvement solution is to apply high-performance collection equipment to obtain more accurate environmental data through high-performance collection equipment.
  • the price of high-performance collection equipment is usually high, and this solution will increase the cost of creating maps, so it cannot be widely used.
  • embodiments of the present disclosure provide a map creation method and device.
  • the confidence level of the environmental data will be determined, and the map will be created based on the environmental data with higher confidence level.
  • the embodiments of the present disclosure can be applied to application scenarios that require creating a map, and the application scenarios may include application scenarios such as assisted driving or automatic driving.
  • the environmental data corresponding to the surrounding environment of the vehicle can be collected by collection equipment such as depth cameras and lidar installed in the vehicle, and through the solution provided by the present disclosure, the environmental data can be used to create high-level data for the vehicle. Precise map so that the map can provide reference for assisted driving or autonomous driving.
  • the device used to implement the map creation method of the embodiment of the present disclosure may be an electronic device such as a computer, an intelligent driving control device, or a server (such as a vehicle-mounted server), and an example diagram of the device is disclosed in FIG. 1 .
  • a map creation device 100 that creates a map based on the solution provided by the embodiment of the present disclosure is connected to a device for collecting environmental data.
  • the device is usually installed on a vehicle and moves with the movement of the vehicle.
  • the devices used to collect environmental data are called mobile devices.
  • the mobile device includes a first device 200 and a second device 300 connected to the map creation device 100 .
  • it may also include more mobile devices connected to the map creation device 100 and used to collect environmental data, which is not limited by this disclosure.
  • the first device 200 and the second device 300 may collect the same type of environmental data, or may collect different types of environmental data.
  • one mobile device may include a depth camera, and the environmental data collected by the mobile device may include depth images corresponding to pavement elements.
  • the other mobile device may include a lidar, and the environmental data collected by the mobile device may include pavement elements.
  • the environmental data collected by the first device 200 and the second device 300 may both include point cloud data sets corresponding to road surface elements, or both may include depth images corresponding to road surface elements.
  • the first device 200 and the second device 300 may be connected to the map creation device 100 in various ways.
  • the map creation device 100 may be electrically connected to the first device 200 and the second device 300.
  • the map creation device 100 is a vehicle-mounted device.
  • the map creation device 100 is connected to the first device 200 and the second device 300.
  • the device 200 and the second device 300 are electrically connected.
  • connection method between the map creation device 100 and the first device 200 and the second device 300 may also be a network connection.
  • connection method between the map creation device 100 and the first device 200, and the map creation device 100 and the second device 300 may be the same connection method, or may be different connection methods, and this disclosure does not limit this.
  • the first device 200 and the second device 300 can collect environmental data for a target area.
  • the solution provided by the embodiment of the present disclosure is applied to the field of assisted driving or automatic driving, then the target area can include the area of the vehicle's surrounding environment,
  • the environmental data may include data indicating the types and locations of road surface elements included in the target area.
  • the environmental data may include a point cloud data set corresponding to the road surface elements, or a depth image corresponding to the road surface elements, or both the road surface and the road surface. Point cloud dataset and depth image corresponding to the element.
  • the environmental data may also include other data that may indicate the type and location of road surface elements, which is not limited by the present disclosure.
  • the first device 200 and the second device 300 may transmit the collected environmental data to the map creation device 100 .
  • the map creation device 100 executes the map creation method provided by the embodiment of the present disclosure to create a high-precision map.
  • the map creation device 100 may be a vehicle-mounted device (such as a vehicle-mounted smart terminal). In this case, the map creation device 100 may create a high-precision map for the vehicle based on the received environmental data. Maps to meet vehicle needs.
  • FIG. 2 is a schematic flowchart of a map creation method provided by an exemplary embodiment of the present disclosure. This embodiment can be applied to electronic devices, as shown in Figure 2, including the following steps:
  • Step S201 Obtain environmental data collected by the mobile device for the target area.
  • the target area usually includes the area where a map needs to be created.
  • the target area usually includes the area surrounding the road where the vehicle is located.
  • the environmental data includes data indicating the type and location of pavement elements.
  • the environmental data may include depth images corresponding to road elements.
  • the mobile device may include a depth camera capable of capturing depth images.
  • the environmental data may include a point cloud data set corresponding to road surface elements, in which case the mobile device may include a lidar capable of generating a point cloud data set.
  • the environmental data may include both depth images and point cloud data sets corresponding to pavement elements, or other data that may indicate the type and location of pavement elements.
  • Step S202 Determine at least one data set by classifying environmental data.
  • each data set corresponds to the same type of pavement elements.
  • the semantics of the environmental data can be determined through the analysis and processing of the environmental data, and the environmental data can be classified accordingly, and the environmental data corresponding to the same type of pavement elements can be divided into the same data collection.
  • the environmental data collected by the mobile device includes a point cloud data set corresponding to the road edge.
  • a data set can be determined, and the environmental data contained in the data set corresponds to the pavement element of the road edge.
  • the environmental data collected by the mobile device includes a point cloud data set corresponding to the road edge, a point cloud data set corresponding to the lane line, a depth image corresponding to the road edge, and a depth image corresponding to the lane line
  • Operation can obtain the data set corresponding to the road edge, which includes the point cloud data set and depth image corresponding to the road edge, and obtain the data set corresponding to the lane line, which includes the point cloud data set corresponding to the lane line. and depth images.
  • Step S203 Determine the confidence level corresponding to each data set based on the three-dimensional spatial coordinates corresponding to each environmental data contained in each data set.
  • the confidence of the data set can be used to characterize the accuracy of the data set.
  • Step S204 Create a high-precision map based on the confidence levels corresponding to each data set.
  • the confidence of each data set can reflect the accuracy of each data set, based on the confidence of each data set, the data set with higher accuracy can be determined, so that a high-precision map can be created.
  • the environmental data after obtaining the environmental data of the target area, the environmental data can be divided into at least one data set.
  • the environmental data contained in each data set corresponds to the same type of road elements, and then it is determined that the environmental data contained in each data set contains environmental data respectively
  • the corresponding three-dimensional spatial coordinates are used to determine the confidence levels corresponding to each data set, and then a high-precision map is created based on the confidence levels.
  • the solution provided by the embodiments of the present disclosure can create a higher-precision map and solve the problem of low accuracy of maps created by the prior art.
  • the solution provided by the embodiments of the present disclosure has lower performance requirements for mobile devices that collect environmental data, and does not require the use of high-performance mobile devices, so that high-precision maps can be obtained at reduced costs and has a wide range of applications.
  • step S203 of the present disclosure the confidence level corresponding to each data set is determined based on the three-dimensional spatial coordinates corresponding to each environmental data contained in each data set.
  • step S203 may include the following steps:
  • Step S2031 Based on the three-dimensional spatial coordinates of each environmental data contained in each data set, determine three relationship curves between the first parameter of the road element and the three second parameters.
  • step S2031 by performing data fitting on each environmental data contained in the same data set, three relationship curves of the pavement elements corresponding to the data set are determined. These three relationship curves respectively represent the relationship between the first parameter and one of the second parameters.
  • the first parameter is the arc length between the observation point of the pavement element and the starting point of the relationship curve
  • each second parameter is the distance in the three-dimensional direction between the observation point and the projection point of the relationship curve respectively.
  • the observation point refers to a point on the surface of the pavement element. If the environmental data includes a point cloud data set corresponding to a pavement element, the laser emitted by the mobile device will be reflected after reaching the observation point of the pavement element. The mobile device receives the emitted laser and generates a point cloud data set accordingly. In addition, if the environmental data includes depth images corresponding to pavement elements, when the mobile device photographs the target environment at different angles, it can obtain depth images of observation points at different angles of the pavement elements.
  • each relationship curve can be represented by a cubic polynomial.
  • t is the first parameter, that is, the arc length between the observation point of the pavement element and the starting point of the relationship curve; rx, ry and rz are all second parameters, and rx represents the distance between the observation point and the relationship curve.
  • the distance between projection points in the z direction; a0, b0, c0, d0, a1, b1, c1, d1, a2, b2, c2 and d2 are the coefficients of the cubic polynomial, x_obs, y_obs and z_obs respectively represent the constants of the relationship curve item.
  • Step S2032 Based on the relationship curve, determine the degree of aggregation of the distribution of each environmental data included in each data set.
  • the environmental data contained in each data set correspond to the same type of road surface elements. That is to say, based on the environmental data contained in a certain data set, a certain type of road surface corresponding to the data set can be determined. The type and position of the element.
  • the environmental data contained in a certain data set is distributed more dispersedly, that is, the degree of aggregation of the environmental data contained in the data set is low, then the accuracy of the environmental data contained in the data set is low. Based on the data There is a large deviation between the categories and locations of pavement elements determined by the set and the actual environment.
  • the accuracy of determining pavement elements based on the environmental data contained in the data set is usually related to the degree of aggregation of the environmental data contained in the data set.
  • the road surface elements include lane lines and road edges.
  • a first data set corresponding to the lane lines and a second data set corresponding to the road edges can be determined.
  • the distribution of lane lines determined based on the environmental data contained in the first data set is shown as the solid line in Figure 4
  • the distribution of road edges determined based on the environmental data contained in the second data set is shown as the dotted line in Figure 4
  • the line containing the arrow is the route the vehicle traveled in this example.
  • the solid lines corresponding to the lane lines are more concentrated in distribution than the dotted lines corresponding to the road edges, so the environmental data distribution included in the first data set has a higher degree of aggregation.
  • the lane line determined based on the environmental data included in the first data set is often closer to the actual environment than the road edge determined based on the environmental data included in the second data set.
  • the distribution of environmental data contained in a certain data set is shown in Figure 5(a), and the distribution of environmental data contained in another data set is shown in Figure 5(b).
  • the environmental data distribution contained in the data set corresponding to Figure 5(b) has a higher degree of aggregation.
  • the accuracy of the pavement elements determined based on the environmental data contained in the data set corresponding to Figure 5(b) is higher.
  • Step S2033 Based on the aggregation degree of each environmental data distribution, determine the confidence level corresponding to each data set.
  • the accuracy of a data set is usually related to the degree of aggregation of each environmental data contained in the data set.
  • the higher the degree of aggregation of each environmental data distribution in a certain data set the higher the confidence level corresponding to the data set.
  • the confidence level corresponding to each data set can be determined based on the degree of aggregation of each environmental data distribution in the data set.
  • the degree of aggregation of the environmental data distribution contained in the data set is determined, and then the degree of aggregation is used Determine the confidence level corresponding to each data set.
  • the accuracy of a data set is usually related to the degree of aggregation of each environmental data contained in the data set, and the confidence of each data set is determined based on the degree of aggregation of each environmental data distribution in the data set. Therefore, through the embodiment of the present disclosure The confidence level determined by the scheme can be used to evaluate the accuracy of the environmental data within each data set.
  • the aggregation degree of each environmental data distribution contained in each data set can be determined through the following steps:
  • Step S20321 Determine the residual matrix corresponding to each data set based on the relationship curve and the three-dimensional spatial coordinates of each environmental data in each data set.
  • the residual matrix corresponding to a certain data set can be expressed by the following formula (4):
  • R represents the residual matrix corresponding to one of the data sets.
  • the data set includes n environmental data, n is a positive integer, (rx0, ry0, rz0), (rx1, ry1, rz1) to (rxn, ryn, rzn ) is the three-dimensional spatial coordinates of n environmental data contained in the data set, and r is a constant.
  • Step S20322 Determine the covariance matrix corresponding to each data set based on the residual matrix corresponding to each data set.
  • the distribution of residuals of environmental data generally approximately conforms to a normal distribution, and the normal distribution generally follows the following equation (5):
  • ⁇ represents the covariance matrix of one of the data sets
  • the residual matrix corresponding to the data set is R
  • R T represents the transpose of the residual matrix R.
  • Step S20323 Based on the trace of the covariance matrix corresponding to each data set, determine the unbiased estimator corresponding to each data set.
  • This unbiased estimator is used to characterize the degree of aggregation of the distribution of each environmental data contained in the data set.
  • the variance of the residual matrix corresponding to a certain data set can be represented by the trace of the covariance matrix corresponding to the data set.
  • the unbiased estimator corresponding to one of the data sets can be expressed by the following equation (7):
  • s 2 represents the unbiased estimator corresponding to the data set
  • n represents the number of environmental data contained in the data set
  • tr( ⁇ ) represents the trace of the covariance matrix corresponding to the data set.
  • the trace of the covariance matrix corresponding to the data set can be used to determine the unbiased estimator corresponding to the data set.
  • the unbiased estimator can represent the degree of aggregation of the environmental data distribution contained in the data set.
  • the confidence level of the data set is usually positively related to the number of times the mobile device collects environmental data.
  • the higher the number of times the environmental data is collected the richer the amount of environmental data contained in the data set. Pavement elements are determined through the data set. The higher the accuracy. In other words, the higher the number of times environmental data is collected, the greater the confidence of the corresponding data set. Therefore, after determining the aggregation degree of each environmental data distribution contained in the data set, the step of determining the confidence level corresponding to each data set based on the aggregation degree of each environmental data distribution can be achieved through the following operations:
  • the confidence level corresponding to each data set is determined.
  • the confidence level of the data set can be determined through the following equation (8):
  • score represents the confidence of the data set
  • the unbiased estimator of the data set is s 2
  • tr( ⁇ ) represents the trace of the covariance matrix corresponding to the data set.
  • the number of environmental data contained in the data set is n
  • n represents the number of environmental data contained in the data set.
  • the mobile device since the number of data contained in the data set is n, and the mobile device usually acquires one piece of environmental data each time it collects environmental data, the number of times the mobile device collects the environmental data contained in the data set is also n.
  • the unbiased estimator corresponding to each data set can be determined based on the relationship curve of each data set.
  • the unbiased estimator can represent the degree of aggregation of the distribution of each environmental data contained in each data set. .
  • the confidence level corresponding to each data set can be determined, so that a high-precision map can be created based on the confidence level of each data set.
  • At least one data set is determined by classifying environmental data. During the actual map creation process, this data collection can be determined in several ways.
  • the environmental data is classified based on the semantics of each environmental data.
  • the environmental data contained in the same data set may include data collected by different mobile devices and correspond to the same type of pavement elements. environmental data.
  • the confidence of each data set represents the environment of each type of pavement element.
  • the confidence levels corresponding to the data respectively.
  • the mobile device used to collect environmental data includes a depth camera and a lidar
  • the collected environmental data includes environmental data corresponding to the road edge and environmental data corresponding to the lane lines
  • two data sets can be determined based on the above implementation, where One data set includes environmental data of road edges collected by depth cameras and lidar, and the other data set includes environmental data of lane lines collected by depth cameras and lidar.
  • the environmental data can be classified through the following steps to determine at least one data set:
  • Step S2021 Classify the environmental data based on the type of mobile device corresponding to each environmental data, and determine at least one intermediate data set;
  • Step S2022 Classify each environmental data contained in each intermediate data set based on the types of pavement elements corresponding to each environmental data, and determine at least one data set.
  • the environmental data is first classified based on the mobile device that collects the environmental data to obtain an intermediate data set.
  • each environmental data contained in the same intermediate data set is composed of the same type.
  • the mobile device collects; and then, based on the type of the pavement element corresponding to the environmental data, classify each environmental data contained in the intermediate data set to determine at least one data set.
  • the environmental data can also be classified first based on the types of pavement elements corresponding to each environmental data, and at least one intermediate data set is determined.
  • each environmental data contained in the same intermediate data set Corresponding to the same type of pavement elements; then, based on the type of mobile device corresponding to each environmental data, each environmental data contained in the intermediate data set is classified, and at least one data set is determined.
  • each environmental data contained in each determined data set is collected by the same type of mobile device and corresponds to the same type of road element.
  • the mobile device used to collect environmental data includes a depth camera and a lidar
  • the collected environmental data includes environmental data corresponding to the road edge and environmental data corresponding to the lane lines
  • four data sets can be determined based on the above implementation.
  • the four data sets include: a data set composed of environmental data of road edges collected by depth cameras, a data set composed of environmental data of road edges collected by lidar, and a data set composed of environmental data of lane lines collected by depth cameras. And a data set consisting of environmental data of lane lines collected by lidar.
  • each environmental data contained in each determined data set is collected by the same type of mobile device and corresponds to the same type of road element, so that it can be based on each data
  • the confidence level of the collection determines the accuracy of each mobile data collection environment data.
  • the data set determined based on the solution provided by the embodiment of the present disclosure includes: a third data set composed of environmental data of the road edge collected by the depth camera, and a fourth data composed of environmental data of the road edge collected by the lidar collection if third data If the confidence level of the set is higher than the confidence level of the fourth data set, it indicates that the accuracy of the environmental data collected by the depth camera is higher than the accuracy of the environmental data collected by the lidar.
  • Step S2023 Determine the first target data set in each data set.
  • the confidence of the first target data set is smaller than the confidence of other data sets in each data set.
  • the number of the first target data set may be N, and N is a preset positive integer.
  • Step S2024 Generate prompt information based on the mobile device corresponding to the first target data set.
  • the prompt information is used to indicate that the mobile device corresponding to the first target data set has a problem of low accuracy when collecting environmental data.
  • the environmental data contained in each determined data set is collected by the same type of mobile device and corresponds to the same type of road element.
  • the confidence level of each data set can reflect each mobile data Accuracy of collecting environmental data.
  • the prompt information can prompt the technician to determine the mobile device's accuracy.
  • the equipment is inspected and repaired in order to improve the accuracy of environmental data collection by the mobile device and further improve the accuracy of map creation.
  • an operation of creating a high-precision map based on the respective confidence levels of each data set is provided.
  • This operation can be implemented in a variety of ways.
  • the operation of creating a high-precision map may include the following steps:
  • a second target data set is determined based on the confidence levels corresponding to each data set.
  • the confidence level of the second target data set is greater than the confidence levels of other data sets.
  • the number of the second target data set can be preset, and the second target data set is determined based on the confidence corresponding to each data set and the preset number of the second target data set.
  • a confidence threshold may be set in advance. If the confidence of a certain data set is greater than the confidence threshold, the data set is determined to be the second target data set.
  • the accuracy of the pavement elements determined through the second target data set is higher.
  • the road surface elements determined through the second target data set are set as high-precision road surface elements, and the high-precision road surface elements are marked and displayed on the high-precision map.
  • the high-precision road surface elements and other high-precision road surface elements in the high-precision map can be marked and displayed. pavement elements.
  • the high-accuracy road element can be displayed in a highlighted form on a high-precision map. Or, in another example, you can use a different color from other pavement elements in an HD map Display this high-accuracy road surface element.
  • a map can be created based on the types and locations of pavement elements determined by each data set, and in this map, pavement elements determined by the second target data can be marked, and the pavement elements determined by the second target data can be marked in the map.
  • the accuracy of the pavement elements determined by the set is high. Therefore, the high-precision map marks and displays the pavement elements with high accuracy, which can distinguish the high-precision pavement elements from other pavement elements in the high-precision map. Facilitates the determination of high-accuracy pavement elements in high-precision maps.
  • the operation of creating a high-precision map may include the following steps:
  • a third target data set is determined, and the confidence level of the third target data set is greater than the confidence levels of other data sets.
  • the map is created only through the third target data set with higher confidence, so the accuracy of the created map is higher.
  • vehicles can navigate through maps and other methods at the same time.
  • vehicles can also navigate through the global positioning system (GPS) and through the inertial navigation system (INS).
  • GPS global positioning system
  • INS inertial navigation system
  • the confidence level of each data set can be determined.
  • the weights of different navigation methods can also be adjusted according to the confidence level of the data set.
  • the at least two navigation methods include a first map navigation method based on high-precision maps, see Figure 9, based on the embodiment shown in Figure 2 above, In the solution provided by the embodiment of the present disclosure, the following steps are also included:
  • Step S205 Determine the weight of applying the first map navigation method based on the confidence level corresponding to each data set and the preset threshold corresponding to each data set.
  • the confidence corresponding to each data set is greater than the preset threshold corresponding to each data set, it indicates that the accuracy of the environmental data contained in each data set is relatively high, and accordingly the accuracy of the high-precision map created based on the data set is higher. The accuracy is high. In this case, the weight of navigation using the first map navigation method can be increased accordingly.
  • the application first can be reduced accordingly.
  • the weight of navigation by map navigation method In this operation, the confidence level corresponding to each data set is not greater than the preset threshold value corresponding to each data set, which means that the confidence level corresponding to any data set in each data set is not greater than the preset threshold value corresponding to the data set. threshold.
  • the corresponding relationship between the first difference value and the weight of applying the first map navigation method can be set in advance.
  • the first difference value is the difference between the confidence level corresponding to each data set and the preset threshold. value.
  • the greater the first difference value the higher the weight of applying the first map navigation method.
  • the weight of applying the first map navigation method can be determined.
  • Step S206 Navigate for the mobile device based on the determined weight of applying the first map navigation method.
  • the weight of applying the first map navigation method for navigation can be adjusted based on the confidence level corresponding to each data set, so that the navigation method can be adjusted based on the confidence level corresponding to each data set to ensure the navigation method. accuracy and improve vehicle safety.
  • the navigation method of the mobile device includes a first map navigation method and an INS navigation method using INS for navigation.
  • the confidence corresponding to each data set is greater than the preset threshold corresponding to the data set, the weight of the first map navigation method is increased, and the INS navigation method is combined with the first map navigation method after the weight is increased. Navigate.
  • the weight of the first map navigation method in various navigation methods is adjusted, which can accordingly improve the accuracy of navigation.
  • the navigation method of the mobile device includes: a first map navigation method, an INS navigation method using INS for navigation, and a GPS navigation method using GPS for navigation.
  • the weight of the first map navigation method is increased.
  • the strength of the GPS signal easily affects the accuracy of GPS navigation, the strength of the current GPS signal can also be determined. If the current GPS signal is weak, the weight of the GPS navigation method is reduced. Then, navigate based on the adjusted weight.
  • the weight of the first map navigation method in various navigation methods is adjusted based on the comparison of the confidence level corresponding to the data set with the preset threshold, and based on the strength of the GPS signal, the GPS navigation method is adjusted in various navigation methods.
  • the weight in the method can accordingly improve the accuracy of navigation.
  • the weight of applying the first map navigation method is determined based on a comparison between the confidence level corresponding to each data set and a preset threshold, where the preset threshold can be determined in a variety of ways.
  • the preset thresholds corresponding to each data set are determined in the following way:
  • the preset thresholds corresponding to each data set are determined.
  • the corresponding relationship between each pavement element and the preset threshold is set in advance, and then based on the corresponding relationship and the pavement elements corresponding to each data set, the preset corresponding to each data set can be determined. threshold.
  • a preset threshold corresponding to the pavement element may be determined based on the volume of the pavement element.
  • the larger the volume of the pavement element the easier it is to be recognized. Therefore, generally, the larger the volume of the pavement element, the larger the corresponding preset threshold in the corresponding relationship.
  • the preset threshold corresponding to the pavement element can be determined based on the degree of importance attached to the pavement element. In this case, the more emphasis is placed on a certain pavement element, the greater the corresponding preset threshold in the correspondence relationship. For example, if the road surface elements corresponding to each data set include road signs, and during the navigation process, the vehicle often needs to determine whether it needs to adjust its direction according to the instructions of the road signs, so it pays more attention to the road signs. In this case, the road sign can be set as the road surface Elements correspond to larger preset thresholds.
  • the preset thresholds corresponding to each data set are determined in the following way:
  • the first step is to determine the influencing parameters of the environmental data, which include at least one of the following parameters: the collection time of the environmental data, lighting and weather.
  • the accuracy of environmental data collected by the same mobile device is often affected by the current environment. For example, if the current illumination intensity is low, there may be obstructions in the surroundings, which will cause the lidar to determine the point cloud data set less accurately, and the lower illumination intensity will often reduce the clarity of the depth image captured by the depth camera; if the current In rainy weather, the clarity of depth images taken by depth cameras usually decreases, and the accuracy of point cloud data sets obtained by lidar also decreases; in addition, different acquisition times correspond to different lighting, for example, the light intensity in the morning is often lower than The light intensity at noon, and when the light intensity is stronger, the accuracy of the depth image captured by the depth camera is usually higher.
  • the preset thresholds corresponding to each data set are determined.
  • the preset thresholds corresponding to each data set will be reduced accordingly; when the influencing parameters of the environmental data will cause the mobile device to collect environmental data When the accuracy of , the preset threshold corresponding to each data set increases accordingly.
  • the weight of the first map navigation method in each navigation method is related to the preset threshold corresponding to each data set.
  • the preset threshold corresponding to each data set can be determined based on the current influencing parameters.
  • the threshold is adjusted. Therefore, the embodiment of the present disclosure can adjust the weight of the first map navigation method based on the current influencing parameters, thereby further improving the accuracy of navigation for mobile devices.
  • FIG. 10 is a schematic structural diagram of a map creation device provided by an exemplary embodiment of the present disclosure.
  • the map creation device can be installed in electronic equipment such as terminal equipment and servers, or on objects such as vehicles, to execute the map creation method of any of the above embodiments of the present disclosure.
  • the map creation device of this embodiment includes: a data acquisition module 201 , a set determination module 202 , a confidence determination module 203 and a map creation module 204 .
  • the data acquisition module 201 is used to acquire the environmental data collected by the mobile device for the target area;
  • the set determination module 202 is configured to determine at least one data set by classifying the environmental data obtained 201 by the data acquisition module.
  • each data set corresponds to the same type of pavement elements
  • the confidence determination module 203 is configured to determine the confidence corresponding to each data set based on the three-dimensional spatial coordinates corresponding to each environmental data contained in each data set determined by the set determination module 202;
  • the map creation module 204 is configured to create a high-precision map based on the confidence levels corresponding to each data set determined by the confidence level determination module 203 .
  • the confidence determination module 203 includes:
  • the curve determination unit 2031 is used to determine three relationship curves between the first parameter of the pavement element and three second parameters respectively based on the three-dimensional spatial coordinates of each environmental data contained in each data set, wherein the first parameter is the first parameter of the pavement element.
  • the arc from the starting point of the observation point distance relationship curve length, each second parameter is the distance in the three-dimensional direction between the observation point and the projection point of the relationship curve respectively;
  • the degree of aggregation determination unit 2032 is configured to determine the degree of aggregation of the distribution of each environmental data contained in each data set based on the relationship curve determined by the curve determination unit 2031;
  • the confidence level determination unit 2033 is configured to determine the confidence level corresponding to each data set based on the aggregation degree of each environmental data distribution determined by the aggregation degree determination unit 2032.
  • the degree of aggregation determining unit 2032 includes:
  • the residual matrix determination subunit is used to determine the residual matrix corresponding to each data set based on the relationship curve and the three-dimensional spatial coordinates of each environmental data in each data set;
  • the covariance matrix determination subunit is used to determine the residual matrix corresponding to each data set determined by the subunit based on the residual matrix, and determine the covariance matrix corresponding to each data set;
  • the unbiased estimator determination subunit is used to determine the trace of the covariance matrix corresponding to each data set determined by the subunit based on the covariance matrix, and determine the unbiased estimator corresponding to each data set.
  • the unbiased estimator is used for Characterizes the degree of aggregation of the distribution of each environmental data contained in the data set.
  • the confidence determination unit 203 includes: a confidence determination subunit, which is used to determine based on the unbiased estimator corresponding to each data set and the number of times the mobile device collects environmental data. The confidence level corresponding to each data set.
  • the set determination module 202 includes:
  • the first classification unit 2021 is used to classify the environmental data based on the type of mobile device corresponding to each environmental data, and determine at least one intermediate data set;
  • the second classification unit 2022 classifies each environmental data contained in each intermediate data set determined by the first classification unit based on the type of the road element corresponding to each environmental data, and determines at least one data set.
  • a prompt module 205 is also included.
  • the prompt module 205 is configured to determine the first target data set in each data set after determining at least one data set.
  • the confidence of the first target data set is smaller than the confidence of other data sets in each data set; and, the The prompt module 205 is also configured to generate prompt information based on the mobile device corresponding to the first target data set.
  • the map creation module 204 includes:
  • the target set determining unit 2041 is configured to determine a second target data set based on the confidence levels corresponding to each data set, where the confidence level of the second target data set is greater than the confidence levels of other data sets;
  • the map creation unit 2042 is configured to create a high-precision map corresponding to each data set, and mark and display the road surface elements determined by the second target data set determined by the target set determination unit 2041 in the high-precision map.
  • the map creation device further includes: a navigation module 206.
  • the navigation module 206 is configured to, after creating a high-precision map based on the confidence level corresponding to each data set, determine to apply the first map navigation method based on the confidence level corresponding to each data set and the preset threshold corresponding to each data set. the weight; and, the navigation module 206 is also used to perform navigation for the mobile device based on the weight.
  • the navigation module 206 may include:
  • the first threshold determination order 2061 is used to determine the preset threshold corresponding to each data set based on the correspondence between the road surface elements and the preset threshold.
  • the navigation module 206 includes:
  • the parameter determination unit 2062 is used to determine the influencing parameters of the environmental data.
  • the influencing parameters include at least one of the following parameters: the collection time of the environmental data, lighting and weather;
  • the second threshold determination unit 2063 is configured to determine the preset threshold corresponding to each data set based on the influence parameter determined by the parameter determination unit 2062.
  • the electronic device may be an electronic device including the map creation device 100 shown in FIG. 1 , or, in another exemplary embodiment of the present disclosure, the electronic device may be an electronic device including the map creation device 100 shown in FIG. 1 Electronic devices of map creation device 100, first device 200 and second device 300 are shown. Of course, the electronic device can also be in other forms, which is not limited by this disclosure.
  • FIG. 12 shows a block diagram of an electronic device provided by an embodiment of the present disclosure.
  • the electronic device 11 includes one or more processors 111 and a memory 112 .
  • the processor 111 may be a central processing unit (CPU) or other form of processing unit with data processing capabilities and/or instruction execution capabilities, and may control other components in the electronic device 11 to perform desired functions.
  • CPU central processing unit
  • the processor 111 may be a central processing unit (CPU) or other form of processing unit with data processing capabilities and/or instruction execution capabilities, and may control other components in the electronic device 11 to perform desired functions.
  • Memory 112 may include one or more computer program products, which may include various forms of computer-readable storage media, such as volatile memory and/or non-volatile memory.
  • Volatile memory may include, for example, random access memory (random access memory, RAM) and/or cache memory (cache), etc.
  • Non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc.
  • One or more computer program instructions may be stored on a computer-readable storage medium, and the processor 111 may execute the program instructions to implement the above map creation methods of various embodiments of the present disclosure and/or other desired functions.
  • Various contents such as environmental data and created high-precision maps may also be stored in the computer-readable storage medium.
  • the electronic device 11 may further include an input device 113 and an output device 114, and these components are interconnected through a bus system and/or other forms of connection mechanisms (not shown).
  • the input device 113 may also include, for example, a keyboard, a mouse, and the like.
  • the output device 114 can output various information to the outside, including created high-precision maps, etc.
  • the output device 114 may include, for example, a display, a speaker, a printer, a communication network and remote output devices connected thereto, and the like.
  • the electronic device 11 may also include any other appropriate components depending on the specific application.
  • embodiments of the present disclosure may also be a computer program product, which includes computer program instructions that, when executed by a processor, cause the processor to perform the “exemplary method” described above in this specification
  • the steps in the map creation method according to various embodiments of the present disclosure are described in Sec.
  • the computer program product may have program code for performing operations of embodiments of the present disclosure written in any combination of one or more programming languages, including object-oriented programming languages such as Java, C++, etc., and Includes conventional procedural programming languages, such as the "C" language or similar programming languages.
  • the program code may execute entirely on the user's computing device, partly on the user's device, as a stand-alone software package, partly on the user's computing device and partly on a remote computing device, or entirely on the remote computing device or server execute on.
  • embodiments of the present disclosure may also be a computer-readable storage medium having computer program instructions stored thereon.
  • the computer program instructions when executed by a processor, cause the processor to execute the above-mentioned “example method” part of this specification.
  • the steps in the map creation method according to various embodiments of the present disclosure are described in .
  • Computer-readable storage media can take the form of any combination of one or more computer-readable media.
  • the readable medium may be a readable signal medium or a readable storage medium.
  • the readable storage medium may include, for example, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or devices, or any combination thereof.
  • readable storage media include: electrical connections with one or more wires, portable disks, hard disks, random access memory (RAM), read-only memory (read -only memory (ROM), erasable programmable read-only memory (EPROM) or flash memory, optical fiber, portable compact disk read-only memory (compact disc read-only memory (CD-ROM)), optical memory device, magnetic memory device, or any suitable combination of the above.

Abstract

本公开提供一种地图创建方法及装置。该方法中，首先获取移动设备针对目标区域采集的环境数据；然后对环境数据分类，获取包含对应同一类路面元素的环境数据的至少一个数据集合；再基于各数据集合内包含的各环境数据对应的三维空间坐标，确定各数据集合分别对应的置信度；基于各数据集合分别对应的置信度，创建高精地图。本公开实施例提供的方案能够创建较高精度的地图，从而解决现有技术创建的地图准确性低的问题，并且，本公开实施例提供的方案无需采用高性能的移动设备，从而能够通过降低的成本获取高精度的地图，具有较广的应用范围。

Description

一种地图创建方法及装置
本公开要求在2022年09月05日提交的、申请号为202211080493.5、发明名称为“一种地图创建方法及装置”的中国专利申请的优先权，其全部内容通过引用结合在本公开中。
技术领域
本公开涉及地图创建技术领域，尤其是一种地图创建方法及装置。
背景技术
在辅助驾驶或自动驾驶等应用场景下，车辆通常需要确定周边环境中包含的路面元素(例如车道线、道路边沿和路标等)的类型和位置，以便据此调整行驶行为。
为了满足车辆的这一需求，目前通常为车辆构建显示周边环境的地图。在构建地图时，首先获取由采集设备(例如深度相机或激光雷达等)采集到的环境数据，然后根据该环境数据创建地图。
但是，该环境数据中可能包含准确度较低的数据，从而导致地图存在准确性低的问题，因此亟需一种能够构建较高精度的地图的方案。
发明内容
为了解决上述技术问题，提出了本公开。本公开的实施例提供了一种地图创建方法及装置。
根据本公开的一个方面，提供了一种地图创建方法，包括：
获取移动设备针对目标区域采集的环境数据；
通过对环境数据进行分类，确定至少一个数据集合，其中，各数据集合内包含的环境数据对应同一类路面元素；
基于各数据集合内包含的各环境数据对应的三维空间坐标，确定各数据集合分别对应的置信度；
基于各数据集合分别对应的置信度，创建高精地图。
根据本公开实施例的又一个方面，提供一种地图创建装置，包括：
数据获取模块，用于获取移动设备针对目标区域采集的环境数据；
集合确定模块，用于通过对数据获取模块获取的环境数据进行分类，确定至少一个数据集合，其中，各数据集合内包含的环境数据对应同一类路面元素；
置信度确定模块，用于基于集合确定模块确定的各数据集合内包含的各环境数据对应的三维空间坐标，确定各数据集合分别对应的置信度；
地图创建模块，用于基于置信度确定模块确定的各数据集合分别对应的置信度，创建高精地图。
根据本公开实施例的又一个方面，提供了一种计算机可读存储介质，存储介质存储有计算机程序，计 算机程序用于执行本公开上述任一实施例的地图创建方法。
根据本公开实施例的又一个方面，提供了一种电子设备，电子设备包括：
处理器；
用于存储处理器可执行指令的存储器；
处理器，用于从存储器中读取可执行指令，并执行指令以实现本公开上述任一实施例的地图创建方法。
基于本公开实施例提供的方案，能够在获取目标区域的环境数据之后，将环境数据划分为至少一个数据集合，各数据集合内包含的环境数据对应同一类路面元素，然后确定各数据集合内包含的环境数据分别对应的三维空间坐标，确定各数据集合分别对应的置信度，再基于该置信度创建高精地图。与现有技术的方案相比，本公开实施例提供的方案能够创建较高精度的地图，解决现有技术创建的地图准确性低的问题。
进一步的，本公开实施例提供的方案对采集环境数据的移动设备的性能要求较低，无需采用高性能的移动设备，从而能够通过降低的成本获取高精度的地图，具有较广的应用范围。
下面通过附图和实施例，对本公开的技术方案做进一步的详细描述。
附图说明
图1是本公开所适用的一种场景图。
图2是本公开一示例性实施例提供的地图创建方法的流程示意图。
图3是本公开另一示例性实施例提供的地图创建方法的流程示意图。
图4是本公开一示例性实施例提供的环境数据分布的示例图。
图5(a)是本公开另一示例性实施例提供的环境数据分布的示例图。
图5(b)是本公开另一示例性实施例提供的环境数据分布的示例图。
图6是本公开另一示例性实施例提供的地图创建方法的流程示意图。
图7是本公开另一示例性实施例提供的地图创建方法的流程示意图。
图8是本公开另一示例性实施例提供的地图创建方法的流程示意图。
图9是本公开另一示例性实施例提供的地图创建方法的流程示意图。
图10是本公开一示例性实施例提供的地图创建装置的结构示意图。
图11是本公开另一示例性实施例提供的地图创建装置的结构示意图。
图12是本公开一示例性实施例提供的电子设备的结构图。
具体实施方式
下面，将参考附图详细地描述根据本公开的示例实施例。显然，所描述的实施例仅仅是本公开的一部分实施例，而不是本公开的全部实施例，应理解，本公开不受这里描述的示例实施例的限制。
应注意到：除非另外具体说明，否则在这些实施例中阐述的部件和步骤的相对布置、数字表达式和数 值不限制本公开的范围。
本公开概述
在辅助驾驶或自动驾驶等应用场景下，车辆往往需要确定周边环境中存在的路面元素(例如车道线、道路边沿和路标等)的类型和位置，以便据此调整车辆的行驶行为。例如，如果确定车辆前方包含路标，可适当根据该路标的指示调整车辆的行驶方向。为了满足车辆的这一需求，目前通常为车辆构建包含周边环境的地图。
在构建地图时，目前通常由采集设备采集周边环境中各个路面元素的环境数据，该环境数据包括用于指示路面元素的类型和位置的数据，示例性的，该环境数据可包括路面元素对应的点云数据集和/或深度图像等；然后通过对环境数据的聚类处理，确定各个路面元素的类型和位置，进一步据此构建指示周边环境的地图。
其中，点云数据集包括多个点，每个点指的是相应的点云数据。在一个示例中，该点云数据集可通过激光雷达测量确定(即采集设备包括激光雷达)，激光雷达发射的激光在遇到路面元素某一位置后发生反射，并由该激光雷达接收反射的光线，那么该点云数据集中的每个点(即点云数据)通常包括该路面元素的某一位置的三维坐标和激光反射强度。
但是，采集设备采集到的环境数据有时准确性较低。例如，如果该环境数据包括路面元素的深度图像，在阴雨天气下，采集设备拍摄的深度图像有时较为模糊，导致深度图像中的路面元素清晰度较低，相应的环境数据的准确性下降。另外，如果周边环境存在树木或建筑物等遮蔽物，路面元素反射的激光可能受到遮蔽物的影响，无法被激光雷达接收，导致点云数据分布稀疏，相应的使激光雷达确定的点云数据集的准确性下降。
当采集到的路面元素的数据不准确时，基于该数据所创建的地图的准确性通常也较低。因此，通过现有方案创建地图时，得到的地图的准确度通常较低。
另外，为了提高地图的精度，目前一种改进方案是应用高性能的采集设备，以便通过高性能的采集设备获取准确度较高的环境数据。但是，高性能的采集设备的价格通常较为高昂，该方案会导致创建地图的成本增加，因此无法得到广泛应用。
有鉴于此，本公开实施例提供一种地图创建方法及装置。通过本公开的方案创建地图的过程中，在获取环境数据后，会确定环境数据的置信度，并基于较高置信度的环境数据创建地图。
通过本公开实施例的方案构建地图时，由于应用置信度较高的环境数据创建地图，因此，能够获取高精度的地图，满足车辆的需求。并且，该方案无需采用高性能的采集设备，创建高精地图的成本较低，因此应用范围较广。
示例性系统
本公开实施例可应用于需要创建地图的应用场景中，该应用场景可包括辅助驾驶或自动驾驶等应用场景。
例如，在辅助驾驶或自动驾驶的应用场景中，可由车辆中设置的深度相机和激光雷达等采集设备采集车辆周边环境对应的环境数据，并通过本公开提供的方案，利用环境数据为车辆创建高精地图，以便该地图为辅助驾驶或自动驾驶提供参考。
用于实现本公开实施例的地图创建方法的设备可为计算机、智能驾驶控制设备或服务器(例如车载服务器)等电子设备，并且在图1中公开一种该设备的示例图。
参见图1，基于本公开实施例提供的方案创建地图的地图创建设备100与用于采集环境数据的设备相连接，该设备通常设置在车辆上，并随车辆的移动而移动，在本公开中，将用于采集环境数据的设备称为移动设备。
在图1中，该移动设备包括与地图创建设备100相连接的第一设备200和第二设备300。在实际应用场景中，还可包括更多与地图创建设备100相连接并用于采集环境数据的移动设备，本公开对此不做限定。
其中，第一设备200和第二设备300可采集同一种类型的环境数据，也可采集不同类型的环境数据。在一个示例中，其中一个移动设备可包括深度相机，该移动设备采集的环境数据可包括路面元素对应的深度图像，另一个移动设备可包括激光雷达，该移动设备采集的环境数据可包括路面元素对应的点云数据集。或者，在另一示例中，第一设备200和第二设备300采集的环境数据均可包括路面元素对应的点云数据集，或均可包括路面元素对应的深度图像。
另外，第一设备200和第二设备300与地图创建设备100的连接方式可包括多种。其中一种可行的方式中，地图创建设备100可与第一设备200、第二设备300电连接，例如，该地图创建设备100为车机设备，这种情况下，地图创建设备100与第一设备200和第二设备300电连接。
在另外一种可行的方式中，地图创建设备100与第一设备200、第二设备300之间的连接方式也可为网络连接。
当然，地图创建设备100与第一设备200，以及地图创建设备100与第二设备300之间的连接方式可为同一种连接方式，也可为不同的连接方式，本公开对此不做限定。
第一设备200和第二设备300可针对目标区域采集环境数据，在一个示例中，本公开实施例提供的方案应用于辅助驾驶或自动驾驶领域，则该目标区域可包括车辆周边环境的区域，该环境数据可包括用于指示目标区域中包括的路面元素的类型及位置的数据，例如，该环境数据可包括路面元素对应的点云数据集，或者路面元素对应的深度图像，或者同时包括路面元素对应的点云数据集和深度图像。当然，该环境数据还可包括其他可指示路面元素的类型及位置的数据，本公开对此不作限定。
第一设备200和第二设备300可将采集的环境数据传输至地图创建设备100。地图创建设备100在获取该环境数据之后，执行本公开实施例提供的地图创建方法，以创建高精地图。
在辅助驾驶或自动驾驶的应用场景中，地图创建设备100可为车机设备(例如车载的智能终端)，这种情况下，地图创建设备100可基于接收到的环境数据，为车辆创建高精地图，满足车辆的需求。
示例性方法
图2是本公开一示例性实施例提供的地图创建方法的流程示意图。本实施例可应用在电子设备上，如图2所示，包括如下步骤：
步骤S201、获取移动设备针对目标区域采集的环境数据。
其中，目标区域通常包括需要创建地图的区域，例如，在辅助驾驶或自动驾驶的应用场景下，该目标区域通常包括车辆所在道路的周边环境的区域。
该环境数据包括用于指示路面元素的类型和位置的数据。示例性的，该环境数据可包括路面元素对应的深度图像，这种情况下，移动设备可包括能够拍摄深度图像的深度相机。或者，在另一示例中，该环境数据可包括路面元素对应的点云数据集，这种情况下，移动设备可包括能够生成点云数据集的激光雷达。或者，在又一示例中，该环境数据可同时包括路面元素对应的深度图像和点云数据集，或者还包括其他可指示路面元素的类型和位置的数据。
步骤S202、通过对环境数据进行分类，确定至少一个数据集合。
其中，各数据集合内包含的环境数据对应同一类路面元素。
在这一步骤中，可通过对环境数据的分析处理，确定环境数据的语义，并据此对环境数据分类，将对应同一类路面元素的环境数据划分至同一个数据集合中。
在一个示例中，移动设备采集的环境数据包括道路边沿对应的点云数据集，通过步骤S202，可确定一个数据集合，该数据集合内包含的环境数据对应道路边沿这一路面元素。
在另一个示例中，移动设备采集的环境数据包括道路边沿对应的点云数据集、车道线对应的点云数据集、道路边沿对应的深度图像和车道线对应的深度图像，则通过步骤S202的操作，可得到道路边沿对应的数据集合，该数据集合中包括道路边沿对应的点云数据集和深度图像，以及得到车道线对应的数据集合，该数据集合中包括车道线对应的点云数据集和深度图像。
步骤S203、基于各数据集合内包含的各环境数据对应的三维空间坐标，确定各数据集合分别对应的置信度。
其中，数据集合的置信度可用于表征数据集合的准确度。在本公开实施例中，通常某一数据集合的置信度越高，则该数据集合的准确度越高。
步骤S204、基于各数据集合分别对应的置信度，创建高精地图。
由于各数据集合的置信度可反映各数据集合的准确度，因此，基于各数据集合的置信度，可确定其中准确度较高的数据集合，从而能够创建高精地图。
基于本公开实施例提供的方案，能够在获取目标区域的环境数据之后，将环境数据划分为至少一个数据集合，各数据集合内包含的环境数据对应同一类路面元素，然后确定各数据集合内包含的环境数据分别 对应的三维空间坐标，确定各数据集合分别对应的置信度，再基于该置信度创建高精地图。
因此，与现有技术的方案相比，本公开实施例提供的方案能够创建较高精度的地图，解决现有技术创建的地图准确性低的问题。
进一步的，本公开实施例提供的方案对采集环境数据的移动设备的性能要求较低，无需采用高性能的移动设备，从而能够通过降低的成本获取高精度的地图，具有较广的应用范围。
在本公开的步骤S203中，基于各数据集合内包含的各环境数据对应的三维空间坐标，确定各数据集合分别对应的置信度。在一种可行的实现方式中，如图3所示，在上述图2所示实施例的基础上，步骤S203可包括如下步骤：
步骤S2031、基于各数据集合内包含的各环境数据的三维空间坐标，确定路面元素的第一参数分别与三个第二参数的三条关系曲线。
在步骤S2031中，通过对同一数据集合内包含的各环境数据进行数据拟合，确定该数据集合对应的路面元素的三条关系曲线。这三条关系曲线分别表征第一参数与其中一个第二参数中之间的关系。
其中，第一参数为路面元素的观测点距离关系曲线的起点的弧长，各第二参数分别为观测点和观测点在关系曲线的投影点之间在三维方向的距离。
在本公开实施例提供的方案中，观测点指的是路面元素表面的一点。如果环境数据包括路面元素对应的点云数据集，移动设备发出的激光在到达路面元素的观测点之后会发生反射，移动设备接收发射的激光，并据此生成点云数据集。另外，如果环境数据包括路面元素对应的深度图像，移动设备在不同角度对目标环境进行拍摄时，可获取针对路面元素不同角度的观测点的深度图像。
基于步骤S2031的操作，可获取各数据集合对应的路面元素的三条关系曲线，其中，各条关系曲线可通过一个三次多项式表示，示例性的，该三条关系曲线可通过以下公式(1)至公式(3)表示：
rx＝a0*t*t*t+b0*t*t+c0*t+d0-x_obs                    (1)；
ry＝a1*t*t*t+b1*t*t+c1*t+d1-y_obs                    (2)；
rz＝a2*t*t*t+b2*t*t+c2*t+d2-z_obs                    (3)。
上述三个公式分别对应三条关系曲线，其中，t为第一参数，即路面元素的观测点距离关系曲线的起点的弧长；rx、ry和rz均为第二参数，rx表示观测点与该观测点在关系曲线的投影点之间在x方向的距离，ry表示观测点与该观测点在关系曲线的投影点之间在y方向的距离，rz表示观测点与该观测点在关系曲线的投影点之间在z方向的距离；a0、b0、c0、d0、a1、b1、c1、d1、a2、b2、c2和d2为三次多项式的系数，x_obs、y_obs和z_obs分别表示关系曲线的常数项。
通过对各数据集合内包含的环境数据进行拟合计算，可以确定a0、b0、c0、d0、a1、b1、c1、d1、a2、 b2、c2、d2、x_obs、y_obs和z_obs的具体数值，从而可以得到相应的关系曲线。
步骤S2032、基于关系曲线，确定各数据集合内包含的各环境数据分布的聚合程度。
在本公开实施例提供的方案中，各数据集合内包含的环境数据对应同一类路面元素，也就是说，基于某一数据集合内包含的环境数据，可确定该数据集合对应的某一类路面元素的类型和位置。
这种情况下，如果某一数据集合内包含的环境数据分布较为集中，则表明该数据集合内包含的环境数据的聚合程度较高，该数据集合内包含的环境数据的准确度较高，基于该数据集合确定的路面元素的类别和位置与实际环境较接近。
相应的，如果某一数据集合内包含的环境数据分布较分散，即该数据集合内包含的环境数据的聚合程度较低，那么该数据集合内包含的环境数据的准确度较低，基于该数据集合确定的路面元素的类别和位置与实际环境之间具有较大的偏差。
也就是说，基于数据集合内包含的环境数据确定路面元素的准确度，通常与该数据集合内包含的环境数据的聚合程度相关，该聚合程度越高，基于该数据集合内包含的环境数据确定路面元素的准确度越高。
参见图4所示的示意图，在一个示例中，路面元素包括车道线和道路边沿，在对环境数据进行分类之后，可确定车道线对应的第一数据集合和道路边沿对应的第二数据集合。基于第一数据集合内包含的环境数据确定的车道线的分布情况如图4中的实线所示，基于第二数据集合内包含的环境数据确定的道路边沿的分布情况如图4中的虚线所示，另外，包含箭头的线为车辆在该示例中的行驶路线。
在图4中，车道线对应的实线与道路边沿对应的虚线相比，分布更为集中，因此第一数据集合内包含的环境数据分布的聚合程度更高。这种情况下，基于第一数据集合内包含的环境数据所确定的车道线，往往比基于第二数据集合内包含的环境数据所确定的道路边沿更接近实际环境。
在另一个示例中，某一数据集合内包含的各环境数据分布情况如图5(a)所示，另一数据集合内包含的各环境数据分布情况如图5(b)所示，二者对比，可确定图5(b)对应的数据集合内包含的环境数据分布的聚合程度更高。相应的，基于图5(b)对应的数据集合内包含的环境数据所确定的路面元素的准确性更高。
步骤S2033、基于各环境数据分布的聚合程度，确定各数据集合分别对应的置信度。
数据集合的准确度通常与该数据集合内包含的各环境数据的聚合程度相关，通常某一数据集合内各环境数据分布的聚合程度越高，该数据集合对应的置信度越高，相应的，通过该数据集合确定的路面元素的准确度越高。因此，在本公开实施例中，可基于数据集合内各环境数据分布的聚合程度，确定各数据集合分别对应的置信度。
在本公开实施例提供的方案中，确定数据集合内包含的环境数据分布的聚合程度，再通过该聚合程度 确定各数据集合分别对应的置信度。其中，数据集合的准确度通常与该数据集合内包含的各环境数据的聚合程度相关，而各数据集合的置信度基于数据集合内各环境数据分布的聚合程度确定，因此，通过本公开实施例的方案所确定的置信度，能够用于评估各数据集合内环境数据的准确度。
参见图6，在本公开另一示例性实施例中，可在上述图3所示实施例的基础上，通过以下步骤确定各数据集合内包含的各环境数据分布的聚合程度：
步骤S20321、基于关系曲线，以及各数据集合内的各环境数据的三维空间坐标，确定各数据集合分别对应的残差矩阵。
在一个示例中，某一数据集合对应的残差矩阵可通过下式(4)表示：
其中，R表示其中一个数据集合对应的残差矩阵，该数据集合包括n个环境数据，n为正整数，(rx0,ry0,rz0)、(rx1,ry1,rz1)至(rxn,ryn,rzn)为该数据集合包含的n个环境数据的三维空间坐标，r为常数。
步骤S20322、基于各数据集合分别对应的残差矩阵，确定各数据集合分别对应的协方差矩阵。
在本公开中，环境数据的残差的分布通常近似符合正态分布，而正态分布通常遵循下式(5)：
上述公式中，x表示rx、ry和rz；rx、ry和rz均为第二参数，rx表示观测点与该观测点在关系曲线的投影点之间在x方向的距离，ry表示观测点与该观测点在关系曲线的投影点之间在y方向的距离，rz表示观测点与该观测点在关系曲线的投影点之间在z方向的距离；μ表示协方差矩阵；T表示环境数据的维度，由于在本公开实施例中，基于环境数据的三维空间坐标确定各数据集合分别对应的残差矩阵，因此k为3；μ表示rx、ry和rz的平均值；T为矩阵的转置符号。
据此，基于公式(4)，数据集合的协方差矩阵可通过下式(6)表示：
∑＝R×RT                    (6)。
在公式(6)中，∑表示其中一个数据集合的协方差矩阵，该数据集合对应的残差矩阵为R，RT表示残差矩阵R的转置。
步骤S20323、基于各数据集合分别对应的协方差矩阵的迹，确定各数据集合分别对应的无偏估计量。
该无偏估计量用于表征数据集合内包含的各环境数据分布的聚合程度。
在本公开实施例中，可通过某一数据集合对应的协方差矩阵的迹表示该数据集合对应的残差矩阵的方差，该方差越小，通常表示该数据集合内各环境数据的聚合程度越好。
在一个示例中，其中一个数据集合对应的无偏估计量可通过下式(7)表示：
其中，s2表示数据集合对应的无偏估计量，n表示该数据集合内包含的环境数据的数量，tr(∑)表示该数据集合对应的协方差矩阵的迹。
基于公式(7)，可利用数据集合对应的协方差矩阵的迹，确定该数据集合对应的无偏估计量，该无偏估计量可表征该数据集合内包含的环境数据分布的聚合程度。
进一步的，数据集合的置信度通常与移动设备采集环境数据的次数正相关，其中，通常采集环境数据的次数越高，该数据集合内包含的环境数据数量越丰富，通过该数据集合确定路面元素的准确度越高。也就是说，采集环境数据的次数越高，相应的该数据集合的置信度越大。因此在确定数据集合内包含的各环境数据分布的聚合程度之后，基于各环境数据分布的聚合程度，确定各数据集合分别对应的置信度的步骤，可通过以下操作实现：
基于数据集合对应的无偏估计量，以及移动设备采集环境数据的次数，确定各数据集合分别对应的置信度。
在一个示例中，可通过下式(8)，确定数据集合的置信度：
上述公式中，score表示数据集合的置信度，该数据集合的无偏估计量为s2，tr(∑)表示该数据集合对应的协方差矩阵的迹。
其中，该数据集合内包含的环境数据的数量为n，n表示该数据集合内包含的环境数据的数量。另外，由于该数据集合内包含的数据的数量为n，并且移动设备通常每采集一次环境数据，可获取一个环境数据，因此，移动设备采集该数据集合内包含的环境数据的次数也为n。
通过本公开上述实施例提供的方案，能够基于各数据集合的关系曲线，确定各数据集合对应的无偏估计量，该无偏估计量可表征各数据集合内包含的各环境数据分布的聚合程度。并且，基于该无偏估计量以及移动设备采集环境数据的次数，可确定各数据集合分别对应的置信度，以便通过各数据集合的置信度，创建高精地图。
在本公开的各实施例中，通过对环境数据的分类，确定至少一个数据集合。在实际的地图创建过程中，可通过多种方式确定该数据集合。
在一种可行的实现方式中，基于各环境数据的语义，对环境数据进行分类，这种情况下，同一个数据集合内包含的环境数据可包括不同移动设备采集，并且对应同一类路面元素的环境数据。
相应的，如果通过上述实现方式确定数据集合，各数据集合的置信度表示的是各类型路面元素的环境 数据分别对应的置信度。
例如，如果用于采集环境数据的移动设备包括深度相机和激光雷达，并且采集的环境数据包括道路边沿对应的环境数据和车道线对应的环境数据，基于上述实现方式可确定两个数据集合，其中一个数据集合中包括由深度相机和激光雷达采集的道路边沿的环境数据，另一个数据集合中包括由深度相机和激光雷达采集的车道线的环境数据。
在另外一种可行的实现方式中，参见图7，可在上述图3所示实施例的基础上，通过以下步骤对环境数据进行分类，确定至少一个数据集合：
步骤S2021、基于各环境数据分别对应的移动设备的类型，对环境数据进行分类，确定至少一个中间数据集合；
步骤S2022、基于各环境数据分别对应的路面元素的类型，对各中间数据集合内包含的各环境数据进行分类，确定至少一个数据集合。
也就是说，在该实现方式中，首先基于采集环境数据的移动设备对环境数据进行分类，以获取中间数据集合，这种情况下，同一个中间数据集合内包含的各环境数据由同一类型的移动设备采集；然后，基于环境数据对应的路面元素的类型，对中间数据集合内包含的各环境数据进行分类，以确定至少一个数据集合。
或者，在该实现方式中，还可首先基于各环境数据分别对应的路面元素的类型对环境数据进行分类，确定至少一个中间数据集合，这种情况下，同一中间数据集合内包含的各环境数据对应同一类路面元素；然后，基于各环境数据分别对应的移动设备的类型，对该中间数据集合中包含的各环境数据分类，确定至少一个数据集合。
通过上述实现方式提供的方案，所确定的各数据集合内包含的各环境数据由同一类型的移动设备采集，并且对应同一类型的路面元素。
例如，如果用于采集环境数据的移动设备包括深度相机和激光雷达，并且采集的环境数据包括道路边沿对应的环境数据和车道线对应的环境数据，基于上述实现方式可确定四个数据集合，这四个数据集合包括：由深度相机采集的道路边沿的环境数据构成的数据集合、由激光雷达采集的道路边沿的环境数据构成的数据集合、由深度相机采集的车道线的环境数据构成的数据集合以及由激光雷达采集的车道线的环境数据构成的数据集合。
进一步的，由于通过本公开实施例提供的确定数据集合的方案，所确定的各数据集合内包含的各环境数据由同一类型的移动设备采集，并且对应同一类型的路面元素，从而能够根据各数据集合的置信度，确定各移动数据采集环境数据的准确度。
例如，如果基于本公开实施例提供的方案确定的数据集合包括：由深度相机采集的道路边沿的环境数据构成的第三数据集合，以及由激光雷达采集的道路边沿的环境数据构成的第四数据集合，如果第三数据 集合的置信度高于第四数据集合的置信度，则表明深度相机采集的环境数据的准确度高于激光雷达采集的环境数据的准确度。
相应的，参见图8，在上述图7所示实施例的基础上，还可包括以下步骤：
步骤S2023、确定各数据集合中的第一目标数据集合。
其中，第一目标数据集合的置信度小于各数据集合中的其他数据集合的置信度。
在本公开实施例中，第一目标数据集合的数量可为N，N为预设的正整数。
步骤S2024、基于第一目标数据集合对应的移动设备，生成提示信息。
该提示信息用于指示该第一目标数据集合对应的移动设备在采集环境数据时存在准确度低的问题。
通过步骤S2021至步骤S2024的操作，所确定的各数据集内包含的环境数据由同一类型的移动设备采集，并且对应同一类型的路面元素，相应的，各数据集合的置信度能够反映各移动数据采集环境数据的准确度。
由于第一目标数据集合的置信度小于其他数据集合的置信度，因此，采集第一目标数据集合内各环境数据的移动设备的准确度较低，通过该提示信息，可提示技术人员对该移动设备进行检修处理，以便提高该移动设备采集环境数据的准确度，进一步提高创建地图的准确度。
在本公开上述实施例中，提供基于各数据集合分别对应的置信度，创建高精地图的操作，该操作可通过多种方式实现。
在其中一种可行的实现方式中，基于各数据集合分别对应的置信度，创建高精地图的操作可包括以下步骤：
首先，基于各数据集合分别对应的置信度，确定第二目标数据集合，第二目标数据集合的置信度大于其他数据集合的置信度。
其中，第二目标数据集合的数量可预先设定，并基于各数据集合分别对应的置信度以及预先设定的第二目标数据集合的数量，确定该第二目标数据集合。或者，可预先设定一个置信度阈值，如果某一数据集合的置信度大于该置信度阈值，则确定该数据集合为第二目标数据集合。
由于第二目标数据集合的置信度大于其他数据集合的置信度，因此，通过第二目标数据集合确定的路面元素的准确度较高。
然后，创建各数据集合对应的高精地图，并在高精地图中标记显示通第二目标数据集合确定的路面元素。
设定通过第二目标数据集合确定的路面元素为高准确度路面元素，在高精地图中对通过该高准确度路面元素进行标记显示，可对高精地图中的高准确度路面元素和其他的路面元素进行区分。
对该高准确度路面元素进行标记显示的方式可包括多种。在一个示例中，可在高精地图中通过高亮的形式显示该高准确度路面元素。或者，在另一个示例中，可通过与其他路面元素不同的色彩在高精地图中 显示该高准确度路面元素。
通过这一实现方式创建高精地图时，可基于各数据集合确定的路面元素的类型和位置创建地图，并在该地图中，标记通过第二目标数据确定的路面元素，而通过第二目标数据集合确定的路面元素的准确度较高，因此，该高精地图对准确度较高的路面元素进行标记显示，可实现对高精地图中的高准确度路面元素和其他的路面元素的区分，便于确定高精地图中的高准确度路面元素。
在另外一种可行的实现方式中，基于各数据集合分别对应的置信度，创建高精地图的操作可包括以下步骤：
首先，基于各数据集合分别对应的置信度，确定第三目标数据集合，第三目标数据集合的置信度大于其他数据集合的置信度。
然后，基于第三目标数据集合创建高精地图。
这一实现方式中，仅通过置信度较高的第三目标数据集合进行地图的创建，因此，所创建的地图的准确度较高。
在实际的应用场景中，为了保障导航的准确度，车辆可同时通过地图和其他方式进行导航。例如，车辆还可通过全球定位系统(global positioning system，GPS)进行导航，以及还可通过惯性导航系统(inertial navigation system，INS)进行导航。
基于本公开上述实施例提供的方案，可确定各数据集合的置信度，这种情况下，还可根据该数据集合的置信度，调整不同导航方式的权重。针对这一情况，如果移动设备包括至少两种导航方式，该至少两种导航方式中包括基于高精地图的第一地图导航方式，参见图9，在上述图2所示实施例的基础上，在本公开实施例提供的方案中，还包括以下步骤：
步骤S205、基于各数据集合分别对应的置信度与各数据集合分别对应的预设阈值，确定应用第一地图导航方式的权重。
其中，如果各数据集合分别对应的置信度大于各数据集合分别对应的预设阈值，则表明各数据集合内包含的环境数据的准确度较高，相应的基于该数据集合创建的高精地图的准确度较高，这种情况下，可相应提高应用第一地图导航方式进行导航的权重。
另外，如果各数据集合分别对应的置信度不大于各数据集合分别对应的预设阈值，则表明各数据集合内包含的环境数据的准确度较低，这种情况下，可相应降低应用第一地图导航方式进行导航的权重。在这一操作中，各数据集合分别对应的置信度不大于各数据集合分别对应的预设阈值，指的是各数据集合中任意一个数据集合对应的置信度不大于该数据集合对应的预设阈值。
在一种可行的示例中，可预先设置第一差值与应用第一地图导航方式的权重之间的对应关系，该第一差值为各数据集合分别对应的置信度与预设阈值的差值。在该对应关系中，第一差值越大，通常应用第一地图导航方式的权重越高。这种情况下，基于该对应关系，即可确定应用第一地图导航方式的权重。
步骤S206、基于确定的应用第一地图导航方式的权重为移动设备进行导航。
通过本公开实施例提供的方案，可基于各数据集合分别对应的置信度，调整应用第一地图导航方式进行导航的权重，从而能够基于各数据集合分别对应的置信度调整导航方式，保障导航方式的准确度，提高车辆的安全性。
为了明确本公开实施例的有益效果，以下公开两个示例。在一个示例中，移动设备的导航方式包括第一地图导航方式和利用INS进行导航的INS导航方式。这种情况下，如果各数据集合分别对应的置信度大于该数据集合对应的预设阈值，则提高第一地图导航方式的权重，并结合INS导航方式与提高权重后的第一地图导航方式共同进行导航。
这一示例中，基于数据集合分别对应的置信度与预设阈值的比较，调整了第一地图导航方式在各种导航方式中的权重，相应能够提高导航的准确度。
在另一示例中，移动设备的导航方式包括：第一地图导航方式、利用INS进行导航的INS导航方式以及利用GPS进行导航的GPS导航方式。这种情况下，如果各数据集合分别对应的置信度大于该数据集合对应的预设阈值，则提高第一地图导航方式的权重。另外，由于GPS信号的强度较易影响GPS导航的准确度，还可确定当前GPS信号的强度，如果当前的GPS信号较弱，则降低GPS导航方式的权重。然后，基于调整权重后的导航方式进行导航。
这一示例中，基于数据集合分别对应的置信度与预设阈值的比较，调整第一地图导航方式在各种导航方式中的权重，并且基于GPS信号的强度，调整GPS导航方式在各种导航方式中的权重，相应能够提高导航的准确度。
在本公开的上述实施例中，基于各数据集合分别对应的置信度与预设阈值之间的比较，确定应用第一地图导航方式的权重，其中，该预设阈值可通过多种方式确定。
在一种可行的实现方式中，各数据集合分别对应的预设阈值通过如下方式确定：
基于路面元素与预设阈值之间的对应关系，确定各数据集合分别对应的预设阈值。
在这一实现方式中，预先设定各路面元素与预设阈值之间的对应关系，然后基于该对应关系，以及各数据集合分别对应的路面元素，即可确定各数据集合分别对应的预设阈值。
示例性的，在确定该对应关系时，可根据路面元素的体积大小，确定该路面元素对应的预设阈值。这种情况下，由于路面元素的体积越大越容易被识别到，因此，通常路面元素的体积越大，在该对应关系中对应的预设阈值越大。
或者，可根据对路面元素的重视程度，确定该路面元素对应的预设阈值，这种情况下，对某一路面元素越重视，该对应关系中对应的预设阈值越大。例如，如果各数据集合对应的路面元素包括路标，而在导航过程中，车辆往往需要根据路标的指示确定是否需要调整方向，因此对路标较为重视，这种情况下，可设定路标这一路面元素对应较大的预设阈值。
或者，在另一种可行的实现方式中，各数据集合分别对应的预设阈值通过如下方式确定：
第一步，确定环境数据的影响参数，该影响参数包括以下参数中的至少一种：环境数据的采集时间、光照和天气。
在实际的应用场景中，同一移动设备采集环境数据的准确度往往会受到当前环境的影响。例如，如果当前光照强度较低，周边可能存在遮蔽物，这将导致激光雷达确定点云数据集的准确度下降，并且较低的光照强度往往会降低深度相机拍摄深度图像的清晰度；如果当前为阴雨天气，深度相机拍摄深度图像的清晰度通常会下降，并且激光雷达获取点云数据集的准确度也会下降；另外，不同的采集时间对应的光照不同，例如早晨的光照强度往往低于中午的光照强度，而光照强度较强的情况下，深度相机拍摄深度图像的准确度通常较高。
第二步，基于影响参数，确定各数据集合分别对应的预设阈值。
该步骤中，通常当环境数据的影响参数会导致移动设备采集环境数据的准确度降低时，各数据集合分别对应的预设阈值相应减小；当环境数据的影响参数会导致移动设备采集环境数据的准确度提高时，各数据集合分别对应的预设阈值相应增大。
第一地图导航方式在各导航方式中的权重与各数据集合分别对应的预设阈值相关，而通过本公开实施例提供的方案，能够基于当前的影响参数，对各数据集合分别对应的预设阈值进行调整，因此，本公开实施例能够基于当前的影响参数，对第一地图导航方式的权重进行调整，进一步提高对移动设备进行导航的准确度。
示例性装置
图10是本公开一示例性实施例提供的地图创建装置的结构示意图。该地图创建装置可以设置于终端设备、服务器等电子设备中，或者车辆等对象上，执行本公开上述任一实施例的地图创建方法。如图10所示，该实施例的地图创建装置包括：数据获取模块201、集合确定模块202、置信度确定模块203和地图创建模块204。
其中，数据获取模块201，用于获取移动设备针对目标区域采集的环境数据；
集合确定模块202，用于通过对数据获取模块获取201的环境数据进行分类，确定至少一个数据集合。
其中，各数据集合内包含的环境数据对应同一类路面元素；
置信度确定模块203，用于基于集合确定模块202确定的各数据集合内包含的各环境数据对应的三维空间坐标，确定各数据集合分别对应的置信度；
地图创建模块204，用于基于置信度确定模块203确定的各数据集合分别对应的置信度，创建高精地图。
进一步的，参见图11所示的结构示意图，在一种可行的示例中，置信度确定模块203包括：
曲线确定单元2031，用于基于各数据集合内包含的各环境数据的三维空间坐标，确定路面元素的第一参数分别与三个第二参数的三条关系曲线，其中，第一参数为路面元素的观测点距离关系曲线的起点的弧 长，各第二参数分别为观测点和观测点在关系曲线的投影点之间在三维方向的距离；
聚合程度确定单元2032，用于基于曲线确定单元2031确定的关系曲线，确定各数据集合内包含的各环境数据分布的聚合程度；
置信度确定单元2033，用于基于聚合程度确定单元2032确定的各环境数据分布的聚合程度，确定各数据集合分别对应的置信度。
进一步的，聚合程度确定单元2032包括：
残差矩阵确定子单元，用于基于关系曲线，以及各数据集合内的各环境数据的三维空间坐标，确定各数据集合分别对应的残差矩阵；
协方差矩阵确定子单元，用于基于残差矩阵确定子单元确定的各数据集合分别对应的残差矩阵，确定各数据集合分别对应的协方差矩阵；
无偏估计量确定子单元，用于基于协方差矩阵确定子单元确定的各数据集合分别对应的协方差矩阵的迹，确定各数据集合分别对应的无偏估计量，该无偏估计量用于表征数据集合内包含的各环境数据分布的聚合程度。
相应的，这种情况下，置信度确定单元203包括：置信度确定子单元，该置信度确定子单元用于基于各数据集合对应的无偏估计量，以及移动设备采集环境数据的次数，确定各数据集合分别对应的置信度。
在一种可行的示例中，该集合确定模块202包括：
第一分类单元2021，用于基于各环境数据分别对应的移动设备的类型，对环境数据进行分类，确定至少一个中间数据集合；
第二分类单元2022，基于各环境数据分别对应的路面元素的类型，对第一分类单元确定的各中间数据集合内包含的各环境数据进行分类，确定至少一个数据集合。
进一步的，在本公开实施例提供的方案中，还包括：提示模块205。该提示模块205用于在确定至少一个数据集合之后，确定各数据集合中的第一目标数据集合，第一目标数据集合的置信度小于各数据集合中的其他数据集合的置信度；并且，该提示模块205还用于基于第一目标数据集合对应的移动设备，生成提示信息。
另外，在一种可行的示例中，该地图创建模块204包括：
目标集合确定单元2041，用于基于各数据集合分别对应的置信度，确定第二目标数据集合，第二目标数据集合的置信度大于其他数据集合的置信度；
地图创建单元2042，用于创建各数据集合对应的高精地图，并在高精地图中标记显示通过目标集合确定单元2041确定的第二目标数据集合确定的路面元素。
进一步的，如果移动设备包括至少两种导航方式，该至少两种导航方式中包括基于高精地图的第一地图导航方式，本公开实施例提供的地图创建装置中，还包括：导航模块206。
该导航模块206用于在基于各数据集合分别对应的置信度，创建高精地图之后，基于各数据集合分别对应的置信度与各数据集合分别对应的预设阈值，确定应用第一地图导航方式的权重；并且，该导航模块206还用于基于该权重为移动设备进行导航。
其中，在一种可行的实现方式中，该导航模块206可包括：
第一阈值确定单2061，用于基于路面元素与预设阈值之间的对应关系，确定各数据集合分别对应的预设阈值。
或者，在另一种可行的实现方式中，该导航模块206包括：
参数确定单元2062，用于确定环境数据的影响参数，影响参数包括以下参数中的至少一种：环境数据的采集时间、光照和天气；
第二阈值确定单元2063，用于基于参数确定单元2062确定的影响参数，确定各数据集合分别对应的预设阈值。
示例性电子设备
下面，参考图12来描述根据本公开实施例的电子设备。在本公开一示例性实施例中，该电子设备可以是包括图1所示的地图创建设备100的电子设备，或者，在本公开另一示例性实施例中，该电子设备可以是包括图1所示的地图创建设备100、第一设备200和第二设备300的电子设备。当然，该电子设备也可以为其他形式，本公开对此不作限定。
图12示出了本公开实施例提供的电子设备的框图，电子设备11包括一个或多个处理器111和存储器112。
处理器111可以是中央处理单元(central processing unit，CPU)或者具有数据处理能力和/或指令执行能力的其他形式的处理单元，并且可以控制电子设备11中的其他组件以执行期望的功能。
存储器112可以包括一个或多个计算机程序产品，计算机程序产品可以包括各种形式的计算机可读存储介质，例如易失性存储器和/或非易失性存储器。易失性存储器例如可以包括随机存取存储器(random access memory，RAM)和/或高速缓冲存储器(cache)等。非易失性存储器例如可以包括只读存储器(read-only memory，ROM)、硬盘、闪存等。在计算机可读存储介质上可以存储一个或多个计算机程序指令，处理器111可以运行程序指令，以实现上文的本公开的各个实施例的地图创建方法以及/或者其他期望的功能。在计算机可读存储介质中还可以存储诸如环境数据和创建的高精地图等各种内容。
在一个示例中，电子设备11还可以包括：输入装置113和输出装置114，这些组件通过总线系统和/或其他形式的连接机构(未示出)互连。
此外，该输入装置113还可以包括例如键盘、鼠标等等。
该输出装置114可以向外部输出各种信息，包括创建的高精地图等。该输出装置114可以包括例如显示器、扬声器、打印机、以及通信网络及其所连接的远程输出设备等等。
当然，为了简化，图12中仅示出了该电子设备11中与本公开有关的组件中的一些，省略了诸如总线、 输入/输出接口等等的组件。除此之外，根据具体应用情况，电子设备11还可以包括任何其他适当的组件。
示例性计算机程序产品和计算机可读存储介质
除了上述方法和设备以外，本公开的实施例还可以是计算机程序产品，其包括计算机程序指令，所述计算机程序指令在被处理器运行时使得所述处理器执行本说明书上述“示例性方法”部分中描述的根据本公开各种实施例的地图创建方法中的步骤。
计算机程序产品可以以一种或多种程序设计语言的任意组合来编写用于执行本公开实施例操作的程序代码，所述程序设计语言包括面向对象的程序设计语言，诸如Java、C++等，还包括常规的过程式程序设计语言，诸如“C”语言或类似的程序设计语言。程序代码可以完全地在用户计算设备上执行、部分地在用户设备上执行、作为一个独立的软件包执行、部分在用户计算设备上部分在远程计算设备上执行、或者完全在远程计算设备或服务器上执行。
此外，本公开的实施例还可以是计算机可读存储介质，其上存储有计算机程序指令，所述计算机程序指令在被处理器运行时使得所述处理器执行本说明书上述“示例性方法”部分中描述的根据本公开各种实施例的地图创建方法中的步骤。
计算机可读存储介质可以采用一个或多个可读介质的任意组合。可读介质可以是可读信号介质或者可读存储介质。可读存储介质例如可以包括但不限于电、磁、光、电磁、红外线、或半导体的系统、装置或器件，或者任意以上的组合。可读存储介质的更具体的例子(非穷举的列表)包括：具有一个或多个导线的电连接、便携式盘、硬盘、随机存取存储器(random access memory，RAM)、只读存储器(read-only memory，ROM)、可擦式可编程只读存储器(erasable programmable read-only memory，EPROM)或闪存、光纤、便携式紧凑盘只读存储器(compact disc read-only memory，CD-ROM)、光存储器件、磁存储器件、或者上述的任意合适的组合。
以上结合具体实施例描述了本公开的基本原理，但是，需要指出的是，在本公开中提及的优点、优势、效果等仅是示例而非限制，不能认为这些优点、优势、效果等是本公开的各个实施例必须具备的。另外，上述公开的具体细节仅是为了示例的作用和便于理解的作用，而非限制，上述细节并不限制本公开为必须采用上述具体的细节来实现。
本说明书中各个实施例均采用递进的方式描述，每个实施例重点说明的都是与其它实施例的不同之处，各个实施例之间相同或相似的部分相互参见即可。对于系统实施例而言，由于其与方法实施例基本对应，所以描述的比较简单，相关之处参见方法实施例的部分说明即可。
为了例示和描述的目的已经给出了以上描述。此外，此描述不意图将本公开的实施例限制到在此公开的形式。尽管以上已经讨论了多个示例方面和实施例，但是本领域技术人员将认识到其某些变型、修改、改变、添加和子组合。

Claims (11)

  1. 一种地图创建方法，包括：
    获取移动设备针对目标区域采集的环境数据；
    通过对所述环境数据进行分类，确定至少一个数据集合，其中，各所述数据集合内包含的所述环境数据对应同一类路面元素；
    基于各所述数据集合内包含的各环境数据对应的三维空间坐标，确定各所述数据集合分别对应的置信度；
    基于各所述数据集合分别对应的置信度，创建高精地图。
  2. 根据权利要求1所述的方法，其中，所述基于各所述数据集合内包含的各环境数据对应的三维空间坐标，确定各所述数据集合分别对应的置信度，包括：
    基于各所述数据集合内包含的各环境数据的三维空间坐标，确定所述路面元素的第一参数分别与三个第二参数的三条关系曲线，其中，所述第一参数为所述路面元素的观测点距离所述关系曲线的起点的弧长，各所述第二参数分别为所述观测点和所述观测点在所述关系曲线的投影点之间在三维方向的距离；
    基于所述关系曲线，确定各所述数据集合内包含的各所述环境数据分布的聚合程度；
    基于各所述环境数据分布的聚合程度，确定各所述数据集合分别对应的置信度。
  3. 根据权利要求2所述的方法，其中，所述基于所述关系曲线，确定各所述数据集合内包含的各所述环境数据分布的聚合程度，包括：
    基于所述关系曲线，以及各所述数据集合内的各环境数据的三维空间坐标，确定各所述数据集合分别对应的残差矩阵；
    基于各所述数据集合分别对应的残差矩阵，确定各所述数据集合分别对应的协方差矩阵；
    基于各所述数据集合分别对应的协方差矩阵的迹，确定各所述数据集合分别对应的无偏估计量，所述无偏估计量用于表征所述数据集合内包含的各所述环境数据分布的聚合程度；
    所述基于各所述环境数据分布的聚合程度，确定各所述数据集合分别对应的置信度，包括：
    基于各所述数据集合对应的无偏估计量，以及所述移动设备采集所述环境数据的次数，确定各所述数据集合分别对应的置信度。
  4. 根据权利要求1所述的方法，其中，所述通过对所述环境数据进行分类，确定至少一个数据集合，包括：
    基于各所述环境数据分别对应的移动设备的类型，对所述环境数据进行分类，确定至少一个中间数据集合；
    基于各所述环境数据分别对应的路面元素的类型，对各所述中间数据集合内包含的各所述环境数据进行分类，确定至少一个数据集合。
  5. 根据权利要求4所述的方法，其中，在所述确定至少一个数据集合之后，还包括：
    确定各所述数据集合中的第一目标数据集合，所述第一目标数据集合的置信度小于各所述数据集合中的其他数据集合的置信度；
    基于所述第一目标数据集合对应的所述移动设备，生成提示信息。
  6. 根据权利要求1至5任一项所述的方法，其中，所述基于各所述数据集合分别对应的置信度，创建高精地图，包括：
    基于各所述数据集合分别对应的置信度，确定第二目标数据集合，所述第二目标数据集合的置信度大于其他数据集合的置信度；
    创建各所述数据集合对应的高精地图，并在所述高精地图中标记显示通过所述第二目标数据集合确定的所述路面元素。
  7. 根据权利要求1至5任一项所述的方法，其中，所述移动设备包括至少两种导航方式，所述至少两种导航方式中包括基于所述高精地图的第一地图导航方式，在所述基于各所述数据集合分别对应的置信度，创建高精地图之后，还包括：
    基于各所述数据集合分别对应的置信度与各所述数据集合分别对应的预设阈值，确定应用所述第一地图导航方式的权重；
    基于所述权重，为所述移动设备进行导航。
  8. 根据权利要求7所述的方法，其中，各所述数据集合分别对应的所述预设阈值通过如下方式确定，包括：
    基于所述路面元素与所述预设阈值之间的对应关系，确定各所述数据集合分别对应的预设阈值；或者，确定所述环境数据的影响参数，所述影响参数包括以下参数中的至少一种：所述环境数据的采集时间、光照和天气；
    基于所述影响参数，确定各所述数据集合分别对应的所述预设阈值。
  9. 一种地图创建装置，包括：
    数据获取模块，用于获取移动设备针对目标区域采集的环境数据；
    集合确定模块，用于通过对所述数据获取模块获取的所述环境数据进行分类，确定至少一个数据集合，其中，各所述数据集合内包含的所述环境数据对应同一类路面元素；
    置信度确定模块，用于基于所述集合确定模块确定的各所述数据集合内包含的各环境数据对应的三维空间坐标，确定各所述数据集合分别对应的置信度；
    地图创建模块，用于基于所述置信度确定模块确定的各所述数据集合分别对应的置信度，创建高精地图。
  10. 一种计算机可读存储介质，所述存储介质存储有计算机程序，所述计算机程序用于执行上述权利要求1-8任一所述的地图创建方法。
  11. 一种电子设备，所述电子设备包括：
    处理器；
    用于存储所述处理器可执行指令的存储器；
    所述处理器，用于从所述存储器中读取所述指令，并执行所述指令以实现上述权利要求1-8任一所述的地图创建方法。
PCT/CN2023/107169 2022-09-05 2023-07-13 一种地图创建方法及装置 WO2024051344A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202211080493.5A CN115930978A (zh) 2022-09-05 2022-09-05 一种地图创建方法及装置
CN202211080493.5 2022-09-05

Publications (1)

Publication Number Publication Date
WO2024051344A1 true WO2024051344A1 (zh) 2024-03-14

Family

ID=86699446

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2023/107169 WO2024051344A1 (zh) 2022-09-05 2023-07-13 一种地图创建方法及装置

Country Status (2)

Country Link
CN (1) CN115930978A (zh)
WO (1) WO2024051344A1 (zh)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115930978A (zh) * 2022-09-05 2023-04-07 北京地平线机器人技术研发有限公司 一种地图创建方法及装置

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111912419A (zh) * 2020-08-28 2020-11-10 清华大学苏州汽车研究院（吴江） 基于激光雷达的高精度语义导航地图构建方法和装置
CN113034566A (zh) * 2021-05-28 2021-06-25 湖北亿咖通科技有限公司 高精度地图构建方法、装置、电子设备及存储介质
CN114022614A (zh) * 2021-11-05 2022-02-08 武汉中海庭数据技术有限公司 一种三维重建目标位置置信度的估计方法及系统
WO2022067647A1 (zh) * 2020-09-30 2022-04-07 华为技术有限公司 一种路面要素确定方法及装置
CN115930978A (zh) * 2022-09-05 2023-04-07 北京地平线机器人技术研发有限公司 一种地图创建方法及装置

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111912419A (zh) * 2020-08-28 2020-11-10 清华大学苏州汽车研究院（吴江） 基于激光雷达的高精度语义导航地图构建方法和装置
WO2022067647A1 (zh) * 2020-09-30 2022-04-07 华为技术有限公司 一种路面要素确定方法及装置
CN113034566A (zh) * 2021-05-28 2021-06-25 湖北亿咖通科技有限公司 高精度地图构建方法、装置、电子设备及存储介质
CN114022614A (zh) * 2021-11-05 2022-02-08 武汉中海庭数据技术有限公司 一种三维重建目标位置置信度的估计方法及系统
CN115930978A (zh) * 2022-09-05 2023-04-07 北京地平线机器人技术研发有限公司 一种地图创建方法及装置

Also Published As

Publication number Publication date
CN115930978A (zh) 2023-04-07

Similar Documents

Publication Publication Date Title
EP3581890B1 (en) Method and device for positioning
CN110095752B (zh) 定位方法、装置、设备和介质
WO2020237996A1 (zh) 一种车辆位姿的修正方法和装置
CN111694903A (zh) 地图构建方法、装置、设备以及可读存储介质
US11144770B2 (en) Method and device for positioning vehicle, device, and computer readable storage medium
CN113989450B (zh) 图像处理方法、装置、电子设备和介质
EP3875905B1 (en) Method, device and medium for detecting environmental change
WO2024051344A1 (zh) 一种地图创建方法及装置
CN109635868B (zh) 障碍物类别的确定方法、装置、电子设备及存储介质
TW202018256A (zh) 多定位系統切換與融合校正方法及其裝置
CN113189610A (zh) 地图增强的自动驾驶多目标追踪方法和相关设备
WO2023138331A1 (zh) 一种构建语义地图的方法及装置
CN112639822A (zh) 一种数据处理方法及装置
CN111337010A (zh) 可移动设备的定位方法、定位装置及电子设备
CN112835063B (zh) 物体动静属性的确定方法、装置、设备及存储介质
CN112097742B (zh) 一种位姿确定方法及装置
CN112304293B (zh) 道路高度检测方法、装置、可读存储介质和电子设备
CN109859254B (zh) 自动驾驶中用于发送信息的方法和装置
CN114386481A (zh) 一种车辆感知信息融合方法、装置、设备及存储介质
CN114202574A (zh) 定位可靠性检测方法及相关设备
CN116168366B (zh) 点云数据生成方法、模型训练方法、目标检测方法和装置
CN113191368B (zh) 一种标志物的匹配方法及装置
US20230062694A1 (en) Navigation apparatus and method
US20240144612A1 (en) Vehicle ar display device and ar service platform
CN115131458A (zh) 一种高精地图数据处理方法和装置