CN111638528A - Positioning method, positioning device, electronic equipment and storage medium - Google Patents

Positioning method, positioning device, electronic equipment and storage medium Download PDF

Info

Publication number
CN111638528A
CN111638528A CN202010455341.3A CN202010455341A CN111638528A CN 111638528 A CN111638528 A CN 111638528A CN 202010455341 A CN202010455341 A CN 202010455341A CN 111638528 A CN111638528 A CN 111638528A
Authority
CN
China
Prior art keywords
point cloud
cloud data
original
precision map
target point
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202010455341.3A
Other languages
Chinese (zh)
Other versions
CN111638528B (en
Inventor
付向宇
万国伟
卢维欣
周尧
宋适宇
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beijing Baidu Netcom Science and Technology Co Ltd
Original Assignee
Beijing Baidu Netcom Science and Technology Co Ltd
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 Beijing Baidu Netcom Science and Technology Co Ltd filed Critical Beijing Baidu Netcom Science and Technology Co Ltd
Priority to CN202010455341.3A priority Critical patent/CN111638528B/en
Publication of CN111638528A publication Critical patent/CN111638528A/en
Application granted granted Critical
Publication of CN111638528B publication Critical patent/CN111638528B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S17/00—Systems using the reflection or reradiation of electromagnetic waves other than radio waves, e.g. lidar systems
    • G01S17/88—Lidar systems specially adapted for specific applications
    • G01S17/93—Lidar systems specially adapted for specific applications for anti-collision purposes
    • G01S17/931—Lidar systems specially adapted for specific applications for anti-collision purposes of land vehicles
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S19/00—Satellite radio beacon positioning systems; Determining position, velocity or attitude using signals transmitted by such systems
    • G01S19/38—Determining a navigation solution using signals transmitted by a satellite radio beacon positioning system
    • G01S19/39—Determining a navigation solution using signals transmitted by a satellite radio beacon positioning system the satellite radio beacon positioning system transmitting time-stamped messages, e.g. GPS [Global Positioning System], GLONASS [Global Orbiting Navigation Satellite System] or GALILEO
    • G01S19/42—Determining position
    • G01S19/45—Determining position by combining measurements of signals from the satellite radio beacon positioning system with a supplementary measurement
    • G01S19/46—Determining position by combining measurements of signals from the satellite radio beacon positioning system with a supplementary measurement the supplementary measurement being of a radio-wave signal type
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S19/00—Satellite radio beacon positioning systems; Determining position, velocity or attitude using signals transmitted by such systems
    • G01S19/38—Determining a navigation solution using signals transmitted by a satellite radio beacon positioning system
    • G01S19/39—Determining a navigation solution using signals transmitted by a satellite radio beacon positioning system the satellite radio beacon positioning system transmitting time-stamped messages, e.g. GPS [Global Positioning System], GLONASS [Global Orbiting Navigation Satellite System] or GALILEO
    • G01S19/42—Determining position
    • G01S19/45—Determining position by combining measurements of signals from the satellite radio beacon positioning system with a supplementary measurement
    • G01S19/47—Determining position by combining measurements of signals from the satellite radio beacon positioning system with a supplementary measurement the supplementary measurement being an inertial measurement, e.g. tightly coupled inertial
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/22—Matching criteria, e.g. proximity measures

Landscapes

  • Engineering & Computer Science (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Remote Sensing (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • Electromagnetism (AREA)
  • Navigation (AREA)

Abstract

The application discloses a positioning method, a positioning device, electronic equipment and a storage medium, and relates to the technical field of automatic driving. The method comprises the following steps: in the original point cloud data, determining a preset number of original point clouds as target point clouds according to the sequence of the weights of the features of the original point clouds from high to low, and determining the target point cloud data according to the target point clouds; generating a high-precision map according to the target point cloud data; and when receiving the request information from the equipment to be positioned, sending a high-precision map to the equipment to be positioned. Because the point cloud data in the high-precision map is not all original point cloud data but the point cloud data with higher weight determined according to the characteristics of the point cloud, the data volume of the point cloud data in the high-precision map can be reduced on the premise of not influencing the positioning accuracy, so that the storage space occupied by the high-precision map is reduced, and the positioning speed when the high-precision map is used for positioning is further improved.

Description

Positioning method, positioning device, electronic equipment and storage medium
Technical Field
The present application relates to the field of automatic driving technologies, and in particular, to a positioning method, an apparatus, an electronic device, and a storage medium.
Background
With the rapid rise of artificial intelligence, autopilot is more and more concerned by users. In an automatic driving system, a vehicle needs to sense the surrounding environment to determine a driving action or a driving route. Where accurate positioning of the vehicle is crucial for the perception of the environment.
In the existing positioning method, a laser radar is arranged on a vehicle, the laser radar is controlled to emit laser to the surrounding environment in the running process of the vehicle, and laser point cloud data around the vehicle is obtained according to the laser reflected by a reflection point. And matching the laser point cloud data around the vehicle with the original laser point cloud data in the high-precision map to obtain the positioning position of the vehicle.
Because the high-precision map contains a large amount of original laser point cloud data, the storage space occupied by the high-precision map is large, and the speed of positioning by using the high-precision map is low.
Disclosure of Invention
The application provides a positioning method, a positioning device, an electronic device and a storage medium, which can reduce the storage space occupied by a high-precision map and improve the positioning speed when the high-precision map is used for positioning.
A first aspect of the present application provides a positioning method, including:
acquiring the characteristics of each original point cloud in the original point cloud data; determining the weight of each original point cloud according to the characteristics of each original point cloud; determining a preset number of original point clouds as target point clouds according to the sequence of the weights from high to low, and determining target point cloud data according to the target point clouds; generating a high-precision map according to the target point cloud data; receiving request information from equipment to be positioned, wherein the request information is used for requesting a high-precision map; and sending the high-precision map to the equipment to be positioned.
In the positioning method in this embodiment, according to the point cloud characteristics of the original point cloud data, point cloud data corresponding to characteristics with a high contribution degree to positioning, that is, characteristics with a high weight, is screened out to generate a high-precision map. On one hand, the use of the high-precision map is not influenced, the positioning accuracy is not influenced, and the data volume of point cloud data in the high-precision map can be reduced, so that the storage space occupied by the high-precision map is reduced, and the positioning speed when the high-precision map is used for positioning is further improved.
A second aspect of the present application provides a positioning method, including: acquiring point cloud data around equipment to be positioned; determining a predicted location of the device to be located; extracting point cloud data at the predicted position from a high-precision map as candidate point cloud data according to the predicted position, wherein the point cloud data in the high-precision map is target point cloud data determined according to the weight of the characteristics of the point cloud in pre-obtained original point cloud data, and acquiring the error distance of the predicted position according to the candidate point cloud data, the point cloud data around the equipment to be positioned and a characteristic matching model, wherein the characteristic matching model is used for representing the corresponding relation between the characteristic matching degree and the error distance of the point cloud data; and determining the position of the equipment to be positioned according to the predicted position and the error distance.
A third aspect of the present application provides a positioning apparatus comprising:
the processing module is used for acquiring the characteristics of each original point cloud in the original point cloud data, determining the weight of each original point cloud according to the characteristics of each original point cloud, determining a preset number of original point clouds as target point clouds according to the sequence from high to low of the weights, determining the target point cloud data according to the target point clouds, and generating a high-precision map according to the target point cloud data; the receiving and sending module is used for receiving request information from equipment to be positioned, and the request information is used for requesting a high-precision map; the transceiver module is further configured to send the high-precision map to the device to be positioned.
The beneficial effects of the positioning device provided by the third aspect and various possible designs can be referred to the beneficial effects brought by the first aspect, which are not described herein again.
A fourth aspect of the present application provides a positioning device, comprising:
the radar module is used for acquiring point cloud data around the equipment to be positioned; the processing module is used for determining a predicted position of the equipment to be positioned, extracting point cloud data at the predicted position from a high-precision map as candidate point cloud data according to the predicted position, acquiring an error distance of the predicted position according to the candidate point cloud data, the point cloud data around the equipment to be positioned and a feature matching model, and determining the position of the equipment to be positioned according to the predicted position and the error distance; the point cloud data in the high-precision map is target point cloud data determined according to the weight of the characteristics of the point cloud in pre-acquired original point cloud data, and the characteristic matching model is used for representing the corresponding relation between the characteristic matching degree and the error distance of the point cloud data.
The above fourth aspect and various possible designs provide a positioning device, and the beneficial effects thereof can be seen from the related description in the embodiments.
A fifth aspect of the present application provides an electronic device, comprising: at least one processor and memory; the memory stores computer-executable instructions; the at least one processor executes computer-executable instructions stored by the memory to cause the electronic device to perform the positioning method of the first and second aspects.
A sixth aspect of the present application provides a computer-readable storage medium having stored thereon computer-executable instructions, which, when executed by a processor, implement the positioning method of the first and second aspects.
Other effects of the above-described alternative will be described below with reference to specific embodiments.
The application discloses a positioning method, a positioning device, electronic equipment and a storage medium, wherein the method comprises the following steps: acquiring point cloud data around equipment to be positioned; in the original point cloud data, determining a preset number of original point clouds as target point clouds according to the sequence of the weights of the features of the original point clouds from high to low, and determining the target point cloud data according to the target point clouds; generating a high-precision map according to the target point cloud data; and when receiving the request information from the equipment to be positioned, sending a high-precision map to the equipment to be positioned. Because the point cloud data in the high-precision map is not all original point cloud data but the point cloud data with higher weight determined according to the characteristics of the point cloud, and when the positioning is carried out according to the high-precision map, the point cloud characteristics in the point cloud data around the equipment to be positioned are matched with the point cloud characteristics of the point cloud data in the high-precision map, the positioning method in the application can reduce the data volume of the point cloud data in the high-precision map on the premise that the high-precision map is used and the positioning accuracy is not influenced, so that the storage space occupied by the high-precision map is reduced, and the positioning speed when the high-precision map is used for positioning is improved.
Drawings
The drawings are included to provide a better understanding of the present solution and are not intended to limit the present application. Wherein:
fig. 1 is a schematic view of a scene to which the positioning method provided in the present application is applied;
fig. 2 is a schematic flowchart of an embodiment of a positioning method provided in the present application;
FIG. 3 is a block diagram of a deep learning model provided herein;
FIG. 4 is a schematic view of a sub-region of a high-precision map provided herein;
FIG. 5 is a schematic diagram of obtaining a feature matching model provided herein;
fig. 6 is a schematic flowchart of another embodiment of a positioning method provided in the present application;
FIG. 7 is a first schematic structural diagram of a positioning device provided in the present application;
fig. 8 is a second schematic structural diagram of the positioning device provided in the present application;
fig. 9 is a schematic structural diagram of an electronic device provided in the present application.
Detailed Description
The following description of the exemplary embodiments of the present application, taken in conjunction with the accompanying drawings, includes various details of the embodiments of the application for the understanding of the same, which are to be considered exemplary only. Accordingly, those of ordinary skill in the art will recognize that various changes and modifications of the embodiments described herein can be made without departing from the scope and spirit of the present application. Also, descriptions of well-known functions and constructions are omitted in the following description for clarity and conciseness.
For convenience of describing the positioning method provided in the present application, a positioning method in the prior art is first described. In robotic systems or autonomous driving systems, the positioning of the robot or vehicle is of critical importance. In the prior art, Positioning systems such as a beidou System and a Global Positioning System (GPS) can provide Positioning services. The GPS is affected by factors such as multipath reflection and signal blocking, and therefore cannot provide stable and reliable positioning service in many scenarios, such as building blocking or indoor. With the continuous development of the positioning technology, the robot or the vehicle can be positioned according to the sensor, and the positioning stability and accuracy can be improved.
Taking a vehicle as an example, an automatic driving system generally adopts a mode of fusing multiple sensors such as a GPS, a laser radar, an inertial navigation and the like to provide stable and reliable vehicle positioning. When the vehicle is positioned, the laser radar can be controlled to emit laser to the periphery of the vehicle, the laser reflected by the reflection point is received, laser point cloud data of the periphery of the vehicle is obtained according to the reflected laser, and the position of the vehicle is determined by combining a high-precision map. Currently, this method can provide stable centimeter-level positioning results.
The method comprises the steps that laser point cloud data and collection positions of the laser point cloud data are stored in a high-precision map, and when the position of a vehicle is determined according to the laser point cloud data around the vehicle by combining the high-precision map, point cloud features in the laser point cloud data around the vehicle are matched with point cloud features of the laser point cloud data in the high-precision map to determine the position of the vehicle. It should be understood that, when the matching degree of the point cloud feature is greater than the threshold matching degree, the collection position of the laser point cloud data in the high-precision map may be taken as the position of the vehicle.
It should be understood that the laser point cloud data stored in the high-precision map is acquired point cloud data of the vehicle in the automatic driving area, and the manner of acquiring the laser point cloud data by the vehicle is the same as the manner of acquiring the laser point cloud data by the vehicle. The difference is that the collection vehicle can collect laser point cloud data at a plurality of position points on the road of the automatic driving area, and then the laser point cloud data collected by the collection vehicle are spliced to obtain a high-precision map. When the laser point cloud data of the vehicle are collected at two adjacent position points, the collected laser point clouds need to be overlapped, and then the point cloud data can be spliced according to the same laser point cloud so as to obtain the laser point cloud data of all the positions in the automatic driving area.
In view of the fact that the high-precision map comprises all the original point cloud data collected by the collection vehicle, the original point cloud data is all the laser point cloud data collected by the collection vehicle. Therefore, the data volume of the original point cloud data in the high-precision map is large, the occupied storage space is also large, and when the vehicle is positioned by using the high-precision map, the number of point clouds needing feature matching is large, so that the positioning speed of the vehicle is influenced.
In order to solve the above problems, the present application provides a positioning method, in which partial point cloud data is screened from original point cloud data included in a high-precision map, so that the data amount of the point cloud data included in the high-precision map can be reduced without affecting the positioning precision when the high-precision map is used for positioning, the storage space occupied by the high-precision map is further reduced, and the positioning speed when the high-precision map is used for positioning is increased.
Fig. 1 is a schematic view of a scene to which the positioning method provided in the present application is applied. As shown in fig. 1, the scenario includes: a device to be positioned and a server. The server is an electronic device with a high-precision map stored therein, and can acquire original point cloud data acquired by a vehicle and generate the high-precision map according to part of point cloud data screened from the original point cloud data. The device to be positioned can be a vehicle in an automatic driving scene, a robot in a robot system and other devices needing positioning. It should be understood that the device to be positioned in the present application may obtain a high-precision map from the server, and then determine the position of the device to be positioned by combining the high-precision map during the driving process. Fig. 1 illustrates an example of a vehicle.
It should be understood that the main execution body for executing the positioning method in the present application is a positioning device, and the positioning device may be a device to be positioned, or a processor, a chip, and the like in the device to be positioned. The positioning means may be implemented by any software and/or hardware.
The following describes a positioning method provided by the present application with reference to specific embodiments from the perspective of interaction between a server and a device to be positioned. Fig. 2 is a schematic flowchart of an embodiment of a positioning method provided in the present application. As shown in fig. 2, the positioning method provided in this embodiment may include:
s201, a server acquires the characteristics of each original point cloud in original point cloud data;
s202, the server determines the weight of each original point cloud according to the characteristics of each original point cloud;
s203, the server determines a preset number of original point clouds as target point clouds according to the sequence of the weights from high to low, and determines target point cloud data according to the target point clouds;
s204, the server generates a high-precision map according to the target point cloud data;
s205, the equipment to be positioned sends request information to a server, and the request information is used for requesting a high-precision map.
S206, the server sends the high-precision map to the equipment to be positioned.
In the above S201, in this embodiment, the original point cloud data includes data of the original point cloud, and the data of the original point cloud may include three-dimensional coordinates and reflection values of the original point cloud in space. It should be understood that the raw point cloud data may be acquired by the test vehicle during travel. The test vehicle can transmit radar signals to the surroundings, and the surrounding point cloud data can be obtained according to the reflected radar signals. The radar signal may be a laser radar signal or a microwave radar signal, and the like, which is not limited in this embodiment. Correspondingly, the original point cloud data in this embodiment may be laser point cloud data or microwave point cloud data.
The method comprises the steps of obtaining the characteristics of each original point cloud, determining the weight of each original point cloud according to the characteristics of each original point cloud, and determining a preset number of original point clouds as target point clouds according to the sequence of the weights from high to low. It should be understood that, in the present embodiment, a preset number of original point clouds before the weight ranking are used as the target point clouds.
In this embodiment, the feature of each original point cloud may be obtained from the original point cloud data, and specifically, the attribute feature of each original point cloud may be extracted according to the data of each original point cloud. The attribute features are determined by point cloud data, such as three-dimensional coordinates and reflection values. Optionally, the obtaining of the attribute feature of each original point cloud may be extracting a 32-dimensional feature vector of each original point cloud. The feature vector extraction method may adopt a feature vector extraction method in the prior art, or a depth feature extraction module (Deep feature extractor) may be disposed in the server in this embodiment to extract a 32-dimensional feature vector of each original point cloud.
In the above S202, in this embodiment, a weight determination model may be stored in advance in the server, and the weight determination model is used to represent the correspondence between the features of the point clouds and the weights, so after the features of each original point cloud are extracted, the features of each point cloud may be input into the weight determination model to obtain the weights of the features of each original point cloud.
It should be appreciated that the weight of the features of each original point cloud may be used to characterize the degree of contribution of the features of each original point cloud to the localization. The larger the weight of the features of the original point cloud is, the larger the contribution degree of the features of the original point cloud to positioning is. Optionally, the weight determination model in this embodiment may be stored in a weight module (weighting _ layer), that is, the weight module is configured to obtain a weight of the feature of each original point cloud according to the feature of each original point cloud.
In S203, in this embodiment, the weights of the features of the point clouds may be ranked in order from high to low, so as to determine the point cloud ranked first. In this embodiment, a preset number of original point clouds before the weight ranking may be used as the target point clouds, and target point cloud data is determined according to the target point clouds, where data corresponding to the target point clouds is the target point cloud data. In this embodiment, because the selected target point cloud data is the point cloud data with a high contribution to positioning in the original point cloud data, the accuracy of positioning of the device to be positioned is not affected by the obtained high-precision map, i.e., the positioning accuracy can be ensured by using the high-precision map for positioning.
In this embodiment, in order to improve the positioning accuracy of the precision map, since some point cloud features may be missing after the target point cloud data is screened from the original point cloud data, in this embodiment, in order to further improve the accuracy of the target point cloud features, that is, in order to improve the positioning accuracy of the precision map, the features of the target point cloud may be supplemented with other original point clouds around the target point cloud features.
Optionally, in this embodiment, the spatial feature of the target point cloud may be obtained according to the attribute feature of the target point cloud, so as to supplement a part of features that may be missing from the target point cloud. The spatial features are determined by the attribute features of the target point cloud and the attribute features of the point cloud around the target point cloud, and further the spatial features, the three-dimensional coordinates and the reflection values of the target point cloud are used as target point cloud data in this embodiment.
In this embodiment, the spatial feature of the target point cloud is obtained according to the attribute feature of the target point cloud and the feature determined by the attribute feature of the point cloud around the target point cloud. Wherein the data of the target point cloud comprises: in this embodiment, the original point cloud within the preset area range of the target point cloud is determined according to the three-dimensional coordinates of the target point cloud, and then the original point cloud within the preset area range of the target point cloud is used as the point cloud around the target point cloud.
The spatial features of the target point cloud can be obtained according to the attribute features of the target point cloud and the attribute features of the original point cloud within the preset area range of the target point cloud, for example, a 32-dimensional feature vector of the target point cloud and a 32-dimensional feature vector of the original point cloud within the preset area range of the target point cloud. Optionally, the spatial feature of the target point cloud may be a 32-dimensional feature vector, and compared with the attribute feature of the above-mentioned target point cloud, the spatial feature in this embodiment combines the attribute features of the point clouds around the target point cloud.
The server of this embodiment stores a spatial feature model, and the spatial feature model is used to represent the attribute features of the point cloud, the attribute features of the point cloud within the preset area range of the point cloud, and the corresponding relationship between the spatial features of the point cloud, that is, after the attribute features of the target point cloud and the attribute features of the original point cloud within the preset area range of the target point cloud are input to the spatial feature model, the spatial features of the target point cloud can be obtained. Alternatively, the spatial Feature model in this embodiment may be stored in a depth Feature fusing module (Deep Feature embedding).
After the spatial features of the target point cloud are obtained, the spatial features, the three-dimensional coordinates and the reflection values of the target point cloud can be used as target point cloud data. Optionally, in this implementation, the three-dimensional coordinates and the reflection value of the target point cloud may be used as 4-dimensional features of the target point cloud, and then the 32-dimensional spatial features of the target point cloud and the 4-dimensional features of the target point cloud are spliced to obtain 36-dimensional features of the target point cloud, and then the 36-dimensional features are used as the features of the target point cloud. Correspondingly, when the point cloud features of the point cloud data in the high-precision map are used for matching, the 36-dimensional features can be adopted for matching with the features of the point cloud in the point cloud data around the equipment to be positioned.
Correspondingly, the target point cloud data in this embodiment includes a spatial feature of the target point cloud and a three-dimensional coordinate and a reflection value of the target point cloud.
Optionally, in this embodiment, the server may integrate the functions of the depth feature extraction module, the weighting module, and the depth feature fusion module into a deep learning model, so as to implement the functions of the depth feature extraction module, the weighting module, and the depth feature fusion module. That is, after the original point cloud data is input to the deep learning model, a high-precision map with target point cloud data can be obtained.
Fig. 3 is a schematic diagram of a deep learning model framework provided in the present application. As shown in fig. 3, a depth feature extraction module, a weighting module and a depth feature fusion module are integrated in the deep learning model, after the original point clouds are input, 32-dimensional attribute features of each original point cloud can be obtained through the depth feature extraction module, then the weighting of the 32-dimensional attribute features of each original point cloud can be obtained through the weighting module, and after the target point cloud data are obtained, the 32-dimensional spatial attributes of the target point cloud data can be obtained through the depth feature fusion module and the attribute features of the original point clouds around the target point cloud data and the attribute features of the target point cloud data. And then combining the 4-dimensional features (three-dimensional coordinates and reflection values) of the target point cloud to obtain the 36-dimensional features of the target point cloud. It should be understood that the point cloud data included in the high-precision map in the present embodiment is target point cloud data including spatial features of the target point cloud and three-dimensional coordinates and reflection values of the target point cloud.
In step S204, in this embodiment, after the target point cloud data is determined in the original point cloud data, a high-precision map may be generated according to the target point cloud data. Optionally, in this embodiment, a high-precision map is generated by using a voxelization (voxel) method.
Wherein, the voxelization refers to storing the target point cloud in the target point cloud data in a voxel form. That is, the high-precision map may be pre-divided into a plurality of sub-regions, each of which has the same volume, and the target point cloud may be mapped into each of the sub-regions according to the target point cloud data.
Fig. 4 is a schematic diagram of a sub-area of a high-precision map provided in the present application. As shown in fig. 4, each subregion includes a fixed number of rows, columns and heights in the X, Y and Z-axis directions, one subregion being referred to as a voxel, each voxel having a fixed volume. In this embodiment, the area covered by the target point cloud may be mapped into multiple sub-areas in the high-precision map, that is, which sub-area each target point cloud is in is determined according to the three-dimensional coordinates of the target point cloud.
In the embodiment, the target point cloud in each sub-region is determined according to the three-dimensional coordinates of each target point cloud, and the relative coordinates of the target point cloud in the sub-region in each sub-region can be determined according to the position of the target point cloud in each sub-region in view of the fact that each sub-region has its own coordinate system. In this embodiment, the target point cloud may be loaded into the corresponding sub-region according to the relative coordinates of the target point cloud in the sub-region, so as to generate the high-precision map. Wherein, the target point cloud data included in the high-precision map is also: features of the target point cloud, and three-dimensional coordinates and reflectance values of the target point cloud.
In S205 and S206, before the vehicle or the robot travels, the device to be positioned may send request information to the server to request a high-precision map. After receiving the request information, the server can send a high-precision map to the equipment to be positioned, so that the equipment to be positioned can determine the position of the equipment to be positioned according to the high-precision map.
Different from the high-precision map in the prior art, the high-precision map in the prior art includes original point cloud data, and the point cloud data included in the high-precision map in the embodiment is target point cloud data determined from the original point cloud data. The data volume of the target point cloud data is smaller than that of the original point cloud data, so that the storage space occupied by the high-precision map can be reduced. In this embodiment, when the device to be positioned is positioned according to the high-precision map, matching is performed according to the characteristics of the point cloud in the point cloud data around the device to be positioned and the characteristics of the point cloud data in the high-precision map, and a position corresponding to the point cloud data in the high-precision map where the matching similarity of the characteristics is greater than the similarity threshold is used as the position of the device to be positioned. It should be noted that, the manner in which the device to be positioned in this embodiment determines the position of the device to be positioned according to the high-precision map may specifically refer to the related description of the following embodiments.
In the embodiment, in the original point cloud data, a preset number of original point clouds are determined as target point clouds according to the sequence from high to low of the weights of the features of the original point clouds, and the target point cloud data are determined according to the target point clouds; generating a high-precision map according to the target point cloud data; and when receiving the request information from the equipment to be positioned, sending a high-precision map to the equipment to be positioned. Because the point cloud data in the high-precision map is not all original point cloud data but the point cloud data with higher weight determined according to the characteristics of the point cloud, the data volume of the point cloud data in the high-precision map can be reduced on the premise of not influencing the positioning accuracy, so that the storage space occupied by the high-precision map is reduced, and the positioning speed when the high-precision map is used for positioning is further improved.
In order to facilitate description of a manner in which the following device to be positioned determines a position of the device to be positioned according to the high-precision map, a manner in which a server acquires a feature matching model is described with reference to fig. 5, where fig. 5 is a schematic diagram of the feature matching model acquisition provided in the present application:
in this embodiment, the training data for training the feature matching model is original point cloud data, an acquisition position of the original point cloud data, and historical point cloud data around the device to be positioned. Optionally, in this embodiment, the basic network architecture of the training feature matching model is an L3Net structure.
In this embodiment, before training the feature matching model with the training data, the training data may be preprocessed. Optionally, in this embodiment, the point cloud data of the ground point cloud in the historical point cloud data may be removed, so as to train the feature matching model according to the point cloud data in the surrounding environment of the device to be positioned, avoid a large amount of point cloud data from occupying the features of the point cloud in the high-precision map, and improve the accuracy of the feature matching model. The point cloud data of the ground point cloud in the historical point cloud data can be identified by adopting a random forest, a point cloud identification model and the like so as to remove the point cloud data of the ground point cloud. It should be understood that the identification model is used to characterize the correspondence of the point cloud to the point cloud type. In this embodiment, the process of point cloud identification using a random forest or the like is not described in detail.
The original point cloud data in this embodiment includes multiple frames of original point cloud data, that is, original point cloud data acquired by acquiring a vehicle at multiple position points, and correspondingly, the acquisition position of the original point cloud data also includes the acquisition position of each frame of original point cloud data. When the training data is preprocessed, a plurality of sample positions corresponding to the acquisition positions of each frame of original point cloud can be generated according to the acquisition positions of each frame of original point cloud data. For example, for acquisition position a, a plurality of sample positions corresponding to the acquisition position a may be generated. Optionally, in this embodiment, at least one error distance may be randomly generated, and a sum or a difference between the acquisition position of the frame original point cloud and each error distance is used as a sample position to generate a plurality of sample positions.
Optionally, the collection position of each frame of original point cloud in this embodiment may be a three-dimensional coordinate in the high-precision map, and the error distance may include an error distance of an abscissa, an error distance of a ordinate, and an error distance of a Z coordinate. Illustratively, if the at least one error distance is (+3, +2, -1) and (+2, +1, 0), and the acquisition position of the original point cloud is (+200, +300, +10), then the corresponding plurality of sample positions of the frame of original point cloud may be (+203, +302, +9) and (+202, +301, + 10).
In the embodiment, the original point cloud data in the preset area range of each sample position can be obtained from the multi-frame original point cloud data, and the historical point cloud data with the same coverage range as the original point cloud data in the preset area can be obtained by randomly sampling in the historical point cloud data.
In this embodiment, the original point cloud data in the preset area range of each sample position and the historical point cloud data with the same coverage as the preset area range may be used as training data, and each sample position and the corresponding collection position are identified in the original point cloud data in the preset area range of each sample position. In the training process, the error distance between each sample position and the acquisition position of each training can be obtained, and then parameters in the feature matching model are adjusted according to the error distance until the error distance between each sample position and the acquisition position is smaller than a distance threshold, so that the feature matching model in the embodiment is obtained, wherein the feature matching model is used for representing the corresponding relation between the feature matching degree and the error distance of the point cloud data, and the error distance is the error distance between the acquisition position of the point cloud data and the position of the equipment to be positioned.
The above is a process of acquiring a feature matching model by the server in this embodiment, and when the device to be positioned is positioned according to the embodiment, the device to be positioned needs to be positioned according to matching between point cloud data around the device to be positioned and point cloud data in a high-precision map. In the matching process, feature extraction and matching are required to be performed on the point cloud in the point cloud data, so that the matching speed is low. When positioning is carried out based on the matching similarity and the similarity threshold, the determination of the similarity threshold is crucial, if the similarity threshold is large, the position of the equipment to be positioned cannot be determined, and if the similarity threshold is small, the position of the equipment to be positioned cannot be determined accurately.
In this embodiment, in order to improve the positioning accuracy when using a high-precision map, a feature matching model may be stored in advance in the device to be positioned, or when requesting a high-precision map from a server, the server may request the feature matching model. The feature matching model is used for representing the corresponding relation between the feature matching degree and the error distance of the point cloud data, so that the error distance between the positions of the point cloud data around the equipment to be positioned and the point cloud data in the high-precision map can be determined according to the feature matching degree of the point cloud data around the equipment to be positioned and the feature matching degree of the point cloud data in the high-precision map, and the accuracy of positioning by using the high-precision map can be improved.
The following describes a manner of determining the position of the device to be positioned based on a high-precision map obtained by the device to be positioned from the perspective of the device to be positioned. Fig. 6 is a schematic flowchart of another embodiment of a positioning method provided in the present application. As shown in fig. 6, the positioning method provided in this embodiment may include:
s601, point cloud data around the equipment to be positioned is obtained.
S602, determining the predicted position of the equipment to be positioned.
And S603, extracting point cloud data at the predicted position from the high-precision map as candidate point cloud data according to the predicted position.
S604, obtaining the error distance of the predicted position according to the candidate point cloud data, the point cloud data around the equipment to be positioned and the feature matching model, wherein the feature matching model is used for representing the corresponding relation between the feature matching degree of the point cloud data and the error distance.
And S605, determining the position of the equipment to be positioned according to the predicted position and the error distance.
It should be understood that in the above S601, during the process of moving or driving the device to be positioned, the radar signal may be controlled to be emitted to the periphery of the device to be positioned, and the point cloud data around the device to be positioned may be obtained according to the reflected radar signal. The radar signal may be a laser radar signal or a microwave radar signal, and correspondingly, the point cloud data around the device to be positioned in this embodiment may be laser point cloud data or microwave point cloud data.
In the above S602, the predicted position of the device to be positioned may be a predicted position where the device to be positioned is located when acquiring the surrounding point cloud data. Alternatively, the predicted position may be determined by a GPS positioning system, and the position of the device to be positioned determined by the GPS positioning system may be used as the predicted position in view of the low accuracy of the position of the device to be positioned determined by the GPS positioning system.
In the above S603, since the high-precision map includes the collection position of the point cloud data, that is, the collection position of the target point cloud data, the point cloud data at the predicted position can be extracted from the high-precision map as candidate point cloud data. In this embodiment, the error distance corresponding to the feature matching degree between the point cloud data around the device to be positioned and the candidate point cloud data may be determined by combining the feature matching model, so as to determine the accurate position of the device to be positioned.
In the above step S604, in this embodiment, the candidate point cloud data and the point cloud data around the device to be positioned may be input to the feature matching model, so as to obtain an error distance corresponding to the feature matching degree of the point cloud data and the candidate point cloud data.
Optionally, in order to reduce the calculation amount of the feature matching model, candidate point cloud data may be screened. The point cloud data comprises three-dimensional coordinates of the point cloud, and the candidate point cloud data also comprises the three-dimensional coordinates of the point cloud. In the embodiment, the number of the point clouds around the device to be positioned in the preset area range of each point cloud in the candidate point cloud data can be determined according to the three-dimensional coordinates of each point cloud in the candidate point cloud data and the three-dimensional coordinates of each point cloud in the point cloud data around the device to be positioned. In this embodiment, the number of point clouds around the device to be positioned existing within 1.5 meters around each point cloud in the candidate point cloud data may be obtained, that is, the number of point clouds around the device to be positioned, which is less than or equal to 1.5 meters away from each point cloud in the candidate point cloud data, may be obtained. The point clouds in the candidate point cloud data with the number of the point clouds around the equipment to be positioned being less than the number threshold existing within 1.5 meters around each point cloud can be deleted, and the point clouds in the candidate point cloud data with the number being more than the number threshold can be obtained.
In this embodiment, the error distance of the predicted position may be obtained according to the point cloud data of the point clouds in the candidate point cloud data whose number is greater than the number threshold, the point cloud data around the device to be positioned, and the feature matching model. Namely, the point cloud data of the point clouds in the candidate point cloud data with the number larger than the number threshold value and the point cloud data around the equipment to be positioned are input into the feature matching model, and the error distance of the predicted position is obtained.
In the above step S605, after determining the error distance corresponding to the predicted position of the device to be positioned, the position of the device to be positioned may be determined according to the predicted position and the error distance. Optionally, the predicted position of the device to be positioned is a three-dimensional coordinate in the high-precision map, and the error distance is an error distance of the three-dimensional coordinate, such as an error distance of an abscissa, an error distance of a ordinate, and an error distance of a Z coordinate. In the embodiment, the position of the equipment to be positioned can be determined according to the three-dimensional coordinates and the error distance of the predicted position in the high-precision map. Specifically, the abscissa, the ordinate, and the Z coordinate of the three-dimensional coordinate of the predicted position in the high-precision map may be respectively added or subtracted with the corresponding error distance of the abscissa, the error distance of the ordinate, and the error distance of the Z coordinate, so as to obtain the accurate three-dimensional coordinate of the device to be positioned in the high-precision map, that is, the position of the device to be positioned.
In the positioning method provided by this embodiment, a feature matching model may be stored in advance in the device to be positioned, and the feature matching model is used to represent a correspondence between a feature matching degree of point cloud data and an error distance, and further, an error distance between positions of the point cloud data and the point cloud data in the high-precision map may be determined according to the feature matching degree of the point cloud data around the device to be positioned and the point cloud data in the high-precision map, and further, the position of the device to be positioned may be determined.
The characteristic matching model stored in the equipment to be positioned can be acquired by the server requested by the equipment to be positioned. Optionally, in this embodiment, before the device to be positioned travels, request information may be sent to the server to request the feature matching model. It should be understood that the request information may be one request information for requesting a high-precision map with the above-mentioned device to be positioned.
After receiving the request information, the server may send the feature matching model to the device to be located. The server may obtain the original point cloud data and historical point cloud data around the device to be positioned, which is obtained by the device to be positioned, through deep learning, which is specifically described in fig. 6.
In the embodiment, the feature matching model for representing the corresponding relation between the feature matching degree and the error distance of the point cloud data can be trained according to the original point cloud data, the acquisition position of the original point cloud data and the historical point cloud data around the equipment to be positioned, so that the problem that the positioning precision is influenced by the set similarity threshold value in the prior art can be solved, and the accuracy of positioning by using the high-precision map can be improved.
Fig. 7 is a first schematic structural diagram of the positioning device provided in the present application. It should be understood that the positioning device may be the server in the above embodiments, for performing the actions of the server. As shown in fig. 7, the positioning apparatus 700 includes: a processing module 701 and a transceiver module 702.
The processing module 701 is used for acquiring the characteristics of each original point cloud in the original point cloud data, determining the weight of each original point cloud according to the characteristics of each original point cloud, determining a preset number of original point clouds as target point clouds according to the sequence from high to low of the weights, determining the target point cloud data according to the target point clouds, and generating a high-precision map according to the target point cloud data;
the transceiver module 702 is configured to receive request information from a device to be positioned, where the request information is used to request a high-precision map.
The transceiver module 702 is further configured to send the high-precision map to the device to be positioned.
In one design, the processing module 702 is specifically configured to determine a weight of each original point cloud according to the features of each original point cloud and a weight determination model, where the weight determination model is used to represent the correspondence between the features and the weights of the point cloud.
In one design, the point cloud data includes three-dimensional coordinates and reflectance values of the point clouds, with each original point cloud having characteristics that are attribute characteristics of the original point cloud, with the attribute characteristics being characteristics determined from the point cloud's own data.
The processing module 702 is specifically configured to obtain a spatial feature of the target point cloud according to the attribute feature of the target point cloud, where the spatial feature is a feature determined by the attribute feature of the target point cloud and the attribute features of point clouds around the target point cloud; and taking the spatial characteristics, the three-dimensional coordinates and the reflection values of the target point cloud as target point cloud data.
In one design, the processing module 702 is specifically configured to determine an original point cloud within a preset area of a target point cloud according to three-dimensional coordinates of the target point cloud; and acquiring the spatial characteristics of the target point cloud according to the attribute characteristics of the target point cloud and the attribute characteristics of the original point cloud in the preset area range of the target point cloud.
In one design, the processing module 702 is specifically configured to obtain the spatial features of the target point cloud according to the attribute features of the target point cloud, the attribute features of the original point cloud within the preset area range of the target point cloud, and a spatial feature model, where the spatial feature model is used to represent the corresponding relationship between the attribute features of the point cloud, the attribute features of the point cloud within the preset area range of the point cloud, and the spatial features of the point cloud.
In one design, the target point cloud is multiple, the high-precision map is divided into multiple sub-areas, and the volume of each sub-area is the same.
A processing module 702, specifically configured to map an area covered by the target point cloud into multiple sub-areas in the high-precision map; determining target point clouds in each sub-area according to the three-dimensional coordinates of each target point cloud; acquiring relative coordinates of the target point cloud in each sub-region in the sub-region; and loading the target point cloud into the corresponding sub-area to generate a high-precision map.
In one design, the processing module 702 is further configured to train a feature matching model according to the original point cloud data, the collection position of the original point cloud data, and historical point cloud data around the device to be positioned as training data, where the feature matching model is used to represent a corresponding relationship between a feature matching degree of the point cloud data and an error distance, and the error distance is an error distance between the collection position of the point cloud data and the position of the device to be positioned.
In one design, the original point cloud data includes point cloud data of multiple frames of original point clouds, and the collection position of the original point cloud data is the collection position of each frame of original point cloud data.
A processing module 702, specifically configured to generate a plurality of sample positions corresponding to the acquisition positions of each frame of original point cloud data according to the acquisition positions of each frame of original point cloud data; acquiring original point cloud data in a preset area range of each sample position from multi-frame original point cloud data; acquiring historical point cloud data with the same coverage range as a preset area; and taking original point cloud data in a preset area range of each sample position and historical point cloud data with the same coverage range as the preset area as training data, training until the error distance between each sample position and the acquisition position is smaller than a distance threshold, and acquiring a feature matching model.
In one design, the processing module 702 is specifically configured to randomly generate at least one error distance; and taking the sum or the difference of the acquisition position of each frame of original point cloud data and each error distance as a sample position to generate a plurality of sample positions.
In one design, the processing module 702 is further configured to remove point cloud data of the ground point cloud from the historical point cloud data.
In one design, the processing module 702 is specifically configured to identify a ground point cloud in the historical point clouds according to a point cloud identification model to remove point cloud data of the ground point cloud, and the identification model is used to represent a corresponding relationship between the point cloud and a point cloud type.
In one design, transceiver module 701 is further configured to receive second request information from the device to be located, and send the feature matching model to the device to be located, where the second request information is used for the feature matching model.
Fig. 8 is a schematic structural diagram of a positioning device provided in the present application. The positioning device may be a device to be positioned in the above embodiment, or a processor, a chip, and the like in the device to be positioned. As shown in fig. 8, the positioning apparatus 800 includes: a radar module 801, a processing module 802 and a transceiver module 803.
And the radar module 801 is used for acquiring point cloud data around the equipment to be positioned.
A processing module 802, configured to determine a predicted position of the device to be positioned, extract point cloud data at the predicted position from a high-precision map as candidate point cloud data according to the predicted position, obtain an error distance of the predicted position according to the candidate point cloud data, the point cloud data around the device to be positioned, and a feature matching model, and determine a position of the device to be positioned according to the predicted position and the error distance; the point cloud data in the high-precision map is target point cloud data determined according to the weight of the characteristics of the point cloud in pre-acquired original point cloud data, and the characteristic matching model is used for representing the corresponding relation between the characteristic matching degree and the error distance of the point cloud data.
In one design, the processing module 802 is specifically configured to determine the number of point clouds around the device to be positioned existing within a preset area range of each point cloud in the candidate point cloud data according to the three-dimensional coordinates of each point cloud in the candidate point cloud data and the three-dimensional coordinates of each point cloud in the point cloud data around the device to be positioned; determining point clouds in the candidate point cloud data with the number larger than a number threshold; and acquiring the error distance of the predicted position according to the point cloud data of the point clouds in the candidate point cloud data with the number larger than the number threshold, the point cloud data around the equipment to be positioned and the characteristic matching model.
In one design, the predicted position is a three-dimensional coordinate of the predicted position in a high-precision map, and the error distance is an error distance of the three-dimensional coordinate.
The processing module 802 is specifically configured to determine the position of the device to be positioned according to the three-dimensional coordinates and the error distance of the predicted position in the high-precision map.
In one design, the transceiver module 803 is used to send request information to the server and receive high-precision maps and feature matching models from the server, the request information being used to request the high-precision maps and feature matching models.
The positioning device provided in this embodiment has similar principles and technical effects to those achieved by the positioning method, and will not be described herein again.
According to an embodiment of the present application, an electronic device and a readable storage medium are also provided. Fig. 9 is a schematic structural diagram of an electronic device provided in the present application. Fig. 9 is a block diagram of an electronic device according to the positioning method of the embodiment of the present application. Electronic device is intended to represent various forms of digital computers, processors, or chips. Such as an on-board computer, an on-board terminal device, a vehicle central control computer, a chip of a processor in a vehicle, a robot, or a processor or chip in a robot, etc. The components shown herein, their connections and relationships, and their functions, are meant to be examples only, and are not meant to limit implementations of the present application that are described and/or claimed herein.
As shown in fig. 9, the electronic apparatus includes: one or more processors 901, memory 902, and interfaces for connecting the various components, including a high-speed interface and a low-speed interface. The various components are interconnected using different buses and may be mounted on a common motherboard or in other manners as desired. The processor may process instructions for execution within the electronic device, including instructions stored in or on the memory to display graphical information of a GUI on an external input/output apparatus (such as a display device coupled to the interface). In other embodiments, multiple processors and/or multiple buses may be used, along with multiple memories and multiple memories, as desired. Also, multiple electronic devices may be connected, with each device providing portions of the necessary operations (e.g., as a server array, a group of blade servers, or a multi-processor system). Fig. 9 illustrates an example of a processor 901.
Memory 902 is a non-transitory computer readable storage medium as provided herein. Wherein the memory stores instructions executable by at least one processor to cause the at least one processor to perform the positioning method provided herein. The non-transitory computer-readable storage medium of the present application stores computer instructions for causing a computer to perform the positioning method provided herein.
The memory 902, which is a non-transitory computer readable storage medium, may be used to store non-transitory software programs, non-transitory computer executable programs, and modules, such as program instructions/modules corresponding to the positioning method in the embodiments of the present application. The processor 901 executes various functional applications of the server and sample processing by executing non-transitory software programs, instructions, and modules stored in the memory 902, so as to implement the positioning method in the above-described method embodiment.
The memory 902 may include a storage program area and a storage sample area, wherein the storage program area may store an operating system, an application program required for at least one function; the stored sample area may store a sample created according to use of an electronic device for performing the positioning method, or the like. Further, the memory 902 may include high speed random access memory, and may also include non-transitory memory, such as at least one magnetic disk storage device, flash memory device, or other non-transitory solid state storage device. In some embodiments, the memory 902 may optionally include memory located remotely from the processor 901, which may be connected via a network to an electronic device for performing the positioning method. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.
Optionally, when the electronic device executing the positioning method is a device to be positioned, the electronic device may further include: an input device 903, an output device 904, and a radar 905. The processor 901, the memory 902, the input device 903 and the output device 904 may be connected by a bus or other means, and fig. 9 illustrates the connection by a bus as an example. The radar 905 may be a laser radar, a microwave radar, or the like, and the radar 905 may be disposed at a head, a tail, two sides, or other positions of the vehicle, for executing the actions of the radar module. The processor 901, the memory 902, the input device 903, the output device 904, and the radar 905 may be connected by a bus in other ways, and fig. 9 illustrates an example of a connection by a bus.
The input device 903 may receive input numeric or character information and generate key signal inputs related to user settings and function control of an electronic apparatus for performing the pointing method, such as an input device like a touch screen, a keypad, a mouse, a track pad, a touch pad, a pointer stick, one or more mouse buttons, a track ball, a joystick, etc. The output devices 904 may include a display device, auxiliary lighting devices (e.g., LEDs), tactile feedback devices (e.g., vibrating motors), and the like. The display device may include, but is not limited to, a Liquid Crystal Display (LCD), a Light Emitting Diode (LED) display, and a plasma display. In some implementations, the display device can be a touch screen.
Various implementations of the systems and techniques described here can be realized in digital electronic circuitry, integrated circuitry, application specific ASICs (application specific integrated circuits), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include: implemented in one or more computer programs that are executable and/or interpretable on a programmable system including at least one programmable processor, which may be special or general purpose, receiving samples and instructions from, and transmitting samples and instructions to, a storage system, at least one input device, and at least one output device.
These computer programs (also known as programs, software applications, or code) include machine instructions for a programmable processor, and may be implemented using high-level procedural and/or object-oriented programming languages, and/or assembly/machine languages. As used herein, the terms "machine-readable medium" and "computer-readable medium" refer to any computer program product, apparatus, and/or device (e.g., magnetic discs, optical disks, memory, Programmable Logic Devices (PLDs)) used to provide machine instructions and/or samples to a programmable processor, including a machine-readable medium that receives machine instructions as a machine-readable signal. The term "machine-readable signal" refers to any signal used to provide machine instructions and/or samples to a programmable processor.
To provide for interaction with a user, the systems and techniques described here can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to a user; and a keyboard and a pointing device (e.g., a mouse or a trackball) by which a user can provide input to the computer. Other kinds of devices may also be used to provide for interaction with a 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 may be received in any form, including acoustic, speech, or tactile input.
The systems and techniques described here can be implemented in a computing system that includes a back-end component (e.g., as a sample server), or that includes a middleware component (e.g., an application server), or that includes a front-end component (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such back-end, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital sample communication (e.g., a communication network). Examples of communication networks include: local Area Networks (LANs), Wide Area Networks (WANs), and the Internet.
The computer system may include clients and servers. A client and server are generally remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other.
It should be understood that various forms of the flows shown above may be used, with steps reordered, added, or deleted. For example, the steps described in the present application may be executed in parallel, sequentially, or in different orders, as long as the desired results of the technical solutions disclosed in the present application can be achieved, and the present invention is not limited herein.
The above-described embodiments should not be construed as limiting the scope of the present application. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions may be made in accordance with design requirements and other factors. Any modification, equivalent replacement, and improvement made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims (20)

1. A method of positioning, comprising:
acquiring the characteristics of each original point cloud in the original point cloud data;
determining the weight of each original point cloud according to the characteristics of each original point cloud;
determining a preset number of original point clouds as target point clouds according to the sequence of the weights from high to low, and determining target point cloud data according to the target point clouds;
generating a high-precision map according to the target point cloud data;
receiving request information from equipment to be positioned, wherein the request information is used for requesting a high-precision map;
and sending the high-precision map to the equipment to be positioned.
2. The method of claim 1, wherein determining the weight for each raw point cloud based on the characteristics of each raw point cloud comprises:
and determining the weight of each original point cloud according to the characteristics of each original point cloud and a weight determination model, wherein the weight determination model is used for representing the corresponding relation between the characteristics and the weight of the point cloud.
3. The method of claim 1 or 2, wherein the feature of each original point cloud is an attribute feature of the original point cloud, the attribute feature being a feature determined from the point cloud's own data, the determining target point cloud data from the target point cloud comprising:
acquiring a spatial feature of the target point cloud according to the attribute feature of the target point cloud, wherein the spatial feature is determined by the attribute feature of the target point cloud and the attribute features of the point clouds around the target point cloud;
and taking the space characteristic, the three-dimensional coordinate and the reflection value of the target point cloud as the target point cloud data.
4. The method of claim 3, wherein the obtaining the spatial feature of the target point cloud according to the attribute feature of the target point cloud comprises:
determining an original point cloud in a preset area range of the target point cloud according to the three-dimensional coordinates of the target point cloud;
and acquiring the spatial characteristics of the target point cloud according to the attribute characteristics of the target point cloud and the attribute characteristics of the original point cloud in the preset area range of the target point cloud.
5. The method of claim 4, wherein the generating the spatial features of the target point cloud according to the attribute features of the target point cloud and the attribute features of the original point cloud within the preset area range of the target point cloud comprises:
and acquiring the spatial characteristics of the target point cloud according to the attribute characteristics of the target point cloud, the attribute characteristics of the original point cloud in the preset area range of the target point cloud and a spatial characteristic model, wherein the spatial characteristic model is used for representing the corresponding relation among the attribute characteristics of the point cloud, the attribute characteristics of the point cloud in the preset area range of the point cloud and the spatial characteristics of the point cloud.
6. The method according to any one of claims 1 to 5, wherein the target point cloud is plural, the high-precision map is divided into a plurality of sub-regions, each sub-region has the same volume, and the generating the high-precision map according to the target point cloud data comprises:
mapping an area covered by the target point cloud into a plurality of sub-areas in the high-precision map;
determining target point clouds in each sub-area according to the three-dimensional coordinates of each target point cloud;
acquiring relative coordinates of the target point cloud in each sub-region in the sub-region;
and loading the target point cloud into a corresponding sub-region to generate the high-precision map.
7. The method of claim 1, further comprising:
and training a feature matching model according to the original point cloud data, the acquisition position of the original point cloud data and the historical point cloud data around the equipment to be positioned as training data, wherein the feature matching model is used for representing the corresponding relation between the feature matching degree of the point cloud data and the error distance, and the error distance is the error distance between the acquisition position of the point cloud data and the position of the equipment to be positioned.
8. The method of claim 7, wherein the original point cloud data comprises point cloud data of a plurality of frames of original point clouds, wherein the acquisition location of the original point cloud data comprises the acquisition location of each frame of original point cloud data, and wherein training the feature matching model according to the original point cloud data, the acquisition location of the original point cloud data, and historical point cloud data around a device to be positioned as training data comprises:
generating a plurality of sample positions corresponding to the acquisition positions of each frame of original point cloud data according to the acquisition positions of each frame of original point cloud data;
acquiring original point cloud data in a preset area range of each sample position in the multi-frame original point cloud data;
acquiring historical point cloud data with the same coverage range as the preset area;
and taking original point cloud data in a preset area range of each sample position and historical point cloud data with the same coverage range as the preset area as training data, training until the error distance between each sample position and the acquisition position is smaller than a distance threshold, and acquiring the feature matching model.
9. The method of claim 8, wherein generating a plurality of sample positions corresponding to the collection position of each frame of the original point cloud data according to the collection position of each frame of the original point cloud data comprises:
randomly generating at least one error distance;
and taking the sum or the difference of the acquisition position of each frame of original point cloud data and each error distance as a sample position to generate a plurality of sample positions.
10. The method according to claim 8 or 9, wherein before the obtaining of the historical point cloud data with the same coverage as the preset area, the method further comprises:
and removing the point cloud data of the ground point cloud in the historical point cloud data.
11. The method of claim 10, wherein removing point cloud data of ground point clouds in the historical point cloud data comprises:
and according to a point cloud identification model, identifying the ground point cloud in the historical point cloud so as to remove the point cloud data of the ground point cloud, wherein the identification model is used for representing the corresponding relation between the point cloud and the point cloud type.
12. The method according to any one of claims 7-11, further comprising:
receiving second request information from the equipment to be positioned, wherein the second request information is used for the feature matching model;
and sending the feature matching model to the equipment to be positioned.
13. A method of positioning, comprising:
acquiring point cloud data around equipment to be positioned;
determining a predicted location of the device to be located;
extracting point cloud data at the predicted position from a high-precision map as candidate point cloud data according to the predicted position, wherein the point cloud data in the high-precision map is target point cloud data determined according to the weight of the characteristics of the point cloud in pre-obtained original point cloud data;
acquiring an error distance of the predicted position according to the candidate point cloud data, the point cloud data around the equipment to be positioned and a feature matching model, wherein the feature matching model is used for representing the corresponding relation between the feature matching degree and the error distance of the point cloud data;
and determining the position of the equipment to be positioned according to the predicted position and the error distance.
14. The method of claim 13, wherein obtaining the error distance of the predicted location from the candidate point cloud data, the point cloud data surrounding the device to be located, and a feature matching model comprises:
determining the number of the point clouds around the equipment to be positioned existing in the preset area range of each point cloud in the candidate point cloud data according to the three-dimensional coordinates of each point cloud in the candidate point cloud data and the three-dimensional coordinates of each point cloud in the point cloud data around the equipment to be positioned;
determining the number of point clouds in the candidate point cloud data which is greater than a number threshold;
and acquiring the error distance of the predicted position according to the point cloud data of the point clouds in the candidate point cloud data with the number larger than the number threshold, the point cloud data around the equipment to be positioned and the feature matching model.
15. The method of claim 13, wherein the predicted location is a three-dimensional coordinate of the predicted location in the high-accuracy map, the error distance is an error distance of the three-dimensional coordinate, and determining the location of the device to be located based on the predicted location and the error distance comprises:
and determining the position of the equipment to be positioned according to the three-dimensional coordinates of the predicted position in the high-precision map and the error distance.
16. The method according to any one of claims 13-15, further comprising:
sending request information to a server, wherein the request information is used for requesting the high-precision map and the feature matching model;
receiving the high-precision map and the feature matching model from the server.
17. A positioning device, comprising:
the processing module is used for acquiring the characteristics of each original point cloud in the original point cloud data, determining the weight of each original point cloud according to the characteristics of each original point cloud, determining a preset number of original point clouds as target point clouds according to the sequence from high to low of the weights, determining the target point cloud data according to the target point clouds, and generating a high-precision map according to the target point cloud data;
the receiving and sending module is used for receiving request information from equipment to be positioned, and the request information is used for requesting a high-precision map;
the transceiver module is further configured to send the high-precision map to the device to be positioned.
18. A positioning device, comprising:
the radar module is used for acquiring point cloud data around the equipment to be positioned;
the processing module is used for determining a predicted position of the equipment to be positioned, extracting point cloud data at the predicted position from a high-precision map as candidate point cloud data according to the predicted position, acquiring an error distance of the predicted position according to the candidate point cloud data, the point cloud data around the equipment to be positioned and a feature matching model, and determining the position of the equipment to be positioned according to the predicted position and the error distance; the point cloud data in the high-precision map is target point cloud data determined according to the weight of the characteristics of the point cloud in pre-acquired original point cloud data, and the characteristic matching model is used for representing the corresponding relation between the characteristic matching degree and the error distance of the point cloud data.
19. An electronic device, comprising:
at least one processor; and
a memory communicatively coupled to the at least one processor; wherein the content of the first and second substances,
the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-16.
20. A non-transitory computer readable storage medium having stored thereon computer instructions for causing the computer to perform the method of any one of claims 1-16.
CN202010455341.3A 2020-05-26 2020-05-26 Positioning method, positioning device, electronic equipment and storage medium Active CN111638528B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010455341.3A CN111638528B (en) 2020-05-26 2020-05-26 Positioning method, positioning device, electronic equipment and storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010455341.3A CN111638528B (en) 2020-05-26 2020-05-26 Positioning method, positioning device, electronic equipment and storage medium

Publications (2)

Publication Number Publication Date
CN111638528A true CN111638528A (en) 2020-09-08
CN111638528B CN111638528B (en) 2023-05-30

Family

ID=72331030

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010455341.3A Active CN111638528B (en) 2020-05-26 2020-05-26 Positioning method, positioning device, electronic equipment and storage medium

Country Status (1)

Country Link
CN (1) CN111638528B (en)

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112462372A (en) * 2021-01-29 2021-03-09 北京主线科技有限公司 Vehicle positioning method and device
CN112543859A (en) * 2020-10-28 2021-03-23 华为技术有限公司 Positioning method, positioning device, electronic equipment and storage medium
CN112859136A (en) * 2021-01-19 2021-05-28 腾讯科技（深圳）有限公司 Positioning method and related device
CN113640779A (en) * 2021-10-15 2021-11-12 北京一径科技有限公司 Radar failure determination method and device, and storage medium
CN114051628A (en) * 2020-10-30 2022-02-15 华为技术有限公司 Method and device for determining target object point cloud set
CN115526936A (en) * 2022-11-29 2022-12-27 长沙智能驾驶研究院有限公司 Training method of positioning model and point cloud data positioning method and device

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108921947A (en) * 2018-07-23 2018-11-30 百度在线网络技术（北京）有限公司 Generate method, apparatus, equipment, storage medium and the acquisition entity of electronic map
CN109064506A (en) * 2018-07-04 2018-12-21 百度在线网络技术（北京）有限公司 Accurately drawing generating method, device and storage medium
CN109141446A (en) * 2018-07-04 2019-01-04 百度在线网络技术（北京）有限公司 For obtaining the method, apparatus, equipment and computer readable storage medium of map
CN109459734A (en) * 2018-10-30 2019-03-12 百度在线网络技术（北京）有限公司 A kind of laser radar locating effect appraisal procedure, device, equipment and storage medium
CN110084272A (en) * 2019-03-26 2019-08-02 哈尔滨工业大学（深圳） A kind of cluster map creating method and based on cluster map and the matched method for relocating of location expression
CN110794392A (en) * 2019-10-15 2020-02-14 上海创昂智能技术有限公司 Vehicle positioning method and device, vehicle and storage medium
CN110873568A (en) * 2018-08-30 2020-03-10 百度在线网络技术（北京）有限公司 High-precision map generation method and device and computer equipment
CN111144483A (en) * 2019-12-26 2020-05-12 歌尔股份有限公司 Image feature point filtering method and terminal

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109064506A (en) * 2018-07-04 2018-12-21 百度在线网络技术（北京）有限公司 Accurately drawing generating method, device and storage medium
CN109141446A (en) * 2018-07-04 2019-01-04 百度在线网络技术（北京）有限公司 For obtaining the method, apparatus, equipment and computer readable storage medium of map
CN108921947A (en) * 2018-07-23 2018-11-30 百度在线网络技术（北京）有限公司 Generate method, apparatus, equipment, storage medium and the acquisition entity of electronic map
CN110873568A (en) * 2018-08-30 2020-03-10 百度在线网络技术（北京）有限公司 High-precision map generation method and device and computer equipment
CN109459734A (en) * 2018-10-30 2019-03-12 百度在线网络技术（北京）有限公司 A kind of laser radar locating effect appraisal procedure, device, equipment and storage medium
CN110084272A (en) * 2019-03-26 2019-08-02 哈尔滨工业大学（深圳） A kind of cluster map creating method and based on cluster map and the matched method for relocating of location expression
CN110794392A (en) * 2019-10-15 2020-02-14 上海创昂智能技术有限公司 Vehicle positioning method and device, vehicle and storage medium
CN111144483A (en) * 2019-12-26 2020-05-12 歌尔股份有限公司 Image feature point filtering method and terminal

Cited By (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112543859A (en) * 2020-10-28 2021-03-23 华为技术有限公司 Positioning method, positioning device, electronic equipment and storage medium
CN112543859B (en) * 2020-10-28 2022-07-15 华为技术有限公司 Positioning method, positioning device, electronic equipment and storage medium
CN114051628A (en) * 2020-10-30 2022-02-15 华为技术有限公司 Method and device for determining target object point cloud set
CN112859136A (en) * 2021-01-19 2021-05-28 腾讯科技（深圳）有限公司 Positioning method and related device
CN112859136B (en) * 2021-01-19 2022-03-25 腾讯科技（深圳）有限公司 Positioning method and related device
CN112462372A (en) * 2021-01-29 2021-03-09 北京主线科技有限公司 Vehicle positioning method and device
CN112462372B (en) * 2021-01-29 2021-06-15 北京主线科技有限公司 Vehicle positioning method and device
CN113640779A (en) * 2021-10-15 2021-11-12 北京一径科技有限公司 Radar failure determination method and device, and storage medium
CN113640779B (en) * 2021-10-15 2022-05-03 北京一径科技有限公司 Radar failure determination method and device, and storage medium
WO2023061481A1 (en) * 2021-10-15 2023-04-20 北京一径科技有限公司 Radar failure determining method and device and storage medium
CN115526936A (en) * 2022-11-29 2022-12-27 长沙智能驾驶研究院有限公司 Training method of positioning model and point cloud data positioning method and device
CN115526936B (en) * 2022-11-29 2023-04-28 长沙智能驾驶研究院有限公司 Training method of positioning model and point cloud data positioning method and device

Also Published As

Publication number Publication date
CN111638528B (en) 2023-05-30

Similar Documents

Publication Publication Date Title
CN111638528B (en) Positioning method, positioning device, electronic equipment and storage medium
CN110556012B (en) Lane positioning method and vehicle positioning system
EP3798974B1 (en) Method and apparatus for detecting ground point cloud points
CN111324115B (en) Obstacle position detection fusion method, obstacle position detection fusion device, electronic equipment and storage medium
CN110806215B (en) Vehicle positioning method, device, equipment and storage medium
JP7328282B2 (en) Method, device, device and storage medium for positioning vehicle
CN111174799A (en) Map construction method and device, computer readable medium and terminal equipment
CN112132829A (en) Vehicle information detection method and device, electronic equipment and storage medium
CN111563450B (en) Data processing method, device, equipment and storage medium
CN111310840B (en) Data fusion processing method, device, equipment and storage medium
CN111220164A (en) Positioning method, device, equipment and storage medium
CN105229490A (en) Use the positional accuracy of satellite visibility data for promoting
CN111352434B (en) Device and method for supporting aircraft approaching airport runway of airport
CN110703732B (en) Correlation detection method, device, equipment and computer readable storage medium
CN111079079B (en) Data correction method, device, electronic equipment and computer readable storage medium
CN113887400B (en) Obstacle detection method, model training method and device and automatic driving vehicle
CN112147632A (en) Method, device, equipment and medium for testing vehicle-mounted laser radar perception algorithm
CN111784837A (en) High-precision map generation method and device
CN112101209A (en) Method and apparatus for determining a world coordinate point cloud for roadside computing devices
CN112344855A (en) Obstacle detection method and device, storage medium and drive test equipment
CN113844463A (en) Vehicle control method and device based on automatic driving system and vehicle
CN110554685A (en) test method, device, equipment and storage medium for planning control system
CN111597987A (en) Method, apparatus, device and storage medium for generating information
CN113177980B (en) Target object speed determining method and device for automatic driving and electronic equipment
CN111612851B (en) Method, apparatus, device and storage medium for calibrating camera

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant