CN112241414B - High-precision map updating method and device - Google Patents
High-precision map updating method and device Download PDFInfo
- Publication number
- CN112241414B CN112241414B CN201910650809.1A CN201910650809A CN112241414B CN 112241414 B CN112241414 B CN 112241414B CN 201910650809 A CN201910650809 A CN 201910650809A CN 112241414 B CN112241414 B CN 112241414B
- Authority
- CN
- China
- Prior art keywords
- attribute
- continuous numerical
- data
- module
- classification data
- 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.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/23—Updating
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/29—Geographical information databases
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Databases & Information Systems (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Remote Sensing (AREA)
- Traffic Control Systems (AREA)
- Instructional Devices (AREA)
- Navigation (AREA)
Abstract
The invention provides a high-precision map updating method and device. The method comprises the following steps: obtaining M x N decomposition results of M continuous numerical attributes of the same map element, wherein N represents the collection times of crowdsourcing data of the same map element, M is more than or equal to 1, and N is more than or equal to 2; according to the M x N decomposition results, denoising each continuous numerical type attribute respectively to obtain a denoising result of each continuous numerical type attribute; and updating the high-precision map according to the denoising result. Compared with the prior art that the high-precision map is updated directly on the basis of crowdsourcing reported data, the updating method improves the precision of the high-precision map.
Description
Technical Field
The invention relates to the field of unmanned driving, in particular to a high-precision map updating method and device.
Background
In recent years, unmanned vehicles are not paid much attention by governments of various countries, the unmanned technology is the materialization of understanding, learning and memorizing the process of 'environment perception-decision and planning-control and execution' by human drivers in long-term driving practice, and the unmanned vehicles are complex intelligent automatic systems with software and hardware combined. In the field of unmanned driving, a high-precision map is used as a service provider of prior environmental information and plays an important role in the processes of high-precision positioning, environment perception assistance, planning and decision making. At present, the industry basically collects laser point clouds and images through a mobile measurement system, and then generates and updates a high-precision map through a manual, semi-automatic or full-automatic production platform, however, the method is high in cost, and the freshness does not meet the requirement.
In view of this, a method for updating a high-precision map by using a crowdsourcing data source is developed in the industry, however, due to the fact that data provided by the crowdsourcing data source is relatively noisy, the updating method in the prior art cannot meet the requirement of the high-precision map on precision.
Disclosure of Invention
The invention provides a high-precision map updating method and device, which are used for improving the precision of a high-precision map.
In a first aspect, the present invention provides a high-precision map updating method, including:
obtaining M x N decomposition results of M continuous numerical attributes of the same map element, wherein N represents the collection times of crowdsourcing data of the same map element, M is more than or equal to 1, and N is more than or equal to 2;
according to the M x N decomposition results, denoising each continuous numerical type attribute respectively to obtain a denoising result of each continuous numerical type attribute;
and updating the high-precision map according to the denoising result.
In a second aspect, the present invention provides a high-precision map updating apparatus, including:
the map feature extraction module is used for extracting M x N decomposition results of M kinds of continuous numerical attributes of the same map element, wherein N represents the collection times of crowdsourcing data of the same map element, M is greater than or equal to 1, and N is greater than or equal to 2;
the denoising module is used for respectively denoising each continuous numerical type attribute according to the M x N decomposition results to obtain a denoising result of each continuous numerical type attribute;
and the updating module is used for updating the high-precision map according to the denoising result.
In a third aspect, the present invention provides a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, implements the above-described high-precision map updating method.
In a fourth aspect, the present invention provides a cloud server, including:
the receiving module is used for receiving N times of crowdsourcing data of the same map element;
the classification module is used for classifying crowdsourcing data each time to obtain M x N classification data of M continuous numerical attributes of the same map element;
the decomposition module is used for decomposing the M x N classified data to obtain M x N decomposition results;
the denoising module is used for respectively denoising each continuous numerical type attribute according to the M x N decomposition results to obtain a denoising result of each continuous numerical type attribute;
and the updating module is used for updating the high-precision map according to the denoising result.
In a fifth aspect, the present invention provides a map updating system, including: the system comprises an intelligent terminal and a cloud server;
wherein, intelligent terminal includes: the device comprises an acquisition module, a classification module, a decomposition module and an uploading module;
the acquisition module is used for acquiring crowdsourcing data of the same map element;
the classification module is used for classifying the crowdsourcing data to obtain classification data of M kinds of continuous numerical attributes of the same map element;
the decomposition module is used for decomposing the classified data to obtain a decomposition result;
the uploading module is used for uploading the decomposition result to the cloud server;
wherein the cloud server comprises: the device comprises a receiving module, a denoising module, an updating module and a sending module;
the receiving module is configured to receive M × N decomposition results uploaded by one or more intelligent terminals, where N represents the number of times that one or more intelligent terminals acquire crowdsourcing data of the same map element;
the denoising module is used for respectively denoising each continuous numerical type attribute according to the M x N decomposition results to obtain a denoising result of each continuous numerical type attribute;
the updating module is used for updating the high-precision map according to the denoising result;
and the sending module is used for sending the updated high-precision map to the intelligent terminal.
The high-precision map updating method and the high-precision map updating device firstly obtain M x N decomposition results of M continuous numerical attributes of the same map element, then carry out denoising treatment on each continuous numerical attribute according to the M x N decomposition results to obtain a denoising result of each continuous numerical attribute, and finally update the high-precision map according to the denoising result. Compared with the prior art that the high-precision map is updated directly on the basis of crowdsourcing reported data, the updating method improves the precision of the high-precision map.
Drawings
Fig. 1 is an optional application scenario diagram of the high-precision map updating method provided by the present invention;
FIG. 2 is a flowchart illustrating a first exemplary embodiment of a high-precision map updating method according to the present invention;
fig. 3 is a first schematic diagram of a cloud server acquiring M × N decomposition results according to the present invention;
fig. 4 is a schematic diagram ii of the cloud server obtaining M × N decomposition results according to the present invention;
FIG. 5 is a flowchart illustrating a second embodiment of a high-precision map updating method according to the present invention;
FIG. 6 is a schematic diagram of a shape attribute decomposition provided by the present invention;
FIG. 7 is a flowchart illustrating a third embodiment of a high-precision map updating method according to the present invention;
FIG. 8 is a schematic structural diagram of an embodiment of a high-precision map updating apparatus provided in the present invention;
fig. 9 is a schematic diagram of a hardware structure of a cloud server provided in the present invention;
fig. 10 is a schematic diagram of a hardware structure of the map updating system provided by the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all embodiments of the present invention. All other embodiments, which can be obtained by a person skilled in the art without inventive step based on the embodiments of the present invention, are within the scope of protection of the present invention.
The terms "first," "second," "third," "fourth," and the like in the description and in the claims, as well as in the drawings, if any, are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used is interchangeable under appropriate circumstances such that the embodiments of the invention described herein are, for example, capable of operation in sequences other than those illustrated or otherwise described herein.
Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
At present, crowdsourcing data reported by a crowdsourcing vehicle is a main data source for updating a high-precision map. However, compared with the data collected by the high-precision data collection vehicle, the crowdsourcing data has higher noise, and the high-precision map is updated directly by taking the crowdsourcing data as a basis, so that the precision requirement of the high-precision map cannot be met.
Based on the technical problem, the invention provides a high-precision map updating method and high-precision map updating equipment. After obtaining M x N decomposition results of M continuous numerical attributes of the same map element, the cloud server respectively performs denoising processing on each continuous numerical attribute according to the M x N decomposition results to obtain a denoising result of each continuous numerical attribute; and then updating the high-precision map according to the denoising result. Compared with the prior art that the high-precision map is updated directly on the basis of crowdsourcing reported data, the updating method improves the precision of the high-precision map.
Fig. 1 is an optional application scenario diagram of the high-precision map updating method provided by the present invention. The application scenario diagram shown in fig. 1 includes: crowdsourcing equipment and cloud servers.
The crowdsourcing car comprises a crowdsourcing car body, a crowdsourcing data acquisition system, a vehicle-mounted sensor and a crowdsourcing data acquisition system, wherein the crowdsourcing car body is provided with the crowdsourcing data acquisition system, the crowdsourcing data acquisition system comprises various types of vehicle-mounted sensors, and the crowdsourcing car body can acquire crowdsourcing data through the sensors.
Wherein, cloud server and crowdsourcing car pass through wireless communication technology and establish the connection, and the crowdsourcing car machine's that connects with cloud server quantity can be a plurality of, and crowdsourcing car machine can upload to cloud server with the crowdsourcing data of gathering or the decomposition result of handling and obtaining. The cloud server can complete the denoising of the continuous numerical attributes and the updating of the high-precision map by the method provided by the invention.
The following describes the technical solutions of the present invention and how to solve the above technical problems with specific embodiments. The following several specific embodiments may be combined with each other, and details of the same or similar concepts or processes may not be repeated in some embodiments. Embodiments of the present invention will be described below with reference to the accompanying drawings.
Fig. 2 is a schematic flowchart of a first embodiment of a high-precision map updating method provided by the present invention. The high-precision map updating method provided by the embodiment can be executed by the cloud server shown in fig. 1. As shown in fig. 2, the high-precision map updating method provided in this embodiment includes:
s201, M x N decomposition results of M kinds of continuous numerical attributes of the same map element are obtained.
Wherein N represents the collection times of crowdsourcing data of the same map element, M is greater than or equal to 1, and N is greater than or equal to 2.
A first implementation manner for the cloud server to obtain the M × N decomposition results is as follows:
and A, receiving N times of crowdsourcing data of the same map element.
The map elements may be elements distributed discretely along the road direction on the map, that is, the map elements herein may be discrete map elements, and the discrete map elements include but are not limited to: traffic signs, shafts, traffic lights and floor print, wherein the floor print may be, for example, arrows or text.
The N-time crowdsourcing data received by the cloud server can be reported after being collected by the same crowdsourcing machine, and can also be reported after being collected by different crowdsourcing machines.
The greater the value of N is required to be greater than or equal to 2, the higher the accuracy of the high-accuracy map obtained by the updating method provided by the embodiment is.
And B, classifying the crowdsourcing data every time to obtain M x N classification data of the M continuous numerical attributes of the same map element.
The attributes of the discrete map elements include two types: continuous numeric attributes and discrete enumerated attributes. Wherein the continuous numerical attribute includes but is not limited to: a location attribute, a shape attribute, and a pose attribute. Discrete enumerated attributes include, but are not limited to: semantic attributes.
Specifically, when the continuous numerical attribute refers to a position attribute, N classification data of the position attribute can be obtained after classifying crowd-sourced data in step B each time; similarly, when the continuous numerical attribute refers to the shape attribute, N classification data of the shape attribute can be obtained after classifying the crowdsourcing data in step B each time; when the continuous numerical attribute refers to the attitude attribute, N classification data of the attitude attribute can be obtained after classifying the crowdsourcing data in the step B each time; when the continuous numerical attribute includes two or more attributes, for example: the continuous numerical type attributes include a position attribute, a shape attribute and a posture attribute, and after the crowd-sourced data is classified in the step B, N classification data (3 × N classification data in total) of three attributes of the position attribute, the shape attribute and the posture attribute can be obtained. For the case that the continuous numerical attribute is other attribute combination, there is no further example here.
And C, decomposing the M x N classified data to obtain M x N decomposition results.
Specifically, when the continuous numerical attribute refers to a position attribute, an orthogonal model may be introduced in step C to decompose N classification data of the position attribute, and three directions of the orthogonal model are: the N classification data of the position attribute are decomposed in the vertical road direction, the along road direction, and the height direction, and each classification data is decomposed into the above three directions.
Specifically, the continuous numerical type attribute refers to a shape attribute, and the N classification data of the shape attribute is a direct expression of the shape attribute, that is, the N classification data of the shape attribute contains shape point information of a map element, such as: the longitude and latitude coordinates of four corner points of the rectangular traffic sign, the longitude and latitude coordinates of each corner point of the rod-shaped object, the longitude and latitude coordinates of each corner point of the ground arrow and the like. The decomposition of the N classification data for the shape attributes may be achieved by converting the shape point information into a modeled representation of the shape attributes.
Specifically, similar to the shape attribute, the continuous numerical attribute refers to the pose attribute, and the N classification data of the pose attribute is a direct expression of the pose attribute, that is, the N classification data of the pose attribute includes shape point information of the map element, such as: longitude and latitude coordinates of four corner points of the rectangular traffic sign, longitude and latitude coordinates of each corner point of the rod-shaped object, longitude and latitude coordinates of each corner point of a ground arrow and the like. The decomposition of the N classification data for the pose attributes may be achieved by converting the shape point information into a modeled representation of the pose attributes.
When the continuous numerical attribute includes two or more attributes, for example: the continuous numerical type attribute includes: the position attribute, the shape attribute and the posture attribute can be combined with the decomposition method corresponding to each attribute to realize the decomposition of the classification data of each attribute, and N decomposition results of each attribute are obtained. The present invention will not be described in detail herein.
In the foregoing implementation manner, referring to fig. 3, the crowdsourcing equipment is configured to collect crowdsourcing data of the map elements, and report the collected crowdsourcing data to the cloud server, and the cloud server is configured to classify and decompose the received crowdsourcing data, so as to obtain the M × N decomposition results.
A second implementation manner for the cloud server to obtain the M × N decomposition results is as follows:
and receiving M x N decomposition results uploaded by the crowdsourcing machine.
In the foregoing implementation manner, referring to fig. 4, the crowdsourcing equipment is configured to collect crowdsourcing data of map elements, classify and decompose the crowdsourcing data collected each time, and upload a decomposition result to the cloud server, so that the cloud server performs denoising processing according to M × N received decomposition results, as described above, the M × N decomposition results may be uploaded by one crowdsourcing equipment or by multiple crowdsourcing equipment, and the number of the crowdsourcing equipment is not limited in the present invention, which is shown in fig. 4.
S202, according to the M x N decomposition results, denoising is respectively carried out on each continuous numerical type attribute, and a denoising result of each continuous numerical type attribute is obtained.
Optionally, the M × N decomposition results may be filtered in combination with the confidence and the confidence interval to implement denoising processing, so as to obtain a denoising result.
S203, updating the high-precision map according to the denoising result.
Specifically, after the denoising result is obtained, the continuous numerical attribute may be restored according to the denoising result, and then the high-precision map may be updated according to the restoration result.
The high-precision map updating method provided in this embodiment includes obtaining M × N decomposition results of M continuous numerical attributes of the same map element, performing denoising processing on each continuous numerical attribute according to the M × N decomposition results to obtain a denoising result of each continuous numerical attribute, and updating the high-precision map according to the denoising result. Compared with the prior art that the high-precision map is updated directly based on crowdsourcing reported data, the updating method improves the precision of the high-precision map.
Referring to the above description, the continuous numerical type attribute of the map element may refer to one attribute, or may include two or more attributes, and the following describes in detail the decomposition process performed by the cloud server in the above embodiment by taking the example that the continuous numerical type attribute includes a position attribute, a shape attribute, and a posture attribute. Fig. 5 is a flowchart illustrating a second embodiment of a high-precision map updating method according to the present invention. As shown in fig. 5, the high-precision map updating method provided by this embodiment includes:
s501, receiving the crowdsourcing data of the same map element for N times.
For the implementation of S501, reference may be made to step a in the above embodiment, and details of the present invention are not repeated herein.
S502, classifying the crowdsourcing data every time to obtain M x N classification data of the M continuous numerical attributes of the same map element.
Since the continuous numerical attributes in this embodiment include the position attribute, the shape attribute, and the posture attribute, after classifying crowd-sourced data each time, 3 × N classification data of the three continuous numerical attributes are obtained.
Optionally, the N classification data of the location attribute may be: and (4) reporting the longitude and latitude coordinates of the central point of the map element by the crowdsourcing machine for N times. The N classification data for the shape attribute may be: and (3) converting N classified data of the shape attributes into modeled expressions of the shape attributes through the following decomposition process under the condition that the longitude and latitude coordinates of each corner point of the map elements are reported by the crowdsourcing machine for N times. The N classification data for the pose attributes may be: and (3) converting N classified data of the attitude attributes into modeled expressions of the attitude attributes through the following decomposition process under the condition that the longitude and latitude coordinates of each corner of the map elements are reported by the crowdsourcing machine for N times.
S503, decomposing the N classification data of the position attribute, the shape attribute and the posture attribute respectively to obtain N decomposition results of the position attribute, the shape attribute and the posture attribute.
The following explains the decomposition process of N classification data of the location attribute:
and step A, determining the average central point of the same map element under a longitude and latitude coordinate system according to the N classification data of the position attribute.
Specifically, as can be seen from the above description, the N classification data of the location attribute may be longitude and latitude coordinates of a center point of the map element that is reported by the crowdsourcing equipment N times, and the implementation manner of determining the coordinate of the average center point may be: firstly, converting N classified data of the position attributes into a geocentric earth-fixed ECEF coordinate system, then solving an average central point in the ECEF coordinate system, and finally converting the ECEF coordinate of the average central point into a longitude and latitude coordinate system, thereby obtaining the average central point of the map element in the longitude and latitude coordinate system.
And B, converting the N classified data into an ENU (north east China coordinate system) with the average central point as an origin to obtain the ENU coordinates of the N classified data.
And step C, determining the road direction of the road corresponding to the average center point.
Optionally, a spatial index may be established for all roads on the high-precision map, the road corresponding to the average center point is found through the spatial index, a perpendicular line is drawn from the center point to the found road to obtain a perpendicular point, and the road direction of the road passing through the perpendicular point on the high-precision map is the road direction to be determined in this step.
And D, determining orthogonal model coordinates of the N classified data according to the ENU coordinates of the N classified data and the road direction, wherein the orthogonal model coordinates of the N classified data are N decomposition results of the position attribute.
Specifically, after the road direction is determined in step C, firstly, an angle α of clockwise rotation of the due north direction to the road direction is determined according to the road direction and the due north direction, and then, the ENU coordinates of the N classification data are clockwise rotated by the angle α, so that the orthogonal model coordinates of the N classification data can be obtained, where the orthogonal model coordinates of the N classification data are the N decomposition results of the position attribute.
The following explains a decomposition process of N classification data of shape attributes:
for each classification data in the N classification data of the shape attribute, firstly, sorting the angular points in each classification data to obtain the serial number of each angular point; then, according to the serial number of each corner point, determining the modeling expression corresponding to each classification data, wherein the modeling expressions corresponding to the N classification data of the shape attribute are N decomposition results of the shape attribute.
Take a rectangular traffic sign as an example: referring to fig. 6, assuming that the classification data includes longitude and latitude coordinates of 4 corner points, the 4 corner points are along the road direction, the upper left corner is set as serial number 1, the upper right corner is set as serial number 2, the lower right corner is set as serial number 3, and the lower left corner is set as serial number 4, and then the classification data is converted into the width and height of the rectangular sign through the following formula, the width and height of the rectangular sign are modeled expressions of the shape attribute of the rectangular sign, and length is used as the length of the rectangular sign ij Represents the distance between two corner points:
the modeled representation of the shape attribute of the rectangular sign in this embodiment is represented by width and height. Alternatively, the modeled representation of a triangular traffic sign (isosceles triangle) may be represented by base width and height; the modeled expression of the shaft can be represented by a shaft length, an upper bottom diameter and a lower bottom diameter; the modeled representation of the floor print, such as an arrow, can be represented by the length and width of the bounding rectangle.
Each classification data in the N classification data of the shape attribute can be decomposed through the method to obtain N modeling expressions of the shape attribute, and the N modeling expressions are N decomposition results of the shape attribute.
The following explains the decomposition process of N classification data of the pose attributes:
and aiming at each classification data in the N classification data of the attitude attribute, adopting quaternion attitude calculation or rotation matrix attitude calculation to convert the corresponding classification data into the modeling expression of the attitude attribute, wherein the modeling expression corresponding to the N classification data of the attitude attribute is N decomposition results of the attitude attribute.
The following describes the decomposition process using quaternion attitude resolution (taking a rectangular traffic sign as an example):
and step A, fitting out a label plane by using longitude and latitude coordinates of each corner point in the reported data through a least square method to obtain a plane equation of the label plane in the three-dimensional space.
Step B, taking the east-righting direction under the EUN coordinate system as the positive direction and the central point of the label as the origin, and calculating the unit vector of the rotary shaft of the plane of the label
And equivalent rotation angle θ, resulting in a quaternion q = (x, y, z, w), where x, y, z, w are:
step C, converting the quaternion q into an attitude angle, wherein the attitude angle is a modeling expression of the attitude attribute of the rectangular sign:
where roll denotes a roll angle, pitch denotes a pitch angle, and yaw denotes a yaw angle.
Each classification data in the N classification data of the attitude attribute can be decomposed through the method, and N modeling expressions of the attitude attribute are obtained, wherein the N modeling expressions are N decomposition results of the attitude attribute.
S504, denoising is conducted according to the position attribute, the shape attribute and the N decomposition results of the posture attribute, and denoising results of the position attribute, the shape attribute and the posture attribute are obtained.
Specifically, after N decomposition results of the position attribute, the shape attribute, and the posture attribute are obtained in S503, the N decomposition results of the three attributes are subjected to denoising processing, so that denoising results of the three attributes can be obtained.
And S505, updating the high-precision map according to the denoising result of the position attribute, the denoising result of the shape attribute and the denoising result of the attitude attribute.
Specifically, the restoration can be performed according to the denoising result of the position attribute to obtain a restoration result of the position attribute; restoring according to the denoising result of the shape attribute to obtain a restoration result of the shape attribute; restoring according to the denoising result of the attitude attribute to obtain a restoration result of the attitude attribute; and then updating the high-precision map according to the restoration results of the three attributes.
The high-precision map updating method provided by the embodiment describes decomposition modes of the position attribute, the shape attribute and the posture attribute, and provides data support for subsequent denoising processing.
The following describes a process of denoising processing in the above embodiments with reference to specific embodiments. Fig. 7 is a flowchart illustrating a third embodiment of the high-precision map updating method according to the present invention. As shown in fig. 7, the high-precision map updating method provided by this embodiment includes:
and S701, receiving the N times of crowdsourcing data of the same map element.
S702, classifying the crowdsourcing data every time to obtain M x N classification data of the M continuous numerical attributes of the same map element.
S703, decomposing the M x N classified data to obtain M x N decomposition results.
The implementation manners of S701-S703 may refer to the above embodiments, and the present invention is not described herein again.
S704, calculating the mean value and the standard deviation of the N decomposition results of each continuous numerical type attribute.
S705, determining a confidence interval according to a preset confidence coefficient, the mean value and the standard deviation.
S706, filtering the N decomposition results according to the confidence interval to obtain filtering results.
And S707, taking the average value of the filtering results.
And S708, taking the mean value as a denoising result corresponding to the continuous numerical attribute.
Specifically, for any one of the above-described position attribute, shape attribute, and posture attribute, the denoising processing may be performed by using the method of S704-S708.
The following describes the above processes of S704-S708 by taking the location attribute as an example:
after N decomposition results of the position attribute are obtained through S703 decomposition, the mean value of the N decomposition results is calculated
Sum standard deviation σ c . Alternatively, it can be at the standard deviation σ c If the value is greater than the predetermined threshold value, S705 is executed. The preset confidence level β in S705 may be selected as required, for example, it may be: 95% or 99%, etc. Taking β =95% as an example, it can be foundThe confidence interval should be
Within the range. And comparing the N decomposition results with the confidence interval, and eliminating the decomposition results of which the N decomposition results are not in the confidence interval to obtain a filtering result. Then, the residual decomposition results (i.e. filtering results) after the elimination process are averaged, and the average value can be used as the denoising result of the position attribute.
And S709, updating the high-precision map according to the denoising result.
Specifically, as described above, after the denoising result is obtained, the continuous numerical attribute may be restored according to the denoising result, and then the high-precision map may be updated according to the restoration result.
The following explains the process of restoring the location attribute:
for the position attribute, the orthogonal model coordinate of the center point of the map element after denoising can be obtained through the decomposition and denoising processes, and the reduction process can be realized through the following modes: and rotating the orthogonal model coordinate of the central point of the map element counterclockwise by the angle alpha in the decomposition process so as to convert the orthogonal model coordinate into an ENU coordinate, and then reducing the ENU coordinate of the central point of the map element into a longitude and latitude coordinate by taking the average central point in the decomposition process as a reference point.
The following explains the process of restoring the shape attribute:
for the shape attribute, through the decomposition and denoising processes, a modeled expression of the denoised shape attribute can be obtained, and the reduction process can be realized through the following method: and calculating the coordinate offset of each corner point of each map element relative to the central point of the map element according to the modeled expression of the denoised shape attribute, and restoring to obtain the longitude and latitude coordinates of each corner point of each map element according to the longitude and latitude coordinates of the central point restored from the position attribute and the coordinate offset obtained by calculation.
The process of restoring the posture attribute is similar to the process of restoring the shape attribute, and the details are not repeated herein.
It should be noted that: the shape attribute and the posture attribute are restored by considering the scene of the back-end application. If the denoised modeled expression can meet the use requirement of the back end, the restoration process is not needed, and the denoised modeled expression is directly updated.
The high-precision map updating method provided by the embodiment describes an implementation mode of denoising processing, and the high-precision map is updated according to a denoising result, so that the precision of the high-precision map is greatly improved, and the requirement of the high-precision map on the precision is met.
Fig. 8 is a schematic structural diagram of an embodiment of a high-precision map updating apparatus provided in the present invention. As shown in fig. 8, the high-precision map updating apparatus provided by the present embodiment includes:
an obtaining module 801, configured to obtain M × N decomposition results of M continuous numerical attributes of the same map element, where N represents a collection frequency of crowdsourcing data of the same map element, M is greater than or equal to 1, and N is greater than or equal to 2;
a denoising module 802, configured to perform denoising processing on each of the continuous numerical attributes according to the M × N decomposition results, to obtain a denoising result of each of the continuous numerical attributes;
and the updating module 803 is configured to update the high-precision map according to the denoising result.
Optionally, the obtaining module 801 is specifically configured to:
receiving N times of crowdsourcing data for the same map element;
classifying crowdsourcing data each time to obtain M x N classification data of the M continuous numerical attributes of the same map element;
and decomposing the M x N classified data to obtain M x N decomposition results.
Optionally, the obtaining module 801 is specifically configured to:
receiving M x N decomposition results uploaded by a crowdsourcing vehicle, wherein the M x N decomposition results are obtained after the crowdsourcing vehicle classifies and decomposes crowdsourcing data acquired each time.
Optionally, the continuous numerical attribute includes: at least one of a position attribute, a shape attribute, and a pose attribute.
Optionally, when the continuous numerical attribute is a position attribute, the obtaining module 801 is specifically configured to:
determining the average center point of the same map element under a longitude and latitude coordinate system according to the N classification data of the position attribute;
converting the N classified data into an ENU (north east China Unit) coordinate system with the average central point as an origin to obtain ENU coordinates of the N classified data;
determining the road direction of the road corresponding to the average center point;
and determining orthogonal model coordinates of the N classification data according to the ENU coordinates of the N classification data and the road direction, wherein the orthogonal model coordinates of the N classification data are N decomposition results of the position attribute.
Optionally, the obtaining module 801 is specifically configured to:
according to the road direction and the due north direction, determining the angle of the due north direction rotating clockwise to the road direction;
and clockwise rotating the ENU coordinates of the N classified data by the angle to obtain orthogonal model coordinates of the N classified data.
Optionally, when the continuous numerical attribute is a shape attribute, the obtaining module 801 is specifically configured to:
sorting the corner points in each classification data of the shape attribute to obtain the serial number of each corner point;
and determining the modeling expression corresponding to each classification data according to the serial number of each corner point, wherein the modeling expressions corresponding to the N classification data of the shape attribute are N decomposition results of the shape attribute.
Optionally, when the continuous numerical attribute is a posture attribute, the obtaining module 801 is specifically configured to:
and aiming at each classification data of the attitude attribute, adopting quaternion attitude calculation or rotation matrix attitude calculation to convert the corresponding classification data into the modeling expression of the attitude attribute, wherein the modeling expression corresponding to the N classification data of the attitude attribute is N decomposition results of the attitude attribute.
Optionally, the denoising module 802 is specifically configured to:
calculating the mean and standard deviation of the N decomposition results of each of the continuous numerical attributes;
determining a confidence interval according to a preset confidence coefficient, the mean value and the standard deviation;
filtering the N decomposition results according to the confidence interval to obtain filtering results;
taking the average value of the filtering results;
and taking the mean value as a denoising result corresponding to the continuous numerical attribute.
Optionally, the update module 803 is specifically configured to:
restoring the continuous numerical type attribute according to the denoising result to obtain a restoration result;
and updating the high-precision map according to the reduction result.
The high-precision map updating apparatus provided in this embodiment may be used to execute the high-precision map updating method described in any of the above embodiments, and the implementation principle and the technical effect are similar, which are not described herein again.
Fig. 9 is a schematic diagram of a hardware structure of a cloud server provided in the present invention. As shown in fig. 9, the cloud server provided by the present invention may include:
a receiving module 901, configured to receive crowdsourcing data of the same map element for N times;
a classification module 902, configured to classify crowdsourcing data each time to obtain M × N classification data of M continuous numerical attributes of the same map element;
a decomposition module 903, configured to decompose the M x N classification data to obtain M x N decomposition results;
a denoising module 904, configured to perform denoising processing on each of the continuous numerical attributes according to the M × N decomposition results, to obtain a denoising result of each of the continuous numerical attributes;
and the updating module 905 is configured to update the high-precision map according to the denoising result.
Each module in the cloud server may be configured to execute each step corresponding to the server side in the embodiment shown in fig. 3, and the implementation principle and the technical effect are similar, which are not described herein again.
The present invention provides a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, implements the high-precision map updating method described in any of the above embodiments.
The present invention also provides a program product including a computer program stored in a readable storage medium, the computer program being readable from the readable storage medium by at least one processor, the computer program being executable by the at least one processor to cause a cloud server to implement the high-precision map updating method described in any of the above embodiments.
Fig. 10 is a schematic diagram of a hardware structure of the map updating system provided by the present invention. As shown in fig. 10, the map updating system provided by the present invention may include: the system comprises an intelligent terminal and a cloud server;
wherein, intelligent terminal includes: the system comprises an acquisition module 1001, a classification module 1002, a decomposition module 1003 and an uploading module 1004;
the acquisition module 1001 is configured to acquire crowdsourcing data of the same map element;
the classification module 1002 is configured to classify the crowdsourcing data to obtain classification data of M kinds of continuous numerical attributes of the same map element;
the decomposition module 1003 is configured to decompose the classified data to obtain a decomposition result;
the uploading module 1004 is configured to upload the decomposition result to the cloud server;
wherein the cloud server comprises: a receiving module 1005, a denoising module 1006, an updating module 1007 and a transmitting module 1008;
the receiving module 1005 is configured to receive M × N decomposition results uploaded by one or more intelligent terminals;
the denoising module 1006 is configured to perform denoising processing on each of the continuous numerical attributes according to the M × N decomposition results, to obtain a denoising result of each of the continuous numerical attributes;
the updating module 1007 is configured to update the high-precision map according to the denoising result;
the sending module 1008 is configured to send the updated high-precision map to the intelligent terminal.
The intelligent terminal can be a mobile phone or a vehicle machine.
Each module in the intelligent terminal may be configured to execute each step corresponding to the vehicle-side of the crowdsourcing equipment in the embodiment shown in fig. 4, and correspondingly, each module in the cloud server may be configured to execute each step corresponding to the cloud server-side in the embodiment shown in fig. 4.
In the embodiments provided in the present invention, it should be understood that the disclosed apparatus and method may be implemented in other ways. For example, the above-described apparatus embodiments are merely illustrative, and for example, the division of the units is only one logical division, and other divisions may be realized in practice, for example, a plurality of units or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some interfaces, devices or units, and may be in an electrical, mechanical or other form.
The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one position, or may be distributed on multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.
In addition, functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit may be implemented in the form of hardware, or in the form of hardware plus a software functional unit.
The integrated unit implemented in the form of a software functional unit may be stored in a computer readable storage medium. The software functional unit is stored in a storage medium and includes several instructions to enable a computer device (which may be a personal computer, a server, or a network device) or a processor (processor) to execute some steps of the methods according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a portable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other media capable of storing program codes.
It should be understood that the Processor may be a Central Processing Unit (CPU), other general purpose Processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), etc. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like. The steps of a method disclosed in this application may be directly implemented by a hardware processor, or implemented by a combination of hardware and software modules in a processor.
Finally, it should be noted that: the above embodiments are only used to illustrate the technical solution of the present invention, and not to limit the same; while the invention has been described in detail and with reference to the foregoing embodiments, it will be understood by those skilled in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some or all of the technical features may be equivalently replaced; and these modifications or substitutions do not depart from the spirit of the corresponding technical solutions of the embodiments of the present invention.
Claims (14)
1. A high-precision map updating method is characterized by comprising the following steps:
obtaining M x N decomposition results of M continuous numerical attributes of the same map element, wherein N represents the collection times of crowdsourcing data of the same map element, M is more than or equal to 1, and N is more than or equal to 2; the continuous numeric attribute comprises a location attribute;
the obtaining of M × N decomposition results of M kinds of continuous numerical attributes of the same map element includes: obtaining M x N classification data of the M continuous numerical attributes of the same map element, and decomposing the M x N classification data to obtain M x N decomposition results;
when the continuous numerical type attribute is a position attribute, obtaining orthogonal model coordinates of N classification data of the position attribute, wherein the orthogonal model coordinates of the N classification data of the position attribute are N decomposition results of the position attribute;
according to the M x N decomposition results, denoising each continuous numerical type attribute respectively to obtain a denoising result of each continuous numerical type attribute;
and updating the high-precision map according to the denoising result.
2. The method according to claim 1, wherein said obtaining M x N classification data of said M consecutive numerical attributes of said same map element comprises:
receiving N times of crowdsourcing data of the same map element;
and classifying the crowdsourcing data every time to obtain M x N classification data of the M continuous numerical attributes of the same map element.
3. The method according to claim 1, wherein the obtaining M × N decomposition results of M kinds of continuous numerical attributes of the same map element further comprises:
receiving M x N decomposition results uploaded by a crowdsourcing vehicle, wherein the M x N decomposition results are obtained after the crowdsourcing vehicle classifies and decomposes crowdsourcing data acquired each time.
4. The method of claim 2, wherein the continuous numeric attribute further comprises: at least one of a shape attribute and a pose attribute.
5. The method of claim 1, wherein said obtaining orthogonal model coordinates of the N classification data of the location attribute comprises:
determining the average central point of the same map element under a longitude and latitude coordinate system according to the N classification data of the position attribute;
converting the N classified data into an ENU (north east Asia) coordinate system with the average central point as an origin to obtain ENU coordinates of the N classified data;
determining the road direction of the road corresponding to the average center point;
and determining the orthogonal model coordinates of the N classification data according to the ENU coordinates of the N classification data and the road direction.
6. The method of claim 5, wherein said determining orthomode coordinates of the N classification data from the ENU coordinates of the N classification data and the road direction comprises:
according to the road direction and the due north direction, determining the angle of the due north direction rotating clockwise to the road direction;
and clockwise rotating the ENU coordinates of the N classified data by the angle to obtain orthogonal model coordinates of the N classified data.
7. The method according to claim 4, wherein when the continuous numerical attribute is a shape attribute, the decomposing the M x N classification data to obtain the M x N decomposition results includes:
sorting the corner points in each classification data of the shape attribute to obtain the serial number of each corner point;
and determining the modeling expression corresponding to each classification data according to the serial number of each corner point, wherein the modeling expressions corresponding to the N classification data of the shape attribute are N decomposition results of the shape attribute.
8. The method according to claim 4, wherein when the continuous numerical attribute is an attitude attribute, said decomposing the M x N classification data to obtain the M x N decomposition results includes:
and aiming at each classification data of the attitude attribute, adopting quaternion attitude calculation or rotation matrix attitude calculation to convert the corresponding classification data into the modeling expression of the attitude attribute, wherein the modeling expression corresponding to the N classification data of the attitude attribute is N decomposition results of the attitude attribute.
9. The method according to any one of claims 1 to 8, wherein the performing denoising processing on each of the continuous numerical attributes according to the Mx N decomposition results to obtain a denoising result of each of the continuous numerical attributes respectively comprises:
calculating the mean and standard deviation of the N decomposition results of each of the continuous numerical attributes;
determining a confidence interval according to a preset confidence coefficient, the mean value and the standard deviation;
filtering the N decomposition results according to the confidence interval to obtain filtering results;
taking the average value of the filtering results;
and taking the mean value as a denoising result corresponding to the continuous numerical attribute.
10. The method of claim 1, wherein the updating the high-precision map according to the denoising result comprises:
restoring the continuous numerical type attribute according to the denoising result to obtain a restoration result;
and updating the high-precision map according to the reduction result.
11. A high-precision map updating apparatus, comprising:
the map feature extraction module is used for extracting M x N decomposition results of M kinds of continuous numerical attributes of the same map element, wherein N represents the collection times of crowdsourcing data of the same map element, M is greater than or equal to 1, and N is greater than or equal to 2; the continuous numerical attribute comprises a location attribute;
the obtaining module is specifically configured to obtain M x N classification data of the M continuous numerical attributes of the same map element, and decompose the M x N classification data to obtain M x N decomposition results; when the continuous numerical type attribute is a position attribute, obtaining orthogonal model coordinates of N classification data of the position attribute, wherein the orthogonal model coordinates of the N classification data of the position attribute are N decomposition results of the position attribute;
the denoising module is used for respectively denoising each continuous numerical type attribute according to the M x N decomposition results to obtain a denoising result of each continuous numerical type attribute;
and the updating module is used for updating the high-precision map according to the denoising result.
12. A computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out the method of any one of claims 1 to 10.
13. A cloud server, comprising:
the receiving module is used for receiving N times of crowdsourcing data of the same map element;
the classification module is used for classifying crowdsourcing data each time to obtain M x N classification data of M continuous numerical attributes of the same map element; the continuous numeric attribute comprises a location attribute;
the decomposition module is used for decomposing the M x N classified data to obtain M x N decomposition results; when the continuous numerical type attribute is a position attribute, obtaining orthogonal model coordinates of N classification data of the position attribute, wherein the orthogonal model coordinates of the N classification data of the position attribute are N decomposition results of the position attribute;
the denoising module is used for respectively denoising each continuous numerical type attribute according to the M x N decomposition results to obtain a denoising result of each continuous numerical type attribute;
and the updating module is used for updating the high-precision map according to the denoising result.
14. A map updating system, comprising: the system comprises an intelligent terminal and a cloud server;
wherein, intelligent terminal includes: the device comprises an acquisition module, a classification module, a decomposition module and an uploading module;
the acquisition module is used for acquiring crowdsourcing data of the same map element;
the classification module is used for classifying the crowdsourcing data to obtain classification data of M continuous numerical attributes of the same map element; the continuous numerical attribute comprises a location attribute;
the decomposition module is used for decomposing the classified data to obtain a decomposition result; when the continuous numerical type attribute is a position attribute, obtaining an orthogonal model coordinate of classification data of the position attribute, wherein the orthogonal model coordinate of the classification data of the position attribute is a decomposition result of the position attribute;
the uploading module is used for uploading the decomposition result to the cloud server;
wherein the cloud server comprises: the device comprises a receiving module, a denoising module, an updating module and a sending module;
the receiving module is used for receiving M x N decomposition results uploaded by one or more intelligent terminals;
the denoising module is used for respectively denoising each continuous numerical type attribute according to the M x N decomposition results to obtain a denoising result of each continuous numerical type attribute;
the updating module is used for updating the high-precision map according to the denoising result;
and the sending module is used for sending the updated high-precision map to the intelligent terminal.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910650809.1A CN112241414B (en) | 2019-07-18 | 2019-07-18 | High-precision map updating method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910650809.1A CN112241414B (en) | 2019-07-18 | 2019-07-18 | High-precision map updating method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112241414A CN112241414A (en) | 2021-01-19 |
| CN112241414B true CN112241414B (en) | 2022-10-21 |
Family
ID=74168361
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910650809.1A Active CN112241414B (en) | 2019-07-18 | 2019-07-18 | High-precision map updating method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112241414B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101645078A (en) * | 2009-09-11 | 2010-02-10 | 中南大学 | System and method for integration of change information collection and spatiotemporal data update] |
| CN107145578A (en) * | 2017-05-08 | 2017-09-08 | 深圳地平线机器人科技有限公司 | Map constructing method, device, equipment and system |
| CN108007453A (en) * | 2017-12-11 | 2018-05-08 | 北京奇虎科技有限公司 | Map updating method, device and electronic equipment based on a cloud |
| CN108709557A (en) * | 2018-05-18 | 2018-10-26 | 东南大学 | Indoor map generation method based on multi-user's track fitting |
| CN109167818A (en) * | 2018-08-06 | 2019-01-08 | 长安大学 | Road evenness detection system based on smart phone crowdsourcing acquisition |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| DE102016217654A1 (en) * | 2016-09-15 | 2018-03-15 | Bayerische Motoren Werke Aktiengesellschaft | Method and data processing system for generating map data of a digital map |
-
2019
- 2019-07-18 CN CN201910650809.1A patent/CN112241414B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101645078A (en) * | 2009-09-11 | 2010-02-10 | 中南大学 | System and method for integration of change information collection and spatiotemporal data update] |
| CN107145578A (en) * | 2017-05-08 | 2017-09-08 | 深圳地平线机器人科技有限公司 | Map constructing method, device, equipment and system |
| CN108007453A (en) * | 2017-12-11 | 2018-05-08 | 北京奇虎科技有限公司 | Map updating method, device and electronic equipment based on a cloud |
| CN108709557A (en) * | 2018-05-18 | 2018-10-26 | 东南大学 | Indoor map generation method based on multi-user's track fitting |
| CN109167818A (en) * | 2018-08-06 | 2019-01-08 | 长安大学 | Road evenness detection system based on smart phone crowdsourcing acquisition |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112241414A (en) | 2021-01-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110119438B (en) | Airborne LiDAR point cloud filtering method based on active learning | |
| CN110471058A (en) | The system and method detected automatically for trailer attribute | |
| CN109903330B (en) | Method and device for processing data | |
| EP3968266B1 (en) | Obstacle three-dimensional position acquisition method and apparatus for roadside computing device | |
| CN111324145B (en) | Unmanned aerial vehicle autonomous landing method, device, equipment and storage medium | |
| CN110887493B (en) | Track calculation method, medium, terminal and device based on local map matching | |
| CN112284400B (en) | Vehicle positioning method and device, electronic equipment and computer readable storage medium | |
| CN112348921A (en) | Mapping method and system based on visual semantic point cloud | |
| US20220270323A1 (en) | Computer Vision Systems and Methods for Supplying Missing Point Data in Point Clouds Derived from Stereoscopic Image Pairs | |
| CN114724117A (en) | Lane line key point data generation method and device, electronic equipment and storage medium | |
| CN112651266B (en) | Pedestrian detection method and device | |
| CN112164138A (en) | Point cloud data screening method and device | |
| CN114677435A (en) | Point cloud panoramic fusion element extraction method and system | |
| CN111627001A (en) | Image detection method and device | |
| CN114612616A (en) | Mapping method and device, electronic equipment and storage medium | |
| CN111982132B (en) | Data processing method, device and storage medium | |
| CN112241414B (en) | High-precision map updating method and device | |
| CN116309943B (en) | Parking lot semantic map road network construction method and device and electronic equipment | |
| US20210134034A1 (en) | Region clipping method and recording medium storing region clipping program | |
| CN111860084B (en) | Image feature matching and positioning method and device and positioning system | |
| CN112732860A (en) | Road extraction method, device, readable storage medium and equipment | |
| CN116823966A (en) | Internal reference calibration method and device for camera, computer equipment and storage medium | |
| KR20220062709A (en) | System for detecting disaster situation by clustering of spatial information based an image of a mobile device and method therefor | |
| CN109657556A (en) | The classification method and system of road and its periphery atural object | |
| CN115661313A (en) | Point cloud map generation method, point cloud map generation device and storage medium |
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 |