CN-116467483-B - Automatic driving scene mining method, device, equipment and storage medium

Abstract

The disclosure provides an automatic driving scene mining method, device, equipment and storage medium, and relates to the technical field of data processing, in particular to the technical fields of automatic driving, cloud computing, deep learning and the like. The method comprises the steps of determining automatic driving data to be mined and a target automatic driving scene according to a received mining task, determining at least a tag variable and a logic variable in a preset mining architecture according to the target automatic driving scene, generating mining logic based on the preset mining architecture according to the tag variable and the logic variable, and mining target data fragments corresponding to the target automatic driving scene from the plurality of data fragments according to the mining logic. According to the technology disclosed by the invention, the mining logic can be conveniently and rapidly generated based on the target automatic driving scene, and the target data fragment corresponding to the target automatic driving scene can be accurately mined from the automatic driving data based on the mining logic.

Inventors

• ZHU JIANHUA

Assignees

• 北京百度网讯科技有限公司

Dates

Publication Date

20260512

Application Date

20221219

Claims (20)

1. 1. An automatic driving scene mining method, comprising: Determining automatic driving data to be mined and a target automatic driving scene according to the received mining task, wherein the automatic driving data comprises a plurality of data fragments with labels; determining at least a tag variable, a logic variable and a constraint variable in a preset mining architecture according to the target automatic driving scene, wherein the constraint variable is used for limiting constraint conditions required to be met for mining based on the plurality of data fragments; generating mining logic based on the preset mining architecture according to the tag variable, the logic variable and the constraint variable, and According to the mining logic, mining target data fragments corresponding to the target automatic driving scene from the plurality of data fragments; the mining, according to the mining logic, a target data segment corresponding to the target autopilot scene from the plurality of data segments, including: According to the tag variable of the mining logic, acquiring candidate data fragments corresponding to the tag variable from the plurality of data fragments; according to the time period logic variable in the logic variables, determining a first data segment conforming to the time period logic variable based on the time period information of the candidate data segment; Determining a target data segment conforming to the comparison operation logic variable based on the time period information of the first data segment according to the comparison operation logic variable in the logic variables; Or alternatively The mining, according to the mining logic, a target data segment corresponding to the target autopilot scene from the plurality of data segments, including: acquiring a first candidate data segment corresponding to the tag variable from the plurality of data segments according to the tag variable of the mining logic; Determining a second candidate data segment conforming to the constraint variable based on the first candidate data segment according to the constraint variable of the mining logic; And determining a target data segment conforming to the logic variable based on the second candidate data segment according to the logic variable of the mining logic.
2. 2. The method of claim 1, determining tag variables and logical variables in a preset mining architecture, comprising: determining a tag variable in a preset mining architecture according to scene objects and scene object behaviors in the target automatic driving scene; And determining logic variables in the preset mining architecture according to the behavior relation in the target automatic driving scene.
3. 3. The method of claim 1, determining constraint variables in a preset mining architecture, comprising: and determining constraint variables in the preset excavation framework according to the road limiting conditions and the behavior limiting conditions in the target automatic driving scene.
4. 4. The method of claim 1, wherein the determining, based on the second candidate data segment, the target data segment that meets the logical variable according to the logical variable of the mining logic comprises: according to the time period logic variable in the logic variables, determining a first data segment conforming to the time period logic variable based on the time period information of the second candidate data segment; and determining a target data segment conforming to the comparison operation logic variable based on the time period information of the first data segment according to the comparison operation logic variable in the logic variables.
5. 5. The method of any one of claims 1 to 4, further comprising: Determining motion parameter information corresponding to the target data segment according to the automatic driving data and the time period information of the target data segment; and returning the time period information of the target data segment and the motion parameter information.
6. 6. The method according to any one of claims 1 to 4, wherein the determining of the autopilot data and the target autopilot scenario to be mined from the received mining task, is preceded by: dividing the automatic driving data to be mined into a plurality of data segments; And determining the label of each data segment in the plurality of data segments according to the preset label dimension.
7. 7. The method of claim 6, wherein the preset tag dimensions comprise at least one of a road network topology dimension, a host vehicle behavior dimension, an obstacle behavior dimension, and an interaction behavior dimension.
8. 8. The method of claim 6, wherein the dividing the autopilot data to be mined into a plurality of data segments comprises: determining main vehicle data and obstacle data from automatic driving data to be mined; Dividing the main vehicle data into a plurality of first data fragments according to driving behavior data in the main vehicle data; Dividing the obstacle data into a plurality of second data pieces according to obstacle behavior data in the obstacle data.
9. 9. The method of claim 8, wherein the determining the tag for each of the plurality of data segments according to a preset tag dimension comprises: determining the main vehicle behavior labels of the first data fragments according to the driving behavior data; Determining obstacle behavior tags of the plurality of second data segments according to the obstacle behavior data; determining road network topology labels of the first data fragments and the second data fragments according to road data in the automatic driving data; and determining interaction behavior labels of the plurality of first data fragments and the plurality of second data fragments according to the interaction data in the automatic driving data.
10. 10. An excavating device of an autopilot scene, comprising: The system comprises a first determining module, a second determining module and a third determining module, wherein the first determining module is used for determining automatic driving data to be mined and a target automatic driving scene according to a received mining task, and the automatic driving data comprises a plurality of data fragments with labels; The second determining module is used for determining at least a tag variable, a logic variable and a constraint variable in a preset mining architecture according to the target automatic driving scene, wherein the constraint variable is used for limiting constraint conditions required to be met for mining based on the plurality of data fragments; A generation module for generating mining logic based on the preset mining architecture according to the tag variable, the logic variable and the constraint variable, and The mining module is used for mining target data fragments corresponding to the target automatic driving scene from the plurality of data fragments according to the mining logic; wherein, the excavation module includes: The first acquisition submodule is used for acquiring candidate data fragments corresponding to the tag variable from the plurality of data fragments according to the tag variable of the mining logic; A fourth determining sub-module, configured to determine, according to a time period logic variable in the logic variables, a first data segment that conforms to the time period logic variable based on time period information of the candidate data segment; Or alternatively Wherein, the excavation module includes: The second obtaining submodule is used for obtaining a first candidate data segment corresponding to the tag variable from the plurality of data segments according to the tag variable of the mining logic; A fifth determination submodule, configured to determine, based on the first candidate data segment, a second candidate data segment that conforms to the constraint variable of the mining logic; And a sixth determining submodule, configured to determine, according to the logic variable of the mining logic, a target data segment according to the logic variable based on the second candidate data segment.
11. 11. The apparatus of claim 10, wherein the second determination module comprises: The first determining submodule is used for determining a tag variable in a preset mining architecture according to scene objects and scene object behaviors in the target automatic driving scene; and the second determining submodule is used for determining logic variables in the preset mining architecture according to the behavior relation in the target automatic driving scene.
12. 12. The apparatus of claim 11, wherein the second determination module further comprises: and the third determining submodule is used for determining constraint variables in the preset excavation framework according to the road limiting conditions and the behavior limiting conditions in the target automatic driving scene.
13. 13. The apparatus of claim 10, wherein the sixth determination submodule is to: according to the time period logic variable in the logic variables, determining a first data segment conforming to the time period logic variable based on the time period information of the second candidate data segment; and determining a target data segment conforming to the comparison operation logic variable based on the time period information of the first data segment according to the comparison operation logic variable in the logic variables.
14. 14. The apparatus of any of claims 10 to 13, further comprising: the third determining module is used for determining motion parameter information corresponding to the target data segment according to the automatic driving data and the time period information of the target data segment; And the feedback module is used for carrying out feedback on the time period information of the target data segment and the motion parameter information.
15. 15. The apparatus of any of claims 10 to 13, further comprising: The division module is used for dividing the automatic driving data to be mined into a plurality of data fragments; And a fourth determining module, configured to determine a tag of each data segment of the plurality of data segments according to a preset tag dimension.
16. 16. The apparatus of claim 15, wherein the preset tag dimensions comprise at least one of a road network topology dimension, a host vehicle behavior dimension, an obstacle behavior dimension, and an interaction behavior dimension.
17. 17. The apparatus of claim 15, wherein the partitioning module is to: determining main vehicle data and obstacle data from automatic driving data to be mined; Dividing the main vehicle data into a plurality of first data fragments according to driving behavior data in the main vehicle data; Dividing the obstacle data into a plurality of second data pieces according to obstacle behavior data in the obstacle data.
18. 18. The apparatus of claim 17, wherein the fourth determination module is configured to: determining the main vehicle behavior labels of the first data fragments according to the driving behavior data; Determining obstacle behavior tags of the plurality of second data segments according to the obstacle behavior data; determining road network topology labels of the first data fragments and the second data fragments according to road data in the automatic driving data; and determining interaction behavior labels of the plurality of first data fragments and the plurality of second data fragments according to the interaction data in the automatic driving data.
19. 19. An electronic device, comprising: at least one processor, and A memory communicatively coupled to the at least one processor, wherein, 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 to 9.
20. 20. A non-transitory computer readable storage medium storing computer instructions for causing the computer to perform the method of any one of claims 1 to 9.

Description

Automatic driving scene mining method, device, equipment and storage medium The application discloses a mining method, a device, equipment and a storage medium for an automatic driving scene, which are named as '202211630717.5', and a divisional application of Chinese cases with the application date of 2022, 12 months and 19 days. Technical Field The disclosure relates to the technical field of data processing, in particular to the technical fields of automatic driving, cloud computing, deep learning and the like. Background The concept of upgrading iteration of the automatic driving system driven by the automatic driving data is a widely accepted scheme in the industry. How to extract high value data from massive autopilot data has been a challenge facing the industry. Disclosure of Invention The disclosure provides an automatic driving scene mining method, device and equipment and a storage medium. According to an aspect of the present disclosure, there is provided an excavating method of an autopilot scene, including: Determining automatic driving data to be mined and a target automatic driving scene according to the received mining task, wherein the automatic driving data comprises a plurality of data fragments with labels; according to the target automatic driving scene, at least determining a tag variable and a logic variable in a preset mining architecture; Generating mining logic based on a preset mining architecture according to the tag variable and the logic variable, and And mining target data fragments corresponding to the target automatic driving scene from the plurality of data fragments according to mining logic. According to another aspect of the present disclosure, there is provided an excavating apparatus of an autopilot scene, including: the automatic driving system comprises a first determining module, a second determining module and a third determining module, wherein the first determining module is used for determining automatic driving data to be mined and a target automatic driving scene according to a received mining task, and the automatic driving data comprises a plurality of data fragments with labels; the second determining module is used for determining at least a tag variable and a logic variable in a preset mining architecture according to the target automatic driving scene; the generation module is used for generating mining logic based on a preset mining architecture according to the tag variable and the logic variable, and And the mining module is used for mining target data fragments corresponding to the target automatic driving scene from the plurality of data fragments according to mining logic. According to another aspect of the present disclosure, there is provided an electronic device including: at least one processor, and A memory communicatively coupled to the at least one processor, wherein, 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 the embodiments of the present disclosure. According to another aspect of the present disclosure, there is provided a non-transitory computer-readable storage medium storing computer instructions for causing the computer to perform a method according to any one of the embodiments of the present disclosure. According to another aspect of the present disclosure, there is provided a computer program product comprising a computer program which, when executed by a processor, implements a method according to any of the embodiments of the present disclosure. According to the technology disclosed by the invention, the mining logic can be conveniently and rapidly generated based on the target automatic driving scene, and the target data fragment corresponding to the target automatic driving scene can be accurately mined from the automatic driving data based on the mining logic. It should be understood that the description in this section is not intended to identify key or critical features of the embodiments of the disclosure, nor is it intended to be used to limit the scope of the disclosure. Other features of the present disclosure will become apparent from the following specification. Drawings The drawings are for a better understanding of the present solution and are not to be construed as limiting the present disclosure. Wherein: FIG. 1 is a flow diagram of a method of mining an autopilot scenario in accordance with an embodiment of the present disclosure; FIG. 2 is a schematic illustration of an autopilot data partitioning data segment of an autopilot scenario mining method according to an embodiment of the present disclosure; FIG. 3 is a flow diagram of a method of mining an autopilot scenario in accordance with another embodiment of the present disclosure; FIG. 4 is a schematic diagram of preset tag dimensions of an mining method of an autopilot scenario in accordance with an embodiment of the present disclosure; FIG. 5 is a schematic