CN115062076B - Automatic driving full-quantization data acquisition method and system - Google Patents
Automatic driving full-quantization data acquisition method and system Download PDFInfo
- Publication number
- CN115062076B CN115062076B CN202210766765.0A CN202210766765A CN115062076B CN 115062076 B CN115062076 B CN 115062076B CN 202210766765 A CN202210766765 A CN 202210766765A CN 115062076 B CN115062076 B CN 115062076B
- Authority
- CN
- China
- Prior art keywords
- data
- vehicle
- platform
- original data
- end platform
- 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
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
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2458—Special types of queries, e.g. statistical queries, fuzzy queries or distributed queries
- G06F16/2465—Query processing support for facilitating data mining operations in structured databases
-
- 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
- G06F16/21—Design, administration or maintenance of databases
- G06F16/215—Improving data quality; Data cleansing, e.g. de-duplication, removing invalid entries or correcting typographical errors
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Databases & Information Systems (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Computational Linguistics (AREA)
- Evolutionary Computation (AREA)
- Biophysics (AREA)
- Computing Systems (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- Fuzzy Systems (AREA)
- Probability & Statistics with Applications (AREA)
- Quality & Reliability (AREA)
- Traffic Control Systems (AREA)
- Radar Systems Or Details Thereof (AREA)
Abstract
The invention discloses an automatic driving full-quantization data acquisition method and system, comprising signal acquisition equipment, a vehicle end platform, a cloud platform and a local research and development platform; the signal acquisition equipment is used for acquiring original data; the vehicle end platform is used for receiving the vehicle body data and the road perception data acquired by the signal acquisition equipment; the cloud platform is used for carrying out data screening on the original data; the local research and development platform is used for providing the vehicle-end platform with original data acquisition strategies under different conditions and providing the cloud-end platform with different data screening models aiming at different types of original data. The invention can fully quantify the data acquisition of the automatic driving vehicle, comprises the acquisition of perception data and vehicle signals, adopts different acquisition methods aiming at different parts of the vehicle, and also carries out linkage acquisition of the vehicle end at the position through a local research and development platform, a cloud platform and a vehicle end platform and transmits the acquired original data to the cloud platform, and the cloud platform can carry out data screening aiming at actual needs.
Description
Technical Field
The invention relates to the field of automatic driving data processing, in particular to an automatic driving full-quantization data acquisition method and system.
Background
The vehicle data acquisition is an important ring in automatic driving research and development, and the acquired data mainly comprises image data, laser radar data, millimeter wave radar data, ultrasonic radar data and whole vehicle motion parameter data, and the data can provide data input samples for algorithm engineers, train and output effective perception results by applying a model, and are applied to automatic driving research and development.
With the development of the automatic driving technology level, the requirements on vehicle data acquisition are higher and higher. In the face of a large number of data acquisition tasks, how to realize efficient and full-scale acquisition of vehicle data becomes a problem that a vehicle enterprise has to think at present. The vehicle-mounted terminal in the current stage continuously collects user data and vehicle data in the background in real time, and packages and transmits the collected data to the background server in real time. The collection of a large number of data collection items and high frequency will generate a large amount of redundant data information, which will increase the load of the CPU and the communication bandwidth, affect the vehicle performance, and in addition, the data collected by the vehicle-mounted terminal will also increase the flow consumption when transmitted to the background server.
The traditional automatic driving test vehicle collects automatic driving related data by using a camera, a laser radar and other sensors, acquires original data of the automatic driving test vehicle and analyzes and processes the original data, and the method aims at some parts of the vehicle and cannot collect the original data, so that the collected vehicle data is not comprehensive enough.
Disclosure of Invention
In order to solve the technical problems, the invention provides a full-quantization data acquisition method and system for automatic driving.
The aim of the invention is achieved by the following technical scheme:
an automatic driving full-quantization data acquisition system comprises signal acquisition equipment, a vehicle-end platform, a cloud platform and a local research and development platform;
the signal acquisition equipment is used for acquiring original data, wherein the original data comprises vehicle body data and road perception data; the signal acquisition equipment comprises a camera, a laser radar, a millimeter wave radar and an ultrasonic radar; the vehicle body data comprises gear information, chassis information and tire information;
the vehicle end platform is used for receiving the vehicle body data and the road perception data acquired by the signal acquisition equipment, controlling the opening of different signal acquisition equipment according to the vehicle body data and the road perception data, and obtaining finished original data under the condition that the least signal acquisition equipment is opened;
the cloud platform is used for receiving the original data forwarded by the vehicle-end platform and carrying out data screening on the original data;
the local research and development platform is used for providing the vehicle-end platform with original data acquisition strategies under different conditions and providing the cloud-end platform with different data screening models aiming at different types of original data.
Further improved, the data screening method comprises the following steps:
1) For duplicate data: performing data de-duplication, slicing and de-framing
2) For the case of lens fouling of the camera: performing image quality analysis, deleting unqualified images, and supplementing radar information of the deleted images;
3) And performing special data mining by using the deployed data screening model and the deployed data acquisition strategy.
The method for carrying out the slicing processing of the data is further improved by adopting a slicing algorithm to extract continuous video frames in a step mode, so that the data redundancy is avoided; the duplicate removal method is to remove data with high similarity.
Further improvements include camera data screening models, laser radar data screening models, millimeter wave radar data screening and ultrasonic radar data screening models, and data comprehensive processing models; the camera data screening model, the laser radar data screening model, the millimeter wave radar data screening model and the ultrasonic radar data screening model are trained neural network models, and data screening is carried out on the camera data, the laser radar data, the millimeter wave radar data and the ultrasonic radar data respectively; the data comprehensive processing model is also a trained neural network model and is used for comprehensively obtaining the vehicle condition and the road condition by the output of the camera data screening model, the laser radar data screening model, the millimeter wave radar data screening model and the ultrasonic radar data screening model so as to provide a driving strategy in real time.
An automatic driving full-quantization data acquisition method comprises the following steps:
step one, providing original data acquisition strategies under different conditions for a vehicle-end platform by a local research and development platform, and providing different data screening models for different types of original data for a cloud-end platform;
step two, the vehicle-end platform collects the original data collected by the signal collection equipment, and adjusts the original data collection strategy according to the collected original data in real time so as to obtain finished original data under the condition that the minimum signal collection equipment is started;
step three, the vehicle-end platform transmits the collected original data to the cloud-end platform, the cloud-end platform performs data screening processing by adopting data screening models of different points aiming at different data, and then the processing results are transmitted to the data comprehensive processing model to provide an automatic driving strategy in real time.
Compared with the prior art, the invention has the following beneficial effects:
1. the full-quantization automatic driving vehicle data acquisition is realized, including perception data and vehicle signal acquisition.
2. Different acquisition methods are adopted for different parts of the vehicle.
3. And carrying out linkage acquisition processing through a local research and development platform, a cloud platform and a vehicle-end platform.
4. The original data acquired by the vehicle end are transmitted to the cloud platform, and the cloud platform can conduct data screening aiming at actual requirements.
Drawings
The invention is further illustrated by the accompanying drawings, the content of which does not constitute any limitation of the invention.
FIG. 1 is a schematic diagram of a system architecture of the present invention;
FIG. 2 is a schematic flow chart of the method of the present invention.
Detailed Description
The invention will be further described in detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the invention more apparent.
A method and a device for collecting full-quantization data of automatic driving are implemented as follows:
sensing data and vehicle signal acquisition: the vehicle-end platform collects by using equipment, and adopts different measures aiming at different situations to obtain complete information data of the vehicle.
1) The device comprises: the sensor comprises a camera, a laser radar, a millimeter wave radar, an ultrasonic radar and the like.
2) And (3) acquiring signals of a vehicle body (gear, chassis, tires and the like).
3) And storing the acquired vehicle body data and road perception data.
Data screening: and sending the original data acquired by the vehicle to a cloud platform, and carrying out data screening by the cloud platform aiming at different data.
1) A large number of duplicate data: performing data de-duplication, slicing, de-framing and the like, reducing data redundancy, improving data quality, and performing step extraction on continuous video frames by a slicing algorithm to avoid the data redundancy; de-duplication is to remove data with high similarity
2) Lens dirt of camera etc. condition: and carrying out image quality analysis, and particularly adopting a self-grinding deep learning classification model.
3) And (3) data mining: and performing special data mining by using the deployed algorithm model and the acquisition strategy.
And (3) data storage: and compressing the processed data and remotely storing the compressed data into a memory of the cloud platform.
And (3) data processing: and acquiring screened data, and selecting corresponding models for training aiming at different data.
Data application: and applying the trained model to a vehicle-end platform for data acquisition, thereby improving acquisition precision and efficiency.
The local research and development platform can provide an algorithm model for the cloud platform so as to facilitate improving the training capacity of the cloud platform, and can also provide an acquisition strategy for the vehicle-end platform, so that the acquisition process is more efficient, and the vehicle performance is improved.
Finally, it should be noted that the above embodiments are only for illustrating the technical solution of the present invention and not for limiting the scope of the present invention, and although the present invention has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that modifications or equivalent substitutions can be made to the technical solution of the present invention without departing from the spirit and scope of the technical solution of the present invention.
Claims (2)
1. The automatic driving full-quantization data acquisition system is characterized by comprising signal acquisition equipment, a vehicle end platform, a cloud platform and a local research and development platform;
the signal acquisition equipment is used for acquiring original data, wherein the original data comprises vehicle body data and road perception data; the signal acquisition equipment comprises a camera, a laser radar, a millimeter wave radar and an ultrasonic radar; the vehicle body data comprises gear information, chassis information and tire information;
the vehicle end platform is used for receiving the vehicle body data and the road perception data acquired by the signal acquisition equipment, controlling the opening of different signal acquisition equipment according to the vehicle body data and the road perception data, and obtaining complete original data under the condition that the least signal acquisition equipment is opened;
the cloud platform is used for receiving the original data forwarded by the vehicle-end platform and carrying out data screening on the original data;
the local research and development platform is used for providing original data acquisition strategies under different conditions for the vehicle-end platform and providing different data screening models for the cloud-end platform aiming at different types of original data; the data screening method comprises the following steps:
1) For duplicate data: performing data de-duplication, slicing and de-framing
2) For the case of lens fouling of the camera: performing image quality analysis, deleting unqualified images, and supplementing radar information of the deleted images;
3) Performing special data mining by using the deployed data screening model and the deployed data acquisition strategy; the method for carrying out the slicing processing of the data is to extract continuous video frames in a step mode by adopting a slicing algorithm, so that the data redundancy is avoided; the duplicate removal method is to remove data with high similarity; the data screening model comprises a camera data screening model, a laser radar data screening model, a millimeter wave radar data screening model, an ultrasonic radar data screening model and a data comprehensive processing model; the camera data screening model, the laser radar data screening model, the millimeter wave radar data screening model and the ultrasonic radar data screening model are trained neural network models, and data screening is carried out on the camera data, the laser radar data, the millimeter wave radar data and the ultrasonic radar data respectively; the data comprehensive processing model is also a trained neural network model and is used for comprehensively obtaining the vehicle condition and the road condition by the output of the camera data screening model, the laser radar data screening model, the millimeter wave radar data screening model and the ultrasonic radar data screening model so as to provide a driving strategy in real time.
2. An automatic driving full-quantization data acquisition method using the automatic driving full-quantization data acquisition system according to claim 1, comprising the steps of:
step one, providing original data acquisition strategies under different conditions for a vehicle-end platform by a local research and development platform, and providing different data screening models for different types of original data for a cloud-end platform;
step two, the vehicle-end platform collects the original data collected by the signal collection equipment, and adjusts the original data collection strategy according to the collected original data in real time so as to obtain finished original data under the condition that the minimum signal collection equipment is started;
step three, the vehicle-end platform transmits the collected original data to the cloud-end platform, the cloud-end platform performs data screening processing by adopting data screening models of different points aiming at different data, and then the processing results are transmitted to the data comprehensive processing model to provide an automatic driving strategy in real time.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210766765.0A CN115062076B (en) | 2022-07-01 | 2022-07-01 | Automatic driving full-quantization data acquisition method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210766765.0A CN115062076B (en) | 2022-07-01 | 2022-07-01 | Automatic driving full-quantization data acquisition method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115062076A CN115062076A (en) | 2022-09-16 |
| CN115062076B true CN115062076B (en) | 2023-08-29 |
Family
ID=83204109
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210766765.0A Active CN115062076B (en) | 2022-07-01 | 2022-07-01 | Automatic driving full-quantization data acquisition method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115062076B (en) |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107846576B (en) * | 2017-09-30 | 2019-12-10 | 北京大学 | Method and system for encoding and decoding visual characteristic data |
| CN110395262A (en) * | 2019-08-07 | 2019-11-01 | 安徽江淮汽车集团股份有限公司 | Driving behavior data collection system and method |
| CN114117133A (en) * | 2021-10-11 | 2022-03-01 | 云度新能源汽车有限公司 | Automatic driving data collection method and system |
-
2022
- 2022-07-01 CN CN202210766765.0A patent/CN115062076B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN115062076A (en) | 2022-09-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108109385B (en) | System and method for identifying and judging dangerous behaviors of power transmission line anti-external damage vehicle | |
| CN110263706B (en) | Method for detecting and identifying dynamic target of vehicle-mounted video in haze weather | |
| CN111340151B (en) | Weather phenomenon recognition system and method for assisting automatic driving of vehicle | |
| WO2022241784A1 (en) | Defect detection method and apparatus, storage medium, and electronic device | |
| CN112633120B (en) | Model training method of intelligent roadside sensing system based on semi-supervised learning | |
| CN110763685A (en) | Artificial intelligent detection method and device for DFB semiconductor laser chip surface defects | |
| CN113436184B (en) | Power equipment image defect discriminating method and system based on improved twin network | |
| CN113159166A (en) | Embedded image identification detection method, system, medium and equipment based on edge calculation | |
| CN112417973A (en) | Unmanned system based on car networking | |
| CN110599459A (en) | Underground pipe network risk assessment cloud system based on deep learning | |
| CN110599458A (en) | Underground pipe network detection and evaluation cloud system based on convolutional neural network | |
| CN115062076B (en) | Automatic driving full-quantization data acquisition method and system | |
| CN111753885A (en) | Privacy enhanced data processing method and system based on deep learning | |
| CN115452376A (en) | Bearing fault diagnosis method based on improved lightweight deep convolution neural network | |
| CN115130519A (en) | Ship structure fault prediction method using convolutional neural network | |
| CN115546191A (en) | Insulator defect detection method and device based on improved RetinaNet | |
| CN115240090A (en) | Unmanned aerial vehicle target detection system based on edge calculation | |
| CN115439954A (en) | Data closed-loop method based on cloud large model | |
| CN111488889B (en) | Intelligent image processor for extracting image edges | |
| CN114353880A (en) | Strain insulator string wind-induced vibration online monitoring system and method | |
| CN114792375A (en) | Terrain classification method based on audio-visual information fusion | |
| CN114120159A (en) | Method and device for detecting pin defects of power transmission line | |
| CN115101092B (en) | Btpnet 21-based classification method for automatic classification model of construction environment sound | |
| CN203876655U (en) | Vehicle-mounted infrared night pedestrian detecting system | |
| CN113627269B (en) | Pest target detection method based on decoupling classification and regression feature optimal layer technology |
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 |