WO2019127085A1 - Processing method, processing apparatus, control device and cloud server - Google Patents

Processing method, processing apparatus, control device and cloud server Download PDF

Info

Publication number
WO2019127085A1
WO2019127085A1 PCT/CN2017/118946 CN2017118946W WO2019127085A1 WO 2019127085 A1 WO2019127085 A1 WO 2019127085A1 CN 2017118946 W CN2017118946 W CN 2017118946W WO 2019127085 A1 WO2019127085 A1 WO 2019127085A1
Authority
WO
WIPO (PCT)
Prior art keywords
vehicle
machine learning
learning model
image
parameters
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.)
Ceased
Application number
PCT/CN2017/118946
Other languages
French (fr)
Inventor
Wei Sun
Bo Hu
Zheng Zhang
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Volkswagen China Investment Co Ltd
Original Assignee
Volkswagen China Investment Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Volkswagen China Investment Co Ltd filed Critical Volkswagen China Investment Co Ltd
Priority to CN201780097995.XA priority Critical patent/CN111542834A/en
Priority to PCT/CN2017/118946 priority patent/WO2019127085A1/en
Publication of WO2019127085A1 publication Critical patent/WO2019127085A1/en
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/50—Context or environment of the image
    • G06V20/56—Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B60—VEHICLES IN GENERAL
    • B60H—ARRANGEMENTS OF HEATING, COOLING, VENTILATING OR OTHER AIR-TREATING DEVICES SPECIALLY ADAPTED FOR PASSENGER OR GOODS SPACES OF VEHICLES
    • B60H1/00—Heating, cooling or ventilating devices
    • B60H1/00642—Control systems or circuits; Control members or indication devices for heating, cooling or ventilating devices
    • B60H1/00735—Control systems or circuits characterised by their input, i.e. by the detection, measurement or calculation of particular conditions, e.g. signal treatment, dynamic models
    • B60H1/008—Control systems or circuits characterised by their input, i.e. by the detection, measurement or calculation of particular conditions, e.g. signal treatment, dynamic models the input being air quality
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B60—VEHICLES IN GENERAL
    • B60H—ARRANGEMENTS OF HEATING, COOLING, VENTILATING OR OTHER AIR-TREATING DEVICES SPECIALLY ADAPTED FOR PASSENGER OR GOODS SPACES OF VEHICLES
    • B60H1/00—Heating, cooling or ventilating devices
    • B60H1/00642—Control systems or circuits; Control members or indication devices for heating, cooling or ventilating devices
    • B60H1/00814—Control systems or circuits characterised by their output, for controlling particular components of the heating, cooling or ventilating installation
    • B60H1/00821—Control systems or circuits characterised by their output, for controlling particular components of the heating, cooling or ventilating installation the components being ventilating, air admitting or air distributing devices
    • B60H1/00835—Damper doors, e.g. position control
    • B60H1/00849—Damper doors, e.g. position control for selectively commanding the induction of outside or inside air
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B60—VEHICLES IN GENERAL
    • B60H—ARRANGEMENTS OF HEATING, COOLING, VENTILATING OR OTHER AIR-TREATING DEVICES SPECIALLY ADAPTED FOR PASSENGER OR GOODS SPACES OF VEHICLES
    • B60H1/00—Heating, cooling or ventilating devices
    • B60H1/00642—Control systems or circuits; Control members or indication devices for heating, cooling or ventilating devices
    • B60H1/00735—Control systems or circuits characterised by their input, i.e. by the detection, measurement or calculation of particular conditions, e.g. signal treatment, dynamic models
    • B60H1/00792—Arrangement of detectors

Definitions

  • the present invention relates to a field of a vehicle and particularly to a processing method, a processing apparatus, a control device and a cloud server.
  • a vehicle usually uses an air quality sensor installed in the vehicle to measure air quality around the vehicle, and when it is measured that the air quality around the vehicle is not good, air intake of the air condition system of the vehicle is switched into an inner loop air intake to prevent pollutants outside the vehicle from flowing into a cabin of the vehicle.
  • VOC Volatile Organic Compound
  • Embodiments of the present invention provide a processing method, a processing apparatus, a control device and a cloud server, which can improve air quality inside a vehicle.
  • a processing method comprises: acquiring an image including an environment ahead of a first vehicle; classifying the acquired image with a trained machine learning model, wherein the trained machine learning model is used to classify an image into a first type image indicating no air pollution or a second type image indicating air pollution; and preventing air pollutants outside the first vehicle from entering into the first vehicle when the acquired image is classified into the second type image.
  • a processing method comprises: re-training a trained machine learning model by using sample images including environments stored in a sample image database to obtain an updated machine learning model, wherein the trained machine learning model is used to classify an image into a first type image indicating no air pollution or a second type image indicating air pollution, and the sample images include first type images indicating no air pollution and second type images indicating air pollution; and transmitting the updated machine learning model to vehicles.
  • a processing apparatus comprises: an acquiring module for acquiring an image including an environment ahead of a first vehicle; a classifying module for classifying the acquired image with a trained machine learning model, wherein the trained machine learning model is used to classify an image into a first type image indicating no air pollution or a second type image indicating air pollution; and a switching module for preventing air pollutants outside the first vehicle from entering into the first vehicle when the acquired image is classified into the second type image.
  • a processing apparatus comprises: a re-training module for re-training a trained machine learning model by using sample images including environments stored in a sample image database to obtain an updated machine learning model, wherein the trained machine learning model is used to classify an image into a first type image indicating no air pollution or a second type image indicating air pollution, and the sample images include first type images indicating no air pollution and second type images indicating air pollution; and a transmitting module for transmitting the updated machine learning model to vehicles.
  • a cloud server comprises: a processor; and a memory for storing executable instructions which cause, when executed, the processor to execute operations included in the above method.
  • a computer readable medium has thereon executable instructions for causing, when executed, a machine to execute operations included in the above method.
  • the above embodiments of the present invention uses the image including an environment ahead of the first vehicle and the machine learning model to determine whether there is air pollution around the first vehicle, which is not necessary to have the pollutants outside the first vehicle to flow into the cabin of the first vehicle.
  • the above embodiments of the present invention can improve air quality inside a vehicle.
  • Fig. 1A illustrates an architecture diagram of an air quality control system according to an embodiment of the present invention
  • Fig. 1B shows some examples of the second type image indicating air pollution
  • Fig. 2 illustrates a flowchart of a method for air quality control according to an embodiment of the present invention
  • Fig. 3B illustrates a flowchart of a processing method according to an embodiment of the present invention
  • Fig. 4A illustrates a flowchart of a processing apparatus according to an embodiment of the present invention
  • Fig. 4B illustrates a flowchart of a processing apparatus according to an embodiment of the present invention
  • Fig. 5 illustrates a flowchart of a control device according to an embodiment of the present invention.
  • Fig. 6 illustrates a flowchart of a cloud server according to an embodiment of the present invention.
  • Fig. 1A illustrates an architecture diagram of an air quality control system according to an embodiment of the present invention.
  • the air quality control system 10 may include a plurality of vehicles 40 and a cloud server 80.
  • Any vehicle 40i of the plurality of vehicles 40 may include an air quality sensor 44, a camera 48 and a control device 52.
  • the air quality sensor 44 of the vehicle 40i may install on the vehicle 40i to measure continually air quality reading around the vehicle 40i.
  • the camera 48 of the vehicle 40i may install in or on the vehicle 40i to continually capture an image including an environment ahead of the vehicle 40i.
  • the control device 52 of the vehicle 40i may be coupled to the air quality sensor 44 and the camera 48 of the vehicle 40i via a wired or wireless communication, and connected to the cloud server 80 via a wireless communication.
  • the control device 52 of the vehicle 40i may receive the air quality reading around the vehicle 40i from the air quality sensor 44 of the vehicle 40i and the image from the camera 48 of the vehicle 40i and transmit the air quality reading around the vehicle 40i and the image to the cloud server 80 if the air quality reading is larger than an air quality reading threshold.
  • the control device 52 of the vehicle 40i may classify the image received from the camera 48 with a trained machine learning model (e.g.
  • a neural network model received from the cloud server 80, wherein the trained machine learning model is used to classify an image into a first type image indicating no air pollution or a second type image indicating air pollution.
  • Fig. 1B shows some examples of the second type image. If the image received from the camera 48 is classified into the second type image indicating air pollution, the control device 52 of the vehicle 40i may switch the air intake of the air condition system of the vehicle 40i into the inner loop air intake to prevent the air outside the vehicle 40i from entering into the vehicle 40i.
  • the cloud server 80 may receive the image and the air quality reading around the vehicle 40i from the control device 52 of the vehicle 40i, and store, in a sample image database, the received image in association with the received air quality reading.
  • the cloud server 80 may re-train the trained machine learning model with the sample images stored in the sample image database to obtain an updated machine learning model, and transmit the updated machine learning model to the vehicle 40i.
  • Fig. 2 illustrates a flowchart of a method for air quality control according to an embodiment of the present invention.
  • the trained machine learning model for classifying an image into the first type image indicating no air pollution or the second type image indicating air pollution has been stored in the cloud server 80 and transmitted to the plurality of vehicles 40 by the cloud server 80.
  • the control device 52 of any vehicle 40i of the plurality of vehicles 40 may receive the image including an environment ahead of the vehicle 40i from the camera 48 of the vehicle 40i and the air quality reading around the vehicle 40i from the air quality sensor 44 of the vehicle 40i. Acquiring of the image by the camera 48 and measuring of the air quality reading by the air quality sensor 44 is basically in the same time.
  • the control device 52 of the vehicle 40i may transmit the received image and the received air quality reading around the vehicle 40i to the cloud server 80 if the received air quality reading is larger than an air quality reading threshold.
  • the received air quality reading is larger than the air quality reading threshold, it indicates that there is air pollution in the environment included in the received image.
  • the control device 52 of the vehicle 40i may obtain values of particular parameters from the received image.
  • the particular parameters are input parameters of the trained machine learning model received from the cloud server 80 and may include: a vehicle emission level, a vehicle type, a vehicle clean level, and a license plate number.
  • the value of the vehicle emission level may correspond to an average of grey values of pixels of the received image.
  • the value of the vehicle type may indicate a large type vehicle if an area of an interesting region including a vehicle extracted from the received image is larger than an area threshold and a small type vehicle if the area of the interesting region is less than the area threshold.
  • the value of the vehicle clean level may correspond to an average of grey values of pixels of an interesting region including a vehicle extracted from the received image.
  • control device 52 of the vehicle 40i may input the values of the particular parameters into the input parameters of the trained machine learning model to classify the received image with the trained machine learning model.
  • the control device 52 of the vehicle 40i may switch the air intake of the air condition system of the vehicle 40i into the inner loop air intake to prevent the air outside the vehicle 40i from entering into the vehicle 40i.
  • control device 52 of the vehicle 40i may notify the driver of the vehicle 40i that the air intake of the air condition system of the vehicle 40i is switched into the inner loop air intake.
  • the cloud server 80 may store, the sample image database, the received image in association with the received air quality reading.
  • the cloud server 80 may check whether the number of the sample images stored in the sample image database is larger than a number threshold.
  • the cloud server 80 may re-train the trained machine learning model with the sample images stored in the sample image database.
  • the cloud server 80 may obtain values of the particular parameters (the particular parameters are input parameters of the trained machine learning model and may include: the vehicle emission level, the vehicle type, the vehicle clean level, and the license plate number) from each image of the sample images stored in the sample image database to acquire the values of the particular parameters for each image of the sample images.
  • the cloud server 80 may weigh the value of the vehicle emission level for each image of the sample images based on the air quality reading stored in association with each image of the sample images, wherein the value of the vehicle emission level is weight more if the air quality reading is larger.
  • the cloud server 80 may then input the values of the particular parameters for each image of the sample images into the input parameters of the trained machine learning model to perform re-training of the trained machine learning model, so that an updated machine learning model is obtained.
  • the cloud server 80 may finally delete the sample images stored in the sample image database after the updated machine learning model is obtained.
  • the cloud server 80 may store the updated machine learning model.
  • the cloud server 80 may transmit the updated machine learning model to the vehicle 40i.
  • control device 52 of the vehicle 40i may replace the trained machine learning model received previously from the cloud server 80 with the updated machine learning model.
  • the solution of the present embodiment uses the image including an environment ahead of a particular vehicle and the machine learning model to determine whether there is air pollution around the particular vehicle, which is not necessary to have the pollutants outside the particular vehicle to flow into the cabin of the particular vehicle.
  • the solution of the present embodiment can improve air quality inside a vehicle.
  • the input parameters of the machine learning model include the vehicle emission level, the vehicle type, the vehicle clean level, and the license plate number, but the present invention is not so limited.
  • the input parameters of the machine learning model may include only a part of the vehicle emission level, the vehicle type, the vehicle clean level, and the license plate number; or, the input parameters of the machine learning model may include other suitable parameter (s) and a part of the vehicle emission level, the vehicle type, the vehicle clean level, and the license plate number; or, the input parameters of the machine learning model may include other suitable parameters except for the vehicle emission level, the vehicle type, the vehicle clean level, and the license plate number.
  • the sample images and the air quality readings required for re-training the trained machine learning model by the cloud server 80 are provided by the plurality of vehicles 40, but the present invention is not so limited.
  • the sample images and the air quality readings required for re-training the trained machine learning model by the cloud server 80 may be provided by other suitable devices except for the plurality of vehicles 40.
  • the other suitable devices may be roadside devices on roads and/or in toll stations.
  • a prerequisite for triggering re-training of the trained machine learning model is that the number of the sample images stored in the sample image database is larger than the number threshold, but the present invention is not so limited. In other embodiments of the present invention, the prerequisite for triggering re-training of the trained machine learning model may be other suitable prerequisite except that the number of the sample images stored in the sample image database is larger than the number threshold.
  • the trained machine learning model used by the plurality of vehicles 40 is provided by the cloud server 80, but the present invention is not so limited. In other embodiments of the present invention, the trained machine learning model may be inputted in the plurality of vehicles 40 by the manufacturers of the plurality of vehicles 40.
  • the air outside the vehicle 40i is prevented from entering into the vehicle 40i by switching the air intake of the air condition system of the vehicle 40i into the inner loop air intake, but the present invention is not so limited. In other embodiments of the present invention, the air outside the vehicle 40i may also be prevented from entering into the vehicle 40i for example by shutting off the air condition system of the vehicle 40i.
  • Fig. 3A illustrates a flowchart of a processing method according to an embodiment of the present invention.
  • the processing method 300 shown in Fig. 3A may be implemented by the control device 52 for example.
  • the processing method 300 may include, at block 302, acquiring an image including an environment ahead of a first vehicle.
  • the processing method 300 may also include, at block 304, classifying the acquired image with a trained machine learning model, wherein the trained machine learning model is used to classify an image into a first type image indicating no air pollution or a second type image indicating air pollution.
  • the processing method 300 may also include, at block 306, preventing air pollutants outside the first vehicle from entering into the first vehicle when the acquired image is classified into the second type image.
  • the trained machine learning model may be received from a cloud server.
  • block 306 may include: switching an air intake of an air condition system of the first vehicle into an inner loop air intake.
  • the processing method 300 may further include: receiving a updated machine learning model from the cloud server; and replacing the trained machine learning model with the updated machine learning model.
  • the processing method 300 may further include: obtaining air quality reading around the first vehicle at the time of acquiring of the acquired image; and transmitting the acquired image and the obtained air quality reading to the cloud server if the obtained air quality reading is larger than an air quality reading threshold.
  • block 304 may include: obtaining values of a plurality of parameters from the acquired image, wherein the plurality of parameters are input parameters of the trained machine learning model; and performing classifying of the acquired image by inputting the values of the plurality of parameters into the input parameters of the trained machine learning model.
  • the plurality of parameters may include a vehicle emission level, a vehicle type, a vehicle clean level, and a license plate number.
  • Fig. 3B illustrates a flowchart of a processing method according to an embodiment of the present invention.
  • the processing method 350 shown in Fig. 3B may be implemented by the cloud server 80 for example.
  • the processing method 350 may include, at block 352, re-training a trained machine learning model by using sample images including environments stored in a sample image database to obtain an updated machine learning model, wherein the trained machine learning model is used to classify an image into a first type image indicating no air pollution or a second type image indicating air pollution, and the sample images include first type images indicating no air pollution and second type images indicating air pollution.
  • the processing method 350 may also include, at block 354, transmitting the updated machine learning model to vehicles.
  • the processing method 350 may further include: receiving, from any vehicle of the vehicles, an image including an environment ahead of the any vehicle and an air quality reading around the any vehicle; and storing, in the sample image database, the received image in association with the received air quality reading.
  • block 352 may include: obtaining values of a plurality of parameters from each of the sample images to acquire the values of the plurality of parameters for each of the sample images, wherein the plurality of parameters is input parameters of the trained machine learning model; weighting one or more of the values of the plurality of parameters for each of the sample images based on an air quality reading stored in association with each of the sample images; and performing re-training of the trained machine learning model by inputting the values of the plurality of parameters for each of the sample images into the input parameters of the trained machine learning model.
  • the plurality of parameters may include a vehicle emission level, a vehicle type, a vehicle clean level, and a license plate number.
  • Fig. 4A illustrates a schematic diagram of a processing apparatus according to an embodiment of the present invention.
  • the processing apparatus 400 shown in Fig. 4A may be implemented in software, hardware or a combination of software and hardware.
  • the processing apparatus 400 may install in the control device 52 for example.
  • the processing apparatus 400 may include an acquiring module 402, a classifying module 404 and a preventing module 406.
  • the acquiring module 402 is configured for acquiring an image including an environment ahead of a first vehicle.
  • the classifying module 404 is configured for classifying the acquired image with a trained machine learning model, wherein the trained machine learning model is used to classify an image into a first type image indicating no air pollution or a second type image indicating air pollution.
  • the preventing module 406 is configured for preventing air pollutants outside the first vehicle from entering into the first vehicle when the acquired image is classified into the second type image.
  • the trained machine learning model may be received from a cloud server.
  • the preventing module 406 may be further configured for switching an air intake of an air condition system of the first vehicle into an inner loop air intake.
  • the processing apparatus 400 may further include: a receiving module for receiving a updated machine learning model from the cloud server; and a replacing module for replacing the trained machine learning model with the updated machine learning model.
  • the processing apparatus 400 may further include: an first obtaining module for obtaining air quality reading around the first vehicle at the time of acquiring of the acquired image; and a transmitting module for transmitting the acquired image and the obtained air quality reading to the cloud server if the obtained air quality reading is larger than an air quality reading threshold.
  • the classifying module 404 may include: an second obtaining module for obtaining values of a plurality of parameters from the acquired image, wherein the plurality of parameters are input parameters of the trained machine learning model; and a performing module for performing classifying of the acquired image by inputting the values of the plurality of parameters into the input parameters of the trained machine learning model.
  • the plurality of parameters may include a vehicle emission level, a vehicle type, a vehicle clean level, and a license plate number.
  • Fig. 4B illustrates a schematic diagram of a processing apparatus according to an embodiment of the present invention.
  • the processing apparatus 450 shown in Fig. 4B may be implemented in software, hardware or a combination of software and hardware.
  • the processing apparatus 450 may install in the cloud server 80 for example.
  • the processing apparatus 450 may include a re-training module 452 and a transmitting module 454.
  • the re-training module 452 is configured for re-training a trained machine learning model by using sample images including environments stored in a sample image database to obtain an updated machine learning model, wherein the trained machine learning model is used to classify an image into a first type image indicating no air pollution or a second type image indicating air pollution, and the sample images include first type images indicating no air pollution and second type images indicating air pollution.
  • the transmitting module 454 is configured for transmitting the updated machine learning model to vehicles.
  • the processing apparatus 450 may further include: a receiving module for receiving, from any vehicle of the vehicles, an image including an environment ahead of the any vehicle and an air quality reading around the any vehicle; and a storing module for storing, in the sample image database, the received image in association with the received air quality reading.
  • the re-training module 452 may include: an obtaining module for obtaining values of a plurality of parameters from each of the sample images to acquire the values of the plurality of parameters for each of the sample images, wherein the plurality of parameters is input parameters of the trained machine learning model; a weighting module for weighting one or more of the values of the plurality of parameters for each of the sample images based on an air quality reading stored in association with each of the sample images; and a performing module for performing re-training of the trained machine learning model by inputting the values of the plurality of parameters for each of the sample images into the input parameters of the trained machine learning model.
  • the plurality of parameters may include a vehicle emission level, a vehicle type, a vehicle clean level, and a license plate number.
  • Fig. 5 illustrates a schematic diagram of a control device according to an embodiment of the present invention.
  • the control device 500 may include a processor 502 and a memory 504.
  • the memory 504 may store executable instructions which cause, when executed, the processor 502 to execute operations included in the method 300.
  • Fig. 6 illustrates a schematic diagram of a cloud server according to an embodiment of the present invention.
  • the cloud server 600 may include a processor 602 and a memory 604.
  • the memory 604 may store executable instructions which cause, when executed, the processor 602 to execute operations included in the method 350.
  • Embodiments of the present invention may provide a machine-readable medium storing thereon executable instructions which, when executed, cause a machine to execute the above methods.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Thermal Sciences (AREA)
  • Mechanical Engineering (AREA)
  • General Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • Theoretical Computer Science (AREA)
  • Traffic Control Systems (AREA)
  • Image Analysis (AREA)

Abstract

The present invention relates to a processing method and a processing apparatus, wherein the processing method comprising: acquiring an image including an environment ahead of a first vehicle; classifying the acquired image with a trained machine learning model, wherein the trained machine learning model is used to classify an image into a first type image indicating no air pollution or a second type image indicating air pollution; and preventing air pollutants outside the first vehicle from entering into the first vehicle when the acquired image is classified into the second type image. The processing method and the processing apparatus can improve air quality inside the vehicle.

Description

PROCESSING METHOD, PROCESSING APPARATUS, CONTROL DEVICE AND CLOUD SERVER TECHNICAL FIELD
The present invention relates to a field of a vehicle and particularly to a processing method, a processing apparatus, a control device and a cloud server.
BACKGROUND
Currently a vehicle usually uses an air quality sensor installed in the vehicle to measure air quality around the vehicle, and when it is measured that the air quality around the vehicle is not good, air intake of the air condition system of the vehicle is switched into an inner loop air intake to prevent pollutants outside the vehicle from flowing into a cabin of the vehicle.
The existed air quality control technology used in the vehicle takes reactive approach, by which some pollutants must have flowed into the cabin of the vehicle already. In many cases, it is very hard to clean the pollutants flowed into the cabin of the vehicle, especially VOC (Volatile Organic Compound) related pollutants.
SUMMARY
Embodiments of the present invention provide a processing method, a processing apparatus, a control device and a cloud server, which can improve air quality inside a vehicle.
A processing method according to embodiments of the present invention comprises: acquiring an image including an environment ahead of a first vehicle; classifying the acquired image with a trained machine learning model, wherein the trained machine learning model is used to classify an image into a first type image indicating no air pollution or a second type image indicating air pollution; and  preventing air pollutants outside the first vehicle from entering into the first vehicle when the acquired image is classified into the second type image.
A processing method according to embodiments of the present invention comprises: re-training a trained machine learning model by using sample images including environments stored in a sample image database to obtain an updated machine learning model, wherein the trained machine learning model is used to classify an image into a first type image indicating no air pollution or a second type image indicating air pollution, and the sample images include first type images indicating no air pollution and second type images indicating air pollution; and transmitting the updated machine learning model to vehicles.
A processing apparatus according to embodiments of the present invention comprises: an acquiring module for acquiring an image including an environment ahead of a first vehicle; a classifying module for classifying the acquired image with a trained machine learning model, wherein the trained machine learning model is used to classify an image into a first type image indicating no air pollution or a second type image indicating air pollution; and a switching module for preventing air pollutants outside the first vehicle from entering into the first vehicle when the acquired image is classified into the second type image.
A processing apparatus according to embodiments of the present invention comprises: a re-training module for re-training a trained machine learning model by using sample images including environments stored in a sample image database to obtain an updated machine learning model, wherein the trained machine learning model is used to classify an image into a first type image indicating no air pollution or a second type image indicating air pollution, and the sample images include first type images indicating no air pollution and second type images indicating air pollution; and a transmitting module for transmitting the updated machine learning model to vehicles.
A control device according to embodiments of the present invention comprises: a processor; and a memory for storing executable instructions which cause, when executed, the processor to execute operations included in the above method.
A cloud server according to embodiments of the present invention comprises: a processor; and a memory for storing executable instructions which cause, when executed, the processor to execute operations included in the above method.
A computer readable medium according to embodiments of the present invention has thereon executable instructions for causing, when executed, a machine to execute operations included in the above method.
The above embodiments of the present invention uses the image including an environment ahead of the first vehicle and the machine learning model to determine whether there is air pollution around the first vehicle, which is not necessary to have the pollutants outside the first vehicle to flow into the cabin of the first vehicle. Thus, compared to the prior art, the above embodiments of the present invention can improve air quality inside a vehicle.
DESCRIPTION OF FIGURES
The above and other features and advantages of the present invention will become more apparent from the following detailed description made with reference to the accompanying drawings. In the drawings:
Fig. 1A illustrates an architecture diagram of an air quality control system according to an embodiment of the present invention;
Fig. 1B shows some examples of the second type image indicating air pollution;
Fig. 2 illustrates a flowchart of a method for air quality control according to an embodiment of the present invention;
Fig. 3A illustrates a flowchart of a processing method according to an embodiment of the present invention;
Fig. 3B illustrates a flowchart of a processing method according to an embodiment of the present invention;
Fig. 4A illustrates a flowchart of a processing apparatus according to an embodiment of the present invention;
Fig. 4B illustrates a flowchart of a processing apparatus according to an embodiment of the present invention;
Fig. 5 illustrates a flowchart of a control device according to an embodiment of the present invention; and
Fig. 6 illustrates a flowchart of a cloud server according to an embodiment of the present invention.
DETAILED DESCRIPTION
Various embodiments are now described with reference to the drawings, wherein like reference numerals are used to refer to like elements throughout. In the following description, for purposes of explanation, numerous specific details are set forth in order to provide a thorough understanding of one or more embodiments. However, it may be evident that such embodiment (s) can be practiced without these specific details.
Fig. 1A illustrates an architecture diagram of an air quality control system according to an embodiment of the present invention. As shown in Fig. 1A, the air quality control system 10 may include a plurality of vehicles 40 and a cloud server 80.
Any vehicle 40i of the plurality of vehicles 40 may include an air quality sensor 44, a camera 48 and a control device 52.
The air quality sensor 44 of the vehicle 40i may install on the vehicle 40i to measure continually air quality reading around the vehicle 40i.
The camera 48 of the vehicle 40i may install in or on the vehicle 40i to continually capture an image including an environment ahead of the vehicle 40i.
The control device 52 of the vehicle 40i may be coupled to the air quality sensor 44 and the camera 48 of the vehicle 40i via a wired or wireless communication, and connected to the cloud server 80 via a wireless communication. The control device 52 of the vehicle 40i may receive the air quality reading around the vehicle 40i from the air quality sensor 44 of the vehicle 40i and the image from the camera 48 of the vehicle 40i and transmit the air quality reading around the vehicle 40i and the image to the cloud server 80 if the air quality reading is larger than an air quality reading threshold. Further, the control device 52 of the vehicle 40i may classify the image  received from the camera 48 with a trained machine learning model (e.g. , a neural network model, a Bayesian model, a support vector machine model, a decision tree model and the like) received from the cloud server 80, wherein the trained machine learning model is used to classify an image into a first type image indicating no air pollution or a second type image indicating air pollution. Fig. 1B shows some examples of the second type image. If the image received from the camera 48 is classified into the second type image indicating air pollution, the control device 52 of the vehicle 40i may switch the air intake of the air condition system of the vehicle 40i into the inner loop air intake to prevent the air outside the vehicle 40i from entering into the vehicle 40i.
The cloud server 80 may receive the image and the air quality reading around the vehicle 40i from the control device 52 of the vehicle 40i, and store, in a sample image database, the received image in association with the received air quality reading. The cloud server 80 may re-train the trained machine learning model with the sample images stored in the sample image database to obtain an updated machine learning model, and transmit the updated machine learning model to the vehicle 40i.
Fig. 2 illustrates a flowchart of a method for air quality control according to an embodiment of the present invention. Here, it is assumed that the trained machine learning model for classifying an image into the first type image indicating no air pollution or the second type image indicating air pollution has been stored in the cloud server 80 and transmitted to the plurality of vehicles 40 by the cloud server 80.
As shown in Fig. 2, at block 202, the control device 52 of any vehicle 40i of the plurality of vehicles 40 may receive the image including an environment ahead of the vehicle 40i from the camera 48 of the vehicle 40i and the air quality reading around the vehicle 40i from the air quality sensor 44 of the vehicle 40i. Acquiring of the image by the camera 48 and measuring of the air quality reading by the air quality sensor 44 is basically in the same time.
At block 204, the control device 52 of the vehicle 40i may transmit the received image and the received air quality reading around the vehicle 40i to the cloud server  80 if the received air quality reading is larger than an air quality reading threshold. Here, if the received air quality reading is larger than the air quality reading threshold, it indicates that there is air pollution in the environment included in the received image.
At block 206, the control device 52 of the vehicle 40i may obtain values of particular parameters from the received image. The particular parameters are input parameters of the trained machine learning model received from the cloud server 80 and may include: a vehicle emission level, a vehicle type, a vehicle clean level, and a license plate number. For example, the value of the vehicle emission level may correspond to an average of grey values of pixels of the received image. The value of the vehicle type may indicate a large type vehicle if an area of an interesting region including a vehicle extracted from the received image is larger than an area threshold and a small type vehicle if the area of the interesting region is less than the area threshold. The value of the vehicle clean level may correspond to an average of grey values of pixels of an interesting region including a vehicle extracted from the received image.
At block 208, the control device 52 of the vehicle 40i may input the values of the particular parameters into the input parameters of the trained machine learning model to classify the received image with the trained machine learning model.
At block 210, if the received image is classified into the second type image indicating air pollution, the control device 52 of the vehicle 40i may switch the air intake of the air condition system of the vehicle 40i into the inner loop air intake to prevent the air outside the vehicle 40i from entering into the vehicle 40i.
At block 212, the control device 52 of the vehicle 40i may notify the driver of the vehicle 40i that the air intake of the air condition system of the vehicle 40i is switched into the inner loop air intake.
At block 214, when receiving from the vehicle 40i the image including the environment ahead of the vehicle 40i and the air quality reading around the vehicle 40i, the cloud server 80 may store, the sample image database, the received image in association with the received air quality reading.
At block 216, the cloud server 80 may check whether the number of the sample images stored in the sample image database is larger than a number threshold.
At block 218, if the check result at block 220 is confirmative, the cloud server 80 may re-train the trained machine learning model with the sample images stored in the sample image database. For example, the cloud server 80 may obtain values of the particular parameters (the particular parameters are input parameters of the trained machine learning model and may include: the vehicle emission level, the vehicle type, the vehicle clean level, and the license plate number) from each image of the sample images stored in the sample image database to acquire the values of the particular parameters for each image of the sample images. The cloud server 80 may weigh the value of the vehicle emission level for each image of the sample images based on the air quality reading stored in association with each image of the sample images, wherein the value of the vehicle emission level is weight more if the air quality reading is larger. The cloud server 80 may then input the values of the particular parameters for each image of the sample images into the input parameters of the trained machine learning model to perform re-training of the trained machine learning model, so that an updated machine learning model is obtained. The cloud server 80 may finally delete the sample images stored in the sample image database after the updated machine learning model is obtained.
At block 220, the cloud server 80 may store the updated machine learning model.
At block 222, the cloud server 80 may transmit the updated machine learning model to the vehicle 40i.
At block 224, after receiving the updated machine learning model from the cloud server 80, the control device 52 of the vehicle 40i may replace the trained machine learning model received previously from the cloud server 80 with the updated machine learning model.
The solution of the present embodiment uses the image including an environment ahead of a particular vehicle and the machine learning model to determine whether there is air pollution around the particular vehicle, which is not necessary to have the  pollutants outside the particular vehicle to flow into the cabin of the particular vehicle. Thus, compared to the prior art, the solution of the present embodiment can improve air quality inside a vehicle.
Other Modifications
Those skilled in the art will appreciate that in the above embodiment, the input parameters of the machine learning model include the vehicle emission level, the vehicle type, the vehicle clean level, and the license plate number, but the present invention is not so limited. In other embodiments of the present invention, the input parameters of the machine learning model may include only a part of the vehicle emission level, the vehicle type, the vehicle clean level, and the license plate number; or, the input parameters of the machine learning model may include other suitable parameter (s) and a part of the vehicle emission level, the vehicle type, the vehicle clean level, and the license plate number; or, the input parameters of the machine learning model may include other suitable parameters except for the vehicle emission level, the vehicle type, the vehicle clean level, and the license plate number.
Those skilled in the art will appreciate that in the above embodiment, the sample images and the air quality readings required for re-training the trained machine learning model by the cloud server 80 are provided by the plurality of vehicles 40, but the present invention is not so limited. In other embodiments of the present invention, the sample images and the air quality readings required for re-training the trained machine learning model by the cloud server 80 may be provided by other suitable devices except for the plurality of vehicles 40. For example, the other suitable devices may be roadside devices on roads and/or in toll stations.
Those skilled in the art will appreciate that in the above embodiment, a prerequisite for triggering re-training of the trained machine learning model is that the number of the sample images stored in the sample image database is larger than the number threshold, but the present invention is not so limited. In other embodiments of the present invention, the prerequisite for triggering re-training of the trained machine learning model may be other suitable prerequisite except that the number of the  sample images stored in the sample image database is larger than the number threshold.
Those skilled in the art will appreciate that in the above embodiment, the trained machine learning model used by the plurality of vehicles 40 is provided by the cloud server 80, but the present invention is not so limited. In other embodiments of the present invention, the trained machine learning model may be inputted in the plurality of vehicles 40 by the manufacturers of the plurality of vehicles 40.
Those skilled in the art will appreciate that in the above embodiment, the air outside the vehicle 40i is prevented from entering into the vehicle 40i by switching the air intake of the air condition system of the vehicle 40i into the inner loop air intake, but the present invention is not so limited. In other embodiments of the present invention, the air outside the vehicle 40i may also be prevented from entering into the vehicle 40i for example by shutting off the air condition system of the vehicle 40i.
Fig. 3A illustrates a flowchart of a processing method according to an embodiment of the present invention. The processing method 300 shown in Fig. 3A may be implemented by the control device 52 for example.
As shown in Fig. 3A, the processing method 300 may include, at block 302, acquiring an image including an environment ahead of a first vehicle.
The processing method 300 may also include, at block 304, classifying the acquired image with a trained machine learning model, wherein the trained machine learning model is used to classify an image into a first type image indicating no air pollution or a second type image indicating air pollution.
The processing method 300 may also include, at block 306, preventing air pollutants outside the first vehicle from entering into the first vehicle when the acquired image is classified into the second type image.
In a first aspect, the trained machine learning model may be received from a cloud server.
In a second aspect, block 306 may include: switching an air intake of an air condition system of the first vehicle into an inner loop air intake.
In a third aspect, the processing method 300 may further include: receiving a updated machine learning model from the cloud server; and replacing the trained machine learning model with the updated machine learning model.
In a fourth aspect, the processing method 300 may further include: obtaining air quality reading around the first vehicle at the time of acquiring of the acquired image; and transmitting the acquired image and the obtained air quality reading to the cloud server if the obtained air quality reading is larger than an air quality reading threshold.
In a fifth aspect, block 304 may include: obtaining values of a plurality of parameters from the acquired image, wherein the plurality of parameters are input parameters of the trained machine learning model; and performing classifying of the acquired image by inputting the values of the plurality of parameters into the input parameters of the trained machine learning model.
In a sixth aspect, the plurality of parameters may include a vehicle emission level, a vehicle type, a vehicle clean level, and a license plate number.
Fig. 3B illustrates a flowchart of a processing method according to an embodiment of the present invention. The processing method 350 shown in Fig. 3B may be implemented by the cloud server 80 for example.
As shown in Fig. 3B, the processing method 350 may include, at block 352, re-training a trained machine learning model by using sample images including environments stored in a sample image database to obtain an updated machine learning model, wherein the trained machine learning model is used to classify an image into a first type image indicating no air pollution or a second type image indicating air pollution, and the sample images include first type images indicating no air pollution and second type images indicating air pollution.
The processing method 350 may also include, at block 354, transmitting the updated machine learning model to vehicles.
In a first aspect, the processing method 350 may further include: receiving, from any vehicle of the vehicles, an image including an environment ahead of the any vehicle and an air quality reading around the any vehicle; and storing, in the sample  image database, the received image in association with the received air quality reading.
In a second aspect, block 352 may include: obtaining values of a plurality of parameters from each of the sample images to acquire the values of the plurality of parameters for each of the sample images, wherein the plurality of parameters is input parameters of the trained machine learning model; weighting one or more of the values of the plurality of parameters for each of the sample images based on an air quality reading stored in association with each of the sample images; and performing re-training of the trained machine learning model by inputting the values of the plurality of parameters for each of the sample images into the input parameters of the trained machine learning model.
In a third aspect, the plurality of parameters may include a vehicle emission level, a vehicle type, a vehicle clean level, and a license plate number.
Fig. 4A illustrates a schematic diagram of a processing apparatus according to an embodiment of the present invention. The processing apparatus 400 shown in Fig. 4A may be implemented in software, hardware or a combination of software and hardware. The processing apparatus 400 may install in the control device 52 for example.
As shown in Fig. 4A, the processing apparatus 400 may include an acquiring module 402, a classifying module 404 and a preventing module 406. The acquiring module 402 is configured for acquiring an image including an environment ahead of a first vehicle. The classifying module 404 is configured for classifying the acquired image with a trained machine learning model, wherein the trained machine learning model is used to classify an image into a first type image indicating no air pollution or a second type image indicating air pollution. The preventing module 406 is configured for preventing air pollutants outside the first vehicle from entering into the first vehicle when the acquired image is classified into the second type image.
In a first aspect, the trained machine learning model may be received from a cloud server.
In a second aspect, the preventing module 406 may be further configured for switching an air intake of an air condition system of the first vehicle into an inner loop air intake.
In a third aspect, the processing apparatus 400 may further include: a receiving module for receiving a updated machine learning model from the cloud server; and a replacing module for replacing the trained machine learning model with the updated machine learning model.
In a fourth aspect, the processing apparatus 400 may further include: an first obtaining module for obtaining air quality reading around the first vehicle at the time of acquiring of the acquired image; and a transmitting module for transmitting the acquired image and the obtained air quality reading to the cloud server if the obtained air quality reading is larger than an air quality reading threshold.
In a fifth aspect, the classifying module 404 may include: an second obtaining module for obtaining values of a plurality of parameters from the acquired image, wherein the plurality of parameters are input parameters of the trained machine learning model; and a performing module for performing classifying of the acquired image by inputting the values of the plurality of parameters into the input parameters of the trained machine learning model.
In a sixth aspect, the plurality of parameters may include a vehicle emission level, a vehicle type, a vehicle clean level, and a license plate number.
Fig. 4B illustrates a schematic diagram of a processing apparatus according to an embodiment of the present invention. The processing apparatus 450 shown in Fig. 4B may be implemented in software, hardware or a combination of software and hardware. The processing apparatus 450 may install in the cloud server 80 for example.
As shown in Fig. 4B, the processing apparatus 450 may include a re-training module 452 and a transmitting module 454. The re-training module 452 is configured for re-training a trained machine learning model by using sample images including environments stored in a sample image database to obtain an updated machine  learning model, wherein the trained machine learning model is used to classify an image into a first type image indicating no air pollution or a second type image indicating air pollution, and the sample images include first type images indicating no air pollution and second type images indicating air pollution. The transmitting module 454 is configured for transmitting the updated machine learning model to vehicles.
In a first aspect, the processing apparatus 450 may further include: a receiving module for receiving, from any vehicle of the vehicles, an image including an environment ahead of the any vehicle and an air quality reading around the any vehicle; and a storing module for storing, in the sample image database, the received image in association with the received air quality reading.
In a second aspect, the re-training module 452 may include: an obtaining module for obtaining values of a plurality of parameters from each of the sample images to acquire the values of the plurality of parameters for each of the sample images, wherein the plurality of parameters is input parameters of the trained machine learning model; a weighting module for weighting one or more of the values of the plurality of parameters for each of the sample images based on an air quality reading stored in association with each of the sample images; and a performing module for performing re-training of the trained machine learning model by inputting the values of the plurality of parameters for each of the sample images into the input parameters of the trained machine learning model.
In a third aspect, the plurality of parameters may include a vehicle emission level, a vehicle type, a vehicle clean level, and a license plate number.
Fig. 5 illustrates a schematic diagram of a control device according to an embodiment of the present invention. As shown in Fig. 5, the control device 500 may include a processor 502 and a memory 504. The memory 504 may store executable instructions which cause, when executed, the processor 502 to execute operations included in the method 300.
Fig. 6 illustrates a schematic diagram of a cloud server according to an  embodiment of the present invention. As shown in Fig. 6, the cloud server 600 may include a processor 602 and a memory 604. The memory 604 may store executable instructions which cause, when executed, the processor 602 to execute operations included in the method 350.
Embodiments of the present invention may provide a machine-readable medium storing thereon executable instructions which, when executed, cause a machine to execute the above methods.
Other embodiments and modifications of this invention will be apparent to those having ordinary skill in the art upon consideration of the specification and practice of the invention disclosed herein. The specification and examples given should be considered exemplary only, and it is contemplated that the appended claims will cover any other such embodiments or modifications as fall within the true scope of the invention.

Claims (25)

  1. A processing method, comprising:
    acquiring an image including an environment ahead of a first vehicle;
    classifying the acquired image with a trained machine learning model, wherein the trained machine learning model is used to classify an image into a first type image indicating no air pollution or a second type image indicating air pollution; and
    preventing air pollutants outside the first vehicle from entering into the first vehicle when the acquired image is classified into the second type image.
  2. The method of claim 1, wherein the trained machine learning model is received from a cloud server.
  3. The method of claim 1, wherein the preventing air pollutants outside the first vehicle from entering into the first vehicle comprising:
    switching an air intake of an air condition system of the first vehicle into an inner loop air intake.
  4. The method of claim 2, wherein further comprising:
    receiving a updated machine learning model from the cloud server; and
    replacing the trained machine learning model with the updated machine learning model.
  5. The method of claim 2, wherein further comprising:
    obtaining air quality reading around the first vehicle at the time of acquiring of the acquired image; and
    transmitting the acquired image and the obtained air quality reading to the cloud server if the obtained air quality reading is larger than an air quality reading threshold.
  6. The method of claim 1, wherein the classifying the acquired image with the trained machine learning model comprising:
    obtaining values of a plurality of parameters from the acquired image, wherein the plurality of parameters are input parameters of the trained machine learning model; and
    performing classifying of the acquired image by inputting the values of the plurality of parameters into the input parameters of the trained machine learning model.
  7. The method of claim 6, wherein the plurality of parameters comprising: a vehicle emission level, a vehicle type, a vehicle clean level, and a license plate number.
  8. A processing method, comprising:
    re-training a trained machine learning model by using sample images including environments stored in a sample image database to obtain an updated machine learning model, wherein the trained machine learning model is used to classify an image into a first type image indicating no air pollution or a second type image indicating air pollution, and the sample images include first type images indicating no air pollution and second type images indicating air pollution; and
    transmitting the updated machine learning model to vehicles.
  9. The method of claim 8, wherein further comprising:
    receiving, from any vehicle of the vehicles, an image including an environment ahead of the any vehicle and an air quality reading around the any vehicle; and
    storing, in the sample image database, the received image in association with the received air quality reading.
  10. The method of claim 8, wherein the re-training a trained machine learning model comprising:
    obtaining values of a plurality of parameters from each of the sample images to acquire the values of the plurality of parameters for each of the sample images, wherein the plurality of parameters is input parameters of the trained machine learning model;
    weighting one or more of the values of the plurality of parameters for each of the sample images based on an air quality reading stored in association with each of the sample images; and
    performing re-training of the trained machine learning model by inputting the values of the plurality of parameters for each of the sample images into the input parameters of the trained machine learning model.
  11. The method of claim 10, wherein the plurality of parameters comprising: a vehicle emission level, a vehicle type, a vehicle clean level, and a license plate number.
  12. A processing apparatus, comprising:
    an acquiring module for acquiring an image including an environment ahead of a first vehicle;
    a classifying module for classifying the acquired image with a trained machine learning model, wherein the trained machine learning model is used to classify an image into a first type image indicating no air pollution or a second type image indicating air pollution; and
    a preventing module for preventing air pollutants outside the first vehicle from entering into the first vehicle when the acquired image is classified into the second type image.
  13. The processing apparatus of claim 12, wherein the trained machine learning model is received from a cloud server.
  14. The processing apparatus of claim 12, wherein the preventing module is  further configured for:
    switching an air intake of an air condition system of the first vehicle into an inner loop air intake.
  15. The processing apparatus of claim 13, wherein further comprising:
    a receiving module for receiving a updated machine learning model from the cloud server; and
    a replacing module for replacing the trained machine learning model with the updated machine learning model.
  16. The processing apparatus of claim 13, wherein further comprising:
    an first obtaining module for obtaining air quality reading around the first vehicle at the time of acquiring of the acquired image; and
    a transmitting module for transmitting the acquired image and the obtained air quality reading to the cloud server if the obtained air quality reading is larger than an air quality reading threshold.
  17. The processing apparatus of claim 12, wherein the classifying module comprising:
    an second obtaining module for obtaining values of a plurality of parameters from the acquired image, wherein the plurality of parameters are input parameters of the trained machine learning model; and
    a performing module for performing classifying of the acquired image by inputting the values of the plurality of parameters into the input parameters of the trained machine learning model.
  18. The processing apparatus of claim 17, wherein the plurality of parameters comprising: a vehicle emission level, a vehicle type, a vehicle clean level, and a license plate number.
  19. A processing apparatus, comprising:
    a re-training module for re-training a trained machine learning model by using sample images including environments stored in a sample image database to obtain an updated machine learning model, wherein the trained machine learning model is used to classify an image into a first type image indicating no air pollution or a second type image indicating air pollution, and the sample images include first type images indicating no air pollution and second type images indicating air pollution; and
    a transmitting module for transmitting the updated machine learning model to vehicles.
  20. The processing apparatus of claim 19, wherein further comprising:
    a receiving module for receiving, from any vehicle of the vehicles, an image including an environment ahead of the any vehicle and an air quality reading around the any vehicle; and
    a storing module for storing, in the sample image database, the received image in association with the received air quality reading.
  21. The processing apparatus of claim 19, wherein the re-training module comprising:
    an obtaining module for obtaining values of a plurality of parameters from each of the sample images to acquire the values of the plurality of parameters for each of the sample images, wherein the plurality of parameters is input parameters of the trained machine learning model;
    a weighting module for weighting one or more of the values of the plurality of parameters for each of the sample images based on an air quality reading stored in association with each of the sample images; and
    a performing module for performing re-training of the trained machine learning model by inputting the values of the plurality of parameters for each of the sample images into the input parameters of the trained machine learning model.
  22. The processing apparatus of claim 21, wherein the plurality of parameters comprising: a vehicle emission level, a vehicle type, a vehicle clean level, and a license plate number.
  23. A control device, comprising:
    a processor; and
    a memory for storing executable instructions which cause, when executed, the processor to execute operations included in the method of any of claims 1-7.
  24. A cloud server, comprising:
    a processor; and
    a memory for storing executable instructions which cause, when executed, the processor to execute operations included in the method of any of claims 8-11.
  25. A machine readable medium having thereon executable instructions for causing, when executed, a machine to execute operations included in the method of any of claims 1-11.
PCT/CN2017/118946 2017-12-27 2017-12-27 Processing method, processing apparatus, control device and cloud server Ceased WO2019127085A1 (en)

Priority Applications (2)

Application Number Priority Date Filing Date Title
CN201780097995.XA CN111542834A (en) 2017-12-27 2017-12-27 Processing method, processing device, control equipment and cloud server
PCT/CN2017/118946 WO2019127085A1 (en) 2017-12-27 2017-12-27 Processing method, processing apparatus, control device and cloud server

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2017/118946 WO2019127085A1 (en) 2017-12-27 2017-12-27 Processing method, processing apparatus, control device and cloud server

Publications (1)

Publication Number Publication Date
WO2019127085A1 true WO2019127085A1 (en) 2019-07-04

Family

ID=67062801

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2017/118946 Ceased WO2019127085A1 (en) 2017-12-27 2017-12-27 Processing method, processing apparatus, control device and cloud server

Country Status (2)

Country Link
CN (1) CN111542834A (en)
WO (1) WO2019127085A1 (en)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112906825A (en) * 2021-03-30 2021-06-04 第四范式（北京）技术有限公司 Method and computing node for realizing distributed training of machine learning model
FR3106304A1 (en) * 2020-01-22 2021-07-23 Psa Automobiles Sa Method and system for managing the air quality of the passenger compartment of a motor vehicle according to the vehicles to which it is close
WO2021244984A1 (en) * 2020-06-02 2021-12-09 Daimler Ag Method for adjusting a supply of external air into an interior compartment of a vehicle
CN114577481A (en) * 2020-12-02 2022-06-03 新智数字科技有限公司 Pollution index monitoring method and device for gas internal combustion engine

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2015051959A1 (en) * 2013-10-10 2015-04-16 Robert Bosch Gmbh Method and apparatus for controlling an air-recirculation operation in a motor vehicle
US20160232423A1 (en) * 2015-02-11 2016-08-11 Qualcomm Incorporated Environmental scene condition detection
US20170166209A1 (en) * 2015-12-09 2017-06-15 Ford Global Technologies, Llc Dust resuspension system for a motor vehicle
US20170270368A1 (en) * 2016-03-15 2017-09-21 Robert Bosch Gmbh Method for detecting a soiling of an optical component of a driving environment sensor used to capture a field surrounding a vehicle; method for automatically training a classifier; and a detection system

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20170022340A (en) * 2015-08-20 2017-03-02 한온시스템 주식회사 Vehicle air conditioning device and method for preventing exhaust gas inflow in vehicle
US10049284B2 (en) * 2016-04-11 2018-08-14 Ford Global Technologies Vision-based rain detection using deep learning
TWI592883B (en) * 2016-04-22 2017-07-21 財團法人車輛研究測試中心 Image recognition system and its adaptive learning method
KR101883743B1 (en) * 2016-05-18 2018-07-31 네이버 주식회사 Apparatus and method for providing vehicle air conditioning system based on vision recognition
CN106203346A (en) * 2016-07-13 2016-12-07 吉林大学 A kind of road environment image classification method towards the switching of intelligent vehicle driving model

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2015051959A1 (en) * 2013-10-10 2015-04-16 Robert Bosch Gmbh Method and apparatus for controlling an air-recirculation operation in a motor vehicle
US20160232423A1 (en) * 2015-02-11 2016-08-11 Qualcomm Incorporated Environmental scene condition detection
US20170166209A1 (en) * 2015-12-09 2017-06-15 Ford Global Technologies, Llc Dust resuspension system for a motor vehicle
US20170270368A1 (en) * 2016-03-15 2017-09-21 Robert Bosch Gmbh Method for detecting a soiling of an optical component of a driving environment sensor used to capture a field surrounding a vehicle; method for automatically training a classifier; and a detection system

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
FR3106304A1 (en) * 2020-01-22 2021-07-23 Psa Automobiles Sa Method and system for managing the air quality of the passenger compartment of a motor vehicle according to the vehicles to which it is close
WO2021244984A1 (en) * 2020-06-02 2021-12-09 Daimler Ag Method for adjusting a supply of external air into an interior compartment of a vehicle
CN115697734A (en) * 2020-06-02 2023-02-03 梅赛德斯-奔驰集团股份公司 Method for regulating the supply of external air to the interior of a vehicle
JP2023526872A (en) * 2020-06-02 2023-06-23 メルセデス・ベンツ グループ アクチェンゲゼルシャフト Method for regulating outside air supply to interior of vehicle
JP7523590B2 (en) 2020-06-02 2024-07-26 メルセデス・ベンツ グループ アクチェンゲゼルシャフト Method for regulating the supply of outside air to the interior of a vehicle - Patents.com
US12522047B2 (en) 2020-06-02 2026-01-13 Mercedes-Benz Group AG Method for adjusting external air intake in an interior of a vehicle
CN114577481A (en) * 2020-12-02 2022-06-03 新智数字科技有限公司 Pollution index monitoring method and device for gas internal combustion engine
CN114577481B (en) * 2020-12-02 2024-01-12 新奥新智科技有限公司 Pollution index monitoring method and device for gas internal combustion engine
CN112906825A (en) * 2021-03-30 2021-06-04 第四范式（北京）技术有限公司 Method and computing node for realizing distributed training of machine learning model

Also Published As

Publication number Publication date
CN111542834A (en) 2020-08-14

Similar Documents

Publication Publication Date Title
US20250245923A1 (en) Photo deformation techniques for vehicle repair analysis
WO2019127085A1 (en) Processing method, processing apparatus, control device and cloud server
KR102599212B1 (en) Adaptive Real-Time Detection and Inspection Network (ARDEN)
EP3673417B1 (en) System and method for distributive training and weight distribution in a neural network
US11164051B2 (en) Image and LiDAR segmentation for LiDAR-camera calibration
US10109196B2 (en) Real-time monitoring and diagnostic processing of traffic control data
JP7111175B2 (en) Object recognition system, recognition device, object recognition method, and object recognition program
US11932274B2 (en) Electronic device and control method therefor
CN113366496A (en) Neural network for coarse and fine object classification
WO2019241224A1 (en) Architectures for vehicle tolling
CN104376303B (en) A kind of vehicle imaging method in the case of low resolution
CN104036253A (en) Lane line tracking method and lane line tracking system
KR102261187B1 (en) System and method for machine learning based surveillance video analysis
CN115776610B (en) Camera shooting control method and device for cargo monitoring of freight vehicle
CN112270333A (en) Elevator car abnormity detection method and system aiming at electric vehicle identification
US20240311704A1 (en) Computer-implemented method, computer program, and device for generating a data-based model copy in a sensor
EP4390836A1 (en) Image enhancement targeted at addressing degradations caused by environmental conditions
CN113348663B (en) A container monitoring method, terminal equipment and storage medium
US10929687B2 (en) Authentication by navigation-correlated sensing
CN119229267B (en) Near-collision detection method and device for vehicle, terminal equipment and storage medium
US12456292B2 (en) Apparatus and method for detecting performance degradation of an autonomous vehicle
US20240303546A1 (en) Determining whether a given input record of measurement data is covered by the training of a trained machine learning model
CN113379591B (en) Speed determination method, speed determination device, electronic equipment and storage medium
Nagarajan et al. Detection of Potholes and Speedbumps by Monitoring Front Traffic
US20250191190A1 (en) Systems and methods for managing segmented image data for vehicles

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 17935969

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 17935969

Country of ref document: EP

Kind code of ref document: A1