CN110276322A - A kind of image processing method and device of the unused resource of combination vehicle device - Google Patents
A kind of image processing method and device of the unused resource of combination vehicle device Download PDFInfo
- Publication number
- CN110276322A CN110276322A CN201910562835.9A CN201910562835A CN110276322A CN 110276322 A CN110276322 A CN 110276322A CN 201910562835 A CN201910562835 A CN 201910562835A CN 110276322 A CN110276322 A CN 110276322A
- Authority
- CN
- China
- Prior art keywords
- detection
- image
- model
- vehicle device
- scene
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000003672 processing method Methods 0.000 title claims abstract description 18
- 238000001514 detection method Methods 0.000 claims abstract description 274
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 48
- 238000000034 method Methods 0.000 claims abstract description 23
- 238000012545 processing Methods 0.000 claims abstract description 14
- 238000007689 inspection Methods 0.000 claims description 12
- 238000012549 training Methods 0.000 claims description 9
- 238000012360 testing method Methods 0.000 claims description 8
- 238000004590 computer program Methods 0.000 claims description 5
- 241000208340 Araliaceae Species 0.000 claims description 2
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 claims description 2
- 235000003140 Panax quinquefolius Nutrition 0.000 claims description 2
- 235000008434 ginseng Nutrition 0.000 claims description 2
- 230000001737 promoting effect Effects 0.000 abstract description 2
- 230000000007 visual effect Effects 0.000 description 11
- 230000006870 function Effects 0.000 description 7
- 235000013399 edible fruits Nutrition 0.000 description 6
- 238000010586 diagram Methods 0.000 description 5
- 230000008901 benefit Effects 0.000 description 4
- 230000000694 effects Effects 0.000 description 3
- 230000006978 adaptation Effects 0.000 description 2
- 238000013135 deep learning Methods 0.000 description 2
- 238000005286 illumination Methods 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 235000015243 ice cream Nutrition 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000000737 periodic effect Effects 0.000 description 1
- 238000012805 post-processing Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/59—Context or environment of the image inside of a vehicle, e.g. relating to seat occupancy, driver state or inner lighting conditions
-
- G—PHYSICS
- G07—CHECKING-DEVICES
- G07C—TIME OR ATTENDANCE REGISTERS; REGISTERING OR INDICATING THE WORKING OF MACHINES; GENERATING RANDOM NUMBERS; VOTING OR LOTTERY APPARATUS; ARRANGEMENTS, SYSTEMS OR APPARATUS FOR CHECKING NOT PROVIDED FOR ELSEWHERE
- G07C5/00—Registering or indicating the working of vehicles
- G07C5/08—Registering or indicating performance data other than driving, working, idle, or waiting time, with or without registering driving, working, idle or waiting time
- G07C5/0841—Registering performance data
- G07C5/085—Registering performance data using electronic data carriers
- G07C5/0866—Registering performance data using electronic data carriers the electronic data carrier being a digital video recorder in combination with video camera
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Data Mining & Analysis (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
- Traffic Control Systems (AREA)
Abstract
The present invention provides the image processing methods and device of a kind of unused resource of combination vehicle device, this method comprises: in regular check vehicle device current each hardware device use state parameter, and obtain based on the use state parameter the unused resource parameters of the vehicle device；Judge whether the unused resource parameters meet the preset condition of the vehicle device operation given scenario detection algorithm；If the unused resource parameters meet the preset condition, the scene-detection algorithms are run in the vehicle device, and preset image detecting element is assisted to carry out vision-based detection to the image data for being input to described image detection unit in real time.It based on scheme provided by the invention, is interconnected between vehicle device and preset image detecting element, assistant images processing function is provided, auxiliary is provided to the correlation detection principle on preset image detecting element, to save the cost while promoting detection accuracy.
Description
Technical field
The present invention relates to automobile technical fields, a kind of image processing method more particularly to unused resource of combination vehicle device and
Device.
Background technique
Currently, the fields such as drive in automatic Pilot and auxiliary, and in automatic parking or active safety function, Duo Huiyong
Assist in identifying surrounding enviroment information to algorithm of target detection, and general algorithm of target detection then generally by machine learning or
Person's deep learning is realized.Function is driven for auxiliary, vision-based inspection algorithm is typically all to operate in proprietary chip to work as
In, but the calculation power of proprietary chip is since cost problem is very limited, so the accuracy and speed of vision algorithm is frequently necessary to
It is balanced according to actual scene, sacrifices a small amount of precision to make speed reach requirement of real-time, thus can not be in vision algorithm
Processing accuracy and requirement of real-time between can not active balance.
Summary of the invention
The present invention provides a kind of image processing method of unused resource of combination vehicle device and device with overcome the above problem or
Person at least is partially solved the above problem.
According to an aspect of the invention, there is provided a kind of image processing method of the unused resource of combination vehicle device, comprising:
The use state parameter of current each hardware device in regular check vehicle device, and obtained based on the use state parameter
The unused resource parameters of the vehicle device；
Judge whether the unused resource parameters meet the preset condition of the vehicle device operation given scenario detection algorithm；
If the unused resource parameters meet the preset condition, the scene detection is run in the vehicle device and is calculated
Method assists preset image detecting element to carry out vision-based detection to the image data for being input to described image detection unit in real time.
Optionally, described that the scene-detection algorithms are run in the vehicle device, assist preset image detecting element pair
The image data for being input to described image detection unit in real time carries out vision-based detection, comprising:
Acquisition is input in described image detection unit via image processor treated image data, in the vehicle device
The scenario parameters of the interior operation scene-detection algorithms identification described image data；
Corresponding image detection model is chosen in the multiple images detection model pre-established based on the scenario parameters,
Output detection is tied after carrying out vision-based detection to described image data using described image detection model by described image detection unit
Fruit.
Optionally, the acquisition is input in described image detection unit via image processor treated picture number
According to running the scenario parameters of scene-detection algorithms identification described image data in the vehicle device, comprising:
Acquisition is input in described image detection unit via image processor treated image data；
The scene-detection algorithms are run in the vehicle device, scene detection is carried out to described image data, identify described
The size of target detection thing and its quantity, requirement of real-time and/or precision are wanted in the scene type of image data, image data
It asks；
Wherein, the scene type includes type locating for weather pattern and/or the geographical location in described image data.
Optionally, described to choose corresponding figure in the multiple images detection model pre-established based on the scenario parameters
As detection model, comprising:
Choose image inspection in the multiple images detection model pre-established according to preset rules based on the scenario parameters
Survey model；And/or
The scenario parameters are inputted into preset disaggregated model, scenario parameters matching inspection is based on by the disaggregated model
Survey the model and model parameter of the image detection model of described image data.
Optionally, it is described based on the scenario parameters according to preset rules in the multiple images detection model pre-established
Choose image detection model, comprising:
Chosen in the multiple images detection model pre-established based on the scenario parameters with the scenario parameters in
The matched image detection model of scene type.
Optionally, described that the scenario parameters are inputted into preset disaggregated model, the field is based on by the disaggregated model
Before the model and model parameter of the image detection model of scape parameter matching detection described image data, further includes:
Construct disaggregated model；
It collects the scenario parameters of different scenes and matches image detection model corresponding with the scene, respectively by each scene
Scenario parameters and corresponding image detection model as the disaggregated model output and input the training disaggregated model.
Optionally, described to choose corresponding figure in the multiple images detection model pre-established based on the scenario parameters
As before detection model, further includes:
The image detection model of corresponding several scenes is constructed and trained in advance.
Optionally, described to choose corresponding figure in the multiple images detection model pre-established based on the scenario parameters
As detection model, after carrying out vision-based detection to described image data using described image detection model by described image detection unit
Output test result, comprising:
Corresponding image detection model is chosen in the multiple images detection model pre-established based on the scenario parameters,
The corresponding image detection model configuration file of described image detection model is called by described image detection unit, runs described image
Detection model is to output test result after described image data progress vision-based detection.
According to another aspect of the present invention, a kind of image processing apparatus of unused resource of combination vehicle device is additionally provided, is wrapped
It includes:
It checks module, is configured to the use state parameter of current each hardware device in regular check vehicle device, and based on described
Use state parameter obtains the unused resource parameters of the vehicle device；
Judgment module is configured to judge whether the unused resource parameters meet the vehicle device operation given scenario detection and calculate
The preset condition of method；
Detection module is run in the vehicle device if being configured to the unused resource parameters meets the preset condition
The scene-detection algorithms assist preset image detecting element to the image data for being input to described image detection unit in real time
Carry out vision-based detection.
Optionally, the detection module is additionally configured to, and acquisition is input in described image detection unit via image processing
Device treated image data runs the scene ginseng of the scene-detection algorithms identification described image data in the vehicle device
Number；
Corresponding image detection model is chosen in the multiple images detection model pre-established based on the scenario parameters,
Output detection is tied after carrying out vision-based detection to described image data using described image detection model by described image detection unit
Fruit.
Optionally, the detection module is additionally configured to, and acquisition is input in described image detection unit via image processing
Device treated image data；
The scene-detection algorithms are run in the vehicle device, scene detection is carried out to described image data, identify described
The size of target detection thing and its quantity, requirement of real-time and/or precision are wanted in the scene type of image data, image data
It asks；
Wherein, the scene type includes type locating for weather pattern and/or the geographical location in described image data.
Optionally, the detection module is additionally configured to, and is being pre-established based on the scenario parameters according to preset rules
Image detection model is chosen in multiple images detection model；And/or
The scenario parameters are inputted into preset disaggregated model, scenario parameters matching inspection is based on by the disaggregated model
Survey the model and model parameter of the image detection model of described image data.
Optionally, the detection module is additionally configured to, based on the scenario parameters in the multiple images detection pre-established
It is chosen in model and the matched image detection model of scene type in the scenario parameters.
Optionally, described device further include: the first building module is configured to building disaggregated model；
It collects the scenario parameters of different scenes and matches image detection model corresponding with the scene, respectively by each scene
Scenario parameters and corresponding image detection model as the disaggregated model output and input the training disaggregated model.
Optionally, described device further includes that the second building module is configured to construct in advance and training corresponds to several scenes
Image detection model.
Optionally, the detection module is additionally configured to, based on the scenario parameters in the multiple images detection pre-established
Corresponding image detection model is chosen in model, and the corresponding image of described image detection model is called by described image detection unit
Detection model configuration file, output detection is tied after operation described image detection model carries out vision-based detection to described image data
Fruit.
According to another aspect of the present invention, a kind of computer storage medium, the computer storage medium are additionally provided
It is stored with computer program code, when the computer program code is run on the computing device, leads to the calculating equipment
Execute the image processing method of the unused resource of combination vehicle device described in any of the above embodiments.
The present invention provides the image processing methods and device of a kind of unused resource of combination vehicle device, are checking current vehicle device
In each hardware device have meet preset condition unused resource when, just can in vehicle device Run-time scenario detection algorithm, auxiliary
Preset image detecting element carries out vision-based detection to the image data for being input to the image detecting element in real time.It is mentioned in the present invention
In the scheme of confession, the main unused resource for utilizing vehicle device is interconnected between vehicle device and preset image detecting element, is provided
Assistant images processing function provides auxiliary to the correlation detection principle on preset image detecting element, to promote detection
Save the cost while precision.
The above description is only an overview of the technical scheme of the present invention, in order to better understand the technical means of the present invention,
And it can be implemented in accordance with the contents of the specification, and in order to allow above and other objects of the present invention, feature and advantage can
It is clearer and more comprehensible, the followings are specific embodiments of the present invention.
According to the following detailed description of specific embodiments of the present invention in conjunction with the accompanying drawings, those skilled in the art will be brighter
The above and other objects, advantages and features of the present invention.
Detailed description of the invention
By reading the following detailed description of the preferred embodiment, various other advantages and benefits are common for this field
Technical staff will become clear.The drawings are only for the purpose of illustrating a preferred embodiment, and is not considered as to the present invention
Limitation.And throughout the drawings, the same reference numbers will be used to refer to the same parts.In the accompanying drawings:
Fig. 1 is the image processing method flow diagram of the unused resource of combination vehicle device according to an embodiment of the present invention；
Fig. 2 is the image processing method flow diagram of the unused resource of combination vehicle device according to the preferred embodiment of the invention；
Fig. 3 is that model according to the preferred embodiment of the invention chooses process schematic；
Fig. 4 is the image processing apparatus structural schematic diagram of the unused resource of combination vehicle device according to an embodiment of the present invention；
Fig. 5 is the image processing apparatus structural schematic diagram of the unused resource of combination vehicle device according to the preferred embodiment of the invention.
Specific embodiment
The exemplary embodiment that the present invention will be described in more detail below with reference to accompanying drawings.Although showing the present invention in attached drawing
Exemplary embodiment, it being understood, however, that may be realized in various forms the present invention without should be by embodiments set forth here
It is limited.It is to be able to thoroughly understand the present invention on the contrary, providing these embodiments, and can be by the scope of the present invention
It is fully disclosed to those skilled in the art.
Fig. 1 is the image processing method flow diagram of the unused resource of combination vehicle device provided according to embodiments of the present invention,
Referring to Fig. 1 it is found that the image processing method of the unused resource of combination vehicle device provided in an embodiment of the present invention may include:
Step S102, the use state parameter of current each hardware device in regular check vehicle device, and shape is used based on above-mentioned
State parameter obtains the unused resource parameters of the vehicle device.Wherein, each hardware device in vehicle device may include CPU, memory, Netowrk tape
Wide, hard disk etc. is set to each hardware device in vehicle device.The use state parameter of current each hardware device in checking vehicle device
When, can be checked with Fixed Time Interval, can also with real-time perfoming inspection, it is specific mainly obtain current network bandwidth,
Hard disk has used memory space and memory use state etc., so extrapolate under present case not using CPU handle energy
The unused resource parameters of power, memory, network bandwidth and hard drive space etc. vehicle device.
Step S104, judges whether above-mentioned unused resource parameters meet the default item of vehicle device operation given scenario detection algorithm
Part.
When judging whether unused resource parameters meet the preset condition of vehicle device operation given scenario detection algorithm, can sentence
Whether each hardware resource such as CPU processing capacity, memory size, network bandwidth, hard drive space etc. reaches default in current vehicle device of breaking
The corresponding requirement of operation given scenario detection algorithm.The preset condition can be wanted for different requirement of real-time or precision
It asks and is configured, the present invention is without limitation.
Step S106, if above-mentioned unused resource parameters meet the preset condition, Run-time scenario detection is calculated in vehicle device
Method assists preset image detecting element to carry out vision-based detection to the image data for being input to the image detecting element in real time.
The embodiment of the invention provides a kind of image processing methods of unused resource of combination vehicle device, are checking current vehicle device
In each hardware device have meet preset condition unused resource when, just can in vehicle device Run-time scenario detection algorithm, auxiliary
Preset image detecting element carries out vision-based detection to the image data for being input to the image detecting element in real time.Of the invention real
In the scheme that example offer is provided, the incative money of vehicle device (such as IHU, Infotainment Head Unit, Infotainment host) is utilized
Source is interconnected between vehicle device and preset image detecting element, provides assistant images processing function, is examined to preset image
The correlation detection principle surveyed on unit provides auxiliary, to save the cost while promoting detection accuracy.
The preset image detecting element referred in the present embodiment can be carry out visual detection algorithm proprietary chip or
It is processor.Wherein, visual detection algorithm can be the visual detection algorithm for automatic parking scene, as parking stall line detection,
The detection of moving Object Detection, pedestrian detection, static-obstacle thing (wheel shelves, ice cream bucket) etc.；It can also be for love interaction safety
The visual detection algorithm of forward sight camera, such as pedestrian and vehicle detection, lane detection, traffic lights and speed(-)limit sign detection
Deng.
Above-mentioned steps S106 is referred to, if the unused resource of vehicle device meets the default of vehicle device operation given scenario detection algorithm
Condition then Run-time scenario detection algorithm assistant images detection unit can carry out vision-based detection in vehicle device.Further, step
S106 may include:
S106-1, acquisition are input in image detecting element via image processor treated image data, in vehicle device
The scenario parameters of the interior operation scene-detection algorithms identification described image data.
In the present embodiment, needing image detecting element to carry out the image data of vision-based detection is via being set to automobile body
Image collecting device (such as camera) acquired image data, and may be provided with image processor in image collecting device
The signal that (Image Signal Processor, abbreviation ISP) exports front-end image sensor does post-processing, major function
Linear correction, noise remove, bad point removal, interpolation, white balance, auto-exposure control etc., could be different dependent on ISP
Live details can be preferably restored under optical condition, ISP technology largely determines the image quality of video camera.
ISP provides safeguard imaging effect of the camera under the extreme cases such as backlight, illumination variation (for example crossing tunnel).For vehicle device
For, when its own condition meets preset condition, by obtaining via image processor treated image data to run
Given scenario detection algorithm identifies the scenario parameters of the image data, can make the identification knot to the scenario parameters of image data
Fruit is more accurate.
Optionally, above-mentioned steps S106-1 can further comprise: acquisition is input in described image detection unit via shadow
As processor treated image data；The scene-detection algorithms are run in vehicle device, and scene detection is carried out to image data,
Identify the scene types of described image data, the size of target detection thing and its quantity in image data, requirement of real-time and/
Or required precision；Wherein, the scene type includes class locating for weather pattern and/or the geographical location in described image data
Type.
For example, the scene-detection algorithms that vehicle device is run in the embodiment of the present invention can identify in image data
Scenario parameters at least may include following one:
1, weather relevant parameter, such as rainy day, cloudy day, dusk, snowy day, intensity of illumination etc.；
2, geographic location type parameter, such as urban road, countryside, high velocity environment etc.；
3, image quality parameter, for example, image it is whether fuzzy, image definition, whether have smear, whether have backlight or
Whether retroreflective regions etc. are had in image；
4, in the real-time and required precision, such as the visual field of target detection small object target quantity and speed and this vehicle
Speed etc. combines judgement.
For example, when there was only wisp in the visual field but when there is no near objects, it is higher to required precision, rate request compared with
It is low；It is higher to detection rate request when there was only nearby big object in the visual field, but detection accuracy requirement is lower.When speed is high,
Requirement of real-time is higher；When speed is low, requirement of real-time is lower.It, can be with by combining both article sizes in speed and the visual field
It carries out exporting corresponding precision using a simple Machine learning classifiers (such as SVM or adaboost etc.) and speed is wanted
It asks, and then selects the model of suitable size and precision.
S106-2 chooses corresponding image inspection based on the scenario parameters in the multiple images detection model pre-established
Model is surveyed, output detects knot after carrying out vision-based detection to described image data using the image detection model by image detecting element
Fruit.
Image detection model in the present embodiment, which can be, uses different deep learning or machine for different scenes
Device learning model can make the precision of visual detection algorithm higher by the different corresponding models of scene specific customization, inspection
It is more accurate to survey result.Deep learning model is mainly the various convolutional Neural nets for realizing target detection, tracking and semantic segmentation
Network, there are commonly SSD, Yolo, DeepLab etc., main customization operations include picture size adaptation, output classification adaptation, mark
Infuse data distribution arrangement, training parameter setting and adjustment etc..
Therefore, the image detection model of corresponding several scenes can be constructed and trained in advance.When above-mentioned steps S106-2 base
After the scenario parameters choose corresponding image detection model in the multiple images detection model pre-established, Ke Yiyou
Image detecting element calls the corresponding image detection model configuration file of the image detection model, runs the image detection model pair
Described image data carry out output test result after vision-based detection.
In practical application, when detecting the scene of current image date based on the scene-detection algorithms that vehicle device is run, just
Corresponding image detection model is selected to be examined according to current scene (combination that can be multiclass), speed and required precision
It surveys.When the scene of continuous image data changes, the switching of image detection model only needs to call not in main program
Same image detection model configuration file and parameter；Image detection model and corresponding parameter under several scenes are to instruct in advance
It perfects and is pre-stored in equipment.
In the selection of image detection model, it can be existed based on the scenario parameters according to preset rules in the embodiment of the present invention
Image detection model is chosen in the multiple images detection model pre-established；And/or the scenario parameters are inputted preset point
Class model, the model and mould of the image detection model by disaggregated model based on the scenario parameters matching detection described image data
Shape parameter.
Wherein, it when carrying out the selection of image detection model according to preset rules, can be based on the scenario parameters pre-
It is chosen in the multiple images detection model first established and the matched image detection model of scene type in the scenario parameters.Letter
For list, it is assumed that detect that, for the rainy day, the model with regard to use for rainy day special training and setting goes to detect；Detect image matter
Situations such as amount is fuzzy, then goes to detect using the model for fuzzy scene picture, backlight, smear is also the same.
For relative complex scene, scenario parameters can be inputted to preset disaggregated model, by the classification mould
The model and model parameter of image detection model of the type based on the scenario parameters matching detection described image data.And herein it
Before, it can first construct disaggregated model；It collects the scenario parameters of different scenes and matches image detection mould corresponding with the scene
Type outputs and inputs instruction using the scenario parameters of each scene and corresponding image detection model as the disaggregated model respectively
Practice the disaggregated model.The model of customization provided in this embodiment may there are many kinds of, such as detect snowy day+high speed+smear+
There is no Small object in the visual field, then the model that will use for this combination customization is detected.
Fig. 2 shows the signals of the image processing method process of the unused resource of combination vehicle device according to the preferred embodiment of the invention
Figure, referring to fig. 2 it is found that method provided in this embodiment may include:
S1, whether the unused resource of engine end periodic detection vehicle device is enough, i.e., whether unused resource, which meets operation, executes scene
The preset condition of detection algorithm；If so, thening follow the steps S202；If it is not, then continuing to repeat the interval detected, detected twice
Time can be judged according to the real-time of scene detection, for example, scene detection function can be in engine end every certain predetermined
It transported within time (30 seconds or one minute, automatically adjusted according to speed, then preset time is short fastly for speed, and then preset time is long slowly for speed)
Row is primary, or interval other times, and the present invention is without limitation.
S2, the Run-time scenario detection algorithm in vehicle device.
In the step, the object of scene-detection algorithms be to camera output image data after, via image detection list
ISP treated image data in member, it is assumed that any one frame or multiple image data are directed to, when needing to carry out to the image data
When scene detection, the image data exported via ISP can be obtained in a manner of wireless connection, or output it by ISP
Image data is transmitted to engine end and carries out scene detection.
S3, when vehicle device executes scene-detection algorithms based on the image data that ISP is exported, as shown in figure 3, can be based on scene
It the parameters such as testing result (i.e. the scene type of image data), requirement of real-time and required precision and existing constructs in advance
Multiple images detection model carries out Data Matching, to choose the image to match with parameter current in multiple scene detection models
Detection model (the model x) in Fig. 3, or the parameters such as scene detection results, requirement of real-time and required precision are inputted in advance
In the disaggregated model first constructed, examined by the final selected image detection model of disaggregated model output, then by selected image
The type transfers of model are surveyed to image detecting element.
Assuming that detect that the corresponding scene type of image data is snowy day, and requirement of real-time and required precision are lower,
The model for snowy day special training and setting can be used to go to detect at this time；Assuming that the image detected is rainy day+high speed+visual field
In only nearby big object can be input in disaggregated model at this time since its parameter is more, by disaggregated model choose correspond to
The image detection model of the carry out vision-based detection of the image data.
S4, image detecting element carries out vision-based detection to current image date using selected model x, to assist carrying out
Automatic parking or safe driving.
In vehicle device operational process, the embodiment of the present invention can be according to scene detection results switching model, model size and essence
Degree, and then achieve the effect that final detection accuracy is promoted.
Based on the same inventive concept, the embodiment of the invention also provides a kind of image procossing of unused resource of combination vehicle device dresses
It sets, as shown in figure 4, the image processing apparatus of the unused resource of combination vehicle device provided in an embodiment of the present invention may include:
It checks module 410, is configured to the use state parameter of current each hardware device in regular check vehicle device, and be based on institute
State the unused resource parameters that use state parameter obtains the vehicle device；
Judgment module 420 is configured to judge whether the unused resource parameters meet the vehicle device operation given scenario inspection
The preset condition of method of determining and calculating；
Detection module 430 is transported in the vehicle device if being configured to the unused resource parameters meets the preset condition
The row scene-detection algorithms, assist preset image detecting element to the picture number for being input to described image detection unit in real time
According to progress vision-based detection.
In an alternate embodiment of the present invention, the detection module 430 is additionally configured to, and acquisition is input to described image detection
Via image processor treated image data in unit, run in the vehicle device described in the scene-detection algorithms identification
The scenario parameters of image data；
Corresponding image detection model is chosen in the multiple images detection model pre-established based on the scenario parameters,
Output detection is tied after carrying out vision-based detection to described image data using described image detection model by described image detection unit
Fruit.
In an alternate embodiment of the present invention, the detection module 430 is additionally configured to, and acquisition is input to described image detection
Via image processor treated image data in unit；
The scene-detection algorithms are run in the vehicle device, scene detection is carried out to described image data, identify described
The size of target detection thing and its quantity, requirement of real-time and/or precision are wanted in the scene type of image data, image data
It asks；
Wherein, the scene type includes type locating for weather pattern and/or the geographical location in described image data.
In an alternate embodiment of the present invention, the detection module 430 is additionally configured to, based on the scenario parameters according to pre-
If rule chooses image detection model in the multiple images detection model pre-established；And/or
The scenario parameters are inputted into preset disaggregated model, scenario parameters matching inspection is based on by the disaggregated model
Survey the model and model parameter of the image detection model of described image data.
In an alternate embodiment of the present invention, the detection module 430 is additionally configured to, based on the scenario parameters preparatory
It is chosen in the multiple images detection model of foundation and the matched image detection model of scene type in the scenario parameters.
In an alternate embodiment of the present invention, as shown in figure 5, described device further include: the first building module 440, configuration
To construct disaggregated model；
It collects the scenario parameters of different scenes and matches image detection model corresponding with the scene, respectively by each scene
Scenario parameters and corresponding image detection model as the disaggregated model output and input the training disaggregated model.
In an alternate embodiment of the present invention, as shown in figure 5, described device further includes, the second building module 450, configuration
For the image detection model for constructing and training corresponding several scenes in advance.
In an alternate embodiment of the present invention, the detection module 430 is additionally configured to, based on the scenario parameters preparatory
Corresponding image detection model is chosen in the multiple images detection model of foundation, and described image is called by described image detection unit
The corresponding image detection model configuration file of detection model, operation described image detection model carry out vision to described image data
Output test result after detection.
Based on the same inventive concept, the embodiment of the invention also provides a kind of computer storage medium, the computer is deposited
Storage media is stored with computer program code, when the computer program code is run on the computing device, leads to the meter
Calculate the image processing method that equipment executes the unused resource of combination vehicle device described in any of the above-described embodiment.
The embodiment of the invention provides a kind of image processing methods of unused resource of combination vehicle device, are checking current vehicle device
In each hardware device have meet preset condition unused resource when, just can in vehicle device Run-time scenario detection algorithm, auxiliary
Preset image detecting element carries out vision-based detection to the image data for being input to the image detecting element in real time.Of the invention real
In the scheme that example offer is provided, using the unused resource of vehicle device, is interconnected, mentioned between vehicle device and preset image detecting element
For assistant images processing function, auxiliary is provided to the correlation detection principle on preset image detecting element, to promote inspection
Save the cost while surveying precision.
In addition, the embodiment of the present invention can also be according to scene detection results switching model, model size and precision, Jin Erda
The effect promoted to final detection accuracy.
It is apparent to those skilled in the art that the specific work of the system of foregoing description, device and unit
Make process, can refer to corresponding processes in the foregoing method embodiment, for brevity, does not repeat separately herein.
In addition, each functional unit in each embodiment of the present invention can be physically independent, can also two or
More than two functional units integrate, and can be all integrated in a processing unit with all functional units.It is above-mentioned integrated
Functional unit both can take the form of hardware realization, can also be realized in the form of software or firmware.
Those of ordinary skill in the art will appreciate that: if the integrated functional unit is realized and is made in the form of software
It is independent product when selling or using, can store in a computer readable storage medium.Based on this understanding,
Technical solution of the present invention is substantially or all or part of the technical solution can be embodied in the form of software products,
The computer software product is stored in a storage medium comprising some instructions, with so that calculating equipment (such as
Personal computer, server or network equipment etc.) various embodiments of the present invention the method is executed when running described instruction
All or part of the steps.And storage medium above-mentioned includes: USB flash disk, mobile hard disk, read-only memory (ROM), random access memory
Device (RAM), the various media that can store program code such as magnetic or disk.
Alternatively, realizing that all or part of the steps of preceding method embodiment can be (all by the relevant hardware of program instruction
Such as personal computer, the calculating equipment of server or network equipment etc.) it completes, described program instruction can store in one
In computer-readable storage medium, when described program instruction is executed by the processor of calculating equipment, the calculating equipment is held
The all or part of the steps of row various embodiments of the present invention the method.
Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention., rather than its limitations；To the greatest extent
Present invention has been described in detail with reference to the aforementioned embodiments for pipe, those skilled in the art should understand that: at this
Within the spirit and principle of invention, it is still possible to modify the technical solutions described in the foregoing embodiments or right
Some or all of the technical features are equivalently replaced；And these are modified or replaceed, and do not make corresponding technical solution de-
From protection scope of the present invention.
Claims (10)
1. a kind of image processing method of the unused resource of combination vehicle device, comprising:
The use state parameter of current each hardware device in regular check vehicle device, and based on described in use state parameter acquisition
The unused resource parameters of vehicle device；
Judge whether the unused resource parameters meet the preset condition of the vehicle device operation given scenario detection algorithm；
If the unused resource parameters meet the preset condition, the scene-detection algorithms are run in the vehicle device, it is auxiliary
Preset image detecting element is helped to carry out vision-based detection to the image data for being input to described image detection unit in real time.
2. described to run the scene-detection algorithms, auxiliary in the vehicle device according to the method described in claim 1, wherein
Preset image detecting element carries out vision-based detection to the image data for being input to described image detection unit in real time, comprising:
Acquisition is input in described image detection unit via image processor treated image data, is transported in the vehicle device
The scenario parameters of the row scene-detection algorithms identification described image data；
Corresponding image detection model is chosen in the multiple images detection model pre-established based on the scenario parameters, by institute
Image detecting element is stated using described image detection model to output test result after described image data progress vision-based detection.
3. according to the method described in claim 2, wherein, the acquisition is input in described image detection unit at via image
Device treated image data is managed, the scene ginseng of the scene-detection algorithms identification described image data is run in the vehicle device
Number, comprising:
Acquisition is input in described image detection unit via image processor treated image data；
The scene-detection algorithms are run in the vehicle device, scene detection is carried out to described image data, identify described image
The size and its quantity, requirement of real-time and/or required precision of target detection thing in the scene types of data, image data；
Wherein, the scene type includes type locating for weather pattern and/or the geographical location in described image data.
4. according to the method described in claim 2, wherein, it is described based on the scenario parameters in the multiple images inspection pre-established
It surveys in model and chooses corresponding image detection model, comprising:
Image detection mould is chosen in the multiple images detection model pre-established according to preset rules based on the scenario parameters
Type；And/or
The scenario parameters are inputted into preset disaggregated model, the scenario parameters matching detection institute is based on by the disaggregated model
State the model and model parameter of the image detection model of image data.
5. described to be pre-established based on the scenario parameters according to preset rules according to the method described in claim 4, wherein
Multiple images detection model in choose image detection model, comprising:
It is chosen in the multiple images detection model pre-established based on the scenario parameters and the scene in the scenario parameters
The image detection model of type matching.
6. it is described that the scenario parameters are inputted into preset disaggregated model according to the method described in claim 4, wherein, by institute
State the model and model parameter of image detection model of the disaggregated model based on the scenario parameters matching detection described image data
Before, further includes:
Construct disaggregated model；
It collects the scenario parameters of different scenes and matches image detection model corresponding with the scene, respectively by the field of each scene
Scape parameter and corresponding image detection model output and input the training disaggregated model as the disaggregated model.
7. according to the described in any item methods of claim 2-6, wherein described more what is pre-established based on the scenario parameters
Before choosing corresponding image detection model in a image detection model, further includes:
The image detection model of corresponding several scenes is constructed and trained in advance.
8. according to the method described in claim 7, wherein, it is described based on the scenario parameters in the multiple images inspection pre-established
It surveys in model and chooses corresponding image detection model, by described image detection unit using described image detection model to the figure
As output test result after data progress vision-based detection, comprising:
Corresponding image detection model is chosen in the multiple images detection model pre-established based on the scenario parameters, by institute
It states image detecting element and calls the corresponding image detection model configuration file of described image detection model, operation described image detection
Model is to output test result after described image data progress vision-based detection.
9. a kind of image processing apparatus of the unused resource of combination vehicle device, comprising:
It checks module, is configured to the use state parameter of current each hardware device in regular check vehicle device, and be based on the use
State parameter obtains the unused resource parameters of the vehicle device；
Judgment module is configured to judge whether the unused resource parameters meet the vehicle device operation given scenario detection algorithm
Preset condition；
Detection module, if being configured to the unused resource parameters meets the preset condition, in the vehicle device described in operation
Scene-detection algorithms assist preset image detecting element to carry out the image data for being input to described image detection unit in real time
Vision-based detection.
10. a kind of computer storage medium, the computer storage medium is stored with computer program code, when the computer
When program code is run on the computing device, the calculating equipment perform claim is caused to require the described in any item combination vehicles of 1-8
The image processing method of the unused resource of machine.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910562835.9A CN110276322B (en) | 2019-06-26 | 2019-06-26 | Image processing method and device combined with vehicle machine idle resources |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910562835.9A CN110276322B (en) | 2019-06-26 | 2019-06-26 | Image processing method and device combined with vehicle machine idle resources |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110276322A true CN110276322A (en) | 2019-09-24 |
| CN110276322B CN110276322B (en) | 2022-01-07 |
Family
ID=67963315
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910562835.9A Active CN110276322B (en) | 2019-06-26 | 2019-06-26 | Image processing method and device combined with vehicle machine idle resources |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110276322B (en) |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103366386A (en) * | 2013-07-14 | 2013-10-23 | 西安电子科技大学 | Parallel image uncompressing system based on multiple processes and multiple threads |
| EP2743861A2 (en) * | 2012-12-12 | 2014-06-18 | Ricoh Company, Ltd. | Method and device for detecting continuous object in disparity direction based on disparity map |
| CN105390021A (en) * | 2015-11-16 | 2016-03-09 | 北京蓝卡科技股份有限公司 | Parking spot state detection method and parking spot state detection device |
| CN106791475A (en) * | 2017-01-23 | 2017-05-31 | 上海兴芯微电子科技有限公司 | Exposure adjustment method and the vehicle mounted imaging apparatus being applicable |
| CN107563512A (en) * | 2017-08-24 | 2018-01-09 | 腾讯科技（上海）有限公司 | A kind of data processing method, device and storage medium |
| CN107563256A (en) * | 2016-06-30 | 2018-01-09 | 北京旷视科技有限公司 | Aid in driving information production method and device, DAS (Driver Assistant System) |
| CN107944375A (en) * | 2017-11-20 | 2018-04-20 | 北京奇虎科技有限公司 | Automatic Pilot processing method and processing device based on scene cut, computing device |
| CN108874615A (en) * | 2017-05-16 | 2018-11-23 | 惠州市德赛西威汽车电子股份有限公司 | A kind of in-vehicle multi-media system device for detecting performance and detection method |
| CN109074069A (en) * | 2016-03-29 | 2018-12-21 | 智动科技有限公司 | Autonomous vehicle with improved vision-based detection ability |
| CN109712431A (en) * | 2017-10-26 | 2019-05-03 | 丰田自动车株式会社 | Drive assistance device and driving assistance system |
| CN109918977A (en) * | 2017-12-13 | 2019-06-21 | 华为技术有限公司 | Determine the method, device and equipment of free time parking stall |
-
2019
- 2019-06-26 CN CN201910562835.9A patent/CN110276322B/en active Active
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2743861A2 (en) * | 2012-12-12 | 2014-06-18 | Ricoh Company, Ltd. | Method and device for detecting continuous object in disparity direction based on disparity map |
| CN103366386A (en) * | 2013-07-14 | 2013-10-23 | 西安电子科技大学 | Parallel image uncompressing system based on multiple processes and multiple threads |
| CN105390021A (en) * | 2015-11-16 | 2016-03-09 | 北京蓝卡科技股份有限公司 | Parking spot state detection method and parking spot state detection device |
| CN109074069A (en) * | 2016-03-29 | 2018-12-21 | 智动科技有限公司 | Autonomous vehicle with improved vision-based detection ability |
| CN107563256A (en) * | 2016-06-30 | 2018-01-09 | 北京旷视科技有限公司 | Aid in driving information production method and device, DAS (Driver Assistant System) |
| CN106791475A (en) * | 2017-01-23 | 2017-05-31 | 上海兴芯微电子科技有限公司 | Exposure adjustment method and the vehicle mounted imaging apparatus being applicable |
| CN108874615A (en) * | 2017-05-16 | 2018-11-23 | 惠州市德赛西威汽车电子股份有限公司 | A kind of in-vehicle multi-media system device for detecting performance and detection method |
| CN107563512A (en) * | 2017-08-24 | 2018-01-09 | 腾讯科技（上海）有限公司 | A kind of data processing method, device and storage medium |
| CN109712431A (en) * | 2017-10-26 | 2019-05-03 | 丰田自动车株式会社 | Drive assistance device and driving assistance system |
| CN107944375A (en) * | 2017-11-20 | 2018-04-20 | 北京奇虎科技有限公司 | Automatic Pilot processing method and processing device based on scene cut, computing device |
| CN109918977A (en) * | 2017-12-13 | 2019-06-21 | 华为技术有限公司 | Determine the method, device and equipment of free time parking stall |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110276322B (en) | 2022-01-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20230096020A1 (en) | Method and system for on-the-fly object labeling via cross modality validation in autonomous driving vehicles | |
| CN110059608B (en) | Object detection method and device, electronic equipment and storage medium | |
| US11790551B2 (en) | Method and system for object centric stereo in autonomous driving vehicles | |
| US20230140540A1 (en) | Method and system for distributed learning and adaptation in autonomous driving vehicles | |
| JP6565967B2 (en) | Road obstacle detection device, method, and program | |
| CN108345875B (en) | Driving region detection model training method, detection method and device | |
| Tan et al. | Color model-based real-time learning for road following | |
| CN110147706A (en) | The recognition methods of barrier and device, storage medium, electronic device | |
| Rachmadi et al. | Adaptive traffic signal control system using camera sensor and embedded system | |
| CN110717918B (en) | Pedestrian detection method and device | |
| WO2020007589A1 (en) | Training a deep convolutional neural network for individual routes | |
| CN113907663A (en) | Obstacle map construction method, cleaning robot and storage medium | |
| CN114694123B (en) | Traffic signal lamp sensing method, device, equipment and storage medium | |
| CN112465868A (en) | Target detection tracking method and device, storage medium and electronic device | |
| US20200097005A1 (en) | Object detection device, object detection method, and vehicle controller | |
| CN110909656A (en) | Pedestrian detection method and system with integration of radar and camera | |
| CN106408593A (en) | Video-based vehicle tracking method and device | |
| CN110276322A (en) | A kind of image processing method and device of the unused resource of combination vehicle device | |
| Bernuy et al. | Adaptive and real-time unpaved road segmentation using color histograms and RANSAC | |
| CN114972731A (en) | Traffic light detection and identification method and device, moving tool and storage medium | |
| CN115240168A (en) | Perception result obtaining method and device, computer equipment and storage medium | |
| CN110177222A (en) | A kind of the camera exposure parameter method of adjustment and device of the unused resource of combination vehicle device | |
| CN116245906A (en) | Mapping anchor point calibration method and system based on moving target and electronic equipment | |
| CN115683125A (en) | Method, system and computer program product for automatically locating a vehicle | |
| CN117576926A (en) | Method, device and storage medium for detecting vehicle violations |
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 | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20220323 Address after: 430051 No. b1336, chuanggu startup area, taizihu cultural Digital Creative Industry Park, No. 18, Shenlong Avenue, Wuhan Economic and Technological Development Zone, Wuhan, Hubei Province Patentee after: Yikatong (Hubei) Technology Co.,Ltd. Address before: No.c101, chuanggu start up area, taizihu cultural Digital Industrial Park, No.18 Shenlong Avenue, Wuhan Economic Development Zone, Hubei Province Patentee before: HUBEI ECARX TECHNOLOGY Co.,Ltd. |
|
| TR01 | Transfer of patent right |