CN108446618A - Car damage identification method, device, electronic equipment and storage medium - Google Patents
Car damage identification method, device, electronic equipment and storage medium Download PDFInfo
- Publication number
- CN108446618A CN108446618A CN201810196561.1A CN201810196561A CN108446618A CN 108446618 A CN108446618 A CN 108446618A CN 201810196561 A CN201810196561 A CN 201810196561A CN 108446618 A CN108446618 A CN 108446618A
- Authority
- CN
- China
- Prior art keywords
- vehicle
- setting loss
- picture
- damage identification
- confidence level
- 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.)
- Pending
Links
Classifications
-
- 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
- G06V20/58—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads
- G06V20/584—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads of vehicle lights or traffic lights
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/20—Administration of product repair or maintenance
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/24323—Tree-organised classifiers
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/08—Insurance
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Strategic Management (AREA)
- Accounting & Taxation (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- Finance (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Economics (AREA)
- Human Resources & Organizations (AREA)
- Molecular Biology (AREA)
- Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Mathematical Physics (AREA)
- Computing Systems (AREA)
- General Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Operations Research (AREA)
- Tourism & Hospitality (AREA)
- Quality & Reliability (AREA)
- Entrepreneurship & Innovation (AREA)
- Development Economics (AREA)
- Technology Law (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
- Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)
Abstract
The present invention provides a kind of car damage identification method, the method includes:Obtain vehicle pictures;Vehicle pictures are input in the parted pattern of trained vehicle position, each position picture of vehicle is partitioned into;Each position picture is input in trained vehicle site tissue damage identification model, identifies the confidence level of the machine setting loss result at each position and the machine setting loss result at each position of output;Confidence level is sent on the user equipment of setting loss personnel less than or equal to the position picture of threshold value so that setting loss personnel carry out setting loss to the position picture, and determine final setting loss result of the confidence level less than the position of threshold value;Position picture by confidence level less than threshold value is added in the training sample of vehicle site tissue damage identification model, re -training vehicle site tissue damage identification model.The present invention also provides a kind of electronic equipment and storage mediums.The present invention can make vehicle site tissue damage identification model be strengthened study, by continually strengthening and updating, improve Model Identification accuracy rate.
Description
Technical field
The present invention relates to artificial intelligence fields more particularly to a kind of car damage identification method, device, electronic equipment and storage to be situated between
Matter.
Background technology
Artificial intelligence technology has been widely used at present in various scenes, such as in vehicle insurance, can be by artificial
Intelligent image identification technology judges deformation and the degree of injury of vehicle, reduces artificial prospecting cost, evades human factor risk.
Machine learning in the prior art has been able to help people's processing and solves most challenge.But it is sharp in the prior art
With the method for machine learning, accuracy of identification is not high, however it remains larger challenge.
Invention content
In view of the foregoing, it is necessary to a kind of car damage identification method, device, electronic equipment and storage medium are provided, can be led to
The thought using HITL manpower interventions is crossed, so that vehicle site tissue damage identification model is strengthened study, forms the adaptive of model
Effect improves vehicle site tissue damage identification model Model Identification accuracy rate by continually strengthening and updating.
A kind of car damage identification method, the method includes:
Obtain vehicle pictures;
Vehicle pictures are input in the parted pattern of trained vehicle position, each position picture of vehicle is partitioned into;
Each position picture is input in trained vehicle site tissue damage identification model, identifies the machine at each position
The confidence level of setting loss result and the machine setting loss result at each position of output;
Position picture by confidence level less than or equal to threshold value is sent on the user equipment of setting loss personnel so that setting loss people
Member carries out setting loss to the position picture, and determines final setting loss result of the confidence level less than the position of threshold value;
Position picture by confidence level less than threshold value is added in the training sample of vehicle site tissue damage identification model, again
Training vehicle site tissue damage identification model.
According to the preferred embodiment of the present invention, in vehicle pictures to be input to trained vehicle position parted pattern, point
Before cutting out each position picture of vehicle, the method further includes:
The vehicle pictures of acquisition are detected, judge whether the vehicle pictures obtained qualified, the content of detection include with
The combination of lower one or more:Picture clarity, shooting angle, the recognizable degree for shooting position, picture, which whether there is, usurps
Change suspicion;
When the vehicle pictures of the acquisition are unqualified, user is prompted to upload vehicle pictures again.
According to the preferred embodiment of the present invention, mould is identified each position picture is input to trained vehicle site tissue damage
Before type, the method further includes:
Car plate position, VIN codes position are identified from vehicle pictures, license plate number is identified from car plate position, from VIN codes
Whether VIN codes are identified at position, be vehicle of insuring using license plate number or VIN code identifications vehicle, when for insure vehicle when, judge
The degree of injury of each position picture of vehicle.
According to the preferred embodiment of the present invention, the method further includes:When being sent to for the position that confidence level is less than to threshold value
When setting loss personnel, user waiting prompt setting loss result.
According to the preferred embodiment of the present invention, the method further includes:Machine setting loss result when the position and artificial setting loss
As a result when different, the final setting loss result using artificial setting loss result as confidence level less than the position of threshold value.
According to the preferred embodiment of the present invention, the method further includes:Machine setting loss by confidence level higher than the position of threshold value
As a result the final setting loss result as confidence level higher than the position of threshold value.
According to the preferred embodiment of the present invention, the method further includes:
When vehicle is to insure vehicle, the vehicle pictures are input in trained vehicle cab recognition model, vehicle is exported
Brand and vehicle;
According to the final setting loss at the brand of vehicle and vehicle and each position of vehicle as a result, determining each portion of vehicle
The mantenance data of position;
According to the mantenance data at each position of vehicle, the maintenance cost of vehicle is calculated, and the user for being sent to user sets
It is standby.
According to the preferred embodiment of the present invention, the method further includes:The insurance data for obtaining vehicle, according to insurance data and
Maintenance cost determines Claims Resolution data, and data of settling a claim are sent to the equipment of the user of the vehicle so that user checks.
A kind of car damage identification device, described device include:
Acquisition module, for obtaining vehicle pictures;
Segmentation module is partitioned into vehicle for vehicle pictures to be input in the parted pattern of trained vehicle position
Each position picture;
Identification module is identified for each position picture to be input in trained vehicle site tissue damage identification model
The confidence level of the machine setting loss result at each position and the machine setting loss result at each position of output;
Sending module, the user equipment for confidence level to be sent to setting loss personnel less than or equal to the position picture of threshold value
Above so that setting loss personnel carry out setting loss to the position picture, and determine final setting loss result of the confidence level less than the position of threshold value;
Training module, the instruction for confidence level to be added to vehicle site tissue damage identification model less than the position picture of threshold value
Practice in sample, re -training vehicle site tissue damage identification model.
A kind of electronic equipment, the electronic equipment include memory and processor, and the memory is for storing at least one
A instruction, the processor is for executing at least one instruction to realize any one of any embodiment car damage identification
Method.
A kind of computer readable storage medium, the computer-readable recording medium storage has at least one instruction, described
Any one of any embodiment car damage identification method is realized at least one instruction when being executed by processor.
As can be seen from the above technical solutions, the present invention obtains vehicle pictures;Vehicle pictures are input to trained vehicle
In position parted pattern, it is partitioned into each position picture of vehicle;Each position picture is input to trained vehicle portion
Bit loss is hindered in identification model, identifies the confidence of the machine setting loss result at each position and the machine setting loss result at each position of output
Degree;Position picture by confidence level less than or equal to threshold value is sent on the user equipment of setting loss personnel so that setting loss personnel are to this
Position picture carries out setting loss, and determines final setting loss result of the confidence level less than the position of threshold value;By confidence level less than threshold value
Position picture is added in the training sample of vehicle site tissue damage identification model, re -training vehicle site tissue damage identification model.
The present invention can make vehicle site tissue damage identification model be strengthened study by using the thought of HITL manpower interventions, form mould
The adaptive effect of type improves vehicle site tissue damage identification model Model Identification accuracy rate by continually strengthening and updating.
Description of the drawings
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technology description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
The embodiment of invention for those of ordinary skill in the art without creative efforts, can also basis
The attached drawing of offer obtains other attached drawings.
Fig. 1 is the flow chart of the first preferred embodiment of car damage identification method of the present invention.
Fig. 2 is the flow chart of the second preferred embodiment of car damage identification method of the present invention.
Fig. 3 is the Program modual graph of the first preferred embodiment of car damage identification device of the present invention.
Fig. 4 is the Program modual graph of the second preferred embodiment of car damage identification device of the present invention.
Fig. 5 is the structural schematic diagram of the preferred embodiment of electronic equipment at least one example of the present invention.
Specific implementation mode
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation describes, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts every other
Embodiment shall fall within the protection scope of the present invention.
In order to make the foregoing objectives, features and advantages of the present invention clearer and more comprehensible, below in conjunction with the accompanying drawings and specific real
Applying mode, the present invention is described in further detail.
In order to enable those skilled in the art to better understand the solution of the present invention, below in conjunction in the embodiment of the present invention
Attached drawing, technical scheme in the embodiment of the invention is clearly and completely described, it is clear that described embodiment is only
The embodiment of a part of the invention, instead of all the embodiments.Based on the embodiments of the present invention, ordinary skill people
The every other embodiment that member is obtained without making creative work should all belong to the model that the present invention protects
It encloses.
Term " first ", " second " and " third " in description and claims of this specification and above-mentioned attached drawing etc. is
For distinguishing different objects, not for description particular order.In addition, term " comprising " and their any deformations, it is intended that
Non-exclusive include in covering.Such as process, method, system, product or the equipment for containing series of steps or unit do not have
It is defined in the step of having listed or unit, but further includes the steps that optionally not listing or unit, or further include optionally
For the intrinsic other steps of these processes, method, product or equipment or unit.
As shown in Figure 1, being the flow chart of the first preferred embodiment of car damage identification method of the present invention.According to different need
It asks, the sequence of step can change in the flow chart, and certain steps can be omitted.
S10, electronic equipment obtain vehicle pictures.
In an alternative embodiment, it takes pictures to the vehicle in the scene of the accident, the vehicle pictures of shooting is sent to high in the clouds,
The vehicle pictures include vehicle panoramic picture, vehicle sections picture etc..
Preferably, the vehicle pictures of acquisition are detected, and judge whether the vehicle pictures obtained are qualified, obtained when described
When the vehicle pictures taken are unqualified, user is prompted to upload vehicle pictures again.Detection content to the picture of the acquisition includes
Picture clarity, shooting angle shoot the recognizable degree at position, picture with the presence or absence of distorting suspicion etc., such as picture is clear
Whether degree is within the scope of the clarity of configuration, and whether shooting angle is in the angular range of configuration, the recognizable journey at shooting position
Whether degree is in the recognizable extent and scope of configuration etc..If when the vehicle pictures qualification obtained, executing S11.It can keep away in this way
Exempt from underproof picture influences caused by subsequent setting loss result, to improve the accuracy of car damage identification.
Vehicle pictures are input in the parted pattern of trained vehicle position by S11, the electronic equipment, are partitioned into vehicle
Each position picture.
In optional implementation, vehicle position parted pattern is used for each position picture of dividing vehicle.The vehicle
The training sample of position parted pattern includes the picture at each position of vehicle, such as door handle, car door, tire, etc..In training
During the parted pattern of vehicle position, vehicle position parted pattern constantly learns the feature at each position of vehicle.Work as vehicle
It after position parted pattern trains, can be identified from vehicle pictures, and be partitioned into each position picture of vehicle, convenient for follow-up
Judge the degree of injury of each position picture.
Preferably, car plate position, VIN codes position are also identified from vehicle pictures, and car plate is identified from car plate position
Number, it identifies VIN codes from VIN codes position, whether is vehicle of insuring using license plate number or VIN code identifications vehicle, when to insure
When vehicle, the degree of injury of each position picture of vehicle is judged.
Optionally, include in the training process of vehicle position parted pattern:
A, configure each position (for example, door handle, car door, tire, left front door, right front door, lobus sinister daughter board, lobus dexter daughter board,
Front bumper, rear bumper etc.) corresponding preset quantity samples pictures;
B, each samples pictures are subjected to picture pretreatment and train picture to obtain training vehicle position parted pattern,
Such as can such as be scaled, after cutting, overturning and/or distortion operation by carrying out picture pretreatment to each samples pictures, make instruction
Practice picture be of the same size and identical visual angle after, just progress model training, to effectively improve the authenticity of model training
And accuracy rate.
C, all trained pictures are divided into the training set of the first ratio (for example, 70%), the second ratio (for example, 30%)
Verification collection;
D, vehicle position parted pattern is trained using the training set;
E, using the accuracy rate of the vehicle position parted pattern of the verification collection verification training, if accuracy rate is more than or waits
In default accuracy rate, then training terminates, if alternatively, accuracy rate increases the corresponding sample in each position less than default accuracy rate
Picture
Quantity simultaneously re-executes above-mentioned steps B, C, D, E, until the accuracy rate of the vehicle position parted pattern of training is more than
Or equal to default accuracy rate.
Each position picture is input in trained vehicle site tissue damage identification model by S12, the electronic equipment, is known
The confidence level of the machine setting loss result at not each position and the machine setting loss result at each position of output.
In an alternative embodiment, the vehicle site tissue damage identification model is used to judge the degree of injury at each position, and
Export the confidence level of the machine setting loss result at each position.The training sample of training vehicle site tissue damage identification model includes each
The picture of the various degree of injury at position.During training vehicle site tissue damage identification model, the identification of vehicle site tissue damage
Model constantly learns the feature of the various degree of injury at each position.It, can be right after vehicle site tissue damage identification model trains
Each position picture carries out setting loss, judges the degree of injury at each position, and export the confidence level of the setting loss result at each position.
Subsequently confidence level can be issued multiple professionals less than the machine setting loss result of threshold value and carry out setting loss.
Preferably, the method further includes:Machine setting loss result using confidence level higher than the position of threshold value is as confidence level
Higher than the final setting loss result at the position of threshold value.
Confidence level is sent to the user of setting loss personnel by S13, the electronic equipment less than or equal to the position picture of threshold value
So that setting loss personnel carry out setting loss to the position picture in equipment, and determine final setting loss knot of the confidence level less than the position of threshold value
Fruit.
In an alternative embodiment, setting loss personnel are one or more.Using more than the setting loss result of default number as people
Work setting loss result.For example, the position picture by confidence level less than threshold value is sent to 5 setting loss personnel, there are 4 setting loss personnel judgements
For level-one degree of injury, 1 setting loss personnel is determined as secondary damage degree, then artificial setting loss result is level-one degree of injury.This
Confidence level is sent to multiple setting loss personnel so that setting loss personnel are to car damage identification, and use big by sample less than the position picture of threshold value
The identical setting loss result of part setting loss personnel is as artificial setting loss as a result, it is possible to prevente effectively from Human disturbance factor.
Preferably, the method further includes:When being sent to setting loss personnel of position that confidence level is less than to threshold value, prompt
User waits for setting loss result.
Preferably, the method further includes:When the machine setting loss result at the position and artificial setting loss result difference, by people
Final setting loss result of the work setting loss result as confidence level less than the position of threshold value.It in this way can be with the setting loss knot of manual intervention vehicle
Fruit improves setting loss precision.
The position picture of S14, the electronic equipment by confidence level less than threshold value is added to vehicle site tissue damage identification model
Training sample in re -training vehicle site tissue damage identification model.
In an alternative embodiment, the position picture by the confidence level less than threshold value is updated to vehicle site tissue damage identification model
Training sample in.For example, the machine setting loss result of the position picture is secondary damage, artificial setting loss result is damaged for level-one,
Then the model parameter of secondary damage needs to strengthen in vehicle site tissue damage identification model, needs the two level at the more positions of study
The feature of damage, the position picture by confidence level less than threshold value is added in secondary damage classification, as secondary damage classification
Training sample.The sample of the classification of machine learning algorithm decision error can be increased in this way, and mould is identified to vehicle site tissue damage
Type re -training makes the feature of the sample of the classification of vehicle site tissue damage identification model study decision error, to make vehicle portion
The model parameter that bit loss hinders identification model can accurately subsequently judge the vehicle damage degree of the wrongheaded classification.
Optionally, include in the training process of the vehicle site tissue damage identification model:
A, (for example, for door handle position, level-one damage journey is respectively configured in the various degree of injury for configuring each position
The various degree of injury of degree, secondary damage degree, three-level degree of injury etc.) corresponding preset quantity samples pictures;
B, each samples pictures are subjected to picture pretreatment to obtain the instruction of the training vehicle site tissue damage identification model
Practice picture, such as can be by carrying out picture pretreatment such as scaling, cutting, overturning and/or distortion operation to each samples pictures
Afterwards, make train picture be of the same size and identical visual angle after, just progress model training, to effectively improve model training
Authenticity and accuracy rate.
C, all trained pictures are divided into the training set of the first ratio (for example, 80%), the second ratio (for example, 20%)
Verification collection;
D, the vehicle site tissue damage identification model is trained using the training set;
E, using it is described verification collection verification training vehicle site tissue damage identification model accuracy rate, if accuracy rate be more than or
Person is equal to default accuracy rate, then training terminates, if alternatively, it is corresponding to increase each position less than default accuracy rate for accuracy rate
Samples pictures;
Quantity simultaneously re-executes above-mentioned steps B, C, D, E, until the accuracy rate of the vehicle site tissue damage identification model of training
More than or equal to default accuracy rate.
In an alternative embodiment, vehicle position parted pattern, the vehicle site tissue damage identification model and the vehicle
Type identification model can be for without the depth convolutional neural networks model of full articulamentum, the depth convolutional neural networks mould
Type includes input layer, convolutional layer, pond layer, up-samples layer and cut layer, in a kind of specific embodiment, the depth volume
Product neural network model is by 1 input layer, and 16 convolutional layers, 5 pond layers, 1 up-sampling layer, 1 cuts layer composition.
Due in traditional classification problem, generally requiring each pictures are exported with full articulamentum belong to each class
Probability, however in semantic segmentation problem, to predict which class to be belonged to by each sample, to inevitably result in efficiency low in this way
Under.Therefore, the identification model in the present embodiment is the depth convolutional neural networks model without full articulamentum, the depth convolution
Neural network model only need to export the classification score of each pixel with a convolutional layer on Conv8.On the layer, feature
Point has the score of different classifications in class num+1 classification each of on figure, therefore the channel exported is also class
Num+1, recognition efficiency greatly improve.
By above-mentioned implementation, confidence level can be added to the identification of vehicle site tissue damage by the present invention less than the position picture of threshold value
Re -training vehicle site tissue damage identification model in the training sample of model makes vehicle site tissue damage identification model be strengthened
It practises, forms the adaptive effect of model, by continually strengthening and updating, it is accurate to improve vehicle site tissue damage identification model Model Identification
True rate.
As shown in Fig. 2, being the flow chart of the second preferred embodiment of car damage identification method of the present invention.According to different need
It asks, the sequence of step can change in the flow chart, and certain steps can be omitted.
S20 to S24 is corresponding with the S10 to S14 in the first preferred embodiment respectively, and this will not be detailed here.
S25, when vehicle is to insure vehicle, the vehicle pictures are input to trained vehicle by the electronic equipment to be known
In other model, the brand and vehicle of vehicle are exported.Preferred embodiment, using the panoramic pictures of the vehicle pictures as the vehicle
The input of type identification model.In training vehicle cab recognition model, the training sample of vehicle cab recognition model is various brand vehicles
The panoramic pictures of a variety of models.After vehicle cab recognition model training is good, vehicle in the panoramic pictures of energy automatic identification input
Brand and vehicle, training algorithm are the prior art, are included, but are not limited to:Convolutional neural networks model.
Optionally, include in the training process of the vehicle cab recognition model:
A, various brand a variety of models are configured (for example, the panoramic pictures of Audi Q5, the panoramic pictures of Audi A3, C grades of benz
Panoramic pictures, E grades of the panoramic pictures etc. of running quickly) corresponding preset quantity samples pictures;
B, each samples pictures are subjected to picture pretreatment and train picture to obtain the training vehicle cab recognition model,
Such as can such as be scaled, after cutting, overturning and/or distortion operation by carrying out picture pretreatment to each samples pictures, make instruction
Practice picture be of the same size and identical visual angle after, just progress model training, to effectively improve the authenticity of model training
And accuracy rate.
C, all trained pictures are divided into the training set of the first ratio (for example, 85%), the second ratio (for example, 15%)
Verification collection;
D, the vehicle cab recognition model is trained using the training set;
E, using the accuracy rate of the vehicle cab recognition model of the verification collection verification training, if accuracy rate is more than or waits
In default accuracy rate, then training terminates, if alternatively, accuracy rate increases the corresponding sample in each position less than default accuracy rate
Picture;
Quantity simultaneously re-executes above-mentioned steps B, C, D, E, until training vehicle cab recognition model accuracy rate be more than or
Equal to default accuracy rate.
S26, the electronic equipment according to the final setting loss at the brand and vehicle of vehicle and each position of vehicle as a result,
Determine the mantenance data at each position of vehicle.
In an alternative embodiment, the mantenance data includes the price data of aftermarket attachment, maintenance man-hours cost data.Respectively
The different degree of injury at a position correspond to different aftermarket attachments and different maintenance man-hours.The brand is read from database
With the mantenance data at each position of the vehicle of vehicle.When each position of the not no vehicle of the brand and vehicle of the database
Mantenance data when, to supplier send inquiry instruction to inquire the repair number at each position of the vehicle of the brand and vehicle
According to.
S27, the electronic equipment calculate the maintenance cost of vehicle, concurrently according to the mantenance data at each position of vehicle
Give the user equipment of user.
In an alternative embodiment, by the price data of the aftermarket attachment at each position of vehicle, maintenance man-hours cost data
It adds up, the maintenance cost as vehicle.
In an alternative embodiment, the method further includes:The insurance data for obtaining vehicle, according to insurance data and maintenance cost
With data of settling a claim are sent to the equipment of the user of the vehicle so that user checks by determining Claims Resolution data.
By the way that in above-described embodiment, the present invention accurately can carry out setting loss to damage vehicle, and according to the setting loss situation of vehicle
And insurance data, the maintenance cost and Claims Resolution data of vehicle are calculated, Claims Resolution efficiency and the transparence of Claims Resolution are improved.
As shown in figure 3, the Program modual graph of the first preferred embodiment of car damage identification device of the present invention.The car damage identification
Device 3 includes, but are not limited to one or more following module:Acquisition module 30, segmentation module 31, identification module 32, training
Module 33, determining module 34, sending module 35 and reminding module 36.The so-called unit of the present invention, which refers to one kind, to be determined by vehicle
The processor of damage device 3 is performed and can complete the series of computation machine program segment of fixed function, is stored in memory
In.Function about each unit will be described in detail in subsequent embodiment.
The acquisition module 30 obtains vehicle pictures.
In an alternative embodiment, it takes pictures to the vehicle in the scene of the accident, the vehicle pictures of shooting is sent to high in the clouds,
The vehicle pictures include vehicle panoramic picture, vehicle sections picture etc..
Preferably, the acquisition module 30 is additionally operable to:The vehicle pictures of acquisition are detected, and judge the vehicle obtained
Whether picture is qualified, when the vehicle pictures of the acquisition are unqualified, user is prompted to upload vehicle pictures again.To the acquisition
Picture detection content include picture clarity, shooting angle, shoot the recognizable degree at position, picture is with the presence or absence of distorting
Suspicion etc., such as whether picture clarity within the scope of the clarity of configuration, shooting angle whether in the angular range of configuration,
The recognizable degree at position is shot whether in the recognizable extent and scope of configuration etc..If when the vehicle pictures qualification obtained,
Execute segmentation module 31.It can be influenced caused by subsequent setting loss result to avoid underproof picture in this way, to improve vehicle
The accuracy of setting loss.
Vehicle pictures are input in the parted pattern of trained vehicle position by the segmentation module 31, are partitioned into vehicle
Each position picture.
In optional implementation, vehicle position parted pattern is used for each position picture of dividing vehicle.The vehicle
The training sample of position parted pattern includes the picture at each position of vehicle, such as door handle, car door, tire, etc..In training
During the parted pattern of vehicle position, vehicle position parted pattern constantly learns the feature at each position of vehicle.Work as vehicle
It after position parted pattern trains, can be identified from vehicle pictures, and be partitioned into each position picture of vehicle, convenient for follow-up
Judge the degree of injury of each position picture.
Preferably, the segmentation module 31 is additionally operable to:Car plate position, VIN codes position are identified from vehicle pictures, from vehicle
Board identifies license plate number in position, and VIN codes are identified from VIN codes position, using license plate number or VIN code identifications vehicle whether be
Insure vehicle, when for insure vehicle when, judge the degree of injury of each position picture of vehicle.
Optionally, training module 33 includes in the training process of vehicle position parted pattern:
A, configure each position (for example, door handle, car door, tire, left front door, right front door, lobus sinister daughter board, lobus dexter daughter board,
Front bumper, rear bumper etc.) corresponding preset quantity samples pictures;
B, each samples pictures are subjected to picture pretreatment and train picture to obtain training vehicle position parted pattern,
Such as can such as be scaled, after cutting, overturning and/or distortion operation by carrying out picture pretreatment to each samples pictures, make instruction
Practice picture be of the same size and identical visual angle after, just progress model training, to effectively improve the authenticity of model training
And accuracy rate.
C, all trained pictures are divided into the training set of the first ratio (for example, 70%), the second ratio (for example, 30%)
Verification collection;
D, vehicle position parted pattern is trained using the training set;
E, using the accuracy rate of the vehicle position parted pattern of the verification collection verification training, if accuracy rate is more than or waits
In default accuracy rate, then training terminates, if alternatively, accuracy rate increases the corresponding sample in each position less than default accuracy rate
Picture;
Quantity simultaneously re-executes above-mentioned steps B, C, D, E, until the accuracy rate of the vehicle position parted pattern of training is more than
Or equal to default accuracy rate.
Each position picture is input to trained vehicle site tissue damage and identifies mould by electronic equipment described in identification module 32
In type, the confidence level of the machine setting loss result at each position and the machine setting loss result at each position of output is identified.
In an alternative embodiment, the vehicle site tissue damage identification model is used to judge the degree of injury at each position, and
Export the confidence level of the machine setting loss result at each position.The training sample of training vehicle site tissue damage identification model includes each
The picture of the various degree of injury at position.During training vehicle site tissue damage identification model, the identification of vehicle site tissue damage
Model constantly learns the feature of the various degree of injury at each position.It, can be right after vehicle site tissue damage identification model trains
Each position picture carries out setting loss, judges the degree of injury at each position, and export the confidence level of the setting loss result at each position.
Subsequently confidence level can be issued multiple professionals less than the machine setting loss result of threshold value and carry out setting loss.
Preferably, the determining module 34 is used for:Machine setting loss result using confidence level higher than the position of threshold value is as setting
Final setting loss result of the reliability higher than the position of threshold value.
The user that confidence level is sent to setting loss personnel by the sending module 35 less than or equal to the position picture of threshold value sets
For upper so that setting loss personnel carry out setting loss to the position picture, and determining confidence level is less than the final setting loss knot at the position of threshold value
Fruit.
In an alternative embodiment, setting loss personnel are one or more.Using more than the setting loss result of default number as people
Work setting loss result.For example, the position picture by confidence level less than threshold value is sent to 5 setting loss personnel, there are 4 setting loss personnel judgements
For level-one degree of injury, 1 setting loss personnel is determined as secondary damage degree, then artificial setting loss result is level-one degree of injury.This
Confidence level is sent to multiple setting loss personnel so that setting loss personnel are to car damage identification, and use big by sample less than the position picture of threshold value
The identical setting loss result of part setting loss personnel is as artificial setting loss as a result, it is possible to prevente effectively from Human disturbance factor.
Preferably, the reminding module 36 is used for:When by confidence level be less than threshold value when being sent to setting loss personnel of position,
User waiting prompt setting loss result.
Preferably, the determining module 34 is used for:When the machine setting loss result at the position and artificial setting loss result difference,
Final setting loss result using artificial setting loss result as confidence level less than the position of threshold value.It can be determined in this way with manual intervention vehicle
Damage is as a result, improve setting loss precision.
Position picture of the training module 33 by confidence level less than threshold value is added to vehicle site tissue damage identification model
Re -training vehicle site tissue damage identification model in training sample.
In an alternative embodiment, position picture of the training module 33 by the confidence level less than threshold value is updated to vehicle portion
Bit loss is hindered in the training sample of identification model.For example, the machine setting loss result of the position picture is secondary damage, artificial setting loss knot
Fruit damages for level-one, then the model parameter of secondary damage needs to strengthen in vehicle site tissue damage identification model, needs study more
The position secondary damage feature, confidence level is added to less than the position picture of threshold value in secondary damage classification, as
The training sample of secondary damage classification.The sample of the classification of machine learning algorithm decision error can be increased in this way, and to vehicle
Site tissue damage identification model re -training makes the spy of the sample of the classification of vehicle site tissue damage identification model study decision error
Sign, to make the model parameter of vehicle site tissue damage identification model that can accurately subsequently judge the vehicle of the wrongheaded classification
Degree of injury.
Optionally, the training module 33 includes in the training process of the vehicle site tissue damage identification model:
A, (for example, for door handle position, level-one damage journey is respectively configured in the various degree of injury for configuring each position
The various degree of injury of degree, secondary damage degree, three-level degree of injury etc.) corresponding preset quantity samples pictures;
B, each samples pictures are subjected to picture pretreatment to obtain the instruction of the training vehicle site tissue damage identification model
Practice picture, such as can be by carrying out picture pretreatment such as scaling, cutting, overturning and/or distortion operation to each samples pictures
Afterwards, make train picture be of the same size and identical visual angle after, just progress model training, to effectively improve model training
Authenticity and accuracy rate.
C, all trained pictures are divided into the training set of the first ratio (for example, 80%), the second ratio (for example, 20%)
Verification collection;
D, the vehicle site tissue damage identification model is trained using the training set;
E, using it is described verification collection verification training vehicle site tissue damage identification model accuracy rate, if accuracy rate be more than or
Person is equal to default accuracy rate, then training terminates, if alternatively, it is corresponding to increase each position less than default accuracy rate for accuracy rate
Samples pictures;
Quantity simultaneously re-executes above-mentioned steps B, C, D, E, until the accuracy rate of the vehicle site tissue damage identification model of training
More than or equal to default accuracy rate.
In an alternative embodiment, vehicle position parted pattern, the vehicle site tissue damage identification model and the vehicle
Type identification model can be for without the depth convolutional neural networks model of full articulamentum, the depth convolutional neural networks mould
Type includes input layer, convolutional layer, pond layer, up-samples layer and cut layer, in a kind of specific embodiment, the depth volume
Product neural network model is by 1 input layer, and 16 convolutional layers, 5 pond layers, 1 up-sampling layer, 1 cuts layer composition.
Due in traditional classification problem, generally requiring each pictures are exported with full articulamentum belong to each class
Probability, however in semantic segmentation problem, to predict which class to be belonged to by each sample, to inevitably result in efficiency low in this way
Under.Therefore, the identification model in the present embodiment is the depth convolutional neural networks model without full articulamentum, the depth convolution
Neural network model only need to export the classification score of each pixel with a convolutional layer on Conv8.On the layer, feature
Point has the score of different classifications in class num+1 classification each of on figure, therefore the channel exported is also class
Num+1, recognition efficiency greatly improve.
By above-mentioned implementation, confidence level can be added to the identification of vehicle site tissue damage by the present invention less than the position picture of threshold value
Re -training vehicle site tissue damage identification model in the training sample of model makes vehicle site tissue damage identification model be strengthened
It practises, forms the adaptive effect of model, by continually strengthening and updating, it is accurate to improve vehicle site tissue damage identification model Model Identification
True rate.
As shown in figure 4, the Program modual graph of the second preferred embodiment of car damage identification device of the present invention.The car damage identification
Device 3 is in addition to including one or more module in first preferably implementation:Acquisition module 30, segmentation module 31, identification module
32, except training module 33, determining module 34, sending module 35 and reminding module 36, the car damage identification device 3 can also wrap
Include one or more following module:Output module 37 and computing module 38.The so-called unit of the present invention refers to that one kind can be by
The processor of car damage identification device 3 is performed and can complete the series of computation machine program segment of fixed function, is stored in
In memory.Function about each unit will be described in detail in subsequent embodiment.
When vehicle is to insure vehicle, the vehicle pictures are input to trained vehicle cab recognition by the output module 37
In model, the brand and vehicle of vehicle are exported.Preferred embodiment, using the panoramic pictures of the vehicle pictures as the vehicle
The input of identification model.In training vehicle cab recognition model, the training sample of vehicle cab recognition model is each of various brand vehicles
The panoramic pictures of kind vehicle.After vehicle cab recognition model training is good, the product of vehicle in the panoramic pictures of energy automatic identification input
Board and vehicle, training algorithm are the prior art, are included, but are not limited to:Convolutional neural networks model.
Optionally, the training module 33 includes in the training process of the vehicle cab recognition model:
A, various brand a variety of models are configured (for example, the panoramic pictures of Audi Q5, the panoramic pictures of Audi A3, C grades of benz
Panoramic pictures, E grades of the panoramic pictures etc. of running quickly) corresponding preset quantity samples pictures;
B, each samples pictures are subjected to picture pretreatment and train picture to obtain the training vehicle cab recognition model,
Such as can such as be scaled, after cutting, overturning and/or distortion operation by carrying out picture pretreatment to each samples pictures, make instruction
Practice picture be of the same size and identical visual angle after, just progress model training, to effectively improve the authenticity of model training
And accuracy rate.
C, all trained pictures are divided into the training set of the first ratio (for example, 85%), the second ratio (for example, 15%)
Verification collection;
D, the vehicle cab recognition model is trained using the training set;
E, using the accuracy rate of the vehicle cab recognition model of the verification collection verification training, if accuracy rate is more than or waits
In default accuracy rate, then training terminates, if alternatively, accuracy rate increases the corresponding sample in each position less than default accuracy rate
Picture;
Quantity simultaneously re-executes above-mentioned steps B, C, D, E, until training vehicle cab recognition model accuracy rate be more than or
Equal to default accuracy rate.
The determining module 34 is according to the final setting loss at the brand and vehicle of vehicle and each position of vehicle as a result, really
Determine the mantenance data at each position of vehicle.
In an alternative embodiment, the mantenance data includes the price data of aftermarket attachment, maintenance man-hours cost data.Respectively
The different degree of injury at a position correspond to different aftermarket attachments and different maintenance man-hours.The brand is read from database
With the mantenance data at each position of the vehicle of vehicle.When each position of the not no vehicle of the brand and vehicle of the database
Mantenance data when, to supplier send inquiry instruction to inquire the repair number at each position of the vehicle of the brand and vehicle
According to.
The computing module 38 calculates the maintenance cost of vehicle according to the mantenance data at each position of vehicle, and sends
To the user equipment of user.
In an alternative embodiment, by the price data of the aftermarket attachment at each position of vehicle, maintenance man-hours cost data
It adds up, the maintenance cost as vehicle.
In an alternative embodiment, the determining module 34 is additionally operable to:The insurance data for obtaining vehicle, according to insurance data and
Maintenance cost determines Claims Resolution data, and data of settling a claim are sent to the equipment of the user of the vehicle so that user checks.
By the way that in above-described embodiment, the present invention accurately can carry out setting loss to damage vehicle, and according to the setting loss situation of vehicle
And insurance data, the maintenance cost and Claims Resolution data of vehicle are calculated, Claims Resolution efficiency and the transparence of Claims Resolution are improved.
The above-mentioned integrated unit realized in the form of software program module, can be stored in one and computer-readable deposit
In storage media.Above-mentioned software program module is stored in a storage medium, including some instructions are used so that a computer
It is each that equipment (can be personal computer, server or the network equipment etc.) or processor (processor) execute the present invention
The part steps of embodiment the method.
As shown in figure 5, the electronic equipment 5 includes at least one sending device 51, at least one processor 52, at least one
A processor 53, at least one reception device 54 and at least one communication bus.Wherein, the communication bus is for realizing this
Connection communication between a little components.
The electronic equipment 5 be it is a kind of can according to the instruction for being previously set or storing, it is automatic carry out numerical computations and/or
The equipment of information processing, hardware include but not limited to microprocessor, application-specific integrated circuit (Application Specific
Integrated Circuit, ASIC), programmable gate array (Field-Programmable Gate Array, FPGA), number
Word processing device (Digital Signal Processor, DSP), embedded device etc..The electronic equipment 5 may also include network
Equipment and/or user equipment.Wherein, the network equipment includes but not limited to single network server, multiple network servers
The server group of composition or the cloud being made of a large amount of hosts or network server for being based on cloud computing (Cloud Computing),
Wherein, cloud computing is one kind of Distributed Calculation, a super virtual computing being made of the computer collection of a group loose couplings
Machine.
The electronic equipment 5, which may be, but not limited to, any type, to pass through keyboard, touch tablet or voice-operated device with user
Etc. modes carry out the electronic product of human-computer interaction, for example, tablet computer, smart mobile phone, personal digital assistant (Personal
Digital Assistant, PDA), intellectual Wearable, picture pick-up device, the terminals such as monitoring device.
Network residing for the electronic equipment 5 includes, but are not limited to internet, wide area network, Metropolitan Area Network (MAN), LAN, virtual
Dedicated network (Virtual Private Network, VPN) etc..
Wherein, the reception device 54 and the sending device 51 can be wired sending ports, or wirelessly set
It is standby, such as including antenna assembly, for other equipment into row data communication.
The memory 52 is for storing program code.The memory 52 can not have physical form in integrated circuit
The circuit with store function, such as RAM (Random-Access Memory, random access memory), FIFO (First In
First Out) etc..Alternatively, the memory 52 can also be the memory with physical form, such as memory bar, TF card
(Trans-flash Card), smart media card (smart media card), safe digital card (secure digital
Card), storage facilities such as flash memory cards (flash card) etc..
The processor 53 may include one or more microprocessor, digital processing unit.The processor 53 is adjustable
With the program code stored in memory 52 to execute relevant function.For example, the modules described in Fig. 3 and Fig. 4 are to deposit
The program code in the memory 52 is stored up, and performed by the processor 53, to realize a kind of car damage identification method.Institute
It states processor 53 and is also known as central processing unit (CPU, Central Processing Unit), be one block of ultra-large integrated electricity
Road is arithmetic core (Core) and control core (Control Unit).
The embodiment of the present invention also provides a kind of computer readable storage medium, is stored thereon with computer instruction, the finger
It enables when being executed by the electronic equipment including one or more processors, electronic equipment is made to execute as described in embodiment of the method above
Car damage identification method.
In conjunction with shown in Fig. 1 and Fig. 2, the memory 52 in the electronic equipment 5 stores multiple instruction to realize one kind
Car damage identification method, the processor 53 can perform the multiple instruction to realize:
Obtain vehicle pictures;Vehicle pictures are input in the parted pattern of trained vehicle position, vehicle is partitioned into
Each position picture;Each position picture is input in trained vehicle site tissue damage identification model, identifies each position
Machine setting loss result and each position of output machine setting loss result confidence level;Confidence level is less than or equal to the portion of threshold value
Bitmap piece is sent on the user equipment of setting loss personnel so that setting loss personnel carry out setting loss to the position picture, and determine confidence level
Less than the final setting loss result at the position of threshold value;Confidence level is added to the identification of vehicle site tissue damage less than the position picture of threshold value
In the training sample of model, re -training vehicle site tissue damage identification model.
The processor also executes when executing the multiple instruction to give an order:It is trained being input to vehicle pictures
It in the parted pattern of vehicle position, is partitioned into before each position picture of vehicle, the vehicle pictures of acquisition is detected, judge
Whether the vehicle pictures of acquisition are qualified, and the content of detection includes the combination of following one or more:Picture clarity, shooting angle
Degree, the recognizable degree at shooting position, picture, which whether there is, distorts suspicion;
When the vehicle pictures of the acquisition are unqualified, user is prompted to upload vehicle pictures again.
The processor also executes when executing the multiple instruction to give an order:Each position picture is being input to training
Before good vehicle site tissue damage identification model, car plate position, VIN codes position are identified from vehicle pictures, from car plate position
In identify license plate number, identify VIN codes from VIN codes position, whether be vehicle of insuring using license plate number or VIN code identifications vehicle
, when for insure vehicle when, judge the degree of injury of each position picture of vehicle.
The processor also executes when executing the multiple instruction to give an order:When the position that confidence level is less than to threshold value
When being sent to setting loss personnel, user waiting prompt setting loss result.
The processor also executes when executing the multiple instruction to give an order:As the machine setting loss result at the position and people
When work setting loss result difference, the final setting loss result using artificial setting loss result as confidence level less than the position of threshold value.
The processor also executes when executing the multiple instruction to give an order:Machine by confidence level higher than the position of threshold value
Final setting loss result of the device setting loss result as confidence level higher than the position of threshold value.
The processor also executes when executing the multiple instruction to give an order:
When vehicle is to insure vehicle, the vehicle pictures are input in trained vehicle cab recognition model, vehicle is exported
Brand and vehicle;
According to the final setting loss at the brand of vehicle and vehicle and each position of vehicle as a result, determining each portion of vehicle
The mantenance data of position;
According to the mantenance data at each position of vehicle, the maintenance cost of vehicle is calculated, and the user for being sent to user sets
It is standby.
The processor also executes when executing the multiple instruction to give an order:The insurance data for obtaining vehicle, according to throwing
Data and maintenance cost are protected, determines Claims Resolution data, data of settling a claim are sent to the equipment of the user of the vehicle so that user checks.
The corresponding multiple instruction of the car damage identification method described in any embodiment is stored in the memory 52, and passes through
The processor 53 executes, and this will not be detailed here.
The characteristic means of present invention mentioned above can be realized by integrated circuit, and control above-mentioned of realization
The function of car damage identification method described in embodiment of anticipating.That is, the integrated circuit of the present invention is installed in the electronic equipment, make institute
Electronic equipment is stated to play the following functions:Obtain vehicle pictures;Vehicle pictures are input to trained vehicle position parted pattern
In, it is partitioned into each position picture of vehicle;Each position picture is input to trained vehicle site tissue damage identification model
In, identify the confidence level of the machine setting loss result at each position and the machine setting loss result at each position of output;Confidence level is low
It is sent on the user equipment of setting loss personnel so that setting loss personnel carry out position picture in or equal to the position picture of threshold value
Setting loss, and determine final setting loss result of the confidence level less than the position of threshold value;Position picture addition by confidence level less than threshold value
Into the training sample of vehicle site tissue damage identification model, re -training vehicle site tissue damage identification model.
Function achieved by the car damage identification method described in any embodiment can be transferred through the integrated circuit of the present invention
It is installed in the electronic equipment, the electronic equipment is made to play achieved by car damage identification method described in any embodiment
Function, this will not be detailed here.
It should be noted that for each method embodiment above-mentioned, for simple description, therefore it is all expressed as a series of
Combination of actions, but those skilled in the art should understand that, the present invention is not limited by the described action sequence because
According to the present invention, certain steps can be performed in other orders or simultaneously.Secondly, those skilled in the art should also know
It knows, embodiment described in this description belongs to preferred embodiment, and involved action and module are not necessarily of the invention
It is necessary.
In the above-described embodiments, it all emphasizes particularly on different fields to the description of each embodiment, there is no the portion being described in detail in some embodiment
Point, it may refer to the associated description of other embodiment.
In several embodiments provided herein, it should be understood that disclosed device, it can be by another way
It realizes.For example, the apparatus embodiments described above are merely exemplary, for example, the unit division, it is only a kind of
Division of logic function, formula that in actual implementation, there may be another division manner, such as multiple units or component can combine or can
To be integrated into another system, or some features can be ignored or not executed.Another point, shown or discussed is mutual
Coupling, direct-coupling or communication connection can be by some interfaces, the INDIRECT COUPLING or communication connection of device or unit,
Can be electrical or other forms.
The unit illustrated as separating component may or may not be physically separated, aobvious as unit
The component shown may or may not be physical unit, you can be located at a place, or may be distributed over multiple
In network element.Some or all of unit therein can be selected according to the actual needs to realize the mesh of this embodiment scheme
's.
In addition, each functional unit in various embodiments of the present invention can be integrated in a processing unit, also may be used
It, can also be during two or more units be integrated in one unit to be that each unit physically exists alone.It is above-mentioned integrated
The form that hardware had both may be used in unit is realized, can also be realized in the form of SFU software functional unit.
If the integrated unit is realized in the form of SFU software functional unit and sells or use as independent product
When, it can be stored in a computer read/write memory medium.Based on this understanding, technical scheme of the present invention is substantially
The all or part of the part that contributes to existing technology or the technical solution can be in the form of software products in other words
It embodies, which is stored in a storage medium, including some instructions are used so that a computer
Equipment (can be personal computer, server or network equipment etc.) execute each embodiment the method for the present invention whole or
Part steps.And storage medium above-mentioned includes:USB flash disk, read-only memory (ROM, Read-Only Memory), arbitrary access are deposited
Reservoir (RAM, Random Access Memory), mobile hard disk, magnetic disc or CD etc. are various can to store program code
Medium.
The above, the above embodiments are merely illustrative of the technical solutions of the present invention, rather than its limitations;Although with reference to before
Stating embodiment, invention is explained in detail, it will be understood by those of ordinary skill in the art that:It still can be to preceding
The technical solution recorded in each embodiment is stated to modify or equivalent replacement of some of the technical features;And these
Modification or replacement, the range for various embodiments of the present invention technical solution that it does not separate the essence of the corresponding technical solution.
Claims (10)
1. a kind of car damage identification method, which is characterized in that the method includes:
Obtain vehicle pictures;
Vehicle pictures are input in the parted pattern of trained vehicle position, each position picture of vehicle is partitioned into;
Each position picture is input in trained vehicle site tissue damage identification model, identifies the machine setting loss at each position
And the confidence level of the machine setting loss result at each position of output as a result;
Position picture by confidence level less than or equal to threshold value is sent on the user equipment of setting loss personnel so that setting loss personnel couple
The position picture carries out setting loss, and determines final setting loss result of the confidence level less than the position of threshold value;
Position picture by confidence level less than threshold value is added in the training sample of vehicle site tissue damage identification model, re -training
Vehicle site tissue damage identification model.
2. car damage identification method as described in claim 1, which is characterized in that vehicle pictures are being input to trained vehicle
In the parted pattern of position, it is partitioned into before each position picture of vehicle, the method further includes:
The vehicle pictures of acquisition are detected, judge whether the vehicle pictures obtained are qualified, and the content of detection includes with next
Kind or a variety of combinations:Picture clarity, shooting angle, the recognizable degree for shooting position, picture, which whether there is, distorts suspicion
It doubts;
When the vehicle pictures of the acquisition are unqualified, user is prompted to upload vehicle pictures again.
3. car damage identification method as described in claim 1, which is characterized in that trained each position picture to be input to
Before vehicle site tissue damage identification model, the method further includes:
Car plate position, VIN codes position are identified from vehicle pictures, license plate number is identified from car plate position, from VIN codes position
Identify VIN codes, whether be vehicle of insuring using license plate number or VIN code identifications vehicle, when for insure vehicle when, judge vehicle
Each position picture degree of injury.
4. car damage identification method as described in claim 1, which is characterized in that the method further includes:When the machine at the position
When setting loss result is with artificial setting loss result difference, the final setting loss using artificial setting loss result as confidence level less than the position of threshold value
As a result.
5. car damage identification method as described in claim 1, which is characterized in that the method further includes:Confidence level is higher than threshold
Final setting loss result of the machine setting loss result at the position of value as confidence level higher than the position of threshold value.
6. car damage identification method as described in claim 1, which is characterized in that the method further includes:
When vehicle is to insure vehicle, the vehicle pictures are input in trained vehicle cab recognition model, vehicle is exported
Brand and vehicle;
According to the final setting loss at the brand of vehicle and vehicle and each position of vehicle as a result, determining each position of vehicle
Mantenance data;
According to the mantenance data at each position of vehicle, the maintenance cost of vehicle is calculated, and is sent to the user equipment of user.
7. car damage identification method as claimed in claim 6, which is characterized in that the method further includes:Obtain insuring for vehicle
Data determine Claims Resolution data according to insurance data and maintenance cost, data of settling a claim are sent to the equipment of the user of the vehicle with
It is checked for user.
8. a kind of car damage identification device, described device include:
Acquisition module, for obtaining vehicle pictures;
Segmentation module is partitioned into each of vehicle for vehicle pictures to be input in the parted pattern of trained vehicle position
Position picture;
Identification module, for each position picture to be input in trained vehicle site tissue damage identification model, identification is each
The confidence level of the machine setting loss result at position and the machine setting loss result at each position of output;
Sending module, for by confidence level less than or equal to threshold value position picture be sent on the user equipment of setting loss personnel with
So that setting loss personnel is carried out setting loss to the position picture, and determines final setting loss result of the confidence level less than the position of threshold value;
Training module, the training sample for confidence level to be added to vehicle site tissue damage identification model less than the position picture of threshold value
In this, re -training vehicle site tissue damage identification model.
9. a kind of electronic equipment, which is characterized in that the electronic equipment includes memory and processor, and the memory is for depositing
At least one instruction is stored up, the processor is for executing at least one instruction to realize such as any one of claim 1 to 7
The car damage identification method.
10. a kind of computer readable storage medium, which is characterized in that the computer-readable recording medium storage has at least one
Instruction, at least one instruction realize the car damage identification side as described in any one of claim 1 to 7 when being executed by processor
Method.
Priority Applications (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810196561.1A CN108446618A (en) | 2018-03-09 | 2018-03-09 | Car damage identification method, device, electronic equipment and storage medium |
PCT/CN2018/082577 WO2019169688A1 (en) | 2018-03-09 | 2018-04-10 | Vehicle loss assessment method and apparatus, electronic device, and storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810196561.1A CN108446618A (en) | 2018-03-09 | 2018-03-09 | Car damage identification method, device, electronic equipment and storage medium |
Publications (1)
Publication Number | Publication Date |
---|---|
CN108446618A true CN108446618A (en) | 2018-08-24 |
Family
ID=63194477
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810196561.1A Pending CN108446618A (en) | 2018-03-09 | 2018-03-09 | Car damage identification method, device, electronic equipment and storage medium |
Country Status (2)
Country | Link |
---|---|
CN (1) | CN108446618A (en) |
WO (1) | WO2019169688A1 (en) |
Cited By (37)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109271908A (en) * | 2018-09-03 | 2019-01-25 | 阿里巴巴集团控股有限公司 | Vehicle damages detection method, device and equipment |
CN109344819A (en) * | 2018-12-13 | 2019-02-15 | 深源恒际科技有限公司 | Vehicle damage recognition methods based on deep learning |
CN109359676A (en) * | 2018-10-08 | 2019-02-19 | 百度在线网络技术(北京)有限公司 | Method and apparatus for generating vehicle damage information |
CN109389169A (en) * | 2018-10-08 | 2019-02-26 | 百度在线网络技术(北京)有限公司 | Method and apparatus for handling image |
CN109410218A (en) * | 2018-10-08 | 2019-03-01 | 百度在线网络技术(北京)有限公司 | Method and apparatus for generating vehicle damage information |
CN109410270A (en) * | 2018-09-28 | 2019-03-01 | 百度在线网络技术(北京)有限公司 | A kind of damage identification method, equipment and storage medium |
CN109523556A (en) * | 2018-09-30 | 2019-03-26 | 百度在线网络技术(北京)有限公司 | Vehicle part dividing method and device |
CN109614935A (en) * | 2018-12-12 | 2019-04-12 | 泰康保险集团股份有限公司 | Car damage identification method and device, storage medium and electronic equipment |
CN109710654A (en) * | 2018-10-17 | 2019-05-03 | 青岛腾信汽车网络科技服务有限公司 | A kind of impaired level identification method of vehicle collision |
CN109740547A (en) * | 2019-01-04 | 2019-05-10 | 平安科技(深圳)有限公司 | A kind of image processing method, equipment and computer readable storage medium |
CN109767339A (en) * | 2018-12-03 | 2019-05-17 | 中国人民财产保险股份有限公司 | A kind of Claims Resolution data determination method, the apparatus and system of accident vehicle |
CN109784171A (en) * | 2018-12-14 | 2019-05-21 | 平安科技(深圳)有限公司 | Car damage identification method for screening images, device, readable storage medium storing program for executing and server |
CN109815997A (en) * | 2019-01-04 | 2019-05-28 | 平安科技(深圳)有限公司 | The method and relevant apparatus of identification vehicle damage based on deep learning |
CN110135437A (en) * | 2019-05-06 | 2019-08-16 | 北京百度网讯科技有限公司 | For the damage identification method of vehicle, device, electronic equipment and computer storage medium |
CN110570435A (en) * | 2018-09-10 | 2019-12-13 | 阿里巴巴集团控股有限公司 | method and device for carrying out damage segmentation on vehicle damage image |
CN110569701A (en) * | 2018-12-29 | 2019-12-13 | 阿里巴巴集团控股有限公司 | computer-implemented vehicle damage assessment method and device |
CN110570317A (en) * | 2018-08-31 | 2019-12-13 | 阿里巴巴集团控股有限公司 | Method and device for vehicle nuclear damage |
CN110660000A (en) * | 2019-09-09 | 2020-01-07 | 平安科技(深圳)有限公司 | Data prediction method, device, equipment and computer readable storage medium |
CN110837779A (en) * | 2019-10-12 | 2020-02-25 | 平安科技(深圳)有限公司 | Vehicle appearance intelligent diagnosis method and device and computer readable storage medium |
WO2020073510A1 (en) * | 2018-10-11 | 2020-04-16 | 平安科技(深圳)有限公司 | Neural network-based vehicle damage determination method, server and medium |
CN111191400A (en) * | 2019-12-31 | 2020-05-22 | 上海钧正网络科技有限公司 | Vehicle part service life prediction method and system based on user fault reporting data |
CN111242070A (en) * | 2020-01-19 | 2020-06-05 | 上海眼控科技股份有限公司 | Target object detection method, computer device, and storage medium |
WO2020140684A1 (en) * | 2019-01-02 | 2020-07-09 | 阿里巴巴集团控股有限公司 | Method and device for evaluating vehicle damage identification model |
CN111489433A (en) * | 2020-02-13 | 2020-08-04 | 北京百度网讯科技有限公司 | Vehicle damage positioning method and device, electronic equipment and readable storage medium |
CN111583215A (en) * | 2020-04-30 | 2020-08-25 | 平安科技(深圳)有限公司 | Intelligent damage assessment method and device for damage image, electronic equipment and storage medium |
CN111666990A (en) * | 2020-05-27 | 2020-09-15 | 平安科技(深圳)有限公司 | Vehicle damage characteristic detection method and device, computer equipment and storage medium |
CN111709352A (en) * | 2020-06-12 | 2020-09-25 | 浪潮集团有限公司 | Vehicle scratch detection method based on neural network |
CN111915446A (en) * | 2020-08-14 | 2020-11-10 | 南京三百云信息科技有限公司 | Accident vehicle damage assessment method and device and terminal equipment |
CN112017065A (en) * | 2020-08-27 | 2020-12-01 | 中国平安财产保险股份有限公司 | Vehicle loss assessment and claim settlement method and device and computer readable storage medium |
CN112085721A (en) * | 2020-09-07 | 2020-12-15 | 中国平安财产保险股份有限公司 | Damage assessment method, device and equipment for flooded vehicle based on artificial intelligence and storage medium |
WO2021012891A1 (en) * | 2019-07-23 | 2021-01-28 | 平安科技(深圳)有限公司 | Vehicle loss assessment method, device, apparatus, and storage medium |
CN112700436A (en) * | 2021-01-13 | 2021-04-23 | 上海微亿智造科技有限公司 | Method, system and medium for improving iteration of industrial quality inspection model |
CN113378619A (en) * | 2021-03-12 | 2021-09-10 | 中国平安财产保险股份有限公司 | Insurance business data processing method and device, electronic equipment and storage medium |
CN113496242A (en) * | 2020-04-07 | 2021-10-12 | 华晨宝马汽车有限公司 | Method and apparatus for classifying damaged portion of vehicle |
CN113627252A (en) * | 2021-07-07 | 2021-11-09 | 浙江吉利控股集团有限公司 | Vehicle damage assessment method and device, storage medium and electronic equipment |
WO2023273296A1 (en) * | 2021-06-30 | 2023-01-05 | 平安科技(深圳)有限公司 | Vehicle image segmentation quality evaluation method and apparatus, device, and storage medium |
CN116168356A (en) * | 2023-04-26 | 2023-05-26 | 威海海洋职业学院 | Vehicle damage judging method based on computer vision |
Families Citing this family (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110889585B (en) * | 2019-10-12 | 2023-08-22 | 中国平安财产保险股份有限公司 | Information classification decision method, device, computer system and readable storage medium |
CN111415018A (en) * | 2020-03-18 | 2020-07-14 | 上海钧正网络科技有限公司 | Method, device, medium and equipment for evaluating maintenance ex-warehouse |
CN111666973B (en) * | 2020-04-29 | 2024-04-09 | 平安科技(深圳)有限公司 | Vehicle damage picture processing method and device, computer equipment and storage medium |
CN112085610B (en) * | 2020-09-07 | 2023-08-22 | 中国平安财产保险股份有限公司 | Target damage assessment method, target damage assessment device, electronic equipment and computer readable storage medium |
CN112712498A (en) * | 2020-12-25 | 2021-04-27 | 北京百度网讯科技有限公司 | Vehicle damage assessment method and device executed by mobile terminal, mobile terminal and medium |
CN113326954B (en) * | 2021-06-25 | 2023-07-07 | 中国平安财产保险股份有限公司 | Vehicle maintenance task scheduling method, device, equipment and storage medium |
CN113744392A (en) * | 2021-08-30 | 2021-12-03 | 深圳壹账通智能科技有限公司 | Three-dimensional model library construction method, device, equipment and medium |
CN114273245A (en) * | 2021-11-10 | 2022-04-05 | 上海艾豚科技有限公司 | Continuous mixed detection method and system for multiple kinds of workpieces of automotive interior parts |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106231257A (en) * | 2016-07-29 | 2016-12-14 | 深圳市永兴元科技有限公司 | Car damage identification method and car damage identification server |
CN106303426A (en) * | 2016-08-17 | 2017-01-04 | 苏州华兴源创电子科技有限公司 | A kind of vehicle remote damage identification method and system |
CN107133876A (en) * | 2017-05-08 | 2017-09-05 | 明觉科技(北京)有限公司 | Car damage identification method and setting loss client |
US20170316262A1 (en) * | 2016-04-30 | 2017-11-02 | Infrared Integrated Systems Limited | System, Method And Apparatus For Occupancy Detection |
CN107358596A (en) * | 2017-04-11 | 2017-11-17 | 阿里巴巴集团控股有限公司 | A kind of car damage identification method based on image, device, electronic equipment and system |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107403424B (en) * | 2017-04-11 | 2020-09-18 | 阿里巴巴集团控股有限公司 | Vehicle loss assessment method and device based on image and electronic equipment |
-
2018
- 2018-03-09 CN CN201810196561.1A patent/CN108446618A/en active Pending
- 2018-04-10 WO PCT/CN2018/082577 patent/WO2019169688A1/en active Application Filing
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20170316262A1 (en) * | 2016-04-30 | 2017-11-02 | Infrared Integrated Systems Limited | System, Method And Apparatus For Occupancy Detection |
CN106231257A (en) * | 2016-07-29 | 2016-12-14 | 深圳市永兴元科技有限公司 | Car damage identification method and car damage identification server |
CN106303426A (en) * | 2016-08-17 | 2017-01-04 | 苏州华兴源创电子科技有限公司 | A kind of vehicle remote damage identification method and system |
CN107358596A (en) * | 2017-04-11 | 2017-11-17 | 阿里巴巴集团控股有限公司 | A kind of car damage identification method based on image, device, electronic equipment and system |
CN107133876A (en) * | 2017-05-08 | 2017-09-05 | 明觉科技(北京)有限公司 | Car damage identification method and setting loss client |
Cited By (50)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110570317A (en) * | 2018-08-31 | 2019-12-13 | 阿里巴巴集团控股有限公司 | Method and device for vehicle nuclear damage |
CN109271908A (en) * | 2018-09-03 | 2019-01-25 | 阿里巴巴集团控股有限公司 | Vehicle damages detection method, device and equipment |
CN109271908B (en) * | 2018-09-03 | 2022-05-13 | 创新先进技术有限公司 | Vehicle loss detection method, device and equipment |
CN110570435A (en) * | 2018-09-10 | 2019-12-13 | 阿里巴巴集团控股有限公司 | method and device for carrying out damage segmentation on vehicle damage image |
TWI709085B (en) * | 2018-09-10 | 2020-11-01 | 開曼群島商創新先進技術有限公司 | Method, device, computer readable storage medium and computing equipment for damage segmentation of vehicle damage image |
WO2020052352A1 (en) * | 2018-09-10 | 2020-03-19 | 阿里巴巴集团控股有限公司 | Method and device for damage segmentation of vehicle damage image |
CN109410270A (en) * | 2018-09-28 | 2019-03-01 | 百度在线网络技术(北京)有限公司 | A kind of damage identification method, equipment and storage medium |
CN109523556A (en) * | 2018-09-30 | 2019-03-26 | 百度在线网络技术(北京)有限公司 | Vehicle part dividing method and device |
CN109389169A (en) * | 2018-10-08 | 2019-02-26 | 百度在线网络技术(北京)有限公司 | Method and apparatus for handling image |
US11392792B2 (en) | 2018-10-08 | 2022-07-19 | Baidu Online Network Technology (Beijing) Co., Ltd. | Method and apparatus for generating vehicle damage information |
CN109410218A (en) * | 2018-10-08 | 2019-03-01 | 百度在线网络技术(北京)有限公司 | Method and apparatus for generating vehicle damage information |
CN109359676A (en) * | 2018-10-08 | 2019-02-19 | 百度在线网络技术(北京)有限公司 | Method and apparatus for generating vehicle damage information |
WO2020073510A1 (en) * | 2018-10-11 | 2020-04-16 | 平安科技(深圳)有限公司 | Neural network-based vehicle damage determination method, server and medium |
CN109710654A (en) * | 2018-10-17 | 2019-05-03 | 青岛腾信汽车网络科技服务有限公司 | A kind of impaired level identification method of vehicle collision |
CN109767339A (en) * | 2018-12-03 | 2019-05-17 | 中国人民财产保险股份有限公司 | A kind of Claims Resolution data determination method, the apparatus and system of accident vehicle |
CN109614935A (en) * | 2018-12-12 | 2019-04-12 | 泰康保险集团股份有限公司 | Car damage identification method and device, storage medium and electronic equipment |
CN109344819A (en) * | 2018-12-13 | 2019-02-15 | 深源恒际科技有限公司 | Vehicle damage recognition methods based on deep learning |
CN109784171A (en) * | 2018-12-14 | 2019-05-21 | 平安科技(深圳)有限公司 | Car damage identification method for screening images, device, readable storage medium storing program for executing and server |
CN110569701A (en) * | 2018-12-29 | 2019-12-13 | 阿里巴巴集团控股有限公司 | computer-implemented vehicle damage assessment method and device |
TWI726364B (en) * | 2018-12-29 | 2021-05-01 | 開曼群島商創新先進技術有限公司 | Computer-executed vehicle damage assessment method and device |
CN110569701B (en) * | 2018-12-29 | 2020-08-07 | 阿里巴巴集团控股有限公司 | Computer-implemented vehicle damage assessment method and device |
WO2020140684A1 (en) * | 2019-01-02 | 2020-07-09 | 阿里巴巴集团控股有限公司 | Method and device for evaluating vehicle damage identification model |
WO2020140610A1 (en) * | 2019-01-04 | 2020-07-09 | 平安科技(深圳)有限公司 | Image processing method and device, and computer-readable storage medium |
CN109740547A (en) * | 2019-01-04 | 2019-05-10 | 平安科技(深圳)有限公司 | A kind of image processing method, equipment and computer readable storage medium |
CN109815997A (en) * | 2019-01-04 | 2019-05-28 | 平安科技(深圳)有限公司 | The method and relevant apparatus of identification vehicle damage based on deep learning |
CN110135437A (en) * | 2019-05-06 | 2019-08-16 | 北京百度网讯科技有限公司 | For the damage identification method of vehicle, device, electronic equipment and computer storage medium |
CN110135437B (en) * | 2019-05-06 | 2022-04-05 | 北京百度网讯科技有限公司 | Loss assessment method and device for vehicle, electronic equipment and computer storage medium |
WO2021012891A1 (en) * | 2019-07-23 | 2021-01-28 | 平安科技(深圳)有限公司 | Vehicle loss assessment method, device, apparatus, and storage medium |
CN110660000A (en) * | 2019-09-09 | 2020-01-07 | 平安科技(深圳)有限公司 | Data prediction method, device, equipment and computer readable storage medium |
CN110837779A (en) * | 2019-10-12 | 2020-02-25 | 平安科技(深圳)有限公司 | Vehicle appearance intelligent diagnosis method and device and computer readable storage medium |
WO2021068566A1 (en) * | 2019-10-12 | 2021-04-15 | 平安科技(深圳)有限公司 | Vehicle appearance intelligent diagnosis method and apparatus, computer device, and computer-readable storage medium |
CN111191400B (en) * | 2019-12-31 | 2023-12-29 | 上海钧正网络科技有限公司 | Vehicle part life prediction method and system based on user fault reporting data |
CN111191400A (en) * | 2019-12-31 | 2020-05-22 | 上海钧正网络科技有限公司 | Vehicle part service life prediction method and system based on user fault reporting data |
CN111242070A (en) * | 2020-01-19 | 2020-06-05 | 上海眼控科技股份有限公司 | Target object detection method, computer device, and storage medium |
CN111489433A (en) * | 2020-02-13 | 2020-08-04 | 北京百度网讯科技有限公司 | Vehicle damage positioning method and device, electronic equipment and readable storage medium |
CN113496242A (en) * | 2020-04-07 | 2021-10-12 | 华晨宝马汽车有限公司 | Method and apparatus for classifying damaged portion of vehicle |
CN111583215A (en) * | 2020-04-30 | 2020-08-25 | 平安科技(深圳)有限公司 | Intelligent damage assessment method and device for damage image, electronic equipment and storage medium |
CN111666990A (en) * | 2020-05-27 | 2020-09-15 | 平安科技(深圳)有限公司 | Vehicle damage characteristic detection method and device, computer equipment and storage medium |
WO2021114809A1 (en) * | 2020-05-27 | 2021-06-17 | 平安科技(深圳)有限公司 | Vehicle damage feature detection method and apparatus, computer device, and storage medium |
CN111709352A (en) * | 2020-06-12 | 2020-09-25 | 浪潮集团有限公司 | Vehicle scratch detection method based on neural network |
CN111709352B (en) * | 2020-06-12 | 2022-10-04 | 浪潮集团有限公司 | Vehicle scratch detection method based on neural network |
CN111915446A (en) * | 2020-08-14 | 2020-11-10 | 南京三百云信息科技有限公司 | Accident vehicle damage assessment method and device and terminal equipment |
CN112017065A (en) * | 2020-08-27 | 2020-12-01 | 中国平安财产保险股份有限公司 | Vehicle loss assessment and claim settlement method and device and computer readable storage medium |
CN112017065B (en) * | 2020-08-27 | 2024-05-24 | 中国平安财产保险股份有限公司 | Method, device and computer readable storage medium for vehicle damage assessment and claim |
CN112085721A (en) * | 2020-09-07 | 2020-12-15 | 中国平安财产保险股份有限公司 | Damage assessment method, device and equipment for flooded vehicle based on artificial intelligence and storage medium |
CN112700436A (en) * | 2021-01-13 | 2021-04-23 | 上海微亿智造科技有限公司 | Method, system and medium for improving iteration of industrial quality inspection model |
CN113378619A (en) * | 2021-03-12 | 2021-09-10 | 中国平安财产保险股份有限公司 | Insurance business data processing method and device, electronic equipment and storage medium |
WO2023273296A1 (en) * | 2021-06-30 | 2023-01-05 | 平安科技(深圳)有限公司 | Vehicle image segmentation quality evaluation method and apparatus, device, and storage medium |
CN113627252A (en) * | 2021-07-07 | 2021-11-09 | 浙江吉利控股集团有限公司 | Vehicle damage assessment method and device, storage medium and electronic equipment |
CN116168356A (en) * | 2023-04-26 | 2023-05-26 | 威海海洋职业学院 | Vehicle damage judging method based on computer vision |
Also Published As
Publication number | Publication date |
---|---|
WO2019169688A1 (en) | 2019-09-12 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108446618A (en) | Car damage identification method, device, electronic equipment and storage medium | |
CN108595585B (en) | Sample data classification method, model training method, electronic equipment and storage medium | |
WO2019243863A1 (en) | Vehicle re-identification techniques using neural networks for image analysis, viewpoint-aware pattern recognition, and generation of multi-view vehicle representations | |
CN112380921A (en) | Road detection method based on Internet of vehicles | |
CN107679997A (en) | Method, apparatus, terminal device and storage medium are refused to pay in medical treatment Claims Resolution | |
CN107590774A (en) | A kind of car plate clarification method and device based on generation confrontation network | |
CN108197326A (en) | A kind of vehicle retrieval method and device, electronic equipment, storage medium | |
CN105069415A (en) | Lane line detection method and device | |
CN106709475A (en) | Obstacle recognition method and device, computer equipment and readable storage medium | |
CN106845416A (en) | Obstacle recognition method and device, computer equipment and computer-readable recording medium | |
CN110689043A (en) | Vehicle fine granularity identification method and device based on multiple attention mechanism | |
CN110263628B (en) | Obstacle detection method, obstacle detection device, electronic apparatus, and storage medium | |
CN108596056A (en) | A kind of taxi operation behavior act recognition methods and system | |
CN104680795A (en) | Vehicle type recognition method and device based on partial area characteristic | |
CN110516691A (en) | A kind of Vehicular exhaust detection method and device | |
CN112132033B (en) | Vehicle type recognition method and device, electronic equipment and storage medium | |
US11120308B2 (en) | Vehicle damage detection method based on image analysis, electronic device and storage medium | |
CN114783021A (en) | Intelligent detection method, device, equipment and medium for wearing of mask | |
CN114972911A (en) | Method and equipment for collecting and processing output data of automatic driving perception algorithm model | |
CN112101246A (en) | Vehicle identification method, device, equipment and medium | |
CN111798047A (en) | Wind control prediction method and device, electronic equipment and storage medium | |
CN115205813A (en) | Multi-view road vehicle matching system and method based on convolutional neural network | |
CN106056150A (en) | System and method for establishing part division remote damage assessment of different vehicle types based on artificial intelligence random forest method | |
CN116861262A (en) | Perception model training method and device, electronic equipment and storage medium | |
CN114863680A (en) | Prediction processing method, prediction processing device, computer equipment and storage medium |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180824 |