CN109242006A - The method and device of identification vehicle damage based on vehicle classification - Google Patents
The method and device of identification vehicle damage based on vehicle classification Download PDFInfo
- Publication number
- CN109242006A CN109242006A CN201810967661.XA CN201810967661A CN109242006A CN 109242006 A CN109242006 A CN 109242006A CN 201810967661 A CN201810967661 A CN 201810967661A CN 109242006 A CN109242006 A CN 109242006A
- Authority
- CN
- China
- Prior art keywords
- vehicle
- damage
- feature
- classification
- damaged
- 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
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
- G06V10/75—Organisation of the matching processes, e.g. simultaneous or sequential comparisons of image or video features; Coarse-fine approaches, e.g. multi-scale approaches; using context analysis; Selection of dictionaries
- G06V10/751—Comparing pixel values or logical combinations thereof, or feature values having positional relevance, e.g. template matching
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- 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)
- General Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Finance (AREA)
- Artificial Intelligence (AREA)
- Accounting & Taxation (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- Databases & Information Systems (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Multimedia (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Development Economics (AREA)
- Economics (AREA)
- Marketing (AREA)
- Strategic Management (AREA)
- Technology Law (AREA)
- General Business, Economics & Management (AREA)
- Traffic Control Systems (AREA)
Abstract
This specification embodiment provides a kind of method and apparatus of identification vehicle damage based on vehicle classification, according to this method embodiment, the field data of damaged vehicle is obtained first, then the vehicle feature and component feature, damage characteristic of damaged vehicle, then at least vehicle damage result based on vehicle feature and component feature, damage characteristic identification damaged vehicle are determined according to field data.In this way, making full use of the field data of damaged vehicle during identifying vehicle damage, the foundation of damaged vehicle judgement is enriched, so as to improve the accuracy of vehicle damage identification.
Description
Technical field
This specification one or more embodiment is related to field of computer technology, more particularly to is based on vehicle classification, passes through
The method and apparatus of computer identification vehicle damage.
Background technique
Vehicle is the general name of various vehicles, such as may include bicycle, car, truck, lorry, train etc..Vehicle
In injury event, often because of damaged vehicle caused by unexpected or artificial mistake, such as vehicle is scraped, collision etc..Damaged vehicle
Subsequent processing often relates to the identification to vehicle damage, to provide foundation to vehicle maintenance, settlement of insurance claim etc..Tradition damages vehicle
The qualification process of wound, often through artificial dam site investigation setting loss.For example, under settlement of insurance claim scene, damaged vehicle is determined
Damage process may include: by insurance company send profession survey setting loss personnel, by surveying setting loss, provide maintenance program and compensation
Indemnity volume, and by setting loss photo upload backstage, so that inspector's core damage in backstage is verified prices.As can be seen that existing by artificial
To the process of car damage identification, not only need to put into a large amount of artificial, it is also necessary to input artificial Damage identification with higher
Ability, and process cycle is longer.
In view of this, some prior arts are damaged using artificial intelligence identification vehicle part.However, utilizing artificial intelligence
The recognition result of vehicle part damage may be because that various factors causes a deviation.For example, the case where being damaged by picture recognition
Under, component more similar for some shapes is easy erroneous judgement.Such as the fender of car and wheel arch, by information collection
The influence of the factors such as distance, angle, light, weather, the partial region of fender and wheel arch may be more similar and be confused.
Accordingly, it would be desirable to there is improved plan, mentioned using the more information of damaged vehicle to be automatically performed setting loss process
The accuracy of height identification vehicle damage.
Summary of the invention
This specification one or more embodiment describes a kind of method and apparatus, can use artificial intelligence, and be based on
Specific vehicle classification identifies vehicle damage, and more damaged vehicle information are utilized, are automatically performed setting loss process, so as to mention
The accuracy of high vehicle damage identification.
According in a first aspect, providing a kind of method of identification vehicle damage based on vehicle classification, comprising: obtain impaired
The field data of vehicle, the field data include at least image or video；The impaired vehicle is determined according to the field data
Component feature and damage characteristic, the component feature is used to divide all parts of the damaged vehicle；According to described existing
Field information determines the vehicle classification of the damaged vehicle, and determines that the vehicle of the damaged vehicle is special according to the vehicle classification
Sign；The vehicle damage of the damaged vehicle is at least identified based on the component feature, the vehicle feature and the damage characteristic
As a result.
In some embodiments, the field data further includes threedimensional model and/or vehicles identifications, wherein the vehicle
Mark includes at least one of the following: vehicle cab recognition code, license plate number, motor number, Vehicle Identify Number.
In some embodiments, determine that the vehicle classification of the damaged vehicle includes: described according to the field data
In the case that field data includes the vehicles identifications, searched in model data library and the vehicles identifications based on mapping ruler
Corresponding vehicle classification, wherein the model data library is for storing multiple vehicles identifications and multiple vehicle classifications, Ge Geche
Mark and the corresponding relationship of each vehicle classification are described by the mapping ruler；In response to retrieving and the vehicle
Corresponding vehicle classification is identified, the vehicle classification retrieved is determined as to the vehicle classification of the damaged vehicle.
In some embodiments, based on mapping ruler retrieved in model data library it is opposite less than with the vehicles identifications
In the case where the vehicle classification answered, new vehicle classification is added in the model data library, and remember in the mapping ruler
Record the vehicle feature of the damaged vehicle and the corresponding relationship of new vehicle classification.
In some embodiments, the vehicle classification that the damaged vehicle is determined according to the field data include: from
The vehicle characteristics of the damaged vehicle are extracted in the field data；The vehicle characteristics are inputted into vehicle classification model, and base
The vehicle classification of the damaged vehicle is determined in the output result of the vehicle classification model.
In some embodiments, the vehicle feature includes at least one of the following: component names, number of components, component
Position feature, component between material feature, component are damaged incidence relation.
In some embodiments, described at least to be known based on the component feature, the vehicle feature and the damage characteristic
The vehicle damage result of the not described damaged vehicle includes: that the component feature, the vehicle feature and the damage characteristic is defeated
Enter the first non-destructive tests model, the vehicle damage result is determined according to the output result of the first non-destructive tests model.
In some embodiments, described at least to be known based on the component feature, the vehicle feature and the damage characteristic
The vehicle damage result of the not described damaged vehicle includes: to be updated using the vehicle feature to the component feature；It will more
Component feature and the damage characteristic after new input the second non-destructive tests model, according to the defeated of the second non-destructive tests model
Result determines the vehicle damage result out.
In some embodiments, described at least to be known based on the component feature, the vehicle feature and the damage characteristic
The vehicle damage result of the not described damaged vehicle includes: by the component feature and damage characteristic input third non-destructive tests
Model determines preliminary damage results according to the output result of the third non-destructive tests model；Using the vehicle feature to institute
It states preliminary damage results to be verified, and generates the vehicle damage result according to check results.
In some embodiments, the vehicle damage result includes: defective component, damage type.
In some embodiments, the method also includes determining maintenance program according to the vehicle damage result and/or estimate
Expense.
According to second aspect, a kind of device of identification vehicle damage based on vehicle classification is provided, comprising: acquiring unit,
It is configured to obtain the field data of damaged vehicle, the field data includes at least image or video；First determination unit, configuration
For the component feature and damage characteristic for determining the damaged vehicle according to the field data, the component feature is for dividing institute
State all parts of damaged vehicle；Second determination unit is configured to determine the vehicle of the damaged vehicle according to the field data
Type classification, and determine according to the vehicle classification vehicle feature of the damaged vehicle；Recognition unit is configured at least based on institute
State the vehicle damage result that component feature, the vehicle feature and the damage characteristic identify the damaged vehicle.
According to the third aspect, a kind of computer readable storage medium is provided, computer program is stored thereon with, when described
When computer program executes in a computer, enable computer execute first aspect method.
According to fourth aspect, a kind of calculating equipment, including memory and processor are provided, which is characterized in that described to deposit
It is stored with executable code in reservoir, when the processor executes the executable code, the method for realizing first aspect.
The method and apparatus provided by this specification embodiment obtain the field data of damaged vehicle, then root first
The component feature, damage characteristic and vehicle feature of damaged vehicle are determined according to field data, then at least special based on component feature, damage
The vehicle damage result of vehicle of seeking peace feature identification damaged vehicle.In this way, during identifying vehicle damage, make full use of by
The field data for damaging vehicle, together identifies vehicle damage using vehicle feature and component feature, damage characteristic, enriches
The foundation of damaged vehicle judgement, so as to improve the accuracy of vehicle damage identification.
Detailed description of the invention
In order to illustrate the technical solution of the embodiments of the present invention more clearly, required use in being described below to embodiment
Attached drawing be briefly described, it should be apparent that, drawings in the following description are only some embodiments of the invention, for this
For the those of ordinary skill of field, without creative efforts, it can also be obtained according to these attached drawings others
Attached drawing.
Fig. 1 shows the implement scene schematic diagram of one embodiment of this specification disclosure；
Fig. 2 shows the method flow diagrams for identifying vehicle damage based on vehicle classification according to one embodiment；
Fig. 3 show according to one embodiment based on vehicle classification, damage characteristic and component feature identify vehicle damage
As a result flow diagram；
Fig. 4 show according to another embodiment based on vehicle classification, damage characteristic and component feature identification vehicle damage
Hurt the flow diagram of result；
Fig. 5 is shown according to another embodiment based on vehicle classification, damage characteristic and component feature identification vehicle damage
Hurt the flow diagram of result；
Fig. 6 shows the schematic block diagram of the device of the identification vehicle damage based on vehicle classification of one embodiment.
Specific embodiment
With reference to the accompanying drawing, the scheme provided this specification is described.
Fig. 1 is the implement scene schematic diagram of one embodiment that this specification discloses.In the implement scene, it is assumed that vehicle by
Damage, can by can collection site information terminal, such as smart phone, camera, sensor etc., to computing platform send by
The field data of vehicle is damaged, computing platform can identify vehicle damage result according to the field data that user sends.
During computing platform identification vehicle damage result: it is possible, firstly, to obtain being acquired by above-mentioned terminal, impaired
The field data of vehicle, such as photo site and/or video, can also be the information such as vehicle deformation；It then, can be according to above-mentioned
Field data determines the vehicle classification of damaged vehicle, and the vehicle feature of damaged vehicle is determined based on vehicle classification, meanwhile, it can be with
The component feature and damage characteristic of damaged vehicle are determined according to field data；Then, above-mentioned vehicle feature and component are at least based on
Feature, damage characteristic identify vehicle damage result.The vehicle damage result for example can be " bumper recess " (defective component+
Damage type) form.In some implementations, computing platform can also be determined according to vehicle damage result maintenance program (such as
Replacement bumper), estimate maintenance cost etc. in it is one or more.
In a possible design, computing platform can be by above-mentioned vehicle feature and component feature, damage characteristic together
Trained first non-destructive tests model in advance is inputted, and result is exported according to the first non-destructive tests model and determines vehicle damage knot
Fruit.
In another possible design, computing platform can also be updated component feature using vehicle feature, will
Updated component feature and damage characteristic input the second non-destructive tests model, according to the output result of the second non-destructive tests model
Determine vehicle damage result.
In another possible design, computing platform can also know component feature and damage characteristic input third damage
Other model determines preliminary damage results according to the output result of third non-destructive tests model, then using vehicle feature to preliminary
Damage results are verified, and generate vehicle damage result according to check results.
In this way, the vehicle classification of the damaged vehicle determined by the field data is utilized, and vehicle classification will be passed through
Determining vehicle feature enriches the effective information of identification of damage as one of the foundation for determining vehicle damage result, comes automatic
Setting loss process is completed, to improve the accuracy of identification vehicle damage.By taking fender and wheel arch in background technique as an example, according to
Vehicle feature will not generate damage results as " wheel arch scraping " if the vehicle of a vehicle does not have wheel arch component, from
And it is more acurrate to the identification of vehicle damage.
Further, in some scenes, the vehicle damage result identified can also be fed back to scene by computing platform
Terminal, so that user refers to subsequent processing.For example, computing platform can be combined with according to vehicle in some scenes
The determining Maintenance price of damage results estimates maintenance price, feeds back to user, and user can determine according to the maintenance price estimated
Whether settlement of insurance claim, whether repair etc..
Fig. 2 shows the method flow diagrams for identifying vehicle damage based on vehicle classification according to one embodiment.This method
Executing subject can be and any there is calculating, the system of processing capacity, unit, platform or server, such as Fig. 1 institute
The computing platform etc. shown.More specifically, e.g. setting loss terminal, or the setting loss clothes of support are provided for car damage identification class application
Business device.In the case where above-mentioned executing subject is setting loss terminal, by taking Fig. 1 as an example, setting loss terminal can be collection site information
Terminal, and computing platform can be a part of setting loss terminal.
As shown in Fig. 2, method includes the following steps: step 21, obtains the field data of damaged vehicle, field data is extremely
It less include image or video；Step 22, the component feature and damage characteristic of damaged vehicle are determined according to field data；Step 23,
The vehicle classification of damaged vehicle is determined according to field data, and the vehicle feature of damaged vehicle is determined according to vehicle classification；Step
24, vehicle damage result is at least identified based on vehicle feature and component feature, damage characteristic.
Firstly, obtaining the field data of damaged vehicle in step 21.It is appreciated that the field data of damaged vehicle can be with
Be acquired from damaged vehicle site, the information of form for describing damaged vehicle.The field data of damaged vehicle can
To be acquired by the various acquisition equipment at damaged vehicle site (scene).Such as smart phone, camera can be passed through
Etc. the equipment acquisition image of damaged vehicle, video information；
In one embodiment, the opposite acquisition datum mark of each point in image can also be acquired by depth transducer etc.
Distance；
In another embodiment, the vehicle of character recognition or the reception user's input of collection in worksite terminal can also be passed through
Mark, such as: vehicle cab recognition code (such as certain brand car 7180AE vehicle), license plate number, motor number, Vehicle Identify Number (Vehicle
Identification Number, VIN) etc. at least one of；
In yet another embodiment, the point cloud chart, etc. of damaged vehicle can also be acquired by laser, radar camera.
Wherein, the threedimensional model of damaged vehicle can be rebuild by point cloud chart.It in other embodiments, can also be according to damaged vehicle
Multiple scene photographs establish the threedimensional model of damaged vehicle, are not repeating herein.
In more embodiments, it can also be believed by the scene of other equipment or the combination acquisition damaged vehicle of equipment
Breath, no longer enumerates herein.
Then, in step 22, the component feature and damage characteristic of damaged vehicle are determined according to field data.It is appreciated that
Component feature can be the feature of the different components for distinguishing damaged vehicle, for example, in photo all parts profile etc..Damage
Hurting feature can be feature for identification of damage region and/or damage type.It is readily appreciated that, for damaged vehicle, pars affecta
Position is compared with other positions, it is possible that the variation of such as color, light, shape, depth etc.Therefore, damage characteristic can
To be the various features of the pixel for having the differences such as color in photo with normal vehicle body, region contour, regional depth or the like, herein
It repeats no more.
On the other hand, in step 23, the vehicle classification of damaged vehicle is determined according to field data, and true according to vehicle classification
Determine the vehicle feature of damaged vehicle.It is appreciated that vehicle feature can be feature relevant to the vehicle of damaged vehicle.For example,
To the car of some model of Mr. Yu's brand, vehicle feature be can include but is not limited to: component names, number of components (such as seat
Number), component material feature, the position feature (such as headlight and bumper positional relationship) between component, the impaired incidence relation of component (such as
Components A is impaired, the impaired probability of part B) etc. in it is one or more.The vehicle classification of damaged vehicle can be by step 1
Acquired field data determines.
In a possible embodiment, it if including the vehicles identifications of damaged vehicle in field data, can be based on
Scheduled mapping ruler retrieves vehicle classification corresponding with above-mentioned vehicles identifications in model data library.Retrieve with it is above-mentioned
In the case where the corresponding vehicle classification of vehicles identifications, using the vehicle classification retrieved as the vehicle classification of damaged vehicle.
Wherein, model data library can store multiple vehicles identifications and multiple vehicle classifications, each vehicles identifications and each vehicle classification
Corresponding relationship can be described by above-mentioned mapping ruler.
In some implementations, above-mentioned vehicles identifications for example can include but is not limited to vehicle cab recognition code, license plate number, engine
Number, Vehicle Identify Number, owner identity card number etc. in it is one or more.Vehicle cab recognition code can usually accurately identify vehicle classification；It is right
In motor vehicles, it may uniquely correspond to these information such as license plate number, motor number, Vehicle Identify Number, owner identity card number and lead to, also
It is to say, is likely to trace back to vehicle classification according to any of these numbers.In this way, information includes above- mentioned information at the scene
In the case where, mapping ruler may include the corresponding relationship of various numbers Yu vehicle classification.It, can be in vehicle based on mapping ruler
A corresponding vehicle classification in above-mentioned number is retrieved in database.
Optionally, in model data library, for each vehicle classification, shared various of category vehicle can also be stored with
Feature, such as list of parts, appearance threedimensional model, geometry feature and other features for vehicle vision identification, such as
Vehicle body length wide high proportion, roof shape, whether there is or not skylights etc..In the embodiment of this specification, can using these features as
Priori knowledge describes the relationship of itself and vehicle classification by mapping ruler, can also be in the vehicle classification for determining damaged vehicle
Afterwards, feature of these features as damaged vehicle is obtained.
In a possible realization, if corresponding vehicle class can not be retrieved in model data library based on mapping ruler
Not, new vehicle classification can also be increased in model data library, the new vehicle classification is corresponding with damaged vehicle.Some
In embodiment, which can also be by manually demarcating vehicle or computer according on Fixed Time Interval and network
Vehicle carries out match cognization vehicle.The corresponding vehicle feature of new vehicle can be various features (such as roof shape of damaged vehicle
Shape, whether there is or not skylights etc.).Optionally, the corresponding relationship of the vehicles identifications of the new vehicle classification and damaged vehicle, which is recorded in, reflects
It penetrates in rule.In concrete practice, multiple vehicle classifications newly increased can also be clustered, each cluster corresponds on one
The vehicle classification of position, vehicle feature of the common trait of the vehicle in cluster as corresponding vehicle classification.
In another possible embodiment, the vehicle of different automobile types can be first passed through in advance as sample training vehicle classification
Model, for carrying out vehicle classification according to vehicle characteristics.When classifying to damaged vehicle, can be extracted from field data
Vehicle characteristics are inputted vehicle classification model, and determine damaged vehicle based on the output result of vehicle classification model by vehicle characteristics
Vehicle classification.During vehicle classification model training, can first against various types of other vehicle sample extraction vehicle characteristics,
Such as roof shape, whether there is or not skylight, vehicle body length wide high proportion, car light position, bumper shape, bumper position, whether there is or not wheel arch
(if any wheel arch position) etc..These features can be obtained by the three-dimensional vehicle model of reconstruction, can also be by manually knowing
Not, this specification embodiment is not construed as limiting this.
In some implementations, vehicle can be trained (without paying close attention to the true classification of sample) by way of unsupervised learning
Disaggregated model.By taking clustering method as an example, the vehicle characteristics of each vehicle sample are clustered, vehicle is divided according to cluster result
Classification.Optionally, manually each vehicle classification in cluster result can also be demarcated or machine recognition cluster centre
The vehicle classification of sample, the vehicle classification as the sample for belonging to same cluster with it.
In other realizations, mode (the true classification of calibration sample in advance) training vehicle of supervised learning can be used
Disaggregated model.By taking vehicle classification model is decision tree (Decision Tree) as an example, vehicle characteristics are extracted from vehicle sample,
By training decision tree, each internal node of decision tree is made to indicate the judgement of a vehicle characteristics, each branch represents one
Judgement output, each leaf node represent a kind of vehicle classification.
According to a possible design, vehicle classification is further divided into multiple levels.Such as: first level divides vehicle
For car, truck, lorry, engineering truck etc.；In second level, car is divided into car, (sports type is practical by SUV
Automobile), offroad vehicle etc., truck is divided into min-truck, light truck, heavy truck etc.. and so on, Ke Yixi
Assign to the concrete model of specific brand or even automatic catch/manual gear of concrete model, high quota/low quota etc..In vehicle classification
In the case where including multiple levels, damaged vehicle can be divided into the vehicle classification for the level that can determine that.For example,
It is car or SUV that one, which is damaged serious vehicle and cannot be distinguished, then can be divided into car classification.
In concrete practice, for disaggregated model, the level between class of vehicle can be embodied by different modes.Example
Such as, for Clustering Model, each cluster can be used as a thinner classification, using closely located multiple clusters as they
Upper hierarchical category, etc..For another example the corresponding vehicle classification of each internal node, leaf can be made for decision-tree model
Node is most thin vehicle classification, etc..
In a possible design, can detecte in field data whether include damaged vehicle vehicles identifications, do not having
In the case where there are the vehicles identifications including damaged vehicle, then determine based on vehicle classification model the vehicle classification of damaged vehicle.It can
To understand, although for these information such as license plate number, motor number, Vehicle Identify Number, owner identity cards number, according to either of which
Item is likely to trace back to vehicle classification, but if not corresponding to accurate vehicle classification in statistics, still can not know
It Chu not specific vehicle classification.It therefore, in one implementation, can also be by the information retrieved (such as brand name) as vehicle point
One of feature of class model, to identify the vehicle classification of damaged vehicle by vehicle classification model.
It in some embodiments, can also basis if can not identify specific vehicle classification using vehicle classification model
Vehicle characteristics increase new vehicle classification, which can be used for updating vehicle classification model.New vehicle classification
Processing it is similar with processing above-mentioned, details are not described herein.
It is worth noting that can have in the vehicle characteristics of vehicle classification and the vehicle feature based on vehicle extraction
There is part or all of identical feature, this specification embodiment is not construed as limiting this.It is appreciated that vehicle characteristics are due to for knowing
Other vehicle classification, therefore may include some and visual correlation, or other features of vehicle can be distinguished, and vehicle feature due to
It lays particular emphasis on and damages for identification, therefore can also include vehicle part title (such as list of designations), component material, component position
The relevant feature such as set.
Then, in step 24, at least based on above-mentioned vehicle feature and above-mentioned component feature, damage characteristic identification vehicle damage
Hurt result.It is appreciated that vehicle damage the result is that the degree of impairment to damaged vehicle judgement.Vehicle damage result may include
Such as damage type (as scraped), defective component (such as bumper), defective component+damage type/degree, damage material (such as steel
Part)+damage type/degree etc. as a result, this specification embodiment is not construed as limiting this.According to a possible implementation
Mode, as shown in figure 3, component feature, vehicle feature and damage characteristic can be inputted the first non-destructive tests model, according to first
The output result of non-destructive tests model determines vehicle damage result.In one embodiment, the output of the first non-destructive tests model
It as a result is exactly vehicle damage result.First non-destructive tests model can be training in advance, based on component, vehicle and damage characteristic
Determine the model of damage results.In one embodiment, which can be trained in advance by the way of having supervision.
It according to another embodiment, can be using vehicle feature as the prior information of non-destructive tests, in vehicle feature
On the basis of carry out non-destructive tests.Specifically, as shown in figure 4, being updated first with vehicle feature to component feature.Here
Update such as can be component feature (such as bezel locations) is corrected, increase description the attribute (material as plastic cement or metal plate
Attribute, damage probability equivalent damage attribute etc.) etc..Then, updated component feature and damage characteristic are inputted second together
Non-destructive tests model determines vehicle damage result according to the output result of the second non-destructive tests model.
In some implementations, the second non-destructive tests model is also possible to damage extent identification model, at this point, can also elder generation's root
Component locating for damage position is determined according to vehicle feature, determines damage type/degree further according to damage extent identification model.Another
In some realizations, the second non-destructive tests model can also be to be trained according to different vehicles, at this point it is possible to first according to vehicle
Feature determines the corresponding second non-destructive tests model of damaged vehicle, damage characteristic is then inputted the second non-destructive tests model, really
Determine vehicle damage result.
According to another embodiment, vehicle feature can also be used for the subsequent place of the recognition result of non-destructive tests model
Reason, to obtain accurate non-destructive tests result.It in this case, as shown in figure 5, can be first by damage characteristic and component feature
Third non-destructive tests model is inputted, determines that the preliminary damage of damaged vehicle is tied according to the output result of third non-destructive tests model
Fruit.Then, the preliminary damage results are verified using vehicle feature, and generates vehicle damage result according to check results.
It specifically, is the equal of to first if determining that the defective component in preliminary damage results is correct according to vehicle feature
Walk the verifying that damage results carry out；It, can if determining that the defective component in preliminary damage results has deviation according to vehicle feature
To be modified by vehicle feature to preliminary damage results, damage type/journey of correct defective component and each component is determined
Degree.For example, preliminary damage results are the part distortions for the component of a plastic cement material, then the preliminary identification can be determined
Result may be mistake, need to be modified.For example, it may is that a possibility that a kind of amendment and determine " damage " position
It on the components, then may be erroneous judgement caused by the reasons such as light, which may be to be not present；Another modified possibility
Property may is that " damage " position in the junction of the component and another component, and according to vehicle feature, actual damage region may be
On another component.In this way, can the setting loss result to routine techniques be modified.
It is appreciated that above-mentioned " the first non-destructive tests model ", " the second non-destructive tests model ", " third non-destructive tests model "
In " first ", " second ", " third " be rather than the restriction to sequence or title in order to distinguish different models.In other words,
These three non-destructive tests models are since input feature vector is different, rather than identical model.In actual use, it is not excluded that it
A possibility that being the same model or being trained using identical machine learning algorithm.
It is worth noting that although being labelled with step 21-24 in this embodiment, however, not suitable to the execution of step
Sequence is defined.By taking step 22 and step 23 as an example, although being step 22 in above-mentioned process preceding, however, as shown in Fig. 2, step
Rapid 22 and step 23 can also execute parallel.In other embodiments, step 22 and step 23 can be executed with reversed order,
That is step 23 can also execute before step 22 in above-mentioned process.
In some be able to achieve, further maintenance program can also be determined (as replacement is insured according to non-destructive tests result
Thick stick) and/or estimate maintenance cost etc..It, can be according to the shop 4S, a large amount of vehicle damage results of repair shop for maintenance program
It is obtained, can also be closed taking human as designated vehicle damage results and the corresponding of maintenance program by machine learning with final maintenance program
System.Similarly, estimating maintenance cost can also pass through according to the shop 4S, a large amount of vehicle damage results of repair shop and final maintenance cost
Machine learning obtains, or artificial determining.In some optinal plans, estimating maintenance cost is done on the basis of maintenance program
Out, such as maintenance program is replacement bumper, then is directed to the maintenance program, obtains the Maintenance price in each shop 4S, repair shop.
Optionally, estimating maintenance cost can be the Maintenance price (such as the quotation in the shop 4S) of a default, can also be according to user
The Maintenance price that the maintenance condition (such as high-quality or price is low) of selection determines.
It is appreciated that can also be tied when determining maintenance program according to non-destructive tests result and/or estimating maintenance cost etc.
Vehicle classification is closed just to correct.For example, for different vehicles, though under identical non-destructive tests result, maintenance program and
Maintenance cost also all may not be identical.
Look back above procedure make full use of the field data of damaged vehicle during identifying vehicle damage, not only from
It is extracted component feature and damage characteristic in field data, also extracts vehicle feature, utilizes vehicle feature and damage characteristic one
It rises and vehicle damage is identified, enrich the foundation of damaged vehicle judgement, so as to improve the accurate of vehicle damage identification
Degree.
According to the embodiment of another aspect, a kind of device of identification vehicle damage based on vehicle classification is also provided.Fig. 6 shows
Out according to the schematic block diagram of the device for the identification vehicle damage based on vehicle classification of one embodiment.As shown in fig. 6,
Device 600 for the identification vehicle damage based on vehicle classification includes: acquiring unit 61, is configured to obtain showing for damaged vehicle
Field information, field data include at least image or video；First determination unit 62 is configured to determine impaired vehicle according to field data
Component feature and damage characteristic, component feature is used to divide all parts of damaged vehicle；Second determination unit 63, configuration
For the vehicle classification for determining damaged vehicle according to field data, and determine according to vehicle classification the vehicle feature of damaged vehicle；Know
Other unit 64 is configured to the vehicle damage at least based on above-mentioned component feature, vehicle feature and damage characteristic identification damaged vehicle
As a result.
It is appreciated that the field data of damaged vehicle can be from damaged vehicle site acquire, for describe by
Damage the information of the form of vehicle.Field data at least may include image and/or video.In some implementations, field data is also
It may include threedimensional model and/or vehicles identifications.Wherein, vehicles identifications can include but is not limited to vehicle cab recognition code, license plate number,
It is one or more in motor number, Vehicle Identify Number, owner identity card number etc..
According to a kind of possible design, the second determination unit 63 may further be configured that information includes vehicle at the scene
In the case where mark, vehicle class corresponding with the vehicles identifications of damaged vehicle is searched in model data library based on mapping ruler
Not；In response to retrieving vehicle classification corresponding with the vehicles identifications of damaged vehicle, the vehicle classification retrieved is determined
For the vehicle classification of damaged vehicle.Wherein, model data library can store multiple vehicles identifications and multiple vehicle classifications, each
Vehicles identifications and the corresponding relationship of each vehicle classification are described by mapping ruler.In some implementations, model data library
In, for each vehicle classification, the shared various features of category vehicle, such as list of parts, appearance three can also be stored with
Dimension module, geometry feature and other features for vehicle vision identification.
According to alternatively possible design, the second determination unit 63 is also configured as: being extracted from field data impaired
The vehicle characteristics of vehicle；By vehicle characteristics input vehicle classification model, and based on the output result of vehicle classification model determine by
Damage the vehicle classification of vehicle.
In some implementations, the second determination unit 63 can first detect the predetermined mark of damaged vehicle from field data,
When not detecting predetermined mark, or can not accurately determine the vehicle classification of impaired vehicle according to predetermined mark, then pass through
Above-mentioned vehicle classification model determines the vehicle classification of damaged vehicle.Optionally, when can not pass through above-mentioned vehicle classification model determine
When the vehicle classification of damaged vehicle, new vehicle classification can also be established according to the vehicle characteristics of damaged vehicle, to vehicle number
Data extending is carried out according to library.
Vehicle feature determined by second determination unit 63 can be feature relevant to the vehicle of damaged vehicle.Vehicle
Feature for example can include but is not limited at least one of following: the position between component names, the material feature of component, component is special
Sign, etc..
According to one embodiment, recognition unit 64 may further be configured that by component feature, the vehicle feature and
Damage characteristic inputs the first non-destructive tests model, determines vehicle damage result according to the output result of the first non-destructive tests model.
According to another embodiment, recognition unit 64 may further be configured that using vehicle feature to component feature
It is updated；Updated component feature and damage characteristic are inputted into the second non-destructive tests model, according to the second non-destructive tests mould
The output result of type determines the vehicle damage result of damaged vehicle.
According to yet another embodiment, recognition unit 64 may further be configured that component feature and damage characteristic is defeated
Enter third non-destructive tests model, preliminary damage results are determined according to the output result of third non-destructive tests model；Utilize vehicle spy
Sign verifies preliminary damage results, and generates vehicle damage result according to check results.
According to a possible design, device 600 further includes third determination unit (not shown), is configured to be damaged according to vehicle
Hurt result and determines maintenance program and/or maintenance cost.
It is worth noting that device 600 shown in fig. 6 be with Fig. 2 shows the corresponding device of embodiment of the method implement
Example, Fig. 2 shows embodiment of the method in it is corresponding describe be equally applicable to device 600, details are not described herein.
The field data of damaged vehicle is made full use of, vehicle is passed through during identifying vehicle damage by apparatus above
Type feature and damage characteristic together identify vehicle damage, make full use of the more information of damaged vehicle, so as to mention
The accuracy of high vehicle damage identification.
According to the embodiment of another aspect, a kind of computer readable storage medium is also provided, is stored thereon with computer journey
Sequence enables computer execute method described in conjunction with Figure 2 when the computer program executes in a computer.
According to the embodiment of another further aspect, a kind of calculating equipment, including memory and processor, the memory are also provided
In be stored with executable code, when the processor executes the executable code, realize the method in conjunction with described in Fig. 2.
Those skilled in the art are it will be appreciated that in said one or multiple examples, function described in the invention
It can be realized with hardware, software, firmware or their any combination.It when implemented in software, can be by these functions
Storage in computer-readable medium or as on computer-readable medium one or more instructions or code transmitted.
Above-described specific embodiment has carried out further the purpose of the present invention, technical scheme and beneficial effects
It is described in detail, it should be understood that being not intended to limit the present invention the foregoing is merely a specific embodiment of the invention
Protection scope, all any modification, equivalent substitution, improvement and etc. on the basis of technical solution of the present invention, done should all
Including within protection scope of the present invention.
Claims (24)
1. a kind of method of the identification vehicle damage based on vehicle classification, which comprises
The field data of damaged vehicle is obtained, the field data includes at least image or video；
The component feature and damage characteristic of the damaged vehicle are determined according to the field data, the component feature is for dividing
The all parts of the damaged vehicle；
The vehicle classification of the damaged vehicle is determined according to the field data, and described be damaged is determined according to the vehicle classification
The vehicle feature of vehicle；
The vehicle damage of the damaged vehicle is at least identified based on the component feature, the vehicle feature and the damage characteristic
As a result.
2. according to the method described in claim 1, wherein, the field data further includes threedimensional model and/or vehicles identifications,
In, the vehicles identifications include at least one of the following: vehicle cab recognition code, license plate number, motor number, Vehicle Identify Number.
3. according to the method described in claim 2, wherein, the vehicle that the damaged vehicle is determined according to the field data
Classification includes:
In the case where the field data includes the vehicles identifications, retrieved in model data library and institute based on mapping ruler
State the corresponding vehicle classification of vehicles identifications, wherein the model data library at least stores multiple vehicles identifications and multiple vehicles
Classification, each vehicles identifications and the corresponding relationship of vehicle classification are described by the mapping ruler；
In response to retrieving vehicle classification corresponding with the vehicles identifications, the vehicle classification retrieved is determined as described
The vehicle classification of damaged vehicle.
4. according to the method described in claim 3, wherein:
The case where being retrieved in model data library less than vehicle classification corresponding with the vehicles identifications based on mapping ruler
Under, new vehicle classification corresponding with the damaged vehicle is added in the model data library.
5. according to the method described in claim 1, wherein, the vehicle that the damaged vehicle is determined according to the field data
Classification includes:
The vehicle characteristics of the damaged vehicle are extracted from the field data；
The vehicle characteristics are inputted into vehicle classification model, and based on the output result of the vehicle classification model determine it is described by
Damage the vehicle classification of vehicle.
6. according to the method described in claim 1, wherein, the vehicle feature includes at least one of the following:
Position feature, component between component names, number of components, the material feature of component, component are damaged incidence relation.
7. described to be at least based on the component feature, the vehicle feature and institute according to the method described in claim 1, wherein
It states damage characteristic and identifies that the vehicle damage result of the damaged vehicle includes:
The component feature, the vehicle feature and the damage characteristic are inputted into the first non-destructive tests model, according to described the
The output result of one non-destructive tests model determines the vehicle damage result.
8. described to be at least based on the component feature, the vehicle feature and institute according to the method described in claim 1, wherein
It states damage characteristic and identifies that the vehicle damage result of the damaged vehicle includes:
The component feature is updated using the vehicle feature；
Updated component feature and the damage characteristic are inputted into the second non-destructive tests model, according to second non-destructive tests
The output result of model determines the vehicle damage result.
9. described to be at least based on the component feature, the vehicle feature and institute according to the method described in claim 1, wherein
It states damage characteristic and identifies that the vehicle damage result of the damaged vehicle includes:
The component feature and the damage characteristic are inputted into third non-destructive tests model, according to the third non-destructive tests model
Output result determine preliminary damage results；
The preliminary damage results are verified using the vehicle feature, and generate the vehicle damage according to check results
As a result.
10. -9 any method according to claim 1, wherein the vehicle damage result includes: defective component, damage
Classification.
11. according to the method described in claim 10, wherein, the method also includes determining according to the vehicle damage result
Maintenance program and/or maintenance cost.
12. a kind of device of the identification vehicle damage based on vehicle classification, described device include:
Acquiring unit is configured to obtain the field data of damaged vehicle, and the field data includes at least image or video；
First determination unit is configured to determine the component feature and damage characteristic of the damaged vehicle according to the field data,
The component feature is used to divide all parts of the damaged vehicle；
Second determination unit is configured to determine the vehicle classification of the damaged vehicle according to the field data, and according to described
Vehicle classification determines the vehicle feature of the damaged vehicle；
Recognition unit, be configured at least based on the component feature, the vehicle feature and damage characteristic identification it is described by
Damage the vehicle damage result of vehicle.
13. device according to claim 12, wherein the field data further includes threedimensional model or vehicles identifications,
In, the vehicles identifications include at least one of the following: vehicle cab recognition code, license plate number, motor number, Vehicle Identify Number.
14. device according to claim 13, wherein second determination unit is further configured to:
In the case where the field data includes the vehicles identifications, searched in model data library and institute based on mapping ruler
State the corresponding vehicle classification of vehicles identifications, wherein the model data library at least stores multiple vehicles identifications and multiple vehicles
Classification, each vehicles identifications and the corresponding relationship of vehicle classification are described by the mapping ruler；
In response to retrieving vehicle classification corresponding with the vehicles identifications, the vehicle classification retrieved is determined as described
The vehicle classification of damaged vehicle.
15. device according to claim 14, wherein described device further includes model data library updating unit, is configured that
The case where being retrieved in model data library less than vehicle classification corresponding with the vehicles identifications based on mapping ruler
Under, new vehicle classification is added in the model data library, and the vehicle of the damaged vehicle is recorded in the mapping ruler
The corresponding relationship of type feature and new vehicle classification.
16. device according to claim 12, wherein second determination unit is additionally configured to:
The vehicle characteristics of the damaged vehicle are extracted from the field data；
The vehicle characteristics are inputted into vehicle classification model, and based on the output result of the vehicle classification model determine it is described by
Damage the vehicle classification of vehicle.
17. device according to claim 12, wherein the vehicle feature includes at least one of the following:
Position feature, components damage incidence relation between component names, number of components, the material feature of component, component.
18. device according to claim 12, wherein the recognition unit is further configured to:
The component feature, the vehicle feature and the damage characteristic are inputted into the first non-destructive tests model, according to described the
The output result of one non-destructive tests model determines the vehicle damage result.
19. device according to claim 12, wherein the recognition unit is further configured to:
The component feature is updated using the vehicle feature；
Updated component feature and the damage characteristic are inputted into the second non-destructive tests model, according to second non-destructive tests
The output result of model determines the vehicle damage result.
20. device according to claim 12, wherein the recognition unit is further configured to:
The component feature and the damage characteristic are inputted into third non-destructive tests model, according to the third non-destructive tests model
Output result determine preliminary damage results；
The preliminary damage results are verified using the vehicle feature, and generate the vehicle damage according to check results
As a result.
21. any device of 2-20 according to claim 1, wherein the vehicle damage result includes: defective component, damage
Hurt classification.
22. device according to claim 21, wherein described device further includes third determination unit, is configured to according to institute
It states vehicle damage result and determines maintenance program and/or maintenance cost.
23. a kind of computer readable storage medium, is stored thereon with computer program, when the computer program in a computer
When execution, computer perform claim is enabled to require the method for any one of 1-11.
24. a kind of calculating equipment, including memory and processor, which is characterized in that be stored with executable generation in the memory
Code realizes method of any of claims 1-11 when the processor executes the executable code.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810967661.XA CN109242006A (en) | 2018-08-23 | 2018-08-23 | The method and device of identification vehicle damage based on vehicle classification |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810967661.XA CN109242006A (en) | 2018-08-23 | 2018-08-23 | The method and device of identification vehicle damage based on vehicle classification |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109242006A true CN109242006A (en) | 2019-01-18 |
Family
ID=65069447
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810967661.XA Pending CN109242006A (en) | 2018-08-23 | 2018-08-23 | The method and device of identification vehicle damage based on vehicle classification |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109242006A (en) |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109934263A (en) * | 2019-01-31 | 2019-06-25 | 德联易控科技(北京)有限公司 | Vehicle identification method, device, electronic equipment and computer readable storage medium |
| CN110264444A (en) * | 2019-05-27 | 2019-09-20 | 阿里巴巴集团控股有限公司 | Damage detecting method and device based on weak segmentation |
| CN110287768A (en) * | 2019-05-06 | 2019-09-27 | 浙江君嘉智享网络科技有限公司 | Digital image recognition car damage identification method |
| CN110569865A (en) * | 2019-02-01 | 2019-12-13 | 阿里巴巴集团控股有限公司 | Method and device for recognizing vehicle body direction |
| CN110738576A (en) * | 2019-10-14 | 2020-01-31 | 支付宝(杭州)信息技术有限公司 | Method and device for generating damage assessment file for damaged vehicle |
| US10783643B1 (en) | 2019-05-27 | 2020-09-22 | Alibaba Group Holding Limited | Segmentation-based damage detection |
| 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 |
| 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 |
| CN113706315A (en) * | 2021-07-28 | 2021-11-26 | 中国银行保险信息技术管理有限公司 | Vehicle damage assessment method, device and equipment |
| CN113850263A (en) * | 2021-08-31 | 2021-12-28 | 深圳壹账通智能科技有限公司 | Vehicle model prediction method, device, equipment and medium based on image recognition |
| CN115880565A (en) * | 2022-12-06 | 2023-03-31 | 江苏凤火数字科技有限公司 | Neural network-based scraped vehicle identification method and system |
Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5950169A (en) * | 1993-05-19 | 1999-09-07 | Ccc Information Services, Inc. | System and method for managing insurance claim processing |
| US6185540B1 (en) * | 1994-12-28 | 2001-02-06 | Automatic Data Processing | Insurance estimating system |
| CN101242379A (en) * | 2008-03-18 | 2008-08-13 | 北京中车检信息技术有限公司 | Car damage identification method based on mobile communication terminal or network terminal |
| CN105223706A (en) * | 2015-09-28 | 2016-01-06 | 大连楼兰科技股份有限公司 | The method of vehicle degree of injury is judged for the intelligent glasses in vehicle repair and maintenance process |
| US20170293894A1 (en) * | 2016-04-06 | 2017-10-12 | American International Group, Inc. | Automatic assessment of damage and repair costs in vehicles |
| 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 |
| CN107392218A (en) * | 2017-04-11 | 2017-11-24 | 阿里巴巴集团控股有限公司 | A kind of car damage identification method based on image, device and electronic equipment |
| CN107748893A (en) * | 2017-09-29 | 2018-03-02 | 阿里巴巴集团控股有限公司 | Lift the method, apparatus and server of car damage identification image recognition result |
| CN107798284A (en) * | 2017-04-11 | 2018-03-13 | 平安科技（深圳）有限公司 | Car damage identification accessory matching process and system |
| CN107977895A (en) * | 2017-12-08 | 2018-05-01 | 北京小米移动软件有限公司 | Vehicle damages the definite method, apparatus and user equipment of information |
| US20180150805A1 (en) * | 2016-11-30 | 2018-05-31 | Thomas Edward Shaver | Method and system for determining condition report of assets |
| CN108364253A (en) * | 2018-03-15 | 2018-08-03 | 北京威远图易数字科技有限公司 | Car damage identification method, system and electronic equipment |
-
2018
- 2018-08-23 CN CN201810967661.XA patent/CN109242006A/en active Pending
Patent Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5950169A (en) * | 1993-05-19 | 1999-09-07 | Ccc Information Services, Inc. | System and method for managing insurance claim processing |
| US6185540B1 (en) * | 1994-12-28 | 2001-02-06 | Automatic Data Processing | Insurance estimating system |
| CN101242379A (en) * | 2008-03-18 | 2008-08-13 | 北京中车检信息技术有限公司 | Car damage identification method based on mobile communication terminal or network terminal |
| CN105223706A (en) * | 2015-09-28 | 2016-01-06 | 大连楼兰科技股份有限公司 | The method of vehicle degree of injury is judged for the intelligent glasses in vehicle repair and maintenance process |
| US20170293894A1 (en) * | 2016-04-06 | 2017-10-12 | American International Group, Inc. | Automatic assessment of damage and repair costs in vehicles |
| US20180150805A1 (en) * | 2016-11-30 | 2018-05-31 | Thomas Edward Shaver | Method and system for determining condition report of assets |
| 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 |
| CN107392218A (en) * | 2017-04-11 | 2017-11-24 | 阿里巴巴集团控股有限公司 | A kind of car damage identification method based on image, device and electronic equipment |
| CN107798284A (en) * | 2017-04-11 | 2018-03-13 | 平安科技（深圳）有限公司 | Car damage identification accessory matching process and system |
| CN107748893A (en) * | 2017-09-29 | 2018-03-02 | 阿里巴巴集团控股有限公司 | Lift the method, apparatus and server of car damage identification image recognition result |
| CN107977895A (en) * | 2017-12-08 | 2018-05-01 | 北京小米移动软件有限公司 | Vehicle damages the definite method, apparatus and user equipment of information |
| CN108364253A (en) * | 2018-03-15 | 2018-08-03 | 北京威远图易数字科技有限公司 | Car damage identification method, system and electronic equipment |
Non-Patent Citations (3)
| Title |
|---|
| 杨少彬: "《新型神龙富康轿车结构与维修》", 31 January 2001 * |
| 欧冬秀: "《交通信息技术》", 31 May 2007 * |
| 高佩华: "《安检仪器使用与维护》", 31 July 2015 * |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109934263A (en) * | 2019-01-31 | 2019-06-25 | 德联易控科技(北京)有限公司 | Vehicle identification method, device, electronic equipment and computer readable storage medium |
| CN110569865A (en) * | 2019-02-01 | 2019-12-13 | 阿里巴巴集团控股有限公司 | Method and device for recognizing vehicle body direction |
| CN110569865B (en) * | 2019-02-01 | 2020-07-17 | 阿里巴巴集团控股有限公司 | Method and device for recognizing vehicle body direction |
| CN110287768A (en) * | 2019-05-06 | 2019-09-27 | 浙江君嘉智享网络科技有限公司 | Digital image recognition car damage identification method |
| CN110264444A (en) * | 2019-05-27 | 2019-09-20 | 阿里巴巴集团控股有限公司 | Damage detecting method and device based on weak segmentation |
| CN110264444B (en) * | 2019-05-27 | 2020-07-17 | 阿里巴巴集团控股有限公司 | Damage detection method and device based on weak segmentation |
| US10783643B1 (en) | 2019-05-27 | 2020-09-22 | Alibaba Group Holding Limited | Segmentation-based damage detection |
| US11004204B2 (en) | 2019-05-27 | 2021-05-11 | Advanced New Technologies Co., Ltd. | Segmentation-based damage detection |
| CN110738576A (en) * | 2019-10-14 | 2020-01-31 | 支付宝(杭州)信息技术有限公司 | Method and device for generating damage assessment file for damaged vehicle |
| CN110738576B (en) * | 2019-10-14 | 2021-08-06 | 支付宝(杭州)信息技术有限公司 | Method and device for generating damage assessment file for damaged vehicle |
| 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 |
| 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 |
| CN113706315A (en) * | 2021-07-28 | 2021-11-26 | 中国银行保险信息技术管理有限公司 | Vehicle damage assessment method, device and equipment |
| CN113850263A (en) * | 2021-08-31 | 2021-12-28 | 深圳壹账通智能科技有限公司 | Vehicle model prediction method, device, equipment and medium based on image recognition |
| CN115880565A (en) * | 2022-12-06 | 2023-03-31 | 江苏凤火数字科技有限公司 | Neural network-based scraped vehicle identification method and system |
| CN115880565B (en) * | 2022-12-06 | 2023-09-05 | 江苏凤火数字科技有限公司 | Neural network-based scraped vehicle identification method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109242006A (en) | The method and device of identification vehicle damage based on vehicle classification | |
| US10628890B2 (en) | Visual analytics based vehicle insurance anti-fraud detection | |
| CN110209866A (en) | A kind of image search method, device, equipment and computer readable storage medium | |
| CN105448103B (en) | Vehicle fake-license detection method and system | |
| CN110753892A (en) | Method and system for instant object tagging via cross-modality verification in autonomous vehicles | |
| CN110869559A (en) | Method and system for integrated global and distributed learning in autonomous vehicles | |
| CN110799982A (en) | Method and system for object-centric stereo vision in an autonomous vehicle | |
| CN110135318B (en) | Method, device, equipment and storage medium for determining passing record | |
| Hinz | Detection and counting of cars in aerial images | |
| CN109740479A (en) | A kind of vehicle recognition methods, device, equipment and readable storage medium storing program for executing again | |
| US20200105009A1 (en) | Methods and Systems for Detecting and Assigning Attributes to Objects of Interest in Geospatial Imagery | |
| CN111126393A (en) | Vehicle appearance refitting judgment method and device, computer equipment and storage medium | |
| CN111369801B (en) | Vehicle identification method, device, equipment and storage medium | |
| CN110378254B (en) | Method and system for identifying vehicle damage image modification trace, electronic device and storage medium | |
| CN111753610B (en) | Weather identification method and device | |
| CN109784171A (en) | Car damage identification method for screening images, device, readable storage medium storing program for executing and server | |
| US20210256738A1 (en) | Computer-implemented method and system for generating a virtual vehicle environment | |
| CN115810134B (en) | Image acquisition quality inspection method, system and device for vehicle insurance anti-fraud | |
| US11120308B2 (en) | Vehicle damage detection method based on image analysis, electronic device and storage medium | |
| CN115830399A (en) | Classification model training method, apparatus, device, storage medium, and program product | |
| CN114140025A (en) | Multi-modal data-oriented vehicle insurance fraud behavior prediction system, method and device | |
| WO2022064486A1 (en) | Vehicle image analysis | |
| CN111259887B (en) | Intelligent quality inspection method, system and device for dumb resource equipment | |
| CN106339683A (en) | Road test traffic light identification method and device | |
| CN114187585A (en) | Intelligent identification method and system for shielded license plate vehicle |
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 | ||
| REG | Reference to a national code |
Ref country code: HK Ref legal event code: DE Ref document number: 40003295 Country of ref document: HK |
|
| TA01 | Transfer of patent application right |
Effective date of registration: 20200925 Address after: Cayman Enterprise Centre, 27 Hospital Road, George Town, Grand Cayman Islands Applicant after: Innovative advanced technology Co.,Ltd. Address before: Cayman Enterprise Centre, 27 Hospital Road, George Town, Grand Cayman Islands Applicant before: Advanced innovation technology Co.,Ltd. Effective date of registration: 20200925 Address after: Cayman Enterprise Centre, 27 Hospital Road, George Town, Grand Cayman Islands Applicant after: Advanced innovation technology Co.,Ltd. Address before: A four-storey 847 mailbox in Grand Cayman Capital Building, British Cayman Islands Applicant before: Alibaba Group Holding Ltd. |
|
| TA01 | Transfer of patent application right | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190118 |
|
| RJ01 | Rejection of invention patent application after publication |