CN111652274B

CN111652274B - Vehicle type recognition method and device based on image recognition and computer equipment

Info

Publication number: CN111652274B
Application number: CN202010351240.1A
Authority: CN
Inventors: 张焘
Original assignee: Ping An Property and Casualty Insurance Company of China Ltd
Current assignee: Ping An Property and Casualty Insurance Company of China Ltd
Priority date: 2020-04-28
Filing date: 2020-04-28
Publication date: 2024-05-14
Anticipated expiration: 2040-04-28
Also published as: CN111652274A

Abstract

The invention relates to artificial intelligence technology, and discloses a vehicle type recognition method, a device, computer equipment and a storage medium based on image recognition, wherein the method receives vehicle photo data uploaded by a user side according to photo uploading guiding information after judging that the distance between the user side positioning information and the vehicle machine positioning information does not exceed a preset distance threshold; the vehicle photo data at least comprises a vehicle front view picture, a vehicle rear view picture and a vehicle interior part picture; and then, identifying the vehicle front view picture and the vehicle rear view picture to determine the vehicle model information, and identifying the vehicle interior trim part picture to determine the vehicle configuration level. Meanwhile, the method relates to a block chain technology, image recognition is carried out based on vehicle photo data uploaded by a user side according to photo uploading guide information, the vehicle type is determined through the obtained vehicle type information and the vehicle configuration grade, and the accuracy of vehicle type recognition is improved.

Description

Vehicle type recognition method and device based on image recognition and computer equipment

Technical Field

The present invention relates to the field of artificial intelligence technologies, and in particular, to a vehicle type recognition method and apparatus based on image recognition, a computer device, and a storage medium.

Background

At present, automobile type recognition is widely applied in many scenes, such as online insurance application (namely, application of uploading vehicle information to a user on an insurance application platform), vehicle violation escape, accident loss and payment and the like.

However, the existing vehicle type recognition is generally based on the characteristics of scale invariant feature conversion and the like, and then training is carried out based on a classifier such as a support vector machine and the like, so that the vehicle type and license plate information are recognized in a limited mode, the accuracy of the existing vehicle type recognition of different manufacturers is extremely low, and the recognition failure rate is high.

Disclosure of Invention

The embodiment of the invention provides a vehicle type recognition method, device, computer equipment and storage medium based on image recognition, and aims to solve the problems that in the prior art, vehicle type recognition is generally based on characteristics such as scale-invariant feature conversion and the like, and then training is carried out based on a classifier such as a support vector machine and the like, so that the accuracy rate of vehicle type recognition is low and the recognition failure rate is high.

In a first aspect, an embodiment of the present invention provides a vehicle type recognition method based on image recognition, including:

Receiving the user side positioning information uploaded by the user side and receiving the vehicle machine positioning information uploaded by the vehicle machine;

judging whether the distance between the user side positioning information and the vehicle-mounted positioning information exceeds a preset distance threshold value or not;

if the distance between the user side positioning information and the vehicle machine positioning information does not exceed the distance threshold, receiving vehicle photo data uploaded by the user side according to photo uploading guiding information; the vehicle photo data at least comprises a vehicle front view picture, a vehicle rear view picture and a vehicle interior part picture; the vehicle interior part pictures at least comprise car machine pictures, steering wheel pictures, skylight pictures, transmission pictures, seat pictures and air conditioner air outlet pictures;

judging whether the vehicle photo data comprises a new vehicle qualification picture or not;

If the vehicle photo data does not comprise the new vehicle qualification picture, carrying out image recognition according to the vehicle forward-looking picture and the vehicle rear-looking picture to obtain a corresponding vehicle feature sequence;

Taking a vehicle characteristic sequence corresponding to the vehicle photo data as input of a first convolutional neural network model trained in advance to obtain vehicle model information corresponding to the vehicle photo data; the first convolutional neural network model is used for identifying a vehicle model;

Performing image recognition according to the vehicle interior part picture to obtain a corresponding vehicle interior part sequence;

Taking a vehicle interior part sequence corresponding to the vehicle photo data as input of a pre-trained second convolutional neural network model to obtain a vehicle configuration grade corresponding to the vehicle photo data; wherein the second convolutional neural network model is used for identifying a vehicle configuration level;

judging whether the vehicle configuration level exceeds a preset configuration level threshold value or not; and

If the vehicle configuration level exceeds the configuration level threshold, adding suspicious vehicle identification to the vehicle data corresponding to the vehicle photo data to obtain current first vehicle data, and storing the current first vehicle data into a preset first storage area.

In a second aspect, an embodiment of the present invention provides a vehicle type recognition device based on image recognition, including:

the first positioning information acquisition unit is used for receiving the user side positioning information uploaded by the user side and receiving the vehicle machine positioning information uploaded by the vehicle machine;

the first distance judging unit is used for judging whether the distance between the user side positioning information and the vehicle-machine positioning information exceeds a preset distance threshold value;

The vehicle photo data receiving unit is used for receiving vehicle photo data uploaded by the user side according to the photo uploading guiding information if the distance between the user side positioning information and the vehicle machine positioning information does not exceed the distance threshold value; the vehicle photo data at least comprises a vehicle front view picture, a vehicle rear view picture and a vehicle interior part picture; the vehicle interior part pictures at least comprise car machine pictures, steering wheel pictures, skylight pictures, transmission pictures, seat pictures and air conditioner air outlet pictures;

the qualification picture judging unit is used for judging whether the vehicle photo data comprise a new vehicle qualification picture or not;

the vehicle feature sequence acquisition unit is used for carrying out image recognition according to the vehicle front view picture and the vehicle rear view picture to obtain a corresponding vehicle feature sequence if the vehicle photo data does not comprise a new vehicle qualification picture;

a vehicle model information obtaining unit, configured to obtain vehicle model information corresponding to the vehicle photograph data by using a vehicle feature sequence corresponding to the vehicle photograph data as an input of a first convolutional neural network model trained in advance; the first convolutional neural network model is used for identifying a vehicle model;

a vehicle interior part sequence acquisition unit, configured to perform image recognition according to the vehicle interior part picture, so as to obtain a corresponding vehicle interior part sequence;

A vehicle configuration level obtaining unit, configured to obtain a vehicle configuration level corresponding to the vehicle photo data by using a vehicle interior part sequence corresponding to the vehicle photo data as an input of a second convolutional neural network model trained in advance; wherein the second convolutional neural network model is used for identifying a vehicle configuration level;

a configuration level judging unit for judging whether the vehicle configuration level exceeds a preset configuration level threshold; and

And the first vehicle data storage unit is used for adding suspicious vehicle identifications to the vehicle data corresponding to the vehicle photo data to obtain current first vehicle data if the vehicle configuration level exceeds the configuration level threshold value, and storing the current first vehicle data into a preset first storage area.

In a third aspect, an embodiment of the present invention further provides a computer device, which includes a memory, a processor, and a computer program stored in the memory and capable of running on the processor, where the processor executes the computer program to implement the vehicle type recognition method based on image recognition according to the first aspect.

In a fourth aspect, an embodiment of the present invention further provides a computer readable storage medium, where the computer readable storage medium stores a computer program, where the computer program when executed by a processor causes the processor to perform the vehicle model recognition method based on image recognition described in the first aspect.

The embodiment of the invention provides a vehicle type recognition method, a vehicle type recognition device, computer equipment and a storage medium based on image recognition, which comprise the steps of receiving user side positioning information uploaded by a user side and receiving vehicle machine positioning information uploaded by a vehicle machine; if the distance between the user side positioning information and the vehicle machine positioning information does not exceed the distance threshold, receiving vehicle photo data uploaded by the user side according to photo uploading guiding information; if the vehicle photo data does not comprise the new vehicle qualification picture, carrying out image recognition according to the vehicle forward-looking picture and the vehicle rear-looking picture to obtain a corresponding vehicle feature sequence; taking a vehicle characteristic sequence corresponding to the vehicle photo data as input of a first convolutional neural network model trained in advance to obtain vehicle model information corresponding to the vehicle photo data; performing image recognition according to the vehicle interior part picture to obtain a corresponding vehicle interior part sequence; taking a vehicle interior part sequence corresponding to the vehicle photo data as input of a pre-trained second convolutional neural network model to obtain a vehicle configuration grade corresponding to the vehicle photo data; if the vehicle configuration level exceeds the configuration level threshold, adding suspicious vehicle identification to the vehicle data corresponding to the vehicle photo data to obtain current first vehicle data, and storing the current first vehicle data into a preset first storage area. The method realizes image recognition based on the vehicle photo data uploaded by the user side according to the photo uploading guide information, and the vehicle model is determined according to the obtained vehicle model information and the vehicle configuration grade, so that the accuracy of vehicle model recognition is improved.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.

Fig. 1 is a schematic view of an application scenario of a vehicle type recognition method based on image recognition according to an embodiment of the present invention;

fig. 2 is a schematic flow chart of a vehicle type recognition method based on image recognition according to an embodiment of the present invention;

fig. 3 is a schematic block diagram of a vehicle type recognition device based on image recognition according to an embodiment of the present invention;

Fig. 4 is a schematic block diagram of a computer device according to an embodiment of the present invention.

Detailed Description

The following description of the embodiments of the present invention will be made clearly and fully with reference to the accompanying drawings, in which it is evident that the embodiments described are some, but not all embodiments of the invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.

It should be understood that the terms "comprises" and "comprising," when used in this specification and the appended claims, specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof.

It is also to be understood that the terminology used in the description of the invention herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. As used in this specification and the appended claims, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise.

It should be further understood that the term "and/or" as used in the present specification and the appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes such combinations.

The application relates to an artificial intelligence technology, referring to fig. 1 and 2, fig. 1 is a schematic diagram of an application scenario of a vehicle type recognition method based on image recognition provided by an embodiment of the application; fig. 2 is a flow chart of a vehicle type recognition method based on image recognition, which is provided by the embodiment of the application, and is applied to a server, and the method is executed by application software installed in the server.

As shown in fig. 2, the method includes steps S101 to S110.

S101, receiving the user side positioning information uploaded by the user side and receiving the vehicle machine positioning information uploaded by the vehicle machine.

In this embodiment, in order to more clearly understand the technical solution, the following describes the related terminal in detail. The application relates to a technical scheme for describing the server.

The user side, that is, the intelligent terminal (such as a smart phone, a tablet personal computer, etc.) used by the user is used for acquiring vehicle photos (such as a vehicle front view picture, a vehicle rear view picture, and a vehicle interior trim component picture), and also can acquire user side positioning information (that is, positioning the longitude and latitude of the current position of the user side), and upload the acquired vehicle photos and the user side positioning information to the server. The general user end establishes connection with the server, namely, firstly, the user name and the password are input after the APP application program provided by the server is downloaded by the user end, so that the APP application program is used as a medium for data interaction between the user end and the server.

The car machine, that is, the multimedia terminal installed at one side of the steering wheel of the car. The vehicle machine can also download the APP application program which is the same as the user side and log in by using the user name and the password which are the same as the user side, and at the moment, the vehicle machine can acquire the positioning information of the vehicle machine and upload the positioning information to the server after logging in the APP application program.

And the server is used for receiving the pictures uploaded by the user side so as to identify the vehicle type. And the method can also judge whether the user is photographed nearby the vehicle currently after receiving the positioning information of the user side and the positioning information of the vehicle machine so as to ensure the authenticity of the picture and prevent the picture uploaded by the user side from being acquired from the Internet or other terminals.

For example, when a user purchases a car and needs to perform online insurance, the server side needs to determine the car type and the car configuration level first, and at this time, the user needs to start the APP application program of the user side and log in, and at the same time, also start the APP application program of the car machine and log in. At this time, the user terminal can automatically upload the positioning information of the user terminal, and the vehicle machine can also automatically upload the positioning information of the vehicle machine. The server receives the user side positioning information uploaded by the user side and receives the vehicle machine positioning information uploaded by the vehicle machine.

S102, judging whether the distance between the user side positioning information and the vehicle machine positioning information exceeds a preset distance threshold.

In this embodiment, when the server receives the user side positioning information uploaded by the user side and receives the vehicle positioning information uploaded by the vehicle, in order to determine whether the user actually shoots around the vehicle and then performs online insurance application, at this time, the server side may acquire a corresponding first positioning point according to the user side positioning information and acquire a corresponding second positioning point according to the vehicle positioning information. At this time, the distance between the first positioning point and the second positioning point is calculated, and then whether the distance between the first positioning point and the second positioning point exceeds the distance threshold is judged (for example, the distance threshold can be set to be 10m, 20m or 30m, etc., and the value is set according to the actual detection precision requirement). If the distance between the user side positioning information and the vehicle machine positioning information does not exceed the preset distance threshold (for example, the distance threshold is set to 0-30 m), which means that the current distance between the user and the vehicle is very close (the user can be understood to be at the side of the vehicle), at this time, the server sends photo uploading guiding information to the user side so as to prompt the user to upload at least the front view picture, the rear view picture and the interior trim part picture of the vehicle.

S103, if the distance between the user side positioning information and the vehicle machine positioning information does not exceed the distance threshold, receiving vehicle photo data uploaded by the user side according to photo uploading guiding information; the vehicle photo data at least comprises a vehicle front view picture, a vehicle rear view picture and a vehicle interior part picture; the vehicle interior part pictures at least comprise car machine pictures, steering wheel pictures, skylight pictures, transmission pictures, seat pictures and air conditioner air outlet pictures.

In this embodiment, the user is generally prompted to upload the front view image, the rear view image, the new license image, and the interior trim part image of the vehicle in the photo upload guide information. In the process of uploading the front view picture, the rear view picture, the new license picture and the interior trim part picture of the vehicle, if some pictures are not available (if the new license picture is not obtained), uploading prompts of the pictures can be ignored, but at least the front view picture, the rear view picture and the interior trim part picture of the vehicle are required to be uploaded.

When the vehicle interior trim part picture is specifically prompted to be uploaded, the vehicle interior trim part picture, the steering wheel picture, the skylight picture, the transmission picture, the seat picture and the air conditioner air outlet picture are mainly prompted to be uploaded. The vehicle configuration level recognition can be further performed by these main in-vehicle photographs.

In one embodiment, step S103 further includes:

Acquiring corresponding picture positioning information according to the photo parameter data included in the vehicle photo data;

Judging whether the distance between the picture positioning information and the car machine positioning information exceeds the distance threshold value or not; if the distance between the picture positioning information and the vehicle-mounted positioning information does not exceed the distance threshold, executing the step of judging whether the vehicle photo data comprise a new vehicle qualification picture or not; if the distance between the picture positioning information and the vehicle-mounted positioning information exceeds the distance threshold, executing the step of sending prompt information for prompting the user to approach the vehicle to the user side;

And sending prompt information for prompting the user to approach the vehicle to the user side.

In this embodiment, in order to further determine that the photo positioning information corresponding to the vehicle photo data uploaded by the user is consistent with the user side positioning information, the vehicle photo data may be analyzed first to obtain the corresponding photo positioning information. The image shot by the camera at the user side generally comprises attribute information such as shooting time, shooting position and the like, and at the moment, each vehicle image in the vehicle photo data is analyzed, so that corresponding image positioning information can be obtained.

And judging whether the distance between the picture positioning information and the vehicle positioning information exceeds the distance threshold, namely acquiring a corresponding third positioning point according to the picture positioning information, and calculating the distance between the third positioning point and the second positioning point at the moment because the vehicle positioning information is known to acquire a corresponding second positioning point, wherein the distance between the third positioning point and the second positioning point can be used as the distance between the picture positioning information and the vehicle positioning information. That is, the distance between the picture positioning information and the car machine positioning information does not exceed the distance threshold, which indicates that the distance between the user and the car is very close (it can be understood that the user is at the car). And if the distance between the picture positioning information and the vehicle-mounted positioning information exceeds the distance threshold, sending prompt information for prompting a user to approach the vehicle.

S104, judging whether the vehicle photo data comprises a new vehicle qualification picture or not.

In this embodiment, if the vehicle photo data does not include a new vehicle license picture, the vehicle type cannot be intuitively identified according to the data of the new vehicle license picture, and at this time, the vehicle type identification can only be performed according to the vehicle front view picture and the vehicle rear view picture as the picture material. If the vehicle photo data comprises the new vehicle qualification picture, the vehicle type can be intuitively identified according to the data of the new vehicle qualification picture.

And S105, if the vehicle photo data does not comprise the new vehicle qualification picture, carrying out image recognition according to the vehicle forward-looking picture and the vehicle rear-looking picture to obtain a corresponding vehicle characteristic sequence.

In this embodiment, if the vehicle type cannot be intuitively identified according to the data of the new vehicle qualification picture, the front cover logo in the vehicle front view picture, the rear cover logo of the vehicle rear view picture, and the corresponding marking values of the displacement are obtained through a preset first marking strategy, so as to form a corresponding vehicle feature sequence.

For example, in a preset first labeling strategy, labeling 1 when the front cover logo is a mass logo, labeling 2 when the front cover logo is honda, labeling 3 when the front cover logo is a bidi logo, and the like; the rear cover car logo is marked 1 when the public car logo is marked 1, marked 2 when the rear cover car logo is in Honda, marked 3 when the rear cover car logo is in a Biedi time mark, and the like; manual displacement is 1.8, automatic displacement is 1.8, 2, manual displacement is 2.0, 3 and automatic displacement is 2.0, 4. For example, if the corresponding vehicle feature sequence is [114] after a certain vehicle is marked, the vehicle is automatically represented as 2.0 of the public.

More specifically, 3 image recognition models (a convolutional neural network model for front cover logo recognition, a convolutional neural network model for rear cover logo recognition, and a vehicle displacement OCR image-text recognition model) can be trained in advance to perform image recognition on the front view image and the rear view image of the vehicle. When the pixel matrix corresponding to the vehicle front view image is used as the input of the convolutional neural network model identified by the vehicle front cover logo, the output result is a front cover logo value; when the pixel matrix corresponding to the vehicle rear view image is used as the input of the convolutional neural network model identified by the vehicle rear cover logo, the output result is a rear cover logo value; and directly identifying a digital character string corresponding to the vehicle displacement in the vehicle rearview picture by using the vehicle displacement OCR image-text identification model, and taking the digital character string as a displacement marking value. The convolutional neural network model for identifying the front cover logo and the convolutional neural network model for identifying the rear cover logo of the automobile are obtained by training specific picture identification scenes, and are the same as the training process of the conventional convolutional neural network, but different in picture set used in the training process. The convolutional neural network model training for front cover logo recognition is training by using a large number of front cover pictures, and the convolutional neural network model training for rear cover logo recognition is training by using a large number of rear cover pictures.

In one embodiment, after step S104, the method further includes:

If the vehicle photo data comprises a new vehicle qualification picture, sequentially carrying out graying, edge detection, binarization and filtering treatment on the new vehicle qualification picture to obtain a candidate region corresponding to the new vehicle qualification picture;

And correspondingly acquiring vehicle model information corresponding to the vehicle photo data according to the characters in the candidate area.

In this embodiment, if the vehicle photograph data includes a new vehicle certification picture, the vehicle model may be intuitively identified according to the data of the new vehicle certification picture. Since the new license image is generally a color image, the color image contains more information, but if the color image is directly processed, the execution speed will be reduced and the storage space will be increased. The graying of the color picture is a basic method for image processing, is widely applied in the field of pattern recognition, and reasonable graying is helpful for the extraction and subsequent processing of image information, can save storage space and accelerate processing speed.

The edge detection method is to examine the gray level change condition of the pixels of the image in a certain field and identify the points with obvious brightness change in the digital image. The edge detection of the image can greatly reduce the data volume, eliminate irrelevant information and save important structural attributes of the image. The operators used for edge detection are a lot, and commonly used are Sobel operators (namely a Sobel operator), laplacian edge detection operators (namely a Laplacian edge detection operator), canny edge detection operators (namely a Canny edge detection operator) and the like.

After the gray level image is subjected to edge detection, characters and edge information on the new vehicle certification image are highlighted. Meanwhile, edge texture features of frames of other non-characters and non-new car qualification pictures are also highlighted, and in order to reduce the influence of noise, binarization processing is required to be carried out on the new car qualification pictures, wherein the binarization is a type of thresholding of images. According to the selection condition of the threshold value, the binarization method can be divided into a global threshold value method, a dynamic threshold value method and a local threshold value method, thresholding is carried out by using a maximum inter-class variance method (also called Otsu algorithm) to remove some pixels with smaller gradient values, the license plate range needing to be searched is reduced, and the pixel value of a license plate image after binarization processing is 0 or 255.

Then, the noise of the target image is suppressed under the condition of retaining the detail characteristics of the image as much as possible, which is an indispensable operation for eliminating the noise in the image processing, and the quality of the processing result directly influences the effectiveness and reliability of processing and analyzing the subsequent image. There are many common filtering methods, such as median filtering, morphological filtering, gaussian filtering, bilateral filtering, etc. And sequentially carrying out graying, edge detection, binarization and filtering treatment on the new vehicle qualification picture to obtain a candidate region corresponding to the new vehicle qualification picture.

After the candidate region is acquired, a character segmentation algorithm may be referenced to identify vehicle model information for the candidate region.

S106, taking a vehicle characteristic sequence corresponding to the vehicle photo data as input of a first convolutional neural network model trained in advance, and obtaining vehicle model information corresponding to the vehicle photo data; wherein the first convolutional neural network model is used to identify a vehicle model.

In this embodiment, when the vehicle feature sequence as in [114] is acquired, the vehicle feature sequence is used as an input of the first convolutional neural network model trained in advance, and the recognition result (i.e., the vehicle model information) can be output. Through the first convolutional neural network model, intelligent recognition of the vehicle model is achieved.

In one embodiment, step S106 further includes:

Acquiring vehicle appearance pictures in the webpage according to the acquisition corresponding to the preset first website list so as to form a vehicle appearance picture set;

According to a preset first labeling strategy, acquiring a front cover logo, a rear cover logo and a corresponding labeling value of displacement of each vehicle appearance picture in the vehicle appearance picture set so as to form each vehicle appearance picture and a corresponding vehicle characteristic sequence;

And taking the vehicle characteristic sequence corresponding to the vehicle appearance picture in the vehicle appearance picture set as input of a first convolutional neural network to be trained, taking the vehicle model information corresponding to the vehicle appearance picture as output of the first convolutional neural network to be trained, and training the first convolutional neural network to be trained to obtain a first convolutional neural network model for identifying the vehicle model.

In this embodiment, in order to train to obtain the first convolutional neural network model, a large number of automobile appearance pictures are acquired by corresponding to a preset first website list, so as to form a training set (i.e., a vehicle appearance picture set) of the first convolutional neural network to be trained. And then combining a first marking strategy and adopting a manual marking mode to realize marking values respectively corresponding to the front cover marks, the rear cover marks and the displacement of each vehicle appearance picture in the vehicle appearance picture set so as to form each vehicle appearance picture and a corresponding vehicle characteristic sequence. Meanwhile, the vehicle model information corresponding to each vehicle appearance picture in the vehicle appearance picture set is marked.

And finally, taking the vehicle characteristic sequence corresponding to the vehicle appearance picture in the vehicle appearance picture set as the input of the first convolutional neural network to be trained, taking the vehicle model information corresponding to the vehicle appearance picture as the output of the first convolutional neural network to be trained, and training the first convolutional neural network to be trained to obtain a first convolutional neural network model. Through the training process, the first convolutional neural network model for identifying the vehicle model according to the vehicle appearance picture can be obtained quickly.

And S107, carrying out image recognition according to the vehicle interior part picture to obtain a corresponding vehicle interior part sequence.

In this embodiment, if the vehicle interior part picture needs to be identified, the corresponding marking values of the vehicle machine, the steering wheel, the sunroof, the transmission, the seat and the air conditioner air outlet in the vehicle interior part picture are obtained through a preset second marking strategy, so as to form a corresponding vehicle interior part sequence.

In one embodiment, step S107 includes:

Identifying and acquiring a vehicle machine, a steering wheel, a skylight, a speed changer, a seat and an air conditioner air outlet in the vehicle interior part picture;

And according to a preset second labeling strategy, obtaining the corresponding labeling values of the vehicle machine, the steering wheel, the skylight, the transmission, the seat and the air conditioner air outlet in the vehicle interior part picture so as to form a vehicle interior part sequence.

In this embodiment, after each vehicle interior part in the vehicle interior part picture is identified and acquired, the listed 6 judgment standards are checked one by one (the 6 judgment standards are preset labeling strategies) to obtain a corresponding vehicle interior part sequence. The second marking strategy is that the marking strategy is marked as 1 when the large-size screen car machine exists, and marked as 0 when the large-size screen car machine does not exist; marking 1 when the multifunctional steering wheel is used and marking 0 when the multifunctional steering wheel is not used; marking 1 when a skylight exists, and marking 0 when no skylight exists; the automatic transmission is marked as 1, and the manual transmission is marked as 0; when the leather seat is marked as 1, the leather seat is not marked as 0; when the rear row is provided with an air conditioner air outlet, the air conditioner air outlet is marked as 1, and when the rear row is not provided with the air conditioner air outlet, the air conditioner air outlet is marked as 0; for example, a vehicle is labeled [111111], and the vehicle interior part is a large-sized screen vehicle, a multifunctional steering wheel, a sunroof, an automatic transmission, a leather seat, and an air-conditioning outlet in the rear row.

Similarly, the process of obtaining the corresponding indication values of the vehicle machine, the steering wheel, the skylight, the transmission, the seat and the air conditioner air outlet in the vehicle interior part picture can also refer to the process of obtaining the corresponding indication values through the corresponding convolutional neural network model.

S108, taking a vehicle interior part sequence corresponding to the vehicle photo data as input of a pre-trained second convolutional neural network model, and obtaining a vehicle configuration grade corresponding to the vehicle photo data; wherein the second convolutional neural network model is used to identify a vehicle configuration level.

In this embodiment, when the vehicle interior part sequence corresponding to the vehicle photo data is acquired, the corresponding vehicle configuration level may be obtained by using the vehicle interior part sequence as the input of the second convolutional neural network model trained in advance. For example, [110100] is input to the second convolutional neural network model at this time, and a corresponding vehicle configuration level (e.g., 2) is obtained. The vehicle configuration grade is effectively judged through the convolutional neural network model, and the actual value of the vehicle is determined in the process of vehicle insurance claim settlement.

S109, judging whether the vehicle configuration level exceeds a preset configuration level threshold.

In this embodiment, in order to prevent the user from retrofitting the vehicle when online insurance is performed on the vehicle, it is necessary to determine whether to retrofit the vehicle according to the vehicle configuration level.

And S110, if the vehicle configuration level exceeds the configuration level threshold, adding suspicious vehicle identification to the vehicle data corresponding to the vehicle photo data to obtain current first vehicle data, and storing the current first vehicle data into a preset first storage area.

In this embodiment, the second convolutional neural network model is trained, and some data of the modified vehicles are added for training, and the vehicle configuration grades corresponding to the modified vehicles are generally higher. If the vehicle configuration level exceeds the configuration level threshold previously described (e.g., the vehicle configuration level is 5 and the configuration level threshold is set to 4), this indicates that the vehicle may be a retrofit vehicle and is not within the range of the application. At the moment, the vehicle data corresponding to the vehicle photo data are added with suspicious vehicle identifications to obtain current first vehicle data, and the current first vehicle data are stored in a preset first storage area. And the first storage area stores all current first vehicle data with suspicious vehicle identifications. Further manual verification may be performed in the server for each current first vehicle data in the first storage area.

In one embodiment, step S110 further includes:

If the vehicle configuration level does not exceed the configuration level threshold, acquiring unique vehicle type data corresponding to the vehicle picture to be identified according to the vehicle model information and the vehicle configuration level, adding the unique vehicle type data to the vehicle data corresponding to the vehicle photo data to obtain current second vehicle data, and storing the current second vehicle data in a preset second storage area.

In an embodiment, the preset first storage area and the preset second storage area are respectively blocks on a blockchain network, the first vehicle data and the second vehicle data are respectively stored in the blocks created on the blockchain, and information sharing among different platforms is achieved through the blockchain.

Blockchains are novel application modes of computer technologies such as distributed data storage, point-to-point transmission, consensus mechanisms, encryption algorithms, and the like. The blockchain (Blockchain), essentially a de-centralized database, is a string of data blocks that are generated in association using cryptographic methods, each of which contains information from a batch of network transactions for verifying the validity (anti-counterfeit) of its information and generating the next block. The blockchain may include a blockchain underlying platform, a platform product services layer, and an application services layer.

In this embodiment, if the vehicle configuration level does not exceed the preset configuration level threshold, it indicates that the user does not reconfigure the vehicle, and at this time, after the vehicle model and the vehicle configuration level corresponding to the vehicle picture to be identified are obtained, the unique vehicle type information corresponding to the vehicle picture to be identified may be automatically obtained, and the actual value of the vehicle may be estimated according to the vehicle model and the vehicle configuration level. Moreover, the information is acquired through image recognition, and manual input of a user is not needed.

The method realizes image recognition based on the vehicle photo data uploaded by the user side according to the photo uploading guide information, and the vehicle model is determined according to the obtained vehicle model information and the vehicle configuration grade, so that the accuracy of vehicle model recognition is improved.

The embodiment of the invention also provides a vehicle type recognition device based on image recognition, which is used for executing any embodiment of the vehicle type recognition method based on image recognition. Specifically, referring to fig. 3, fig. 3 is a schematic block diagram of a vehicle type recognition device based on image recognition according to an embodiment of the present invention. The image recognition-based vehicle type recognition apparatus 100 may be configured in a server.

As shown in fig. 3, the image recognition-based vehicle type recognition apparatus 100 includes a first positioning information acquisition unit 101, a first distance determination unit 102, a vehicle photograph data reception unit 103, a certification picture determination unit 104, a vehicle feature sequence acquisition unit 105, a vehicle model information acquisition unit 106, a vehicle interior part sequence acquisition unit 107, a vehicle configuration level acquisition unit 108, a configuration level determination unit 109, and a first vehicle data storage unit 110.

The first positioning information obtaining unit 101 is configured to receive the user side positioning information uploaded by the user side, and receive the vehicle positioning information uploaded by the vehicle.

In this embodiment, for example, when a user purchases a car and needs to perform online insurance, the server needs to determine the car type and the car configuration level first, and at this time, the user needs to start the APP application program of the user side and log in, and at the same time, also start the APP application program of the car machine and log in. At this time, the user terminal can automatically upload the positioning information of the user terminal, and the vehicle machine can also automatically upload the positioning information of the vehicle machine. The server receives the user side positioning information uploaded by the user side and receives the vehicle machine positioning information uploaded by the vehicle machine.

The first distance determining unit 102 is configured to determine whether a distance between the user side positioning information and the vehicle positioning information exceeds a preset distance threshold.

A vehicle photo data receiving unit 103, configured to receive vehicle photo data uploaded by the user side according to the photo uploading guiding information if the distance between the user side positioning information and the vehicle positioning information does not exceed the distance threshold; the vehicle photo data at least comprises a vehicle front view picture, a vehicle rear view picture and a vehicle interior part picture; the vehicle interior part pictures at least comprise car machine pictures, steering wheel pictures, skylight pictures, transmission pictures, seat pictures and air conditioner air outlet pictures.

In an embodiment, the image recognition-based vehicle type recognition apparatus 100 further includes:

The second positioning information acquisition unit is used for acquiring corresponding picture positioning information according to the photo parameter data included in the vehicle photo data;

The second distance judging unit is used for judging whether the distance between the picture positioning information and the vehicle-machine positioning information exceeds the distance threshold value; if the distance between the picture positioning information and the vehicle-mounted positioning information does not exceed the distance threshold, executing the step of judging whether the vehicle photo data comprise a new vehicle qualification picture or not; if the distance between the picture positioning information and the vehicle-mounted positioning information exceeds the distance threshold, executing the step of sending prompt information for prompting the user to approach the vehicle to the user side;

and the prompt information sending unit is used for sending prompt information for prompting the user to approach the vehicle to the user side.

And the certification picture judging unit 104 is used for judging whether the vehicle photo data comprises a new vehicle certification picture.

And the vehicle feature sequence obtaining unit 105 is configured to, if the vehicle photo data does not include a new vehicle qualification picture, perform image recognition according to the vehicle forward-looking picture and the vehicle backward-looking picture, and obtain a corresponding vehicle feature sequence.

The candidate region positioning unit is used for sequentially carrying out graying, edge detection, binarization and filtering treatment on the new vehicle qualification picture if the vehicle photograph data comprise the new vehicle qualification picture, so as to obtain a candidate region corresponding to the new vehicle qualification picture;

And the candidate area character recognition unit is used for correspondingly acquiring the vehicle model information corresponding to the vehicle photo data according to the characters in the candidate area.

A vehicle model information obtaining unit 106, configured to obtain vehicle model information corresponding to the vehicle photo data by using a vehicle feature sequence corresponding to the vehicle photo data as an input of a first convolutional neural network model trained in advance; wherein the first convolutional neural network model is used to identify a vehicle model.

The picture set acquisition unit is used for acquiring vehicle appearance pictures in the webpage according to the acquisition corresponding to the preset first website list so as to form a vehicle appearance picture set;

The image marking unit is used for acquiring the front cover marks, the rear cover marks and the marking values respectively corresponding to the displacement of each vehicle appearance image in the vehicle appearance image set according to a preset first marking strategy so as to form each vehicle appearance image and a corresponding vehicle characteristic sequence;

The vehicle appearance picture set comprises a first convolutional neural network training unit, a first convolutional neural network training unit and a second convolutional neural network training unit, wherein the first convolutional neural network training unit is used for taking a vehicle characteristic sequence corresponding to a vehicle appearance picture in the vehicle appearance picture set as input of a first convolutional neural network to be trained, taking vehicle model information corresponding to the vehicle appearance picture as output of the first convolutional neural network to be trained, and training the first convolutional neural network to be trained to obtain a first convolutional neural network model for identifying the vehicle model.

The vehicle interior part sequence obtaining unit 107 is configured to perform image recognition according to the vehicle interior part picture, so as to obtain a corresponding vehicle interior part sequence.

In one embodiment, the vehicle interior component sequence acquisition unit 107 includes:

The interior trim picture identification unit is used for identifying and acquiring a vehicle machine, a steering wheel, a skylight, a transmission, a seat and an air conditioner air outlet in the interior trim part picture of the vehicle;

the vehicle interior trim part marking unit is used for acquiring marking values respectively corresponding to a vehicle machine, a steering wheel, a skylight, a transmission, a seat and an air conditioner air outlet in the vehicle interior trim part picture according to a preset second marking strategy so as to form a vehicle interior trim part sequence.

A vehicle arrangement level obtaining unit 108 configured to obtain a vehicle arrangement level corresponding to the vehicle photograph data by using a vehicle interior part sequence corresponding to the vehicle photograph data as an input of a second convolutional neural network model trained in advance; wherein the second convolutional neural network model is used to identify a vehicle configuration level.

A configuration level judging unit 109 for judging whether the vehicle configuration level exceeds a preset configuration level threshold.

The first vehicle data storage unit 110 is configured to increase the suspicious vehicle identifier to the vehicle data corresponding to the vehicle photo data to obtain the current first vehicle data if the vehicle configuration level exceeds the configuration level threshold, and store the current first vehicle data to a preset first storage area.

And the second vehicle data storage unit is used for acquiring the unique vehicle type data corresponding to the vehicle picture to be identified according to the vehicle model information and the vehicle configuration level if the vehicle configuration level does not exceed the configuration level threshold, adding the unique vehicle type data to the vehicle data corresponding to the vehicle photo data to obtain current second vehicle data, and storing the current second vehicle data in a preset second storage area.

The device realizes image recognition based on the vehicle photo data uploaded by the user side according to the photo uploading guide information, and the vehicle model is determined according to the obtained vehicle model information and the vehicle configuration grade, so that the accuracy of vehicle model recognition is improved.

The above-described image recognition-based vehicle type recognition apparatus may be implemented in the form of a computer program that can be run on a computer device as shown in fig. 4.

Referring to fig. 4, fig. 4 is a schematic block diagram of a computer device according to an embodiment of the present invention. The computer device 500 is a server, and the server may be a stand-alone server or a server cluster formed by a plurality of servers.

With reference to FIG. 4, the computer device 500 includes a processor 502, memory, and a network interface 505, connected by a system bus 501, where the memory may include a non-volatile storage medium 503 and an internal memory 504.

The non-volatile storage medium 503 may store an operating system 5031 and a computer program 5032. The computer program 5032, when executed, may cause the processor 502 to perform an image recognition-based vehicle model recognition method.

The processor 502 is used to provide computing and control capabilities to support the operation of the overall computer device 500.

The internal memory 504 provides an environment for the execution of a computer program 5032 in the non-volatile storage medium 503, which computer program 5032, when executed by the processor 502, causes the processor 502 to perform an image recognition-based vehicle model recognition method.

The network interface 505 is used for network communication, such as providing for transmission of data information, etc. It will be appreciated by those skilled in the art that the architecture shown in fig. 4 is merely a block diagram of some of the architecture relevant to the present inventive arrangements and is not limiting of the computer device 500 to which the present inventive arrangements may be implemented, and that a particular computer device 500 may include more or fewer components than shown, or may combine some of the components, or have a different arrangement of components.

The processor 502 is configured to execute a computer program 5032 stored in a memory, so as to implement the vehicle type recognition method based on image recognition disclosed in the embodiment of the present invention.

Those skilled in the art will appreciate that the embodiment of the computer device shown in fig. 4 is not limiting of the specific construction of the computer device, and in other embodiments, the computer device may include more or less components than those shown, or certain components may be combined, or a different arrangement of components. For example, in some embodiments, the computer device may include only a memory and a processor, and in such embodiments, the structure and function of the memory and the processor are consistent with the embodiment shown in fig. 4, and will not be described again.

It should be appreciated that in embodiments of the present invention, the Processor 502 may be a central processing unit (Central Processing Unit, CPU), the Processor 502 may also be other general purpose processors, digital signal processors (DIGITAL SIGNAL processors, DSPs), application SPECIFIC INTEGRATED Circuits (ASICs), off-the-shelf Programmable gate arrays (Field-Programmable GATEARRAY, FPGA) or other Programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or the like. Wherein the general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.

In another embodiment of the invention, a computer-readable storage medium is provided. The computer readable storage medium may be a non-volatile computer readable storage medium. The computer readable storage medium stores a computer program, wherein the computer program when executed by a processor implements the vehicle model recognition method based on image recognition disclosed in the embodiment of the present invention.

It will be clearly understood by those skilled in the art that, for convenience and brevity of description, specific working procedures of the apparatus, device and unit described above may refer to corresponding procedures in the foregoing method embodiments, which are not repeated herein. Those of ordinary skill in the art will appreciate that the elements and algorithm steps described in connection with the embodiments disclosed herein may be embodied in electronic hardware, in computer software, or in a combination of the two, and that the elements and steps of the examples have been generally described in terms of function in the foregoing description to clearly illustrate the interchangeability of hardware and software. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.

In the several embodiments provided by the present invention, it should be understood that the disclosed apparatus, device and method may be implemented in other manners. For example, the apparatus embodiments described above are merely illustrative, and for example, the division of the units is merely a logical function division, there may be another division manner in actual implementation, or units having the same function may be integrated into one unit, for example, multiple units or components may be combined or may be integrated into another system, or some features may be omitted, or not performed. In addition, the coupling or direct coupling or communication connection shown or discussed with each other may be an indirect coupling or communication connection via some interfaces, devices, or elements, or may be an electrical, mechanical, or other form of connection.

The units described as separate units may or may not be physically separate, and units shown as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the embodiment of the present invention.

In addition, each functional unit in the embodiments of the present invention may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit. The integrated units may be implemented in hardware or in software functional units.

The integrated units may be stored in a storage medium if implemented in the form of software functional units and sold or used as stand-alone products. Based on such understanding, the technical solution of the present invention is essentially or a part contributing to the prior art, or all or part of the technical solution may be embodied in the form of a software product stored in a storage medium, comprising several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a magnetic disk, an optical disk, or other various media capable of storing program codes.

While the invention has been described with reference to certain preferred embodiments, it will be understood by those skilled in the art that various changes and substitutions of equivalents may be made and equivalents will be apparent to those skilled in the art without departing from the scope of the invention. Therefore, the protection scope of the invention is subject to the protection scope of the claims.

Claims

1. The vehicle type recognition method based on image recognition is characterized by comprising the following steps of:

If the vehicle configuration level exceeds the configuration level threshold, adding suspicious vehicle identification to the vehicle data corresponding to the vehicle photo data to obtain current first vehicle data, and storing the current first vehicle data into a preset first storage area;

after the judging whether the vehicle configuration level exceeds the preset configuration level threshold, the method further comprises the following steps:

If the vehicle configuration level does not exceed the configuration level threshold, acquiring unique vehicle type data corresponding to the vehicle picture to be identified according to the vehicle model information and the vehicle configuration level, adding the unique vehicle type data to the vehicle data corresponding to the vehicle photo data to obtain current second vehicle data, and storing the current second vehicle data in a preset second storage area;

If the distance between the user side positioning information and the vehicle machine positioning information does not exceed the distance threshold, after receiving the vehicle photo data uploaded by the user side according to the photo uploading guiding information, the method further comprises the following steps:

2. The image recognition-based vehicle model recognition method according to claim 1, characterized by further comprising:

3. The vehicle model recognition method based on image recognition according to claim 1, wherein the performing image recognition according to the vehicle interior part picture to obtain a corresponding vehicle interior part sequence includes:

4. The image recognition-based vehicle type recognition method according to claim 1, wherein after the determining whether the vehicle photograph data includes a new vehicle certification picture, further comprises:

5. The image recognition-based vehicle type recognition method according to claim 1, wherein the preset first storage area and the preset second storage area are blocks on a blockchain network, respectively.

6. A vehicle model recognition device based on image recognition, characterized by comprising:

The first vehicle data storage unit is used for adding suspicious vehicle identifications to vehicle data corresponding to the vehicle photo data to obtain current first vehicle data if the vehicle configuration level exceeds the configuration level threshold value, and storing the current first vehicle data into a preset first storage area;

The second vehicle data storage unit is used for acquiring unique vehicle type data corresponding to the vehicle picture to be identified according to the vehicle model information and the vehicle configuration level if the vehicle configuration level does not exceed the configuration level threshold, adding the unique vehicle type data to the vehicle data corresponding to the vehicle photo data to obtain current second vehicle data, and storing the current second vehicle data in a preset second storage area;

7. A computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor implements the image recognition-based vehicle model recognition method according to any one of claims 1 to 5 when executing the computer program.

8. A computer-readable storage medium, characterized in that the computer-readable storage medium stores a computer program which, when executed by a processor, causes the processor to execute the image recognition-based vehicle model recognition method according to any one of claims 1 to 5.