CN110598749A - Image vehicle recognition implementation system based on convolutional neural network AI algorithm - Google Patents
Image vehicle recognition implementation system based on convolutional neural network AI algorithm Download PDFInfo
- Publication number
- CN110598749A CN110598749A CN201910746195.7A CN201910746195A CN110598749A CN 110598749 A CN110598749 A CN 110598749A CN 201910746195 A CN201910746195 A CN 201910746195A CN 110598749 A CN110598749 A CN 110598749A
- Authority
- CN
- China
- Prior art keywords
- model
- identification
- vehicle
- network
- recognition
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- 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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/08—Detecting or categorising vehicles
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Mathematical Physics (AREA)
- Computational Linguistics (AREA)
- Health & Medical Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses an image vehicle recognition implementation system based on a convolutional neural network AI algorithm, which comprises the following steps: the input module is used for acquiring an image to be identified; the image identification module comprises a data unit, an AI model network unit and a result unit; the data unit is provided with a mass data training set and is used for acquiring and storing all automobile information on the market and establishing a recognition model; the AI model network unit comprises a vehicle angle identification subunit, a brand identification subunit and a model identification subunit; the vehicle angle identification subunit is used for identifying the vehicle angle in the image to be identified through the identification model; the brand identification subunit is used for further identifying the automobile brand in the image to be identified through the identification model; the model identification subunit is used for identifying models in the same brand through an identification model so as to finish identification; the result unit is used for acquiring the recognition result after the AI model network unit completes recognition; and the output module is used for outputting the identification result.
Description
Technical Field
The invention relates to the technical field of image recognition, in particular to an image vehicle recognition implementation system based on a convolutional neural network AI algorithm.
Background
With the increasing of market competition, commercial vehicle manufacturers enter the military vehicle market, autonomous vehicle manufacturers continuously release high-end vehicle types to compete with joint ventures, joint ventures non-luxury manufacturers design low-cost vehicle types to enlarge audience of consumers, luxury manufacturers reduce the access threshold, and small luxury vehicles are continuously released to extrude the joint ventures non-luxury high-end vehicle types. The number of models existing on the market is more and more, and according to statistics, more than 200 automobile brands, more than 2000 automobile models and more than 40000 automobile model numbers exist on the market. Therefore, in the past, the method for identifying the vehicle type through human experience is not feasible, the efficiency is low, and generally experienced professionals are difficult to ensure the accuracy of the model granularity. With the accumulation of a large number of pictures, a picture recognition model is trained by applying a deep learning method, so that the model and parameter values of the automobile can be distinguished, the future direction is gradually formed, and the problem which is greatly needed to be solved by the automobile industry is solved.
Disclosure of Invention
The invention provides an image vehicle identification implementation system based on a convolutional neural network AI algorithm, which aims to solve the technical problem that the specific model and the parameter value of a vehicle cannot be quickly identified in an image in the prior art, so that an AI identification model is established based on the convolutional neural network AI algorithm, the image is quickly identified, and the specific model and the parameter value of the vehicle are quickly identified in the image.
In order to solve the above technical problem, an embodiment of the present invention provides an image vehicle identification implementation system based on a convolutional neural network AI algorithm, including:
the input module is used for acquiring an image to be identified;
the image identification module comprises a data unit, an AI model network unit and a result unit; the data unit is provided with a mass data training set and is used for acquiring and storing all automobile information on the market and establishing an identification model to cover all the automobiles of all brands in the whole market so as to improve the generalization capability of the model; according to the model and the actual requirement, angle marking is carried out on the obtained picture, and the obtained picture is divided into the front direction, the side direction, the rear direction, the car lamp, the hub, the interior trim and the like, so that the model training pertinence is stronger, and the identification is more accurate; the AI model network unit comprises a vehicle angle identification subunit, a brand identification subunit and a model identification subunit; the vehicle angle identification subunit is used for identifying the vehicle angle in the image to be identified through the identification model; the brand identification subunit is used for further identifying the automobile brand in the image to be identified through the identification model; the model identification subunit is used for identifying models in the same brand through the identification model so as to finish identification; the result unit is used for acquiring the recognition result after the AI model network unit completes recognition;
the output module is used for outputting the identification result;
the image recognition module further comprises a network structure selection and modification unit, wherein the network structure selection and modification unit is used for selecting acceptance-res-v 2 as a basic network, importing weights pre-trained on a large data set ImageNet by the network, modifying the network into a proper network according to the requirements of the automobile picture data set, establishing a proper loss function and optimizing the network.
Preferably, the car information stored in the data unit includes: the method comprises the steps of obtaining a large number of real pictures of a second-hand vehicle source, a large number of automobile brands and automobile type information of the second-hand vehicle source, covering various colors of automobile types with the pictures, and simultaneously comprising various shooting angles and shooting environments.
Preferably, the identifying the automobile angle in the image to be identified through the identification model includes: and screening the pictures of the same vehicle source at different angles, and only selecting the front picture with the largest information amount as a classification identifier to enter the model, so that the model achieves the purpose of identifying the whole vehicle through one picture.
As a preferred scheme, the mass data training set comprises a standard new car picture library, and the standard new car picture library is used for storing standard pictures of a new car, comparing and examining details of classification information of original pictures, and ensuring correctness of the classification information of the pictures.
Preferably, the identifying the model in the same brand by the model identifying subunit includes: combining the models with the same appearance through a standard new vehicle picture library to obtain a standard model group; the model group comprises vehicles with different models and different year types but completely same appearance characteristics.
Preferably, the data unit further comprises a standard industry database, and the industry database comprises official guide prices, annual money information and time to market of mass automobiles.
As a preferred scheme, the data unit further comprises a segmentation data subunit, and the segmentation data subunit is used for segmenting the massive data training set into a training set, a verification set and a test set according to a preset proportion; the training set is used for training the model, the verification set is used for selecting a proper network structure and adjusting the hyper-parameters, and the test set is used for testing the precision and generalization capability of the model.
Preferably, the AI model network unit further comprises a data enhancement unit, wherein the data enhancement unit is configured to perform data enhancement processing on the picture entering the recognition model, so that the recognition model can adapt to different compression ratios, picture quality, shooting environments and shooting angles, the generalization of the model is enhanced, and the risk of overfitting is reduced.
Preferably, the data enhancement processing on the picture includes: and (4) carrying out noise addition, cutting, rotation, turnover and stretching on the picture.
As a preferred scheme, the network structure selecting and modifying unit is further configured to: by means of the fine tuning learning and back propagation method, the coefficients are fine tuned layer by layer, and the effect of optimizing the whole network parameters is achieved.
Compared with the prior art, the embodiment of the invention has the following beneficial effects:
the method and the device establish the AI identification model based on the convolutional neural network AI algorithm, quickly identify the image, and solve the technical problem that the specific model and the parameter value of the automobile cannot be quickly identified in the image in the prior art, so that the specific model and the parameter value of the automobile in the image can be quickly identified.
Drawings
FIG. 1: the data interaction flow diagram of the image vehicle recognition system in the embodiment of the invention is shown.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Referring to fig. 1, a preferred embodiment of the present invention provides an image recognition vehicle implementation system based on a convolutional neural network AI algorithm, including:
the input module is used for acquiring an image to be identified;
the image identification module comprises a data unit, an AI model network unit and a result unit; the data unit is provided with a mass data training set and is used for acquiring and storing all automobile information on the market and establishing a recognition model;
in this embodiment, the car information stored in the data unit includes: the method comprises the steps of obtaining a large number of real pictures of a second-hand vehicle source, a large number of automobile brands and automobile type information of the second-hand vehicle source, covering various colors of automobile types with the pictures, and simultaneously comprising various shooting angles and shooting environments.
In this embodiment, the data unit further includes a standard industry database, and the industry database includes official guide prices, annual money information, and time to market for a large number of automobiles.
In this embodiment, the data unit further includes a segmentation data subunit, where the segmentation data subunit is configured to segment the massive data training set into a training set, a verification set, and a test set according to a preset ratio; the training set is used for training the model, the verification set is used for selecting a proper network structure and adjusting the hyper-parameters, and the test set is used for testing the precision and generalization capability of the model.
The AI model network unit comprises a vehicle angle identification subunit, a brand identification subunit and a model identification subunit; the vehicle angle identification subunit is used for identifying the vehicle angle in the image to be identified through the identification model;
in this embodiment, the identifying, by the identification model, the automobile angle in the image to be identified includes: and screening the pictures of the same vehicle source at different angles, and only selecting the front picture with the largest information amount as a classification identifier to enter the model, so that the model achieves the purpose of identifying the whole vehicle through one picture.
The brand identification subunit is used for further identifying the automobile brand in the image to be identified through the identification model;
the model identification subunit is used for identifying models in the same brand through the identification model so as to finish identification; the result unit is used for acquiring the recognition result after the AI model network unit completes recognition;
in this embodiment, the mass data training set includes a standard new car picture library, and the standard new car picture library is used for storing a standard picture of a new car, and performing detail comparison and examination on classification information of an original picture to ensure correctness of the classification information of the picture.
In this embodiment, the identifying the model in the same brand by the model identifying subunit includes: combining the models with the same appearance through a standard new vehicle picture library to obtain a standard model group; the model group comprises vehicles with different models and different year types but completely same appearance characteristics.
In this embodiment, the AI model network unit further includes a data enhancement unit, where the data enhancement unit is configured to perform data enhancement processing on the picture entering the recognition model, so that the recognition model can adapt to different compression ratios, picture quality, shooting environments and shooting angles, enhance generalization of the model, and reduce risk of overfitting.
In this embodiment, the performing data enhancement processing on the picture includes: and (4) carrying out noise addition, cutting, rotation, turnover and stretching on the picture.
The output module is used for outputting the identification result;
the image recognition module further comprises a network structure selection and modification unit, wherein the network structure selection and modification unit is used for selecting acceptance-res-v 2 as a basic network, importing weights pre-trained on a large data set ImageNet by the network, modifying the network into a proper network according to the requirements of the automobile picture data set, establishing a proper loss function and optimizing the network.
In this embodiment, the network structure selecting and modifying unit is further configured to: by means of the fine tuning learning and back propagation method, the coefficients are fine tuned layer by layer, and the effect of optimizing the whole network parameters is achieved.
The present invention will be described in detail with reference to specific examples.
The image recognition based on artificial intelligence is based on image training of an ultra-large number scale, realizes vehicle type recognition of high-precision fine and minimum granularity, and recognizes model groups with different colors and consistent shapes at high precision. By inputting the vehicle pictures of different angles such as the front side and the side surface, the brand name, the model name, the annual payment range, the model set, the body color, the identification accuracy and other parameter values of the vehicle are output within 0.5 s. Through three times of algorithm iteration and long-time training, the vehicle type recognition accuracy of the current image vehicle recognition model is as high as 98.2%, the annual pattern range recognition accuracy is as high as 93.1%, and the vehicle body color recognition accuracy is as high as 91.6%. The model algorithm can adapt to different picture background environments and has the capabilities of resisting light reflection and resisting obstruction. The model is suitable for recognizing different car body colors and different background environments; the method can be suitable for vehicle type image recognition at different angles; the method can be used for identifying different elevation angles and depression angles; the method has good identification effect in the reflective and dark environment of the underground garage environment. The method specifically comprises the following steps:
a super-large scale data training set, wherein the real pictures of 3000 ten thousand second-hand vehicles from different channels are obtained; covering 2005 to date, including more than 200 automobile brands and more than 2000 vehicle models; the picture covers various colors of the vehicle type and simultaneously comprises various shooting angles and shooting environments; and acquiring a large number of used vehicle pictures shot by the user and vehicle information filled by the user from the used vehicle picture database.
The standard industry database covers the standard information database of most models in the market from 1980 to now; the detailed information comprises official guide price, annual payment information, time to market and time to market; after the image classification result is matched, more detailed information behind the image can be obtained.
And combining the models, namely combining the models with the same appearance by using a standard new car photo library and combining the industry experience to form a standard model group. The model group comprises vehicles of different models and different year types, but the appearance characteristics of the vehicles are completely the same.
And (4) checking the pictures, namely performing detail comparison checking on the classification information of the original pictures by using a standard new car picture library and combining with industry experience to ensure the correctness of the classification information of the pictures. Only correctly classified pictures are reserved, and a certain number of correct pictures are selected for training in each model group.
And (4) angle screening, namely screening different angle pictures of the same vehicle source by using a trained vehicle angle classification model, and only selecting the front picture (including the oblique front) with the largest information amount as a classification identifier to enter the model. The purpose of recognizing the whole vehicle by one picture is achieved by the model.
The data set is divided into three parts, namely a training set, a verification set and a test set according to experience proportion. And when the images are cut, the images are cut on the granularity of the vehicle source, so that the images of different angles of the same vehicle source only exist in one data set. The training set is used as a training model, the verification set is used for selecting a proper network structure and adjusting the hyper-parameters, and the test set is used as the precision and generalization capability of the test model.
Data enhancement is handled, has adopted "data enhancement" technique, carries out data enhancement to the picture that gets into the model, adds the work such as making an uproar, tailorring, rotation, upset, tensile to the picture for the model can adapt to different compression ratios, picture quality, shooting environment and shooting angle, and the generalization of reinforcing model reduces the risk of overfitting, can increase training sample volume simultaneously.
Selecting and transforming a network structure, and finally selecting an initiation-res-v 2 with the best effect as a basic network through a plurality of tests, importing weights pre-trained on a large data set ImageNet by the network, transforming the network into a proper network according to the requirement of an automobile picture data set, and establishing a proper loss function (cross entropy):
and by utilizing a fine tuning learning and back propagation method, the coefficients are fine tuned layer by layer, and the effect of optimizing the parameters of the whole network is achieved. And finally, selecting the parameter combination with the best effect as the parameter of the network.
The application scenario and the artificial intelligence algorithm for image recognition of the vehicle have wide application scenarios, such as: scanning and searching vehicles, assisting accurate estimation, vehicle statistical monitoring, fast estimation and the like. The 'scanning and searching for vehicles' is to help users accurately search the vehicle attribute and the renting and selling information of related new vehicles and used vehicles by utilizing the existing or randomly shot vehicle pictures, to link the vehicles of the entity with the market and help users to know more abundant market information. The auxiliary accurate estimation value is used for indicating that the second-hand car evaluator is helped to quickly and accurately identify the car type, so that the economic loss and the sales risk caused by wrong judgment are reduced. The vehicle statistical monitoring means that various attributes of vehicles are rapidly identified and counted through video monitoring in dealers, residential districts, public places, tourist attractions, highways and other cities, so that intelligent monitoring and management are realized. The rapid valuation method utilizes information such as vehicle types, colors and annual money output by the vehicle type recognition model, combines information such as a city on a license plate and the like, carries out rapid valuation on the vehicle types by means of the second-hand vehicle valuation model, and can be used for rapid asset valuation of enterprises and the like.
The method and the device establish the AI identification model based on the convolutional neural network AI algorithm, quickly identify the image, and solve the technical problem that the specific model and the parameter value of the automobile cannot be quickly identified in the image in the prior art, so that the specific model and the parameter value of the automobile in the image can be quickly identified.
The above-mentioned embodiments are provided to further explain the objects, technical solutions and advantages of the present invention in detail, and it should be understood that the above-mentioned embodiments are only examples of the present invention and are not intended to limit the scope of the present invention. It should be understood that any modifications, equivalents, improvements and the like, which come within the spirit and principle of the invention, may occur to those skilled in the art and are intended to be included within the scope of the invention.
Claims (10)
1. An image vehicle recognition implementation system based on a convolutional neural network AI algorithm is characterized by comprising:
the input module is used for acquiring an image to be identified;
the image identification module comprises a data unit, an AI model network unit and a result unit; the data unit is provided with a mass data training set and is used for acquiring and storing all automobile information on the market and establishing a recognition model; the AI model network unit comprises a vehicle angle identification subunit, a brand identification subunit and a model identification subunit; the vehicle angle identification subunit is used for identifying the vehicle angle in the image to be identified through the identification model; the brand identification subunit is used for further identifying the automobile brand in the image to be identified through the identification model; the model identification subunit is used for identifying models in the same brand through the identification model so as to finish identification; the result unit is used for acquiring the recognition result after the AI model network unit completes recognition;
the output module is used for outputting the identification result;
the image recognition module further comprises a network structure selection and modification unit, wherein the network structure selection and modification unit is used for selecting acceptance-res-v 2 as a basic network, importing weights pre-trained on a large data set ImageNet by the network, designing a proper network according to the requirements of an automobile picture data set, and establishing a proper loss function to optimize the network.
2. The convolutional neural network AI algorithm-based image recognition system of claim 1, wherein the car information stored in the data unit comprises: the method comprises the steps of obtaining a large number of real pictures of a second-hand vehicle source, a large number of automobile brands and automobile type information of the second-hand vehicle source, covering various colors of automobile types with the pictures, and simultaneously comprising various shooting angles and shooting environments.
3. The convolutional neural network AI algorithm-based image recognition vehicle implementation system of claim 1, wherein recognizing the vehicle angle in the image to be recognized through the recognition model comprises: and screening the pictures of the same vehicle source at different angles, and only selecting the front picture with the largest information amount as a classification identifier to enter the model, so that the model achieves the purpose of identifying the whole vehicle through one picture.
4. The convolutional neural network AI algorithm-based image recognition implementation system of claim 1, wherein the massive data training set comprises a standard new car picture library, and the standard new car picture library is used for storing standard pictures of new cars, performing detail comparison examination on classification information of original pictures, and ensuring correctness of the picture classification information.
5. The convolutional neural network AI algorithm-based image recognition vehicle implementation system of claim 4, wherein the model identification subunit identifies models within the same brand through the identification model, and comprises: combining the models with the same appearance through a standard new vehicle picture library to obtain a standard model group; the model group comprises vehicles with different models and different year types but completely same appearance characteristics.
6. The convolutional neural network AI algorithm-based image recognition vehicle implementation system of claim 1, wherein the data unit further comprises a standard industry database, the industry database comprising official guide prices, annuity information, and time to market for a large number of vehicles.
7. The convolutional neural network AI algorithm-based image recognition implementation system of claim 1, wherein the data unit further comprises a data segmentation subunit, the data segmentation subunit being configured to segment the massive data training set into a training set, a validation set, and a test set according to a preset ratio; the training set is used for training the model, the verification set is used for selecting a proper network structure and adjusting the hyper-parameters, and the test set is used for testing the precision and generalization capability of the model.
8. The convolutional neural network AI algorithm-based image recognition implementation system of claim 1, wherein the AI model network unit further comprises a data enhancement unit for performing data enhancement processing on the picture entered into the recognition model, so that the recognition model can adapt to different compression ratios, picture quality, shooting environment and shooting angle, the generalization of the model is enhanced, and the risk of overfitting is reduced.
9. The convolutional neural network AI algorithm-based image recognition implementation system of claim 8, wherein the data enhancement processing on the picture comprises: and (4) carrying out noise addition, cutting, rotation, turnover and stretching on the picture.
10. The convolutional neural network AI algorithm-based image recognition implementation system of claim 1, wherein the network structure selection and modification unit is further configured to: by means of the fine tuning learning and back propagation method, the coefficients are fine tuned layer by layer, and the effect of optimizing the whole network parameters is achieved.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910746195.7A CN110598749A (en) | 2019-08-13 | 2019-08-13 | Image vehicle recognition implementation system based on convolutional neural network AI algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910746195.7A CN110598749A (en) | 2019-08-13 | 2019-08-13 | Image vehicle recognition implementation system based on convolutional neural network AI algorithm |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110598749A true CN110598749A (en) | 2019-12-20 |
Family
ID=68854064
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910746195.7A Pending CN110598749A (en) | 2019-08-13 | 2019-08-13 | Image vehicle recognition implementation system based on convolutional neural network AI algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110598749A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111709884A (en) * | 2020-04-29 | 2020-09-25 | 高新兴科技集团股份有限公司 | License plate key point correction method, system, equipment and storage medium |
| CN115797923A (en) * | 2023-02-13 | 2023-03-14 | 广东天圣网络科技有限公司 | Method and system for recognizing vehicle type through Al intelligent algorithm based on vehicle model |
Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101196979A (en) * | 2006-12-22 | 2008-06-11 | 四川川大智胜软件股份有限公司 | Method for recognizing vehicle type by digital picture processing technology |
| CN105046255A (en) * | 2015-07-16 | 2015-11-11 | 北京交通大学 | Vehicle tail character recognition based vehicle type identification method and system |
| CN105335710A (en) * | 2015-10-22 | 2016-02-17 | 合肥工业大学 | Fine vehicle model identification method based on multi-stage classifier |
| CN105787437A (en) * | 2016-02-03 | 2016-07-20 | 东南大学 | Vehicle brand type identification method based on cascading integrated classifier |
| CN105938560A (en) * | 2016-03-23 | 2016-09-14 | 吉林大学 | Convolutional-neural-network-based vehicle model refined classification system |
| CN105975941A (en) * | 2016-05-31 | 2016-09-28 | 电子科技大学 | Multidirectional vehicle model detection recognition system based on deep learning |
| CN106022285A (en) * | 2016-05-30 | 2016-10-12 | 北京智芯原动科技有限公司 | Vehicle type identification method and vehicle type identification device based on convolutional neural network |
| CN106295541A (en) * | 2016-08-03 | 2017-01-04 | 乐视控股（北京）有限公司 | Vehicle type recognition method and system |
| CN106354735A (en) * | 2015-07-22 | 2017-01-25 | 杭州海康威视数字技术股份有限公司 | Image target searching method and device |
| CN106469299A (en) * | 2016-08-31 | 2017-03-01 | 北京邮电大学 | A kind of vehicle search method and device |
| CN108549926A (en) * | 2018-03-09 | 2018-09-18 | 中山大学 | A kind of deep neural network and training method for refining identification vehicle attribute |
| CN109948610A (en) * | 2019-03-14 | 2019-06-28 | 西南交通大学 | A kind of vehicle fine grit classification method in the video based on deep learning |
-
2019
- 2019-08-13 CN CN201910746195.7A patent/CN110598749A/en active Pending
Patent Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101196979A (en) * | 2006-12-22 | 2008-06-11 | 四川川大智胜软件股份有限公司 | Method for recognizing vehicle type by digital picture processing technology |
| CN105046255A (en) * | 2015-07-16 | 2015-11-11 | 北京交通大学 | Vehicle tail character recognition based vehicle type identification method and system |
| CN106354735A (en) * | 2015-07-22 | 2017-01-25 | 杭州海康威视数字技术股份有限公司 | Image target searching method and device |
| CN105335710A (en) * | 2015-10-22 | 2016-02-17 | 合肥工业大学 | Fine vehicle model identification method based on multi-stage classifier |
| CN105787437A (en) * | 2016-02-03 | 2016-07-20 | 东南大学 | Vehicle brand type identification method based on cascading integrated classifier |
| CN105938560A (en) * | 2016-03-23 | 2016-09-14 | 吉林大学 | Convolutional-neural-network-based vehicle model refined classification system |
| CN106022285A (en) * | 2016-05-30 | 2016-10-12 | 北京智芯原动科技有限公司 | Vehicle type identification method and vehicle type identification device based on convolutional neural network |
| CN105975941A (en) * | 2016-05-31 | 2016-09-28 | 电子科技大学 | Multidirectional vehicle model detection recognition system based on deep learning |
| CN106295541A (en) * | 2016-08-03 | 2017-01-04 | 乐视控股（北京）有限公司 | Vehicle type recognition method and system |
| CN106469299A (en) * | 2016-08-31 | 2017-03-01 | 北京邮电大学 | A kind of vehicle search method and device |
| CN108549926A (en) * | 2018-03-09 | 2018-09-18 | 中山大学 | A kind of deep neural network and training method for refining identification vehicle attribute |
| CN109948610A (en) * | 2019-03-14 | 2019-06-28 | 西南交通大学 | A kind of vehicle fine grit classification method in the video based on deep learning |
Non-Patent Citations (4)
| Title |
|---|
| KUN HUANG 等: "Fine-grained Vehicle Recognition by Deep Convolutional Neural Network", 《CISP-BMEI 2016》 * |
| 杨露菁 等: "《智能图像处理及应用》", 31 March 2019 * |
| 贾振堂: "基于约束卷积神经网络的轿车款式识别", 《上海电力学院学报》 * |
| 高志强 等: "《深度学习 从入门到实战》", 30 June 2018 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111709884A (en) * | 2020-04-29 | 2020-09-25 | 高新兴科技集团股份有限公司 | License plate key point correction method, system, equipment and storage medium |
| CN115797923A (en) * | 2023-02-13 | 2023-03-14 | 广东天圣网络科技有限公司 | Method and system for recognizing vehicle type through Al intelligent algorithm based on vehicle model |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110378236B (en) | Vehicle identity recognition model construction and recognition method and system based on deep learning | |
| US20180364727A1 (en) | Methods, Protocol and System for Customizing Self-driving Motor Vehicles | |
| CN109558823B (en) | Vehicle identification method and system for searching images by images | |
| CN110147726A (en) | Business quality detecting method and device, storage medium and electronic device | |
| CN111898523A (en) | Remote sensing image special vehicle target detection method based on transfer learning | |
| CN109740479A (en) | A kind of vehicle recognition methods, device, equipment and readable storage medium storing program for executing again | |
| CN111444952A (en) | Method and device for generating sample identification model, computer equipment and storage medium | |
| CN110532990A (en) | The recognition methods of turn signal use state, device, computer equipment and storage medium | |
| CN111126396A (en) | Image recognition method and device, computer equipment and storage medium | |
| CN111311540A (en) | Vehicle damage assessment method and device, computer equipment and storage medium | |
| CN110942015A (en) | Crowd density estimation method | |
| US11966829B2 (en) | Convolutional artificial neural network based recognition system in which registration, search, and reproduction of image and video are divided between and performed by mobile device and server | |
| CN110598749A (en) | Image vehicle recognition implementation system based on convolutional neural network AI algorithm | |
| CN114220458B (en) | Voice recognition method and device based on array hydrophone | |
| CN113627229B (en) | Target detection method, system, device and computer storage medium | |
| CN108323209A (en) | Information processing method, system, cloud processing device and computer program product | |
| CN111523352A (en) | Method for intelligently and rapidly identifying illegal modified vehicle and monitoring system thereof | |
| CN110188828A (en) | A kind of image sources discrimination method based on virtual sample integrated study | |
| CN111010668A (en) | Information sharing method based on vehicle-mounted terminal position, terminal device and server | |
| CN113673533A (en) | Model training method and related equipment | |
| CN115408710A (en) | Image desensitization method and related device | |
| US20230142507A1 (en) | Systems and methods for estimating monetary loss to an accident damaged vehicle | |
| CN114529552A (en) | Remote sensing image building segmentation method based on geometric contour vertex prediction | |
| CN116994164A (en) | Multi-mode aerial image fusion and target detection combined learning method | |
| CN116611891A (en) | Content information recommendation method, device, server and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |