CN105335710A - Fine vehicle model identification method based on multi-stage classifier - Google Patents
Fine vehicle model identification method based on multi-stage classifier Download PDFInfo
- Publication number
- CN105335710A CN105335710A CN201510697287.2A CN201510697287A CN105335710A CN 105335710 A CN105335710 A CN 105335710A CN 201510697287 A CN201510697287 A CN 201510697287A CN 105335710 A CN105335710 A CN 105335710A
- Authority
- CN
- China
- Prior art keywords
- vehicle
- layer
- vehicle model
- level sorter
- class
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/56—Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
- G06V20/58—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads
- G06V20/584—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads of vehicle lights or traffic lights
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/285—Selection of pattern recognition techniques, e.g. of classifiers in a multi-classifier system
-
- 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)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a fine vehicle model identification method based on multi-stage classifiers. The fine vehicle model identification method comprises the steps of: collecting positive samples containing vehicle faces and negative samples not containing a vehicle at first, and training an A-stage classifier through an LBP+Adaboost method which is used for positioning the vehicle faces and intercepting images of the vehicle faces as samples to be input in the next step; classifying the images of the vehicle faces according to similarity degrees thereof, classifying the exactly similar samples into one category, designing a B-stage convolutional neural network classifier for training the samples, and outputting coarse categories, wherein the samples are classified into K categories in total; and collecting samples of fine vehicle models in the coarse categories for each coarse category classified in the last step, and designing a C-stage convolutional neural network for further training the samples, thereby identifying fine vehicle models of the samples. According to the fine vehicle model identification method based on the multi-stage classifiers, the specific identification scheme is provided for fine vehicle model identification in gateway images, the accuracy of identification results is very high, and the actual requirement of an intelligent traffic system can be met.
Description
Technical field
The present invention relates to computer vision target identification technology field, specifically a kind of meticulous vehicle model recognition methods based on multistage classifier.
Background technology
Vehicle cab recognition is an important branch in the application of intelligent transportation system computer vision, be widely used in traffic flow analysis, the management of the vehicle that hits out against theft, case-involving car tracing, specification traffic order, large parking lot, the field such as highway automatic charging.And meticulous vehicle model be identified in the process of public safety traffic management department by criminal case involving public security, vehicle false-trademark steals board, car plate is stained, vehicle appearance is stained etc. cause relate to car case time, play very important booster action.
In the existing patent about " vehicle cab recognition " or " vehicle model identification ", be the identification to type of vehicle, or to one of vehicle model rough identification, fail to identify meticulous vehicle model.As: (1) patent 201510071919.4 adopts the method for convolutional neural networks to carry out the identification of vehicle, and the type finally exported is truck, minibus, car and bus; (2) patent 201210494065.7 is based on the waveform character value of ground induction coil signal intensity sequence and corresponding sorter, carry out the identification of vehicle model, the result finally exported is: station wagon 1, station wagon 2, station wagon 3 and jubilee wagon 1, jubilee wagon 2, jubilee wagon 3 etc., concrete recognition result depends on the type of preset template, can not the identical and identical vehicle of ground induction coil signal intensity sequence produced of specification configuration, as: popular FV7187ZADBG and Kia YQZ7142AE5; (2) patent 201310239172.X carries out the identification of vehicle model based on high-resolution remote sensing image, and the vehicle that can identify is limited to the quantity of vehicle texture formwork equally, and can not distinguish from the similar vehicle of the texture of top viewing angles.
Although patent 201310299354.6,201410009098.7 and 201410109999.3 target is the comparison and the identification that realize vehicle model, but because front 2 kinds of methods are all based on operations such as grey level histogram, filtering, expansions, 1 kind of method next carries out geometry correction based on two dimensional discrete Fourier transform, therefore be subject to ambient lighting, impact that shade blocks, robustness is not strong.And patent 201410109999.3 needs to carry out on the basis of License Plate, be subject to the impact of License Plate degree of accuracy and the interference of stained car plate.
Because concrete vehicle model is very many, and the recognizer of meticulous vehicle model needs to block lower maintenance robustness at various illumination condition, shade, and therefore, the identification of meticulous vehicle model is still a very challenging problem.
summary of the inventionthe object of this invention is to provide a kind of meticulous vehicle model recognition methods based on multistage classifier, to solve the meticulous vehicle model identification problem in bayonet socket image.
In order to achieve the above object, the technical solution adopted in the present invention is:
Based on a meticulous vehicle model recognition methods for multistage classifier, it is characterized in that: design the multistage classifier based on Adaboost algorithm and convolutional neural networks, for carrying out meticulous identification to the vehicle model in Gate System image, comprising the following steps:
(1), reclassify device is designed, wherein: A level sorter is based on Adaboost algorithm, positive sample is the image comprising face before vehicle, negative sample is the image not comprising face before vehicle, LBP feature is adopted to train, for the vehicle in limited bayonet image, and vehicle sections is intercepted out, form new data set; Vehicle like front appearance, based on convolutional neural networks, for carrying out broad classification to vehicle model, is classified as a class by B level sorter; For the classification results of B level sorter, train multiple C level sorter, C level sorter based on convolutional neural networks, for again carrying out sophisticated category to every class classification results of B level sorter.
(2), for the image obtained from Gate System, adopt multistage classifier to carry out the identification of vehicle model, identification step is as follows:
(a), first utilize A level sorter to judge whether comprise car face in bayonet socket image, if had, then intercept car face image, perform next step; If no, then deterministic process terminates;
(b), car face image is carried out rude classification according to its similar programs, be divided into the K class altogether such as popular category-A, popular category-B, Audi's category-A, B level sorter is utilized to judge the car face image intercepted in step (a), identify which kind of it belongs to, the near vehicle model of several front appearance in each class of recognition result, can be comprised;
(c), for the rude classification in step (b), utilize for such training C level sorter further judge, recognition result is meticulous vehicle model.
Described a kind of meticulous vehicle model recognition methods based on multistage classifier, it is characterized in that, in step (1), the design process of B level sorter is as follows:
(1), sample be intercept car face image, it is divided into K class according to the similarity degree of car face.
(2), the design of convolutional neural networks is as follows: in input layer, the size of input picture is 227*227.Neural network entirety adopts two-part structure, and first paragraph is continuous print two convolutional layers, and first convolutional layer uses the convolution kernel that 64 sizes are 7*7, and second convolutional layer uses the convolution kernel that 128 sizes are 5*5, connects a pond layer after each convolutional layer; First time connection is entirely carried out to the result extracted, obtains the feature of 4608 dimensions; Second segment connects a convolutional layer and pond layer again after second convolutional layer, and the convolutional layer of this one-phase uses the convolution kernel that 256 sizes are 3*3; Result is carried out second time entirely to connect, obtain 4096 dimensional features.The result entirely connected twice is simultaneously as the input of softmax sorter.
Described a kind of meticulous vehicle model recognition methods based on multistage classifier, it is characterized in that, in step (1), the design of C level sorter is as follows:
(1), in B level sorter vehicle model is divided in order to K class, remember that the meticulous vehicle model one in the i-th class has n
ikind; Prepare this n
iplant the sample of vehicle, at least 5000 often kind, adopt C level sorter to these sample training, the sorter obtained after training is designated as C
i; Then for the rude classification of K class vehicle model, C level sorter K to be trained altogether; Based on the class of vehicle adopting B level sorter to judge, call corresponding C level sorter and carry out sophisticated category.
(2), the design of convolutional neural networks is as follows: ground floor is input layer; The second layer is convolutional layer, adopts 64 sizes to be the convolution kernel of 11*11; Third layer is pond layer; 4th layer is convolutional layer, adopts the convolution kernel of 128 5*5; Layer 5 is pond layer; Six, seven layers is convolutional layer, adopts the convolution kernel of 256 3*3; 8th layer is convolutional layer, adopts the convolution kernel of 128 6*6; 9th layer is pond layer; Ten, eleventh floor is full articulamentum, and dimension is 4608; Floor 12 is full articulamentum, and dimension is n
i.
Compared with the prior art, beneficial effect of the present invention is embodied in:
1, in recognition methods:
(1) recognition methods based on multistage convolutional neural networks of the present invention's proposition, using the vehicle frontal image under various illumination condition as sample, take full advantage of the feature that its front face feature of vehicle of different automobile types is inconsistent, and not easily by the impact of illumination, weather, there is very strong robustness.
(2) adopt multistage classifier, know method for distinguishing by different level, by first identifying thick class, then segmenting in thick class, identifying concrete vehicle model, the method overcomes the identification difficult problem that vehicle model too much brings in a sense.
2, on recognition effect, identify based on the front face image in bayonet socket image, the present invention can obtain very high discrimination, can meet the needs of actual intelligent transportation system.
Accompanying drawing explanation
Fig. 1 is a kind of process flow diagram of the meticulous vehicle model recognition methods based on multistage classifier.
Fig. 2 is the structural drawing of the convolutional neural networks used in B level sorter.
Fig. 3 is the structural drawing of the convolutional neural networks used in C level sorter.
Embodiment
As shown in Figure 1, a kind of meticulous vehicle type recognition method based on multistage convolutional neural networks, comprises the following steps:
(1) reclassify device is designed, wherein: A level sorter is based on Adaboost algorithm, positive sample is the image comprising face before vehicle, negative sample is the image not comprising face before vehicle, LBP feature is adopted to train, for the vehicle in limited bayonet image, and vehicle sections is intercepted out, form new data set; Vehicle like front appearance, based on convolutional neural networks, for carrying out broad classification to vehicle model, is classified as a class by B level sorter; For the classification results of B level sorter, train multiple C level sorter, C level sorter based on convolutional neural networks, for again carrying out sophisticated category to every class classification results of B level sorter.
Wherein, the detailed design of B level sorter is as follows:
A () sample is the car face image intercepted, it is divided into K class according to the similarity degree of car face, as: popular category-A, popular category-B, Audi's category-A etc.Sample is normalized.
B the design of () convolutional neural networks as shown in Figure 2, specifically describes as follows: in input layer, the size of input picture is 227*227.Neural network entirety adopts two-part structure, and first paragraph is continuous print two convolutional layers, and first convolutional layer uses the convolution kernel that 64 sizes are 7*7, and second convolutional layer uses the convolution kernel that 128 sizes are 5*5, connects a pond layer after each convolutional layer; First time connection is entirely carried out to the result extracted, obtains the feature of 4608 dimensions.Second segment connects a convolutional layer and pond layer again after second convolutional layer, and the convolutional layer of this one-phase uses the convolution kernel that 256 sizes are 3*3; Result is carried out second time entirely to connect, obtain 4096 dimensional features.The result entirely connected twice is simultaneously as the input of softmax sorter.
The detailed design of C level sorter is as follows:
A vehicle model divides in order to K class by () in B level sorter, remember that the meticulous vehicle model one in the i-th class has n
ikind.Prepare this n
iplant the sample of vehicle, at least 5000 often kind, adopt C level sorter to these sample training, the sorter obtained after training is designated as C
i.Then for the rude classification of K class vehicle model, C level sorter K to be trained altogether.Based on the class of vehicle adopting B level sorter to judge, call corresponding C level sorter and carry out sophisticated category.
B the design of () convolutional neural networks as shown in Figure 3, specifically describes as follows: ground floor is input layer; The second layer is convolutional layer, adopts 64 sizes to be the convolution kernel of 11*11; Third layer is pond layer; 4th layer is convolutional layer, adopts the convolution kernel of 128 5*5; Layer 5 is pond layer; Six, seven layers is convolutional layer, adopts the convolution kernel of 256 3*3; 8th layer is convolutional layer, adopts the convolution kernel of 128 6*6; 9th layer is pond layer; Ten, eleventh floor is full articulamentum, and dimension is 4608; Floor 12 is full articulamentum, and dimension is n
i.
(2) training of sorter: based on the image obtained from Gate System, collect Sample Storehouse, carry out the training of reclassify device respectively, wherein, A, B level sorter all only has one; C level sorter has K, for the thick class that B level sorter divides, and the corresponding C level sorter of every class.
(3) for the image obtained from Gate System, adopt multistage classifier to carry out the identification of vehicle model, identification step is as follows:
The first step: first utilize A level sorter to judge whether comprise car face in bayonet socket image, if had, then intercepts car face image, performs second step; If no, then deterministic process terminates.
Second step: car face image is carried out rude classification according to its similar programs, is divided into the K class altogether such as popular category-A, popular category-B, Audi's category-A.Utilize B level sorter to judge the car face image intercepted in previous step, identify which kind of it belongs to.The near vehicle model of several front appearance can be comprised in each class of recognition result.
3rd step: for the rude classification in second step, utilize the C level sorter for such training further to judge, recognition result is meticulous vehicle model, as: popular FV7187ZADBG, popular FV7207ZBDBG.
Unique distinction of the present invention is embodied in:
1. the use of multistage classifier, by first locating, then can carry out rough sort, then the vehicle model that the mode identification carrying out disaggregated classification is meticulous.
2. by collecting the sample under various environment, various illumination condition, its generic being trained, decreasing the influence degree of this algorithm by environment, illumination etc., adding the robustness of algorithm.
To sum up, the present invention, by design multistage classifier, is carried out the location of front face, is then carried out rough sort and the disaggregated classification of vehicle model by two-stage convolutional neural networks respectively, achieve the identification of meticulous vehicle model by LBP+Adaboost method.
Claims (3)
1. based on a meticulous vehicle model recognition methods for multistage classifier, it is characterized in that: design the multistage classifier based on Adaboost algorithm and convolutional neural networks, for carrying out meticulous identification to the vehicle model in Gate System image, comprising the following steps:
(1), reclassify device is designed, wherein: A level sorter is based on Adaboost algorithm, positive sample is the image comprising face before vehicle, negative sample is the image not comprising face before vehicle, LBP feature is adopted to train, for the vehicle in limited bayonet image, and vehicle sections is intercepted out, form new data set; Vehicle like front appearance, based on convolutional neural networks, for carrying out broad classification to vehicle model, is classified as a class by B level sorter; For the classification results of B level sorter, train multiple C level sorter, C level sorter based on convolutional neural networks, for again carrying out sophisticated category to every class classification results of B level sorter;
(2), for the image obtained from Gate System, adopt multistage classifier to carry out the identification of vehicle model, identification step is as follows:
(a), first utilize A level sorter to judge whether comprise car face in bayonet socket image, if had, then intercept car face image, perform next step; If no, then deterministic process terminates;
(b), car face image is carried out rude classification according to its similar programs, be divided into the K class altogether such as popular category-A, popular category-B, Audi's category-A, B level sorter is utilized to judge the car face image intercepted in step (a), identify which kind of it belongs to, the near vehicle model of several front appearance in each class of recognition result, can be comprised;
(c), for the rude classification in step (b), utilize for such training C level sorter further judge, recognition result is meticulous vehicle model.
2. a kind of meticulous vehicle model recognition methods based on multistage classifier according to claim 1, it is characterized in that, in step (1), the design of B level sorter is as follows:
(1), sample be intercept car face image, it is divided into K class according to the similarity degree of car face;
(2), the design of convolutional neural networks is as follows: in input layer, the size of input picture is 227*227, neural network entirety adopts two-part structure, first paragraph is continuous print two convolutional layers, first convolutional layer uses the convolution kernel that 64 sizes are 7*7, second convolutional layer uses the convolution kernel that 128 sizes are 5*5, connects a pond layer after each convolutional layer; First time connection is entirely carried out to the result extracted, obtains the feature of 4608 dimensions; Second segment connects a convolutional layer and pond layer again after second convolutional layer, and the convolutional layer of this one-phase uses the convolution kernel that 256 sizes are 3*3; Result is carried out second time entirely to connect, obtain 4096 dimensional features, the result entirely connected twice is simultaneously as the input of softmax sorter.
3. a kind of meticulous vehicle model recognition methods based on multistage classifier according to claim 1, it is characterized in that, in step (1), the design of C level sorter is as follows:
(1), in B level sorter vehicle model is divided in order to K class, remember that the meticulous vehicle model one in the i-th class has n
ikind; Prepare this n
iplant the sample of vehicle, at least 5000 often kind, adopt C level sorter to these sample training, the sorter obtained after training is designated as C
i; Then for the rude classification of K class vehicle model, C level sorter K to be trained altogether; Based on the class of vehicle adopting B level sorter to judge, call corresponding C level sorter and carry out sophisticated category;
(2), the design of convolutional neural networks is as follows: ground floor is input layer; The second layer is convolutional layer, adopts 64 sizes to be the convolution kernel of 11*11; Third layer is pond layer; 4th layer is convolutional layer, adopts the convolution kernel of 128 5*5; Layer 5 is pond layer; Six, seven layers is convolutional layer, adopts the convolution kernel of 256 3*3; 8th layer is convolutional layer, adopts the convolution kernel of 128 6*6; 9th layer is pond layer; Ten, eleventh floor is full articulamentum, and dimension is 4608; Floor 12 is full articulamentum, and dimension is n
i.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510697287.2A CN105335710A (en) | 2015-10-22 | 2015-10-22 | Fine vehicle model identification method based on multi-stage classifier |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510697287.2A CN105335710A (en) | 2015-10-22 | 2015-10-22 | Fine vehicle model identification method based on multi-stage classifier |
Publications (1)
Publication Number | Publication Date |
---|---|
CN105335710A true CN105335710A (en) | 2016-02-17 |
Family
ID=55286226
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510697287.2A Pending CN105335710A (en) | 2015-10-22 | 2015-10-22 | Fine vehicle model identification method based on multi-stage classifier |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN105335710A (en) |
Cited By (29)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105938560A (en) * | 2016-03-23 | 2016-09-14 | 吉林大学 | Convolutional-neural-network-based vehicle model refined classification system |
CN105957238A (en) * | 2016-05-20 | 2016-09-21 | 聚龙股份有限公司 | Banknote management method and 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 |
CN106056751A (en) * | 2016-05-20 | 2016-10-26 | 聚龙股份有限公司 | Prefix number identification method and system |
CN106372624A (en) * | 2016-10-15 | 2017-02-01 | 杭州艾米机器人有限公司 | Human face recognition method and human face recognition system |
CN106446949A (en) * | 2016-09-26 | 2017-02-22 | 成都通甲优博科技有限责任公司 | Vehicle model identification method and apparatus |
CN106446784A (en) * | 2016-08-30 | 2017-02-22 | 东软集团股份有限公司 | Image detection method and apparatus |
CN106570477A (en) * | 2016-10-28 | 2017-04-19 | 中国科学院自动化研究所 | Vehicle model recognition model construction method based on depth learning and vehicle model recognition method based on depth learning |
CN107909017A (en) * | 2017-11-06 | 2018-04-13 | 余帝乾 | The method, apparatus and system of Car license recognition under a kind of complex background |
CN108304882A (en) * | 2018-02-07 | 2018-07-20 | 腾讯科技(深圳)有限公司 | A kind of image classification method, device and server, user terminal, storage medium |
CN108764022A (en) * | 2018-04-04 | 2018-11-06 | 链家网(北京)科技有限公司 | A kind of image-recognizing method and system |
CN108764134A (en) * | 2018-05-28 | 2018-11-06 | 江苏迪伦智能科技有限公司 | A kind of automatic positioning of polymorphic type instrument and recognition methods suitable for crusing robot |
CN108764051A (en) * | 2018-04-28 | 2018-11-06 | Oppo广东移动通信有限公司 | Image processing method, device and mobile terminal |
CN108875803A (en) * | 2018-05-30 | 2018-11-23 | 长安大学 | A kind of detection of harmful influence haulage vehicle and recognition methods based on video image |
CN108921811A (en) * | 2018-04-03 | 2018-11-30 | 阿里巴巴集团控股有限公司 | Detect method and apparatus, the article damage detector of article damage |
CN109406118A (en) * | 2018-10-15 | 2019-03-01 | 华中科技大学 | A kind of mechanical failure prediction method based on level convolutional neural networks |
CN109509228A (en) * | 2017-09-15 | 2019-03-22 | 安讯士有限公司 | Method for positioning one or more candidate digital images |
CN109583275A (en) * | 2017-09-29 | 2019-04-05 | 大连恒锐科技股份有限公司 | A kind of weight determination method that is barefoot or wearing sock print based on CNN |
CN109934080A (en) * | 2017-12-15 | 2019-06-25 | 财团法人工业技术研究院 | The method and device of countenance identification |
CN109977766A (en) * | 2019-02-18 | 2019-07-05 | 南京航空航天大学 | A kind of fine recognition methods of vehicle based on local feature |
CN110110748A (en) * | 2019-03-29 | 2019-08-09 | 广州思德医疗科技有限公司 | A kind of recognition methods of original image and device |
CN110110652A (en) * | 2019-05-05 | 2019-08-09 | 达闼科技(北京)有限公司 | A kind of object detection method, electronic equipment and storage medium |
CN110598749A (en) * | 2019-08-13 | 2019-12-20 | 广东数鼎科技有限公司 | Image vehicle recognition implementation system based on convolutional neural network AI algorithm |
CN110852795A (en) * | 2019-10-30 | 2020-02-28 | 中国科学院自动化研究所南京人工智能芯片创新研究院 | Mobile phone valuation system and valuation method based on computer vision |
CN110858303A (en) * | 2018-08-22 | 2020-03-03 | 西门子(中国)有限公司 | Method and device for detecting license plate number |
WO2020048265A1 (en) * | 2018-09-06 | 2020-03-12 | 北京市商汤科技开发有限公司 | Methods and apparatuses for multi-level target classification and traffic sign detection, device and medium |
CN111222388A (en) * | 2018-12-11 | 2020-06-02 | 图灵通诺(北京)科技有限公司 | Settlement method and system based on visual identification |
CN113109666A (en) * | 2021-04-09 | 2021-07-13 | 河南省博海大数据科技有限公司 | Track circuit fault diagnosis method based on deep convolutional neural network |
US11200447B2 (en) * | 2016-01-13 | 2021-12-14 | Ford Global Technologies, Llc | Low- and high-fidelity classifiers applied to road-scene images |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20140085475A1 (en) * | 2011-05-19 | 2014-03-27 | The Regents Of The University Of California | Dynamic bayesian networks for vehicle classification in video |
CN104298976A (en) * | 2014-10-16 | 2015-01-21 | 电子科技大学 | License plate detection method based on convolutional neural network |
CN104657748A (en) * | 2015-02-06 | 2015-05-27 | 中国石油大学(华东) | Vehicle type recognition method based on convolutional neural network |
CN104820831A (en) * | 2015-05-13 | 2015-08-05 | 沈阳聚德视频技术有限公司 | Front vehicle face identification method based on AdaBoost license plate location |
-
2015
- 2015-10-22 CN CN201510697287.2A patent/CN105335710A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20140085475A1 (en) * | 2011-05-19 | 2014-03-27 | The Regents Of The University Of California | Dynamic bayesian networks for vehicle classification in video |
CN104298976A (en) * | 2014-10-16 | 2015-01-21 | 电子科技大学 | License plate detection method based on convolutional neural network |
CN104657748A (en) * | 2015-02-06 | 2015-05-27 | 中国石油大学(华东) | Vehicle type recognition method based on convolutional neural network |
CN104820831A (en) * | 2015-05-13 | 2015-08-05 | 沈阳聚德视频技术有限公司 | Front vehicle face identification method based on AdaBoost license plate location |
Cited By (42)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US11200447B2 (en) * | 2016-01-13 | 2021-12-14 | Ford Global Technologies, Llc | Low- and high-fidelity classifiers applied to road-scene images |
CN105938560A (en) * | 2016-03-23 | 2016-09-14 | 吉林大学 | Convolutional-neural-network-based vehicle model refined classification system |
CN105938560B (en) * | 2016-03-23 | 2017-03-22 | 吉林大学 | Convolutional-neural-network-based vehicle model refined classification system |
CN105957238A (en) * | 2016-05-20 | 2016-09-21 | 聚龙股份有限公司 | Banknote management method and system |
CN106056751A (en) * | 2016-05-20 | 2016-10-26 | 聚龙股份有限公司 | Prefix number identification method and system |
CN105957238B (en) * | 2016-05-20 | 2019-02-19 | 聚龙股份有限公司 | A kind of paper currency management method and its system |
CN106056751B (en) * | 2016-05-20 | 2019-04-12 | 聚龙股份有限公司 | The recognition methods and system of serial number |
US10930105B2 (en) | 2016-05-20 | 2021-02-23 | Julong Co., Ltd. | Banknote management method and 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 |
CN106446784A (en) * | 2016-08-30 | 2017-02-22 | 东软集团股份有限公司 | Image detection method and apparatus |
CN106446949A (en) * | 2016-09-26 | 2017-02-22 | 成都通甲优博科技有限责任公司 | Vehicle model identification method and apparatus |
CN106446949B (en) * | 2016-09-26 | 2019-05-21 | 成都通甲优博科技有限责任公司 | A kind of vehicle model recognition methods and device |
CN106372624A (en) * | 2016-10-15 | 2017-02-01 | 杭州艾米机器人有限公司 | Human face recognition method and human face recognition system |
CN106570477A (en) * | 2016-10-28 | 2017-04-19 | 中国科学院自动化研究所 | Vehicle model recognition model construction method based on depth learning and vehicle model recognition method based on depth learning |
CN106570477B (en) * | 2016-10-28 | 2019-07-16 | 中科视语(北京)科技有限公司 | Vehicle cab recognition model building method and model recognizing method based on deep learning |
CN109509228A (en) * | 2017-09-15 | 2019-03-22 | 安讯士有限公司 | Method for positioning one or more candidate digital images |
CN109583275B (en) * | 2017-09-29 | 2021-06-18 | 大连恒锐科技股份有限公司 | Weight judgment method of barefoot footprint based on CNN |
CN109583275A (en) * | 2017-09-29 | 2019-04-05 | 大连恒锐科技股份有限公司 | A kind of weight determination method that is barefoot or wearing sock print based on CNN |
CN107909017A (en) * | 2017-11-06 | 2018-04-13 | 余帝乾 | The method, apparatus and system of Car license recognition under a kind of complex background |
CN109934080A (en) * | 2017-12-15 | 2019-06-25 | 财团法人工业技术研究院 | The method and device of countenance identification |
CN108304882A (en) * | 2018-02-07 | 2018-07-20 | 腾讯科技(深圳)有限公司 | A kind of image classification method, device and server, user terminal, storage medium |
CN108921811A (en) * | 2018-04-03 | 2018-11-30 | 阿里巴巴集团控股有限公司 | Detect method and apparatus, the article damage detector of article damage |
US10929717B2 (en) | 2018-04-03 | 2021-02-23 | Advanced New Technologies Co., Ltd. | Article damage detection |
CN108764022A (en) * | 2018-04-04 | 2018-11-06 | 链家网(北京)科技有限公司 | A kind of image-recognizing method and system |
CN108764051A (en) * | 2018-04-28 | 2018-11-06 | Oppo广东移动通信有限公司 | Image processing method, device and mobile terminal |
CN108764134A (en) * | 2018-05-28 | 2018-11-06 | 江苏迪伦智能科技有限公司 | A kind of automatic positioning of polymorphic type instrument and recognition methods suitable for crusing robot |
CN108875803B (en) * | 2018-05-30 | 2022-06-17 | 长安大学 | Hazardous chemical substance transport vehicle detection and identification method based on video image |
CN108875803A (en) * | 2018-05-30 | 2018-11-23 | 长安大学 | A kind of detection of harmful influence haulage vehicle and recognition methods based on video image |
CN110858303A (en) * | 2018-08-22 | 2020-03-03 | 西门子(中国)有限公司 | Method and device for detecting license plate number |
WO2020048265A1 (en) * | 2018-09-06 | 2020-03-12 | 北京市商汤科技开发有限公司 | Methods and apparatuses for multi-level target classification and traffic sign detection, device and medium |
CN110879950A (en) * | 2018-09-06 | 2020-03-13 | 北京市商汤科技开发有限公司 | Multi-stage target classification and traffic sign detection method and device, equipment and medium |
CN109406118A (en) * | 2018-10-15 | 2019-03-01 | 华中科技大学 | A kind of mechanical failure prediction method based on level convolutional neural networks |
CN111222388A (en) * | 2018-12-11 | 2020-06-02 | 图灵通诺(北京)科技有限公司 | Settlement method and system based on visual identification |
CN111222388B (en) * | 2018-12-11 | 2023-09-19 | 图灵通诺(北京)科技有限公司 | Settlement method and system based on visual recognition |
CN109977766A (en) * | 2019-02-18 | 2019-07-05 | 南京航空航天大学 | A kind of fine recognition methods of vehicle based on local feature |
CN110110748B (en) * | 2019-03-29 | 2021-08-17 | 广州思德医疗科技有限公司 | Original picture identification method and device |
CN110110748A (en) * | 2019-03-29 | 2019-08-09 | 广州思德医疗科技有限公司 | A kind of recognition methods of original image and device |
CN110110652A (en) * | 2019-05-05 | 2019-08-09 | 达闼科技(北京)有限公司 | A kind of object detection method, electronic equipment and storage medium |
CN110598749A (en) * | 2019-08-13 | 2019-12-20 | 广东数鼎科技有限公司 | Image vehicle recognition implementation system based on convolutional neural network AI algorithm |
CN110852795A (en) * | 2019-10-30 | 2020-02-28 | 中国科学院自动化研究所南京人工智能芯片创新研究院 | Mobile phone valuation system and valuation method based on computer vision |
CN113109666A (en) * | 2021-04-09 | 2021-07-13 | 河南省博海大数据科技有限公司 | Track circuit fault diagnosis method based on deep convolutional neural network |
CN113109666B (en) * | 2021-04-09 | 2024-03-15 | 河南省博海大数据科技有限公司 | Rail circuit fault diagnosis method based on deep convolutional neural network |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN105335710A (en) | Fine vehicle model identification method based on multi-stage classifier | |
CN109816024B (en) | Real-time vehicle logo detection method based on multi-scale feature fusion and DCNN | |
CN101859382B (en) | License plate detection and identification method based on maximum stable extremal region | |
CN103279756B (en) | Vehicle detection based on integrated classifier analyzes system and determination method thereof | |
CN105335702B (en) | A kind of bayonet model recognizing method based on statistical learning | |
CN103810505B (en) | Vehicles identifications method and system based on multiple layer description | |
CN103258213B (en) | A kind of for the dynamic vehicle model recognizing method in intelligent transportation system | |
CN109190444B (en) | Method for realizing video-based toll lane vehicle feature recognition system | |
CN103971097B (en) | Vehicle license plate recognition method and system based on multiscale stroke models | |
CN109409337B (en) | Muck vehicle feature identification method based on convolutional neural network | |
CN104463241A (en) | Vehicle type recognition method in intelligent transportation monitoring system | |
CN105184271A (en) | Automatic vehicle detection method based on deep learning | |
CN106127107A (en) | The model recognizing method that multi-channel video information based on license board information and vehicle's contour merges | |
CN108269407B (en) | Security robot capable of automatically managing people stream and logistics | |
CN103488973A (en) | Method and system for recognizing vehicle brand based on image | |
CN107016362B (en) | Vehicle weight recognition method and system based on vehicle front windshield pasted mark | |
CN103390167A (en) | Multi-characteristic layered traffic sign identification method | |
CN103093249A (en) | Taxi identifying method and system based on high-definition video | |
CN108960074B (en) | Small-size pedestrian target detection method based on deep learning | |
CN112613344B (en) | Vehicle track occupation detection method, device, computer equipment and readable storage medium | |
CN103679205A (en) | Preceding car detection method based on shadow hypothesis and layered HOG (histogram of oriented gradient) symmetric characteristic verification | |
CN104200670B (en) | A kind of vehicle model recognizing method based on vehicle earth induction feature | |
CN102184412B (en) | Licence plate number and letter identification method based on minimum-error-rate Bayes classifier | |
CN110516691A (en) | A kind of Vehicular exhaust detection method and device | |
CN111209923A (en) | Deep learning technology-based muck truck cover or uncover identification method |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination |