CN112085018A - License plate recognition system based on neural network - Google Patents
License plate recognition system based on neural network Download PDFInfo
- Publication number
- CN112085018A CN112085018A CN202010876999.1A CN202010876999A CN112085018A CN 112085018 A CN112085018 A CN 112085018A CN 202010876999 A CN202010876999 A CN 202010876999A CN 112085018 A CN112085018 A CN 112085018A
- Authority
- CN
- China
- Prior art keywords
- license plate
- vehicle
- image
- neural network
- license
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/60—Type of objects
- G06V20/62—Text, e.g. of license plates, overlay texts or captions on TV images
- G06V20/63—Scene text, e.g. street names
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/2431—Multiple classes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/253—Fusion techniques of extracted features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
- G06V30/10—Character recognition
- G06V30/14—Image acquisition
- G06V30/148—Segmentation of character regions
- G06V30/153—Segmentation of character regions using recognition of characters or words
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/60—Type of objects
- G06V20/62—Text, e.g. of license plates, overlay texts or captions on TV images
- G06V20/625—License plates
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
- G06V30/10—Character recognition
Abstract
The invention discloses a license plate recognition system based on a neural network.A image acquisition module acquires an image of a right front visual angle in the driving process of a currently driven vehicle; the vehicle detection module is used for detecting vehicles with different sizes in the image in a targeted manner and determining the positions of the vehicles in the image; the license plate rough regression module processes the detected vehicles, learns the approximate position information of the license plate by using a residual error network and obtains the approximate position information of the license plate of each vehicle in the image; the license plate fine classification regression module performs fine classification regression on license plates, a residual error network is used for learning the category information of the license plates and the accurate position information of the license plates, the license plate recognition module divides characters of the license plates according to the category information of the license plates and the accurate position information of the license plates, and then the neural network is used for recognizing the divided characters. On the premise of ensuring the license plate recognition precision, the generalization capability of license plate recognition is improved, and the calculation complexity of a license plate module is reduced.
Description
Technical Field
The invention relates to a license plate recognition system based on a neural network, and belongs to the technical field of license plate recognition systems.
Background
Research shows that automatic license plate detection and recognition is one of key tools for traffic toll collection, vehicle violation tracking and processing and the like, and meanwhile, the method has a long-term effect in vehicle auxiliary driving, and the vehicle auxiliary driving is abbreviated as ADAS. Most existing solutions are limited in nature, such as a license plate recognition charging system in an access control system, which employs a working camera in a static state, and performs license plate recognition processing for a specific license plate template by using a specific viewing angle and a specific resolution. However, in a wider real scene, the license plate identification is difficult due to the complex diversity of roads, the complex diversity of vehicle types and the shielding deformation of the license plate.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provide a license plate recognition system based on a neural network, wherein a set of license plate recognition system with stronger generalization capability and wider applicability is designed through a neural network model, and the accurate information of a license plate, including the type of the license plate, the accurate regression frame of the license plate and the specific information of the license plate, can be obtained, so that the performance of vehicle-mounted ADAS is improved.
In order to achieve the above object, the present invention provides a license plate recognition system based on a neural network, comprising the steps of:
the image acquisition module acquires an image of a right front visual angle in the running process of a currently driven vehicle;
the vehicle detection module is used for detecting vehicles with different sizes in the image in a targeted manner based on an FPN network and an anchor-free end-to-end detection network framework;
the license plate rough regression module processes the detected vehicles in the image, and learns the approximate position information of the license plate in the image by using a residual error network, so as to obtain the approximate position information of the license plate of each vehicle in the image;
the license plate fine classification regression module performs fine classification regression on the license plate according to the rough position information of the license plate obtained by the license plate coarse regression module, learns the category information and the accurate position information of the license plate by using a residual error network,
the license plate recognition module is used for segmenting characters of the license plate in the image according to the category information of the license plate and the accurate position information of the license plate, and then recognizing the segmented characters by utilizing the neural network.
Preferentially, the license plate fine classification regression module adopts 4 corner points of the license plate as characteristic points for determining the accurate position information of the license plate.
Preferably, the image acquisition module comprises a camera, which is mounted on the front side of the car.
Preferentially, the license plate fine classification regression module classifies license plates into a first class, a second class and a third class, wherein the first class comprises a common blue plate, a learner-driven vehicle yellow plate, a police vehicle white plate and a police vehicle black plate, and the license plates of the first class are all 7 digits; the second category comprises new energy green plates, and the license plates of the second category are 8 digits; category three includes yellow cards.
Preferentially, an FPN network is used as a backbone network, the FPN network is used as the backbone network, the image acquired by the camera is zoomed and then input into the backbone network to obtain the characteristic information of each vehicle in the image to obtain the vehicle characteristics, the vehicle characteristics are fused by using an image characteristic pyramid, and the approximate position information of the vehicle in the image is detected by using the fused vehicle characteristics;
in the vehicle detection module, distinguishing the background in the image from the vehicle in the image, judging whether each pixel point in the image belongs to the vehicle, and in the training process of the neural network, learning and distinguishing whether each pixel point belongs to the vehicle or the background by using a Focal local Loss function:
wherein, y is 1 to represent that the current pixel belongs to the vehicle, y is 0 to represent that the current pixel belongs to the background, and y' represents the output result of the neural network classification node; l isflRepresenting true value of current pixel and neural network predictionLoss between measurements;
in the vehicle detection module, in order to obtain the approximate position information of the vehicle in the image, an IOU Loss function is used in the training process of the neural network:
wherein A represents the actual precise position of the vehicle, B represents the vehicle position predicted by the neural network, interaction (A, B) represents the Intersection of the actual precise position of the vehicle and the predicted vehicle position, and Union (A, B) represents the Union of the actual position and the predicted position.
Preferentially, the license plate fine classification regression module is used for accurately distinguishing whether the license plate belongs to one of a blue plate, a green plate, a yellow plate or a background, and accurately positioning the accurate position information of the license plate.
Preferably, α is 0.25 and γ is 2.
The invention achieves the following beneficial effects:
in a real scene, the license plate accounts for a small proportion of the whole image, so the first step is to detect the positions of all vehicles in the image by using a vehicle detector based on a convolutional neural network. And secondly, performing a coarse license plate regression operation according to the obtained positions of the vehicles, so that the possible license plate areas can be narrowed again. And thirdly, performing classification regression operation on the rough position of the license plate obtained by the rough regression operation to obtain the accurate position and the category of the license plate. The algorithm of the invention is mainly divided into four parts: vehicle detection, coarse license plate regression, fine license plate classification regression and license plate identification. And finally, identifying the obtained license plate. The invention designs a license plate recognition system based on a convolutional neural network, which is mainly applied to a vehicle-mounted ADAS recognition scene. The method effectively improves the generalization capability of license plate recognition on the premise of ensuring the precision of license plate recognition, simultaneously maximizes the calculation complexity of the license plate module, and meets the application requirements.
Drawings
FIG. 1 is a framework diagram of the present invention;
FIG. 2 is a flow chart of the present invention;
FIG. 3 is a block diagram of an FPN network in accordance with the present invention;
fig. 4 is a structural diagram of a residual error network in the present invention.
Detailed Description
The following examples are only for illustrating the technical solutions of the present invention more clearly, and the protection scope of the present invention is not limited thereby.
The license plate recognition system based on the neural network comprises the following steps:
the image acquisition module acquires an image of a right front visual angle in the running process of a currently driven vehicle;
the vehicle detection module is used for detecting vehicles with different sizes in the image in a targeted manner based on an FPN network and an anchor-free end-to-end detection network framework;
the license plate rough regression module processes the detected vehicles in the image, and learns the approximate position information of the license plate in the image by using a residual error network, so as to obtain the approximate position information of the license plate of each vehicle in the image;
the license plate fine classification regression module performs fine classification regression on the license plate according to the rough position information of the license plate obtained by the license plate coarse regression module, learns the category information and the accurate position information of the license plate by using a residual error network,
the license plate recognition module is used for segmenting characters of the license plate in the image according to the category information of the license plate and the accurate position information of the license plate, and then recognizing the segmented characters by utilizing the neural network.
Furthermore, the license plate fine classification regression module adopts 4 corner points of the license plate as characteristic points for determining the accurate position information of the license plate.
Further, the image acquisition module comprises a camera, and the camera is installed on the front side of the automobile.
Further, the license plate fine classification regression module classifies the license plates into a first class, a second class and a third class, wherein the first class comprises a common blue plate, a learner-driven vehicle yellow plate, a police vehicle white plate and a police vehicle black plate, and the license plates of the first class are all 7 digits; the second category comprises new energy green plates, and the license plates of the second category are 8 digits; category three includes yellow cards.
Further, an FPN network is used as a backbone network, the FPN network is used as the backbone network, the image acquired by the camera is zoomed and then input into the backbone network to obtain the characteristic information of each vehicle in the image to obtain the vehicle characteristics, the vehicle characteristics are fused by using an image characteristic pyramid, and the approximate position information of the vehicle in the image is detected by using the fused vehicle characteristics;
in the vehicle detection module, distinguishing the background in the image from the vehicle in the image, judging whether each pixel point in the image belongs to the vehicle, and in the training process of the neural network, learning and distinguishing whether each pixel point belongs to the vehicle or the background by using a Focal local Loss function:
wherein, y is 1 to represent that the current pixel belongs to the vehicle, y is 0 to represent that the current pixel belongs to the background, and y' represents the output result of the neural network classification node; l isflRepresenting the loss between the true value of the current pixel point and the predicted value of the neural network;
in the vehicle detection module, in order to obtain the approximate position information of the vehicle in the image, an IOU Loss function is used in the training process of the neural network:
wherein A represents the actual precise position of the vehicle, B represents the vehicle position predicted by the neural network, interaction (A, B) represents the Intersection of the actual precise position of the vehicle and the predicted vehicle position, and Union (A, B) represents the Union of the actual position and the predicted position.
Furthermore, the license plate fine classification regression module is used for accurately distinguishing whether the license plate belongs to one of a blue plate, a green plate, a yellow plate or a background, and accurately positioning the accurate position information of the license plate.
Further, α is 0.25 and γ is 2.
It can be seen from the formula that L is the closer y is to yflThe smaller the value, the better the neural network learning. Since the vehicle occupancy is small and the background occupancy is large in the entire image, imbalance of positive and negative samples is caused. We use the balance factor a of 0.25 to equalize the proportion of positive and negative samples. While another factor γ of 2 is used to reduce the classification of simple samples by the neural network and pay more attention to the classification of difficult samples. Smaller values of IOUloss represent closer real and predicted locations.
The FPN mainly solves the multi-scale problem in object detection, greatly improves the performance of small object detection under the condition of basically not increasing the calculated amount of an original model through simple network connection change, and the left side of the FPN is called from bottom to top and the right side of the FPN is called from top to bottom in the figure 3.
The bottom-up process is the normal forward propagation process of the neural network, and the feature map is calculated by a convolution kernel and generally becomes smaller and smaller. The top-down process is to up-sample a high-level feature map with higher abstraction and stronger semantic meaning and then transversely connect the feature to the feature of the previous layer, so that the high-level feature is enhanced, the feature map used by each layer of prediction is fused with the features with different resolution ratios and different semantic strengths, the detection of an object with the corresponding resolution ratio can be completed, and each layer is ensured to have proper resolution ratio and strong semantic features.
It is worth noting that: the features of the two layers that are connected laterally are identical in spatial dimension, which allows the use of the underlying positioning detail information.
The method comprises the following specific steps:
(1) the vehicle-mounted monocular camera is mounted on the window glass, and vehicle condition information right in front of the vehicle is collected in the running process of the vehicle. The input size of the captured picture was 1280 × 720P.
(2) The vehicle detection model uses an end-to-end detection network framework based on an FPN network and anchor-free, and is mainly structurally shown in FIG. 3, the image feature pyramid structure can effectively detect targets with different sizes, and the recall rate of vehicle detection can be improved as much as possible.
Zooming the image acquired by the camera, inputting the zoomed image into a backbone network to obtain the position information of each vehicle in the image, fusing the vehicle characteristics by using an image characteristic pyramid, and detecting the vehicle position in the license plate image by using the fused vehicle characteristics; in a vehicle detection module, in order to judge whether each pixel point in an image is a vehicle, a background and the vehicle in the image need to be distinguished, and in the training process of a neural network, a Focal local Loss function is used for learning and distinguishing whether each pixel point is on the vehicle or in the background image:
wherein, y equals 1 to represent that the current pixel belongs to the vehicle, and y equals 0 to represent that the current pixel belongs to the background. y' represents the output of the neural network classification node. L isflRepresenting the loss between the true value of the current pixel point and the predicted value of the neural network. It can be seen from the formula that L is the closer y is to yflThe smaller the value, the better the neural network learning. Since the vehicle occupancy is small and the background occupancy is large in the entire image, imbalance of positive and negative samples is caused. We use the balance factor a of 0.25 to equalize the proportion of positive and negative samples. While another factor γ of 2 is used to reduce the classification of simple samples by the neural network and pay more attention to the classification of difficult samples.
In the vehicle detection module, in order to obtain the position information of the vehicle in the image, an IOU Loss function is used in the training process of the neural network:
wherein A represents the position information of the target vehicle, B represents the position information predicted by the neural network, the intersection (A, B) represents the intersection of the real position and the predicted position, and the Union (A, B) represents the Union of the real position and the predicted position. Smaller values of IOUloss represent closer real and predicted locations.
(3) The license plate rough regression module processes the detected vehicles and learns the approximate position information of the license plate by using a residual error network, so that the approximate position information of the license plate of each vehicle in the whole image is obtained. Residual network architecture fig. 4.
The coarse license plate regression network performs coarse regression on the vehicle position information acquired by the detection network, so that approximate position information of the license plate in the image is acquired.
(4) The license plate fine classification regression module performs fine classification regression on the license plate according to the approximate license plate position information obtained by the license plate coarse regression module, and learns the category information and the accurate position information of the license plate by using a residual error network. The invention divides the license plate into three categories according to the width and the height of the license plate:
class 1: the number of the license plate is 7, and the size of the license plate is 440mm to 140 mm; class 2: the number of new energy green plates is 8, and the size of the plates is 480mm x 140 mm; class 3: yellow cards, license plate size 440mm 220 mm.
The classification is beneficial to determining the real width of the license plate by the subsequent vehicle ADAS module according to the category information of the license plate so as to accurately measure the distance.
Considering the conditions of license plate deformation and the like caused by imaging, 4 corner points of a license plate are used as characteristic points for determining license plate position information.
(5) The license plate recognition module performs character segmentation on the license plate according to the previous license plate type and the license plate position information to obtain an independent Chinese character or letter, then recognizes the segmented characters by utilizing a neural network to obtain the Chinese characters and the letters which are sequentially arranged from left to right, and finally obtains the detailed information of the license plate, wherein the output detailed information comprises the license plate number and the license plate color.
The above description is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, several modifications and variations can be made without departing from the technical principle of the present invention, and these modifications and variations should also be regarded as the protection scope of the present invention.
Claims (7)
1. The license plate recognition system based on the neural network is characterized by comprising the following steps of:
the image acquisition module acquires an image of a right front visual angle in the running process of a currently driven vehicle;
the vehicle detection module is used for detecting vehicles with different sizes in the image in a targeted manner based on an FPN network and an anchor-free end-to-end detection network framework;
the license plate rough regression module processes the detected vehicles in the image, and learns the approximate position information of the license plate in the image by using a residual error network, so as to obtain the approximate position information of the license plate of each vehicle in the image;
the license plate fine classification regression module performs fine classification regression on the license plate according to the rough position information of the license plate obtained by the license plate coarse regression module, learns the category information and the accurate position information of the license plate by using a residual error network,
the license plate recognition module is used for segmenting characters of the license plate in the image according to the category information of the license plate and the accurate position information of the license plate, and then recognizing the segmented characters by utilizing the neural network.
2. The neural network-based license plate recognition system of claim 1, wherein the license plate fine classification regression module employs 4 corner points of the license plate as feature points for determining precise location information of the license plate.
3. The neural network-based license plate recognition system of claim 1, wherein the image acquisition module comprises a camera, and the camera is mounted on the front side of the automobile.
4. The neural network-based license plate recognition system of claim 1, wherein the license plate fine classification regression module classifies license plates into a first class, a second class and a third class, the first class comprises a common blue plate, a learner-driven vehicle yellow plate, a police vehicle white plate and a police vehicle black plate, and the license plates of the first class are all 7 digits; the second category comprises new energy green plates, and the license plates of the second category are 8 digits; category three includes yellow cards.
5. The neural network-based license plate recognition system of claim 1, wherein an FPN network is used as a backbone network, the FPN network is used as the backbone network, an image acquired by a camera is scaled and then input to the backbone network to obtain feature information of each vehicle in the image to obtain vehicle features, the vehicle features are fused by using an image feature pyramid, and the approximate position information of the vehicle in the image is detected by using the fused vehicle features;
in a vehicle detection module, distinguishing a background in an image from a vehicle in the image, judging whether each pixel point in the image belongs to the vehicle, and in the training process of a neural network, learning and distinguishing whether each pixel point belongs to the vehicle or the background by using a FocalLoss loss function:
wherein, y is 1 to represent that the current pixel belongs to the vehicle, y is 0 to represent that the current pixel belongs to the background, and y' represents the output result of the neural network classification node; l isflRepresenting the loss between the true value of the current pixel point and the predicted value of the neural network;
in the vehicle detection module, in order to obtain the approximate position information of the vehicle in the image, an IOU Loss function is used in the training process of the neural network:
wherein A represents the actual precise position of the vehicle, B represents the vehicle position predicted by the neural network, interaction (A, B) represents the Intersection of the actual precise position of the vehicle and the predicted vehicle position, and Union (A, B) represents the Union of the actual position and the predicted position.
6. The neural network-based license plate recognition system of claim 4, wherein the license plate fine classification regression module is used for accurately distinguishing whether the license plate belongs to one of a blue plate, a green plate, a yellow plate or a background so as to accurately position the accurate position information of the license plate.
7. The neural network-based license plate recognition system of claim 5, wherein α is 0.25 and γ is 2.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010876999.1A CN112085018A (en) | 2020-08-27 | 2020-08-27 | License plate recognition system based on neural network |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010876999.1A CN112085018A (en) | 2020-08-27 | 2020-08-27 | License plate recognition system based on neural network |
Publications (1)
Publication Number | Publication Date |
---|---|
CN112085018A true CN112085018A (en) | 2020-12-15 |
Family
ID=73729628
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010876999.1A Pending CN112085018A (en) | 2020-08-27 | 2020-08-27 | License plate recognition system based on neural network |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN112085018A (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113177503A (en) * | 2021-05-13 | 2021-07-27 | 中国人民解放军陆军炮兵防空兵学院 | Arbitrary orientation target twelve parameter detection method based on YOLOV5 |
CN115376119A (en) * | 2022-10-25 | 2022-11-22 | 珠海亿智电子科技有限公司 | License plate recognition method and device, license plate recognition equipment and storage medium |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107085696A (en) * | 2016-10-15 | 2017-08-22 | 安徽百诚慧通科技有限公司 | A kind of vehicle location and type identifier method based on bayonet socket image |
CN111126399A (en) * | 2019-12-28 | 2020-05-08 | 苏州科达科技股份有限公司 | Image detection method, device and equipment and readable storage medium |
CN111582261A (en) * | 2020-04-30 | 2020-08-25 | 浙江大华技术股份有限公司 | License plate recognition method and device for non-motor vehicle |
-
2020
- 2020-08-27 CN CN202010876999.1A patent/CN112085018A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107085696A (en) * | 2016-10-15 | 2017-08-22 | 安徽百诚慧通科技有限公司 | A kind of vehicle location and type identifier method based on bayonet socket image |
CN111126399A (en) * | 2019-12-28 | 2020-05-08 | 苏州科达科技股份有限公司 | Image detection method, device and equipment and readable storage medium |
CN111582261A (en) * | 2020-04-30 | 2020-08-25 | 浙江大华技术股份有限公司 | License plate recognition method and device for non-motor vehicle |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113177503A (en) * | 2021-05-13 | 2021-07-27 | 中国人民解放军陆军炮兵防空兵学院 | Arbitrary orientation target twelve parameter detection method based on YOLOV5 |
CN115376119A (en) * | 2022-10-25 | 2022-11-22 | 珠海亿智电子科技有限公司 | License plate recognition method and device, license plate recognition equipment and storage medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110969160B (en) | License plate image correction and recognition method and system based on deep learning | |
CN111368687B (en) | Sidewalk vehicle illegal parking detection method based on target detection and semantic segmentation | |
CN109977812B (en) | Vehicle-mounted video target detection method based on deep learning | |
CN111178291B (en) | Parking payment system and parking payment method | |
CN103279756B (en) | Vehicle detection based on integrated classifier analyzes system and determination method thereof | |
CN103824452B (en) | A kind of peccancy parking detector based on panoramic vision of lightweight | |
CN102722704B (en) | Method and system for recognizing vehicle license plate by integrating video dynamic tracking | |
CN103824066B (en) | A kind of licence plate recognition method based on video flowing | |
CA2885019C (en) | Robust windshield detection via landmark localization | |
CN110619279B (en) | Road traffic sign instance segmentation method based on tracking | |
CN110378236A (en) | Testing vehicle register identification model construction, recognition methods and system based on deep learning | |
CN102915433B (en) | Character combination-based license plate positioning and identifying method | |
CN103903018A (en) | Method and system for positioning license plate in complex scene | |
CN106128115A (en) | A kind of fusion method based on twin camera detection Traffic Information | |
CN106709530A (en) | License plate recognition method based on video | |
CN102902983B (en) | A kind of taxi identification method based on support vector machine | |
CN111965636A (en) | Night target detection method based on millimeter wave radar and vision fusion | |
CN106934374A (en) | The recognition methods of traffic signboard and system in a kind of haze scene | |
CN112085018A (en) | License plate recognition system based on neural network | |
CN112115800A (en) | Vehicle combination recognition system and method based on deep learning target detection | |
Nguwi et al. | Number plate recognition in noisy image | |
Wen et al. | A rear-vehicle detection system for static images based on monocular vision | |
CN115601717B (en) | Deep learning-based traffic offence behavior classification detection method and SoC chip | |
CN116796024A (en) | Method and system for processing evidence chain of fee-escaping vehicle | |
CN115376082A (en) | Lane line detection method integrating traditional feature extraction and deep neural network |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20201215 |
|
RJ01 | Rejection of invention patent application after publication |