CN108009543B - License plate recognition method and device - Google Patents
License plate recognition method and device Download PDFInfo
- Publication number
- CN108009543B CN108009543B CN201711223109.1A CN201711223109A CN108009543B CN 108009543 B CN108009543 B CN 108009543B CN 201711223109 A CN201711223109 A CN 201711223109A CN 108009543 B CN108009543 B CN 108009543B
- Authority
- CN
- China
- Prior art keywords
- license plate
- candidate frame
- target
- neural network
- image
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 44
- 238000012545 processing Methods 0.000 claims abstract description 79
- 239000013598 vector Substances 0.000 claims abstract description 73
- 230000011218 segmentation Effects 0.000 claims abstract description 48
- 238000013528 artificial neural network Methods 0.000 claims abstract description 28
- 238000013527 convolutional neural network Methods 0.000 claims description 65
- 238000010586 diagram Methods 0.000 claims description 37
- 230000007613 environmental effect Effects 0.000 claims description 20
- 230000005764 inhibitory process Effects 0.000 claims description 12
- 238000010606 normalization Methods 0.000 claims description 8
- 238000000354 decomposition reaction Methods 0.000 claims description 7
- 238000003709 image segmentation Methods 0.000 claims description 6
- 230000009467 reduction Effects 0.000 claims description 3
- 238000004590 computer program Methods 0.000 description 18
- 239000010410 layer Substances 0.000 description 17
- 230000006870 function Effects 0.000 description 13
- 238000012549 training Methods 0.000 description 10
- 239000002356 single layer Substances 0.000 description 9
- 230000008569 process Effects 0.000 description 8
- 238000012986 modification Methods 0.000 description 5
- 230000004048 modification Effects 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000002950 deficient Effects 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 206010039203 Road traffic accident Diseases 0.000 description 1
- 230000004913 activation Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000006740 morphological transformation Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000012552 review Methods 0.000 description 1
- 230000001629 suppression Effects 0.000 description 1
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
- 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
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
- G08G1/017—Detecting movement of traffic to be counted or controlled identifying vehicles
- G08G1/0175—Detecting movement of traffic to be counted or controlled identifying vehicles by photographing vehicles, e.g. when violating traffic rules
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Character Discrimination (AREA)
- Character Input (AREA)
- Image Analysis (AREA)
- Traffic Control Systems (AREA)
Abstract
The application provides a license plate recognition method and a license plate recognition device, wherein the method comprises the following steps: the method comprises the steps of obtaining an image to be processed, inputting the image to be processed to a first convolution neural network to obtain a first license plate candidate frame set and a first boundary frame regression vector, calibrating the first license plate candidate frame set through the first boundary frame regression vector, conducting de-overlapping processing on the first license plate candidate frame set to obtain a second license plate candidate frame set, inputting the second license plate candidate frame set to a second convolution neural network to eliminate wrong candidate frames to obtain a first target license plate candidate frame, obtaining a target license plate type and a corresponding second boundary frame regression vector at the same time, conducting regression calibration on the first target license plate candidate frame according to the second boundary frame regression vector to obtain a second target candidate frame, namely a final license plate area, feeding the target license plate type and the final license plate area back to the second convolution neural network to conduct character segmentation to obtain a plurality of characters, and conducting character recognition to obtain a target license plate. The license plate positioning and recognizing method and device can improve license plate positioning and recognizing precision.
Description
Technical Field
The invention relates to the technical field of image processing, in particular to a license plate recognition method and device.
Background
With the rapid development of the transportation industry, people go out more and more conveniently, but the traffic safety problem becomes more and more a key concern of people along with the phenomena of frequent traffic accidents, traffic jam and the like. In the background of modern technology, an Intelligent Transportation System (ITS) is introduced, and a Vehicle License Plate is very important as an attribute of a currently only externally marked Vehicle, so that a Vehicle License Plate Recognition technology (VLPR) is a very critical technology for the current Transportation industry, and the construction of a License Plate Recognition System (LPRS) is brought forward.
Specifically, in a real application scene, the existing license plate information is usually taken pictures or photographs by a camera, and particularly, because the vehicle is in a traveling process, the influence of street lamps or vehicle lamps at night, the influence of various environments such as weather and climate and the like are likely to cause that the photographed imaging picture is unclear, so that the positioning of the license plate, the feature extraction of the license plate characters and the recognition work of the license plate have great challenges, and therefore, the problem of how to improve the license plate recognition accuracy needs to be solved urgently.
Disclosure of Invention
The embodiment of the invention provides a license plate recognition method and device, which can improve the license plate recognition precision.
The first aspect of the embodiments of the present invention provides a license plate recognition method, including:
acquiring an image to be processed;
inputting the image to be processed into a first convolution neural network for processing to obtain a first license plate candidate frame set and a first boundary frame regression vector;
calibrating the first license plate candidate frame set through the first boundary frame regression vector, and performing de-overlapping processing on the calibrated first license plate candidate frame set to obtain a second license plate candidate frame set;
inputting the second license plate candidate frame set into a second convolutional neural network for processing to obtain a first target candidate frame, a second boundary frame regression vector and a target license plate type;
calibrating the first target candidate frame through the second bounding box regression vector to obtain a second target candidate frame;
and feeding back the type of the target license plate and the second target candidate frame to the second convolutional neural network for character segmentation to obtain a plurality of characters, and performing character recognition on the plurality of characters to obtain the target license plate.
A second aspect of an embodiment of the present invention provides a license plate recognition apparatus, including:
the first acquisition unit is used for acquiring an image to be processed;
the first processing unit is used for inputting the image to be processed into a first convolution neural network for processing to obtain a first license plate candidate frame set and a first boundary frame regression vector;
the first calibration unit is used for calibrating the first license plate candidate frame set through the first boundary frame regression vector and performing de-overlapping processing on the calibrated first license plate candidate frame set to obtain a second license plate candidate frame set;
the second processing unit is used for inputting the second license plate candidate frame set into a second convolutional neural network for processing to obtain a first target candidate frame, a second boundary frame regression vector and a target license plate type;
the second calibration unit is used for calibrating the first target candidate frame through the second boundary frame regression vector to obtain a second target candidate frame;
and the recognition unit is used for feeding the target license plate type and the second target candidate frame back to the second convolutional neural network for character segmentation to obtain a plurality of characters, and performing character recognition on the plurality of characters to obtain the target license plate.
In a third aspect, an embodiment of the present invention provides a mobile terminal, including: a processor and a memory; and one or more programs stored in the memory and configured to be executed by the processor, the programs including instructions for some or all of the steps as described in the first aspect.
In a fourth aspect, the present invention provides a computer-readable storage medium, where the computer-readable storage medium is used for storing a computer program, where the computer program is used to make a computer execute some or all of the steps described in the first aspect of the present invention.
In a fifth aspect, embodiments of the present invention provide a computer program product, wherein the computer program product comprises a non-transitory computer-readable storage medium storing a computer program, the computer program being operable to cause a computer to perform some or all of the steps as described in the first aspect of embodiments of the present invention. The computer program product may be a software installation package.
The embodiment of the invention has the following beneficial effects:
it can be seen that, through the embodiment of the present invention, an image to be processed is obtained, the image to be processed is input to a first convolutional neural network for processing, a first license plate candidate frame set and a first bounding box regression vector are obtained, the first license plate candidate frame set is calibrated through the first bounding box regression vector, the calibrated first license plate candidate frame set is subjected to de-overlapping processing, a second license plate candidate frame set is obtained, the second license plate candidate frame set is input to a second convolutional neural network for processing, a first target candidate frame, a second bounding box regression vector and a target license plate type are obtained, the first target candidate frame is calibrated through the second bounding box regression vector, a second target candidate frame is obtained, the target license plate type and the second target candidate frame are fed back to the second convolutional neural network for character segmentation, a plurality of characters are obtained, and character recognition is performed on the plurality of characters, and obtaining the target license plate, thereby positioning and classifying the license plate based on the secondary convolution neural network and improving the license plate identification precision.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings needed to be used in the description of the embodiments are briefly introduced below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.
Fig. 1 is a schematic flowchart of a license plate recognition method according to a first embodiment of the present invention;
fig. 1a is a schematic flowchart of another embodiment of a license plate recognition method according to the present invention;
fig. 2 is a schematic flowchart of a license plate recognition method according to a second embodiment of the present invention;
fig. 3a is a schematic structural diagram of a license plate recognition device according to an embodiment of the present invention;
FIG. 3b is a schematic structural diagram of a first processing unit of the license plate recognition device depicted in FIG. 3a according to an embodiment of the present invention;
FIG. 3c is a schematic diagram of another structure of the license plate recognition device depicted in FIG. 3a according to an embodiment of the present invention;
fig. 4 is a schematic structural diagram of a license plate recognition device according to an embodiment of the present invention.
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 some, not all, embodiments of the present invention. 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.
The terms "first," "second," "third," and "fourth," etc. in the description and claims of the invention and in the accompanying drawings are used for distinguishing between different objects and not for describing a particular order. Furthermore, the terms "include" and "have," as well as any variations thereof, are intended to cover non-exclusive inclusions. For example, a process, method, system, article, or apparatus that comprises a list of steps or elements is not limited to only those steps or elements listed, but may alternatively include other steps or elements not listed, or inherent to such process, method, article, or apparatus.
Reference herein to "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the invention. The appearances of the phrase in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments. It is explicitly and implicitly understood by one skilled in the art that the embodiments described herein can be combined with other embodiments.
The license plate recognition device described in the embodiment of the present invention may include a smart Phone (such as an Android Phone, an iOS Phone, a Windows Phone, etc.), a tablet computer, a video matrix, a monitoring platform, a vehicle-mounted device, a satellite, a palm computer, a notebook computer, a Mobile Internet device (MID, Mobile Internet Devices), a wearable device, etc., which are examples and not exhaustive, and include but are not limited to the above Devices, and of course, the license plate recognition device may also be a server.
It should be noted that the present invention includes a first convolutional neural network and a second convolutional neural network, where the first convolutional neural network may be a full convolutional neural network composed of 4 convolutional layers, and the main task is to perform sliding window on an input image, perform coarse classification (license plate/non-license plate) on a sliding window region, and perform regression on a license plate candidate frame. The second convolutional neural network can be a two-branch network, and Batch Normalization is added for Normalization, the main task can be to perform subdivision on the license plate candidate frame (for example, blue plate, single-layer yellow plate, double-layer yellow plate, black plate, white plate, green plate or non-license plate), further regression is performed on the candidate frame according to the license plate type to obtain an accurate license plate area, and character segmentation can be performed on the license plate area by utilizing the license plate type. Both convolutional neural networks use the PReLU as an activation function. Before the embodiment of the invention is implemented, two convolutional neural networks can be trained through a large number of license plate sample sets to obtain a first convolutional neural network and a second convolutional neural network, and three types of training samples can be used in the training process: the method comprises the steps of obtaining a license plate sample, a non-license plate sample and a partial license plate sample (a defective license plate, for example, the complete license plate is YuALN 673, and the defective license plate is YuALN 67), wherein the three training samples can be obtained by randomly sliding a window on an original image containing the license plate. In the embodiment of the present invention, that is, before the method provided by the following embodiment of the present invention is implemented, in the training process, the overlapping degree IOU between the license plate sample and the real license plate label frame is greater than 0.7, the overlapping degree IOU between the non-license plate sample and the real license plate label frame is less than 0.3, and the overlapping degree IOU between a part of license plates and the real license plate label frame is greater than 0.4 and less than 0.7.
Further, in a network training stage, a license plate classification task is trained by utilizing license plate and non-license plate samples, and license plate classification adopts a loss function of a softmax regression algorithm:
wherein, k is a classification mark,
for the probability of a sample i being classified as a class j,
is the true label of the specimen and,
as an illustrative function, when the sample is labeled
Equal to category j, the function value is 1, otherwise the function value is 0.
Since the first convolution neural network only carries out coarse classification of the license plate/non-license plate, the value of the category j is 0 and 1. And the second convolutional neural network needs to subdivide the license plate candidate frame into a blue plate, a single-layer yellow plate, a double-layer yellow plate, a black plate, a white plate, a green plate or a non-license plate, and the like, so that the value of the category j is 0, 1, 2, 3, 4, 5 or 6.
Training a license plate classification task, simultaneously training a license plate frame regression task by using a license plate and part of license plate samples, and calculating a regression loss function through Euclidean:
wherein,
in order to obtain the coordinates of the license plate frame through network prediction, y is the coordinates of a real license plate label frame,
and y is a quadruple (top left x, top left y, wide, high).
Throughout the training process, the overall training learning objective can be expressed as minimizing the following function:
where N is the number of training samples, αjWhich indicates the importance of the task or tasks,
is indicated for the type of sample, an
αdet＝1，αbox＝0.5，LjAs a corresponding loss function.
Based on this, the embodiment of the invention provides a license plate identification method, which comprises the following steps:
acquiring an image to be processed;
inputting the image to be processed into a first convolution neural network for processing to obtain a first license plate candidate frame set and a first boundary frame regression vector;
calibrating the first license plate candidate frame set through the first boundary frame regression vector, and performing de-overlapping processing on the calibrated first license plate candidate frame set to obtain a second license plate candidate frame set;
inputting the second license plate candidate frame set into a second convolutional neural network for processing to obtain a first target candidate frame, a second boundary frame regression vector and a target license plate type;
calibrating the first target candidate frame through the second bounding box regression vector to obtain a second target candidate frame;
and feeding back the type of the target license plate and the second target candidate frame to the second convolutional neural network for character segmentation to obtain a plurality of characters, and performing character recognition on the plurality of characters to obtain the target license plate.
It can be seen that, through the embodiment of the present invention, an image to be processed is obtained, the image to be processed is input to a first convolutional neural network for processing, a first license plate candidate frame set and a first bounding frame regression vector are obtained, the first license plate candidate frame set is calibrated through the first bounding frame regression vector, the calibrated first license plate candidate frame set is subjected to de-overlapping processing, a second license plate candidate frame set is obtained, the second license plate candidate frame set is input to a second convolutional neural network for processing, a first target candidate frame, a second bounding frame regression vector and a target license plate type are obtained, the first target candidate frame is calibrated through the second bounding frame regression vector, a second target candidate frame is obtained, the target license plate type and the second target candidate frame are fed back to the second convolutional neural network for character segmentation, a plurality of characters are obtained, and character recognition is performed on the plurality of characters, and obtaining the target license plate, thereby positioning and classifying the license plate based on the secondary convolution neural network and improving the license plate identification precision.
Therefore, the license plate recognition device provided by the embodiment of the invention can be implemented in various environments such as the influence of street lamps or vehicle lamps at night, weather and climate and the like in the process of vehicle traveling, so that the license plate recognition precision is improved.
Fig. 1 is a schematic flowchart illustrating a license plate recognition method according to a first embodiment of the present invention. The license plate recognition method described in the embodiment includes the following steps:
101. and acquiring an image to be processed.
The image to be processed may be an image including a license plate, and the license plate may be at least one of the following: blue cards, single-layer yellow cards, double-license plates (e.g., hong kong inbound vehicles), double-layer yellow cards, black cards, white cards, green cards, and the like.
102. And inputting the image to be processed into a first convolution neural network for processing to obtain a first license plate candidate frame set and a first boundary frame regression vector.
The first convolution neural network can be a full convolution neural network consisting of 4 convolution layers, and is mainly used for performing window sliding on an input image, performing rough classification (license plate/non-license plate) on a window sliding area and performing regression on a license plate candidate frame.
Optionally, in the step 102, inputting the image to be processed into the first convolutional neural network for processing, so as to obtain the first candidate frame set of the license plate and the first regression vector of the bounding box, which may include the following steps:
21. performing multi-scale decomposition on the image to be processed according to the attribute parameters of the license plate to obtain a plurality of sub-images;
22. inputting each sub-image in the plurality of sub-images into the first convolution neural network respectively to obtain a plurality of thermodynamic diagrams of license plate confidence degrees and a first boundary box regression vector corresponding to the thermodynamic diagrams;
23. and traversing the plurality of thermodynamic diagrams, and performing reduction operation on areas corresponding to points in the plurality of thermodynamic diagrams, wherein the points are larger than a confidence coefficient threshold value, so as to obtain the first license plate candidate frame set.
Wherein, the attribute parameters may include at least one of the following: the size ratio of the license plate, the position of the license plate, the shape and the color of the license plate, the positions of the license plate screws, the number of the license plate screws and the like.
For example, in the implementation of the embodiment of the present invention, the input image may be subjected to multi-scale transformation by using a minimum license plate detection size constraint condition to obtain a pyramid image (including multiple images), each layer of the pyramid image is input to the first convolutional neural network, and since the first convolutional neural network has only 4 convolutional layers and the size of the grid is small, a thermodynamic diagram of the license plate confidence and corresponding regression vectors (lx, ly, rx, ry) of the boundary frame may be obtained quickly, the thermodynamic diagram of the license plate confidence is traversed, and the region corresponding to the point in the thermodynamic diagram that is greater than the confidence threshold is restored, so that the corresponding license plate candidate frame may be obtained.
103. And calibrating the first license plate candidate frame set through the first boundary frame regression vector, and performing de-overlapping processing on the calibrated first license plate candidate frame set to obtain a second license plate candidate frame set.
The first license plate candidate frame set can be calibrated through the first boundary frame regression vector, so that the position of the first license plate candidate frame set is more accurate, in addition, the first license plate candidate frame set after calibration is subjected to overlap removal processing, on one hand, the later-stage data processing amount can be reduced, and on the other hand, the license plate positioning accuracy can also be improved.
Optionally, the step 103 of performing de-overlapping processing on the calibrated first license plate candidate frame set specifically includes the following steps:
eliminating the candidate frames with the overlapping degree larger than the overlapping degree threshold value in the first license plate candidate frame set through non-maximum value inhibition, and then carrying out non-maximum value inhibition on the eliminated first license plate candidate frame set
For candidate frames obtained from each layer of input images, firstly, correcting the license plate candidate frames by using a bounding box regression vector, wherein the coordinates of the corrected license plate candidate frames are as follows:
new_x1＝x1+lx*w
new_y1＝y1+ly*h
new_x2＝x1+rx*w
new_y2＝y1+ry*h
the new _ x1, the new _ y1, the new _ x2 and the new _ y2 are coordinates of the upper left corner and the lower right corner of the corrected candidate frame, x1, y1, w and h are coordinates of the upper left corner, width and height of the candidate frame respectively, then the candidate frame with high overlapping degree can be eliminated through non-maximum value inhibition, the license plate candidate frame with the highest neighborhood confidence coefficient is obtained, and finally the non-maximum value inhibition is carried out on all candidate frames obtained by the pyramid image, so that the candidate frame most possibly representing the license plate is obtained.
104. And inputting the second license plate candidate frame set into a second convolutional neural network for processing to obtain a first target candidate frame, a second boundary frame regression vector and a target license plate type.
The second convolutional neural network can be a two-branch network, and Batch Normalization is added for Normalization, the main task can be to perform subdivision on the license plate candidate frame (for example, blue plate, single-layer yellow plate, double-layer yellow plate, black plate, white plate, green plate or non-license plate), further regress the candidate frame according to the license plate type to obtain an accurate license plate area, and character segmentation can be performed on the license plate area by utilizing the license plate type. For example, the obtained license plate candidate frames resize to 56 × 16 are input to the second convolutional neural network, so as to obtain classification confidence coefficients and bounding box regression vectors of 7 types (blue plate, single-layer yellow plate, double-layer yellow plate, black plate, white plate, green plate or non-license plate) of the license plate candidate frames, and the license plate candidate frames are subjected to regression correction and character segmentation according to the types of the license plates.
105. And calibrating the first target candidate frame through the second boundary frame regression vector to obtain a second target candidate frame.
The first target candidate frame may be calibrated by using the second bounding box regression vector, which is beneficial to improving the accuracy of the first target candidate frame.
106. And feeding back the type of the target license plate and the second target candidate frame to the second convolutional neural network for character segmentation to obtain a plurality of characters, and performing character recognition on the plurality of characters to obtain the target license plate.
And performing character recognition on the second target candidate frame according to the type of the target license plate to obtain the target license plate, wherein different types of the target license plate correspond to different character segmentation modes.
Optionally, in the step 106, the target license plate type and the second target candidate frame are fed back to the second convolutional neural network for character segmentation, which may be implemented as follows:
and acquiring a character segmentation mode corresponding to the target license plate type, and feeding back the character segmentation mode and the second target candidate frame to the second convolutional neural network for character segmentation.
The embodiment of the invention can pre-store the mapping relation between the character segmentation mode and the license plate type, further determine the character segmentation mode corresponding to the target license plate type according to the mapping relation, feed back (backtrack) the character segmentation mode corresponding to the target license plate type and the second target candidate frame to the second convolutional neural network for character segmentation to obtain a plurality of characters, and perform character recognition on the plurality of characters to obtain the target license plate. The characters of the license plate can be segmented according to the license plate type (the license plate type corresponding to the maximum classification confidence).
For example, if the license plate is a single-layer license plate of a blue plate, a single-layer yellow plate, a black plate, a white plate or a green plate, the license plate is subjected to character segmentation by directly using a morphological transformation and vertical projection method, and then character recognition is performed; if the license plate is of a double-layer yellow plate type, the license plate is horizontally projected, the double-layer license plate is divided into an upper layer and a lower layer, the upper layer and the lower layer of the double-layer license plate are subjected to left-right equal-height splicing to obtain a single-layer license plate, a single-layer license plate dividing method is used for dividing characters, and finally character recognition is carried out.
For example, as shown in fig. 1a, an image to be processed may be input, where the image to be processed includes a license plate, pyramid transformation may be performed on the image to be processed to obtain a pyramid image, the pyramid image is input to a first convolutional neural network for processing, a boundary frame regression and non-maximum suppression are performed on a processing result, a result is input to a second convolutional neural network to obtain a target license plate type and a candidate frame, the target license plate type and the candidate frame may be fed back to the second convolutional neural network, and a license plate frame regression, character segmentation, and character recognition may be performed on the result to obtain a license plate.
It can be seen that, through the embodiment of the present invention, an image to be processed is obtained, the image to be processed is input to a first convolutional neural network for processing, a first license plate candidate frame set and a first bounding box regression vector are obtained, the first license plate candidate frame set is calibrated through the first bounding box regression vector, the calibrated first license plate candidate frame set is subjected to de-overlapping processing, a second license plate candidate frame set is obtained, the second license plate candidate frame set is input to a second convolutional neural network for processing, a first target candidate frame, a second bounding box regression vector and a target license plate type are obtained, the first target candidate frame is calibrated through the second bounding box regression vector, a second target candidate frame is obtained, the target license plate type and the second target candidate frame are fed back to the second convolutional neural network for character segmentation, a plurality of characters are obtained, and character recognition is performed on the plurality of characters, and obtaining the target license plate, thereby positioning and classifying the license plate based on the secondary convolution neural network and improving the license plate identification precision.
In accordance with the above, please refer to fig. 2, which is a flowchart illustrating a license plate recognition method according to a second embodiment of the present invention. The license plate recognition method described in the embodiment includes the following steps:
201. and acquiring the environmental parameters.
Wherein, the environmental parameters may include but are not limited to: ambient brightness, weather conditions, geographical location, dirt condition of the camera, shooting distance, driving speed, resolution, temperature, air pressure, and the like.
202. And judging whether the environmental parameters meet preset conditions.
The preset condition may be set by the user, for example, the ambient brightness is greater than 50.
203. And when the environmental parameters meet the preset conditions, shooting to obtain a shot image.
When the environmental parameters meet the preset conditions, shooting can be performed to obtain a shot image.
204. And carrying out image enhancement processing on the shot image, and carrying out image segmentation on the shot image after the image enhancement processing to obtain an image to be processed.
The image enhancement processing may be: histogram equalization, gray scale stretching, image denoising, local enhancement, and the like.
205. And inputting the image to be processed into a first convolution neural network for processing to obtain a first license plate candidate frame set and a first boundary frame regression vector.
206. And calibrating the first license plate candidate frame set through the first boundary frame regression vector, and performing de-overlapping processing on the calibrated first license plate candidate frame set to obtain a second license plate candidate frame set.
207. And inputting the second license plate candidate frame set into a second convolutional neural network for processing to obtain a first target candidate frame, a second boundary frame regression vector and a target license plate type.
208. And calibrating the first target candidate frame through the second boundary frame regression vector to obtain a second target candidate frame.
209. And feeding back the type of the target license plate and the second target candidate frame to the second convolutional neural network for character segmentation to obtain a plurality of characters, and performing character recognition on the plurality of characters to obtain the target license plate.
The detailed descriptions of the steps 205 to 209 may refer to the corresponding steps 101 to 106 of the license plate recognition method described in fig. 1, and are not repeated herein.
It can be seen that, according to the embodiment of the present invention, an environmental parameter is obtained, whether the environmental parameter meets a preset condition is determined, when the environmental parameter meets the preset condition, a shot image is obtained by shooting, the shot image is subjected to image enhancement processing, the shot image after the image enhancement processing is subjected to image segmentation to obtain an image to be processed, the image to be processed is input to a first convolution neural network to be processed to obtain a first candidate frame set of a license plate and a first regression vector of a boundary frame, the first candidate frame set of the license plate is calibrated through the first regression vector of the boundary frame, the calibrated first candidate frame set of the license plate is subjected to de-overlap processing to obtain a second candidate frame set of the license plate, the second candidate frame set of the license plate is input to a second convolution neural network to be processed to obtain a first candidate frame, a second regression vector of the boundary frame and a type of the license plate, the first target candidate frame is calibrated through a second boundary frame regression vector to obtain a second target candidate frame, the type of the target license plate and the second target candidate frame are fed back to a second convolutional neural network to perform character segmentation to obtain a plurality of characters, and the characters are recognized to obtain the target license plate.
In accordance with the above, the following is a device for implementing the license plate recognition method, specifically as follows:
fig. 3a is a schematic structural diagram of a license plate recognition device according to an embodiment of the present invention. The license plate recognition device described in this embodiment includes: the first obtaining unit 301, the first processing unit 302, the first calibration unit 303, the second processing unit 304, the second calibration unit 305, and the identification unit 306 are specifically as follows:
a first acquiring unit 301, configured to acquire an image to be processed;
the first processing unit 302 is configured to input the image to be processed into a first convolutional neural network for processing, so as to obtain a first license plate candidate frame set and a first bounding box regression vector;
a first calibration unit 303, configured to calibrate the first candidate frame set through the first bounding box regression vector, and perform de-overlapping processing on the calibrated first candidate frame set to obtain a second candidate frame set;
the second processing unit 304 is configured to input the second candidate license plate frame set to a second convolutional neural network for processing, so as to obtain a first target candidate frame, a second bounding box regression vector, and a target license plate type;
a second calibration unit 305, configured to calibrate the first target candidate box through the second bounding box regression vector to obtain a second target candidate box;
the recognition unit 306 is configured to feed back the type of the target license plate and the second target candidate frame to the second convolutional neural network for character segmentation to obtain a plurality of characters, and perform character recognition on the plurality of characters to obtain the target license plate.
Optionally, as shown in fig. 3b, fig. 3b is a detailed structure of the first processing unit 302 in the license plate recognition device depicted in fig. 3a, where the first processing unit 302 may include: the decomposing module 3021, the input module 3022, and the restoring unit 3023 are as follows:
the decomposition module 3021 is configured to perform multi-scale decomposition on the image to be processed according to the attribute parameters of the license plate to obtain a plurality of sub-images;
the input module 3022 is configured to input each of the multiple sub-images into the first convolutional neural network, so as to obtain multiple thermodynamic diagrams of license plate confidence and a first bounding box regression vector corresponding to the thermodynamic diagrams;
a restoring module 3023, configured to traverse the plurality of thermodynamic diagrams, and perform a restoring operation on a region corresponding to a point in the plurality of thermodynamic diagrams, where the point is greater than a confidence threshold, to obtain the first license plate candidate frame set.
Optionally, in terms of performing de-overlapping processing on the calibrated first license plate candidate frame set, the first calibration unit 303 is specifically configured to:
and eliminating the candidate frames with the overlapping degree larger than the overlapping degree threshold value in the first license plate candidate frame set through non-maximum value inhibition, and then carrying out non-maximum value inhibition on the eliminated first license plate candidate frame set.
Optionally, the identifying unit 306 is specifically configured to:
and acquiring a character segmentation mode corresponding to the target license plate type, and feeding back the character segmentation mode and the second target candidate frame to the second convolutional neural network for character segmentation.
Alternatively, as shown in fig. 3c, fig. 3c is a further modified structure of the license plate recognition device depicted in fig. 3a, and fig. 3c is compared with fig. 3a, and further includes: the second acquiring unit 307, the determining unit 308, the shooting unit 309 and the processing unit 310 are specifically as follows:
a second obtaining unit 307, configured to obtain an environmental parameter;
a determining unit 308, configured to determine whether the environmental parameter meets a preset condition;
a shooting unit 309, configured to shoot when the environmental parameter meets the preset condition, so as to obtain a shot image;
the processing unit 310 is configured to perform image enhancement processing on the captured image, and perform image segmentation on the captured image after the image enhancement processing to obtain the to-be-processed image.
It can be seen that, with the license plate recognition device described in the embodiment of the present invention, an image to be processed is obtained, the image to be processed is input to a first convolutional neural network to be processed, a first candidate frame set of a license plate and a first bounding box regression vector are obtained, the first candidate frame set of the license plate is calibrated with the first bounding box regression vector, the calibrated first candidate frame set of the license plate is subjected to de-overlapping processing, a second candidate frame set of the license plate is obtained, the second candidate frame set of the license plate is input to a second convolutional neural network to be processed, a first target candidate frame, a second bounding box regression vector and a target license plate type are obtained, the first target candidate frame is calibrated with the second bounding box regression vector, a second target candidate frame is obtained, the target license plate type and the second target candidate frame are fed back to the second convolutional neural network for character segmentation, and obtaining a plurality of characters, and performing character recognition on the characters to obtain a target license plate, so that the license plate is positioned and classified based on a secondary convolution neural network, and the license plate recognition precision is improved.
In accordance with the above, please refer to fig. 4, which is a schematic structural diagram of an embodiment of a license plate recognition device according to an embodiment of the present invention. The license plate recognition device described in this embodiment includes: at least one input device 1000; at least one output device 2000; at least one processor 3000, e.g., a CPU; and a memory 4000, the input device 1000, the output device 2000, the processor 3000, and the memory 4000 being connected by a bus 5000.
The input device 1000 may be a touch panel, a physical button, or a mouse.
The output device 2000 may be a display screen.
The memory 4000 may be a high-speed RAM memory or a non-volatile memory (e.g., a magnetic disk memory). The memory 4000 is used for storing a set of program codes, and the input device 1000, the output device 2000 and the processor 3000 are used for calling the program codes stored in the memory 4000 to execute the following operations:
the processor 3000 is configured to:
acquiring an image to be processed;
inputting the image to be processed into a first convolution neural network for processing to obtain a first license plate candidate frame set and a first boundary frame regression vector;
calibrating the first license plate candidate frame set through the first boundary frame regression vector, and performing de-overlapping processing on the calibrated first license plate candidate frame set to obtain a second license plate candidate frame set;
inputting the second license plate candidate frame set into a second convolutional neural network for processing to obtain a first target candidate frame and a second boundary frame regression vector;
calibrating the first target candidate frame through the second bounding box regression vector to obtain a second target candidate frame;
and feeding back the type of the target license plate and the second target candidate frame to the second convolutional neural network for character segmentation to obtain a plurality of characters, and performing character recognition on the plurality of characters to obtain the target license plate.
Optionally, the processor 3000 inputs the image to be processed into a first convolutional neural network for processing, so as to obtain a first candidate frame set of the license plate and a first regression vector of the bounding box, which includes:
performing multi-scale decomposition on the image to be processed according to the attribute parameters of the license plate to obtain a plurality of sub-images;
inputting each sub-image in the plurality of sub-images into the first convolution neural network respectively to obtain a plurality of thermodynamic diagrams of license plate confidence degrees and a first boundary box regression vector corresponding to the thermodynamic diagrams;
and traversing the plurality of thermodynamic diagrams, and performing reduction operation on areas corresponding to points in the plurality of thermodynamic diagrams, wherein the points are larger than a confidence coefficient threshold value, so as to obtain the first license plate candidate frame set.
Optionally, the processor 3000 performs de-overlapping processing on the calibrated first license plate candidate frame set, including:
and eliminating the candidate frames with the overlapping degree larger than the overlapping degree threshold value in the first license plate candidate frame set through non-maximum value inhibition, and then carrying out non-maximum value inhibition on the eliminated first license plate candidate frame set.
Optionally, the processor 3000 feeds back the target license plate type and the second target candidate frame to the second convolutional neural network for character segmentation, where the character segmentation includes:
and acquiring a character segmentation mode corresponding to the target license plate type, and feeding back the character segmentation mode and the second target candidate frame to the second convolutional neural network for character segmentation.
Optionally, the processor 3000 is further specifically configured to:
acquiring an environmental parameter;
judging whether the environmental parameters meet preset conditions or not;
when the environmental parameters meet the preset conditions, shooting to obtain a shot image;
and carrying out image enhancement processing on the shot image, and carrying out image segmentation on the shot image after the image enhancement processing to obtain the image to be processed.
The embodiment of the invention also provides a computer storage medium, wherein the computer storage medium can store a program, and the program comprises part or all of the steps of any license plate recognition method described in the method embodiment when executed.
Embodiments of the present invention further provide a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program, where the computer program is operable to cause a computer to execute some or all of the steps of any of the license plate recognition methods described in the above method embodiments.
While the invention has been described in connection with various embodiments, other variations to the disclosed embodiments can be understood and effected by those skilled in the art in practicing the claimed invention, from a review of the drawings, the disclosure, and the appended claims. In the claims, the word "comprising" does not exclude other elements or steps, and the word "a" or "an" does not exclude a plurality. A single processor or other unit may fulfill the functions of several items recited in the claims. The mere fact that certain measures are recited in mutually different dependent claims does not indicate that a combination of these measures cannot be used to advantage.
As will be appreciated by one skilled in the art, embodiments of the present invention may be provided as a method, apparatus (device), or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein. A computer program stored/distributed on a suitable medium supplied together with or as part of other hardware, may also take other distributed forms, such as via the Internet or other wired or wireless telecommunication systems.
The present invention has been described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (devices) and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable license plate location device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable license plate location device, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable license plate location device to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable license plate location device to cause a series of operational steps to be performed on the computer or other programmable device to produce a computer implemented process such that the instructions which execute on the computer or other programmable device provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
While the invention has been described in conjunction with specific features and embodiments thereof, it will be evident that various modifications and combinations can be made thereto without departing from the spirit and scope of the invention. Accordingly, the specification and figures are merely exemplary of the invention as defined in the appended claims and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of the invention. It will be apparent to those skilled in the art that various changes and modifications may be made in the present invention without departing from the spirit and scope of the invention. Thus, if such modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include such modifications and variations.
Claims (8)
1. A license plate recognition method is characterized by comprising the following steps:
acquiring an image to be processed;
inputting the image to be processed into a first convolution neural network for processing to obtain a first license plate candidate frame set and a first boundary frame regression vector, wherein the first convolution neural network is a full convolution neural network consisting of 4 convolution layers and has the task of performing window sliding on the input image, performing coarse classification on a window sliding area and performing regression on the license plate candidate frame;
calibrating the first license plate candidate frame set through the first boundary frame regression vector, and performing de-overlapping processing on the calibrated first license plate candidate frame set to obtain a second license plate candidate frame set;
inputting the second license plate candidate frame set into a second convolutional neural network for processing to obtain a first target candidate frame, a second boundary frame regression vector and a target license plate type, wherein the second convolutional neural network is a two-branch network, is added with Batch Normalization for Normalization, and has the task of finely classifying the license plate candidate frames, further regressing the candidate frames according to the license plate types to obtain an accurate license plate area, and performing character segmentation on the license plate area by utilizing the license plate types;
calibrating the first target candidate frame through the second bounding box regression vector to obtain a second target candidate frame;
feeding the target license plate type and the second target candidate frame back to the second convolutional neural network for character segmentation to obtain a plurality of characters, and performing character recognition on the plurality of characters to obtain a target license plate;
feeding the target license plate type and the second target candidate frame back to the second convolutional neural network for character segmentation, wherein the character segmentation comprises the following steps:
and acquiring a character segmentation mode corresponding to the target license plate type, and feeding back the character segmentation mode and the second target candidate frame to the second convolutional neural network for character segmentation.
2. The method of claim 1, wherein inputting the image to be processed into a first convolutional neural network for processing to obtain a first candidate box set of the license plate and a first regression vector of the bounding box, comprises:
performing multi-scale decomposition on the image to be processed according to the attribute parameters of the license plate to obtain a plurality of sub-images;
inputting each sub-image in the plurality of sub-images into the first convolution neural network respectively to obtain a plurality of thermodynamic diagrams of license plate confidence degrees and a first boundary box regression vector corresponding to the thermodynamic diagrams;
and traversing the plurality of thermodynamic diagrams, and performing reduction operation on areas corresponding to points in the plurality of thermodynamic diagrams, wherein the points are larger than a confidence coefficient threshold value, so as to obtain the first license plate candidate frame set.
3. The method of claim 1 or 2, wherein the de-overlapping the calibrated first set of license plate candidate frames comprises:
and eliminating the candidate frames with the overlapping degree larger than the overlapping degree threshold value in the first license plate candidate frame set through non-maximum value inhibition, and then carrying out non-maximum value inhibition on the eliminated first license plate candidate frame set.
4. The method according to claim 1 or 2, characterized in that the method further comprises:
acquiring an environmental parameter;
judging whether the environmental parameters meet preset conditions or not;
when the environmental parameters meet the preset conditions, shooting to obtain a shot image;
and carrying out image enhancement processing on the shot image, and carrying out image segmentation on the shot image after the image enhancement processing to obtain the image to be processed.
5. A license plate recognition device, comprising:
the first acquisition unit is used for acquiring an image to be processed;
the first processing unit is used for inputting the image to be processed into a first convolution neural network for processing to obtain a first license plate candidate frame set and a first boundary frame regression vector, wherein the first convolution neural network is a full convolution neural network consisting of 4 convolution layers and has the task of performing sliding window on the input image, performing coarse classification on a sliding window area and performing regression on the license plate candidate frame;
the first calibration unit is used for calibrating the first license plate candidate frame set through the first boundary frame regression vector and performing de-overlapping processing on the calibrated first license plate candidate frame set to obtain a second license plate candidate frame set;
the second processing unit is used for inputting the second license plate candidate frame set into a second convolutional neural network for processing to obtain a first target candidate frame, a second boundary frame regression vector and a target license plate type, wherein the second convolutional neural network is a two-branch network, is added with Batch Normalization for Normalization processing, and has the tasks of finely classifying the license plate candidate frames, further regressing the candidate frames according to the license plate types to obtain an accurate license plate area, and performing character segmentation on the license plate area by utilizing the license plate types;
the second calibration unit is used for calibrating the first target candidate frame through the second boundary frame regression vector to obtain a second target candidate frame;
the recognition unit is used for feeding the target license plate type and the second target candidate frame back to the second convolutional neural network for character segmentation to obtain a plurality of characters, and performing character recognition on the plurality of characters to obtain a target license plate;
wherein, in the aspect of feeding back the target license plate type and the second target candidate frame to the second convolutional neural network for character segmentation, the identification unit is specifically configured to:
and acquiring a character segmentation mode corresponding to the target license plate type, and feeding back the character segmentation mode and the second target candidate frame to the second convolutional neural network for character segmentation.
6. The apparatus of claim 5, wherein the first processing unit comprises:
the decomposition module is used for carrying out multi-scale decomposition on the image to be processed according to the attribute parameters of the license plate to obtain a plurality of sub-images;
the input module is used for respectively inputting each sub-image in the plurality of sub-images into the first convolution neural network to obtain a plurality of thermodynamic diagrams of the confidence degree of the license plate and a first boundary box regression vector corresponding to the thermodynamic diagrams;
and the restoring module is used for traversing the plurality of thermodynamic diagrams and carrying out restoring operation on the areas corresponding to the points which are greater than the confidence coefficient threshold value in the plurality of thermodynamic diagrams to obtain the first license plate candidate frame set.
7. The apparatus according to claim 5 or 6, wherein, in the de-overlapping the calibrated first set of license plate candidate frames, the first calibration unit is specifically configured to:
and eliminating the candidate frames with the overlapping degree larger than the overlapping degree threshold value in the first license plate candidate frame set through non-maximum value inhibition, and then carrying out non-maximum value inhibition on the eliminated first license plate candidate frame set.
8. The apparatus of claim 5 or 6, further comprising:
a second obtaining unit, configured to obtain an environmental parameter;
the judging unit is used for judging whether the environmental parameters meet preset conditions or not;
the shooting unit is used for shooting when the environmental parameters meet the preset conditions to obtain a shot image;
and the processing unit is used for carrying out image enhancement processing on the shot image and carrying out image segmentation on the shot image after the image enhancement processing to obtain the image to be processed.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711223109.1A CN108009543B (en) | 2017-11-29 | 2017-11-29 | License plate recognition method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711223109.1A CN108009543B (en) | 2017-11-29 | 2017-11-29 | License plate recognition method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108009543A CN108009543A (en) | 2018-05-08 |
| CN108009543B true CN108009543B (en) | 2020-05-26 |
Family
ID=62054447
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711223109.1A Active CN108009543B (en) | 2017-11-29 | 2017-11-29 | License plate recognition method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108009543B (en) |
Families Citing this family (33)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110490186B (en) * | 2018-05-15 | 2022-03-25 | 杭州海康威视数字技术股份有限公司 | License plate recognition method and device and storage medium |
| CN110490179B (en) * | 2018-05-15 | 2022-08-05 | 杭州海康威视数字技术股份有限公司 | License plate recognition method and device and storage medium |
| 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 |
| CN110543882A (en) * | 2018-05-29 | 2019-12-06 | 北京深鉴智能科技有限公司 | Character string recognition method and device |
| US10726555B2 (en) * | 2018-06-06 | 2020-07-28 | International Business Machines Corporation | Joint registration and segmentation of images using deep learning |
| CN110827558B (en) * | 2018-06-19 | 2020-12-22 | 嘉兴觅特电子商务有限公司 | Traffic signal lamp recognition device based on vehicle speed, height of highlight area and image |
| CN110659643B (en) * | 2018-06-29 | 2022-08-26 | 杭州海康威视数字技术股份有限公司 | Container number identification method and device |
| CN110738225B (en) * | 2018-07-19 | 2023-01-24 | 杭州海康威视数字技术股份有限公司 | Image recognition method and device |
| CN109190687A (en) * | 2018-08-16 | 2019-01-11 | 新智数字科技有限公司 | A kind of nerve network system and its method for identifying vehicle attribute |
| CN109255350B (en) * | 2018-08-29 | 2021-12-10 | 南京邮电大学 | New energy license plate detection method based on video monitoring |
| CN109214386B (en) * | 2018-09-14 | 2020-11-24 | 京东数字科技控股有限公司 | Method and apparatus for generating image recognition model |
| CN110942073A (en) * | 2018-09-25 | 2020-03-31 | 杭州海康威视数字技术股份有限公司 | Container trailer number identification method and device and computer equipment |
| CN110956169B (en) * | 2018-09-27 | 2024-09-06 | 杭州海康威视数字技术股份有限公司 | License plate recognition method and device and electronic equipment |
| CN110163199A (en) * | 2018-09-30 | 2019-08-23 | 腾讯科技（深圳）有限公司 | Licence plate recognition method, license plate recognition device, car license recognition equipment and medium |
| CN111027555B (en) * | 2018-10-09 | 2023-09-26 | 杭州海康威视数字技术股份有限公司 | License plate recognition method and device and electronic equipment |
| CN111027534B (en) * | 2018-10-09 | 2023-05-26 | 杭州海康威视数字技术股份有限公司 | Compact double license plate detection method and device |
| CN109784327B (en) * | 2018-12-04 | 2021-04-16 | 北京达佳互联信息技术有限公司 | Boundary box determining method and device, electronic equipment and storage medium |
| CN109637153A (en) * | 2019-01-25 | 2019-04-16 | 合肥市智信汽车科技有限公司 | A kind of vehicle-mounted mobile violation snap-shooting system based on machine vision |
| SG10201905273VA (en) * | 2019-06-10 | 2019-08-27 | Alibaba Group Holding Ltd | Method and system for evaluating an object detection model |
| CN110717493B (en) * | 2019-09-16 | 2022-04-01 | 浙江大学 | License plate recognition method containing stacked characters based on deep learning |
| CN112651417B (en) * | 2019-10-12 | 2024-06-14 | 杭州海康威视数字技术股份有限公司 | License plate recognition method, device, equipment and storage medium |
| CN110766017B (en) * | 2019-10-22 | 2023-08-04 | 国网新疆电力有限公司信息通信公司 | Mobile terminal text recognition method and system based on deep learning |
| CN110807459B (en) * | 2019-10-31 | 2022-06-17 | 深圳市捷顺科技实业股份有限公司 | License plate correction method and device and readable storage medium |
| CN110796144B (en) * | 2019-11-08 | 2023-04-25 | 北京字节跳动网络技术有限公司 | License plate detection method, device, equipment and storage medium |
| CN111046923B (en) * | 2019-11-26 | 2023-02-28 | 佛山科学技术学院 | Image target detection method and device based on bounding box and storage medium |
| CN111126383A (en) * | 2019-12-06 | 2020-05-08 | 广州烽火众智数字技术有限公司 | License plate detection method, system, device and storage medium |
| CN111444911B (en) * | 2019-12-13 | 2021-02-26 | 珠海大横琴科技发展有限公司 | Training method and device of license plate recognition model and license plate recognition method and device |
| CN110929727A (en) * | 2020-02-12 | 2020-03-27 | 成都数联铭品科技有限公司 | Image labeling method and device, character detection method and system and electronic equipment |
| CN111401360B (en) * | 2020-03-02 | 2023-06-20 | 杭州雄迈集成电路技术股份有限公司 | Method and system for optimizing license plate detection model, license plate detection method and system |
| CN111310850B (en) * | 2020-03-02 | 2023-06-16 | 杭州雄迈集成电路技术股份有限公司 | License plate detection model construction method and system, license plate detection method and system |
| CN111914845A (en) * | 2020-07-02 | 2020-11-10 | 浙江大华技术股份有限公司 | Character layering method and device in license plate and electronic equipment |
| CN112560856B (en) * | 2020-12-18 | 2024-04-12 | 深圳赛安特技术服务有限公司 | License plate detection and identification method, device, equipment and storage medium |
| CN113627526B (en) * | 2021-08-10 | 2023-07-25 | 北京百度网讯科技有限公司 | Vehicle identification recognition method and device, electronic equipment and medium |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106650740A (en) * | 2016-12-15 | 2017-05-10 | 深圳市华尊科技股份有限公司 | License plate identification method and terminal |
| CN106845487A (en) * | 2016-12-30 | 2017-06-13 | 佳都新太科技股份有限公司 | A kind of licence plate recognition method end to end |
-
2017
- 2017-11-29 CN CN201711223109.1A patent/CN108009543B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106650740A (en) * | 2016-12-15 | 2017-05-10 | 深圳市华尊科技股份有限公司 | License plate identification method and terminal |
| CN106845487A (en) * | 2016-12-30 | 2017-06-13 | 佳都新太科技股份有限公司 | A kind of licence plate recognition method end to end |
Non-Patent Citations (2)
| Title |
|---|
| Joint Face Detection and Alignment using Multi-task Cascaded Convolutional Networks;Kaipeng Zhang et al.;《IEEE》;20151231;第1-5页 * |
| 基于SVM和排列模型的白色车牌分类;张旭光 等;《计算机工程与应用》;20101231;第207-210页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108009543A (en) | 2018-05-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108009543B (en) | License plate recognition method and device | |
| CN109815997B (en) | Method and related device for identifying vehicle damage based on deep learning | |
| CN110414507B (en) | License plate recognition method and device, computer equipment and storage medium | |
| CN109657533B (en) | Pedestrian re-identification method and related product | |
| CN110738125B (en) | Method, device and storage medium for selecting detection frame by Mask R-CNN | |
| CN107944450B (en) | License plate recognition method and device | |
| CN108960211B (en) | Multi-target human body posture detection method and system | |
| US8792722B2 (en) | Hand gesture detection | |
| US8750573B2 (en) | Hand gesture detection | |
| CN111027504A (en) | Face key point detection method, device, equipment and storage medium | |
| WO2020134528A1 (en) | Target detection method and related product | |
| CN106971185B (en) | License plate positioning method and device based on full convolution network | |
| CN108154149B (en) | License plate recognition method based on deep learning network sharing | |
| CN112560698B (en) | Image processing method, device, equipment and medium | |
| CN111310826B (en) | Method and device for detecting labeling abnormality of sample set and electronic equipment | |
| CN110910445B (en) | Object size detection method, device, detection equipment and storage medium | |
| CN112257692A (en) | Pedestrian target detection method, electronic device and storage medium | |
| CN112997190A (en) | License plate recognition method and device and electronic equipment | |
| CN111127358B (en) | Image processing method, device and storage medium | |
| CN112991349B (en) | Image processing method, device, equipment and storage medium | |
| CN111709377B (en) | Feature extraction method, target re-identification method and device and electronic equipment | |
| CN112926426A (en) | Ship identification method, system, equipment and storage medium based on monitoring video | |
| CN111738272B (en) | Target feature extraction method and device and electronic equipment | |
| WO2024174726A1 (en) | Handwritten and printed text detection method and device based on deep learning | |
| CN110826364B (en) | Library position identification method and device |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |