CN117132969A - Recognition method, device, computer and storage medium based on multinational license plate - Google Patents
Recognition method, device, computer and storage medium based on multinational license plate Download PDFInfo
- Publication number
- CN117132969A CN117132969A CN202210783601.9A CN202210783601A CN117132969A CN 117132969 A CN117132969 A CN 117132969A CN 202210783601 A CN202210783601 A CN 202210783601A CN 117132969 A CN117132969 A CN 117132969A
- Authority
- CN
- China
- Prior art keywords
- license plate
- character
- network
- recognition
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 37
- 238000001514 detection method Methods 0.000 claims abstract description 79
- 230000006870 function Effects 0.000 claims description 42
- 239000010410 layer Substances 0.000 claims description 26
- 238000004590 computer program Methods 0.000 claims description 8
- PXFBZOLANLWPMH-UHFFFAOYSA-N 16-Epiaffinine Natural products C1C(C2=CC=CC=C2N2)=C2C(=O)CC2C(=CC)CN(C)C1C2CO PXFBZOLANLWPMH-UHFFFAOYSA-N 0.000 claims description 3
- 230000002146 bilateral effect Effects 0.000 claims description 3
- 239000002356 single layer Substances 0.000 claims description 3
- 230000009466 transformation Effects 0.000 claims description 3
- 238000005516 engineering process Methods 0.000 abstract description 4
- 238000013135 deep learning Methods 0.000 description 4
- 238000012545 processing Methods 0.000 description 4
- 230000009286 beneficial effect Effects 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 2
- 238000003702 image correction Methods 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 238000005070 sampling Methods 0.000 description 2
- 239000013598 vector Substances 0.000 description 2
- 238000003491 array Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000001413 cellular effect Effects 0.000 description 1
- 230000001788 irregular Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 238000011176 pooling Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 238000012549 training Methods 0.000 description 1
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
- G06V20/625—License plates
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/22—Image preprocessing by selection of a specific region containing or referencing a pattern; Locating or processing of specific regions to guide the detection or recognition
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/24—Aligning, centring, orientation detection or correction of the image
- G06V10/247—Aligning, centring, orientation detection or correction of the image by affine transforms, e.g. correction due to perspective effects; Quadrilaterals, e.g. trapezoids
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
-
- 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
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/07—Target detection
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Computing Systems (AREA)
- Databases & Information Systems (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Character Discrimination (AREA)
Abstract
The invention discloses a recognition method, a device, a computer and a storage medium based on multinational license plates, which relate to an image recognition technology and comprise the following steps: acquiring an image to be identified; positioning a license plate in the image to be identified by utilizing a target detection network to obtain license plate frame coordinates and license plate images in the license plate frame coordinates; inputting the license plate image into a license plate multi-attribute classification network to obtain preset type attribute information corresponding to a license plate; inputting the license plate image into a character detection and recognition network to obtain character information in a license plate; inputting the attribute information of the preset type and the character information into a preset template function for matching; if the matching is successful, the identification result of the license plate is output. The invention can simultaneously support the recognition of the multinational license plates and keep better recognition speed and recognition accuracy.
Description
Technical Field
The present invention relates to the field of image recognition, and in particular, to a recognition method, apparatus, computer and storage medium based on a multi-country license plate.
Background
The license plate recognition technology is one of important components in a modern intelligent traffic system, and is based on technologies such as graphic processing, pattern recognition, computer vision, deep learning algorithm and the like, and is used for analyzing a vehicle image or video sequence shot by a camera to obtain a unique license plate number of each vehicle. At present, license plate recognition technology is mostly applied to some management systems, such as parking lot charge management, road violation snapshot, high-speed function automatic supervision, port traffic records and the like, and has important significance for maintaining traffic safety and urban public security and realizing traffic automatic management.
Most of the existing license plate recognition algorithms are based on domestic license plates, the domestic license plate system is relatively standard, and foreign license plates have the characteristics of irregular license plates, more types and the like, so that the license plate recognition difficulty is high, and furthermore, the situation that multiple national license plates pass in the same area exists in the middle east from the trade area, and correspondingly, the recognition of the multiple national license plates needs to be supported simultaneously.
There are few techniques for identifying multiple national license plates simultaneously. In recent years, license plate recognition algorithms based on deep learning are emerging, and the method has the characteristics of relatively high recognition precision, low false recognition rate, simple flow, no need of character segmentation and the like. However, the existing license plate recognition algorithm based on deep learning has poor effect when the license plate recognition algorithm is used for small targets, multi-layer license plates and needs to support more license plate types.
Disclosure of Invention
The invention provides a multi-country license plate based recognition method, a multi-country license plate based recognition device, a multi-country license plate based recognition computer and a multi-country license plate based storage medium, wherein the multi-country license plate can be recognized at the same time by the recognition method, and the multi-country license plate based recognition method is suitable for recognizing small targets, multi-layer license plates and a plurality of kinds of license plates.
Aiming at the problem that the prior art is difficult to support the simultaneous identification of the multinational license plates, the invention provides a multinational license plate-based identification method, a multinational license plate-based identification device, a multinational license plate-based identification computer and a multinational license plate-based storage medium.
The technical scheme provided by the invention for the technical problems is as follows:
the invention provides a recognition method based on multinational license plates, which comprises the following steps:
acquiring an image to be identified;
positioning a license plate in the image to be identified by utilizing a target detection network to obtain license plate frame coordinates and license plate images in the license plate frame coordinates;
inputting the license plate image into a license plate multi-attribute classification network to obtain preset type attribute information corresponding to the license plate, wherein the preset type attribute information at least comprises license plate country category and license plate color;
inputting the license plate image into a character detection and recognition network to obtain character information in a license plate, wherein the character information comprises character coordinates and character category information;
inputting the attribute information of the preset type and the character information into a preset template function for matching;
if the matching is successful, the identification result of the license plate is output.
Preferably, the method further comprises:
acquiring license plate corner points of license plates in the license plate frame coordinates by utilizing the target detection network;
correcting the license plate image by utilizing the license plate corner points to obtain a license plate screenshot;
the inputting the license plate image into a license plate multi-attribute classification network to obtain the preset type attribute information corresponding to the license plate comprises the following steps: inputting the license plate screenshot into a license plate multi-attribute classification network to obtain preset type attribute information corresponding to the license plate;
the inputting the license plate image into a character detection and recognition network to obtain character information in the license plate comprises the following steps: inputting the license plate screenshot into a character detection and recognition network to obtain all character coordinates and character information in the license plate.
Preferably, the correcting the license plate image by using the license plate corner point to obtain a license plate screenshot includes:
correcting the license plate image through affine transformation by utilizing the license plate corner points to obtain a license plate screenshot.
Preferably, the method further comprises:
and defining the country category and the corresponding license plate character structure to obtain the preset template function.
Preferably, the target detection network adopts a YOLOv5 detection network added with key point regression;
the license plate multi-attribute classification network adopts a resnet18 network, and output layers of the resnet18 network are a country category attribute output layer and a color attribute output layer respectively.
Preferably, the classification mode of the character class information at least comprises one or more of the following:
performing character detection by using the character detection network to obtain a character recognition result, and classifying license plates into Chinese characters and Chinese character-free classes according to whether the character recognition result has Chinese characters or not;
performing character detection by using the character detection network to obtain a character recognition result, and dividing the license plate into Arabic number letters and Arabic number letter-free types according to whether Arabic number letters exist in the character recognition result;
performing character detection by using the character detection network to obtain a character recognition result, and dividing the license plate into Arabic numeral letters and Arabic numeral letters according to whether Arabic numeral letters exist in the character recognition result;
performing character detection by using the character detection network to obtain a character recognition result, and dividing the license plate into a single-layer structure and a double-layer structure according to whether the character result is of a double-layer structure;
performing character detection by using the character detection network to obtain a character recognition result, and dividing the license plate into a symmetrical structure and an asymmetrical structure according to whether the character result is a bilateral symmetry structure or not;
and carrying out character detection by utilizing the character detection network to obtain a character recognition result, and dividing the license plate into a character string with a country sign and a character string without a country sign according to whether the character recognition result has the character string with the country sign.
Preferably, after the inputting the preset type attribute information and the character information into a preset template function for matching, the method further includes:
if the matching is unsuccessful, storing the identification result of the current license plate, and defining the identification result to be used as a new template function to be newly added into a template function library.
The invention provides a recognition device based on a multinational license plate, which comprises:
the acquisition module is used for acquiring an image to be identified;
the detection module is used for positioning the license plate in the image to be identified by utilizing a target detection network so as to obtain license plate frame coordinates and license plate images in the license plate frame coordinates;
the attribute identification module is used for inputting the license plate image into a license plate multi-attribute classification network to obtain preset type attribute information corresponding to the license plate, wherein the preset type attribute information at least comprises license plate country category and license plate color;
the character recognition module is used for inputting the license plate image into a character detection recognition network through the character recognition module so as to acquire character information in the license plate, wherein the character information comprises character coordinates and character category information;
the matching module is used for inputting the attribute information of the preset type and the character information into a preset template function to match;
and the output module outputs the identification result of the license plate when the matching is successful.
The invention also provides a computer device comprising a processor for implementing the steps of the multi-country license plate based identification method as described above when executing a computer program stored in a memory.
The present invention also provides a storage medium having stored thereon a computer program which, when executed by a processor, performs the steps of the multi-country license plate based identification method as described above.
The technical scheme provided by the embodiment of the invention has the beneficial effects that:
the recognition method based on the multinational license plate provided by the invention can simultaneously support the recognition of the multinational license plate by acquiring the license plate image part in the image, carrying out multiattribute recognition and character recognition on the part, inputting corresponding recognition information into a preset template function for matching, and outputting corresponding recognition results when the matching is successful, is suitable for recognizing small targets, multi-layer license plates and a plurality of license plates, and keeps better recognition speed and recognition precision. In addition, the multi-attribute-based recognition network is beneficial to analyzing license plates from multiple dimensions and utilizing recognized attribute data to carry out post expansion.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly described below, and it is apparent that the drawings in the following description are only some embodiments of the present invention, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a flowchart of a multi-country license plate-based recognition method provided by the invention.
Fig. 2 is a schematic diagram of a license plate outside-cut rectangular frame and a license plate inside the rectangular frame, which are obtained by detecting an image by using a target detection network.
Fig. 3 is a functional block diagram of the recognition device based on the multi-country license plate provided by the invention.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the present invention more apparent, the embodiments of the present invention will be described in further detail with reference to the accompanying drawings.
Referring to fig. 1, a flowchart of a multi-country license plate-based identification method provided by the invention is provided. The recognition method based on the multinational license plates is mainly applied to a traffic recognition system, can simultaneously support the recognition of the multinational license plates, has better recognition accuracy and recognition speed, and is suitable for recognizing small targets, multi-layer license plates and license plates of more varieties.
As shown in fig. 1, the multi-country license plate based identification method may include the following steps:
s101: the image to be identified is acquired, and the image to be identified may be a video and/or a picture stored in a specific memory, or may be a video and/or a picture acquired in real time, which is not limited herein.
S102: and positioning the license plate in the image to be identified by utilizing a target detection network to obtain license plate frame coordinates and license plate images in the license plate frame coordinates.
In this step, the target detection network may be a YOLOv5 detection network with increased keypoint regression, and the license plate frame coordinate may be, but is not limited to, a license plate outside-cut rectangular frame coordinate. Specifically, the YOLOv5 detection network of the present embodiment modifies the number of output channels of the predicted output layer. Further, the present invention may be, but is not limited to, using YOLOv5 detection networks, the improved method herein being applicable to all detection type network models.
As shown in fig. 2, the license plate outer section rectangular frame 11 is inscribed with four vertices of the license plate 12, and the four vertices can be used as key points of the image of the license plate 12. In this step, the target detection network may be further used to obtain license plate corner points of the license plate 12 in the coordinates of the rectangular frame 11, where the license plate corner points are four vertices of the license plate 12. And correcting the license plate image by utilizing the license plate corner points to obtain a license plate screenshot.
Here, the license plate corner point can be utilized to correct the license plate image through affine transformation so as to obtain a license plate screenshot.
It will be appreciated that the license plate 12 illustrated in fig. 2 is an image of a portion of a picture, and that the license plate 12 is tilted in the image. In the actual captured picture/video, the license plate 12 may also be in any other pose, but should be limited to an image that is capable of fully displaying the structure of the license plate.
The invention discloses a YOLOv5 detection network structure, which comprises an input end, a backup network unit, a network unit and a prediction network unit, wherein 640 multiplied by 3 images are input into a backbone network from the input end, the backbone network comprises a focus module, a CBL module, a CSP module and an SPP module, the input images are sliced through the focus module to obtain feature images, and then the CBL module and the CSP module are continuously utilized to obtain more features, and the SPP module is used for pooling operation to obtain feature vectors with fixed sizes, and the feature vectors are input into the network unit for feature processing and are transmitted to the prediction network unit for prediction.
Here, it should be noted that, in the conventional YOLOv5 detection network element in the network structure, the output channel count rule is no=na× (nc+5). Setting 3 anchors for each scale output layer, and setting na to be 3; when the output class of the original network is 1 (only 1 class is needed for license plate positioning), nc is 1; and 5 represents the confidence of the coordinates x, y, w, h and box of the rectangular frame of the predicted license plate. After the invention is modified for the prediction network element, na is kept unchanged and is still 3; the number of key points (4 categories and x, y coordinates of each category) is increased by 4, and the upper left corner, the upper right corner, the lower right corner and the lower left corner are sequentially increased, so nc=1+4+4×2=13.
S103, inputting the license plate image into a license plate multi-attribute classification network to obtain preset type attribute information corresponding to the license plate, wherein the preset type attribute information at least comprises license plate country category and license plate color.
Corresponding to the license plate screenshot obtained by adopting license plate corner points and image correction, the step can be further that the license plate screenshot is input into a license plate multi-attribute classification network to obtain preset type attribute information corresponding to the license plate.
In this step, the license plate multi-attribute classification network adopts a resnet18 network, wherein the output layers of the resnet18 network are a country category attribute output layer and a color attribute output layer, and multiple attributes of the license plate, such as blue, white, yellow, green, black, red and other color attribute information and salet, egypt, jordan and other country category attribute information, can be obtained simultaneously through the network.
S104: inputting the license plate image into a character detection and recognition network to obtain character information in the license plate, wherein the character information comprises character coordinates and character category information.
Corresponding to the license plate screenshot obtained by adopting the license plate corner point and the image correction, the step can be that the license plate screenshot is input into a character detection and recognition network to obtain all character coordinates and character information in the license plate.
In this step, the character detection and recognition network can obtain the position information of all the characters and the category information of the characters contained in the license plate.
The classification mode of the character class information at least comprises one or more of the following:
performing character detection by using the character detection network to obtain a character recognition result, and classifying license plates into Chinese characters and Chinese character-free classes according to whether the character recognition result has Chinese characters or not;
performing character detection by using the character detection network to obtain a character recognition result, and dividing the license plate into Arabic number letters and Arabic number letter-free types according to whether Arabic number letters exist in the character recognition result;
performing character detection by using the character detection network to obtain a character recognition result, and dividing the license plate into Arabic numeral letters and Arabic numeral letters according to whether Arabic numeral letters exist in the character recognition result;
performing character detection by using the character detection network to obtain a character recognition result, and dividing the license plate into a single-layer structure and a double-layer structure according to whether the character result is of a double-layer structure;
performing character detection by using the character detection network to obtain a character recognition result, and dividing the license plate into a symmetrical structure and an asymmetrical structure according to whether the character result is a bilateral symmetry structure or not;
and carrying out character detection by utilizing the character detection network to obtain a character recognition result, and dividing the license plate into a character string with a country sign and a character string without a country sign according to whether the character recognition result has the character string with the country sign.
S105: and inputting the attribute information of the preset type and the character information into a preset template function for matching.
In this step, the preset template function may be defined according to country types and corresponding license plate character structures, that is, license plates of respective countries are obtained in advance, the country types and corresponding license plate character structures are determined, and then the country types and corresponding license plate character structures are defined to obtain the preset template function. It will be appreciated that the information capable of characterizing the country category may be embodied as a combination of two or three of color, character, logo, or the like, or any of them.
When the step is specifically executed, a plurality of template functions are defined according to the country category and the license plate character structure. And traversing all the template functions, and inputting the obtained preset type attribute information and character information into the template functions. Finally, if the definition of the template function is met, the ordered character strings can be obtained, the ordered character strings are used as specific license plate recognition results, and the traversal is stopped.
It will be appreciated that if the definition of the template function is not satisfied, then the next template function is continued to be traversed, and so on. Here, the template function is preset as a function satisfying rules of a certain character class, an order of characters, a country class, and the like.
S106: if the matching is successful, the identification result of the license plate is output.
In this step, the recognition result may include the country category of the license plate and corresponding character information, or may be information that the recognition is passed or failed.
Notably, step S106 includes the sub-steps of: if the matching is unsuccessful, sampling the current recognition result and storing the current recognition result into a template function library. That is, when the matching is unsuccessful, the recognition result of the current license plate can be stored, and the recognition result is defined to be newly added into the template function library as a new template function, so that the expansion optimization of the template function library is realized.
The recognition method based on the multinational license plates provided by the invention can simultaneously support the recognition of the multinational license plates, maintain good recognition speed and recognition accuracy and is suitable for recognizing small targets, multi-layer license plates and a plurality of license plates by acquiring the license plate image part in the image, carrying out multi-attribute recognition and character recognition on the part, inputting corresponding recognition information into a preset template function for matching, and outputting corresponding recognition results when the matching is successful. In addition, the multi-attribute-based recognition network is beneficial to analyzing license plates from multiple dimensions and utilizing recognized attribute data to carry out post expansion.
The platform for realizing the recognition method based on the multinational license plate can be implemented as any embedded hardware platform or computer platform with neural network reasoning capability, an operating system can be but not limited to linux or windows, and a model training framework can be but not limited to a pytorch deep learning framework, and is not limited herein.
Referring to fig. 3, a schematic diagram of a functional module of the multi-country license plate recognition device provided by the invention is provided. The recognition device 200 based on the multi-country license plate comprises an acquisition module 21, a detection module 22, an attribute recognition module 23, a character recognition module 24, a matching module 25 and an output module 26, wherein:
the acquiring module 21 is mainly configured to acquire an image to be identified, where the image to be identified may be a video and/or a picture stored in a specific memory, or may be a video and/or a picture acquired in real time, which is not limited herein.
The detection module 22 is mainly used for positioning the license plate in the image to be identified by utilizing a target detection network to obtain a license plate frame coordinate and a license plate image in the license plate frame coordinate. The target detection network may be a YOLOv5 detection network with increased key point regression, and the license plate frame coordinate may be, but is not limited to, a license plate outside-cut rectangular frame coordinate. Specifically, the YOLOv5 detection network of the present embodiment modifies the number of output channels of the predicted output layer.
The attribute identification module 23 is mainly configured to input the license plate image into a license plate multi-attribute classification network to obtain preset type attribute information corresponding to a license plate, where the preset type attribute information at least includes a license plate country category and a license plate color. The license plate multi-attribute classification network adopts a resnet18 network, wherein the output layers of the resnet18 network are a country category attribute output layer and a color attribute output layer respectively, and a plurality of attributes of the license plate, such as blue, white, yellow, green, black, red and other color attribute information and salet, egypt, yodan and other country category attribute information, can be obtained simultaneously through the network.
The character recognition module 24 is mainly used for inputting the license plate image into a character detection recognition network to obtain character information in the license plate, wherein the character information comprises character coordinates and character category information. And acquiring the position information of all the characters and the category information of the characters contained in the license plate through the character detection and recognition network.
The matching module 25 is mainly configured to input the preset type attribute information and the character information into a preset template function for matching. If the definition of the template function is not satisfied, the next template function is traversed continuously, and so on. Here, the template function is preset as a function satisfying rules of a certain character class, an order of characters, a country class, and the like.
The output module 26 is mainly used for outputting the recognition result of the license plate when the matching is successful. If the matching is unsuccessful, sampling the current recognition result and storing the current recognition result into a template function library. That is, when the matching is unsuccessful, the recognition result of the current license plate can be stored, and the recognition result is defined to be newly added into the template function library as a new template function, so that the expansion optimization of the template function library is realized.
It should be understood that the effect achieved after the corresponding module performs the corresponding function may be the same as the image searching method described above, and thus will not be described herein.
Furthermore, the present invention provides a computer device comprising a processor for implementing the steps in the above-described multi-country license plate based identification method when executing a computer program stored in a memory.
The processor may be a central processing unit (Central Processing Unit, CPU), but may also be other general purpose processors, digital signal processors (Digital Signal Processor, DSP), application specific integrated circuits (Application Specific Integrated Circuit, ASIC), off-the-shelf programmable gate arrays (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or the like. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like that is a control center of the computer device, connecting various parts of the overall computer device using various interfaces and lines.
The memory may be used to store the computer program and/or modules, and the processor may implement various functions of the computer device by running or executing the computer program and/or modules stored in the memory, and invoking data stored in the memory. The memory may mainly include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required for at least one function, and the like; the storage data area may store data created according to the use of the cellular phone, etc. In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as a hard disk, memory, plug-in hard disk, smart Media Card (SMC), secure Digital (SD) Card, flash Card (Flash Card), at least one disk storage device, flash memory device, or other volatile solid-state storage device.
The present invention also provides a storage medium having stored thereon a computer program which, when executed by a processor, performs the steps of the multi-country license plate based identification method described above.
The foregoing description of the preferred embodiments of the invention is not intended to limit the invention to the precise form disclosed, and any such modifications, equivalents, and alternatives falling within the spirit and scope of the invention are intended to be included within the scope of the invention.
Claims (10)
1. The recognition method based on the multinational license plate is characterized by comprising the following steps of:
acquiring an image to be identified;
positioning a license plate in the image to be identified by utilizing a target detection network to obtain license plate frame coordinates and license plate images in the license plate frame coordinates;
inputting the license plate image into a license plate multi-attribute classification network to obtain preset type attribute information corresponding to the license plate, wherein the preset type attribute information at least comprises license plate country category and license plate color;
inputting the license plate image into a character detection and recognition network to obtain character information in a license plate, wherein the character information comprises character coordinates and character category information;
inputting the attribute information of the preset type and the character information into a preset template function for matching;
if the matching is successful, the identification result of the license plate is output.
2. The multi-country license plate based identification method of claim 1, further comprising:
acquiring license plate corner points of license plates in the license plate frame coordinates by utilizing the target detection network;
correcting the license plate image by utilizing the license plate corner points to obtain a license plate screenshot;
the inputting the license plate image into a license plate multi-attribute classification network to obtain the preset type attribute information corresponding to the license plate comprises the following steps: inputting the license plate screenshot into a license plate multi-attribute classification network to obtain preset type attribute information corresponding to the license plate;
the inputting the license plate image into a character detection and recognition network to obtain character information in the license plate comprises the following steps: inputting the license plate screenshot into a character detection and recognition network to obtain all character coordinates and character information in the license plate.
3. The multi-country license plate-based recognition method of claim 2, wherein the correcting the license plate image by using the license plate corner points to obtain a license plate screenshot comprises:
correcting the license plate image through affine transformation by utilizing the license plate corner points to obtain a license plate screenshot.
4. The method of claim 1, further comprising:
and defining the country category and the corresponding license plate character structure to obtain the preset template function.
5. The multi-country license plate based identification method of claim 1, wherein the target detection network employs a YOLOv5 detection network with increased key point regression;
the license plate multi-attribute classification network adopts a resnet18 network, and output layers of the resnet18 network are a country category attribute output layer and a color attribute output layer respectively.
6. The multi-country license plate-based recognition method of claim 1, wherein the classification of the character class information comprises at least one or more of the following:
performing character detection by using the character detection network to obtain a character recognition result, and classifying license plates into Chinese characters and Chinese character-free classes according to whether the character recognition result has Chinese characters or not;
performing character detection by using the character detection network to obtain a character recognition result, and dividing the license plate into Arabic number letters and Arabic number letter-free types according to whether Arabic number letters exist in the character recognition result;
performing character detection by using the character detection network to obtain a character recognition result, and dividing the license plate into Arabic numeral letters and Arabic numeral letters according to whether Arabic numeral letters exist in the character recognition result;
performing character detection by using the character detection network to obtain a character recognition result, and dividing the license plate into a single-layer structure and a double-layer structure according to whether the character result is of a double-layer structure;
performing character detection by using the character detection network to obtain a character recognition result, and dividing the license plate into a symmetrical structure and an asymmetrical structure according to whether the character result is a bilateral symmetry structure or not;
and carrying out character detection by utilizing the character detection network to obtain a character recognition result, and dividing the license plate into a character string with a country sign and a character string without a country sign according to whether the character recognition result has the character string with the country sign.
7. The multi-country license plate based recognition method of claim 1, wherein after the inputting the preset type attribute information and the character information into a preset template function for matching, the method further comprises:
if the matching is unsuccessful, storing the identification result of the current license plate, and defining the identification result to be used as a new template function to be newly added into a template function library.
8. A multi-country license plate-based identification device, the device comprising:
the acquisition module is used for acquiring an image to be identified;
the detection module is used for positioning the license plate in the image to be identified by utilizing a target detection network so as to obtain license plate frame coordinates and license plate images in the license plate frame coordinates;
the attribute identification module is used for inputting the license plate image into a license plate multi-attribute classification network to obtain preset type attribute information corresponding to the license plate, wherein the preset type attribute information at least comprises license plate country category and license plate color;
the character recognition module is used for inputting the license plate image into a character detection recognition network through the character recognition module so as to acquire character information in the license plate, wherein the character information comprises character coordinates and character category information;
the matching module is used for inputting the attribute information of the preset type and the character information into a preset template function to match;
and the output module outputs the identification result of the license plate when the matching is successful.
9. A computer device comprising a processor for implementing the steps of the multi-national license plate based identification method according to any one of claims 1-7 when executing a computer program stored in a memory.
10. A storage medium having stored thereon a computer program, which when executed by a processor performs the steps of the multi-national license plate based identification method as claimed in any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210783601.9A CN117132969A (en) | 2022-07-05 | 2022-07-05 | Recognition method, device, computer and storage medium based on multinational license plate |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210783601.9A CN117132969A (en) | 2022-07-05 | 2022-07-05 | Recognition method, device, computer and storage medium based on multinational license plate |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117132969A true CN117132969A (en) | 2023-11-28 |
Family
ID=88849612
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210783601.9A Pending CN117132969A (en) | 2022-07-05 | 2022-07-05 | Recognition method, device, computer and storage medium based on multinational license plate |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117132969A (en) |
-
2022
- 2022-07-05 CN CN202210783601.9A patent/CN117132969A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11455805B2 (en) | Method and apparatus for detecting parking space usage condition, electronic device, and storage medium | |
| CN108009543B (en) | License plate recognition method and device | |
| CN107944450B (en) | License plate recognition method and device | |
| US11380104B2 (en) | Method and device for detecting illegal parking, and electronic device | |
| WO2021051601A1 (en) | Method and system for selecting detection box using mask r-cnn, and electronic device and storage medium | |
| CN111078946A (en) | Bayonet vehicle retrieval method and system based on multi-target regional characteristic aggregation | |
| CN111738036B (en) | Image processing method, device, equipment and storage medium | |
| CN112215190A (en) | Illegal building detection method based on YOLOV4 model | |
| CN111860219B (en) | High-speed channel occupation judging method and device and electronic equipment | |
| CN110889388A (en) | Violation identification method, device, equipment and storage medium | |
| CN113989744A (en) | Pedestrian target detection method and system based on oversized high-resolution image | |
| Parvin et al. | Vehicle number plate detection and recognition techniques: a review | |
| CN110298302B (en) | Human body target detection method and related equipment | |
| Zhang et al. | Image-based approach for parking-spot detection with occlusion handling | |
| CN112241736A (en) | Text detection method and device | |
| CN111292331B (en) | Image processing method and device | |
| Moseva et al. | Development of a System for Fixing Road Markings in Real Time | |
| CN114882490B (en) | Unlimited scene license plate detection and classification method based on point-guided positioning | |
| CN112101135A (en) | Moving target detection method and device and terminal equipment | |
| Qin et al. | A vehicle path tracking system with cooperative recognition of license plates and traffic network big data | |
| CN111709377A (en) | Feature extraction method, target re-identification method and device and electronic equipment | |
| EP4332910A1 (en) | Behavior detection method, electronic device, and computer readable storage medium | |
| CN117132969A (en) | Recognition method, device, computer and storage medium based on multinational license plate | |
| CN115953744A (en) | Vehicle identification tracking method based on deep learning | |
| CN115731477A (en) | Image recognition method, illicit detection method, terminal device, and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |