CN112580531B - Identification detection method and system for true and false license plates - Google Patents
Identification detection method and system for true and false license plates Download PDFInfo
- Publication number
- CN112580531B CN112580531B CN202011537265.7A CN202011537265A CN112580531B CN 112580531 B CN112580531 B CN 112580531B CN 202011537265 A CN202011537265 A CN 202011537265A CN 112580531 B CN112580531 B CN 112580531B
- Authority
- CN
- China
- Prior art keywords
- license plate
- image
- detection
- image group
- images
- 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
- 238000001514 detection method Methods 0.000 title claims abstract description 241
- 238000000034 method Methods 0.000 claims abstract description 37
- 238000012549 training Methods 0.000 claims description 50
- 230000002776 aggregation Effects 0.000 claims description 25
- 238000004220 aggregation Methods 0.000 claims description 25
- 238000009826 distribution Methods 0.000 claims description 20
- 238000002372 labelling Methods 0.000 claims description 7
- 238000013528 artificial neural network Methods 0.000 claims description 6
- 238000010586 diagram Methods 0.000 description 4
- 230000008878 coupling Effects 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 238000003860 storage Methods 0.000 description 3
- 238000004891 communication Methods 0.000 description 2
- 238000001914 filtration Methods 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 1
- 238000005520 cutting process Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000007639 printing Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/107—Static hand or arm
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2413—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on distances to training or reference patterns
- G06F18/24133—Distances to prototypes
- G06F18/24137—Distances to cluster centroïds
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
-
- 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
- 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
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/107—Static hand or arm
- G06V40/117—Biometrics derived from hands
Abstract
The embodiment of the application discloses a method and a system for identifying and detecting a real and false license plate, which are used for improving the accuracy of judging the real and false license plate of a vehicle. The method comprises the following steps: acquiring a pre-detection license plate image group of the same vehicle; judging whether the images in the pre-detection license plate image group have human hand characteristics or not; if yes, key points are set on the hand characteristics of the human body respectively, and a target detection image group is generated; picking a first target detection image from the target detection image group; generating a detection result by the first target detection image through a target model; judging whether the detection result is larger than a first threshold value; if not, counting the number of sheets of the target detection image group, wherein the distance between the center point of the image license plate and the corresponding key point is smaller than a second threshold value; judging whether the number of images smaller than a second threshold is not smaller than half of the total number of images in the target detection image group; if yes, determining that the license plate in the target detection image group is a fake license plate.
Description
Technical Field
The embodiment of the application relates to the field of intelligent monitoring, in particular to a method and a system for identifying and detecting a real license plate and a fake license plate.
Background
With the occurrence of motor vehicle license plate sleeving, fake motor vehicle license plate and other phenomena, the requirements on the true and false recognition technology of the license plate are increasingly urgent. In recent years, with the development of scientific technology, an unattended scheme based on license plate recognition technology appears, so that the traffic efficiency of a parking lot is improved, and the expense of a large number of staff is saved for a property manager.
In the parking lot, the system can release the brake of the user vehicle only by identifying the license plate. Because of unattended operation, a few users can take measures such as displaying license plate images of vehicles or printing license plates of vehicles by a mobile phone to impersonate a month truck to carry out fee evasion in order to reduce the expenditure of the parking fee of the own vehicle in a parking lot.
Disclosure of Invention
The embodiment of the application provides a method and a system for identifying and detecting a real and false license plate, which are used for improving the accuracy of judging the real and false license plate of a vehicle.
The application provides a method for identifying and detecting a true license plate and a false license plate in a first aspect, which comprises the following steps:
acquiring a pre-detection license plate image group of the same vehicle;
judging whether the images in the pre-detection license plate image group have human hand characteristics or not;
if yes, key points are set on the hand features of the human body respectively, and a target detection image group is generated;
picking a first target detection image from the target detection image group; generating a detection result of the first target detection image through a target model, wherein the target model is a model trained by central point characteristic information of license plate images in a learning sample set, and the detection result is a distance length value;
judging whether the detection result is larger than a first threshold value or not;
if not, counting the number of sheets of the target detection image group, wherein the distance between the center point of the image license plate and the corresponding key point is smaller than a second threshold value;
judging whether the number of images smaller than a second threshold is not smaller than half of the total number of images in the target detection image group;
if yes, determining that the license plate in the target detection image group is a fake license plate.
Optionally, before the first target detection image is generated into a detection result by using a target model, the detection method further includes:
acquiring historical license plate image data of a vehicle through a camera;
acquiring a license plate image sample set through the historical vehicle license plate image data, wherein the license plate image sample set is a sample set provided with a license plate center point;
inputting the image set of the license plate image sample set into an initial model to generate training data, wherein the initial model is a model established based on a neural network, and the training data is central point coordinate data of a license plate of a vehicle;
judging whether an aggregation area exists in the distribution of license plate center points according to the training data;
if yes, determining the initial model as a target model;
and determining a region center point according to the aggregation region.
Optionally, after determining whether the distribution of the license plate center points has the aggregation area according to the training data, the detection method further includes:
if not, secondarily acquiring a license plate image sample set, and performing the following steps: and inputting the image set of the license plate image sample acquired secondarily into an initial model to generate training data, and judging whether an aggregation area exists in the distribution of license plate center points according to the training data generated secondarily.
Optionally, the acquiring the pre-detected license plate image group of the same vehicle includes:
acquiring a video containing pre-detection license plate information of the same vehicle through a camera;
and intercepting at least three pre-detection license plate images according to the video containing the detection license plate information, and generating a pre-detection license plate image group in a collecting way.
Optionally, the picking the first target detection image from the target detection image group includes:
respectively labeling the frames of license plates on images in the target detection image group;
comparing the sizes of corresponding license plates displayed on the image according to the frames;
and picking a first target detection image according to the comparison result, wherein the first target detection image is the image with the largest frame in the comparison result.
Optionally, after the determining whether the image in the pre-detected license plate image group has the hand feature of the human body, the detection method further includes:
if not, determining that the license plate in the pre-detection license plate image group is a true license plate.
Optionally, after the determining whether the detection result is greater than the first threshold, the detection method further includes:
if yes, determining that the license plate in the first target detection image is a fake license plate.
Optionally, after the determining whether the number of images smaller than the second threshold is not smaller than half of the total number of images in the target detection image group, the detection method further includes:
if not, determining that the license plate in the target detection image group is a true license plate.
The present application provides in a second aspect a system for identifying and detecting a genuine license plate, comprising:
the first acquisition unit is used for acquiring a pre-detection license plate image group of the same vehicle;
the first judging unit is used for judging whether the images in the pre-detection license plate image group have hand characteristics of a human body or not;
the first execution unit is used for setting key points for the hand features of the human body respectively when the first judgment unit determines that the hand features of the human body exist in the images in the pre-detection license plate image group, and generating a target detection image group;
the picking unit is used for picking a first target detection image from the target detection image group;
the generation unit is used for generating a detection result of the first target detection image through a target model, wherein the target model is a model obtained by training central point characteristic information of license plate images in a learning sample set, and the detection result is a distance length value;
the second judging unit is used for judging whether the detection result is larger than a first threshold value or not;
the second execution unit is used for counting the number of the image license plate center points in the target detection image group, the distance between the image license plate center points and the corresponding key points of which is smaller than a second threshold value when the second judgment unit determines that the detection result is not larger than the first threshold value;
a third judging unit for judging whether the number of images smaller than the second threshold is not smaller than half of the total number of images in the target detection image group;
and the third execution unit is used for determining that the license plate in the target detection image group is a fake license plate when the third judgment unit determines that the number of images smaller than the second threshold is not smaller than half of the total number of images of the target detection image group.
Optionally, the detection system further includes:
the third acquisition unit is used for acquiring historical vehicle license plate image data through the camera;
a fourth obtaining unit, configured to obtain a license plate image sample set according to the historical vehicle license plate image data, where the license plate image sample set is a sample set provided with a license plate center point;
the training unit is used for inputting the image set of the license plate image sample into an initial model to generate training data, wherein the initial model is a model established based on a neural network, and the training data is center point coordinate data of a license plate of a vehicle;
the fourth judging unit is used for judging whether an aggregation area exists in the distribution of license plate center points according to the training data;
the fourth execution unit is used for determining the initial model as a target model when the fourth judgment unit determines that the distribution of license plate center points has an aggregation area according to the training data;
and the region center determining unit is used for determining a region center point according to the aggregation region.
Optionally, the detection system further includes:
the fifth execution unit is configured to, when the fourth judgment unit determines that the distribution of license plate center points does not have an aggregation area according to the training data, secondarily obtain a license plate image sample set, and perform the steps of: and inputting the image set of the license plate image sample acquired secondarily into an initial model to generate training data, and judging whether an aggregation area exists in the distribution of license plate center points according to the training data generated secondarily.
Optionally, the first obtaining unit includes:
the second acquisition module is used for acquiring videos of the same vehicle, which contain pre-detection license plate information, through the camera;
and the intercepting module is used for intercepting at least three pre-detected license plate images according to the video containing the detected license plate information and generating a pre-detected license plate image group in a collecting way.
Optionally, the picking unit includes:
the marking module is used for marking the rims of license plates on the images in the target detection image group respectively;
the comparison module is used for comparing the sizes of the corresponding license plates displayed on the image according to the frames;
the picking module is used for picking a first target detection image according to the comparison result, wherein the first target detection image is the image with the largest frame in the comparison result.
Optionally, the detection system further includes:
and the sixth execution unit is used for determining that the license plate in the pre-detection license plate image group is a true license plate when the first judgment unit determines that the image in the pre-detection license plate image group does not have the hand characteristics of the human body.
Optionally, the detection method further includes:
and the seventh execution unit is used for determining that the license plate in the first target detection image is a fake license plate when the second judgment unit determines that the detection result is larger than a first threshold value.
Optionally, the detection method further includes:
and the eighth execution unit is used for determining that the license plate in the target detection image group is a true license plate when the third judgment unit determines that the number of images smaller than the second threshold is smaller than half of the total number of images in the target detection image group.
From the above technical solutions, the embodiments of the present application have the following advantages:
the embodiment of the application provides a method and a system for identifying and detecting real and false license plates, which are characterized in that whether the acquired pre-detected license plate images of the same vehicle have hand features of a human body is judged, if yes, key points are set on the hand features of the human body to generate a target detection image group, a first target detection image is picked from the target detection image group and input into a target model, whether the numerical value of the detection result is larger than a first threshold value is judged according to the detection result of the target model, if not, the distances between license plate center points of all images in the target detection image group and corresponding hand feature key points are counted to be smaller than the number of the second threshold value, and if the counted number exceeds or is equal to half of the total image number of the image group, the license plates in the target detection image group are determined to be false license plates. The method and the system can realize the filtration of most of the fake license plates under the unattended condition, thereby improving the accuracy of judging the true license plates and the fake license plates.
Drawings
FIG. 1 is a schematic flow chart of an embodiment of a method for identifying and detecting a license plate in an embodiment of the present application;
fig. 2-1 and fig. 2-2 are schematic flow diagrams of another embodiment of a method for identifying and detecting a license plate in the embodiments of the present application;
FIG. 3 is a schematic structural diagram of an embodiment of a system for identifying and detecting a license plate in an embodiment of the present application;
FIG. 4 is a schematic structural diagram of another embodiment of a system for identifying and detecting a license plate in an embodiment of the present application;
fig. 5 is a schematic structural diagram of another embodiment of a system for identifying and detecting a license plate in an embodiment of the present application.
Detailed Description
In order to better understand the technical solutions of the present invention, the following description will clearly and completely describe the technical solutions of the embodiments of the present invention with reference to the drawings in the embodiments of the present invention, and it is apparent that the described embodiments are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, shall fall within the scope of the invention.
The embodiment of the application provides a method and a system for identifying and detecting a true license plate and a false license plate, which are used for filtering most of false license plates under the unattended condition, so that the accuracy of judging the true license plates and the false license plates is improved.
In this embodiment, the method for identifying and detecting the true license plate and the false license plate can be implemented in a system, can be implemented in a server, and can also be implemented in the system without specific limitation. For ease of description, embodiments of the present application are described using a system for performing a subject example.
Referring to fig. 1, an embodiment of a method for identifying and detecting a real license plate in an embodiment of the present application includes:
101. the system acquires a pre-detection license plate image group of the same vehicle;
in the embodiment of the application, when the system recognizes that a vehicle enters a parking lot, a sensing area of a camera at an entrance of the parking lot is set, and when the vehicle passes through the sensing area to the entrance of the parking lot, a license plate image group of the vehicle is acquired through the camera, so that identity information of the vehicle is recognized.
The method for obtaining the license plate image group may be to shoot a photo group of the vehicle passing through the sensing area in a preset time by using a high-level camera, or may be to record a video of the vehicle passing through the sensing area, and then intercept a predetermined frame number image from the video as the image group, which is not limited herein.
102. The system judges whether the images in the pre-detection license plate image group have human hand characteristics or not; if yes, go to step 103;
after the system identifies the identity information of the vehicle through the license plate image set, in order to detect whether the human arm is around the license plate, whether the license plate image set acquired through the high-order camera has the hand characteristics of the human body needs to be identified, if yes, step 103 is executed.
103. The system sets key points for the hand characteristics of the human body respectively and generates a target detection image group;
when the hand features of the human body are identified from the license plate image group, in order to detect whether the human body arm is around the license plate, a plurality of key points are required to be set for the hand features of the human body to serve as reference points for detecting the distance from the human body arm to the center point of the license plate, then the set key points are marked in the images of the pre-detected license plate image group, and the processed images generate a target detection image group.
There are various methods for setting the key points, and a skeleton may be generally used as a set of coordinate points for describing the posture of the human body by connecting. Each coordinate point in the skeleton is a joint, which is also called a key point. The effective connection between two keypoints is called a "pair". The method of setting the key point is not particularly limited herein.
104. The system picks a first target detection image from the target detection image group;
in order to improve accuracy of the subsequent target model on identification of the true license plate and the false license plate, the system needs to pick an image with highest license plate identification degree in the image group from the target detection image group subjected to the key point labeling processing as a first target detection image.
105. The system generates a detection result through a target model by using a first target detection image;
the target model is a model trained in advance by license plate position data of parking lot history, for example: and acquiring image data of a normal vehicle playing position of the parking lot for one month, and training four clustering centers, wherein each clustering center represents a playing position area possibly occurring in the normal vehicle driving process. The number of cluster centers present is generally determined by the height of the vehicle itself and the driving habits of the driver. When the system inputs the image into the target model, the target model can generate a detection result according to the distance between the license plate center point in the image and the center point of each cluster.
The training method for the target model may further train the model by constructing another neural network framework besides the clustering algorithm described above, which is not limited herein.
106. The system judges whether the detection result is larger than a first threshold value; if not, go to step 107;
in the embodiment of the present application, in order to detect the authenticity of the license plate of the target vehicle, the system needs to determine, according to the detection result output by the target model, whether the distance between the license plate and the cluster in the first target detection image of the target vehicle is greater than a certain extent, so that a threshold value of the distance needs to be set, if so, it is determined that the license plate is a fake license plate, and if not, it is determined that the license plate is a fake license plate, and step 107 is executed.
107. The system counts the number of sheets that the distance between the center point of the image license plate in the target detection image group and the corresponding key point is smaller than a second threshold;
when the system determines that the detection result is not greater than the first threshold, the system needs to count the number of sheets of the target detection image group, wherein the distance between the center point of the image license plate and the corresponding key point is smaller than a second threshold, and the second threshold refers to a distance value.
108. The system judges whether the number of images smaller than a second threshold is not smaller than half of the total number of images in the target detection image group; if yes, go to step 109;
the system carries out second judgment on the license plate so as to improve the judgment precision of the true license plate and the false license plate.
After the number of images greater than a certain distance in the target detection image group is obtained, it is determined whether the number of images exceeds half of the total number of images, if yes, it is determined that the hand is in the vicinity of the license plate, and step 109 is executed.
109. The system determines that the license plate in the target detection image group is a fake license plate.
And when the system determines that the number of the images smaller than the second threshold is not smaller than half of the total number of the images in the target detection image group, determining that the license plate in the target detection image group is a fake license plate.
In the embodiment of the application, the parking lot system judges whether the license plate image to be detected has the hand features or not by judging the license plate image to be detected, if so, key points are set through the hand features, the processed image is input into a pre-trained model, and the authenticity of the license plate in the license plate image is determined according to the output result of the model, so that most of fake license plates are filtered under the unattended condition, and the accuracy of judging the fake license plates is improved.
Another embodiment of the method for detecting the identification of the genuine license plate is described in detail below with reference to fig. 2-1 to 2-2.
Referring to fig. 2-1 and fig. 2-2, in an embodiment of the present application, another embodiment of a method for identifying and detecting a real license plate includes:
201. the system acquires videos of the same vehicle, which contain pre-detection license plate information, through a camera;
202. the system intercepts at least three pre-detection license plate images according to the video containing the detection license plate information, and generates a pre-detection license plate image group in a collecting way;
in the embodiment of the application, the system needs the camera to record the video of the same vehicle passing through the sensing area of the camera, and then intercepts the pre-detection license plate image according to the video to identify the identity information of the vehicle, and takes the identity information as the model input data of the subsequent input target. And when the cut pre-detected license plate images are at least three, and images containing clear license plates are cut from the video according to preset time intervals, stopping cutting the video, and collecting the cut images to generate a pre-detected license plate image group.
203. The system judges whether the images in the pre-detection license plate image group have human hand characteristics or not; if yes, go to step 204; if not, go to step 205;
204. the system sets key points for the hand characteristics of the human body respectively and generates a target detection image group;
after this step is performed, step 206 is performed.
Steps 203 and 204 in this embodiment are similar to steps 102 and 103 in the previous embodiment, and will not be repeated here.
205. The system determines that license plates in the pre-detection license plate image group are true license plates;
when the system determines that the images in the pre-detection license plate image group do not have human hand features, the system determines the true license plate of the license plate in the pre-detection license plate image group.
206. The system marks the rims of license plates on images in the target detection image group respectively;
207. the system compares the sizes of corresponding license plates displayed on the image according to the frames;
208. the system picks a first target detection image according to the comparison result, wherein the first target detection image is the image with the largest frame in the comparison result;
in order to improve accuracy of detecting license plate authenticity of the system, a most clear image of a license plate is required to be selected from a pre-detected license plate image group to serve as a target detection image, therefore, a license plate frame is required to be marked on the image in the image group along the outer edge of the license plate in sequence, after all the images are marked by the system, the system needs to determine the definition ranking of the license plate according to the sizes of frames on the images, and the image with the largest ranking is selected to serve as a first target detection image.
209. The system acquires historical license plate image data of the vehicle through a camera;
before the system inputs the image into the target model for detection, an initial model frame is constructed, and then the initial model is trained to obtain the target model.
In the embodiment of the application, the method for training the initial model adopts a clustering algorithm for training, so that the system needs to acquire historical vehicle license plate image data of a near time period or several time periods as a comparison sample.
210. The system acquires a license plate image sample set through historical vehicle license plate image data, wherein the license plate image sample set is a sample set provided with a license plate center point;
after the system acquires the historical vehicle license plate image data as a comparison sample, in order to obtain a reference point of the comparison sample, license plate center point labeling processing is required to be carried out on the comparison sample, and then the processed images are assembled into a license plate image sample set.
211. The system inputs the image set of the license plate image sample into an initial model to generate training data;
212. the system judges whether an aggregation area exists in the distribution of license plate center points according to the training data; if yes, go to step 213; if not, go to step 214;
213. the system determines the initial model as a target model; step 215 is performed after step 213 is performed;
214. the system acquires a license plate image sample set for the second time, and carries out the following steps: inputting the image set of the license plate image sample acquired secondarily into an initial model to generate training data, and judging whether an aggregation area exists in the distribution of license plate center points according to the training data generated secondarily;
215. the system determines a region center point according to the aggregation region;
the system can construct an initial model by adopting a clustering algorithm framework, and the training method is as follows: and (3) intensively inputting the license plate image sample set processed by the marked center point into an initial model to generate training data, wherein the training data is the distribution data of the center points, then judging whether the distribution data of the center points form a point aggregation area, if so, determining that the initial model training is finished into a target model, and if not, secondarily acquiring the license plate image sample set and inputting the license plate image sample set into the initial model for training until the system judges that the center points of the license plate are distributed with the point aggregation area. And then confirming the center point of the area according to the point gathering area to be used as a detection calculation basis when the pre-detection license plate image is input into the target model.
216. The system generates a detection result through a target model by using a first target detection image;
217. the system judges whether the detection result is larger than a first threshold value; if yes, go to step 218; if not, go to step 219;
steps 216 and 217 in this embodiment are similar to steps 105 and 106 in the previous embodiment, and will not be repeated here.
218. The system determines that the license plate in the first target detection image is a fake license plate;
when the system determines that the detection result is larger than the first threshold value, the license plate in the first target detection image is determined to be a fake license plate.
219. The system counts the number of sheets that the distance between the center point of the image license plate in the target detection image group and the corresponding key point is smaller than a second threshold;
220. the system judges whether the number of images smaller than a second threshold is not smaller than half of the total number of images in the target detection image group; if yes, go to step 221; if not, go to step 222;
221. the system determines that the license plate in the target detection image group is a fake license plate;
steps 219 and 221 in this embodiment are similar to steps 107 and 109 in the previous embodiment, and will not be repeated here.
222. The system determines that the license plate in the target detection image set is a true license plate.
And when the system determines that the number of images smaller than the second threshold is smaller than half of the total number of images in the target detection image group, determining that the license plate in the target detection image group is a true license plate.
According to the embodiment of the application, the parking lot system can be used as a sample set to put into an initial model for training according to historical vehicle license plate data, and the parking lot has robustness on the anti-counterfeiting effect of the vehicle license plate through big data and deep learning.
The method for identifying and detecting the true license plate and the false license plate in the embodiment of the application is described above, and the system for identifying and detecting the true license plate and the false license plate in the embodiment of the application is described below:
referring to fig. 3, an embodiment of a system for identifying and detecting a real license plate in an embodiment of the present application includes:
a first acquiring unit 301, configured to acquire a pre-detected license plate image group of the same vehicle;
a first judging unit 302, configured to judge whether the image in the pre-detection license plate image set has a hand feature of a human body;
the first execution unit 303 is configured to set key points for hand features of a human body when the first judgment unit 302 determines that the hand features of the human body exist in the images in the pre-detection license plate image group, and generate a target detection image group;
a picking unit 304, configured to pick a first target detection image from the target detection image group;
the generating unit 305 is configured to generate a detection result from the first target detection image through a target model, where the target model is a model trained by feature information of a central point of a license plate image in the learning sample set, and the detection result is a distance length value;
a second judging unit 306, configured to judge whether the detection result is greater than a first threshold;
the second executing unit 307 is configured to, when the second judging unit 306 determines that the detection result is not greater than the first threshold, count the number of sheets of the target detection image group in which the distance between the center point of the image license plate and the corresponding key point is less than the second threshold;
a third judging unit 308 for judging whether the number of images smaller than the second threshold is not smaller than half of the total number of images in the target detection image group;
and a third executing unit 309, configured to determine that the license plate in the target detection image group is a pseudo license plate when the third judging unit 308 determines that the number of images smaller than the second threshold is not less than half of the total number of images in the target detection image group.
In this embodiment, after the first obtaining unit 301 obtains the pre-detected license plate image set of the same vehicle, the first judging unit 302 judges whether the images in the pre-detected license plate image set have human hand features, if so, the first executing unit 303 sets key points for the human hand features and generates a target detection image set, the picking unit 304 is used for picking the first target detection image in the target detection image set, the generating unit 305 inputs the first target detection image into the target model to generate a detection result, the second executing unit 307 counts the number of images with the distance between the central point of the license plate of the image and the corresponding key point being smaller than the second threshold when the second judging unit 306 determines that the detection result is not greater than the first threshold, if the number of images with the distance between the central point of the license plate of the image and the corresponding key point is not less than half of the total number of images in the target detection image set, the third executing unit 309 can determine that the license plate in the target detection image set is a fake license plate.
Referring to fig. 4, another embodiment of a system for identifying and detecting a real license plate in an embodiment of the present application includes:
a first obtaining unit 401, configured to obtain a pre-detected license plate image group of the same vehicle;
a first judging unit 402, configured to judge whether the image in the pre-detection license plate image group has a hand feature of a human body;
the first execution unit 403 is configured to set key points for hand features of a human body when the first judgment unit 402 determines that the hand features of the human body exist in the images in the pre-detection license plate image set, and generate a target detection image set;
a sixth execution unit 404, configured to determine that the license plate in the pre-detection license plate image set is a real license plate when the first determination unit 402 determines that the image in the pre-detection license plate image set does not have the hand feature of the human body;
a picking unit 405, configured to pick a first target detection image from the target detection image group;
a third obtaining unit 406, configured to obtain historical license plate image data of the vehicle through a camera;
a fourth obtaining unit 407, configured to obtain a license plate image sample set according to historical license plate image data of the vehicle, where the license plate image sample set is a sample set provided with a license plate center point;
the training unit 408 is configured to input an image set of the license plate image sample into an initial model to generate training data, where the initial model is a model built based on a neural network, and the training data is central point coordinate data of a license plate of the vehicle;
a fourth judging unit 409, configured to judge whether an aggregation area exists in the distribution of license plate center points according to the training data;
a fourth execution unit 410, configured to determine, when the fourth determination unit 409 determines that the distribution of license plate center points has an aggregation area according to the training data, the initial model as the target model;
a fifth execution unit 411, configured to, when the fourth judgment unit 409 determines that the distribution of license plate center points does not have an aggregation area according to the training data, secondarily acquire a license plate image sample set, and perform the steps of: inputting the image set of the license plate image sample acquired secondarily into an initial model to generate training data, and judging whether an aggregation area exists in the distribution of license plate center points according to the training data generated secondarily;
a region center determining unit 412 for determining a region center point from the aggregated region;
a generating unit 413, configured to generate a detection result from the first target detection image through a target model, where the target model is a model trained by feature information of a central point of the license plate image in the learning sample set, and the detection result is a distance length value;
a second judging unit 414, configured to judge whether the detection result is greater than a first threshold;
the second executing unit 415 is configured to, when the second judging unit 414 determines that the detection result is not greater than the first threshold, count the number of sheets of the target detection image group in which the distance between the center point of the image license plate and the corresponding key point is less than the second threshold;
a seventh executing unit 416, configured to determine that the license plate in the first target detection image is a pseudo license plate when the second judging unit 414 determines that the detection result is greater than the first threshold;
a third judging unit 417 for judging whether the number of images smaller than the second threshold is not smaller than half of the total number of images in the target detection image group;
a third executing unit 418, configured to determine that the license plate in the target detection image group is a pseudo license plate when the third judging unit 417 determines that the number of images smaller than the second threshold is not smaller than half of the total number of images of the target detection image group;
an eighth execution unit 419 is configured to determine that the license plate in the target detection image group is a true license plate when the third determination unit 417 determines that the number of images smaller than the second threshold is smaller than half of the total number of images in the target detection image group.
In the embodiment of the present application, the first acquisition unit 401 includes a second acquisition module 4011 and an interception module 4012.
The second acquiring module 4011 is configured to acquire, through a camera, a video of the same vehicle including pre-detected license plate information;
the intercepting module 4012 is used for intercepting at least three pre-detected license plate images according to the video containing the detected license plate information, and generating a pre-detected license plate image group in a collecting way.
In the embodiment of the present application, the picking unit 405 includes a labeling module 4051, a comparing module 4052 and a picking module 4053.
The labeling module 4051 is used for labeling the frames of license plates on the images in the target detection image group respectively;
the comparison module 4052 is used for comparing the sizes of the corresponding license plates displayed on the image according to the frames;
the picking module 4053 is configured to pick a first target detection image according to the comparison result, where the first target detection image is an image with a largest frame in the comparison result.
In the above embodiment, the functions of each unit and module correspond to the steps in the embodiment shown in fig. 2, and are not repeated here.
Referring to fig. 5, another embodiment of a method and a system for identifying and detecting a real license plate in the embodiment of the present application includes:
a processor 501, a memory 502, an input/output unit 503, and a bus 504;
the processor 501 is connected to the memory 502, the input/output unit 503, and the bus 504;
the processor 501 specifically performs the following operations:
acquiring a pre-detection license plate image group of the same vehicle;
judging whether the images in the pre-detection license plate image group have human hand characteristics or not;
if yes, key points are set on the hand characteristics of the human body respectively, and a target detection image group is generated;
picking a first target detection image from the target detection image group;
generating a detection result from the first target detection image through a target model, wherein the target model is a model obtained by training central point characteristic information of license plate images in a learning sample set, and the detection result is a distance length value;
judging whether the detection result is larger than a first threshold value;
if not, counting the number of sheets of the target detection image group, wherein the distance between the center point of the image license plate and the corresponding key point is smaller than a second threshold value;
judging whether the number of images smaller than a second threshold is not smaller than half of the total number of images in the target detection image group;
if yes, determining that the license plate in the target detection image group is a fake license plate.
In this embodiment, the functions of the processor 501 correspond to the steps in the embodiments described in fig. 1 to 4, and are not described herein.
It will be clear to those skilled in the art that, for convenience and brevity of description, specific working procedures of the above-described systems, apparatuses and units may refer to corresponding procedures in the foregoing method embodiments, which are not repeated herein.
In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods may be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative, e.g., the division of the units is merely a logical function division, and there may be additional divisions when actually implemented, e.g., multiple units or components may be combined or integrated into another system, or some features may be omitted or not performed. Alternatively, the coupling or direct coupling or communication connection shown or discussed with each other may be an indirect coupling or communication connection via some interfaces, devices or units, which may be in electrical, mechanical or other form.
The units described as separate units may or may not be physically separate, and units shown as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
In addition, each functional unit in each embodiment of the present application may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit. The integrated units may be implemented in hardware or in software functional units.
The integrated units, if implemented in the form of software functional units and sold or used as stand-alone products, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present application may be embodied in essence or a part contributing to the prior art or all or part of the technical solution in the form of a software product stored in a storage medium, including several instructions to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in the embodiments of the present application. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a read-only memory (ROM), a random access memory (RAM, random access memory), a magnetic disk, or an optical disk, or other various media capable of storing program codes.
Claims (10)
1. The method for identifying and detecting the true license plate is characterized by comprising the following steps:
acquiring a pre-detection license plate image group of the same vehicle;
judging whether the images in the pre-detection license plate image group have human hand characteristics or not;
if yes, key points are set on the hand features of the human body respectively, and a target detection image group is generated;
picking a first target detection image from the target detection image group;
generating a detection result of the first target detection image through a target model, wherein the target model is a model trained by central point characteristic information of license plate images in a learning sample set, and the detection result is a distance length value;
judging whether the detection result is larger than a first threshold value or not;
if not, counting the number of sheets of the target detection image group, wherein the distance between the center point of the image license plate and the corresponding key point is smaller than a second threshold value;
judging whether the number of images smaller than a second threshold is not smaller than half of the total number of images in the target detection image group;
if yes, determining that the license plate in the target detection image group is a fake license plate.
2. The method according to claim 1, wherein before the first object detection image is generated into a detection result by an object model, the method further comprises:
acquiring historical license plate image data of a vehicle through a camera;
acquiring a license plate image sample set through the historical vehicle license plate image data, wherein the license plate image sample set is a sample set provided with a license plate center point;
inputting the image set of the license plate image sample set into an initial model to generate training data, wherein the initial model is a model established based on a neural network, and the training data is central point coordinate data of a license plate of a vehicle;
judging whether an aggregation area exists in the distribution of license plate center points according to the training data;
if yes, determining the initial model as a target model;
and determining a region center point according to the aggregation region.
3. The method according to claim 2, wherein after determining whether the distribution of license plate center points has an aggregation area according to the training data, the method further comprises:
if not, secondarily acquiring a license plate image sample set, and performing the following steps: and inputting the image set of the license plate image sample acquired secondarily into an initial model to generate training data, and judging whether an aggregation area exists in the distribution of license plate center points according to the training data generated secondarily.
4. The method according to claim 3, wherein the acquiring the pre-detected license plate image group of the same vehicle includes:
acquiring a video containing pre-detection license plate information of the same vehicle through a camera;
and intercepting at least three pre-detection license plate images according to the video containing the pre-detection license plate information, and generating a pre-detection license plate image group in a collecting way.
5. The method of claim 4, wherein picking a first object detection image from the set of object detection images comprises:
respectively labeling the frames of license plates on images in the target detection image group;
comparing the sizes of corresponding license plates displayed on the image according to the frames;
and picking a first target detection image according to the comparison result, wherein the first target detection image is the image with the largest frame in the comparison result.
6. The method according to any one of claims 1 to 5, wherein after the determining whether the images in the pre-detected license plate image group have hand features of a human body, the method further comprises:
if not, determining that the license plate in the pre-detection license plate image group is a true license plate.
7. The detection method according to any one of claims 1 to 5, wherein after the determination as to whether the detection result is greater than a first threshold value, the detection method further comprises:
if yes, determining that the license plate in the first target detection image is a fake license plate.
8. The detection method according to any one of claims 1 to 5, wherein after the determination as to whether the number of images smaller than the second threshold is not smaller than half of the total number of images in the target detection image group, the detection method further comprises:
if not, determining that the license plate in the target detection image group is a true license plate.
9. A system for identifying and detecting a genuine license plate, comprising:
the first acquisition unit is used for acquiring a pre-detection license plate image group of the same vehicle;
the first judging unit is used for judging whether the images in the pre-detection license plate image group have hand characteristics of a human body or not;
the first execution unit is used for setting key points for the hand features of the human body respectively when the first judgment unit determines that the hand features of the human body exist in the images in the pre-detection license plate image group, and generating a target detection image group;
the picking unit is used for picking a first target detection image from the target detection image group;
the generation unit is used for generating a detection result of the first target detection image through a target model, wherein the target model is a model obtained by training central point characteristic information of license plate images in a learning sample set, and the detection result is a distance length value;
the second judging unit is used for judging whether the detection result is larger than a first threshold value or not;
the second execution unit is used for counting the number of the image license plate center points in the target detection image group, the distance between the image license plate center points and the corresponding key points of which is smaller than a second threshold value when the second judgment unit determines that the detection result is not larger than the first threshold value;
a third judging unit for judging whether the number of images smaller than the second threshold is not smaller than half of the total number of images in the target detection image group;
and the third execution unit is used for determining that the license plate in the target detection image group is a fake license plate when the third judgment unit determines that the number of images smaller than the second threshold is not smaller than half of the total number of images of the target detection image group.
10. The detection system of claim 9, the detection system further comprising:
the third acquisition unit is used for acquiring historical vehicle license plate image data through the camera;
a fourth obtaining unit, configured to obtain a license plate image sample set according to the historical vehicle license plate image data, where the license plate image sample set is a sample set provided with a license plate center point;
the training unit is used for inputting the image set of the license plate image sample into an initial model to generate training data, wherein the initial model is a model established based on a neural network, and the training data is center point coordinate data of a license plate of a vehicle;
the fourth judging unit is used for judging whether an aggregation area exists in the distribution of license plate center points according to the training data;
the fourth execution unit is used for determining the initial model as a target model when the fourth judgment unit determines that the distribution of license plate center points has an aggregation area according to the training data;
and the region center determining unit is used for determining a region center point according to the aggregation region.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011537265.7A CN112580531B (en) | 2020-12-23 | 2020-12-23 | Identification detection method and system for true and false license plates |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011537265.7A CN112580531B (en) | 2020-12-23 | 2020-12-23 | Identification detection method and system for true and false license plates |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112580531A CN112580531A (en) | 2021-03-30 |
| CN112580531B true CN112580531B (en) | 2024-04-02 |
Family
ID=75139414
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011537265.7A Active CN112580531B (en) | 2020-12-23 | 2020-12-23 | Identification detection method and system for true and false license plates |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112580531B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113205069B (en) * | 2021-05-27 | 2023-05-30 | 上海高德威智能交通系统有限公司 | False license plate detection method and device and computer storage medium |
| CN113409502B (en) * | 2021-06-28 | 2023-03-24 | 杭州海康威视数字技术股份有限公司 | Method, device and system for controlling vehicle to pass and electronic equipment |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR101111724B1 (en) * | 2011-06-07 | 2012-02-24 | 주식회사 두원전자통신 | Automatic motoring method and system of illegal parking vehicles |
| CN106778741A (en) * | 2016-12-08 | 2017-05-31 | 中国人民解放军国防科学技术大学 | A kind of license plate locating method in the complex background image shot based on handheld terminal |
| KR101784449B1 (en) * | 2016-12-29 | 2017-11-06 | 주식회사 경림이앤지 | Parking enforcement system and method embedded with the number plate detection alarm service module of a vehicle |
| CN109993056A (en) * | 2019-02-25 | 2019-07-09 | 平安科技（深圳）有限公司 | A kind of method, server and storage medium identifying vehicle violation behavior |
| CN110533039A (en) * | 2019-09-04 | 2019-12-03 | 深圳市捷顺科技实业股份有限公司 | A kind of true-false detection method of license plate, device and equipment |
| CN111079751A (en) * | 2019-12-16 | 2020-04-28 | 深圳市芊熠智能硬件有限公司 | Method and device for identifying authenticity of license plate, computer equipment and storage medium |
| CN111340045A (en) * | 2020-02-12 | 2020-06-26 | 浙江大华技术股份有限公司 | License plate number recognition method and device and storage medium |
| CN111797713A (en) * | 2020-06-16 | 2020-10-20 | 浙江大华技术股份有限公司 | License plate recognition method and photographing device |
| CN111860477A (en) * | 2020-01-22 | 2020-10-30 | 北京嘀嘀无限科技发展有限公司 | Vehicle information verification method and system, terminal, and computer-readable storage medium |
-
2020
- 2020-12-23 CN CN202011537265.7A patent/CN112580531B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR101111724B1 (en) * | 2011-06-07 | 2012-02-24 | 주식회사 두원전자통신 | Automatic motoring method and system of illegal parking vehicles |
| CN106778741A (en) * | 2016-12-08 | 2017-05-31 | 中国人民解放军国防科学技术大学 | A kind of license plate locating method in the complex background image shot based on handheld terminal |
| KR101784449B1 (en) * | 2016-12-29 | 2017-11-06 | 주식회사 경림이앤지 | Parking enforcement system and method embedded with the number plate detection alarm service module of a vehicle |
| CN109993056A (en) * | 2019-02-25 | 2019-07-09 | 平安科技（深圳）有限公司 | A kind of method, server and storage medium identifying vehicle violation behavior |
| CN110533039A (en) * | 2019-09-04 | 2019-12-03 | 深圳市捷顺科技实业股份有限公司 | A kind of true-false detection method of license plate, device and equipment |
| CN111079751A (en) * | 2019-12-16 | 2020-04-28 | 深圳市芊熠智能硬件有限公司 | Method and device for identifying authenticity of license plate, computer equipment and storage medium |
| CN111860477A (en) * | 2020-01-22 | 2020-10-30 | 北京嘀嘀无限科技发展有限公司 | Vehicle information verification method and system, terminal, and computer-readable storage medium |
| CN111340045A (en) * | 2020-02-12 | 2020-06-26 | 浙江大华技术股份有限公司 | License plate number recognition method and device and storage medium |
| CN111797713A (en) * | 2020-06-16 | 2020-10-20 | 浙江大华技术股份有限公司 | License plate recognition method and photographing device |
Non-Patent Citations (2)
| Title |
|---|
| PROVID: Progressive and Multimodal Vehicle Reidentification for Large-Scale Urban Surveillance;X. Liu等;《IEEE Transactions on Multimedia》;20180331;第20卷(第3期);645-658 * |
| 变造车牌图像的特征分析与识别研究;宋迪;《中国优秀硕士学位论文全文数据库工程科技Ⅱ辑》;20200315(第3期);C034-725 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112580531A (en) | 2021-03-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102945366B (en) | A kind of method and device of recognition of face | |
| CN112580531B (en) | Identification detection method and system for true and false license plates | |
| CN105631418A (en) | People counting method and device | |
| CN111192461B (en) | License plate recognition method, server, parking charging method and system | |
| CN110517500B (en) | Man-vehicle association processing method and device | |
| CN110633289A (en) | License plate number determining method, server and parking management system | |
| CN110378254B (en) | Method and system for identifying vehicle damage image modification trace, electronic device and storage medium | |
| CN108694399A (en) | Licence plate recognition method, apparatus and system | |
| CN112633255B (en) | Target detection method, device and equipment | |
| CN111027534A (en) | Compact double-license-plate detection method and device | |
| CN113408364B (en) | Temporary license plate recognition method, system, device and storage medium | |
| CN114943750A (en) | Target tracking method and device and electronic equipment | |
| CN111163071A (en) | Unknown industrial protocol recognition engine | |
| CN111753642B (en) | Method and device for determining key frame | |
| CN109522796A (en) | Fault car reminding method and Related product in high speed | |
| CN110599796B (en) | Parking space state judgment method, server and parking management system | |
| CN111241918A (en) | Vehicle anti-tracking method and system based on face recognition | |
| CN112001280A (en) | Real-time online optimization face recognition system and method | |
| CN116503820A (en) | Road vehicle type based detection method and detection equipment | |
| CN109345824A (en) | The Lane determining method and Related product of fault car in highway | |
| CN112633168B (en) | Garbage truck and method and device for identifying garbage can overturning action of garbage truck | |
| CN115880632A (en) | Timeout stay detection method, monitoring device, computer-readable storage medium, and chip | |
| CN112633163B (en) | Detection method for realizing illegal operation vehicle detection based on machine learning algorithm | |
| CN112950961B (en) | Traffic flow statistical method, device, equipment and storage medium | |
| CN113283303A (en) | License plate recognition 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 |