CN108537223B - License plate detection method, system and equipment and storage medium - Google Patents

License plate detection method, system and equipment and storage medium Download PDF

Info

Publication number
CN108537223B
CN108537223B CN201810338018.0A CN201810338018A CN108537223B CN 108537223 B CN108537223 B CN 108537223B CN 201810338018 A CN201810338018 A CN 201810338018A CN 108537223 B CN108537223 B CN 108537223B
Authority
CN
China
Prior art keywords
license plate
character
target
test sample
features
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
Application number
CN201810338018.0A
Other languages
Chinese (zh)
Other versions
CN108537223A (en
Inventor
王涛
王静
程良伦
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Guangdong University of Technology
Original Assignee
Guangdong University of Technology
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Guangdong University of Technology filed Critical Guangdong University of Technology
Priority to CN201810338018.0A priority Critical patent/CN108537223B/en
Publication of CN108537223A publication Critical patent/CN108537223A/en
Application granted granted Critical
Publication of CN108537223B publication Critical patent/CN108537223B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/60—Type of objects
    • G06V20/62—Text, e.g. of license plates, overlay texts or captions on TV images
    • G06V20/63—Scene text, e.g. street names
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/50—Extraction of image or video features by performing operations within image blocks; by using histograms, e.g. histogram of oriented gradients [HoG]; by summing image-intensity values; Projection analysis
    • G06V10/507—Summing image-intensity values; Histogram projection analysis
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/56—Extraction of image or video features relating to colour
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/60—Type of objects
    • G06V20/62—Text, e.g. of license plates, overlay texts or captions on TV images
    • G06V20/625—License plates

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Data Mining & Analysis (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • General Engineering & Computer Science (AREA)
  • Artificial Intelligence (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Image Analysis (AREA)

Abstract

The invention discloses a license plate detection method, a system and equipment and a computer readable storage medium, wherein the method comprises the following steps: obtaining a test sample, and extracting edge characteristics, HOG characteristics and HSV characteristics of the test sample; calculating Euclidean distance between the test sample and each character in the training set according to the edge characteristics, and determining the character corresponding to the minimum value in all the Euclidean distances as a target character; acquiring the license plate type through a training set, and determining the target license plate type by utilizing a multi-classification logistic regression algorithm and the HOG characteristic; acquiring license plate color features through a training set, constructing an SVM (support vector machine) classifier according to the license plate color features, and inputting the HSV features into the SVM classifier to obtain target license plate color features; and determining the license plate detection result of the test sample according to the target characters, the target license plate type and the target license plate color characteristics. The license plate detection method provided by the invention improves the precision of license plate detection.

Description

License plate detection method, system and equipment and storage medium
Technical Field
The present invention relates to the field of image processing technologies, and in particular, to a license plate detection method, system and device, and a computer-readable storage medium.
Background
In recent years, license plate recognition devices have been widely used in parking lots, urban roads, and other areas to automatically capture and recognize license plates of vehicles. The license plate detection mode in the prior art mainly aims at single characteristic detection, and the detection precision is not high in complex environments such as too strong noise, blurring, half-shielding and similar license plates.
Therefore, how to improve the license plate detection accuracy is a problem to be solved by those skilled in the art.
Disclosure of Invention
The invention aims to provide a license plate detection method, a license plate detection system, license plate detection equipment and a computer readable storage medium, and the license plate detection precision is improved.
In order to achieve the above object, the present invention provides a license plate detection method, comprising:
obtaining a test sample, and extracting edge features, HOG features and HSV features of the test sample;
calculating Euclidean distance between the test sample and each character in the training set according to the edge characteristics, and determining the character corresponding to the minimum value in all the Euclidean distances as a target character;
acquiring the license plate type through the training set, and determining the target license plate type by using a multi-classification logistic regression algorithm and the HOG characteristic;
acquiring license plate color features through the training set, constructing an SVM (support vector machine) classifier according to the license plate color features, and inputting the HSV features into the SVM classifier to obtain target license plate color features;
and determining the license plate detection result of the test sample according to the target characters, the target license plate type and the target license plate color characteristics.
Calculating the Euclidean distance between the test sample and each character in the training set according to the edge features, and determining the character corresponding to the minimum value in all the Euclidean distances as a target character, wherein the method comprises the following steps:
calculating the Euclidean distance between the test sample and each character in the training set according to the edge characteristics, and normalizing all the Euclidean distances;
and calculating a first basic probability distribution of the Euclidean distance after the normalization processing, and determining a character corresponding to the maximum value in all the basic probabilities as the target character.
Determining the type of the target license plate by using a multi-classification logistic regression algorithm and the HOG characteristics, wherein the method comprises the following steps:
and calculating a second basic distribution probability of the license plate types by utilizing a multi-classification logistic regression algorithm according to the HOG characteristics, and determining the license plate type corresponding to the maximum value in all the basic probabilities as the target license plate type.
Inputting the HSV features into the SVM classifier to obtain a target license plate color type, wherein the method comprises the following steps of:
and calculating a third basic probability distribution of the license plate color type by using the SVM classifier according to the HSV characteristics, and determining the license plate color type corresponding to the maximum value in all the basic probabilities as the target license plate color type.
Wherein, still include:
and calculating the trust degree of the license plate detection result according to the first basic probability distribution, the second basic probability distribution and the third basic probability distribution and a preset fusion rule.
Wherein the first fundamental probability m1(Ai) The distribution is as follows:
Figure BDA0001629787690000021
wherein n isiFor the ith character in the training set, diThe Euclidean distance between the test sample and the ith character; alpha is 1 or 2;
the second fundamental probability distribution m2(Bi) Comprises the following steps:
Figure BDA0001629787690000022
wherein K is the total number of license plate types, and K belongs to (0, K)]W is a matrix of weights, WTIs the transpose of W, b is the bias of the multiple classification logistic regression algorithm, xiFor the ith test specimen, yiThe license plate type of the ith test sample is obtained;
the third fundamental probability m3(Ci) The distribution is as follows:
Figure BDA0001629787690000031
wherein n isiIs the ith vehicle color feature, xiFor the ith test specimen, yiThe color characteristic of the ith license plate.
Wherein the fusion rule is:
Figure BDA0001629787690000032
wherein the content of the first and second substances,
Figure BDA0001629787690000033
phi is an identification frame, and s is a set of evidences supporting license plate identification in the identification frame.
In order to achieve the above object, the present invention provides a license plate detection system, comprising:
the extraction module is used for obtaining a test sample and extracting the edge feature, the HOG feature and the HSV feature of the test sample;
the first determining module is used for calculating the Euclidean distance between the test sample and each character in the training set according to the edge characteristics, and determining the character corresponding to the minimum value in all the Euclidean distances as a target character;
the second determining module is used for acquiring the license plate type through the training set and determining the target license plate type by utilizing a multi-classification logistic regression algorithm and the HOG characteristic;
the third determination module is used for acquiring license plate color features through the training set, constructing an SVM classifier according to the license plate color features, and inputting the HSV features into the SVM classifier to obtain target license plate color features;
and the detection module is used for determining the license plate detection result of the test sample according to the target characters, the target license plate type and the target license plate color characteristics.
In order to achieve the above object, the present invention provides a license plate detecting apparatus, comprising:
a memory for storing a computer program;
and the processor is used for realizing the steps of the license plate detection method when executing the computer program.
To achieve the above object, the present invention provides a computer-readable storage medium having stored thereon a computer program, which, when being executed by a processor, implements the steps of the above license plate detection method.
According to the scheme, the license plate detection method provided by the invention comprises the following steps: obtaining a test sample, and extracting edge features, HOG features and HSV features of the test sample; calculating Euclidean distance between the test sample and each character in the training set according to the edge characteristics, and determining the character corresponding to the minimum value in all the Euclidean distances as a target character; acquiring the license plate type through the training set, and determining the target license plate type by using a multi-classification logistic regression algorithm and the HOG characteristic; acquiring license plate color features through the training set, constructing an SVM (support vector machine) classifier according to the license plate color features, and inputting the HSV features into the SVM classifier to obtain target license plate color features; and determining the license plate detection result of the test sample according to the target characters, the target license plate type and the target license plate color characteristics.
The license plate detection method provided by the invention extracts the edge characteristics, the HOG characteristics and the HSV characteristics of the test sample, adopts different classification modes for different characteristics in a pertinence manner to obtain the target characters, the type of the target license plate and the color characteristics of the target license plate of the test sample, and determines the final detection result. Compared with the scheme of detecting a single feature in a single classification mode in the prior art, the method combines multiple features and multiple classification methods to obtain a detection result, and the integration of multiple data improves the license plate detection precision of multiple viewpoints and shielding problems in a multi-lane traffic mode scene. The invention also discloses a license plate detection system and equipment and a computer readable storage medium, and the technical effects can be realized.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
FIG. 1 is a flowchart of a license plate detection method disclosed in an embodiment of the present invention;
FIG. 2 is a flow chart of another license plate detection method disclosed in the embodiment of the present invention;
FIG. 3 is a block diagram of a license plate detection system according to an embodiment of the present invention;
fig. 4 is a structural diagram of a license plate detection device disclosed in the embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The embodiment of the invention discloses a license plate detection method, which improves the license plate detection precision.
Referring to fig. 1, a flowchart of a license plate detection method disclosed in the embodiment of the present invention, as shown in fig. 1, includes:
s101: obtaining a test sample, and extracting edge features, HOG features and HSV features of the test sample;
in a specific implementation, the training sample is actually a vehicle image, where the test sample is acquired by a general camera, or other cameras such as high definition cameras and ultra high definition cameras, or of course, other devices capable of acquiring images may acquire the acquired images, and the method is not limited specifically herein. After the vehicle image is obtained, preprocessing operation needs to be performed on the image so as to extract edge features, HOG (Histogram of oriented gradients, chinese) features, and HSV (Hue, Saturation, Value, a color space created by a.r. smith in 1978 according to intuitive characteristics of colors, which is also called a hexagonal cone model), and the preprocessing operation herein includes conventional technical means in the field of image processing.
S102: calculating Euclidean distance between the test sample and each character in the training set according to the edge characteristics, and determining the character corresponding to the minimum value in all the Euclidean distances as a target character;
it can be understood that the edge feature is a license plate character in the vehicle picture. The test set contains a plurality of license plates, each license plate containing a plurality of characters, e.g., A, B, Beijing, Yue, etc. And representing the similarity between the license plate characters and each character in the test set by using the Euclidean distance, wherein the character with the highest similarity, namely the character corresponding to the smallest Euclidean distance, is the target character of the license plate characters.
S103: acquiring the license plate type through the training set, and determining the target license plate type by using a multi-classification logistic regression algorithm and the HOG characteristic;
in the specific implementation, the license plates in the test set are used for acquiring all kinds of license plate types, such as 8-bit character new energy license plates, armed police license plates, coach license plates, hong Kong and Macau license plates and the like. The HOG features depict local gradient amplitude and direction features of an image, and the HOG allows blocks to be overlapped with each other, so that the HOG features are insensitive to illumination change and small-amount deviation, can effectively depict edge features, and has good detection effect on car logos with large angles. The license plate types of the test samples are represented through the HOG characteristics, and the license plate types of the test samples are judged by using a multi-classification logistic regression algorithm, so that the multi-classification logistic regression is ensured to have the best performance, and the overfitting condition during the training of an LR classifier is reduced.
S104: acquiring license plate color features through the training set, constructing an SVM (support vector machine) classifier according to the license plate color features, and inputting the HSV features into the SVM classifier to obtain target license plate color features;
in a specific implementation, all kinds of license plate color characteristics, such as white, blue, yellow, etc., are acquired by using the license plates in the test set. The color features of the test sample are represented by HSV features, and an SVM (Support Vector Machine, Chinese full name, and a fast mode recognition method) classifier is used for judging which ground color the license plate of the test sample belongs to.
It should be noted that the execution processes of S102, S103, and S104 are not in sequence.
S105: and determining the license plate detection result of the test sample according to the target characters, the target license plate type and the target license plate color characteristics.
The target characters, the target license plate types and the target license plate color characteristics obtained in the steps form a license plate detection result of the test sample.
According to the license plate detection method provided by the embodiment of the invention, the edge characteristics, the HOG characteristics and the HSV characteristics of the test sample are extracted, different classification modes are adopted for different characteristics in a targeted manner, so that the target characters, the type of the target license plate and the color characteristics of the target license plate of the test sample are obtained, and the final detection result is determined. Compared with the scheme of detecting a single feature in a single classification mode in the prior art, the method combines multiple features and multiple classification methods to obtain a detection result, and the integration of multiple data improves the license plate detection precision of multiple viewpoints and shielding problems in a multi-lane traffic mode scene.
The embodiment of the invention discloses a license plate detection method, and compared with the previous embodiment, the technical scheme is further explained and optimized in the embodiment. Specifically, the method comprises the following steps:
referring to fig. 2, a flowchart of another license plate detection method provided in the embodiment of the present invention is shown in fig. 2, and includes:
s201: obtaining a test sample, and extracting edge features, HOG features and HSV features of the test sample;
s221: calculating the Euclidean distance between the test sample and each character in the training set according to the edge characteristics, and normalizing all the Euclidean distances;
s222: calculating a first basic probability distribution of the Euclidean distance after normalization processing, and determining characters corresponding to the maximum value in all the basic probabilities as target characters;
it will be appreciated that since the true value of the Euclidean distance is [0, ∞ ], normalization of the Euclidean distance is required, i.e. normalization
Figure BDA0001629787690000071
Normalized to (0,1), where j ∈ [1, i ]]。
The first fundamental probability m1(Ai) The distribution is as follows:
Figure BDA0001629787690000072
wherein n isiFor the ith character in the training set, diThe Euclidean distance between the test sample and the ith character; alpha is 1 or 2;
that is, if the euclidean distance is 0, the similarity is the maximum, and at this time, the confidence is 1, and if the euclidean distance is not 0, the confidence corresponding to the character is calculated according to the above formula.
S203: acquiring the license plate type through the training set, calculating a second basic distribution probability of the license plate type by utilizing a multi-classification logistic regression algorithm according to the HOG characteristics, and determining the license plate type corresponding to the maximum value in all the basic probabilities as a target license plate type;
the second fundamental probability distribution m2(Bi) Comprises the following steps:
Figure BDA0001629787690000073
wherein K is the total number of license plate types, and K belongs to (0, K)]W is a matrix of weights, WTIs the transpose of W, b is the bias of the multiple classification logistic regression algorithm, xiFor the ith test specimen, yiFor the ith test specimenThe license plate type of (1);
s204: acquiring license plate color features through the training set, calculating third basic probability distribution of the license plate color features according to the HSV features by utilizing the SVM classifier, and determining the license plate color features corresponding to the maximum value of all the basic probabilities as the target license plate color features;
the third fundamental probability m3(Ci) The distribution is as follows:
Figure BDA0001629787690000081
wherein n isiIs the ith vehicle color feature, xiFor the ith test specimen, yiThe color characteristic of the ith license plate.
In the specific implementation, the SVM classifier takes a Sigmod function as a connection function, the output range is 0-1, and the probability distribution of the license plate color feature can be calculated by the SVM classifier by using the formula.
S205: determining a license plate detection result of the test sample according to the target character, the target license plate type and the target license plate color characteristic;
s206: and calculating the trust degree of the license plate detection result according to the first basic probability distribution, the second basic probability distribution and the third basic probability distribution and a preset fusion rule.
On the recognition frame theta, there are three evidences A independent from each otheri、Bi、CiThe basic probability assignment function is m1(Ai)、m2(Bi)、m3(Ci) The person skilled in the art can select an appropriate fusion rule for fusion according to the actual situation, which is not specifically limited herein, and preferably:
Figure BDA0001629787690000082
wherein the content of the first and second substances,
Figure BDA0001629787690000083
phi is an identification frame, and s is a set of evidences supporting license plate identification in the identification frame.
In the following, a license plate detection system provided by an embodiment of the present invention is introduced, and a license plate detection system described below and a license plate detection method described above may be referred to each other.
Referring to fig. 3, a structural diagram of a license plate detection system provided in an embodiment of the present invention is shown in fig. 3, and includes:
an extraction module 301, configured to obtain a test sample, and extract an edge feature, an HOG feature, and an HSV feature of the test sample;
a first determining module 302, configured to calculate a euclidean distance between the test sample and each character in the training set according to the edge feature, and determine a character corresponding to a minimum value in all the euclidean distances as a target character;
a second determining module 303, configured to obtain a license plate type through the training set, and determine a target license plate type by using a multi-classification logistic regression algorithm and the HOG feature;
a third determining module 304, configured to obtain a license plate color feature through the training set, construct an SVM classifier according to the license plate color feature, and input the HSV feature into the SVM classifier to obtain a target license plate color feature;
the detection module 305 is configured to determine a license plate detection result of the test sample according to the target character, the target license plate type, and the target license plate color feature.
The license plate detection system provided by the embodiment of the invention extracts the edge characteristics, the HOG characteristics and the HSV characteristics of the test sample, adopts different classification modes for different characteristics in a pertinence manner to obtain the target characters, the type of the target license plate and the color characteristics of the target license plate of the test sample, and determines the final detection result. Compared with the scheme of detecting a single feature in a single classification mode in the prior art, the method combines multiple features and multiple classification methods to obtain a detection result, and the integration of multiple data improves the license plate detection precision of multiple viewpoints and shielding problems in a multi-lane traffic mode scene.
On the basis of the foregoing embodiment, as a preferred implementation, the first determining module 302 includes:
the normalization unit is used for calculating the Euclidean distance between the test sample and each character in the training set according to the edge characteristics and carrying out normalization processing on all the Euclidean distances;
and the determining unit is used for calculating a first basic probability distribution of the Euclidean distance after the normalization processing, and determining the character corresponding to the maximum value in all the basic probabilities as the target character.
On the basis of the foregoing embodiment, as a preferred embodiment, the second determining module 303 is a module that obtains the license plate type through the training set, calculates a second basic distribution probability of the license plate type by using a multi-classification logistic regression algorithm according to the HOG features, and determines the license plate type corresponding to the maximum value among all the basic probabilities as the target license plate type.
On the basis of the foregoing embodiment, as a preferred implementation, the third determining module 304 includes:
the construction unit is used for acquiring license plate color characteristics through the training set and constructing an SVM classifier according to the license plate color characteristics;
and the calculating unit is used for calculating a third basic probability distribution of the license plate color features by utilizing the SVM classifier according to the HSV features, and determining the license plate color feature corresponding to the maximum value in all the basic probabilities as the target license plate color feature.
In addition to the above embodiments, as a preferred embodiment, the method further includes:
and the calculation module is used for calculating the trust degree of the license plate detection result according to the first basic probability distribution, the second basic probability distribution and the third basic probability distribution and a preset fusion rule.
On the basis of the above-mentioned embodiment, as a preferred implementation, the first fundamental probability m1(Ai) The distribution is as follows:
Figure BDA0001629787690000101
wherein n isiFor the ith character in the training set, diThe Euclidean distance between the test sample and the ith character; alpha is 1 or 2;
the second fundamental probability distribution m2(Bi) Comprises the following steps:
Figure BDA0001629787690000102
wherein K is the total number of license plate types, and K belongs to (0, K)]W is a matrix of weights, WTIs the transpose of W, b is the bias of the multiple classification logistic regression algorithm, xiFor the ith test specimen, yiThe license plate type of the ith test sample is obtained;
the third fundamental probability m3(Ci) The distribution is as follows:
Figure BDA0001629787690000103
wherein n isiIs the ith vehicle color feature, xiFor the ith test specimen, yiThe color characteristic of the ith license plate.
On the basis of the above embodiment, as a preferred implementation, the fusion rule is:
Figure BDA0001629787690000104
wherein the content of the first and second substances,
Figure BDA0001629787690000105
phi is an identification frame, and s is a set of evidences supporting license plate identification in the identification frame.
The present application further provides a license plate detection device, referring to fig. 4, a structure diagram of a license plate detection device provided in an embodiment of the present invention, as shown in fig. 4, includes:
a memory 401 for storing a computer program;
the processor 402, when executing the computer program, may implement the steps provided by the above embodiments. Of course, the license plate detection device may further include various network interfaces, power supplies, and other components.
The license plate detection equipment provided by the embodiment of the invention extracts the edge characteristics, the HOG characteristics and the HSV characteristics of the test sample, adopts different classification modes for different characteristics in a pertinence manner to obtain the target characters, the type of the target license plate and the color characteristics of the target license plate of the test sample, and determines the final detection result. Compared with the scheme of adopting a single classification mode for detecting single characteristics in the prior art, the license plate detection precision is improved. Therefore, the license plate detection equipment provided by the embodiment of the invention obtains the detection result by combining a plurality of characteristics and a plurality of classification methods, and the integration of multiple data improves the license plate detection precision of multiple viewpoints and shielding problems in a multi-lane traffic mode scene.
The present application also provides a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, may implement the steps provided by the above-described embodiments. The storage medium may include: various media capable of storing program codes, such as a usb disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk, or an optical disk.
The embodiments in the present description are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
The embodiments are described in a progressive manner in the specification, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other. For the system disclosed by the embodiment, the description is relatively simple because the system corresponds to the method disclosed by the embodiment, and the relevant points can be referred to the method part for description. It should be noted that, for those skilled in the art, it is possible to make several improvements and modifications to the present application without departing from the principle of the present application, and such improvements and modifications also fall within the scope of the claims of the present application.
It is further noted that, in the present specification, relational terms such as first and second, and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.

Claims (9)

1. A license plate detection method is characterized by comprising the following steps:
obtaining a test sample, and extracting edge features, HOG features and HSV features of the test sample;
calculating Euclidean distances between the test sample and each character in the training set according to the edge features, and determining the character corresponding to the minimum value in all the Euclidean distances as a target character;
acquiring the license plate type through the training set, and determining the target license plate type by using a multi-classification logistic regression algorithm and the HOG characteristic;
acquiring license plate color features through the training set, constructing an SVM (support vector machine) classifier according to the license plate color features, and inputting the HSV features into the SVM classifier to obtain target license plate color features;
determining a license plate detection result of the test sample according to the target character, the target license plate type and the target license plate color characteristic;
calculating the Euclidean distance between the test sample and each character in the training set according to the edge features, and determining the character corresponding to the minimum value in all the Euclidean distances as a target character, wherein the method comprises the following steps:
calculating the Euclidean distance between the test sample and each character in the training set according to the edge characteristics, and normalizing all the Euclidean distances;
calculating a first basic probability distribution of the Euclidean distance after normalization processing, and determining a character corresponding to the maximum value in all the basic probabilities as the target character;
wherein the first fundamental probability m1(Ai) The distribution is as follows:
Figure FDA0003369466180000011
wherein n isiFor the ith character in the training set, p (i) is the fundamental probability of the ith character, diThe Euclidean distance between the test sample and the ith character; alpha is 1 or 2.
2. The license plate detection method of claim 1, wherein determining the target license plate type using a multi-classification logistic regression algorithm and the HOG features comprises:
and calculating a second basic probability distribution of the license plate types by utilizing a multi-classification logistic regression algorithm according to the HOG characteristics, and determining the license plate type corresponding to the maximum value in all the basic probabilities as the target license plate type.
3. The license plate detection method of claim 2, wherein inputting the HSV features into the SVM classifier to obtain target license plate color features comprises:
and calculating a third basic probability distribution of the license plate color features by utilizing the SVM classifier according to the HSV features, and determining the license plate color feature corresponding to the maximum value in all the basic probabilities as the target license plate color feature.
4. The license plate detection method of claim 3, further comprising:
and calculating the trust degree of the license plate detection result according to the first basic probability distribution, the second basic probability distribution and the third basic probability distribution and a preset fusion rule.
5. The license plate detection method of claim 4, wherein the second base probability distribution m2(Bi) Comprises the following steps:
Figure FDA0003369466180000021
wherein K is the total number of license plate types, and K belongs to (0, K)]，WkA matrix formed by weights corresponding to the kth license plate type,
Figure FDA0003369466180000025
is WkTranspose of (b)kA bias value, x, of the multi-classification logistic regression algorithm corresponding to the kth license plate typeiFor the ith test specimen, yiThe license plate type of the ith test sample is obtained;
the third fundamental probability m3(Ci) The distribution is as follows:
Figure FDA0003369466180000022
wherein n isiIs the ith vehicle color feature, x is the test sample, y is the license plate color feature, xiFor the ith test specimen, yiThe color characteristic of the ith license plate.
6. The license plate detection method of claim 5, wherein the fusion rule is:
Figure FDA0003369466180000023
wherein the content of the first and second substances,
Figure FDA0003369466180000024
phi is an identification frame, and s is a set of evidences supporting license plate identification in the identification frame.
7. A license plate detection system, comprising:
the extraction module is used for obtaining a test sample and extracting the edge feature, the HOG feature and the HSV feature of the test sample;
the first determining module is used for calculating the Euclidean distance between the test sample and each character in the training set according to the edge characteristics, and determining the character corresponding to the minimum value in all the Euclidean distances as a target character;
the second determining module is used for acquiring the license plate type through the training set and determining the target license plate type by utilizing a multi-classification logistic regression algorithm and the HOG characteristic;
the third determination module is used for acquiring license plate color features through the training set, constructing an SVM classifier according to the license plate color features, and inputting the HSV features into the SVM classifier to obtain target license plate color features;
the detection module is used for determining the license plate detection result of the test sample according to the target characters, the target license plate type and the target license plate color characteristics;
wherein the first determining module comprises:
the normalization unit is used for calculating the Euclidean distance between the test sample and each character in the training set according to the edge characteristics and carrying out normalization processing on all the Euclidean distances;
the determining unit is used for calculating a first basic probability distribution of the Euclidean distance after normalization processing, and determining a character corresponding to the maximum value in all the basic probabilities as the target character;
wherein the first fundamental probability m1(Ai) The distribution is as follows:
Figure FDA0003369466180000031
wherein n isiFor the ith character in the training set, p (i) is the fundamental probability of the ith character, diThe Euclidean distance between the test sample and the ith character; alpha is 1 or 2.
8. A license plate detection apparatus, comprising:
a memory for storing a computer program;
a processor for implementing the steps of the license plate detection method according to any one of claims 1 to 6 when executing the computer program.
9. A computer-readable storage medium, having stored thereon a computer program which, when being executed by a processor, carries out the steps of the license plate detection method according to any one of claims 1 to 6.
CN201810338018.0A 2018-04-16 2018-04-16 License plate detection method, system and equipment and storage medium Active CN108537223B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201810338018.0A CN108537223B (en) 2018-04-16 2018-04-16 License plate detection method, system and equipment and storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201810338018.0A CN108537223B (en) 2018-04-16 2018-04-16 License plate detection method, system and equipment and storage medium

Publications (2)

Publication Number Publication Date
CN108537223A CN108537223A (en) 2018-09-14
CN108537223B true CN108537223B (en) 2022-02-15

Family

ID=63481142

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201810338018.0A Active CN108537223B (en) 2018-04-16 2018-04-16 License plate detection method, system and equipment and storage medium

Country Status (1)

Country Link
CN (1) CN108537223B (en)

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111046891A (en) * 2018-10-11 2020-04-21 杭州海康威视数字技术股份有限公司 Training method of license plate recognition model, and license plate recognition method and device
CN111199225B (en) * 2018-11-20 2023-10-31 浙江宇视科技有限公司 License plate calibration method and device
CN111145365A (en) * 2019-12-17 2020-05-12 北京明略软件系统有限公司 Method, device, computer storage medium and terminal for realizing classification processing
CN111563504B (en) * 2020-07-16 2020-10-30 平安国际智慧城市科技股份有限公司 License plate recognition method and related equipment

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104156704A (en) * 2014-08-04 2014-11-19 胡艳艳 Novel license plate identification method and system
CN105224956A (en) * 2015-10-16 2016-01-06 深圳市捷顺科技实业股份有限公司 A kind of car plate classification based on multi-feature fusion recognition methods and system
CN107590500A (en) * 2017-07-20 2018-01-16 济南中维世纪科技有限公司 A kind of color recognizing for vehicle id method and device based on color projection classification

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102903110B (en) * 2012-09-29 2015-11-25 宁波大学 To the dividing method of image with deep image information
CN104463134B (en) * 2014-12-19 2019-02-12 深圳市捷顺科技实业股份有限公司 A kind of detection method of license plate and system
US9536315B2 (en) * 2015-01-13 2017-01-03 Xerox Corporation Annotation free license plate recognition method and system
CN105260735A (en) * 2015-10-10 2016-01-20 安徽清新互联信息科技有限公司 License number character segmentation method based on typical position
CN106650731B (en) * 2016-12-23 2020-02-21 中山大学 Robust license plate and vehicle logo recognition method

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104156704A (en) * 2014-08-04 2014-11-19 胡艳艳 Novel license plate identification method and system
CN105224956A (en) * 2015-10-16 2016-01-06 深圳市捷顺科技实业股份有限公司 A kind of car plate classification based on multi-feature fusion recognition methods and system
CN107590500A (en) * 2017-07-20 2018-01-16 济南中维世纪科技有限公司 A kind of color recognizing for vehicle id method and device based on color projection classification

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
License Plate Characters Segmentation Using Projection and Template Matching;Deng Hongyao 等;《2009 International Conference on Information Technology and Computer Science》;20090726;第534-537页 *
基于HSV模型和改进AdaBoost算法的车牌检测;王毅 等;《电子科技》;20150228;第28卷(第2期);第107-111页 *

Also Published As

Publication number Publication date
CN108537223A (en) 2018-09-14

Similar Documents

Publication Publication Date Title
CN109740478B (en) Vehicle detection and identification method, device, computer equipment and readable storage medium
CN108537223B (en) License plate detection method, system and equipment and storage medium
CN107679078B (en) Bayonet image vehicle rapid retrieval method and system based on deep learning
WO2020173022A1 (en) Vehicle violation identifying method, server and storage medium
CN109117879B (en) Image classification method, device and system
EP2701098B1 (en) Region refocusing for data-driven object localization
Wang et al. An effective method for plate number recognition
CN108268867B (en) License plate positioning method and device
CN108491498B (en) Bayonet image target searching method based on multi-feature detection
JP2014232533A (en) System and method for ocr output verification
JP2016134175A (en) Method and system for performing text-to-image queries with wildcards
CN112016605B (en) Target detection method based on corner alignment and boundary matching of bounding box
Türkyılmaz et al. License plate recognition system using artificial neural networks
CN105404886A (en) Feature model generating method and feature model generating device
CN109740417B (en) Invoice type identification method, invoice type identification device, storage medium and computer equipment
CN107194393B (en) Method and device for detecting temporary license plate
CN110309810B (en) Pedestrian re-identification method based on batch center similarity
CN108323209B (en) Information processing method, system, cloud processing device and computer storage medium
CN106778742B (en) Car logo detection method based on Gabor filter background texture suppression
CN112613344B (en) Vehicle track occupation detection method, device, computer equipment and readable storage medium
CN110826415A (en) Method and device for re-identifying vehicles in scene image
CN112784754A (en) Vehicle re-identification method, device, equipment and storage medium
CN111860219A (en) High-speed road occupation judging method and device and electronic equipment
Hoang et al. Scalable histogram of oriented gradients for multi-size car detection
CN106778765B (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