CN108710857B - Method and device for identifying people and vehicles based on infrared supplementary lighting - Google Patents
Method and device for identifying people and vehicles based on infrared supplementary lighting Download PDFInfo
- Publication number
- CN108710857B CN108710857B CN201810495514.7A CN201810495514A CN108710857B CN 108710857 B CN108710857 B CN 108710857B CN 201810495514 A CN201810495514 A CN 201810495514A CN 108710857 B CN108710857 B CN 108710857B
- Authority
- CN
- China
- Prior art keywords
- face
- area
- license plate
- vehicle
- region
- 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
Images
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
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/161—Detection; Localisation; Normalisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
-
- 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
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/08—Detecting or categorising vehicles
Abstract
The invention provides a method for identifying a person and a vehicle based on infrared supplementary lighting, which comprises the following steps: acquiring a scene image; calculating the brightness mean value of the scene image, if the brightness mean value is smaller than a first threshold value, starting infrared light supplement, and acquiring the scene image again; carrying out vehicle detection and license plate detection on the scene image to obtain a vehicle area and a license plate area; if the license plate area exists, performing license plate identification on the license plate area to obtain a license plate identification result; if the license plate area does not exist, vehicle type recognition is carried out on the vehicle area, and a vehicle type recognition result is obtained; setting a vehicle window area, carrying out face detection on the vehicle window area to obtain a face area, identifying the face area and obtaining a face identification result; and outputting a license plate or vehicle type recognition result and a face recognition result. Compared with the existing human-vehicle identification technology, the identification rate can be improved through infrared supplementary lighting, and the human-vehicle identification technology has no influence on human eyes.
Description
Technical Field
The invention relates to image processing, video monitoring and intelligent parking, in particular to a human-vehicle identification method and a human-vehicle identification device.
Background
With the development of economic technology and the increasing of motor vehicles, the traditional manual management mode cannot meet the actual requirement more and more. Meanwhile, with the development of science and technology, intelligent traffic management systems are becoming mature day by day, and gradually replace the traditional manual management mode. The license plate recognition technology is the basis for realizing intelligent traffic management and is paid more and more attention.
In recent years, in order to improve the safety of parking lots, more and more parking lot management requires integration of human and vehicle identification. However, in an actual scene, night illumination is low, and recognition of a human face and a vehicle is affected. The current people's car identification scheme uses gaseous flashing light filling at night generally, and gaseous flashing light filling that explodes not only can influence driver's people's eye, still can produce the problem of light pollution.
In summary, there is an urgent need to provide a human-vehicle identification method that can solve the low illumination effect, and does not affect human eyes and generate light pollution.
Disclosure of Invention
In view of the above, the main purpose of the present invention is to realize human-vehicle recognition with high recognition rate.
To achieve the above object, according to a first aspect of the present invention, there is provided a method for identifying a human and a vehicle based on infrared supplementary lighting, the method comprising:
the method comprises the following steps of firstly, acquiring a scene image;
the second step, calculating the brightness mean value of the scene image, if the brightness mean value is smaller than a first threshold value, starting the infrared light supplement, and acquiring the scene image again;
thirdly, carrying out vehicle detection and license plate detection on the scene image to acquire a vehicle region and a license plate region;
step four, if a license plate area exists, license plate identification is carried out on the license plate area, and a license plate identification result is obtained; if the license plate area does not exist, vehicle type recognition is carried out on the vehicle area, and a vehicle type recognition result is obtained;
step five, setting a vehicle window area, carrying out face detection on the vehicle window area to obtain a face area, identifying the face area and obtaining a face identification result;
and a sixth step of outputting a license plate or vehicle type recognition result and a face recognition result.
Further, the fifth step includes:
a vehicle window area setting step, namely setting a vehicle window area according to the position of the license plate area; if the license plate region does not exist, setting a vehicle window region according to the position of the vehicle region;
a face detection step, namely acquiring a face area from a vehicle window area by adopting a face detection method;
and a face recognition step, comparing the face area with face data logged in a database by adopting a face recognition method, if the comparison is consistent, taking the user information in the database as the face information of the face area, if the comparison is inconsistent, recording the user information as a visitor, and compiling visitor number in the database and storing the visitor face information.
Further, the window area setting step includes:
and a vehicle window setting step based on the license plate, if the license plate region exists, acquiring the left boundary x ═ pl, the right boundary x ═ pr, the upper boundary y ═ pt and the lower boundary y ═ pb of the license plate region, and setting the left boundary of the vehicle window region as
The right boundary is
The upper boundary is
The lower boundary is
WpThe width of a license plate area is W, the width of a scene image is W, and lambda 3 is less than lambda 2;
and based on the vehicle window setting step, if no license plate region exists, acquiring the upper partial region of the vehicle region as a vehicle window region.
Further, the face recognition step may include: selecting a marked face sample image, training a convolutional neural network, and acquiring a trained face network identification model; and comparing the face area with the face data of the user logged in the database by adopting a face network identification model, if the comparison is consistent, taking the user information in the database as the face information of the face area, if the comparison is inconsistent, recording the face information as the visitor, and compiling visitor number in the database and storing the face information of the visitor.
The marked face sample image comprises: the labeled and classified face images with different angles, different brightness and different postures under the non-infrared light supplement lamp, and the labeled and classified face images with different angles and different postures under the infrared light supplement lamp.
Further, the face recognition step may further include:
a first network recognition model training step, namely selecting a marked general face sample image, training a first convolution neural network, and acquiring a trained first network recognition model;
a second network identification model classification step, namely selecting a marked infrared human face sample image, training a second convolutional neural network, and acquiring a trained second network identification model;
a face region identification step, wherein when the infrared light supplement is not started, the face region is compared with the face data of the user logged in the database by adopting a first network identification model, and when the infrared light supplement is started, the face region is compared with the face data of the user logged in the database by adopting a second network identification model; and if the comparison is consistent, the user information in the database is used as the face information of the face area, if the comparison is inconsistent, the visitor is recorded, the visitor number is compiled in the database, and the face information of the visitor is stored.
The marked generic face sample image includes: and the labeled and classified face images with different angles, different brightness and different postures under the non-infrared light supplementing lamp. The marked infrared face sample image comprises: and the labeled and classified face images at different angles and different postures under the infrared light supplement lamp.
According to another aspect of the present invention, there is provided a human-vehicle recognition device based on infrared supplementary lighting, the device comprising:
the scene image acquisition module is used for acquiring a scene image;
the infrared light supplement control module is used for calculating the brightness mean value of the scene image, starting infrared light supplement if the brightness mean value is smaller than a first threshold value, and acquiring the scene image again;
the vehicle license plate detection module is used for carrying out vehicle detection and license plate detection on the scene image to acquire a vehicle area and a license plate area;
the license plate and vehicle type recognition module is used for recognizing the license plate of the license plate area if the license plate area exists, and acquiring a license plate recognition result; if the license plate area does not exist, vehicle type recognition is carried out on the vehicle area, and a vehicle type recognition result is obtained;
the vehicle window area human face recognition module is used for setting a vehicle window area, carrying out human face detection on the vehicle window area, acquiring a human face area, recognizing the human face area and acquiring a human face recognition result;
and the human-vehicle recognition result output module is used for outputting a license plate or vehicle type recognition result and a face recognition result.
Further, the vehicle window area face recognition module comprises:
the vehicle window area setting module is used for setting a vehicle window area according to the position of the license plate area; if the license plate region does not exist, setting a vehicle window region according to the position of the vehicle region;
the face detection module is used for acquiring a face area from the car window area by adopting a face detection method;
and the face recognition module is used for comparing the face area with face data logged in the database by adopting a face recognition method, if the comparison is consistent, the user information in the database is used as the face information of the face area, if the comparison is inconsistent, the visitor is recorded as the visitor, and the visitor number is compiled in the database and the face information of the visitor is stored.
Further, the window area setting module includes:
the vehicle window setting module based on the license plate is used for acquiring the left boundary x ═ pl, the right boundary x ═ pr, the upper boundary y ═ pt and the lower boundary y ═ pb of the license plate area if the license plate area exists, and setting the left boundary of the vehicle window area as
The right boundary is
The upper boundary is
The lower boundary is
WpThe width of a license plate area is W, the width of a scene image is W, and lambda 3 is less than lambda 2;
and the vehicle window setting module is used for acquiring the upper part area of the vehicle area as a vehicle window area if the license plate area does not exist.
Further, the face recognition module may include: the face recognition model is used for selecting a marked face sample image, training the convolutional neural network and obtaining a trained face network recognition model; the system comprises a face network identification model, a database and a visitor, wherein the face network identification model is used for comparing a face area with face data of a user logged in the database, if the comparison is consistent, the user information in the database is used as the face information of the face area, if the comparison is inconsistent, the visitor is recorded, and the visitor number is compiled in the database and the face information of the visitor is stored.
The marked face sample image comprises: the labeled and classified face images with different angles, different brightness and different postures under the non-infrared light supplement lamp, and the labeled and classified face images with different angles and different postures under the infrared light supplement lamp.
Further, the face recognition module may further include:
the first network recognition model training module is used for selecting a marked general face sample image, training a first convolutional neural network and acquiring a trained first network recognition model;
the second network recognition model classification module is used for selecting the marked infrared human face sample image, training a second convolutional neural network and acquiring a trained second network recognition model;
the face region identification module is used for comparing the face region with face data of a user logged in a database by adopting a first network identification model when infrared light supplement is not started, and comparing the face region with the face data of the user logged in the database by adopting a second network identification model when the infrared light supplement is started; and if the comparison is consistent, the user information in the database is used as the face information of the face area, if the comparison is inconsistent, the visitor is recorded, the visitor number is compiled in the database, and the face information of the visitor is stored.
The marked generic face sample image includes: and the labeled and classified face images with different angles, different brightness and different postures under the non-infrared light supplementing lamp. The marked infrared face sample image comprises: and the labeled and classified face images at different angles and different postures under the infrared light supplement lamp.
Compared with the existing human-vehicle identification technology, the human-vehicle identification method and device based on infrared supplementary lighting can improve the accuracy of human-vehicle identification at night by performing infrared supplementary lighting according to the image quality, and meanwhile, due to the adoption of the infrared supplementary lighting, the human-vehicle identification method and device has no influence on human eyes and does not cause light pollution.
Drawings
Fig. 1 shows a flowchart of a method for identifying a human and a vehicle based on infrared supplementary lighting according to the present invention.
Fig. 2 is a block diagram of a human-vehicle recognition device based on infrared supplementary lighting according to the present invention.
Detailed Description
To further clarify the structure, characteristics and other objects of the present invention, those skilled in the art will now describe in detail the preferred embodiments of the present invention with reference to the attached drawings, which are provided for the purpose of describing the technical solutions of the present invention only and are not intended to limit the present invention.
Fig. 1 is a flowchart of a method for identifying a human and a vehicle based on infrared supplementary lighting according to the present invention. As shown in fig. 1, the method for identifying a human and a vehicle based on infrared supplementary lighting according to the present invention includes:
a first step S1 of acquiring a scene image;
a second step S2, calculating the brightness mean value of the scene image, if the brightness mean value is smaller than a first threshold value, starting infrared supplementary lighting, and re-collecting the scene image;
step S3, carrying out vehicle detection and license plate detection on the scene image to obtain a vehicle area and a license plate area;
a fourth step S4, if a license plate region exists, carrying out license plate recognition on the license plate region to obtain a license plate recognition result; if the license plate area does not exist, vehicle type recognition is carried out on the vehicle area, and a vehicle type recognition result is obtained;
a fifth step S5, setting a window area, performing face detection on the window area to obtain a face area, and recognizing the face area to obtain a face recognition result;
and a sixth step S6 of outputting a license plate or vehicle type recognition result and a face recognition result.
The first step S1 may be implemented by an existing image capturing device or apparatus to capture an image of a scene. In an embodiment, a gun camera is installed at an entrance and an exit of a parking lot for capturing a scene image at the entrance and the exit.
Further, the value range of the first threshold is 90-110. The infrared light supplement is realized through the existing infrared light supplement lamp. In an embodiment, the second step S2 is: and calculating the average value of the brightness values of all the pixel points in the scene image, and if the average value of the brightness values of all the pixel points is less than 100, controlling the infrared light supplement lamp to be started, and simultaneously, re-collecting the scene image.
Further, the third step S3 includes: a vehicle detection step S31, wherein a vehicle detection method is adopted to carry out vehicle detection on the scene image and obtain a vehicle detection area; and a license plate detection step S32, wherein a license plate detection method is adopted to carry out license plate detection on the scene image and acquire a license plate detection area. The vehicle detection method and the license plate detection method can be realized by the existing vehicle detection technology and the license plate detection technology.
In the fourth step S4, the existing license plate recognition technology may be used to realize license plate recognition, and the existing vehicle type recognition technology may be used to realize vehicle type recognition.
Further, the fifth step S5 includes:
a window area setting step S51, setting a window area according to the position of the license plate area; if the license plate region does not exist, setting a vehicle window region according to the position of the vehicle region;
a human face detection step S52, wherein a human face detection method is adopted to obtain a human face area from the car window area;
and a face recognition step S53, comparing the face region with face data logged in the database by adopting a face recognition method, if the comparison is consistent, taking the user information in the database as the face information of the face region, if the comparison is inconsistent, recording the user information as a visitor, and compiling a visitor number in the database and storing the face information of the visitor.
Further, the window area setting step S51 includes:
and a vehicle window setting step S511 based on the license plate, if the license plate region exists, acquiring a left boundary x ═ pl, a right boundary x ═ pr, an upper boundary y ═ pt and a lower boundary y ═ pb of the license plate region, and setting a left boundary of the vehicle window region as
The right boundary is
The upper boundary is
The lower boundary is
WpThe width of a license plate area is W, the width of a scene image is W, and lambda 3 is less than lambda 2;
and a step S512 of setting a vehicle window, wherein if the license plate region does not exist, the upper part region of the vehicle region is acquired as a vehicle window region.
Further, λ 1 ∈ [1.2,1.8], λ 2 ∈ [4.2,4.8], λ 3 ∈ [0.3,0.8 ]. In the examples, λ 1 is selected to be 1.5, λ 2 is selected to be 4.5, and λ 3 is selected to be 0.5.
In an embodiment, the vehicle window setting step S512 is: if the license plate area does not exist, the upper part of the vehicle area is obtainedSetting the left boundary, the right boundary and the upper boundary of the window area as the left boundary, the right boundary and the upper boundary of the vehicle area respectively, and setting the lower boundary of the window area as the lower boundary of the vehicle area
The face recognition method in the face recognition step S53 may be an existing face recognition method or a face comparison method. Further, the face recognition method adopts a face recognition method based on a convolutional neural network.
Further, the face recognition step S53 may include: selecting a marked face sample image, training a convolutional neural network, and acquiring a trained face network identification model; and comparing the face area with the face data of the user logged in the database by adopting a face network identification model, if the comparison is consistent, taking the user information in the database as the face information of the face area, if the comparison is inconsistent, recording the face information as the visitor, and compiling visitor number in the database and storing the face information of the visitor.
The marked face sample image comprises: the labeled and classified face images with different angles, different brightness and different postures under the non-infrared light supplement lamp, and the labeled and classified face images with different angles and different postures under the infrared light supplement lamp.
Further, the face recognition step S53 may further include:
a first network identification model training step S531, selecting a marked general face sample image, training a first convolution neural network, and acquiring a trained first network identification model;
a second network recognition model classification step S532, selecting the marked infrared human face sample image, training a second convolutional neural network, and acquiring a trained second network recognition model;
a face region identification step S533, when the infrared light compensation is not started, comparing the face region with the face data of the user logged in the database by using a first network identification model, and when the infrared light compensation is started, comparing the face region with the face data of the user logged in the database by using a second network identification model; and if the comparison is consistent, the user information in the database is used as the face information of the face area, if the comparison is inconsistent, the visitor is recorded, the visitor number is compiled in the database, and the face information of the visitor is stored.
The marked generic face sample image includes: and the labeled and classified face images with different angles, different brightness and different postures under the non-infrared light supplementing lamp. The marked infrared face sample image comprises: and the labeled and classified face images at different angles and different postures under the infrared light supplement lamp.
The convolutional neural network, the first convolutional neural network, and the second convolutional neural network may be existing convolutional neural networks.
Further, the sixth step S6 includes: if the license plate recognition result exists, outputting the license plate recognition result and the face recognition result; and if the license plate recognition result does not exist, outputting a vehicle type recognition result and a face recognition result.
Fig. 2 is a block diagram of the human-vehicle identification device based on infrared supplementary lighting according to the present invention. As shown in fig. 2, the device for identifying a person and a vehicle based on infrared supplementary lighting according to the present invention includes:
a scene image acquisition module 1 for acquiring a scene image;
the infrared light supplement control module 2 is used for calculating a brightness mean value of the scene image, starting infrared light supplement if the brightness mean value is smaller than a first threshold value, and acquiring the scene image again;
the vehicle license plate detection module 3 is used for performing vehicle detection and license plate detection on the scene image to acquire a vehicle area and a license plate area;
the license plate and vehicle type recognition module 4 is used for recognizing the license plate of the license plate area if the license plate area exists, and acquiring a license plate recognition result; if the license plate area does not exist, vehicle type recognition is carried out on the vehicle area, and a vehicle type recognition result is obtained;
the vehicle window area human face recognition module 5 is used for setting a vehicle window area, performing human face detection on the vehicle window area, acquiring a human face area, recognizing the human face area and acquiring a human face recognition result;
and the human-vehicle recognition result output module 6 is used for outputting a license plate or vehicle type recognition result and a face recognition result.
Further, the value range of the first threshold is 90-110. The scene image acquisition module 1 is an existing image acquisition device or apparatus.
Further, the vehicle license plate detection module 3 includes: the vehicle detection module 31 is configured to perform vehicle detection on the scene image by using a vehicle detection method, and acquire a vehicle detection area; and the license plate detection module 32 is used for detecting the license plate of the scene image by adopting a license plate detection method and acquiring a license plate detection area.
The license plate and vehicle type recognition module 4 can realize license plate recognition by adopting the existing license plate recognition equipment or device, and realize vehicle type recognition by adopting the existing vehicle type recognition equipment or device.
Further, the window area face recognition module 5 includes:
the vehicle window area setting module 51 is used for setting a vehicle window area according to the position of the license plate area; if the license plate region does not exist, setting a vehicle window region according to the position of the vehicle region;
the face detection module 52 is configured to acquire a face region from the window region by using a face detection method;
and the face recognition module 53 is configured to compare the face region with face data logged in the database by using a face recognition method, if the comparison is consistent, use user information in the database as face information of the face region, and if the comparison is inconsistent, record the face information as a visitor, and code a visitor number in the database and store the face information of the visitor.
Further, the window area setting module 51 includes:
the vehicle window setting module 511 is configured to, if a license plate region exists, obtain a left boundary x ═ pl, a right boundary x ═ pr, an upper boundary y ═ pt, and a lower boundary y ═ pb of the license plate region, and set a vehicle window regionThe left boundary of the domain is
The right boundary is
The upper boundary is
The lower boundary is
WpThe width of a license plate area is W, the width of a scene image is W, and lambda 3 is less than lambda 2;
and a vehicle-based window setting module 512, configured to acquire an upper partial area of the vehicle area as a window area if the license plate area does not exist.
Further, λ 1 ∈ [1.2,1.8], λ 2 ∈ [4.2,4.8], λ 3 ∈ [0.3,0.8 ].
Further, the face recognition module 53 may include: the face recognition model is used for selecting a marked face sample image, training the convolutional neural network and obtaining a trained face network recognition model; the system comprises a face network identification model, a database and a visitor, wherein the face network identification model is used for comparing a face area with face data of a user logged in the database, if the comparison is consistent, the user information in the database is used as the face information of the face area, if the comparison is inconsistent, the visitor is recorded, and the visitor number is compiled in the database and the face information of the visitor is stored.
The marked face sample image comprises: the labeled and classified face images with different angles, different brightness and different postures under the non-infrared light supplement lamp, and the labeled and classified face images with different angles and different postures under the infrared light supplement lamp.
Further, the face recognition module 53 may further include:
the first network recognition model training module 531 is configured to select a marked general face sample image, train a first convolutional neural network, and obtain a trained first network recognition model;
the second network recognition model classification module 532 is used for selecting the marked infrared face sample image, training the second convolutional neural network and acquiring a trained second network recognition model;
a face region identification module 533, configured to compare the face region with user face data logged in the database by using a first network identification model when infrared light compensation is not turned on, and compare the face region with user face data logged in the database by using a second network identification model when infrared light compensation is turned on; and if the comparison is consistent, the user information in the database is used as the face information of the face area, if the comparison is inconsistent, the visitor is recorded, the visitor number is compiled in the database, and the face information of the visitor is stored.
The marked generic face sample image includes: and the labeled and classified face images with different angles, different brightness and different postures under the non-infrared light supplementing lamp. The marked infrared face sample image comprises: and the labeled and classified face images at different angles and different postures under the infrared light supplement lamp.
Further, the human-vehicle recognition result output module 6 includes: the face recognition module is used for outputting a license plate recognition result and a face recognition result if the license plate recognition result exists; and if the license plate recognition result does not exist, outputting a vehicle type recognition result and a face recognition result.
Compared with the existing human-vehicle identification technology, the human-vehicle identification method and device based on infrared supplementary lighting can improve the accuracy of human-vehicle identification at night by performing infrared supplementary lighting according to the image quality, and meanwhile, due to the adoption of the infrared supplementary lighting, the human-vehicle identification method and device has no influence on human eyes and does not cause light pollution.
While the foregoing is directed to the preferred embodiment of the present invention, and is not intended to limit the scope of the invention, it will be understood that the invention is not limited to the embodiments described herein, which are described to assist those skilled in the art in practicing the invention. Further modifications and improvements may readily occur to those skilled in the art without departing from the spirit and scope of the invention, and it is intended that the invention be limited only by the terms and scope of the appended claims, as including all alternatives and equivalents which may be included within the spirit and scope of the invention as defined by the appended claims.
Claims (6)
1. The method for identifying the people and the vehicles based on the infrared supplementary lighting is characterized by comprising the following steps:
the method comprises the following steps of firstly, acquiring a scene image;
the second step, calculating the brightness mean value of the scene image, if the brightness mean value is smaller than a first threshold value, starting the infrared light supplement, and acquiring the scene image again;
thirdly, carrying out vehicle detection and license plate detection on the scene image to acquire a vehicle region and a license plate region;
step four, if a license plate area exists, license plate identification is carried out on the license plate area, and a license plate identification result is obtained; if the license plate area does not exist, vehicle type recognition is carried out on the vehicle area, and a vehicle type recognition result is obtained;
step five, setting a vehicle window area, carrying out face detection on the vehicle window area to obtain a face area, identifying the face area and obtaining a face identification result;
a sixth step of outputting a license plate or vehicle type recognition result and a face recognition result;
further, the fifth step includes:
a vehicle window area setting step, namely setting a vehicle window area according to the position of the license plate area; if the license plate region does not exist, setting a vehicle window region according to the position of the vehicle region;
a face detection step, namely acquiring a face area from a vehicle window area by adopting a face detection method;
a face recognition step, comparing a face region with face data logged in a database by adopting a face recognition method, if the comparison is consistent, taking user information in the database as face information of the face region, if the comparison is inconsistent, recording the face information as a visitor, and compiling visitor number in the database and storing the face information of the visitor;
furthermore, the face recognition method is a face recognition method based on a convolutional neural network;
further, the face recognition step may include: selecting a marked face sample image, training a convolutional neural network, and acquiring a trained face network identification model; comparing the face area with user face data logged in a database by adopting a face network recognition model, if the comparison is consistent, taking the user information in the database as the face information of the face area, if the comparison is inconsistent, recording the user information as a visitor, and compiling visitor number in the database and storing the visitor face information;
wherein the labeled face sample image comprises: the marked and classified face images with different angles, different brightness and different postures under the non-infrared light supplement lamp, and the marked and classified face images with different angles and different postures under the infrared light supplement lamp;
further, the face recognition step in the fifth step may further include:
a first network recognition model training step, namely selecting a marked general face sample image, training a first convolution neural network, and acquiring a trained first network recognition model;
a second network identification model classification step, namely selecting a marked infrared human face sample image, training a second convolutional neural network, and acquiring a trained second network identification model;
a face region identification step, wherein when the infrared light supplement is not started, the face region is compared with the face data of the user logged in the database by adopting a first network identification model, and when the infrared light supplement is started, the face region is compared with the face data of the user logged in the database by adopting a second network identification model; if the comparison is consistent, the user information in the database is used as the face information of the face area, if the comparison is inconsistent, the visitor is recorded, the visitor number is compiled in the database, and the face information of the visitor is stored;
wherein the marked generic face sample image comprises: the labeled and classified face images with different angles, different brightness and different postures under the non-infrared light supplementing lamp; the marked infrared face sample image comprises:
and the labeled and classified face images at different angles and different postures under the infrared light supplement lamp.
2. The method of claim 1, wherein the first threshold value ranges from 90 to 110.
3. The method of claim 1, wherein the window region setting step comprises:
and a vehicle window setting step based on the license plate, if the license plate region exists, acquiring the left boundary x ═ pl, the right boundary x ═ pr, the upper boundary y ═ pt and the lower boundary y ═ pb of the license plate region, and setting the left boundary of the vehicle window region as
The right boundary is
The upper boundary is
The lower boundary is
WpThe width of a license plate area is W, the width of a scene image is W, and lambda 3 is less than lambda 2;
and based on the vehicle window setting step, if no license plate region exists, acquiring the upper partial region of the vehicle region as a vehicle window region.
4. The method of claim 3, wherein the vehicle-based window setting step comprises: if the license plate region does not exist, acquiring the upper boundary y ═ vt and the lower boundary y ═ vb of the vehicle region, setting the left boundary, the right boundary and the upper boundary of the vehicle window region as the left boundary, the right boundary and the upper boundary of the vehicle region respectively, and setting the vehicle window regionThe lower boundary is
。
5. The method of claim 3, wherein λ 1 ∈ [1.2,1.8], λ 2 ∈ [4.2,4.8], λ 3 ∈ [0.3,0.8 ].
6. People's car recognition device based on infrared light filling, its characterized in that, the device includes:
the scene image acquisition module is used for acquiring a scene image;
the infrared light supplement control module is used for calculating the brightness mean value of the scene image, starting infrared light supplement if the brightness mean value is smaller than a first threshold value, and acquiring the scene image again;
the vehicle license plate detection module is used for carrying out vehicle detection and license plate detection on the scene image to acquire a vehicle area and a license plate area;
the license plate and vehicle type recognition module is used for recognizing the license plate of the license plate area if the license plate area exists, and acquiring a license plate recognition result; if the license plate area does not exist, vehicle type recognition is carried out on the vehicle area, and a vehicle type recognition result is obtained;
the vehicle window area human face recognition module is used for setting a vehicle window area, carrying out human face detection on the vehicle window area, acquiring a human face area, recognizing the human face area and acquiring a human face recognition result;
the human-vehicle recognition result output module is used for outputting a license plate or vehicle type recognition result and a human face recognition result;
further, the vehicle window area face recognition module comprises:
the vehicle window area setting module is used for setting a vehicle window area according to the position of the license plate area; if the license plate region does not exist, setting a vehicle window region according to the position of the vehicle region;
the face detection module is used for acquiring a face area from the car window area by adopting a face detection method;
the face recognition module is used for comparing the face area with face data logged in a database by adopting a face recognition method, if the comparison is consistent, the user information in the database is used as the face information of the face area, if the comparison is inconsistent, the visitor is recorded as the visitor, and the visitor number is compiled in the database and the face information of the visitor is stored;
further, the face recognition module may include: the face recognition model is used for selecting a marked face sample image, training the convolutional neural network and obtaining a trained face network recognition model; the system comprises a face network identification model, a database and a visitor, wherein the face network identification model is used for comparing a face area with face data of a user logged in the database, if the comparison is consistent, the user information in the database is used as the face information of the face area, if the comparison is inconsistent, the visitor is recorded as the visitor, and the visitor number is compiled in the database and the face information of the visitor is stored;
wherein the labeled face sample image comprises: the marked and classified face images with different angles, different brightness and different postures under the non-infrared light supplement lamp, and the marked and classified face images with different angles and different postures under the infrared light supplement lamp;
further, the face recognition module may further include:
the first network recognition model training module is used for selecting a marked general face sample image, training a first convolution neural network and obtaining a trained first network recognition model;
the second network recognition model classification module is used for selecting the marked infrared human face sample image, training a second convolutional neural network and acquiring a trained second network recognition model;
the face area recognition module is used for comparing the face area with the face data of the user logged in the database by adopting a first network recognition model when the infrared light supplement is not started, and comparing the face area with the face data of the user logged in the database by adopting a second network recognition model when the infrared light supplement is started; if the comparison is consistent, the user information in the database is used as the face information of the face area, if the comparison is inconsistent, the visitor is recorded, the visitor number is compiled in the database, and the face information of the visitor is stored;
wherein the marked generic face sample image comprises: the labeled and classified face images with different angles, different brightness and different postures under the non-infrared light supplementing lamp; the marked infrared face sample image comprises: and the labeled and classified face images at different angles and different postures under the infrared light supplement lamp.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810495514.7A CN108710857B (en) | 2018-05-22 | 2018-05-22 | Method and device for identifying people and vehicles based on infrared supplementary lighting |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810495514.7A CN108710857B (en) | 2018-05-22 | 2018-05-22 | Method and device for identifying people and vehicles based on infrared supplementary lighting |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108710857A CN108710857A (en) | 2018-10-26 |
| CN108710857B true CN108710857B (en) | 2022-05-17 |
Family
ID=63868389
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810495514.7A Active CN108710857B (en) | 2018-05-22 | 2018-05-22 | Method and device for identifying people and vehicles based on infrared supplementary lighting |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108710857B (en) |
Families Citing this family (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2020057869A (en) * | 2018-09-28 | 2020-04-09 | パナソニックｉ−ＰＲＯセンシングソリューションズ株式会社 | Imaging apparatus |
| CN109658527A (en) * | 2018-12-14 | 2019-04-19 | 广东安居宝数码科技股份有限公司 | Vehicle parking charge management method, device, equipment and storage medium |
| CN109766854A (en) * | 2019-01-15 | 2019-05-17 | 济南浪潮高新科技投资发展有限公司 | A kind of robust human face recognizer based on two stages complementary networks |
| CN109961026A (en) * | 2019-03-12 | 2019-07-02 | 天津瑟威兰斯科技有限公司 | Integrated license plate face recognition method and module based on raspberry group |
| CN110163089A (en) * | 2019-04-11 | 2019-08-23 | 西安艾润物联网技术服务有限责任公司 | Auth method and device |
| CN111160299A (en) * | 2019-12-31 | 2020-05-15 | 上海依图网络科技有限公司 | Living body identification method and device |
| CN111696243A (en) * | 2020-07-08 | 2020-09-22 | 北京大兴投资集团有限公司 | Access & exit management and control equipment with multiple recognition function |
| CN112215234A (en) * | 2020-10-16 | 2021-01-12 | 神思电子技术股份有限公司 | Method for checking information of vehicle personnel at road interface |
| CN112861645A (en) * | 2021-01-18 | 2021-05-28 | 盛视科技股份有限公司 | Infrared camera dim light environment compensation method and device and electronic equipment |
| CN114187695B (en) * | 2021-12-08 | 2023-12-08 | 上海世茂物联网科技有限公司 | Barrier gate control method, device, system, equipment and medium for garage |
| CN114333329B (en) * | 2022-01-26 | 2023-09-29 | 杭州海康威视数字技术股份有限公司 | Parking detection method, device and storage medium |
| CN114255600A (en) * | 2022-01-26 | 2022-03-29 | 杭州海康威视数字技术股份有限公司 | Parking detection method, parking detection equipment and storage medium |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103781261A (en) * | 2014-02-28 | 2014-05-07 | 深圳英飞拓科技股份有限公司 | Infrared lamp control method for infrared network camera |
| CN105430347A (en) * | 2015-11-30 | 2016-03-23 | 燕山大学 | Compressed sensing based wireless sensor image capturing and transmitting system |
| CN105844202A (en) * | 2015-01-12 | 2016-08-10 | 芋头科技（杭州）有限公司 | Image recognition system and method |
| CN106909882A (en) * | 2017-01-16 | 2017-06-30 | 广东工业大学 | A kind of face identification system and method for being applied to security robot |
| CN107103903A (en) * | 2017-05-05 | 2017-08-29 | 百度在线网络技术（北京）有限公司 | Acoustic training model method, device and storage medium based on artificial intelligence |
| CN107635099A (en) * | 2017-10-09 | 2018-01-26 | 深圳市天视通电子科技有限公司 | A kind of double optical-fiber network video cameras of human body sensing and safety defense monitoring system |
| WO2018076523A1 (en) * | 2016-10-25 | 2018-05-03 | 科世达（上海）管理有限公司 | Gesture recognition method and apparatus, and in-vehicle system |
Family Cites Families (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2000149184A (en) * | 1998-11-10 | 2000-05-30 | Nippon Dainamatto Kk | Overloading control unit |
| CN103093517A (en) * | 2011-10-28 | 2013-05-08 | 台湾色彩与影像科技股份有限公司 | Human face recognition visitor management method |
| CN103093528A (en) * | 2012-12-28 | 2013-05-08 | 中山大学 | Intelligent system for monitoring vehicles and people in community |
| CN103544481A (en) * | 2013-10-24 | 2014-01-29 | 深圳市飞瑞斯科技有限公司 | Road gate monitoring method and system based on human face recognition |
| CN104317397A (en) * | 2014-10-14 | 2015-01-28 | 奇瑞汽车股份有限公司 | Vehicle-mounted man-machine interactive method |
| CN104301623B (en) * | 2014-10-28 | 2017-10-31 | 浙江宇视科技有限公司 | A kind of mixing light compensation method and device |
| CN104966066A (en) * | 2015-06-26 | 2015-10-07 | 武汉大学 | Traffic block port monitoring oriented in-car human face detection method and system |
| CN105184873B (en) * | 2015-08-17 | 2018-03-30 | 深圳北斗应用技术研究院有限公司 | Intelligent managing system for parking lot |
| CN105530437A (en) * | 2016-01-29 | 2016-04-27 | 生迪智慧科技有限公司 | Environment brightness measurement and calculation method and shooting device |
| CN106056071B (en) * | 2016-05-30 | 2019-05-10 | 北京智芯原动科技有限公司 | A kind of driver makes a phone call the detection method and device of behavior |
| KR102483642B1 (en) * | 2016-08-23 | 2023-01-02 | 삼성전자주식회사 | Method and apparatus for liveness test |
| CN106485224B (en) * | 2016-10-13 | 2019-07-16 | 北京智芯原动科技有限公司 | A kind of seatbelt wearing recognition methods and device |
| CN106503707B (en) * | 2016-10-21 | 2019-08-02 | 浙江宇视科技有限公司 | Licence plate recognition method and device under a kind of infrared lighting condition |
| CN206741505U (en) * | 2017-03-17 | 2017-12-12 | 中控智慧科技股份有限公司 | Identify management equipment and device |
| CN107301384A (en) * | 2017-06-09 | 2017-10-27 | 湖北天业云商网络科技有限公司 | A kind of driver takes phone behavioral value method and system |
| CN107248289A (en) * | 2017-07-01 | 2017-10-13 | 广东艾科智泊科技股份有限公司 | A kind of unidentified the self-service of car plate vehicle in many parking lots opens a sluice gate system |
| CN107452104A (en) * | 2017-07-26 | 2017-12-08 | 北京声迅电子股份有限公司 | A kind of control method for vehicle and system of the vehicle bayonet socket based on intelligent monitoring |
| CN107491750A (en) * | 2017-08-14 | 2017-12-19 | 深圳前海华夏智信数据科技有限公司 | A kind of testimony of a witness comparison method and device |
| CN107370956A (en) * | 2017-08-21 | 2017-11-21 | 深圳市天视通电子科技有限公司 | A kind of method and starlight video camera for realizing the switching of starlight video camera diurnal pattern |
-
2018
- 2018-05-22 CN CN201810495514.7A patent/CN108710857B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103781261A (en) * | 2014-02-28 | 2014-05-07 | 深圳英飞拓科技股份有限公司 | Infrared lamp control method for infrared network camera |
| CN105844202A (en) * | 2015-01-12 | 2016-08-10 | 芋头科技（杭州）有限公司 | Image recognition system and method |
| CN105430347A (en) * | 2015-11-30 | 2016-03-23 | 燕山大学 | Compressed sensing based wireless sensor image capturing and transmitting system |
| WO2018076523A1 (en) * | 2016-10-25 | 2018-05-03 | 科世达（上海）管理有限公司 | Gesture recognition method and apparatus, and in-vehicle system |
| CN106909882A (en) * | 2017-01-16 | 2017-06-30 | 广东工业大学 | A kind of face identification system and method for being applied to security robot |
| CN107103903A (en) * | 2017-05-05 | 2017-08-29 | 百度在线网络技术（北京）有限公司 | Acoustic training model method, device and storage medium based on artificial intelligence |
| CN107635099A (en) * | 2017-10-09 | 2018-01-26 | 深圳市天视通电子科技有限公司 | A kind of double optical-fiber network video cameras of human body sensing and safety defense monitoring system |
Non-Patent Citations (3)
| Title |
|---|
| Face recognition using both visible light image and near-infrared image and a deep network;KaiGuo 等;《CAAI Transactions on Intelligence Technology》;20170331;第2卷(第1期);第39-47页 * |
| 基于GA-BP的安全带佩戴识别方法;葛如海;《江苏大学学报》;20140331;第35卷(第2期);第125-131页 * |
| 基于GA-BP神经网络的计算机智能化图像识别技术探究;曹永峰 等;《应用激光》;20170228;第37卷(第1期);第139-143页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108710857A (en) | 2018-10-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108710857B (en) | Method and device for identifying people and vehicles based on infrared supplementary lighting | |
| CN107622229B (en) | Video vehicle re-identification method and system based on fusion features | |
| US8995723B2 (en) | Detecting and recognizing traffic signs | |
| US9842266B2 (en) | Method for detecting driver cell phone usage from side-view images | |
| CN104778444B (en) | The appearance features analysis method of vehicle image under road scene | |
| CN110781883B (en) | Parking space size recognition system and method based on 360-degree looking-around camera | |
| WO2020029444A1 (en) | Method and system for detecting attention of driver while driving | |
| KR101374139B1 (en) | Monitoring method through image fusion of surveillance system | |
| CN106022243B (en) | A kind of retrograde recognition methods of the car lane vehicle based on image procossing | |
| US11117570B1 (en) | Parking assistance using a stereo camera and an added light source | |
| CN112364842B (en) | Double-shot face recognition method and device | |
| CN106585530A (en) | Vehicle-mounted snapshooting device and control method thereof | |
| CN110450706A (en) | A kind of adaptive distance light lamp control system and image processing algorithm | |
| WO2019085930A1 (en) | Method and apparatus for controlling dual-camera apparatus in vehicle | |
| CN113034378B (en) | Method for distinguishing electric automobile from fuel automobile | |
| JP5789578B2 (en) | Eye opening / closing judgment method and apparatus, program, and surveillance video system | |
| CN108805184B (en) | Image recognition method and system for fixed space and vehicle | |
| CN104102921A (en) | Video detection method and video detection device based on face recognition | |
| CN109508659A (en) | A kind of face identification system and method for crossing | |
| CN111553214B (en) | Method and system for detecting smoking behavior of driver | |
| CN112861797A (en) | Method and device for identifying authenticity of license plate and related equipment | |
| KR102309106B1 (en) | Situation linkage type image analysis device | |
| KR101337554B1 (en) | Apparatus for trace of wanted criminal and missing person using image recognition and method thereof | |
| CN114266993A (en) | Image-based road environment detection method and device | |
| CN109878290B (en) | AI technology-based intelligent vehicle-mounted air conditioner control method and system |
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 |