CN111881792A - Mobile micro-bayonet system and working method thereof - Google Patents
Mobile micro-bayonet system and working method thereof Download PDFInfo
- Publication number
- CN111881792A CN111881792A CN202010685677.9A CN202010685677A CN111881792A CN 111881792 A CN111881792 A CN 111881792A CN 202010685677 A CN202010685677 A CN 202010685677A CN 111881792 A CN111881792 A CN 111881792A
- Authority
- CN
- China
- Prior art keywords
- vehicle
- license plate
- image
- subsystem
- face
- 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.)
- Withdrawn
Links
Images
Classifications
-
- 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
- G06V20/56—Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
- G06V20/58—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads
- G06V20/584—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads of vehicle lights or traffic lights
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/253—Fusion techniques of extracted features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/047—Probabilistic or stochastic networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- 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
- G06V20/59—Context or environment of the image inside of a vehicle, e.g. relating to seat occupancy, driver state or inner lighting conditions
-
- 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
- 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
- G06V40/168—Feature extraction; Face representation
-
- 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
- G06V40/172—Classification, e.g. identification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/60—Type of objects
- G06V20/62—Text, e.g. of license plates, overlay texts or captions on TV images
- G06V20/625—License plates
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Multimedia (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Molecular Biology (AREA)
- Human Computer Interaction (AREA)
- Biomedical Technology (AREA)
- Computing Systems (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Probability & Statistics with Applications (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a mobile micro-bayonet system and a working method thereof, belonging to the technical field of face recognition and vehicle recognition, wherein the working method of the mobile micro-bayonet system comprises the following steps: identifying a license plate number and a license plate color from a road image acquired in real time; the collected road image and the collected in-vehicle image are respectively operated by adopting an automatic supervision grid decoder, the three-dimensional shape facial information pixel by pixel can be rapidly predicted, the face image in the road image and the in-vehicle image is identified, and each 512-dimensional face feature code is stored through Hash coding, so that the working efficiency is improved; identifying vehicle attribute information; the point location name of the vehicle, the area of the vehicle and the driving direction of the vehicle are analyzed, and the corresponding time, the point location name, the area, the direction, the license plate number, the license plate color, the face image, the vehicle attribute information and the network map are mapped and stored, so that the working efficiency can be quickly improved, the working time is greatly shortened, and the working cost is reduced.
Description
Technical Field
The invention belongs to the technical field of face recognition and vehicle recognition, and particularly relates to a mobile micro-bayonet system and a working method thereof.
Background
With the development of economy, vehicles have entered ordinary people's families as tools for transportation, and the increase of vehicles brings more and more potential safety hazards.
In order to ensure good traffic order or social security, the license plate recognition equipment based on the license plate recognition technology is widely applied, and the inventor finds that the existing product is widely applied to positions of intersections, property districts, commercial buildings, government institutions and the like through research. In the license plate detection process, the license plate is detected from the image and the position of the license plate in each frame of image is determined, so that a large amount of time is consumed, and the working efficiency is low.
Disclosure of Invention
In order to at least solve the technical problems, the invention provides a mobile micro-bayonet system and a working method thereof.
According to a first aspect of the present invention, there is provided a mobile microsocket system, mountable on a vehicle, comprising:
the image acquisition subsystem is used for acquiring road images in real time and acquiring images in a vehicle in real time;
the license plate detection subsystem is used for identifying the license plate number and the license plate color from the collected road image;
the face recognition subsystem is used for respectively operating the acquired road image and the acquired in-vehicle image by adopting an automatic supervision grid decoder, predicting three-dimensional shape face information pixel by pixel, recognizing the face image in the road image and the in-vehicle image, and storing each 512-dimensional face feature code through Hash coding;
the vehicle attribute analysis subsystem is used for identifying the area where the vehicle logo is located from the collected road image, detecting the vehicle logo to obtain vehicle brand information, identifying the vehicle color and the vehicle type, and taking the vehicle brand information, the vehicle color and the vehicle type as the vehicle attribute information;
and the business subsystem is respectively connected with the license plate detection subsystem, the face recognition subsystem and the vehicle attribute analysis subsystem and is used for analyzing the point location name of the vehicle, the area of the vehicle and the driving direction of the vehicle, and establishing mapping storage of corresponding time, the point location name, the area, the direction, the license plate number, the license plate color, the face image, the vehicle attribute information and the network map.
Further, the system further comprises: and the human-vehicle retrieval subsystem is connected with the service subsystem and is used for acquiring the matched license plate number, time, point location name, area, direction and vehicle attribute information when receiving a retrieval request containing at least one of a time range, a point location name, a license plate number and a face image.
Further, the 512-dimensional face feature codes are stored through hash codes, the 512-dimensional face feature codes are stored in a database through the hash codes, 512-dimensional features are subjected to barreling through a local sensitive hash function, so that the candidate feature set retrieved each time is relatively reduced, the calculation complexity of searching similar users is reduced, and meanwhile, an elastic search frame is adopted for query, and the search efficiency is improved.
The system further comprises a track query subsystem connected with the service subsystem and used for generating the motion track of the target face in the time range and identifying the track in an electronic map and a form under the condition of acquiring the name, the time range, the target face and the license plate number.
Further, the system further comprises: and the control subsystem is connected with the service subsystem and is used for controlling, inquiring and early warning the license plate number.
Furthermore, the control subsystem is used for storing the number of the license plate, time, place, control reason and control person; and the system is also used for inquiring corresponding information such as license plate number, time, place, arrangement reason, arrangement person and the like under the condition of receiving arrangement inquiry.
Further, the deployment and control subsystem is used for sending the early warning information to the deployment and control person mobile phone in real time for alarming under the condition that suspicious vehicles appear.
According to a second aspect of the present invention, there is provided a method for operating a mobile micro-bayonet system, comprising:
acquiring a road image and an in-vehicle image in real time;
identifying a license plate number and a license plate color from the collected road image;
respectively operating the collected road image and the collected in-vehicle image by adopting an automatic supervision grid decoder, predicting three-dimensional shape facial information pixel by pixel, identifying face images in the road image and the in-vehicle image, and storing each 512-dimensional face feature code through Hash coding;
identifying the area where the vehicle logo is located from the acquired road image, detecting the vehicle logo to obtain vehicle brand information, identifying the vehicle color and the vehicle type, and taking the vehicle brand information, the vehicle color and the vehicle type as vehicle attribute information;
analyzing the point location name of the vehicle, the area of the vehicle and the driving direction of the vehicle, and establishing mapping storage of the corresponding time, the point location name, the area, the direction, the license plate number, the license plate color, the face image, the vehicle attribute information and the network map.
According to a third aspect of the present invention, there is provided a computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing the steps of the method as described above when executing the program.
According to a fourth aspect of the present invention, there is provided a computer readable storage medium storing a program which, when executed, is capable of implementing the method as described above.
The invention has the beneficial effects that: the system is installed on a vehicle, road images and images in the vehicle are collected in real time, license plate numbers and license plate colors can be identified from the collected road images, the collected road images and images in the vehicle are operated respectively based on an automatic supervision grid decoder, three-dimensional shape facial information pixel by pixel can be rapidly predicted, face images in the road images and the images in the vehicle are identified, and each 512-dimensional face feature code is stored through Hash codes, so that the working efficiency is improved. The system can analyze the point location name of the vehicle, the area of the vehicle and the driving direction of the vehicle by identifying the area of the vehicle mark from the acquired road image, and further realizes real-time information acquisition, analysis and alarm of remote people and vehicles on the street and the inside of the vehicle by the people and vehicle retrieval subsystem and the track query subsystem, thereby playing an important role in handling emergency cases, face identification, vehicle identification, people and vehicle track analysis and the like.
Drawings
The above and/or additional aspects and advantages of the present invention will become apparent and readily appreciated from the following description of the embodiments, taken in conjunction with the accompanying drawings of which,
FIG. 1 is a schematic structural diagram of a mobile micro-bayonet system according to the present invention;
FIG. 2 is a flow chart of a working method of the mobile micro-bayonet system provided by the present invention;
FIG. 3 is a schematic diagram of a mobile micro-bayonet system according to the present invention;
FIG. 4 is a schematic diagram of a mobile micro-checkpoint system for face and vehicle identification and retrieval according to the present invention;
FIG. 5 is a schematic diagram of a track query result of a mobile micro-bayonet system according to the present invention;
fig. 6 is a schematic diagram of a mobile micro-bayonet system for table search according to the present invention.
Detailed Description
Reference will now be made in detail to embodiments of the present invention, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to the same or similar elements or elements having the same or similar function throughout. The embodiments described below with reference to the drawings are illustrative only and should not be construed as limiting the invention.
In order to more clearly illustrate the invention, the invention is further described below with reference to preferred embodiments and the accompanying drawings. Similar parts in the figures are denoted by the same reference numerals. It is to be understood by persons skilled in the art that the following detailed description is illustrative and not restrictive, and is not to be taken as limiting the scope of the invention.
In a first aspect of the present invention, there is provided a mobile microsocket system, mountable on a vehicle, as shown in fig. 1, comprising:
the image acquisition subsystem 200 is used for acquiring road images in real time and acquiring images in a vehicle in real time;
the license plate detection subsystem 201 is used for identifying license plate numbers and license plate colors from the collected road images;
in the embodiment of the present invention, the license plate detection subsystem 201 is specifically configured to perform image enhancement by performing filtering, contrast enhancement, and image operator scanning on the acquired road image, process the data after the image enhancement processing by using a simplified neural network, improve robustness of deep learning, complete license plate region detection, and directly extract features from a license plate region detection result to perform license plate character recognition and license plate color recognition, so as to obtain a license plate number and a license plate color.
The face recognition subsystem 202 is used for respectively operating the acquired road image and the acquired in-vehicle image by adopting an automatic supervision grid decoder, predicting three-dimensional shape face information pixel by pixel, recognizing the face image in the road image and the in-vehicle image, and storing each 512-dimensional face feature code through Hash coding;
in the embodiment of the invention, the face recognition subsystem 202 is used for respectively adopting the self-supervision grid decoder branch to replace the convolution operation for the collected in-vehicle image and road image, so that the operation speed can be improved on the premise of not influencing the accuracy, meanwhile, the grid convolution and the up-sampling characteristics of partial convolution layers are fused, the three-dimensional shape face information of each pixel is predicted, and the detection speed is greatly improved.
On the face recognition model, the characteristic distances of different faces are increased by updating the am-softmax loss function, and the training results are optimized by adopting a simulated annealing method, so that the algorithm which can only be applied to a hundred thousand level data set originally can be applied to ten million level face data, and the working efficiency is greatly improved.
The vehicle attribute analysis subsystem 203 is used for identifying the area where the vehicle logo is located from the collected road image, detecting the vehicle logo to obtain vehicle brand information, identifying the vehicle color and the vehicle type, and taking the vehicle brand information, the vehicle color and the vehicle type as the vehicle attribute information;
in the embodiment of the invention, the vehicle region in the road image is subjected to region confusion mechanism operation, the spatial layout of the local image region is damaged, the semantic region which has the greatest influence on the detection result is extracted as the main semantic output of the visual content of the vehicle region, the region is tested to be the region where the vehicle logo is located according to the algorithm flow, the vehicle brand information can be obtained by inputting the neural network for classification after the vehicle logo is detected in the mode, and meanwhile, the network branches are added to complete the detection of the vehicle color and the vehicle type.
And the service subsystem 204 is respectively connected with the license plate detection subsystem 201, the face identification subsystem 202 and the vehicle attribute analysis subsystem 203, and is used for analyzing the point location name of the vehicle, the area of the vehicle and the driving direction of the vehicle, and establishing mapping and storing the corresponding time, the point location name, the area, the direction, the license plate number, the license plate color, the face image, the vehicle attribute information and the network map.
In the embodiment of the invention, pictures of one year are stored according to the estimation of the data volume of actual operation, and the storage capacity of at least 100TB is required. To provide storage capability as well as retrieval capability. Further, in the implementation of the present invention, an external interface is provided to facilitate interfacing with other systems of the public security or to extend new functions.
In another embodiment of the present invention, there is provided a mobile micro-card port system, including:
the image acquisition subsystem 200 is respectively connected with the license plate detection subsystem 201, the face identification subsystem 202 and the vehicle attribute analysis subsystem 203, and is used for acquiring road images in real time and acquiring images in a vehicle in real time;
the license plate detection subsystem 201 is used for identifying license plate numbers and license plate colors from the collected road images;
in the embodiment of the present invention, the license plate detection subsystem 201 is specifically configured to perform image enhancement by performing filtering, contrast enhancement, and image operator scanning on the acquired road image, process the data after the image enhancement processing by using a simplified neural network, improve robustness of deep learning, complete license plate region detection, and directly extract features from a license plate region detection result to perform license plate character recognition and license plate color recognition, so as to obtain a license plate number and a license plate color.
The face recognition subsystem 202 is used for respectively operating the acquired road image and the acquired in-vehicle image by adopting an automatic supervision grid decoder, predicting three-dimensional shape face information pixel by pixel, recognizing the face image in the road image and the in-vehicle image, and storing each 512-dimensional face feature code through Hash coding;
in the embodiment of the invention, the face recognition subsystem 202 is used for respectively adopting the self-supervision grid decoder branch to replace the convolution operation for the collected in-vehicle image and road image, so that the operation speed can be improved on the premise of not influencing the accuracy, meanwhile, the grid convolution and the up-sampling characteristics of partial convolution layers are fused, the three-dimensional shape face information of each pixel is predicted, and the detection speed is greatly improved.
On the face recognition model, the characteristic distances of different faces are increased by updating the am-softmax loss function, and the training results are optimized by adopting a simulated annealing method, so that the algorithm which can only be applied to a hundred thousand level data set originally can be applied to ten million level face data, and the working efficiency is greatly improved.
And the vehicle attribute analysis subsystem 203 is used for identifying the area where the vehicle logo is located from the collected road image, detecting the vehicle logo, obtaining the vehicle brand information, identifying the vehicle color and the vehicle type, and taking the vehicle brand information, the vehicle color and the vehicle type as the vehicle attribute information.
In the embodiment of the invention, the vehicle region in the road image is subjected to region confusion mechanism operation, the spatial layout of the local image region is damaged, the semantic region which has the greatest influence on the detection result is extracted as the main semantic output of the visual content of the vehicle region, the region is tested to be the region where the vehicle logo is located according to the algorithm flow, the vehicle brand information can be obtained by inputting the neural network for classification after the vehicle logo is detected in the mode, and meanwhile, the network branches are added to complete the detection of the vehicle color and the vehicle type.
And the service subsystem 204 is respectively connected with the license plate detection subsystem 201, the face identification subsystem 202 and the vehicle attribute analysis subsystem 203, and is used for analyzing the point location name of the vehicle, the area of the vehicle and the driving direction of the vehicle, and establishing mapping and storing the corresponding time, the point location name, the area, the direction, the license plate number, the license plate color, the face image, the vehicle attribute information and the network map.
In the embodiment of the invention, pictures of one year are stored according to the estimation of the data volume of actual operation, and the storage capacity of at least 100TB is required. To provide storage capability as well as retrieval capability. Further, in the implementation of the present invention, an external interface is provided to facilitate interfacing with other systems of the public security or to extend new functions.
And the human-vehicle retrieval subsystem 205 is connected with the service subsystem 204 and is used for acquiring the matched license plate number, time, point location name, region, direction and vehicle attribute information when receiving a retrieval request containing at least one of a time range, the point location name, the license plate number and a face image.
In the embodiment of the present invention, the human-vehicle retrieval subsystem 205 is specifically configured to store each 512-dimensional human face feature code in a database through a hash code, and bucket the 512-dimensional features through a locality sensitive hash function, so that candidate feature sets retrieved each time are relatively small, the computational complexity of finding similar users is reduced, and meanwhile, an elastic search framework is adopted for query, so as to improve the search efficiency.
And the track query subsystem 206 is connected with the service subsystem 204 and is used for generating a motion track of the target face within the time range and identifying the track in an electronic map and a form under the condition of acquiring the name, the time range, the target face and the license plate number.
In another embodiment of the present invention, the system further comprises a control subsystem, connected to the service subsystem 204, for controlling, querying, and warning the license plate number;
in the embodiment of the invention, the control subsystem is used for storing the number, time, place, control reason and control person of the controlled license plate; and the system is also used for inquiring corresponding information such as license plate number, time, place, arrangement reason, arrangement person and the like under the condition of receiving arrangement inquiry. And the system is also used for sending the early warning information to the mobile phone of the deployment and control person in real time to give an alarm under the condition that suspicious vehicles appear. Further, the present embodiment can view the picture information of the vehicle.
The name, identification number, license plate number, vehicle type of the suspect, and the time, reason and information of the suspect for adding control to the suspect can be inquired in the table. When the vehicle of the suspect enters and exits the parking lot, the vehicle of the suspect sends a short message to the deployment and control person for reminding.
Furthermore, the corresponding detailed information such as name, ID card, adding time and the like can be checked.
In a second aspect of the present invention, there is provided a method for operating a mobile micro-bayonet system, as shown in fig. 2, including:
step 301: acquiring a road image and an in-vehicle image in real time;
step 302: identifying a license plate number and a license plate color from the collected road image;
in the embodiment of the invention, the collected road image can be subjected to image enhancement in a filtering, contrast enhancement and image operator scanning mode, the simplified neural network is adopted to process the data after the image enhancement processing, the robustness of deep learning is improved, the license plate region detection is completed, the features are directly extracted from the license plate region detection result to perform license plate character recognition and license plate color recognition, and the license plate number and the license plate color are obtained.
Step 303: respectively operating the collected road image and the collected in-vehicle image by adopting an automatic supervision grid decoder, predicting three-dimensional shape facial information pixel by pixel, identifying face images in the road image and the in-vehicle image, and storing each 512-dimensional face feature code through Hash coding;
in the embodiment of the invention, the collected in-car images and road images can be respectively subjected to self-supervision grid decoder branches instead of convolution operation, so that the operation speed can be improved on the premise of not influencing the accuracy, meanwhile, grid convolution and up-sampling characteristics of partial convolution layers are fused, the three-dimensional shape face information of each pixel is predicted, and the detection speed is greatly improved.
On the face recognition model, the characteristic distances of different faces are increased by updating the am-softmax loss function, and the training results are optimized by adopting a simulated annealing method, so that the algorithm which can only be applied to a hundred thousand level data set originally can be applied to ten million level face data, and the working efficiency is greatly improved.
Step 304: identifying the area where the vehicle logo is located from the acquired road image, detecting the vehicle logo to obtain vehicle brand information, identifying the vehicle color and the vehicle type, and taking the vehicle brand information, the vehicle color and the vehicle type as vehicle attribute information;
in the embodiment of the invention, pictures of one year are stored according to the estimation of the data volume of actual operation, the storage capacity of at least 100TB is required, the data transmitted by an external network convergence system is received by adopting the storage of a distributed file system, the instant secondary analysis and processing of the data are carried out, and the processed data are stored and written into a platform database in real time for displaying and applying by a service platform. To provide storage capability as well as retrieval capability. Further, in the implementation of the present invention, an external interface is provided to facilitate interfacing with other systems of the public security or to extend new functions.
Step 305: analyzing the point location name of the vehicle, the area of the vehicle and the driving direction of the vehicle, and establishing mapping storage of the corresponding time, the point location name, the area, the direction, the license plate number, the license plate color, the face image, the vehicle attribute information and the network map.
In the embodiment of the invention, each 512-dimensional face feature code is stored in the database through hash codes, 512-dimensional features are subjected to barrel partitioning through a local sensitive hash function, so that the alternative feature set retrieved each time is relatively reduced, the calculation complexity for searching similar users is reduced, and meanwhile, an elastic search frame is adopted for query, so that the search efficiency is improved.
In another embodiment of the present invention, a method for operating a mobile micro-bayonet system is provided, including:
step 401: acquiring a road image and an in-vehicle image in real time;
step 402: identifying a license plate number and a license plate color from the collected road image;
in the embodiment of the invention, the collected road image can be subjected to image enhancement in a filtering, contrast enhancement and image operator scanning mode, the simplified neural network is adopted to process the data after the image enhancement processing, the robustness of deep learning is improved, the license plate region detection is completed, the features are directly extracted from the license plate region detection result to perform license plate character recognition and license plate color recognition, and the license plate number and the license plate color are obtained.
Step 403: respectively operating the collected road image and the collected in-vehicle image by adopting an automatic supervision grid decoder, predicting three-dimensional shape facial information pixel by pixel, identifying face images in the road image and the in-vehicle image, and storing each 512-dimensional face feature code through Hash coding;
in the embodiment of the invention, the collected in-car images and road images can be respectively subjected to self-supervision grid decoder branches instead of convolution operation, so that the operation speed can be improved on the premise of not influencing the accuracy, meanwhile, grid convolution and up-sampling characteristics of partial convolution layers are fused, the three-dimensional shape face information of each pixel is predicted, and the detection speed is greatly improved.
On the face recognition model, the characteristic distances of different faces are increased by updating the am-softmax loss function, and the training results are optimized by adopting a simulated annealing method, so that the algorithm which can only be applied to a hundred thousand level data set originally can be applied to ten million level face data, and the working efficiency is greatly improved.
Step 404: analyzing the point location name of the vehicle, the area of the vehicle and the driving direction of the vehicle, and establishing mapping storage of the corresponding time, the point location name, the area, the direction, the license plate number, the license plate color, the face image, the vehicle attribute information and the network map.
In the embodiment of the invention, pictures of one year are stored according to the estimation of the data volume of actual operation, the storage capacity of at least 100TB is required, the data transmitted by an external network convergence system is received by adopting the storage of a distributed file system, the instant secondary analysis and processing of the data are carried out, and the processed data are stored and written into a platform database in real time for displaying and applying by a service platform. To provide storage capability as well as retrieval capability. Further, in the implementation of the present invention, an external interface is provided to facilitate interfacing with other systems of the public security or to extend new functions.
Step 405: and when a retrieval request containing at least one of a time range, a point location name, a license plate number and a face image is received, acquiring the matched license plate number, time, point location name, area, direction and vehicle attribute information.
In the embodiment of the invention, each 512-dimensional face feature code is stored in the database through hash codes, 512-dimensional features are subjected to barrel partitioning through a local sensitive hash function, so that the alternative feature set retrieved each time is relatively reduced, the calculation complexity for searching similar users is reduced, and meanwhile, an elastic search frame is adopted for query, so that the search efficiency is improved.
Step 406: under the condition of acquiring the name, the time range, the target face and the license plate number, generating the motion track of the target face in the time range, and identifying the track in an electronic map and a form.
In another embodiment of the invention, the license plate number can be controlled, queried and early-warned;
furthermore, the number, time, place, arrangement reason and arrangement person of the arranged and controlled license plate can be saved; and the corresponding information such as the license plate number, time, place, arrangement reason, arrangement person and the like can be inquired under the condition of receiving the arrangement inquiry. And the early warning information can be sent to the mobile phone of the deployment and control person to give an alarm in real time under the condition that suspicious vehicles appear. Further, the present embodiment can view the picture information of the vehicle.
The name, identification number, license plate number, vehicle type of the suspect, and the time, reason and information of the suspect for adding control to the suspect can be inquired in the table. When the vehicle of the suspect enters and exits the parking lot, the vehicle of the suspect sends a short message to the deployment and control person for reminding.
Furthermore, the corresponding detailed information such as name, ID card, adding time and the like can be checked.
The invention provides another working method of a mobile micro-bayonet system corresponding to the embodiment, wherein the mobile micro-bayonet device can be installed on a taxi, image node data of real-time road and vehicle interior human face and vehicle attribute are obtained through the motion of a vehicle and the data acquisition of the micro-bayonet device, the data are transmitted back to the cloud for storage, and an electronic map is used as a carrier, and various algorithms are fused to realize intelligent applications such as human face recognition, vehicle attribute analysis, visual display, statistics, retrieval, control and alarm arrangement and the like on the data.
The invention provides a mobile micro-card port system by adopting a face and vehicle recognition technology based on deep learning and by means of an Internet of things technology. The street remote information acquisition, analysis and alarm system realizes real-time street remote information acquisition, analysis and alarm in vehicles and plays an important role in handling emergency cases, face recognition, vehicle recognition, human and vehicle track analysis and the like.
In the embodiment of the invention, the data circulation system is divided into an external network convergence system and an internal network service system. The system architecture is shown in fig. 3, the external network convergence system mainly performs a uniform convergence on collected data sets of vehicles and human faces and regularly enters an internet boundary server, and the external network convergence system has the functions of load balancing, data caching, format unification and the like.
The intranet service system mainly performs secondary identification, analysis and processing on core application data of the mobile micro-card platform, and has the main functions of data storage and stream data analysis and provides data interfaces and expansion interfaces with other public security systems. Wherein the data storage comprises: the storage of a year of pictures based on an estimation of the amount of data actually running requires at least 100TB of storage capacity, and a common file system has difficulty in storing and retrieving data of this magnitude and adopts a distributed file system for storage. The flow analysis comprises the following steps: and receiving data transmitted by the external network convergence system, performing instant secondary analysis and processing on the data, storing the processed data, and writing the data into the platform database in real time for displaying and applying by the service platform. The external interface includes: the system is convenient to be docked with other systems of the public security or new functions are expanded.
As used herein, the singular forms "a", "an", and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms "comprises" and/or "comprising," when used in this specification, specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof. It will be understood that when an element is referred to as being "connected" or "coupled" to another element, it can be directly connected or coupled to the other element or intervening elements may also be present. Further, "connected" or "coupled" as used herein may include wirelessly connected or wirelessly coupled. As used herein, the term "and/or" includes all or any element and all combinations of one or more of the associated listed items.
It will be understood by those skilled in the art that, unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. It will be further understood that terms, such as those defined in commonly used dictionaries, should be interpreted as having a meaning that is consistent with their meaning in the context of the prior art and will not be interpreted in an idealized or overly formal sense unless expressly so defined herein.
It should be understood that the above detailed description of the technical solution of the present invention with the help of preferred embodiments is illustrative and not restrictive. On the basis of reading the description of the invention, a person skilled in the art can modify the technical solutions described in the embodiments, or make equivalent substitutions for some technical features; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (10)
1. A mobile microcard system, mountable on a vehicle, comprising:
the image acquisition subsystem is used for acquiring road images in real time and acquiring images in a vehicle in real time;
the license plate detection subsystem is used for identifying the license plate number and the license plate color from the collected road image;
the face recognition subsystem is used for respectively operating the acquired road image and the acquired in-vehicle image by adopting an automatic supervision grid decoder, predicting three-dimensional shape face information pixel by pixel, recognizing the face image in the road image and the in-vehicle image, and storing each 512-dimensional face feature code through Hash coding;
the vehicle attribute analysis subsystem is used for identifying the area where the vehicle logo is located from the collected road image, detecting the vehicle logo to obtain vehicle brand information, identifying the vehicle color and the vehicle type, and taking the vehicle brand information, the vehicle color and the vehicle type as the vehicle attribute information;
and the business subsystem is respectively connected with the license plate detection subsystem, the face recognition subsystem and the vehicle attribute analysis subsystem and is used for analyzing the point location name of the vehicle, the area of the vehicle and the driving direction of the vehicle, and establishing mapping storage of corresponding time, the point location name, the area, the direction, the license plate number, the license plate color, the face image, the vehicle attribute information and the network map.
2. The system of claim 1,
the system further comprises: and the human-vehicle retrieval subsystem is connected with the service subsystem and is used for acquiring the matched license plate number, time, point location name, area, direction and vehicle attribute information when receiving a retrieval request containing at least one of a time range, a point location name, a license plate number and a face image.
3. The system of claim 2,
the 512-dimensional face feature codes are stored through hash codes, the 512-dimensional face feature codes are stored in a database through the hash codes, 512-dimensional features are barreled through a local sensitive hash function, so that the candidate feature set retrieved each time is relatively reduced, the computational complexity of searching similar users is reduced, and meanwhile, an elastic search frame is adopted for query, and the search efficiency is improved.
4. The system of claim 1,
the system also comprises a track query subsystem which is connected with the service subsystem and used for generating the motion track of the target face in the time range and identifying the track in the electronic map and the form under the condition of acquiring the name, the time range, the target face and the license plate number.
5. The system of claim 1,
the system further comprises: and the control subsystem is connected with the service subsystem and is used for controlling, inquiring and early warning the license plate number.
6. The system of claim 5,
the control subsystem is used for storing the number, time, place, control reason and control person of the controlled license plate; and the system is also used for inquiring corresponding information such as license plate number, time, place, arrangement reason, arrangement person and the like under the condition of receiving arrangement inquiry.
7. The system of claim 5,
and the control subsystem is used for sending the early warning information to the mobile phone of the control person in real time for alarming under the condition that suspicious vehicles appear.
8. A working method of a mobile micro-bayonet system is characterized by comprising the following steps:
acquiring a road image and an in-vehicle image in real time;
identifying a license plate number and a license plate color from the collected road image;
respectively operating the collected road image and the collected in-vehicle image by adopting an automatic supervision grid decoder, predicting three-dimensional shape facial information pixel by pixel, identifying face images in the road image and the in-vehicle image, and storing each 512-dimensional face feature code through Hash coding;
identifying the area where the vehicle logo is located from the acquired road image, detecting the vehicle logo to obtain vehicle brand information, identifying the vehicle color and the vehicle type, and taking the vehicle brand information, the vehicle color and the vehicle type as vehicle attribute information;
analyzing the point location name of the vehicle, the area of the vehicle and the driving direction of the vehicle, and establishing mapping storage of the corresponding time, the point location name, the area, the direction, the license plate number, the license plate color, the face image, the vehicle attribute information and the network map.
9. A computer device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor,
the steps of the method of claim 8 are implemented when the program is executed by the processor.
10. A computer-readable storage medium, characterized in that the computer-readable storage medium stores a program which, when executed, is capable of implementing the method of claim 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010685677.9A CN111881792A (en) | 2020-07-16 | 2020-07-16 | Mobile micro-bayonet system and working method thereof |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010685677.9A CN111881792A (en) | 2020-07-16 | 2020-07-16 | Mobile micro-bayonet system and working method thereof |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111881792A true CN111881792A (en) | 2020-11-03 |
Family
ID=73155617
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010685677.9A Withdrawn CN111881792A (en) | 2020-07-16 | 2020-07-16 | Mobile micro-bayonet system and working method thereof |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111881792A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112328820A (en) * | 2020-11-16 | 2021-02-05 | 青岛以萨数据技术有限公司 | Method, system, terminal and medium for searching vehicle image through face image |
-
2020
- 2020-07-16 CN CN202010685677.9A patent/CN111881792A/en not_active Withdrawn
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112328820A (en) * | 2020-11-16 | 2021-02-05 | 青岛以萨数据技术有限公司 | Method, system, terminal and medium for searching vehicle image through face image |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| AU2019101142A4 (en) | A pedestrian detection method with lightweight backbone based on yolov3 network | |
| CN108388888B (en) | Vehicle identification method and device and storage medium | |
| CN108875548B (en) | Character track generation method and device, storage medium and electronic equipment | |
| CN109241349B (en) | Monitoring video multi-target classification retrieval method and system based on deep learning | |
| CN108197326B (en) | Vehicle retrieval method and device, electronic equipment and storage medium | |
| CN113033604B (en) | Vehicle detection method, system and storage medium based on SF-YOLOv4 network model | |
| US20210382933A1 (en) | Method and device for archive application, and storage medium | |
| CN110619277A (en) | Multi-community intelligent deployment and control method and system | |
| CN111931683B (en) | Image recognition method, device and computer readable storage medium | |
| CN111553355B (en) | Monitoring video-based method for detecting and notifying store outgoing business and managing store owner | |
| CN108182218B (en) | Video character recognition method and system based on geographic information system and electronic equipment | |
| CN109871789A (en) | Vehicle checking method under a kind of complex environment based on lightweight neural network | |
| CN112949578B (en) | Vehicle lamp state identification method, device, equipment and storage medium | |
| CN110348463A (en) | The method and apparatus of vehicle for identification | |
| Sikirić et al. | Image representations on a budget: Traffic scene classification in a restricted bandwidth scenario | |
| CN109993049A (en) | A kind of video image structure analysis system towards intelligent security guard field | |
| CN117011635A (en) | Model training method, image data processing device and computer equipment | |
| CN114550051A (en) | Vehicle loss detection method and device, computer equipment and storage medium | |
| US20180260401A1 (en) | Distributed video search with edge computing | |
| CN113673527B (en) | License plate recognition method and system | |
| CN114385662A (en) | Road network updating method and device, storage medium and electronic equipment | |
| CN111881792A (en) | Mobile micro-bayonet system and working method thereof | |
| CN116413740A (en) | Laser radar point cloud ground detection method and device | |
| CN113343903B (en) | License plate recognition method and system in natural scene | |
| CN112529116B (en) | Scene element fusion processing method, device and equipment and computer storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WW01 | Invention patent application withdrawn after publication | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20201103 |