CN106780886B - Vehicle identification system and vehicle entrance and exit identification method - Google Patents
Vehicle identification system and vehicle entrance and exit identification method Download PDFInfo
- Publication number
- CN106780886B CN106780886B CN201611168658.9A CN201611168658A CN106780886B CN 106780886 B CN106780886 B CN 106780886B CN 201611168658 A CN201611168658 A CN 201611168658A CN 106780886 B CN106780886 B CN 106780886B
- Authority
- CN
- China
- Prior art keywords
- vehicle
- license plate
- identification
- plate recognition
- primary
- 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.)
- Expired - Fee Related
Links
- 238000000034 method Methods 0.000 title claims abstract description 57
- 238000012795 verification Methods 0.000 claims abstract description 53
- 238000000605 extraction Methods 0.000 claims abstract description 21
- 238000013459 approach Methods 0.000 claims description 14
- 230000004888 barrier function Effects 0.000 abstract description 8
- 230000008569 process Effects 0.000 description 8
- 238000012549 training Methods 0.000 description 7
- 238000013526 transfer learning Methods 0.000 description 4
- 239000013589 supplement Substances 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 239000003086 colorant Substances 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 238000009434 installation Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 241000083879 Polyommatus icarus Species 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000002360 explosive Substances 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 239000000047 product Substances 0.000 description 1
- 230000003252 repetitive effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G07—CHECKING-DEVICES
- G07C—TIME OR ATTENDANCE REGISTERS; REGISTERING OR INDICATING THE WORKING OF MACHINES; GENERATING RANDOM NUMBERS; VOTING OR LOTTERY APPARATUS; ARRANGEMENTS, SYSTEMS OR APPARATUS FOR CHECKING NOT PROVIDED FOR ELSEWHERE
- G07C9/00—Individual registration on entry or exit
- G07C9/30—Individual registration on entry or exit not involving the use of a pass
- G07C9/38—Individual registration on entry or exit not involving the use of a pass with central registration
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/60—Type of objects
- G06V20/62—Text, e.g. of license plates, overlay texts or captions on TV images
- G06V20/63—Scene text, e.g. street names
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/60—Type of objects
- G06V20/62—Text, e.g. of license plates, overlay texts or captions on TV images
- G06V20/625—License plates
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Traffic Control Systems (AREA)
Abstract
The invention discloses a vehicle identification system, a vehicle entrance identification method and a vehicle exit identification method, wherein the vehicle identification system comprises a front-end identification device, a vehicle verification server, a control machine and a barrier gate, the front-end identification device is used for capturing and identifying vehicles entering or exiting to obtain a primary license plate identification result, the vehicle verification server is used for carrying out feature extraction and vehicle verification on the vehicles, the control machine is connected with the front-end identification device and the vehicle verification server and is used for fusing the primary vehicle identification result of the front-end identification device with the vehicle verification result of the vehicle verification server, and the barrier gate is controlled to be opened when the primary vehicle identification result is consistent with a record in a fixed vehicle list or the primary vehicle identification result is matched with a record in the fixed vehicle list to reach a threshold value and the vehicle verification result reaches the threshold value. The license plate identification is carried out by combining the license plate identification of the fixed vehicle with the vehicle verification, so that the problems of safety and difficulty in identifying the license plates of partial fixed vehicles are solved.
Description
Technical Field
The invention relates to the technical field of image processing and video analysis, in particular to a vehicle identification system and a vehicle entrance and exit identification method.
Background
At present, due to the explosive growth of vehicles caused by the rapid popularization of the vehicles, the requirements for vehicle management are higher and higher, the vehicles need to pass through a registration gateway quickly, and meanwhile, the registration gateway can also obtain the license plate information of the passing vehicles quickly. Therefore, the license plate recognition device has been widely used in parking lots, urban roads, highways, and other areas to automatically capture and recognize the license plate of a vehicle.
In the past, because the recognition rate of the license plate is not high, the owner mainly brushes the card when entering and exiting the parking lot, and the license plate recognition is more used as a supplement of a monthly card or a temporary card or is only used for storing evidence and uploading recognition information to a traffic police station. In recent years, as the license plate recognition technology is rapidly developed, the recognition rate of the license plate is improved, and the pure license plate recognition and the unattended scheme become feasible.
However, in the process of recognizing the license plate, many times, some vehicles cannot be recognized or are difficult to recognize due to various objective reasons, such as dark environment, improper light supplement measures, direct sunlight in the daytime, damage and fouling of the license plate, and the like. The current license plate recognition result is used as a unique certificate for entering and leaving, and once the license plate recognition result is wrong, vehicles cannot pass through.
Even in other locations, vehicle management is more stringent. For example, property communities have high requirements on safety of vehicles, and many properties select a mode of card-plus-license plate identification at present. The card is an IC card or an RFID card, and the owner must simultaneously ensure that the card information and the license plate identification information are correct and matched to enter and exit each time. The IC card needs to be swiped by an owner during parking, so that congestion is easily caused during the peak of work; the reading distance of the RFID card is not easy to control, and misreading is easy to cause. Therefore, it is a great challenge to improve the traffic speed while ensuring the recognition rate.
Disclosure of Invention
The invention aims to provide a vehicle identification system, a vehicle entrance identification method and a vehicle exit identification method, which solve the problems of safety and difficulty in identifying a part of fixed vehicle license plates.
In order to solve the above technical problem, an embodiment of the present invention provides a vehicle identification system, which includes a front-end identification device, a vehicle authentication server, a controller and a gateway, the front-end recognition equipment is used for capturing and recognizing vehicles entering or leaving the parking lot to obtain a primary license plate recognition result, the vehicle authentication server is used for carrying out feature extraction and vehicle authentication on the vehicle, the control machine is connected with the front-end identification equipment and the vehicle authentication server, for fusing a primary vehicle identification result of the front-end identification device with a vehicle authentication result of the vehicle authentication server, and controlling the barrier gate to be opened when the primary vehicle identification result is consistent with the record in the fixed vehicle list, or the primary vehicle identification result is matched with the record in the fixed vehicle list and reaches a threshold value, and the vehicle verification result reaches the threshold value.
The front-end recognition equipment is a license plate recognition all-in-one machine or a license plate recognizer carrying a high-definition camera.
The ground sensor is used for sending a trigger signal after a vehicle stops and sending the trigger signal to the front-end identification equipment connected with the ground sensor for capturing and identifying.
Wherein the ground sense is a ring coil type ground sense, a square coil type ground sense, an ultrasonic type ground sense, a radar type ground sense or a video type ground sense.
In addition, the embodiment of the invention also provides a vehicle approach identification method, which comprises the following steps:
carrying out license plate recognition on an incoming vehicle to obtain a primary license plate recognition result;
judging whether the primary license plate recognition result is consistent with records in a fixed vehicle list or not;
if so, opening the brake and releasing the vehicle; if not, judging that the matching of the primary license plate recognition result and the record in the fixed vehicle list reaches a threshold value;
if not, judging and recording that the vehicle is a temporary entering vehicle, and opening and releasing the vehicle; if so, carrying out feature extraction on the vehicle to obtain vehicle features;
judging whether the matching degree of the vehicle characteristics and prestored vehicle characteristic records of the background server reaches a threshold value or not;
if not, judging and recording that the vehicle is a temporary entering vehicle, and opening and releasing the vehicle; and if so, modifying the primary license plate recognition result, and opening and releasing the vehicle.
The method comprises the following steps of carrying out license plate recognition on an approaching vehicle to obtain a primary license plate recognition result, wherein the license plate recognition comprises the following steps:
when the ground sensor detects that the vehicle stops, a trigger signal is sent to the front-end identification device;
and the front-end recognition equipment receives the trigger signal, and performs vehicle license plate recognition on the vehicle parked on the ground sensor to obtain a primary license plate recognition result.
In addition, the embodiment of the invention also provides a vehicle departure identification method, which comprises the following steps:
carrying out vehicle license plate recognition on vehicles leaving the field to obtain a primary license plate recognition result;
judging whether the primary license plate recognition result is consistent with records in a fixed vehicle list or not;
if so, opening the brake of the vehicle and releasing the brake, and if not, judging whether the matching degree of the primary license plate recognition result and the records in the fixed vehicle list reaches a threshold value;
if not, determining that the vehicle is a temporary departure vehicle, and if so, performing feature extraction on the vehicle to obtain vehicle features;
the background server verifies the vehicle characteristics and judges whether the matching degree of the vehicle characteristics and vehicle characteristic records prestored in the background server reaches a threshold value;
if so, modifying the primary license plate recognition result, and opening and releasing the vehicle; if not, the vehicle is judged to be a temporary departure vehicle.
The vehicle license plate recognition is carried out on the vehicles leaving the field, and a primary license plate recognition result is obtained, and the method comprises the following steps:
when the ground sensor detects that the vehicle stops, a trigger signal is sent to the front-end identification device;
and the front-end recognition equipment receives the trigger signal, and performs vehicle license plate recognition on the vehicle parked on the ground sensor to obtain a primary license plate recognition result.
Wherein, the matching degree of the first license plate recognition result and the records in the fixed vehicle list does not reach a threshold value, and the vehicle is determined to be a temporary vehicle, and the method comprises the following steps:
searching a temporary card entrance record, and judging whether the vehicles have the same entrance record;
if so, opening the brake of the vehicle and releasing the vehicle, and if not, extracting the characteristics of the vehicle to obtain the characteristics of the vehicle;
judging whether the matching degree of the vehicle characteristics and the prestored vehicle characteristic records reaches a threshold value by using a background server;
if so, modifying the primary license plate recognition result, and opening the brake and releasing the vehicle; if not, the vehicle is manually registered and then is opened and released.
Wherein, the matching degree of the first license plate recognition result and the record in the fixed vehicle list reaches a threshold value, and the matching degree of the vehicle characteristics and the pre-stored vehicle characteristic record does not reach the threshold value, and the vehicle is determined to be a temporary vehicle, including:
searching a temporary card entrance record, and judging whether the vehicles have the same entrance record;
if not, the vehicle is manually registered and then is opened and released, and if yes, the primary license plate recognition result is modified, and the vehicle is opened and released.
Compared with the prior art, the vehicle identification system, the vehicle entrance identification method and the vehicle exit identification method provided by the embodiment of the invention have the following advantages:
the vehicle identification system comprises a front-end identification device, a vehicle verification server, a control machine and a barrier gate, wherein the front-end identification device is used for capturing and identifying vehicles entering or leaving a field to obtain a primary license plate identification result, the vehicle verification server is used for performing feature extraction and vehicle verification on the vehicles, the control machine is connected with the front-end identification device and the vehicle verification server and is used for fusing a primary vehicle identification result of the front-end identification device with a vehicle verification result of the vehicle verification server, and the barrier gate is controlled to be opened when the primary vehicle identification result is consistent with records in a fixed vehicle list or the primary vehicle identification result is matched with records in the fixed vehicle list to reach a threshold value and the vehicle verification result reaches the threshold value.
The vehicle approach identification method provided by the embodiment of the invention comprises the following steps:
carrying out license plate recognition on an incoming vehicle to obtain a primary license plate recognition result;
judging whether the primary license plate recognition result is consistent with records in a fixed vehicle list or not;
if so, opening the brake and releasing the vehicle; if not, judging that the matching of the primary license plate recognition result and the record in the fixed vehicle list reaches a threshold value;
if not, judging and recording that the vehicle is a temporary entering vehicle, and opening and releasing the vehicle; if so, carrying out feature extraction on the vehicle to obtain vehicle features;
judging whether the matching degree of the vehicle characteristics and prestored vehicle characteristic records of the background server reaches a threshold value or not;
if not, judging and recording that the vehicle is a temporary entering vehicle, and opening and releasing the vehicle; and if so, modifying the primary license plate recognition result, and opening and releasing the vehicle.
The vehicle departure identification method provided by the embodiment of the invention comprises the following steps:
carrying out vehicle license plate recognition on vehicles leaving the field to obtain a primary license plate recognition result;
judging whether the primary license plate recognition result is consistent with records in a fixed vehicle list or not;
if so, opening the brake of the vehicle and releasing the brake, and if not, judging whether the matching degree of the primary license plate recognition result and the records in the fixed vehicle list reaches a threshold value;
if not, determining that the vehicle is a temporary departure vehicle, and if so, performing feature extraction on the vehicle to obtain vehicle features;
the background server verifies the vehicle characteristics and judges whether the matching degree of the vehicle characteristics and vehicle characteristic records prestored in the background server reaches a threshold value;
if so, modifying the primary license plate recognition result, and opening and releasing the vehicle; if not, the vehicle is judged to be a temporary departure vehicle
The vehicle identification system, the vehicle entrance identification method and the vehicle exit identification method are characterized in that the license plate of a vehicle entering and exiting is identified through front-end identification equipment, a primary license plate identification result is obtained, and if the primary license plate identification result is consistent with records in a fixed vehicle list, the license plate number in the primary license plate identification result exists in the fixed vehicle list. And (5) indicating that the identification is successful, and opening the brake of the vehicle to release. If the primary license plate recognition result is inconsistent with the record in the fixed vehicle list, but the matching degree of the primary license plate recognition result and the record in the fixed vehicle list reaches a threshold value, namely the primary license plate recognition result is very similar to the record in the fixed vehicle list, and the license plate is likely to be damaged or stained, then a vehicle verification server is used for verification, if the verification is passed, the license plate of the vehicle is the vehicle in the fixed vehicle list, and the vehicle is opened and released.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to these drawings without creative efforts.
FIG. 1 is a schematic structural diagram of an embodiment of a vehicle identification system according to an embodiment of the present invention;
FIG. 2 is a flowchart illustrating steps of a method for identifying a vehicle approach according to an embodiment of the present invention;
FIG. 3 is a flowchart illustrating steps of another embodiment of a method for identifying a vehicle approach according to an embodiment of the present invention;
FIG. 4 is a flowchart illustrating steps of a specific implementation of a vehicle departure identification method according to an embodiment of the present invention;
FIG. 5 is a flowchart illustrating steps of another embodiment of a vehicle departure identification method according to an embodiment of the present invention;
fig. 6 is a schematic step flow chart of a further specific implementation of the vehicle departure identification method according to the embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Referring to fig. 1 to 6, fig. 1 is a schematic structural diagram of a specific implementation of a vehicle identification system according to an embodiment of the invention; FIG. 2 is a flowchart illustrating steps of a method for identifying a vehicle approach according to an embodiment of the present invention; FIG. 3 is a flowchart illustrating steps of another embodiment of a method for identifying a vehicle approach according to an embodiment of the present invention; FIG. 4 is a flowchart illustrating steps of a specific implementation of a vehicle departure identification method according to an embodiment of the present invention; FIG. 5 is a flowchart illustrating steps of another embodiment of a vehicle departure identification method according to an embodiment of the present invention; fig. 6 is a schematic step flow chart of a further specific implementation of the vehicle departure identification method according to the embodiment of the present invention.
In one embodiment, the vehicle identification system includes a front-end identification device 10, a vehicle authentication server 20, a control machine 30, and a barrier 40, the front-end recognition device 10 is used for capturing and recognizing vehicles entering or leaving the vehicle to obtain a primary license plate recognition result, the vehicle authentication server 20 is used for performing feature extraction and vehicle authentication on a vehicle, the control machine 30 is connected with the front-end identification device 10 and the vehicle authentication server 20, for fusing the primary vehicle identification result of the front-end identification device 10 with the vehicle authentication result of the vehicle authentication server 20, and controlling the barrier gate 40 to open when the primary vehicle identification result is consistent with the record in the fixed vehicle list, or the primary vehicle identification result is matched with the record in the fixed vehicle list to reach a threshold value and the vehicle verification result reaches the threshold value.
The vehicle verification in the invention is to collect the characteristics of the vehicle, obtain the vehicle characteristics of the vehicle, then carry out matching verification with the pre-stored vehicle characteristic record, if the threshold value is reached, the vehicle is indicated to be a vehicle in a fixed vehicle list, and the vehicle characteristics, including the vehicle type, the color of the vehicle and the like, can indicate or indicate the content of the characteristic information of the vehicle.
By adopting the method for recognizing the license plate by the front-end recognition device 10 and verifying the vehicle by the vehicle verification server 20, the license plate is recognized by the front-end device, and the vehicle is verified by the vehicle verification server 20. By combining license plate recognition and vehicle verification, the problem of safety can be solved on one hand, and the problem of difficulty in recognizing the license plate of a part of fixed vehicles can be solved on the other hand.
In a specific embodiment, the front-end recognition device 10 is a license plate recognition all-in-one machine or a license plate recognizer with a high-definition camera.
The license plate recognition all-in-one machine is also called an integrated camera, and is specially used for cameras which can automatically focus and are built in lenses. Compared with the traditional camera, the integrated camera has the advantages of small and exquisite volume, attractive appearance, convenience in installation and use, wide monitoring range, high cost performance and the like. The license plate recognition all-in-one machine is characterized in that a license plate recognition function is added to an integrated camera, the license plate recognition all-in-one machine is based on an embedded intelligent high-definition license plate recognition all-in-one machine product which is released aiming at the parking lot industry, license plate recognition, camera shooting, front-end storage, light supplement and the like are integrated, and the license plate automatic exposure control algorithm is based on the license plate, so that the imaging is excellent. Has the characteristics of excellent performance, multiple functions, high adaptability, strong stability and the like.
Support for license plate types: common blue cards, black cards, yellow cards, double-layer yellow cards, police car license plates, new armed police license plates, new military license plates, embassy license plates, and hong Kong and Australia entry and exit continent license plates.
It should be noted that the front-end recognition device 10 may also use other license plate recognition devices, and the present invention is not limited to this, and the license plate recognition device may use a video stream or a single frame recognition mode to recognize the entrance or exit of the vehicle.
In a specific embodiment, the vehicle identification system further comprises a ground sensor 50, and the ground sensor 50 is used for sending a trigger signal after the vehicle stops, and sending the trigger signal to the front-end identification device 10 connected with the ground sensor 50 for capturing and identifying. In a general video stream mode, the detection of a license plate in a single frame image is mainly used, missing recognition can be caused to a vehicle without the license plate, and the purpose of using the ground sense 50 is to detect the vehicle without the license plate. The general single frame recognition mode is based on the vehicle driving on the ground sensor 50, and when the vehicle stops on the ground sensor 50, the ground sensor 50 transmits a trigger signal to be captured and recognized by the front-end recognition device 10.
The ground sensor 50 is a loop coil type ground sensor, a square coil type ground sensor, an ultrasonic type ground sensor, a radar type ground sensor, or a video type ground sensor.
Therefore, the license plate is recognized in a mode of combining video streaming or single-frame recognition with the ground sensor 50, the ground sensor 50 transmits a trigger signal only when a vehicle needs to enter or exit, the front-end recognition device 10 recognizes the video streaming or the single frame, the situation that the front-end recognition device 10 is always in a license plate recognition state is avoided, and the recognition efficiency is improved.
It should be noted that the present invention does not specifically limit the type and installation manner of the floor covering 50.
Certainly, in the present invention, in order to improve the recognition efficiency of the vehicle, in the current vehicle recognition process, if the similarity between the primary license plate recognition result and the license plate number of one of the fixed vehicle lists reaches the threshold (the similarity is provided by the user for facility, and may require that 7 is 5 bits of the license plate number and may also be 6 bits of the license plate number), and in the subsequent vehicle verification process, the matching degree also reaches the threshold, which indicates that the primary license plate recognition result is wrong.
It should be noted that, in the present invention, the vehicle verification server 20 may separately photograph or photograph the vehicle entering or leaving, and obtain the vehicle information of the vehicle from the current frame image, and perform the feature extraction, or obtain the vehicle information of the vehicle from the frame image of the vehicle obtained from the front-end recognition device 10, and perform the feature extraction, which is not limited in the present invention.
In addition, the embodiment of the invention also provides a vehicle approach identification method, which comprises the following steps:
step 101, identifying the license plate of an approaching vehicle to obtain a primary license plate identification result;
step 102, judging whether the primary license plate recognition result is consistent with records in a fixed vehicle list or not;
if yes, step 103, opening the brake of the vehicle and releasing the vehicle; if not, step 104, judging that the matching of the primary license plate recognition result and the record in the fixed vehicle list reaches a threshold value;
if not, step 108, judging and recording that the vehicle is a temporary entering vehicle, and opening the brake to release the vehicle; if yes, step 105, performing feature extraction on the vehicle to obtain vehicle features;
step 106, judging whether the matching degree of the vehicle characteristics and prestored vehicle characteristic records of the background server reaches a threshold value;
if not, turning to step 108, judging and recording that the vehicle is a temporary entering vehicle, and opening the brake to release the vehicle; and if so, step 107, modifying the primary license plate recognition result, and opening and releasing the vehicle.
In the invention, the judgment that the matching of the primary license plate recognition result and the record in the fixed vehicle list reaches the threshold value is to judge whether the similarity between the primary license plate recognition result of the vehicle and one or more of the fixed vehicle list reaches the threshold value, wherein the similarity can be set by a user, and can also be 7 or 5 or 6 in the license plate number.
In the invention, the vehicle features are obtained by extracting the features of the vehicle, and the information content such as the type, the color and the like of the vehicle, which can represent that the vehicle is different from other vehicles, can be extracted.
When a vehicle enters a field, the license plate recognition and vehicle verification modes are adopted, only part of license plate information of the vehicle can be recognized in the license plate recognition process, the characteristics of the license plate are extracted through the two verification servers, the vehicle verification is carried out, when the matching degree of the vehicle characteristic records in the two verification servers and the pre-stored vehicle characteristic records reaches a threshold value, the vehicle is one in a fixed vehicle list, and the vehicle is opened and released. Meanwhile, because the license plate recognition result of the vehicle entering the field is wrong, the vehicle is quickly recognized for the next time, the passing efficiency is improved, the primary license plate recognition result is modified, the barrier gate is closed after the vehicle enters the field, and the process flow of the vehicle is finished.
In order to improve the vehicle identification efficiency, only the approaching vehicle is identified, in step 101, the license plate identification is performed on the approaching vehicle to obtain a first license plate identification result, including:
step 1011, when the ground sensor detects that the vehicle stops, sending a trigger signal to the front-end identification equipment;
and step 1012, receiving the trigger signal by the front-end recognition equipment, and recognizing the license plate of the vehicle parked on the ground sensor to obtain a primary license plate recognition result.
The general video stream mode mainly detects the license plate in a single-frame image, and leads to missing identification for vehicles without license plates. The ground sense is used for detecting vehicles without license plates. The general single-frame recognition mode is based on the ground sense of the vehicle, and when the vehicle stops on the ground sense, the ground sense emits a trigger signal which is captured and recognized by front-end recognition equipment.
The ground sense is a ring coil type ground sense, a square coil type ground sense, an ultrasonic wave type ground sense, a radar type ground sense or a video type ground sense.
Therefore, the license plate is recognized in a mode of combining video streaming or single-frame recognition and ground sensing, the ground sensing is used for transmitting the trigger signal only when the vehicle needs to enter or exit, the front-end recognition equipment is used for recognizing the video streaming or the single frame, the situation that the front-end recognition equipment is always in a license plate recognition state is avoided, and the recognition efficiency is improved.
In addition, the embodiment of the invention also provides a vehicle departure identification method, which comprises the following steps:
step 201, recognizing the license plate of a vehicle leaving a house to obtain a primary license plate recognition result;
step 202, judging whether the primary license plate recognition result is consistent with records in a fixed vehicle list;
if so, step 203, the vehicle is opened and released, and if not, step 204, whether the matching degree of the primary license plate recognition result and the records in the fixed vehicle list reaches a threshold value is judged;
if not, step 208, determining that the vehicle is a temporary departure vehicle, and if so, step 205, performing feature extraction on the vehicle to obtain vehicle features;
step 206, judging whether the matching degree of the vehicle characteristic record prestored in the background server reaches a threshold value;
if yes, step 207, modifying the primary license plate recognition result, and opening and releasing the vehicle; if not, go to step 208, and determine that the vehicle is a temporary departure vehicle.
According to the vehicle departure identification method, after the primary license plate identification result is obtained, whether the primary license plate identification result is consistent with the record in the fixed vehicle list or not is judged, namely whether the primary license plate identification result is the license plate number of one vehicle in the fixed vehicle list or not is judged. If so, the identification is successful, otherwise, whether the similarity between the vehicle and one or more records in the fixed vehicle list reaches a threshold value is judged, namely whether the complete license plate information of the vehicle can not be obtained once because of the primary license plate identification result, and only a few bits (set by a user) are obtained. And then, next, feature extraction is carried out on the vehicle, feature information such as the color, the vehicle type and the like of the vehicle is extracted, the feature information is matched with the vehicle feature record memorability stored in the vehicle verification server, if the threshold value is reached, the vehicle is a vehicle in a fixed vehicle list, and if the threshold value is not reached, the vehicle is determined to be a temporary vehicle. Therefore, a mode of combining license plate recognition and vehicle verification is adopted, the vehicle can be recognized without complete correct recognition, the recognition efficiency is improved, even if the license plate of the vehicle is stained, the vehicle can be verified through feature extraction, whether the vehicle is in a fixed vehicle list or not is judged, and the recognition efficiency and the vehicle passing efficiency are greatly improved.
In addition, in order to conveniently and quickly identify the vehicle at the next time, if the primary license plate identification result is not consistent with the record in the fixed vehicle list, the similarity reaches a threshold value, and meanwhile, the threshold value is reached in the vehicle verification process, the vehicle in the fixed vehicle list is indicated, the primary license plate identification result is indicated to be wrong, the primary license plate identification result is modified, and then the record is carried out, so that the next quick identification of the vehicle is facilitated.
Step 201, recognizing the license plate of the vehicle leaving the field to obtain a primary license plate recognition result, including:
step 2011, when the ground sensor detects that the vehicle stops, a trigger signal is sent to the front-end identification device;
step 2012, the front-end recognition device receives the trigger signal, and performs vehicle license plate recognition on the vehicle parked on the ground sensor to obtain a primary license plate recognition result.
The video stream mode in the general license plate recognition mainly detects the license plate in a single-frame image, and leads to missing recognition for vehicles without license plates. The ground sense is used for detecting vehicles without license plates. The general single-frame recognition mode is based on the ground sense of the vehicle, and when the vehicle stops on the ground sense, the ground sense emits a trigger signal which is captured and recognized by front-end recognition equipment. The ground sense is a ring coil type ground sense, a square coil type ground sense, an ultrasonic wave type ground sense, a radar type ground sense or a video type ground sense.
Therefore, the license plate is recognized in a mode of combining video streaming or single-frame recognition and ground sensing, the ground sensing is used for transmitting the trigger signal only when the vehicle needs to enter or exit, the front-end recognition equipment is used for recognizing the video streaming or the single frame, the situation that the front-end recognition equipment is always in a license plate recognition state is avoided, and the recognition efficiency is improved.
The matching degree of the primary license plate recognition result and the records in the fixed vehicle list does not reach a threshold value, and the vehicle is judged to be a temporary vehicle which is out of the field, and the method comprises the following steps:
step 2081, searching temporary card entrance records, and judging whether the vehicles have the same entrance records;
if so, step 2082, performing brake opening and releasing on the vehicle, and if not, step 2083, performing feature extraction on the vehicle to obtain vehicle features;
step 2084, judging whether the matching degree of the vehicle characteristics and the prestored vehicle characteristic records reaches a threshold value by using a background server;
if yes, turning to step 207, modifying the primary license plate recognition result, and opening the brake of the vehicle for releasing; if not, step 2085, after manual registration, the vehicle is opened and released.
When the vehicle is judged to be a temporary vehicle, whether the vehicle finishes a temporary approach procedure or not is judged, whether a record exists in the temporary card approach record or not is checked, if yes, repeated registration is not needed, and the vehicle can be rapidly released. Otherwise, processing is carried out according to the initial vehicle passing through the inspection gateway for the first time.
The matching degree of the primary license plate recognition result and the records in the fixed vehicle list reaches a threshold value, and the matching degree of the vehicle characteristics and the pre-stored vehicle characteristic records does not reach the threshold value, which indicates that the vehicle is only a vehicle with high license plate similarity in the fixed vehicle name list but not a vehicle in the fixed vehicle list, and the vehicle is judged to be a temporary vehicle, including:
step 2061, searching a temporary card approach record, and judging whether the vehicles have the same approach record;
if not, in step 2085, the vehicle is manually registered and then is opened and released, and if so, the step 207 is switched to, the primary license plate recognition result is modified, and the vehicle is opened and released.
In the invention, the primary license plate recognition result is preferentially used for judging whether the vehicle is the vehicle in the fixed vehicle list, because the primary license plate recognition result is simpler to extract the characteristics of the vehicle and verify the vehicle, and only the result that the vehicle verification can not be recognized due to the fact that the vehicle types, colors and the like of the vehicles are completely consistent is used.
In the invention, if the vehicle is not identified by adopting the modes of the feature extraction and the vehicle verification, the type and the number of the feature extraction are less, the content of the feature extraction of the vehicle can be increased, and the invention does not specifically limit the added feature information of the vehicle.
It should be noted that, in the present invention, the license plate information of the vehicle is implemented in the front-end recognition device, the front-end recognition device has low calculation capability, and only the features of the area near the license plate are taken, without training, and the features are compared.
Vehicle verification: the vehicle verification server is realized, has strong calculation capability, and can extract the characteristics of the vehicle to train the vehicle verification server.
The vehicle characteristic record is obtained by repeatedly extracting and verifying the vehicle information in all the fixed vehicle lists in advance, so that the vehicle identification efficiency can be effectively improved.
When the vehicle verification server performs vehicle verification, a vehicle feature identification method based on deep learning needs to be provided with a model. It is understood that an area, such as a cell or a business center, is trained to become a vehicle verification model, which may be understood as a "proprietary" model of the area. With each vehicle as one of the classes in the model. The features that the model needs to use are determined by the user, such as the vehicle type, the vehicle logo or the vehicle color, etc. E.g. 500 stationary vehicles in the XX garden, a model is trained that contains 500 classes, e.g. where class 101 is a car "yue B12345, white, toyota, cameiry 2015" of a certain owner. The model can classify the above 500 classes and give the classification result. For example, when a vehicle enters or exits, the model first extracts the features of the vehicle from the image and then identifies the vehicle. If the recognition result is correct, the confidence coefficient of the 101 th class is displayed to be the highest, and the software outputs the license plate number corresponding to the 101 th class.
The specific process of training and maintaining the model comprises the following steps:
(1) before the equipment leaves a factory, a deep convolution network is designed according to vehicle characteristics required to be used, such as vehicle types, vehicle logos or vehicle colors and the like and image sizes.
The deep convolutional network is a research field which rises rapidly in recent years, has the characteristics of flexibility and changeability in network weight sharing and network architecture, and can be designed into a better network structure aiming at different application fields. The present invention is primarily directed to systems and therefore the design of deep convolutional networks is not described here.
(2) And after the equipment is installed on site, loading a fixed vehicle information list on the front-end equipment. When a fixed vehicle enters and exits a field (the fixed vehicle is matched with the fixed vehicle information table), a plurality of images of the vehicle are shot by the license plate recognition equipment. For vehicles which are difficult to identify and easy to make mistakes, the vehicle owner is required to take pictures and upload pictures in a matching manner.
(3) Collected over time, a deep convolutional network is trained using the collected images of all vehicles in the area (e.g., a cell or a business center), and the network weights for each layer are obtained. Each vehicle may be considered a class. The training is carried out at the back end, the system only supports front-end recognition when the system is started, and the method of vehicle license plate recognition and vehicle verification is used after the network training at the back end is finished.
(4) Samples of vehicles newly purchased by the user, or certain non-stationary vehicles that require significant attention may be collected continuously. After enough training samples of the new vehicle are collected, the system adds the new vehicle information to the model using transfer learning (transfer learning). What the transfer learning needs to do is to train N +1 classifiers by using the N types of classifiers which are trained before, so that a large amount of cost of manual labeling is saved. In order to make the learning rate of the convolutional layer 0 without modifying the characteristic parameters of the previous convolutional layer and to achieve a fast training success, only the fully-connected layer is trained to fit the new class. The transfer learning algorithm is also not detailed in the present invention. Meanwhile, in the actual use process, vehicles with poor recognition effects can be continuously provided with vehicle samples, and the vehicle samples can be repeatedly studied.
(5) And if the fixed car owner or the special license plate car owner needs to cancel the fixed car, only the fixed car is deleted in the fixed car information list. I.e. the vehicle remains in the trained model and is simply not used. Using procedures (4) and (5) may eliminate network repetitive training.
The procedure for verifying the vehicle using the model is:
1. using an image containing the vehicles entering and exiting as input, and extracting the characteristics of the vehicles in the image by using a trained deep convolutional network;
2. recognizing the vehicle by using a trained deep convolutional network, outputting confidence degrees of a plurality of classes by a system, and selecting the class with the highest confidence degree;
3. and (4) according to the result obtained in the step (2), if the confidence coefficient reaches a threshold value, outputting the license plate number corresponding to the type. Otherwise, a result without a match is returned.
In summary, in the vehicle identification system, the vehicle entering identification method, and the vehicle exiting identification method provided in the embodiments of the present invention, the front-end identification device identifies the license plate of the vehicle entering and exiting to obtain the initial license plate identification result, and if the initial license plate identification result is consistent with the record in the fixed vehicle list, the license plate number in the initial license plate identification result exists in the fixed vehicle list. And (5) indicating that the identification is successful, and opening the brake of the vehicle to release. If the primary license plate recognition result is inconsistent with the record in the fixed vehicle list, but the matching degree of the primary license plate recognition result and the record in the fixed vehicle list reaches a threshold value, namely the primary license plate recognition result is very similar to the record in the fixed vehicle list, and the license plate is likely to be damaged or stained, then a vehicle verification server is used for verification, if the verification is passed, the license plate of the vehicle is the vehicle in the fixed vehicle list, and the vehicle is opened and released.
The vehicle identification system, the vehicle approach identification method and the vehicle departure identification method provided by the invention are described in detail above. The principles and embodiments of the present invention are explained herein using specific examples, which are presented only to assist in understanding the method and its core concepts. It should be noted that, for those skilled in the art, it is possible to make various improvements and modifications to the present invention without departing from the principle of the present invention, and those improvements and modifications also fall within the scope of the claims of the present invention.
Claims (10)
1. The utility model provides a vehicle identification system, its characterized in that, includes front end identification equipment, vehicle verification server, controller and banister, front end identification equipment is used for taking a candid photograph and discerning the vehicle of entering or leaving, obtains first license plate identification result, vehicle verification server is used for carrying out feature extraction and vehicle verification to the vehicle, the controller with front end identification equipment vehicle verification server connects, is used for with the first vehicle identification result of front end identification equipment with vehicle verification server's vehicle verification result fuses first vehicle identification result is unanimous with the record in the fixed vehicle list, or first vehicle identification result matches with the record in the fixed vehicle list and reaches the threshold value just when vehicle verification result reaches the threshold value, control the banister is opened the floodgate.
2. The vehicle identification system of claim 1, wherein the front-end identification device is a license plate identification all-in-one machine or a license plate identifier with a high-definition camera.
3. The vehicle identification system of claim 1, further comprising a ground sensor, wherein the ground sensor is used for sending a trigger signal after a vehicle stops, and sending the trigger signal to the front-end identification device connected with the ground sensor for capturing and identifying.
4. The vehicle identification system of claim 3, wherein the ground feel is a toroidal coil type ground feel, a quad coil type ground feel, an ultrasonic type ground feel, a radar type ground feel, or a video type ground feel.
5. A vehicle approach identification method, comprising:
carrying out license plate recognition on an incoming vehicle to obtain a primary license plate recognition result;
judging whether the primary license plate recognition result is consistent with records in a fixed vehicle list or not;
if so, opening the brake and releasing the vehicle; if not, judging that the matching of the primary license plate recognition result and the record in the fixed vehicle list reaches a threshold value;
if not, judging and recording that the vehicle is a temporary entering vehicle, and opening and releasing the vehicle; if so, carrying out feature extraction on the vehicle to obtain vehicle features;
judging whether the matching degree of the vehicle characteristics and prestored vehicle characteristic records of the background server reaches a threshold value or not;
if not, judging and recording that the vehicle is a temporary entering vehicle, and opening and releasing the vehicle; and if so, modifying the primary license plate recognition result, and opening and releasing the vehicle.
6. The vehicle approach recognition method of claim 5, wherein the performing license plate recognition on the approaching vehicle to obtain a primary license plate recognition result comprises:
when the ground sensor detects that the vehicle stops, a trigger signal is sent to the front-end identification device;
and the front-end recognition equipment receives the trigger signal, and performs vehicle license plate recognition on the vehicle parked on the ground sensor to obtain a primary license plate recognition result.
7. A vehicle departure identification method is characterized by comprising the following steps:
carrying out vehicle license plate recognition on vehicles leaving the field to obtain a primary license plate recognition result;
judging whether the primary license plate recognition result is consistent with records in a fixed vehicle list or not;
if so, opening the brake of the vehicle and releasing the brake, and if not, judging whether the matching degree of the primary license plate recognition result and the records in the fixed vehicle list reaches a threshold value;
if not, determining that the vehicle is a temporary departure vehicle, and if so, performing feature extraction on the vehicle to obtain vehicle features;
the background server verifies the vehicle characteristics and judges whether the matching degree of the vehicle characteristics and vehicle characteristic records prestored in the background server reaches a threshold value;
if so, modifying the primary license plate recognition result, and opening and releasing the vehicle; if not, the vehicle is judged to be a temporary departure vehicle.
8. The vehicle departure recognition method according to claim 7, wherein said recognizing the number plate of the vehicle to obtain the primary number plate recognition result comprises:
when the ground sensor detects that the vehicle stops, a trigger signal is sent to the front-end identification device;
and the front-end recognition equipment receives the trigger signal, and performs vehicle license plate recognition on the vehicle parked on the ground sensor to obtain a primary license plate recognition result.
9. The vehicle leaving identification method of claim 7, wherein the step of determining that the vehicle is a temporary leaving vehicle if the matching degree of the primary license plate identification result and the records in the fixed vehicle list does not reach a threshold value comprises:
searching a temporary card entrance record, and judging whether the vehicles have the same entrance record;
if so, opening the brake of the vehicle and releasing the vehicle, and if not, extracting the characteristics of the vehicle to obtain the characteristics of the vehicle;
judging whether the matching degree of the vehicle characteristics and the prestored vehicle characteristic records reaches a threshold value by using a background server;
if so, modifying the primary license plate recognition result, and opening the brake and releasing the vehicle; if not, the vehicle is manually registered and then is opened and released.
10. The vehicle departure identification method according to claim 7, wherein the determining that the vehicle is a temporary departure vehicle if the degree of matching between the primary license plate identification result and the records in the fixed vehicle list reaches a threshold value and the degree of matching between the vehicle feature and the pre-stored vehicle feature record does not reach the threshold value comprises:
searching a temporary card entrance record, and judging whether the vehicles have the same entrance record;
if not, the vehicle is manually registered and then is opened and released, and if yes, the primary license plate recognition result is modified, and the vehicle is opened and released.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611168658.9A CN106780886B (en) | 2016-12-16 | 2016-12-16 | Vehicle identification system and vehicle entrance and exit identification method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611168658.9A CN106780886B (en) | 2016-12-16 | 2016-12-16 | Vehicle identification system and vehicle entrance and exit identification method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106780886A CN106780886A (en) | 2017-05-31 |
| CN106780886B true CN106780886B (en) | 2020-02-04 |
Family
ID=58892258
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611168658.9A Expired - Fee Related CN106780886B (en) | 2016-12-16 | 2016-12-16 | Vehicle identification system and vehicle entrance and exit identification method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106780886B (en) |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| DE102017216562A1 (en) * | 2017-09-19 | 2019-03-21 | Robert Bosch Gmbh | Concept for determining occupancy states of parking spaces and for identifying a parked on a parking space motor vehicle |
| CN111191481B (en) * | 2018-11-14 | 2023-08-25 | 杭州海康威视数字技术股份有限公司 | Vehicle identification method and system |
| CN109584389A (en) * | 2018-11-29 | 2019-04-05 | 上海振华重工(集团)股份有限公司 | A kind of management method in garage parking, device, system and computer storage medium |
| CN110059748A (en) * | 2019-04-18 | 2019-07-26 | 北京字节跳动网络技术有限公司 | Method and apparatus for output information |
| CN112907770A (en) * | 2019-11-19 | 2021-06-04 | 上海擎感智能科技有限公司 | License plate recognition and verification method, system, medium, parking server and client |
| CN111767917A (en) * | 2019-12-24 | 2020-10-13 | 杭州海康威视数字技术股份有限公司 | Violation evidence picture detection method and device and electronic equipment |
| CN113011540A (en) * | 2021-03-01 | 2021-06-22 | 北京骑胜科技有限公司 | Identification method, device, readable storage medium and electronic equipment |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2004133616A (en) * | 2002-10-09 | 2004-04-30 | Toshiba Corp | Toll collecting/receiving system and toll collecting/receiving method |
| CN1632825A (en) * | 2004-12-23 | 2005-06-29 | 中华电信股份有限公司 | Digital monitoring system using dual identification of license plate number and vehicle characteristics |
| WO2014188446A2 (en) * | 2013-05-09 | 2014-11-27 | Tata Consultancy Services Limited | Method and apparatus for image matching |
| CN104915644B (en) * | 2015-05-29 | 2018-01-19 | 安徽四创电子股份有限公司 | A kind of car plate intelligent distinguishing method rebuild based on difference data |
| CN204884185U (en) * | 2015-09-07 | 2015-12-16 | 苏州市世跃智能科技有限公司 | District access & exit control management system |
| CN105279971B (en) * | 2015-10-09 | 2017-11-14 | 北京易华录信息技术股份有限公司 | A kind of information of vehicles matching process, system and supervising device |
-
2016
- 2016-12-16 CN CN201611168658.9A patent/CN106780886B/en not_active Expired - Fee Related
Also Published As
| Publication number | Publication date |
|---|---|
| CN106780886A (en) | 2017-05-31 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106780886B (en) | Vehicle identification system and vehicle entrance and exit identification method | |
| CN106485801B (en) | Vehicles management method, device and its system based on ETC and Car license recognition | |
| CN106384513B (en) | A kind of fake-licensed car capture system and method based on intelligent transportation | |
| WO2021047187A1 (en) | Face recognition-based vehicle charging method, related device and storage medium | |
| CN106652469B (en) | A kind of vehicle marches into the arena, vehicle appearance licence plate recognition method and Vehicle License Plate Recognition System | |
| CN108986539A (en) | Parking management system, method, vehicle information collecting device and management server | |
| CN105741560A (en) | Vehicle identification method used for gas station | |
| CN108986242B (en) | Secondary license plate recognition system and method for expressway non-stop mobile payment lane | |
| CN107305627A (en) | A kind of automobile video frequency monitoring method, server and system | |
| CN104766479A (en) | Automobile identity recognition method and device based on ultrahigh frequency radio frequency and video image dual-recognition matching | |
| CN107591018A (en) | A kind of open section parking management method and system | |
| US11482012B2 (en) | Method for driving assistance and mobile device using the method | |
| CN106097722A (en) | Video is utilized to carry out the system and method for trackside parking stall automatization supervision | |
| CN111898485A (en) | Parking space vehicle detection processing method and device | |
| CN111508091A (en) | Method for quickly searching vehicle information based on ETC vehicle departure | |
| CN103914898B (en) | Vehicles while passing enters control system and method for control | |
| CN105976438A (en) | Method for automatically identifying vehicles for charging in parking lot | |
| CN105225489B (en) | The method and apparatus that non-bus take bus dedicated Lanes evidence collection | |
| CN110766963A (en) | Reverse vehicle searching method, server, device and storage medium | |
| CN103914896B (en) | Vehicles while passing enters the tolerant system of control and the method for control using this tolerant system | |
| CN107730897A (en) | Vehicle identification method and device | |
| CN106327876B (en) | A kind of fake-licensed car capture system and method based on automobile data recorder | |
| CN111696360A (en) | Parking monitoring method, system and camera | |
| JP5991886B2 (en) | Parking system | |
| CN106530739A (en) | License plate recognition method, device and system thereof based on multiple camera device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20200204 |