CN108921095A - A kind of parking occupancy management system neural network based, method and parking stall - Google Patents
A kind of parking occupancy management system neural network based, method and parking stall Download PDFInfo
- Publication number
- CN108921095A CN108921095A CN201810711316.XA CN201810711316A CN108921095A CN 108921095 A CN108921095 A CN 108921095A CN 201810711316 A CN201810711316 A CN 201810711316A CN 108921095 A CN108921095 A CN 108921095A
- Authority
- CN
- China
- Prior art keywords
- gate inhibition
- column
- vehicle
- distance
- neural network
- 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- Withdrawn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
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
- 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
- G07—CHECKING-DEVICES
- G07B—TICKET-ISSUING APPARATUS; FARE-REGISTERING APPARATUS; FRANKING APPARATUS
- G07B15/00—Arrangements or apparatus for collecting fares, tolls or entrance fees at one or more control points
- G07B15/02—Arrangements or apparatus for collecting fares, tolls or entrance fees at one or more control points taking into account a variable factor such as distance or time, e.g. for passenger transport, parking systems or car rental systems
- G07B15/04—Arrangements or apparatus for collecting fares, tolls or entrance fees at one or more control points taking into account a variable factor such as distance or time, e.g. for passenger transport, parking systems or car rental systems comprising devices to free a barrier, turnstile, or the like
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/14—Traffic control systems for road vehicles indicating individual free spaces in parking areas
- G08G1/149—Traffic control systems for road vehicles indicating individual free spaces in parking areas coupled to means for restricting the access to the parking space, e.g. authorization, access barriers, indicative lights
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Multimedia (AREA)
- Business, Economics & Management (AREA)
- Finance (AREA)
- Computer Security & Cryptography (AREA)
- Image Analysis (AREA)
- Traffic Control Systems (AREA)
Abstract
The present invention provides a kind of parking occupancy management system neural network based, method and parking stall, which passes through neural network and prejudge vehicle running path, realizes that controlling corresponding gate inhibition's column opens.Solves the technical issues of existing parking stall management waste of manpower, time.It realizes unmanned management, saves manpower, and use neural network algorithm, calculate precisely, operation stability is high, False Rate is low.
Description
Technical field
The invention belongs to parking management area, relate more specifically to parking occupancy management system neural network based, side
Method and parking stall.
Background technique
With the economic development with science and technology, car ownership increases year by year, and corresponding automobile accessories infrastructure is also continuous
Perfect, e.g., parking access control system is locked to parking stall.Existing Parking Stall is all behavior adjustment management mostly, and parking stall is mostly all
It is to be distinguished by way of scribing line.Also there is minority managing parking space by the way of kickstand.By the way of kickstand, generally
Personal management is needed, when having parking to need, administrative staff can open kickstand, and stop good to vehicle and then locks kickstand,
Whole operation process is troublesome, and waste of manpower wastes time.
Summary of the invention
The technical issues of in order to solve existing parking management waste of manpower, time, the present invention provide a kind of based on nerve
Parking occupancy management system, method and the parking stall of network.
Technical solution of the invention：
A kind of parking stall management method neural network based, includes the following steps：
Step S110, setting gate inhibition's column open distance, open the lateral distance that distance is vehicle distances camera；
Whether step S120, real-time capture image, and real time contrast's former frame and current frame image are identical, if they are the same, then
Continue step S120；If it is different, then carrying out step S130；
Step S130, identifies image, judge it is different whether be because there is vehicle to drive into, if recognizing vehicle,
Carry out step S140；If unidentified arrive vehicle, continue step S120；
Step S140 chooses t moment to the image set at t+1 moment, calculates separately t moment to t+1 moment camera and vehicle
Transverse and longitudinal distance, obtain sampled data set；Sampled data concentrates the cross for including Δ x (t) for t moment vehicle distances camera
To distance, Δ x (t+1) be the lateral distance of t+1 moment vehicle distances camera, Δ y (t) is t moment vehicle distances camera
Fore-and-aft distance and Δ y (t+1) be t+1 moment vehicle distances camera lateral distance；
Step S150 is found the mapping relations of Δ x (t) Yu Δ y (t) by sampled data set, obtains matched curve；
Step S160, according to matched curve, when gate inhibition's column that the lateral distance delta x (t) of vehicle is equal to setting opens distance
When, calculate the corresponding numerical value of anticipation fore-and-aft distance Δ y (t)；
Step S170, real time monitoring captures image, and calculates the transverse and longitudinal distance between vehicle and camera, and sentenced
Disconnected, when actual lateral separation is less than or equal to, gate inhibition's column opens distance, carries out step S180；When actual lateral separation is not up to door
Prohibit column and open distance, continues step S170；
Step S180 is allocated gate inhibition's column according to calculated anticipation fore-and-aft distance Δ y (t), after send commands to
Corresponding gate inhibition's column controller, gate inhibition's column controller will drive the work of gate inhibition's mechanical structure, and corresponding gate inhibition's column is fallen, and vehicle is logical
Row.
In a preferred embodiment of parking stall management method neural network based according to the present invention,
Identification method in step S130 is：The preparatory a large amount of vehicle pictures of typing, model, using image processing algorithm, into
Row vehicle identification.
In a preferred embodiment of parking stall management method neural network based according to the present invention,
Step S150 is specifically included：
Step S151, defining Δ x (t) is input quantity, and Δ y (t) is output quantity；Have between input quantity and output quantity one it is hidden
Layer is hidden, hidden layer uses weighting/sigmoid functional form,Respectively hidden layer input layer weight and hidden layer are defeated
Layer weight out；The calculation formula of output quantity Δ y (t) is：
Q in formulai- evaluation i-th of hidden layer output signal of network；pi- evaluation i-th of hidden layer input signal of network；
Nch- hide number of layers；
Step S152 then constantly regulate neural network weight according to gradient descent method, fits Δ x (t) and Δ y to reach
(t) mapping relations curve；Wherein the calculation formula of neural network weight regulative mode is：
In formula, α (0<α<It 1) is commutation factor/delay index, lcIt (t) is learning rate lc(t)>0。
In a preferred embodiment of parking stall management method neural network based according to the present invention,
It further include step S210：In vehicle pass-through, step S120, S130 is continued to execute, if recognizing vehicle, repeats to walk
Rapid S180, corresponding gate inhibition's column keep full state；If recognizing vehicle, access control system gate inhibition's column liter is sent commands to
It rises.
In a preferred embodiment of parking stall management method neural network based according to the present invention,
Step S180 is specially：According to the spacing and camera distance between anticipation fore-and-aft distance Δ y (t), gate inhibition's column
The distance of first gate inhibition's column, can determine the method for salary distribution of gate inhibition's column.
A kind of parking occupancy management system neural network based of the invention, including monitoring system, gate inhibition's column control system
With the gate inhibition's column controller for rising landing for access control column；
The monitoring system includes image capture module, image processing module, monitoring communication module, and described image acquires mould
Block：The image on real-time capture gate inhibition's column periphery；
Image processing module includes：
Module one：Vehicle is judged whether there is by image to drive into；
Module two：To the image set of the t moment for thering is vehicle to drive into t+1 moment by image calculation method calculate vehicle away from
Horizontal Distance (Δ x (t), Δ y (t)) from image processing module, obtains sampled data set；
Module three：It is constructed according to sampled data set with Δ x (t) as input quantity, Δ y (t) is the matched curve of output quantity；Root
According to matched curve, when gate inhibition's column that the lateral distance delta x (t) of vehicle is equal to setting open apart from when, calculate anticipation it is longitudinal away from
Data corresponding from Δ y (t)；
Monitor communication module：For identification, distance calculating numerical value to be sent to gate inhibition's column control system；
Gate inhibition's column control system includes calculating and storage unit and communication control module, described to calculate and store list
Member：According to the data that monitoring communication module provides, corresponding gate inhibition's column task is distributed；Communication control module：Gate inhibition's column work
The controller of corresponding gate inhibition's column is sent to as task.
The acquisition pattern of three simulation curve of module is：Defining Δ x (t) is input quantity, and Δ y (t) is output quantity；Input quantity with
There is a hidden layer between output quantity, hidden layer uses weighting/sigmoid functional form,Respectively hidden layer is defeated
Enter layer weight and hidden layer output layer weight；The calculation formula of output quantity Δ y (t) is：
Q in formulai- evaluation i-th of hidden layer output signal of network；pi- evaluation i-th of hidden layer input signal of network；
Nch- hide number of layers；
Neural network weight is then constantly regulate according to gradient descent method, to reach the mapping for fitting Δ x (t) Yu Δ y (t)
Relation curve；Wherein the calculation formula of neural network weight regulative mode is：
In formula, α (0<α<It 1) is commutation factor/delay index, lcIt (t) is learning rate lc(t)>0。
In a preferred embodiment of parking occupancy management system neural network based according to the present invention, Men Jinzhu
Controller is motor control or hydraulic control.
In a preferred embodiment of parking occupancy management system neural network based according to the present invention, the door
Single or multiple gate inhibition's column controllers can be controlled by prohibiting column control system.
A kind of gate inhibition's column parking stall, including multiple parking stalls, it is characterised in that：It further include gate inhibition's column and claim 6-
Gate inhibition's column control system described in 9, the inlet on each parking stall are provided with multiple gate inhibition's columns uniformly put.
Of the invention has the advantage that：
1, neural network gate inhibition column control method provided by the present invention realizes unmanned management, saves manpower.
2, neural network gate inhibition column control method provided by the present invention calculates precisely operation using neural network algorithm
Stability is high, False Rate is low.
3, a kind of neural network gate inhibition column control method provided by the present invention prejudges vehicle driving road by neural network
Diameter controls corresponding gate inhibition's column and opens.
Detailed description of the invention
The specific embodiment part provided and referring to the drawings, the features and advantages of the present invention will become more
It is readily appreciated that, in the accompanying drawings：
Fig. 1 neural network gate inhibition's column control method operation schematic diagram；
Fig. 2 neural network gate inhibition's column control system schematic diagram；
Fig. 3 neural network gate inhibition's column control method work flow diagram；
Fig. 4 sampled point transverse and longitudinal distance calculates schematic diagram；
Fig. 5 gate inhibition's column control method neural network model figure；
Fig. 6 transverse and longitudinal is apart from mapping relations matched curve.
Specific embodiment
Exemplary embodiments of the present invention are described in detail with reference to the accompanying drawings.Illustrative embodiments are retouched
It states merely for the sake of demonstration purpose, and is definitely not to the present invention and its application or the limitation of usage.
Embodiment 1：A kind of gate inhibition's column parking lot, including multiple parking stalls further include gate inhibition's column and gate inhibition's column control system
System, the inlet on each parking stall is provided with multiple gate inhibition's columns uniformly put.Gate inhibition's column control system 3 is mountable to gate inhibition
Near column 2,4 be monitoring system, and monitoring system is mountable near gate inhibition's column, and monitoring system is apart from the lateral, vertical of gate inhibition's column
It is known during installation to distance.The operation schematic diagram in parking lot is as shown in Figure 1,1 be vehicle to be driven into figure, 2 be vertical gate inhibition
Column, quantity can be able to be for one it is multiple, gate inhibition's column by motor or hydraulic control, it can be achieved that the lifting of single gate inhibition's column, when rise, vehicle
It can not drive into, when falling, vehicle can be driven into.3 be gate inhibition's column control system, can control single or multiple gate inhibition's columns to go up and down, door
Prohibit column control system 3 to be mountable near gate inhibition's column 2,4 be monitoring system, and monitoring system is mountable near gate inhibition's column, prison
Transverse direction of the control system apart from gate inhibition's column, fore-and-aft distance are known during installation.
Embodiment 2：The working principle of neural network gate inhibition's column control system is as shown in Fig. 2, gate inhibition's column 2 includes gate inhibition's column
Mechanical structure 201, gate inhibition's column controller 202, controller can be motor control, also can be hydraulic control, access control system 3 wraps
It is calculated containing 301 and storage element, communication control module 302, monitoring system 4 includes camera 401, image processing module
402, communication module 403 is monitored.For shooting video image, image processing module 402 is used for the camera 401 of monitoring system 4
Image recognition, positional distance calculate, and monitoring communication module 403 will be for that will identify, distance calculates data and is sent to access control system
The data that the communication control module 302 of system 3, the calculating of access control system 3 and storage element will be provided according to monitoring system 4,
Gate inhibition's column task is distributed, and task is sent to the controller 202 of corresponding gate inhibition's column 2, is connected to gate inhibition's column control of task
Device 202 processed will drive corresponding gate inhibition's column mechanical structure, and access control column rises, falls.
Embodiment 3：Neural network gate inhibition's column control method process is as shown in figure 3, specific step is as follows：
Step S110, setting gate inhibition's column open distance；
Whether step S120, real-time capture image, and real time contrast's former frame and current frame image are identical, if they are the same, then
Continue step S120；If it is different, then carrying out step S130；
Step S130, identifies image, judge it is different whether be because there is vehicle to drive into, if recognizing vehicle,
Carry out step S140；If unidentified arrive vehicle, continue step S120；
Step S140 chooses t moment to the image set at t+1 moment, calculates separately t moment to t+1 moment camera and vehicle
Transverse and longitudinal distance, obtain sampled data set；Sampled data concentrates the cross for including Δ x (t) for t moment vehicle distances camera
To distance, Δ x (t+1) be the lateral distance of t+1 moment vehicle distances camera, Δ y (t) is t moment vehicle distances camera
Fore-and-aft distance and Δ y (t+1) be t+1 moment vehicle distances camera lateral distance；
Step S150 is found the mapping relations of Δ x (t) Yu Δ y (t) by sampled data set, obtains matched curve；
Step S160, according to matched curve, when gate inhibition's column that the lateral distance delta x (t) of vehicle is equal to setting opens distance
When, calculate the corresponding numerical value of anticipation fore-and-aft distance Δ y (t)；
Step S170 step S170, real time monitoring captures image, and calculates the transverse and longitudinal distance between vehicle and camera,
And judged, when actual lateral separation is less than or equal to, gate inhibition's column opens distance, carries out step S180；Work as actual lateral separation
Not up to gate inhibition's column opens distance, continues step S170；
Step S180 is allocated gate inhibition's column according to calculated anticipation fore-and-aft distance Δ y (t), afterwards and sends order
To corresponding gate inhibition's column controller, gate inhibition's column controller will drive the work of gate inhibition's mechanical structure, and corresponding gate inhibition's column will fall, and allow
Vehicle pass-through.
Embodiment 4：In order to improve the accuracy of matched curve, advanced optimize for：
Step S110：Gate inhibition's column is set and opens distance, that is, after vehicle reaches the distance of setting at a distance from gate inhibition's column,
Gate inhibition's column will open (falling).Under original state, all gate inhibition's columns all close (rise).
Step S120：The camera 401 of monitoring system 4 opens work, Real-time captured video image.Image processing module will
Image captured by real time contrast's former frame and present frame shows if it find that current frame image is had any different with previous frame image
There is dynamic image (having mobile object), then, in next step, if it find that current frame image is not different with previous frame image, shows
Without dynamic image (no mobile object), step is repeated.
Step S130：Vehicle identification.After monitoring system 4 captures dynamic image, image processing module 402 can be to Dynamic Graph
As being identified, vehicle is judged whether it is.Judgment mode is a large amount of vehicle pictures of typing, model in image processing module, benefit
With image processing algorithm, vehicle identification is carried out.If recognizing vehicle, show have vehicle that will drive into, in next step, if do not known
It is clipped to vehicle, then, repeats step 2.
Step S140：Image processing module 402 will utilize image processing algorithm in real time to the transverse and longitudinal of vehicle and gate inhibition's column
Distance is calculated.Continuous secondary (N the is natural number) image of N of choosing successively calculates the lateral distance in sampled point as sampled point
Δ x (t), fore-and-aft distance Δ y (t).As shown in figure 4, Δ x (t) is the lateral distance of t moment vehicle distances camera, Δ x (t+
It 1) is the lateral distance of t+1 moment vehicle distances camera, Δ y (t) is the fore-and-aft distance of t moment vehicle distances camera, Δ y
It (t+1) is the lateral distance of t+1 moment vehicle distances camera.Image processing module is taken as N sub-picture is chosen a little, i.e.,
The image that selection t to t+1 moment, camera moment takes calculates separately the cross of t to t+1 camera and vehicle as sampled point
Fore-and-aft distance.
Step S150：Neural network algorithm is embedded in image processing module 402.The purpose of system is to pass through hits
According to, find the corresponding relationship (mapping relations) of Δ x (t) Yu Δ y (t), by calculate mapping relations, judge the following Δ x (t) and Δ
The position distribution situation of y (t), and then judge the movement routine of vehicle.Neural network model is as shown in Figure 5.In figure, Δ x (t) is
Input quantity, Δ y (t) are output quantity.There is a hidden layer between input quantity and output quantity, hidden layer uses weighting/sigmoid letter
Number form formula,Respectively hidden layer input layer weight and hidden layer output layer weight.The purpose of neural network is to pass through
To hidden layer weight Adjusting, fit the corresponding relationship of Δ x (t) Yu Δ y (t).The calculating of output quantity Δ y (t)
Formula is：
Q in formulai- evaluation i-th of hidden layer output signal of network；pi- evaluation i-th of hidden layer input signal of network；
Nch- hide number of layers.
Then according to gradient descent method, calculation formula is neural network weight regulative mode：
In formula, α (0<α<It 1) is commutation factor/delay index, lcIt (t) is learning rate lc(t)>0。
Step S160：System constantly regulate according to the calculation formula of step 5 neural network weight, is fitted with reaching
The mapping relations curve (Fig. 6) of Δ x (t) and Δ y (t).
Step S170：According to matched curve, calculate when vehicle reach with the restriction of gate inhibition's column apart from when, camera and vehicle
Fore-and-aft distance.That is, when Δ x (t) be equal to limit apart from when, the corresponding numerical value of Δ y (t).Know the corresponding numerical value of Δ y (t),
Just work will be opened by being aware of which gate inhibition's column, that is, is carried out position and is judged in advance.Monitoring system 4 is right by calculated Δ y (t)
Numerical value is answered to be sent to the communication module 302 of access control system 3 by communication module 403.
Step 180：Monitoring system 4 persistently captures vehicle, counts to the transverse and longitudinal distance of vehicle and camera
Calculate, when the lateral distance of vehicle and camera be less than or equal to limit apart from when, the communication module 403 of monitoring system 4 sends signal
To the communication module 302 of access control system 3, inform that access control system opens work, if the transverse direction of vehicle and camera
Distance not up to limits distance, then, repeats step 7.
Step 190：After access control system 3 is connected to the information of monitoring system, gate inhibition's column will be allocated, concurrently be lost one's life
It enables to corresponding gate inhibition's column controller 202, gate inhibition's column controller 202 will drive gate inhibition's mechanical structure 201 to work, corresponding gate inhibition
Column will be fallen, and vehicle is allowed to go together.
Step S200：The camera 401 of monitoring system 4 continues working, and image processing module 402 persistently catches vehicle
It catches, catching mode is identical as 2, step 3 is captured.If capturing vehicle, show that vehicle still not entirely through gate inhibition's column, repeats to walk
Rapid 9, corresponding gate inhibition's column keeps full state.If not capturing vehicle, show that vehicle had passed through gate inhibition's column, then, and monitoring
The communication module 403 of system 4 sends commands to the communication module 302 of access control system 3, and access control system 3 will send and order
It enables to corresponding gate inhibition's column controller 202, gate inhibition's column controller 202 will drive corresponding gate inhibition's column mechanical structure 201, driving gate
Prohibit column to rise.
Although referring to illustrative embodiments, invention has been described, but it is to be understood that the present invention does not limit to
The specific embodiment that Yu Wenzhong is described in detail and shows, without departing from claims limited range, this
Field technical staff can make various improvement or modification to illustrative embodiments.
Claims (10)
1. a kind of parking stall management method neural network based, it is characterised in that include the following steps：
Step S110, setting gate inhibition's column open distance, open the lateral distance that distance is vehicle distances camera；
Whether step S120, real-time capture image, and real time contrast's former frame are identical as current frame image, if they are the same, then continue
Carry out step S120；If it is different, then carrying out step S130；
Step S130, identifies image, judges whether different be, if recognizing vehicle, to carry out because there is vehicle to drive into
Step S140；If unidentified arrive vehicle, continue step S120；
Step S140 chooses t moment to the image set at t+1 moment, calculates separately t moment to t+1 moment camera and vehicle
Transverse and longitudinal distance, obtains sampled data set；Sampled data concentrate include Δ x (t) be t moment vehicle distances camera laterally away from
It is the lateral distance of t+1 moment vehicle distances camera from, Δ x (t+1), Δ y (t) is the vertical of t moment vehicle distances camera
It is the lateral distance of t+1 moment vehicle distances camera to distance and Δ y (t+1)；
Step S150 is found the mapping relations of Δ x (t) Yu Δ y (t) by sampled data set, obtains matched curve；
Step S160, according to matched curve, when gate inhibition's column that the lateral distance delta x (t) of vehicle is equal to setting open apart from when, meter
Calculate the corresponding numerical value of anticipation fore-and-aft distance Δ y (t)；
Step S170, real time monitoring captures image, and calculates the transverse and longitudinal distance between vehicle and camera, and judged,
When actual lateral separation is less than or equal to, gate inhibition's column opens distance, carries out step S180；When actual lateral separation is not up to gate inhibition
Column opens distance, continues step S170；
Step S180 is allocated gate inhibition's column according to calculated anticipation fore-and-aft distance Δ y (t), after send commands to it is corresponding
Gate inhibition's column controller, gate inhibition's column controller will drive the work of gate inhibition's mechanical structure, and corresponding gate inhibition's column falls, vehicle pass-through.
2. parking stall management method neural network based according to claim 1, which is characterized in that
Identification method in step S130 is：The a large amount of vehicle pictures of preparatory typing, model carry out vehicle using image processing algorithm
Identification.
3. parking stall management method neural network based according to claim 1, which is characterized in that
Step S150 is specifically included：
Step S151, defining Δ x (t) is input quantity, and Δ y (t) is output quantity；There is one to hide between input quantity and output quantity
Layer, hidden layer use weighting/sigmoid functional form,Respectively hidden layer input layer weight and hidden layer output
Layer weight；The calculation formula of output quantity Δ y (t) is：
Q in formulai- evaluation i-th of hidden layer output signal of network；pi- evaluation i-th of hidden layer input signal of network；Nch- hidden
Hide number of layers；
Step S152, then constantly regulate neural network weight according to gradient descent method, fits Δ x (t) and Δ y (t) to reach
Mapping relations curve；Wherein the calculation formula of neural network weight regulative mode is：
In formula, α (0<α<It 1) is commutation factor/delay index, lcIt (t) is learning rate lc(t)>0。
4. parking stall management method neural network based according to claim 1, it is characterised in that further include step
S210：In vehicle pass-through, step S120, S130 is continued to execute, if recognizing vehicle, repeats step S180, corresponding gate inhibition
Column keeps full state；If recognizing vehicle, the rise of access control system gate inhibition's column is sent commands to.
5. parking stall management method neural network based according to claim 1, it is characterised in that：Step step S180
Specially：According to the distance of anticipation fore-and-aft distance Δ y (t), the spacing between gate inhibition's column and camera distance the first gate inhibition column,
It can determine the method for salary distribution of gate inhibition's column.
6. a kind of parking occupancy management system neural network based, it is characterised in that：Including monitoring system, gate inhibition's column control system
With the gate inhibition's column controller for rising landing for access control column；
The monitoring system includes image capture module, image processing module, monitoring communication module, described image acquisition module：
The image on real-time capture gate inhibition's column periphery；
Image processing module includes：
Module one：Vehicle is judged whether there is by image to drive into；
Module two：Vehicle distances figure is calculated by image calculation method to the image set of the t moment for having vehicle to drive into t+1 moment
As the Horizontal Distance (Δ x (t), Δ y (t)) of processing module, sampled data set is obtained；
Module three：It is constructed according to sampled data set with Δ x (t) as input quantity, Δ y (t) is the matched curve of output quantity；According to quasi-
Close curve, when the lateral distance delta x (t) of vehicle be equal to setting gate inhibition's column open apart from when, calculate anticipation fore-and-aft distance Δ y
(t) corresponding data；
Monitor communication module：For identification, distance calculating numerical value to be sent to gate inhibition's column control system；
Gate inhibition's column control system includes calculating and storage unit and communication control module, the calculating and storage unit：Root
According to the data that monitoring communication module provides, corresponding gate inhibition's column task is distributed；Communication control module：Gate inhibition's column task
It is sent to the controller of corresponding gate inhibition's column.
7. parking occupancy management system neural network based according to claim 6, it is characterised in that：Module three simulates song
The acquisition pattern of line is：Defining Δ x (t) is input quantity, and Δ y (t) is output quantity；There is one to hide between input quantity and output quantity
Layer, hidden layer use weighting/sigmoid functional form,Respectively hidden layer input layer weight and hidden layer output
Layer weight；The calculation formula of output quantity Δ y (t) is：
Q in formulai- evaluation i-th of hidden layer output signal of network；pi- evaluation i-th of hidden layer input signal of network；Nch- hidden
Hide number of layers；
Neural network weight is then constantly regulate according to gradient descent method, to reach the mapping relations for fitting Δ x (t) Yu Δ y (t)
Curve；Wherein the calculation formula of neural network weight regulative mode is：
In formula, α (0<α<It 1) is commutation factor/delay index, lcIt (t) is learning rate lc(t)>0。
8. parking occupancy management system neural network based according to claim 6 or 7, it is characterised in that：Gate inhibition's column control
Device processed is motor control or hydraulic control.
9. parking occupancy management system neural network based according to claim 8, it is characterised in that：Gate inhibition's column control
System processed can control single or multiple gate inhibition's column controllers.
10. a kind of gate inhibition's column parking stall, including multiple parking stalls, it is characterised in that：It further include gate inhibition's column and claim 6-9
Gate inhibition's column control system, the inlet on each parking stall are provided with multiple gate inhibition's columns uniformly put.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810711316.XA CN108921095A (en) | 2018-07-03 | 2018-07-03 | A kind of parking occupancy management system neural network based, method and parking stall |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810711316.XA CN108921095A (en) | 2018-07-03 | 2018-07-03 | A kind of parking occupancy management system neural network based, method and parking stall |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108921095A true CN108921095A (en) | 2018-11-30 |
Family
ID=64423322
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810711316.XA Withdrawn CN108921095A (en) | 2018-07-03 | 2018-07-03 | A kind of parking occupancy management system neural network based, method and parking stall |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108921095A (en) |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050077085A1 (en) * | 2003-10-14 | 2005-04-14 | Rudolf Zeller | Tracking positions of personnel, vehicles, and inanimate objects |
| CN1684105A (en) * | 2004-04-13 | 2005-10-19 | 清华大学 | Automatic registration method for large-scale three-dimensional scene multi-view laser scanning data |
| US20080040029A1 (en) * | 1997-10-22 | 2008-02-14 | Intelligent Technologies International, Inc. | Vehicle Position Determining System and Method |
| CN203982593U (en) * | 2014-07-09 | 2014-12-03 | 黄晓伟 | A kind of multifunctional parking area parking stall manager |
| CN107202982A (en) * | 2017-05-22 | 2017-09-26 | 徐泽宇 | A kind of beacon arrangement calculated based on UAV position and orientation and image processing method |
| CN107545239A (en) * | 2017-07-06 | 2018-01-05 | 南京理工大学 | A kind of deck detection method matched based on Car license recognition with vehicle characteristics |
| CN107578061A (en) * | 2017-08-16 | 2018-01-12 | 哈尔滨工业大学深圳研究生院 | Based on the imbalanced data classification issue method for minimizing loss study |
| CN107688764A (en) * | 2016-08-03 | 2018-02-13 | 浙江宇视科技有限公司 | Detect the method and device of vehicle peccancy |
| CN107767413A (en) * | 2017-09-20 | 2018-03-06 | 华南理工大学 | A kind of image depth estimation method based on convolutional neural networks |
| CN107945566A (en) * | 2017-11-17 | 2018-04-20 | 张慧 | Curb parking management system and method based on multiple target tracking and deep learning |
| CN107967817A (en) * | 2017-11-17 | 2018-04-27 | 张慧 | Intelligent managing system for parking lot and method based on multi-path camera deep learning |
-
2018
- 2018-07-03 CN CN201810711316.XA patent/CN108921095A/en not_active Withdrawn
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080040029A1 (en) * | 1997-10-22 | 2008-02-14 | Intelligent Technologies International, Inc. | Vehicle Position Determining System and Method |
| US20050077085A1 (en) * | 2003-10-14 | 2005-04-14 | Rudolf Zeller | Tracking positions of personnel, vehicles, and inanimate objects |
| CN1684105A (en) * | 2004-04-13 | 2005-10-19 | 清华大学 | Automatic registration method for large-scale three-dimensional scene multi-view laser scanning data |
| CN203982593U (en) * | 2014-07-09 | 2014-12-03 | 黄晓伟 | A kind of multifunctional parking area parking stall manager |
| CN107688764A (en) * | 2016-08-03 | 2018-02-13 | 浙江宇视科技有限公司 | Detect the method and device of vehicle peccancy |
| CN107202982A (en) * | 2017-05-22 | 2017-09-26 | 徐泽宇 | A kind of beacon arrangement calculated based on UAV position and orientation and image processing method |
| CN107545239A (en) * | 2017-07-06 | 2018-01-05 | 南京理工大学 | A kind of deck detection method matched based on Car license recognition with vehicle characteristics |
| CN107578061A (en) * | 2017-08-16 | 2018-01-12 | 哈尔滨工业大学深圳研究生院 | Based on the imbalanced data classification issue method for minimizing loss study |
| CN107767413A (en) * | 2017-09-20 | 2018-03-06 | 华南理工大学 | A kind of image depth estimation method based on convolutional neural networks |
| CN107945566A (en) * | 2017-11-17 | 2018-04-20 | 张慧 | Curb parking management system and method based on multiple target tracking and deep learning |
| CN107967817A (en) * | 2017-11-17 | 2018-04-27 | 张慧 | Intelligent managing system for parking lot and method based on multi-path camera deep learning |
Non-Patent Citations (4)
| Title |
|---|
| HAIHUI XIE 等: ""Vehicle Detection in Open Parks Using a Convolutional Neural Network"", 《2015 SIXTH INTERNATIONAL CONFERENCE ON INTELLIGENT SYSTEMS DESIGN AND ENGINEERING APPLICATIONS (ISDEA)》 * |
| 周培义: ""单目视觉自动泊车控制系统研究"", 《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》 * |
| 孙玉芹: ""车牌自动识别方法研究"", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
| 张波: ""复杂场景下基于视频的停车检测关键技术研"", 《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106652551B (en) | Parking space detection method and equipment | |
| CN108960426A (en) | Road grade Synthesize estimation system based on BP neural network | |
| CN103294775B (en) | Police service cloud image recognition vehicle administrating system based on geographic space-time constraint | |
| CN104851280B (en) | A kind of vehicle travel control method, device, system and relevant device | |
| CN103854292B (en) | A kind of number and the computational methods and device in crowd movement direction | |
| CN106228242B (en) | A kind of high-precision flood routing method based on optimization algorithm | |
| CN105469105A (en) | Cigarette smoke detection method based on video monitoring | |
| CN107168292B (en) | Submarine navigation device circuit failure diagnosis method based on ELM algorithm | |
| CN109726699A (en) | Electric bicycle based on artificial intelligence occupies car lane recognition methods | |
| CN106934378A (en) | A kind of dazzle light identifying system and method based on video depth study | |
| CN102509083A (en) | Detection method for body conflict event | |
| CN106339687A (en) | Self-adaptive pedestrian street-crossing signal calculating method based on video | |
| CN113529667B (en) | Automatic control method and system for integrated gate passing flow | |
| CN108168204A (en) | Controlling method for refrigerator and system | |
| CN110909657A (en) | Method for identifying apparent tunnel disease image | |
| CN102567782A (en) | Neural-network-based automobile engine torque estimation method | |
| CN109034211A (en) | A kind of parking space state detection method based on machine learning | |
| CN110562810B (en) | Elevator dispatching method, device, computer equipment and storage medium | |
| CN108921044A (en) | Driver's decision feature extracting method based on depth convolutional neural networks | |
| CN111223303B (en) | Charging pile parking space management method, device and system | |
| CN108921095A (en) | A kind of parking occupancy management system neural network based, method and parking stall | |
| CN105975982A (en) | Front vehicle detection method | |
| CN109033964A (en) | It is a kind of judgement vehicle to departure from port event method, system and equipment | |
| CN111028519B (en) | Self-adaptive control method based on video flow detector | |
| CN113505346B (en) | Urban street lamp data processing and combined regulation and control system based on artificial intelligence |
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: 20181130 |