The content of the invention
To solve the shortcomings of the prior art, the invention discloses the Itellectualized uptown electric automobile based on Intelligent Recognition to have
Sequence charging management system, first, electric automobile intelligent identification technology is relied on, structure garden electric automobile manages platform concentratedly, real
Charging reservation, in order charging, unified management and scheduling electric automobile are showed;Secondly, Use barriers thing intelligent identification technology, structure
Unified intelligent parking management platform, United Dispatching management is carried out to the parking stall in whole garden;Finally, using to charging process
In real-time status data analysis, realize to the Intellectualized monitoring of charging process and actively protect.
To achieve the above object, concrete scheme of the invention is as follows:
The orderly charging method of Itellectualized uptown electric automobile based on Intelligent Recognition, including:
Realized based on algorithm for pattern recognition to electric automobile Intelligent Recognition:It is unified to establish the electric automobile based on number plate of vehicle
Administrative model, the algorithm of locating license plate of vehicle being combined using color and texture analysis realize that car plate number plate identifies;
Major key of the car plate number plate as management electric automobile in garden, establish the unified management data of electric automobile in garden
Storehouse, basic data is provided for active defense;
Barrier Intelligent Recognition is realized based on obstacle classification device:Capture danger picture signal, using the barrier based on SVM algorithm
Hinder thing grader to realize that data sample is classified, by being trained to sample data, construct the emergency processing under different situations
Prediction scheme, form active defense ATL;
Active defense is realized based on real-time data analysis:Charging pile all in garden is connected by data monitoring platform,
Extract in real time and analyze the status data of charging pile, corresponding emergency processing in active defense ATL is called according to different situations
Prediction scheme and then realize to the active defense of charging pile state.
Two Intelligent Recognitions and an active defense are all based on what recognizer and Internet of Things technology were realized, wherein first
Recognition methods mainly using graph image recognizer identification car plate, car plate can be used as in garden management electric automobile major key,
The unified management database of electric automobile in garden is established by first Intelligent Recognition, it can provide basis for active defense
Data, such as by establishing the electric automobile being good for based on car plate, it is possible to dynamically store battery mould on specific electric automobile
The status information of block, it is orderly charging service, and then data is provided for the active defense based on real-time data analysis.
The second part is relatively independent, but is also that the sources of data is provided for Part III, it break the barriers sensing,
Identification, classification, to judge whether parking stall free time, charge position have situations such as obstruction.Charge, charged in order for Part III
Journey security provides data and supported.
The linear classifier provided using SVM algorithm, by being trained to sample data, is constructed under different situations
Plan for emergency handling, form active defense ATL.
Further, when realizing number plate identification, each entrance and exit passage in garden is installed to be taken the photograph with high definition
As the control machine of head, when being waited based on the principle of " car of a bar one ", each vehicle into garden before bar, high-definition camera meeting
Trick-plate is taken pictures, and photo upload is handled to system background, and system background uses the intelligence based on algorithm of locating license plate of vehicle
Can recognition methods progress Car license recognition.
Further, the intelligent identification Method based on algorithm of locating license plate of vehicle includes:
A, using color space distance and Similarity Measure, desired color region is partitioned into from image;
B, background colour and license plate number are isolated from region using color contrast algorithm;
C, number plate of vehicle is identified using graph and image processing algorithm.
Further, SVM realizes that data sample is classified by constructing optimal separating hyper plane, is built using linear kernel function
Model.
Further, when capture, cognitive disorders thing, high-definition camera and laser range finder capture image information, carry out
Image procossing, the standard picture information that can be used to parse is formed, the master sample of current object is obtained by image analysis algorithm
Data, basic data is provided so as to screen barrier for next step sample training.
Further, image information is captured, carries out image procossing, the pretreatment work of image is carried out first, by static state
Image is processed into the image information of LINEAR CONTINUOUS by the comparison of front and rear frame, then passes through noise, the processing of filtering and definition
Reduction, ultimately form the standard picture information that can be used to parse.
Further, barrier examination is handled, and barrier, event handling are screened using the obstacle classification device based on SVM
Function.Event handling function, based on SVM algorithm, by the training to dynamic sample data, obtains hyperplane (danger when realizing
Dangerous event boundaries) result set, and then utilize this result set, build hazard event classification model, utilize different result sets
Performance at difference structure template, each classification model is corresponding with specific settling mode, ultimately forms at dynamic event
Put function.
Further, real time data inspecting platform is built, monitor supervision platform also includes:It is extra by being installed additional on electric automobile
Battery cell monitoring sensor device, realize the real-time acquisition of the status information of the vehicle-mounted battery of electric automobile to be charged to needs, and
Analysis is extracted by the implementation data monitoring platform of system, and then realizes the state analysis to charging process on-vehicle battery.
The orderly charging system of Itellectualized uptown electric automobile based on Intelligent Recognition, including:
Electric automobile intelligent identification module:The electric automobile unified management model based on number plate of vehicle is established, utilizes color
Realize that number plate identifies with the algorithm of locating license plate of vehicle that texture analysis is combined;
Barrier intelligent identification module:Capture danger picture signal, number is realized using the obstacle classification device based on SVM algorithm
According to sample classification;
Active defense module:Charging pile all in garden is connected by data monitoring platform, extracts and analyzes in real time and fill
The status data of electric stake, and then realize the monitoring to charging pile state;
The linear classifier provided using SVM algorithm, by being trained to sample data, is constructed under different situations
Plan for emergency handling, form active defense ATL.
Further, when electric automobile intelligent identification module realizes number plate identification, each entrance and exit in garden
Passage all installs the control machine with high-definition camera, and based on the principle of " car of a bar one ", each vehicle into garden is in bar
During preceding wait, high-definition camera meeting trick-plate is taken pictures, and photo upload to system background is handled, system background uses
Intelligent identification module based on algorithm of locating license plate of vehicle carries out Car license recognition.
Further, the intelligent identification module based on algorithm of locating license plate of vehicle includes:
Using color space distance and Similarity Measure, desired color region is partitioned into from image;
Background colour and license plate number are isolated from region using color contrast algorithm;
Number plate of vehicle is identified using graph and image processing algorithm.
Further, in barrier intelligent identification module, in addition to:
Sample classification module, SVM realizes that data sample is classified by constructing optimal separating hyper plane, using linear kernel function
Build model.
Image collection module, capture, cognitive disorders thing when, high-definition camera and laser range finder capture image information,
Image procossing is carried out, forms the standard picture information that can be used to parse, the standard of current object is obtained by image analysis algorithm
Sample data, basic data is provided so as to screen barrier for next step sample training.
Image processing module, image information is captured, carry out image procossing, carry out the pretreatment work of image first, will
Still image is processed into the image information of LINEAR CONTINUOUS by the comparison of front and rear frame, then passes through noise, the processing of filtering and clear
The reduction of clear degree, ultimately form the standard picture information that can be used to parse.
Barrier screens processing module, and barrier, event handling function are screened using the obstacle classification device based on SVM.
Further, real time data inspecting platform is built in active defense module, monitor supervision platform also includes:By electronic
Install extra battery cell monitoring sensor device on automobile additional, realize the status information of the vehicle-mounted battery of electric automobile to be charged to needs
Real-time acquisition, and analysis is extracted by the implementation data monitoring platform of system, and then realize to charging process on-vehicle battery
State analysis.
Beneficial effects of the present invention:
, can be to electric automobile in whole garden, from entering 1. from the lifting of simple charge management module to whole process supervision
Garden, parking, charging process monitor in real time, improve the intelligent level of garden, improve the security cocoa in charging process
Monitoring property.
2. being designed for concrete scene, level of integrated system is high, has significant work to the intelligent level for lifting whole garden
With.
3. the active defense based on real time data, by early warning passive during dangerous generation during former charging electric vehicle
Mode is changed to the active defense mode for integrating " monitoring, early warning, dangerous disposal ", will be endangered in intelligent aspect as far as possible
Danger disposes.
Embodiment:
The present invention is described in detail below in conjunction with the accompanying drawings:
The orderly charging management system of Itellectualized uptown is to be managed collectively electric automobile parking, the information of charging in Itellectualized uptown to put down
Platform, system include Intelligent Recognition, in order charging, the broad aspect of active defense three.Wherein Intelligent Recognition is the basis of system, by building
It is vertical to electric automobile identification, charging obstacle recognition, three proofings, temperature control identification, realize the garden that come in and gone out to electric automobile, parking,
The unified monitoring of charging and management, reduce human intervention, greatly improve the automatization level of Itellectualized uptown and electronic vapour
Car charging safety.
Intelligent Recognition:It is real based on the intelligent identification module of the technique constructions such as Internet of Things, pattern-recognition, small sample statistical analysis
Now electric automobile is come in and gone out the Intelligent Recognition of garden, in garden parking stall idle management, the intelligent monitoring of charging process, greatly
Improve the automaticity of Itellectualized uptown.
Unified management:Itellectualized uptown electric automobile management system is established, unified pipe is carried out to the electric automobile for the garden that comes in and goes out
Reason, including vehicle discrepancy garden record, charging service management, parking service etc..
Active defense:By in charging process, the real-time monitoring and analysis of the status data such as charger, battery, utilizing
Intelligent identification Method, the complete monitoring to charging process is realized, effectively raise the security of charging process.
Intelligent identification Method is summarized:Intelligent Recognition is that one kind is integrated with the skills such as Internet of Things, pattern-recognition, small sample statistical analysis
The recognition methods of art, wherein Internet of Things technology provide a kind of mode of various physical equipments in UNICOM garden.It is pair that pattern, which is not known,
The process that subject or behavior in perception data image, video, sound etc. are differentiated and explained.Realized with machine simulation
The recognition capability of people, it is the vital task of pattern-recognition information processing.
The advantages of intelligent identification Method:
1. cost is cheap, processing is efficient
Compared with traditional manual identified, Intelligent Recognition depends on the disposal ability of machine, has benefited from microelectronics industry
Fast development, only needs several common servers now, and some external sensors can be achieved with being previously required to a large amount of human and material resources
The identification work that could be completed.Use cost is greatly reduced, while is had in the efficiency and precision of identification very big
Lifting.
2. level of integrated system is high
Intelligent Recognition is the process of machine simulation manual identified, briefly, exactly obtains external data, special by one group
Fixed software recognizer can realizes the process management of identification, therefore Intelligent Recognition day is born with the spy with Seamless integration-
Point, with reference to corresponding operation flow, you can easily realize operation flow automation, process management and control.
Application study of the intelligent identification Method in the orderly charging management system of Itellectualized uptown electric automobile
In this section, the orderly charging management system case of Itellectualized uptown electric automobile will be combined to study intelligent identification Method
Application.
The above learns that Intelligent Recognition is the aggregate of multiple technologies, the reality for the business that charged in order with reference to Itellectualized uptown
Border situation, in terms of Intelligent Recognition is mainly reflected in following three:1st, the electric automobile intelligence based on pattern (image) recognizer
Identification;2nd, the barrier Intelligent Recognition based on obstacle classification device;3rd, the active protection method based on real-time data analysis.
Electric automobile Intelligent Recognition based on image recognition algorithm:
Electric automobile is the business datum for having system most crucial, on this core data, can easily be realized such as:In advance
The about a series of business such as charging, battery detecting tracking.Therefore how efficient unique information for confirming electric automobile turns into electricity
The key problem of electrical automobile Intelligent Recognition.
Electric automobile unified management model based on number plate of vehicle:
Based on this, it is necessary to first confirm vehicle unique sign, it is clear that for one can be with the automobile on legal upper road for,
Number plate of vehicle can uniquely confirms this car.Number plate of vehicle identification is computer vision and mode identification technology in intelligent friendship
One of important topic of logical field application, can apply to the occasions such as Itellectualized uptown room entry/exit management, garage management.If can machine reality
When identify vehicle, can be realized in the case where other special devices need not be installed additional for vehicle to the automatic monitoring of vehicle and pipe
Reason.
The algorithm of locating license plate of vehicle being combined based on color and texture analysis:
During license auto-recognition system, License Plate, Character segmentation, character recognition are its key technologies, and with car
Board is positioned as core technology.But weather condition of License Plate when being limited by shooting, humidity, car plate clean level, car plate are multiple
The conditions such as miscellaneous degree, positioning licence plate difficulty is very big, existing extraction licence plate method include artificial neural network mathematical morphology,
The methods of Hough transform, signal transacting, but these methods are generally preferable to edge color treatments, to background and number plate size
Processing shows slightly deficiency, therefore the License Plate being combined by the experimental comparison of reality, Systematic selection color and texture analysis
Algorithm.
Operation principle:The car plate pattern of China has hard and fast rule requirement, and civilian car plate divides blue bottom wrongly written or mispronounced character or yellow bottom surplus,
Its specific array format is:X1X2X3X4X5X6X7 (only discuss here the special car plate such as civilian car plate, army, police except), its
In, X1 is the Chinese character abbreviation of provinces, autonomous regions and municipalities, and X2 is English alphabet, represents districts and cities' code (word in vehicle ownership province
Female O and W represent car for public affairs), such as:Shandong A represents Jinan, and Shandong B represents Qingdao.X3X4X5X6X7 is English or Arabic numerals
Combination, represent the physical number of car plate.In addition to Chinese character, English alphabet connects with Arabic numerals in vertical direction.Its
It is secondary, the length of car plate and width homeostasis, and the mutual distance of intercharacter is fixed in car plate, the width of character with it is highly identical.
In summary, the operation principle of number plate fixation and recognition can simply be understood as the size edge of trick-plate image, face
Color, word, the identification process of numeral, its operation principle realize that logical process is as follows:1st, color space distance and similarity are utilized
Calculate, desired color region is partitioned into from image;2nd, background colour and car are isolated from region using color contrast algorithm
The trade mark;3rd, number plate of vehicle is identified using graph and image processing algorithm.
System is realized
Based on operation principle, system realizes that number plate identifies using high-definition camera auxiliary.
Each entrance and exit passage in garden installs the control machine with high-definition camera, based on " a bar one
When the principle of car ", each vehicle into garden wait before bar, high-definition camera meeting trick-plate is taken pictures, by photo
System background is passed to be handled, system background carries out Car license recognition using the intelligent identification module based on algorithm of locating license plate of vehicle,
Realize the intelligent recognition function to vehicle.
Realize Vehicular intelligent be it is other system can be that each vehicle for coming in and going out garden establish vehicle archive information simultaneously,
Number plate identification technology is relied on, realizes that the automation for the garden that come in and gone out to vehicle is managed collectively.
Barrier Intelligent Recognition algorithm based on obstacle classification device:
Parking management is one of Core Feature of system, and system has had been built up the vehicle management system of the overall situation, but nothing
Method is solved the problems, such as to electric automobile parking management.Among reality, parking stall quantity and position are relatively-stationary in garden, therefore
The problem of parking management, can simply be converted into the searching problem to idle parking stall, be exactly to judge whether parking stall is occupied in brief
With the problem of, therefore, system provides the parking management module based on barrier Intelligent Recognition.
Operation principle:It is general during analysis because the experimental data that barrier may be referred to is less in actual motion
All over being faced with this insufficient problem of sample data, thus system using the obstacle classification device based on SVM algorithm as barrier
The core of identification.
SVMs (Support Vector Machine, SVM) be the 1990s mid-term by Vapnik V
Et al. the machine learning algorithm based on Statistical Learning Theory that proposes, it is minimizing sample point according to limited sample information
While error, the upper bound of model predictive error is reduced, so as to improve the generalization ability of model, main application fields are pattern
Identification, regression forecasting, PDF estimation etc..
SVM is to realize that data sample is classified by constructing optimal separating hyper plane.Studied at present in terms of classification more
Also more common kernel function has:1) linear kernel function;2) Polynomial kernel function;3) Radial basis kernel function;4) Sigmoid cores letter
Number.The system is using linear kernel function structure model.
Implementation:The core missions of grader be exactly capture, cognitive disorders thing.Therefore system needs the function bag realized
Containing two parts:One image signal processing blocks being made up of high-definition camera, one is that the barrier that grader forms is discriminated
Other processing module.Wherein image signal processing blocks mainly realize two functions, and one is that the picture signal of potential barrier is caught
Catch, one is image processing function after capturing signal.Barrier screens processing module and also includes two functions, and one is to make
Barrier is screened with the obstacle classification device based on SVM, one is event handling function.Image signal processing blocks core is appointed
Business is barrier sample collection and pretreatment.
System coordinates the function of reaching sample collection, wherein high definition using two hardware of high-definition camera and laser range finder
Camera is used for capturing danger picture signal.As shown in figure 1, completely realize that logic is as follows:
Capture images, image preprocessing, image procossing, noise, filtering, definition processing, output, standard picture information.
Wherein high-definition camera and laser range finder capture image information, carry out the pretreatment work of image first, and
Still image is processed into the image information of LINEAR CONTINUOUS by the comparison of front and rear frame, second passes through noise, the processing of filtering
And the picture of the reduction of definition, especially haze sky shooting, it is necessary to by processing, ultimately form the standard that can be used to parse
Image information.
Standard picture information is to extract the premise of sample data, and the mark of current object can be obtained by image analysis algorithm
Quasi- sample data.Basic data is provided so as to screen barrier for next step sample training.System realizes that figure is as shown in Figure 2:Fig. 2
It is the Operation interface diagram of system, system provides some conventional image processing algorithms.
First photo and second photo are the original graphs of the image capture in running and passed through at image
Standard picture after reason.
Active protection method based on real-time data analysis:
Charging is one of three Core Features of system to electric automobile in order, and how to ensure that charging process is autonomous safely and enter
Row is the difficult point of charging security protection., it is necessary to monitor the situation of batteries of electric automobile, the shape of charger in real time in charging process
State, to prevent the accident that may occur, found by the analysis to actual charging process, most important security protection concentrates on three
Anti-, car is in battery temperature control etc..
In consideration of it, system provides a kind of charging pile, the active protection method of on-vehicle battery real-time status data analysis comes
Ensure the active safety of charging process.
System is realized:In order to realize this target, two modules of system constructing, first is that real time data inspecting is put down
Platform.Second is active defense template.Using the real time data in charging process, realized by active defense template to entirely filling
The active defense of electric process.
Each charging pile has the state monitoring module of oneself, and these data can react the current shape of charging pile in real time
State, charging pile all in garden is connected by data monitoring platform, extracts in real time and analyzes the status data of charging pile, and then
Realize the monitoring to charging pile state.
The vehicle-mounted battery of electric automobile is another important step that safe charging process needs to monitor, by electronic vapour
Install extra battery cell monitoring sensor device on car additional, realize the status information of the vehicle-mounted battery of electric automobile to be charged to needs
Obtain in real time, and analysis is extracted by the implementation data monitoring platform of system, and then realize the shape to charging process on-vehicle battery
State is analyzed.
Active defense template is similar to emergency plan, is the process of a supervised machine study, is provided using SVM algorithm
Linear classifier, by being trained to sample data, the plan for emergency handling under different situations is constructed, forms active defense
ATL.Fig. 3 is to the active defense training result of an actual charging process in system.
By the training to sample data, it is superimposed by the probability of more statistic algorithms, whether prejudge can produce danger,
Such as upper figure is exactly that anticipation has greater probability to cause danger, system can call emergency preplan to interrupt charging process, ensure charging peace
Entirely.
Intelligent identification Method is applied in the orderly charging management system of electric automobile, by establishing three intelligent protection subsystems
System, the system for integrating vehicle room entry/exit management, parking management, charging process protection etc. is realized, effectively raises intelligence
The automaticity of garden electric automobile management, reduces potential safety hazard in charging process, for being widely applied for Itellectualized uptown
Provide positive case.
Although above-mentioned the embodiment of the present invention is described with reference to accompanying drawing, model not is protected to the present invention
The limitation enclosed, one of ordinary skill in the art should be understood that on the basis of technical scheme those skilled in the art are not
Need to pay various modifications or deformation that creative work can make still within protection scope of the present invention.