CN207731293U - A kind of high robust image identification system based on SVG dynamic charts - Google Patents
A kind of high robust image identification system based on SVG dynamic charts Download PDFInfo
- Publication number
- CN207731293U CN207731293U CN201721708122.1U CN201721708122U CN207731293U CN 207731293 U CN207731293 U CN 207731293U CN 201721708122 U CN201721708122 U CN 201721708122U CN 207731293 U CN207731293 U CN 207731293U
- Authority
- CN
- China
- Prior art keywords
- box
- parsing
- cable
- usb
- identification system
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Landscapes
- Character Input (AREA)
Abstract
The utility model is related to a kind of high robust image identification system based on SVG dynamic charts, designed for the technical issues of existing similar image-recognizing method lacks load mounting device is solved.The image identification system is set to the memory in parsing box, that is, the circuit board parsed in box is equipped with memory;Its main points is that the parsing box is in regular hexagon, box is parsed to connect by the cable interface of cable and computer, that is one end of parsing box is equipped with cable connecting port, the other end is equipped with cable connector lug and USB connector lugs, the cable for connecting computer is connect with the cable connecting port of parsing box, the network interface card of the cable connector lug and computer that parse box connects, and the USB interface of the USB connector lugs and computer that parse box connects.Meanwhile parsing the circuit board in box and being equipped with lithium battery and mobile wifi routers, parsing box is equipped with operating mode switching switch, usb expansion mouth and work light, and the side for parsing box is equipped with card reader.
Description
Technical field
The utility model is related to the mounting boxs of image identification system, are a kind of high robust figures based on SVG dynamic charts
As identifying system.
Background technology
Currently due to protecting the needs of data, many websites not to directly read data by the page, but by data conversion
To show in the page for the chart of SVG formats.SVG formats are a kind of vector graphics languages of the dynamic generation when the page loads,
User directly describes image with code, with any word processing tool open SVG image, makes figure by changing section code
It as having interactive function, and is inserted into HTML and is watched by browser at any time, so traditional crawler technology can not be effective
Crawl SVG chart-informations.And the case where data often will appear shortage of data is read in the way of image recognition.Work as data
When missing, chart can normally be shown, but information can't see;In addition the accuracy rate of image recognition also leverages data acquisition
Accuracy, the influence of a variety of noises is combined together, and makes to crawl SVG chart datas that there are few may.Simultaneously as SVG charts
Belong to dynamic chart, traditional mode effect for crawling static website is little;It is carried out for each coordinate points in SVG charts
Image recognition, due to accuracy rate, identification is more, and the data of mistake are more.And identify elapsed time with coordinate
Increase and sharply increase;The value that mistake is shown in value or SVG charts for the loss of data in SVG charts can not
Effectively handled.Main cause is:The SVG chart web page contents of dynamic load are not handled;To the mistake of picture recognition
Tolerance is relatively low, can not effectively avoid mistake.But existing similar image-recognizing method is more difficult to be used for HTML static text contents, not
Using picture OCR identification technologies, the real data acquisition methods of data screening and coordinate value are not good enough.Meanwhile existing congenic method
It is less to use high robust image-recognizing method, it is so-called " robustness ", refer to control system certain(Structure, size)Parameter
Under perturbation, the characteristic of other certain performances is maintained.Such as computer software input error, disk failure, network over loading or
In the case of intentional attack, it can not crash, not collapse, be exactly the robustness of the software.According to the different definition to performance, it is divided into
Stability robustness and performance robustness, the static controller obtained using the robustness of closed-loop system as target design are known as robust
Controller.
Invention content
It is to provide a kind of height based on SVG dynamic charts to this field to overcome above-mentioned deficiency, the purpose of this utility model
Robustness image identification system makes it mainly solve existing similar image-recognizing method and lacks load mounting device and image
Identification, data screening, data acquisition use more inconvenient technical problem.The purpose is to what is be achieved through the following technical solutions.
A kind of high robust image identification system based on SVG dynamic charts, the image identification system are set to parsing box
Interior memory, that is, the circuit board parsed in box are equipped with memory;It is characterized in that the parsing box passes through cable and computer
Network interface card connection, that is, parse box one end be equipped with cable connecting port, the other end be equipped with cable connector lug and USB connector lugs, connection
The cable of computer is connect with the cable connecting port of parsing box, and the network interface card of the cable connector lug and computer that parse box connects, solution
The USB interface of the USB connector lugs and computer of analysing box connects.High robust image is equipped in memory in above-mentioned parsing box to know
Other software systems conciliate code authentication system, and the high robust image recognition software system to parse memory in box passes through USB
Connector lug is loaded into computer automatically, and the web page address opened by computer is loaded directly into the memory for being recorded in parsing box
In, while parsing box is as Network Cable Connector, there is gateway-filter and writing function;Meanwhile supporting the image-recognizing method
The legal use and systematic difference, sale of design software.
Circuit board in the parsing box is equipped with lithium battery and mobile wifi routers, and parsing box switches equipped with operating mode
Switch, usb expansion mouth and work light, operating mode switching switch, usb expansion mouth and work light pass through circuit and electricity
Road plate connection.Consequently facilitating parsing box is used as USB connection expanders, and mobile wifi routers.
The parsing box is in regular hexagon, and cable connector lug and USB connector lugs are located at regular hexagon the same side of parsing box
Side, symmetrical another side are equipped with cable connecting port.To which usb expansion mouth is arranged in remaining side, parsing one side plane of box is set
Set operating mode switching switch and work light;Or remaining a side setting operating mode switches switch, other sides
It is set to parsing one side plane of box setting usb expansion mouth and work light and changes side plane.
The side of the parsing box is equipped with card reader, and card reader is connect by circuit with circuit board.Also to the parsing box
It can be used as card reader use.
The utility model makes image identification system application, installation, using more convenient using multi-functional parsing box, on
SVG chart-informations are realized automatic crawl and are read in network process;It is suitable for the image recognitions of SVG dynamic charts and portable
USB connections expander, mobile wifi routers, Network Cable Connector, card reader use.
Description of the drawings
Fig. 1 is the process blocks schematic diagram of the utility model.
Fig. 2 is that certain website of the utility model includes SVG dynamic chart content schematic diagrames.
Fig. 3 is that the simulation of the utility model is clicked to obtain maximum value prompt curve graph.
Fig. 4 is that the simulation of the utility model is clicked to obtain minimum value prompt curve graph.
Fig. 5 is that the SVG dynamic charts of the utility model parse box structural schematic diagram, and dotted line is cable connecting port in figure.
Figure number and title:1, box is parsed, 2, cable connecting port, 3, cable connector lug, 4, USB connector lugs, 5, work
Mode selector switch, 6, usb expansion mouth, 7, work light.
Specific implementation mode
In conjunction with attached drawing, the utility model structure and use are further described.The overall flow of the image recognition side
As shown in Figure 1, using the SVG dynamic charts content that certain website includes as specific implementation case, as shown in Figure 2.Separately below into
Row is described in detail:1, SVG dynamic charts are extracted:By " developer's tool " function of browser, SVG dynamic charts are found
Label;Analyzing tags content finds the coordinate value of each coordinate points;2, coordinate value group is chosen:It is chosen in all coordinates
Two points of maximum value and minimum value, then three groups of coordinate values are randomly selected, five groups of coordinate values are obtained altogether;This five groups of coordinate values are protected
Card cannot repeat, if it find that repeating, randomly select again;3, coordinate points real data display diagram is obtained:It is controlled by program
Browser loads corresponding URL(Uniform Resource Locator, uniform resource locator)Content, include in load
After the content page of SVG dynamic charts, need to find out target SVG coordinate points position according to analysis;It is simulated using program actual
Mouse click event promotes page dynamic load to include the HTML of practical business meaning numerical value(Hyper Text Markup
Language, hypertext markup language)Object layer.Simulation, which is clicked, obtains maximum value prompt:As shown in Figure 3;Simulation is clicked and is obtained
Minimum value prompts:As shown in Figure 4.According to the position rule of dynamic reminding object that analysis obtains, its boundary of automatic program identification,
Realize the picture interception in dynamic reminding region.
4, OCR identifies the practical expression data of every group of coordinate, that is, the picture after intercepting needs to carry out two-value for ease of identification
Then the pretreatment of change, picture amplification, interpolation etc. carries out the number identification of picture, obtains SVG dynamic chart ordinate scale generations
The numerical value of table.In this process, since OCR recognition accuracies cannot ensure to reach a satisfactory value, and previous step obtains
Real data display diagram itself will appear abnormal number or void value.So to be sieved to result after OCR identifications
Choosing is handled, and gets rid of abnormal results.By random selection coordinate points before, it is enough to ensure that this step is bound to get
The value of information, and then ensure that the accuracy of data acquisition.
5, the practical ratio indicated between number and coordinate is calculated:Five groups of coordinate values are pairwise grouping, " formula is utilized
1 " calculates the practical ratio indicated between number and coordinate.
(Formula 1)
Wherein, the actual numerical value that v denotation coordinations represent, y represent the ordinate of coordinate points.By vertical seat before calculating every time
Mark is compared, and ensures that the symbol of final result is correct.If each group of coordinate, which all identifies, accurately will produce completely the same 20
A ratio value.As shown in the table:
10000 | 10000 | 10000 | 10000 | ||
10000 | 10000 | 10000 | 10000 | ||
10000 | 10000 | 10000 | 8160 | ||
10000 | 10000 | 10000 | 10000 | ||
10000 | 10000 | 10000 | 10000 |
Method by setting a threshold value, occurrence number are decided to be final ratio value more than threshold value, exclude knowledge
Not Bu Zhun or data exception coordinate value(Bold numerals in table), it is further ensured that accuracy.Meanwhile it will not be chosen in table
The value taken is as exceptional value(,), the corresponding coordinate of exceptional value is set as abnormal coordinate.Finally, selection one is non-
The reference coordinate that abnormal coordinate is calculated as next step(In table,,Middle selection).
6, the real data of all coordinate values is derived:Although the method that can utilize image recognition recycles all coordinate points
It carries out similar process, but when chart coordinate points are more, seriously affects efficiency.For optimization processing speed, analogy thinking is taken
Mode come calculate other each coordinate points ordinate represent numerical value.Specific formula such as " formula 2 " is as follows:
(Formula 2)
Wherein, y is the ordinate to be calculated, and v and k are between the real data of reference coordinate and real data and coordinate
Ratio;The corresponding statistical result value of all coordinates by being derived by statistical graph, and then complete statistical result is obtained, it is real
Now capture purpose.
In addition, the above-mentioned high robust image-recognizing method is installed on by software form load in computer, the Gao Lu
Stick image-recognizing method can also be fabricated to software and be nested in parsing box 1 while sell and using.As shown in figure 5, parsing box
In regular hexagon, the concrete structure for parsing box is as follows:The one end for parsing box is equipped with cable connecting port 2, and the other end connects equipped with cable
The end of a thread 3 and USB connector lugs 4, the cable for connecting computer are connect with the cable connecting port of parsing box, parse the cable connector lug of box
It is connect with the network interface card of computer, the USB interface of the USB connector lugs and computer that parse box connects.Meanwhile parsing the circuit in box
Plate is equipped with lithium battery and mobile wifi routers, and parsing box is equipped with operating mode switching switch 5, usb expansion mouth 6 and work instruction
Lamp 7, operating mode switching switch, usb expansion mouth and work light are connect by circuit with circuit board.Above-mentioned parsing box passes through
USB connector lugs load in the software to computer for installing the high robust image-recognizing method, and the cable by parsing box connects
The end of a thread and cable connecting port connected with network cable, meanwhile, parsing box can also be used as webpage record in addition to being used as software installation disk
Device, mobile memory, USB connections expander, mobile wifi routers, Network Cable Connector use;The side for parsing box is equipped with reading
Card device, card reader are connect by circuit with circuit board, to which the parsing box can also be used as card reader use.
In conclusion the utility model create a kind of high speed, high robust SVG dynamic charts data content analysis side
Method meets industry-by-industry and encounters targeted website chart in data acquisition using SVG dynamics towards general web crawlers aspect
Graph mode is realized, the concrete methods of realizing of dynamic chart data content is captured using particular technology gimmick.The utility model base
In the multi-point recognizing method of SVG dynamic charts, the component values by obtaining SVG charts derive all numerical value;I.e. this practicality is new
Type knows the robustness that method for distinguishing improves system using multiple spot OCR, and ratio value between number and coordinate is indicated using reality is calculated
Mode, avoid the identification for carrying out all coordinates, improve the speed of service, reduce run time.The utility model identifies multiple spot
Technology is applied to the acquisition of SVG dynamic chart contents, reduces the influence of abnormal data and OCR wrong identifications to data acquisition, than
Existing technology accuracy rate change.Method by randomly selecting coordinate points, and then extrapolate chart practical significance numerical value and seat
Association between punctuate;The data represented by the coordinate points in all figures are calculated, the identification one by one to all data is avoided, than
Existing technical speed is faster.
Claims (4)
1. a kind of high robust image identification system based on SVG dynamic charts, the image identification system are set to parsing box(1)
Interior memory, that is, the circuit board parsed in box are equipped with memory;It is characterized in that the parsing box(1)Pass through cable and calculating
The network interface card of machine connects, that is, the one end for parsing box is equipped with cable connecting port(2), the other end is equipped with cable connector lug(3)With USB wiring
Head(4), connect the cable of computer and connect with the cable connecting port of parsing box, parse the cable connector lug of box and the net of computer
The USB interface of card connection, the USB connector lugs and computer that parse box connects.
2. the high robust image identification system of SVG dynamic charts according to claim 1, it is characterised in that the parsing
Box(1)Interior circuit board is equipped with lithium battery and mobile wifi routers, and parsing box is equipped with operating mode switching switch(5), USB expand
Open up mouth(6)And work light(7), operating mode switching switch, usb expansion mouth and work light pass through circuit and circuit board
Connection.
3. the high robust image identification system of SVG dynamic charts according to claim 2, it is characterised in that the parsing
Box(1)In regular hexagon, cable connector lug(3)With USB connector lugs(4)Positioned at the same side of regular hexagon of parsing box, symmetrically
Another side be equipped with cable connecting port(2).
4. the high robust image identification system of SVG dynamic charts according to claim 2, it is characterised in that the parsing
Box(1)Side be equipped with card reader, card reader connect by circuit with circuit board.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201721708122.1U CN207731293U (en) | 2017-12-11 | 2017-12-11 | A kind of high robust image identification system based on SVG dynamic charts |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201721708122.1U CN207731293U (en) | 2017-12-11 | 2017-12-11 | A kind of high robust image identification system based on SVG dynamic charts |
Publications (1)
Publication Number | Publication Date |
---|---|
CN207731293U true CN207731293U (en) | 2018-08-14 |
Family
ID=63098131
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201721708122.1U Active CN207731293U (en) | 2017-12-11 | 2017-12-11 | A kind of high robust image identification system based on SVG dynamic charts |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN207731293U (en) |
-
2017
- 2017-12-11 CN CN201721708122.1U patent/CN207731293U/en active Active
Similar Documents
Publication | Publication Date | Title |
---|---|---|
JP2019520615A (en) | Character recognition method, device, server and storage medium of claim document for damages | |
US10049096B2 (en) | System and method of template creation for a data extraction tool | |
CN104252410A (en) | Method and equipment for testing control in page | |
CN104391786A (en) | Webpage automatic test system and method thereof | |
CN103491116A (en) | Method and device for processing text-related structural data | |
CN104820634A (en) | Method and device for detecting page test track | |
CN107844468A (en) | The cross-page recognition methods of form data, electronic equipment and computer-readable recording medium | |
CN105354138A (en) | Method for determining specific control attribute of mobile phone page | |
CN104424318A (en) | Method and device for controlling page elements | |
CN105488232A (en) | Graphical display method for breakdown maintenance of electronic equipment | |
CN104750835A (en) | Text classification method and device | |
CN110990010A (en) | Software interface code generation method and device | |
CN107871128A (en) | A kind of high robust image-recognizing method based on SVG dynamic charts | |
CN111931771A (en) | Bill content identification method, device, medium and electronic equipment | |
CN103177096A (en) | Page element positioning method based on text attribute and page element positioning device based on text attribute | |
CN103365877A (en) | Method and server for making directory after webpage is transcoded | |
CN113283231B (en) | Method for acquiring signature bit, setting system, signature system and storage medium | |
EP3178009A1 (en) | Replicating a web technology | |
CN110782005A (en) | Image annotation method and system for tracking based on weak annotation data | |
CN207731293U (en) | A kind of high robust image identification system based on SVG dynamic charts | |
CN102722621A (en) | Method for visualizing computed result of finite element method | |
CN105653669A (en) | HTML (Hypertext Marked Language) generating method and device | |
CN110688842B (en) | Analysis method, device and server for document title level | |
US20140196000A1 (en) | System and method for checking signal transmission line | |
CN105183730A (en) | Method and device for processing webpage information |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
GR01 | Patent grant | ||
GR01 | Patent grant |