CN102708167B - Web-based semantic annotation system and Web-based semantic annotation method for high resolution SAR (synthetic aperture radar) image interpretation - Google Patents
Web-based semantic annotation system and Web-based semantic annotation method for high resolution SAR (synthetic aperture radar) image interpretation Download PDFInfo
- Publication number
- CN102708167B CN102708167B CN201210127743.6A CN201210127743A CN102708167B CN 102708167 B CN102708167 B CN 102708167B CN 201210127743 A CN201210127743 A CN 201210127743A CN 102708167 B CN102708167 B CN 102708167B
- Authority
- CN
- China
- Prior art keywords
- mark
- image
- web
- sar
- user
- 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
- Information Transfer Between Computers (AREA)
Abstract
The invention provides a Web-based semantic annotation system and a Web-based semantic annotation method for high resolution SAR (synthetic aperture radar) image interpretation, and aims to provide an effective measure for building a dataset of complex SAR images. Different labeling manners are designed according to different application requirements of the SAR image interpretation so as to lower labeling difficulty. A grid labeling method based on regions and segmentations is provided. By using Web techniques, a cross-system labeling system on webpages is realized, labeling results can be shared effectively, utilization rate is maximized, and manpower cost is lowered. In addition, specialized research and analysis channels for verifications and experiments are provided, and a simple SAR image interpretation research platform is provided.
Description
Technical field
The present invention relates to image interpretation field, relate to a kind of labeling system for the decipher of High-resolution SAR Images scene and target detection and method of sing on web web technologies more specifically.
Background technology
Synthetic-aperture radar (Synthetic Aperture Radar, SAR) be a kind of earth observation means of round-the-clock, round-the-clock, earth's surface high-resolution, the image of large scene and detailed mapping information can be obtained, be all widely used in many occasions such as military and civilian.The satellite being loaded with high resolution SAR in recent years progressively increases, and is more the increase in the demand of the scene decomposition method development of SAR image.The decipher of research High-resolution SAR Images scene, is the basic work detecting target and identification, utilizes the related information between target and scene, promote the effect of target identification; Build intelligent database, the basis of management mass data SAR image; Simultaneously also for solving practical problems provides data processing model and scheme, such as urban development change detects, disaster is emergent, anti-terrorism stability maintenance etc.Information outside the less consideration image pixel of existing many SAR video scenery decomposition methods, the more information process brought for image high resolving power is not very effective, in addition manually the positive negative sample of mark is operated in image interpretation field and is difficult to efficient solution, is therefore difficult to expand in the application of mass data complex class.
Mark refers to that third party is to former document creation and increase data.These data can be notes, symbol, picture, sound bite etc.Existing mask method, main to pay close attention to intuitively, naked eyes is easy to the natural image that judges, use polygon notation methods to mark the things in little scene, build the natural image labeled data collection of magnanimity.And the mark mentioned in the present invention referring on SAR image by adding pictorial symbolization and semantic description as annotation to description object, making intricate SAR image be easy to the personnel that are read and understanding, provide sample for computing machine carries out SAR image interpretation simultaneously.
Summary of the invention
But the object of the invention is to build the active demand of data set mark difficulty and the huge present situation of human cost for SAR image, propose a kind of labeling system for the decipher of High-resolution SAR Images scene and target detection and method of sing on web web technologies, provide effective means for intricate SAR image builds data set.
The present invention is directed to the feature of High-resolution SAR Images decipher, require the ability having process large scene image, for the demand of SAR image interpretation different application, design realizes different notation methods, to reduce mark difficulty, here improve mask method, propose a kind of grid notation methods based on region and piecemeal; Meanwhile, utilize Web technology, webpage realizes the labeling system of cross-system, annotation results is shared effectively, utilization factor is maximized, reach the effect alleviating human cost.In addition, the invention provides and specialize in analysis channel, for checking, experiment, is an easy SAR image interpretation research platform.
The present invention proposes a kind of semantic tagger system of the sing on web for High-resolution SAR Images decipher, comprise following four unit:
(1) browsing image unit, the piecemeal association for large scene SAR image is browsed and is switched with random scene, simultaneously synchronized update image markup information;
(2) unit is marked, for carrying out labeling operation to the SAR image shown in this browsing image unit;
(3) storage unit, the information that image and mark generate is sent into this storage unit and is stored, and provides image upload and download interface and markup information download interface simultaneously;
(4) communication unit, for realizing telecommunication, described this labeling system erection server, this communication unit of server by utilizing accesses this memory module, SAR image and relevant markup information are provided, and image and mark document pass through this communication unit default storage in the storage unit of server, the resource such as image and markup information is maximized by network and shares by server simultaneously;
And mark unit completes the XML format write of markup information, and browsing image unit completes the parsing of markup information XML document, and graphic software platform, also tracking display is on raw video in location, and the semantic information according to mark provides painted lines, helps interpretation directly perceived.
Wherein, browsing image unit comprises web interface design subelement, webpage functional realiey subelement and image capturing subelement further.
Described web interface design subelement comprises image module, functional module and markup information module further.Described image module is for showing raw video and mark, and mark color selects contrast comparatively large and chromatic colour according to mark semanteme, thus mark can highlight on the SAR image of black and white, and classification is distinct; Described functional module, for placing types of functionality button, comprises " amplification ", " reducing ", " recovery ", " dimension model switches (polygon/grid) ", " image switching ", " user's login ", " image is uploaded " etc.; Described markup information module is used for pressing the tree-shaped display markup information of semantic tagger classification, scalable, provides " XML download " and " image is downloaded " interface.
Described webpage functional realiey subelement comprises further browses module and processing module.Described module of browsing is mainly used in convergent-divergent, the recovery of image and switches, user can carry out convergent-divergent and recovery to image by the button or rolling slide-mouse clicking webpage functional area, user clicks " image switching ", server is according to user model and current image information, select association image or random image, and markup information is selected in association.Described processing module is mainly used in notation methods switching and selects with user model, and image module and markup information module upgrade simultaneously.Above webpage functional realiey subelement is relevant to mark unit, storage unit and communication unit.
It is carry out notation methods selection according to the demand of High-resolution SAR Images decipher different background that described notation methods switches, described notation methods comprises polygon notation methods and grid notation methods, by clicking " dimension model switching ", selection is provided with " the polygon mark " for " target detection " or " the grid mark " for " scene decipher ", imagery zone makes response according to selection, changes.
Described polygon mark is the profile using polygon to describe target, uses semantic key words to carry out the mark of Pixel-level.This notation methods can select target subtly, for follow-up study provides convenient.
Described grid mark, for the scene decipher research and design in SAR image interpretation, according to user's request, raw video is partitioned into the rectangular area of fixed size, utilize the semantic description of mark localized region, one or more semantic key words is used to mark, by association semantic net realization scene decipher from top to down.Its sizing grid can according to user's request, under different resolution, different functional region description changes.As, under three meters of resolution, urban function district is described, select the grid of 32*32 size to mark, the urban function district being labeled as 1 hectare of size of generation describes, can from macroscopically dividing scene, then according to different scene feature, more careful atural object is divided.Therefore, it is mainly suitable for crowd is researchist.This notation methods serves primarily in terrain classification research in the decipher of SAR image large scene, and relative to polygon mark, it greatly reduces the human cost of mark.
Described user model comprises interpretation dimension model and mark generate pattern.Described interpretation dimension model utilizes abundant Internet resources, marks arbitrary object, and mainly for the checking marking correctness and reliability, under this pattern, all marks are visible; Described mark generate pattern is for follow-up study personnel set up interface, by the key entry of user name, conceals other people mark, generates special XML document after mark, be convenient to import, carry out confirmatory study.
Wherein, mark unit comprises mark reading subelement further, mark mouse gestures controls subelement and markup information is synchronously located and shows subelement.
Described mark reads subelement and comprises markup information acquisition and markup information parsing further.It is utilize communication module that markup information obtains, access memory module, extracts relevant markup information XML document; It is according to predefined keyword that markup information is resolved, as " Object ", " pt " etc., and analyzing XML file.
Described mark mouse gestures controls subelement and comprises click further, slips over and cooperate three parts with key mouse.According to mouse gestures, judge mouse position and current mark state, as: whether whether " being positioned on mark ", " being polygon mode ", " whether polygon closes " etc., complete mark work, and interrelated logic function---play the operations such as window, filling, write, renewal.In order to complete complicated labeling operation and display, the present invention uses five layers of design to imagery zone on webpage, is from bottom to top respectively " image bearing layer ", " clathrum ", " mark presentation layer ", " mark layer ", " wait layer "." image bearing layer " deposits SAR image; " clathrum " switches for notation methods; " mark presentation layer " shows existing mark; " mark layer " is for labeling operation; " wait layer ", for giving tacit consent to display state, is hyaline layer, protection mark layer.Above five layers, move up and down according to mouse gestures.
Described markup information is synchronously located and is shown subelement, by the parsing of markup information XML document and the judgement to current web page and web browser state, mark is positioned on image, completes corresponding convergent-divergent simultaneously, make mark object and mark remain one_to_one corresponding.
All web interfaces and functional realiey all use html language and JavaScript language to realize above.Wherein, html language is used for page layout, and JavaScript language is used for realizing logic function and the design of part web page graphics.
Wherein, storage unit comprises mark write subelement further, mark descends subelements and image is uploaded and lower subelements.
Described mark write subelement comprises mark parsing further and writes with markup information.Described mark resolves to, and for according to webpage and browser status, mouse gestures is resolved to markup information; Described markup information is written as, and delivers to server for the markup information of resolving acquisition, with predefined form with in the relevant XML document of keyword write, is placed in suitable position.During amendment mark, complete relevant information amendment.
Subelements under described mark, for when user clicks " XML download " button, server, according to user right and webpage current state, provides the download function of markup information XML document.
Image uploads and downloads subelement, and for selecting " image is uploaded " as user, server audits this image according to access rights and presentation content, and performs the operations such as piecemeal, cutting, classification; " image download " button being positioned at markup information region supports that user downloads the current image browsed.
Wherein, communication unit comprises server erection subelement, access rights setting subelement and image and markup information addressing subelement further.
Described server erection subelement, native system realizes under a linux operating system, Apache instrument and Perl language is used to realize remote communicating function, the compatibility of various browser is considered during webpage design, being applicable to the web browser of various core, is a cross-platform SAR image semantic tagger system.
Described access rights setting subelement, be mainly used in protecting carrying out the individual research i.e. user profile of " mark generate pattern ", other information are all transparent.
Described image and markup information addressing subelement, use the pattern of text to classify to addressable image/markup information, user profile, associate, set up access rule simultaneously, while making the present invention become open cooperation SAR decipher labeling system, for the data encasement of individual research provides service.
In the present invention, except raw video data, other information comprise markup information, image information, user profile etc., all use XML document to carry out integrated storage administration, and associate with raw video with file name.
Present invention also offers a kind of semanteme marking method of the sing on web for High-resolution SAR Images decipher, comprise the following steps:
(1) user browses webpage, selecting user model, inputting user name and logging in, entering " mark generate pattern ", only show the mark of active user by whether clicking " login ", and SAR browsing image switch mode is " association switches " simultaneously; " publishing " and then enter " interpretation dimension model ", is all user annotations of display current SAR image, and SAR browsing image switch mode is " switching at random " simultaneously;
(2) carry out mark to former image and add operation, user by selecting notation methods, can select polygon notation methods or grid notation methods, carries out mark add SAR image;
(3), after mark generates, this markup information is shown in the image top layer of web interface with polygonal form, and when streaking this polygon frame, polygon will be filled, and shows this mark semantic description; Click this polygon frame by the amendment of ejection mark and deletion frame, mark amendment and mark deletion action can be carried out to former image;
(4) after labeling operation completes, in xml format markup information is write server, user downloads SAR image and markup information XML document by interface from server.
Notation methods wherein described in step (2), if select " polygon mark ", user selects polygon angle point successively by clicking the mouse, after forming closed polygon, eject semantic description frame, after input describes keyword, complete the interpolation to mark.If select " grid mark ", system carries out gridding piecemeal to image voluntarily, clicks the mouse and namely form the closed quadrilateral in this region in certain region, ejects semantic description frame, after input describes keyword, completes the interpolation to mark.Described grid mark, for the scene decipher research and design in SAR image interpretation, according to user's request, raw video is partitioned into the rectangular area of fixed size, utilize the semantic description of mark localized region, one or more semantic key words is used to mark, by association semantic net realization scene decipher from top to down.
The semantic description of the mark wherein described in step (3) is according to semantic classification, and tree-shaped being shown in marks list area.When streaking semantic description tree, relevant labelling polygons will be filled; Click the leaf (single mark) of semantic description tree by the amendment of ejection mark and deletion frame.
Described mark retouching operation is, under specific operation authority, modifies to the region marked or semantic description.After mark is modified, its " Yes/No was revised " attribute modification is 1.
Described mark deletion action is, under specific operation authority, delete mark.Mark deleted after, XML Central Plains mark be retained, its " Yes/No deletion " attribute modification is 1.
The present invention be directed to the semantic tagger system and method for the sing on web of High-resolution SAR Images decipher.Different notation methods is set, reduces the human cost of SAR image mark; Make full use of internet resource, improves mark utilization factor, reduce repeated labor; Set up researchist's passage, information integral (XML), for subsequent analysis research provides convenient.The present invention, while setting up SAR image interpretation data set, provides the SAR image interpretation research platform of an open cooperation.
Accompanying drawing explanation
Fig. 1 application mode schematic diagram of the present invention;
Fig. 2 web interface schematic diagram of the present invention;
Fig. 3 A-3B XML document structural representation of the present invention;
Fig. 4 access connection diagram of the present invention;
Fig. 5 use procedure schematic diagram of the present invention.
For foregoing of the present invention can be become apparent, shown in hereafter coordinating, accompanying drawing elaborates.
Embodiment
As shown in Figure 1, user browses webpage to the application mode that the present invention will realize, and server provides SAR image and relevant markup information (XML), user is by the specific notation methods of model selection, by labeling system, produce markup information (XML), be synchronously showed on webpage.Webpage provides raw video and markup information to download, and is provided with image simultaneously and uploads interface.
The semantic tagger system of a kind of sing on web for High-resolution SAR Images decipher that the present invention proposes, comprises following four unit:
(1) browsing image unit, the piecemeal association for large scene SAR image is browsed and is switched with random scene, simultaneously synchronized update image markup information;
(2) unit is marked, for carrying out labeling operation to the SAR image shown in this browsing image unit;
(3) storage unit, the information that image and mark generate is sent into this storage unit and is stored, and provides image upload and download interface and markup information download interface simultaneously;
(4) communication unit, for realizing telecommunication, described this labeling system erection server, this communication unit of server by utilizing accesses this memory module, SAR image and relevant markup information are provided, and image and mark document pass through this communication unit default storage in the storage unit of server, the resource such as image and markup information is maximized by Web and shares by server simultaneously;
And mark unit completes the XML format write of markup information, and browsing image unit completes the parsing of markup information XML document, and graphic software platform, also tracking display is on raw video in location, and the semantic information according to mark provides painted lines, helps interpretation directly perceived.
Wherein, browsing image unit comprises web interface design subelement, webpage functional realiey subelement and image capturing subelement further.
Described web interface design subelement as shown in Figure 2, comprises image module, functional module and markup information module further.Described image module is for showing raw video and mark, and mark color selects contrast comparatively large and chromatic colour according to mark semanteme, thus mark can highlight on the SAR image of black and white, and classification is distinct; Described functional module, for placing types of functionality button, comprises " amplification ", " reducing ", " recovery ", " dimension model switches (polygon/grid) ", " image switching ", " user's login ", " image is uploaded " etc.; Described markup information module is used for pressing the tree-shaped display markup information of semantic tagger classification, scalable, provides " XML download " and " image is downloaded " interface.
Described webpage functional realiey subelement comprises further browses module and processing module.Described module of browsing is mainly used in convergent-divergent, the recovery of image and switches, user can carry out convergent-divergent and recovery to image by the button or rolling slide-mouse clicking webpage functional area, user clicks " image switching ", server is according to user model and current image information, select association image or random image, and markup information is selected in association.Described processing module is mainly used in notation methods switching and selects with user model, and image module and markup information module upgrade simultaneously.Above webpage functional realiey subelement is relevant to mark unit, storage unit and communication unit.
Described notation methods switching is the demand according to High-resolution SAR Images decipher different background, described notation methods comprises polygon notation methods and grid notation methods, by clicking " dimension model switching ", selection is provided with " the polygon mark " for " target detection " or " the grid mark " for " scene decipher ", imagery zone makes response according to selection, changes.
Described " polygon mark " is, uses polygon to describe the profile of target, uses semantic key words to carry out the mark of Pixel-level.This notation methods can select target subtly, for follow-up study provides convenient.
Described " grid mark " is, according to user's request, raw video is partitioned into the rectangular area of fixed size, uses one or more semantic key words to mark.This notation methods serves primarily in terrain classification research in the decipher of SAR image large scene, and relative to " polygon mark ", it greatly reduces the human cost of mark.
Described user model comprises " interpretation dimension model " and " mark generate pattern ".Described " interpretation dimension model " utilizes abundant Internet resources, marks arbitrary object, and mainly for the checking marking correctness and reliability, under this pattern, all marks are visible; Described " mark generate pattern " is for follow-up study personnel set up interface, by the key entry of user name, conceals other people mark, generates special XML document after mark, be convenient to import, carry out confirmatory study.
Wherein, mark unit comprises mark reading subelement further, mark mouse gestures controls subelement and markup information is synchronously located and shows subelement.
Described mark reads subelement and comprises markup information acquisition and markup information parsing further.It is utilize communication module that markup information obtains, access memory module, extracts relevant markup information XML document; It is according to predefined keyword that markup information is resolved, as " Object ", " pt " etc., and analyzing XML file.Wherein XML document structural representation as shown in Figure 3.
Described mark mouse gestures controls subelement and comprises click further, slips over and cooperate three parts with key mouse.According to mouse gestures, judge mouse position and current mark state, as: whether whether " being positioned on mark ", " being polygon mode ", " whether polygon closes " etc., complete mark work, and interrelated logic function---play the operations such as window, filling, write, renewal.In order to complete complicated labeling operation and display, the present invention uses five layers of design to imagery zone on webpage, is from bottom to top respectively " image bearing layer ", " clathrum ", " mark presentation layer ", " mark layer ", " wait layer "." image bearing layer " deposits SAR image; " clathrum " switches for notation methods; " mark presentation layer " shows existing mark; " mark layer " is for labeling operation; " wait layer ", for giving tacit consent to display state, is hyaline layer, protection mark layer.Above five layers, move up and down according to mouse gestures.
Described markup information is synchronously located and is shown subelement, by the parsing of markup information XML document and the judgement to current web page and web browser state, mark is positioned on image, completes corresponding convergent-divergent simultaneously, make mark object and mark remain one_to_one corresponding.
All web interfaces and functional realiey all use html language and JavaScript language to realize above.Wherein, html language is used for page layout, and JavaScript language is used for realizing logic function and the design of part web page graphics.
Wherein, storage unit comprises mark write subelement further, subelements and image upload and download subelement to mark down.
Described mark write subelement comprises mark parsing further and writes with markup information.Described mark resolves to, and for according to webpage and browser status, mouse gestures is resolved to markup information; Described markup information is written as, and delivers to server for the markup information of resolving acquisition, with predefined form with in the relevant XML document of keyword write, is placed in suitable position.During amendment mark, complete relevant information amendment.
Subelements under described mark, for when user clicks " XML download " button, server, according to user right and webpage current state, provides the download function of markup information XML document.
Image uploads and downloads subelement, and for selecting " image is uploaded " as user, server audits this image according to access rights and presentation content, and performs the operations such as piecemeal, cutting, classification; " image download " button being positioned at markup information region supports that user downloads the current image browsed.
Wherein, communication unit comprises server erection subelement, access rights setting subelement and image and markup information addressing subelement further.Concrete communications access mode as shown in Figure 4, described this labeling system erection server, this communication unit of server by utilizing accesses this memory module, SAR image and relevant markup information are provided, and image and mark document default storage are in server, the resource such as image and markup information are maximized by Web simultaneously and share.
Described server erection subelement, native system realizes under a linux operating system, Apache instrument and Perl language is used to realize remote communicating function, the compatibility of various browser is considered during webpage design, being applicable to the web browser of various core, is a cross-platform SAR image semantic tagger system.
Described access rights setting subelement, be mainly used in protecting carrying out the individual research i.e. user profile of " mark generate pattern ", other information are all transparent.
Described image and markup information addressing subelement, use the pattern of text to classify to addressable image/markup information, user profile, associate, set up access rule simultaneously, while making the present invention become open cooperation SAR decipher labeling system, for the data encasement of individual research provides service.
In the present invention, except raw video data, other information comprise markup information, image information, user profile etc., all use XML document to carry out integrated storage administration, and associate with raw video with file name.
Present invention also offers a kind of semanteme marking method of the sing on web for High-resolution SAR Images decipher, as shown in Figure 5, the method comprises the following steps:
(1) user browses webpage, selecting user model, inputting user name and logging in, entering " mark generate pattern ", only show the mark of active user by whether clicking " login ", and SAR browsing image switch mode is " association switches " simultaneously; " publishing " and then enter " interpretation dimension model ", is all user annotations of display current SAR image, and SAR browsing image switch mode is " switching at random " simultaneously;
(2) carry out mark to former image and add operation, user by selecting notation methods, can select polygon notation methods or grid notation methods, carries out mark add SAR image;
(3), after mark generates, this markup information is shown in the image top layer of web interface with polygonal form, and when streaking this polygon frame, polygon will be filled, and shows this mark semantic description; Click this polygon frame by the amendment of ejection mark and deletion frame, mark amendment and mark deletion action can be carried out to former image;
(4) after labeling operation completes, in xml format markup information is write server, user downloads SAR image and markup information XML document by interface from server.
Notation methods wherein described in step (2), if select " polygon mark ", user selects polygon angle point successively by clicking the mouse, after forming closed polygon, eject semantic description frame, after input describes keyword, complete the interpolation to mark.If select " grid mark ", system carries out gridding piecemeal to image voluntarily, clicks the mouse and namely form the closed quadrilateral in this region in certain region, ejects semantic description frame, after input describes keyword, completes the interpolation to mark.
The semantic description of the mark wherein described in step (3) is according to semantic classification, and tree-shaped being shown in marks list area.When streaking semantic description tree, relevant labelling polygons will be filled; Click the leaf (single mark) of semantic description tree by the amendment of ejection mark and deletion frame.
Described mark retouching operation is, under specific operation authority, modifies to the region marked or semantic description.After mark is modified, its " Yes/No was revised " attribute modification is 1.
Described mark deletion action is, under specific operation authority, delete mark.Mark deleted after, XML Central Plains mark be retained, its " Yes/No deletion " attribute modification is 1.
The present invention be directed to the semantic tagger system and method for the sing on web of High-resolution SAR Images decipher.Different notation methods is set, reduces the human cost of SAR image mark; Make full use of internet resource, improves mark utilization factor, reduce repeated labor; Set up researchist's passage, information integral (XML), for subsequent analysis research provides convenient.The present invention, while setting up SAR image interpretation data set, provides the SAR image interpretation research platform of an open cooperation.
Although the present invention discloses as above with preferred embodiment; so itself and be not used to limit the present invention, anyly have the knack of this those skilled in the art, without departing from the spirit and scope of the present invention; when doing a little change and retouching, therefore protection scope of the present invention is when being as the criterion depending on the claim person of defining.
Claims (8)
1. for a semanteme marking method for the sing on web of High-resolution SAR Images decipher, it is characterized in that, the method comprises the following steps:
(1) mark is carried out to SAR image and add operation, adopt grid notation methods to carry out mark to add, described grid notation methods is the rectangular area original SAR Image Segmentation being become some fixed sizes, utilizing mark to carry out semantic description to needing this block rectangular area marked, using predefined one or more semantic key words to carry out mark and adding;
(2) be that the predefined form of unit and keyword write in XML document with rectangular area, be placed in the server set up in advance; Also provide interface from server, download SAR image and markup information XML document, realize server and the resource comprising image and markup information is maximized by network shared;
Wherein, before step (1), according to user, whether " login " selects user model to semantic tagger system, thus selects different markup informations and image information display mode;
Described user model comprises mark generate pattern and interpretation dimension model,
When receiving user's " login " information, system enters mark generate pattern, only shows the mark of active user, and SAR browsing image switch mode is " association switches " simultaneously;
" publish " information when receiving user, system enters interpretation dimension model, all user annotations of display current SAR image, and SAR browsing image switch mode is " switching at random " simultaneously;
After step (1), after mark generates, this markup information is shown in the image top layer of web interface with polygonal form, and when streaking this polygonal frame, polygon will be filled, and show the semantic description of this markup information; Click this polygonal frame by the amendment of ejection mark and deletion frame, mark amendment and mark deletion action are carried out to former image.
2. as claimed in claim 1 for the semanteme marking method of the sing on web of High-resolution SAR Images decipher, it is characterized in that, notation methods in step (1) also comprises polygon notation methods, described polygon notation methods, being the polygon angle point that system is selected by clicking the mouse successively according to user, after forming closed polygon, ejecting semantic description frame, after input describes keyword, complete the interpolation to mark.
3. for a semantic tagger system for the sing on web of High-resolution SAR Images decipher, it is characterized in that, comprise following four unit:
(1) browsing image unit, the piecemeal association for large scene SAR image is browsed and is switched with random scene, simultaneously synchronized update image markup information;
(2) unit is marked, for carrying out labeling operation to the SAR image shown in this browsing image unit;
(3) storage unit, the information that image and mark generate is sent into this storage unit and is stored, and provides image upload and download interface and markup information download interface simultaneously;
(4) communication unit, for realizing telecommunication, this labeling system erection server, this communication unit of server by utilizing accesses this storage unit, SAR image and relevant markup information are provided, and image and mark document pass through this communication unit default storage in the storage unit of server, image and markup information maximize shared by server by network simultaneously;
Wherein, according to user, whether " login " selects user model to semantic tagger system, thus selects different markup informations and image information display mode;
Described user model comprises mark generate pattern and interpretation dimension model,
When receiving user's " login " information, system enters mark generate pattern, only shows the mark of active user, and SAR browsing image switch mode is " association switches " simultaneously;
" publish " information when receiving user, system enters interpretation dimension model, all user annotations of display current SAR image, and SAR browsing image switch mode is " switching at random " simultaneously;
Complete the XML write of markup information at mark unit after, browsing image unit completes the parsing of markup information XML document, and graphic software platform, location and tracking display on raw video, semantic information according to mark provides the painted lines that can help interpretation directly perceived, and this markup information is shown in the image top layer of web interface, when streaking this polygonal frame with polygonal form, polygon will be filled, and show the semantic description of this markup information; Click this polygonal frame by the amendment of ejection mark and deletion frame, mark amendment and mark deletion action are carried out to former image.
4. as claimed in claim 3 for the semantic tagger system of the sing on web of High-resolution SAR Images decipher, it is characterized in that, described browsing image unit comprises web interface design subelement, webpage functional realiey subelement and image capturing subelement further;
Described web interface design subelement comprises image module, functional module and markup information module further, described image module is for showing raw video and mark, this image module uses five layers of design, from bottom to top be respectively " image bearing layer ", " clathrum ", " mark presentation layer ", " mark layer ", " wait layer ", wherein, " image bearing layer " deposits SAR image; " clathrum " switches for notation methods; " mark presentation layer " shows existing mark; " mark layer " is for labeling operation; " wait layer ", for giving tacit consent to display state, is hyaline layer, protection mark layer.
5. as claimed in claim 4 for the semantic tagger system of the sing on web of High-resolution SAR Images decipher, it is characterized in that, described webpage functional realiey subelement comprises further browses module and processing module, this webpage functional realiey subelement is relevant to mark unit, storage unit and communication unit
Described module of browsing is mainly used in convergent-divergent, the recovery of image and switches;
Described processing module is mainly used in notation methods switching and selects with user model, and image module and markup information module upgrade simultaneously.
6. as claimed in claim 5 for the semantic tagger system of the sing on web of High-resolution SAR Images decipher, it is characterized in that, notation methods comprises polygon mark and grid mark,
Described polygon mark is the profile using polygon to describe target, uses semantic key words to carry out the mark of Pixel-level;
Described grid mark is according to user's request, raw video is partitioned into the rectangular area of fixed size, utilize the semantic description of mark localized region, use one or more semantic key words to mark, by association semantic net realization scene decipher from top to down.
7. as claimed in claim 5 for the semantic tagger system of the sing on web of High-resolution SAR Images decipher, it is characterized in that, described user model comprises interpretation dimension model and mark generate pattern,
Described interpretation dimension model utilizes Internet resources, marks arbitrary object, and in this mode, all marks are visible;
Described mark generate pattern is for follow-up study personnel provide interface, conceals other people mark, generates special XML document after mark.
8. as claimed in claim 3 for the semantic tagger system of the sing on web of High-resolution SAR Images decipher, it is characterized in that, this system is except raw video data, markup information, image information, user profile all use XML document to carry out integrated storage administration, and associate with raw video with file name.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210127743.6A CN102708167B (en) | 2012-04-27 | 2012-04-27 | Web-based semantic annotation system and Web-based semantic annotation method for high resolution SAR (synthetic aperture radar) image interpretation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210127743.6A CN102708167B (en) | 2012-04-27 | 2012-04-27 | Web-based semantic annotation system and Web-based semantic annotation method for high resolution SAR (synthetic aperture radar) image interpretation |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102708167A CN102708167A (en) | 2012-10-03 |
| CN102708167B true CN102708167B (en) | 2015-06-24 |
Family
ID=46900933
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201210127743.6A Active CN102708167B (en) | 2012-04-27 | 2012-04-27 | Web-based semantic annotation system and Web-based semantic annotation method for high resolution SAR (synthetic aperture radar) image interpretation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102708167B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103294792B (en) * | 2013-05-13 | 2016-02-10 | 西安电子科技大学 | Based on the polarization SAR terrain classification method of semantic information and polarization decomposing |
| CN108776772B (en) * | 2018-05-02 | 2022-02-08 | 北京佳格天地科技有限公司 | Cross-time building change detection modeling method, detection device, method and storage medium |
| CN110377338A (en) * | 2019-07-23 | 2019-10-25 | 网易无尾熊（杭州）科技有限公司 | Information labeling method, medium, device and calculating equipment |
| CN110704416A (en) * | 2019-10-10 | 2020-01-17 | 云南中烟工业有限责任公司 | Method for data management by using data center |
| CN111860370A (en) * | 2020-07-24 | 2020-10-30 | 浙江商汤科技开发有限公司 | Semantic annotation method, device and system |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101739415A (en) * | 2008-11-25 | 2010-06-16 | 华中师范大学 | Browser-oriented webpage labeling system |
| CN101894275A (en) * | 2010-06-29 | 2010-11-24 | 武汉大学 | Weakly supervised method for classifying SAR images |
| CN102135997A (en) * | 2011-03-23 | 2011-07-27 | 华中科技大学 | Method for managing digital learning resource based on body |
-
2012
- 2012-04-27 CN CN201210127743.6A patent/CN102708167B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101739415A (en) * | 2008-11-25 | 2010-06-16 | 华中师范大学 | Browser-oriented webpage labeling system |
| CN101894275A (en) * | 2010-06-29 | 2010-11-24 | 武汉大学 | Weakly supervised method for classifying SAR images |
| CN102135997A (en) * | 2011-03-23 | 2011-07-27 | 华中科技大学 | Method for managing digital learning resource based on body |
Also Published As
| Publication number | Publication date |
|---|---|
| CN102708167A (en) | 2012-10-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Boulos et al. | Web GIS in practice VIII: HTML5 and the canvas element for interactive online mapping | |
| Cecconi et al. | Adaptive zooming in web cartography | |
| CN105447101B (en) | Map engine implementation method and device | |
| US8196035B2 (en) | Adaptation of a website to mobile web browser | |
| Sun et al. | Development of a Web-based visualization platform for climate research using Google Earth | |
| CN102708167B (en) | Web-based semantic annotation system and Web-based semantic annotation method for high resolution SAR (synthetic aperture radar) image interpretation | |
| CN109614601A (en) | Form processing method, device, equipment and readable medium based on web | |
| Beaudette et al. | Soil-Web: an online soil survey for California, Arizona, and Nevada | |
| CA2817554A1 (en) | Mobile content management system | |
| CN109816750B (en) | Image processing method and device based on meteorological data and computer equipment | |
| CN106886586A (en) | A kind of Internet map application development platform based on open source projects | |
| Maravelakis et al. | 3DSYSTEK web-based point cloud viewer | |
| Jin et al. | [Retracted] Application of VR Technology in Jewelry Display | |
| Chen et al. | Data visualization in smart grid and low‐carbon energy systems: A review | |
| Cao | Development and design case function comparison of panoramic roaming system of virtual museum based on Pano2VR | |
| WO2024164589A1 (en) | Information display method and apparatus, electronic device, computer readable storage medium, and computer program product | |
| Sha et al. | Design of service-oriented architecture for spatial data integration and its application in building web-based GIS systems | |
| CN206117732U (en) | Cartographic data server system | |
| CN104036028A (en) | Processing system, generation method, processing method and differential displaying method for electronic document information fragments | |
| CN102110166B (en) | Browser-based body 3D (3-demensional) visualizing and editing system and method | |
| Yao et al. | Interoperable internet mapping—an open source approach | |
| Wu et al. | Digital protection and inheritance of culture under the background of “The Belt and Road Initiative” | |
| Brovelli et al. | Valorisation of Como historical cadastral maps through modern web geoservices | |
| CN111815495B (en) | CIM platform decision method, system, equipment and storage medium based on mixed reality | |
| Kong et al. | A web-based geographic hypermedia system: Data model, system design and prototype applications |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |