Summary of the invention
The object of the invention is to be achieved through the following technical solutions.
According to the embodiment of the present invention, a kind of Distributed Virtual Reality System is proposed, described system comprises multiple distributed mobile terminal and cloud server, and described mobile terminal is provided with virtual machine, and described mobile terminal is connected by LTE-A wireless network with described cloud server; Wherein,
Described virtual machine comprises a LTE-A wireless transmit/receive units, response interface, instruction demodulating unit, outdoor scene acquisition unit, standard point determining unit, information package transmission unit, realtime graphic frame decoding demodulating unit, real-time virtual image image-drawing unit;
Described cloud server comprises the 2nd LTE-A wireless transmit/receive units, unit is set up in packet demodulating unit, actual situation conversion, assembled unit, virtual reality construction unit, virtual reality collision detection unit, virtual reality reliability judging unit, actual situation upgrade regulon and realtime graphic frame coded modulation unit;
A described LTE-A wireless transmit/receive units input end connects the 2nd LTE-A wireless transmit/receive units, output terminal connects the 2nd LTE-A wireless transmit/receive units, response interface and realtime graphic frame decoding demodulating unit respectively, link order demodulating unit, outdoor scene obtain unit, standard point determining unit, information package transmission unit to the output terminal of described response interface successively, the output terminal of described information package transmission unit connects a LTE-A wireless transmit/receive units, and the output terminal of described realtime graphic frame decoding demodulating unit connects real-time virtual image image-drawing unit;
The output terminal connection packet demodulating unit successively of described 2nd LTE-A wireless transmit/receive units, unit is set up in actual situation conversion, assembled unit, virtual reality construction unit, the output terminal of virtual reality construction unit connects virtual reality collision detection unit respectively, virtual reality reliability judging unit and actual situation upgrade regulon, described virtual reality collision detection unit and virtual reality reliability judging unit obtain output terminal and are connected actual situation renewal regulon respectively, the output terminal that described actual situation upgrades regulon connects realtime graphic frame coded modulation unit, the output terminal of described realtime graphic frame coded modulation unit connects the 2nd LTE-A wireless transmit/receive units.
According to the embodiment of the present invention, a described LTE-A wireless transmit/receive units is used for realizing receiving and dispatching with cloud server and other mobile terminals operating;
Described response interface, for the cooperation request message of the operation requests instruction or other mobile terminals that receive cloud server;
Described instruction demodulating unit, for being demodulated into the execution instruction of virtual machine by the operation requests instruction of cloud server or the cooperation request message of other mobile terminals;
Described outdoor scene obtains the Image Acquisition execution instruction that unit is used for demodulating according to instruction demodulating unit, obtains the on-the-spot true picture that will carry out actual situation conversion;
Described standard point determining unit is used for according to the images acquisition locations of described outdoor scene and view direction, and obtain the standard point information of standard point in preset range and correspondence, described standard point information comprises the positional information of corresponding standard point;
Described information package transmission unit, the standard point determined for the on-the-spot true picture that obtained by outdoor scene unit and standard point determining unit and corresponding standard point information are packed, and are sent to cloud server by a LTE-A wireless transmit/receive units.
According to the embodiment of the present invention, described realtime graphic frame decoding demodulating unit, decodes and demodulation for the accurate virtual view data sent the cloud server received;
Described real-time virtual image image-drawing unit, for decoding demodulation after map data and show interactive virtual scene.
According to the embodiment of the present invention, described 2nd LTE-A wireless transmit/receive units is for realizing the radio communication with distributed mobile terminal;
Described packet demodulating unit, for storing and demodulation the packing data bag received, and sends to the conversion of described actual situation to set up unit by the true picture information demodulated and standard point information;
The conversion of described actual situation sets up unit for setting up actual situation mapping graph, by obtained standard point according to the mutual alignment relationship map between standard point on described actual situation mapping graph, and add corresponding standard point information identification marking in the position being mapped with standard point;
Described assembled unit is used for the actual situation mapping graph and described real scene image that are added with standard point information identification marking to combine, to form gain combination image.
According to the embodiment of the present invention, described virtual reality construction unit, the constructing plan preset for basis and described gain combination image, perform corresponding computing and virtual drawing, and virtual drawing sent to actual situation to upgrade regulon, virtual reality collision detection unit and virtual reality reliability judging unit; Described virtual reality construction unit comprises the virtual image-drawing unit of an associating be made up of graphics processing unit group; After all virtual realities of task convert, the virtual image-drawing unit of described associating carries out the virtual drawing of 3D, and virtual for 3D drawing is sent to actual situation renewal regulon, virtual reality collision detection unit and virtual reality reliability judging unit.
According to the embodiment of the present invention, described virtual reality collision detection unit is used for performing virtual reality collision detection according to the virtual drawing of virtual reality construction unit, and testing result is exported to actual situation renewal regulon;
Described virtual reality reliability judging unit is used for the virtual detection performed the virtual image reliability obtained of drawing according to virtual reality construction unit, and testing result is exported to actual situation renewal regulon.
According to the embodiment of the present invention, described actual situation upgrades regulon and is used for carrying out renewal adjustment, to generate accurate virtual image according to the described virtual drawing of virtual reality construction unit generation and the testing result of virtual reality collision detection unit and virtual reality reliability judging unit to virtual drawing;
Described realtime graphic frame coded modulation unit, for carrying out coded modulation to the accurate virtual image generated, and sends it to the realtime graphic frame decoding demodulating unit of described mobile terminal.
Distributed Virtual Reality System of the present invention comprises multiple distributed mobile terminal and cloud server, described mobile terminal is provided with virtual machine, and described virtual machine comprises outdoor scene and obtains unit, standard point determining unit, information package transmission unit, realtime graphic frame decoding demodulating unit, real-time virtual image image-drawing unit; Described cloud server comprises actual situation conversion and sets up unit, assembled unit, virtual reality construction unit, virtual reality collision detection unit, virtual reality reliability judging unit, actual situation renewal regulon and realtime graphic frame coded modulation unit, by design of the present invention, solve the conflict in actual situation conversion and unreliability problem, improve Consumer's Experience.
Embodiment
Below with reference to accompanying drawings illustrative embodiments of the present disclosure is described in more detail.Although show illustrative embodiments of the present disclosure in accompanying drawing, however should be appreciated that can realize the disclosure in a variety of manners and not should limit by the embodiment of setting forth here.On the contrary, provide these embodiments to be in order to more thoroughly the disclosure can be understood, and complete for the scope of the present disclosure can be conveyed to those skilled in the art.
According to the embodiment of the present invention, a kind of Distributed Virtual Reality System is proposed, as shown in Figure 1, described system comprises multiple distributed mobile terminal and cloud server, described mobile terminal is provided with virtual machine, and described mobile terminal is connected by LTE-A wireless network with described cloud server; Wherein,
Described virtual machine comprises a LTE-A wireless transmit/receive units, response interface, instruction demodulating unit, outdoor scene acquisition unit, standard point determining unit, information package transmission unit, realtime graphic frame decoding demodulating unit, real-time virtual image image-drawing unit;
Described cloud server comprises the 2nd LTE-A wireless transmit/receive units, unit is set up in packet demodulating unit, actual situation conversion, assembled unit, virtual reality construction unit, virtual reality collision detection unit, virtual reality reliability judging unit, actual situation upgrade regulon and realtime graphic frame coded modulation unit;
A described LTE-A wireless transmit/receive units input end connects the 2nd LTE-A wireless transmit/receive units, output terminal connects the 2nd LTE-A wireless transmit/receive units, response interface and realtime graphic frame decoding demodulating unit respectively, link order demodulating unit, outdoor scene obtain unit, standard point determining unit, information package transmission unit to the output terminal of described response interface successively, the output terminal of described information package transmission unit connects a LTE-A wireless transmit/receive units, and the output terminal of described realtime graphic frame decoding demodulating unit connects real-time virtual image image-drawing unit;
The output terminal connection packet demodulating unit successively of described 2nd LTE-A wireless transmit/receive units, unit is set up in actual situation conversion, assembled unit, virtual reality construction unit, the output terminal of virtual reality construction unit connects virtual reality collision detection unit respectively, virtual reality reliability judging unit and actual situation upgrade regulon, described virtual reality collision detection unit and virtual reality reliability judging unit obtain output terminal and are connected actual situation renewal regulon respectively, the output terminal that described actual situation upgrades regulon connects realtime graphic frame coded modulation unit, the output terminal of described realtime graphic frame coded modulation unit connects the 2nd LTE-A wireless transmit/receive units.
According to the embodiment of the present invention, a described LTE-A wireless transmit/receive units is used for realizing receiving and dispatching with cloud server and other mobile terminals operating;
Described response interface, for the cooperation request message of the operation requests instruction or other mobile terminals that receive cloud server;
Described instruction demodulating unit, for being demodulated into the execution instruction of virtual machine by the operation requests instruction of cloud server or the cooperation request message of other mobile terminals;
Outdoor scene obtains the Image Acquisition execution instruction that unit is used for demodulating according to instruction demodulating unit, obtains the on-the-spot true picture that will carry out actual situation conversion;
Standard point determining unit is used for according to the images acquisition locations of described outdoor scene and view direction, and obtain the standard point information of standard point in preset range and correspondence, described standard point information comprises the positional information of corresponding standard point;
Described information package transmission unit, the standard point determined for the on-the-spot true picture that obtained by outdoor scene unit and standard point determining unit and corresponding standard point information are packed, and are sent to cloud server by a LTE-A wireless transmit/receive units;
Described realtime graphic frame decoding demodulating unit, decodes and demodulation for the accurate virtual view data sent the cloud server received;
Described real-time virtual image image-drawing unit, for decoding demodulation after map data and show interactive virtual scene.
According to the embodiment of the present invention, described 2nd LTE-A wireless transmit/receive units is for realizing the radio communication with distributed mobile terminal;
Described packet demodulating unit, for storing and demodulation the packing data bag received, and sends to the conversion of described actual situation to set up unit by the true picture information demodulated and standard point information;
The conversion of described actual situation sets up unit for setting up actual situation mapping graph, by obtained standard point according to the mutual alignment relationship map between standard point on described actual situation mapping graph, and add corresponding standard point information identification marking in the position being mapped with standard point;
Described assembled unit is used for the actual situation mapping graph and described real scene image that are added with standard point information identification marking to combine, to form gain combination image;
Described virtual reality construction unit, the constructing plan preset for basis and described gain combination image, perform corresponding computing and virtual drawing, and virtual drawing sent to actual situation to upgrade regulon, virtual reality collision detection unit and virtual reality reliability judging unit; Described virtual reality construction unit comprises the virtual image-drawing unit of an associating be made up of graphics processing unit group; After all virtual realities of task convert, the virtual image-drawing unit of described associating carries out the virtual drawing of 3D, and virtual for 3D drawing is sent to actual situation renewal regulon, virtual reality collision detection unit and virtual reality reliability judging unit;
Described virtual reality collision detection unit is used for performing virtual reality collision detection according to the virtual drawing of virtual reality construction unit, and testing result is exported to actual situation renewal regulon;
Described virtual reality reliability judging unit is used for the virtual detection performed the virtual image reliability obtained of drawing according to virtual reality construction unit, and testing result is exported to actual situation renewal regulon;
Described actual situation upgrades regulon and is used for carrying out renewal adjustment, to generate accurate virtual image according to the described virtual drawing of virtual reality construction unit generation and the testing result of virtual reality collision detection unit and virtual reality reliability judging unit to virtual drawing;
Described realtime graphic frame coded modulation unit, for carrying out coded modulation to the accurate virtual image generated, and sends it to the realtime graphic frame decoding demodulating unit of described mobile terminal.
According to the embodiment of the present invention, the conversion of described actual situation is set up unit and is specifically comprised:
Compartmentalization unit, for setting up described actual situation mapping graph, and is divided into multiple uniform region by described actual situation mapping graph;
Standard point map unit, for being respectively mapped in the corresponding region of described actual situation mapping graph according to the mutual alignment relation between standard point by sampling transformation by obtained standard point;
Identification marking adding device, for adding corresponding standard point information identification marking on the region being mapped with standard point.
According to the embodiment of the present invention, described actual situation conversion is set up unit and can also be comprised further: region parameter setting unit, and for arranging the parameter in described region, described parameter comprises when showing described virtual image and shows described region or hiding described region.
According to the embodiment of the present invention, described assembled unit specifically comprises:
Target image sets up unit, for choosing the image of a secondary outdoor scene as target image;
Basis angle set up unit, for calculate described actual situation mapping graph be visually parallel to the plane in described target image time, the angle between described actual situation mapping graph and physical plane, sets angle based on this angle;
Deflection matrix operation unit, for real scene image described in computing and the deflection matrix between described target image;
Recombiner unit, for according to described deflection matrix and described basic angle, is compound in the actual situation mapping graph being added with standard point information identification marking in described real scene image.
According to the embodiment of the present invention, described virtual reality collision detection unit specifically comprises according to the virtual drawing execution virtual reality collision detection of virtual reality construction unit:
The calibration point of A1, target frame extracts, and extracts target two field picture, carry out Iamge Segmentation in the picture from virtual Drawing image, obtains the coordinate (X of calibration point in displaing coordinate system
1, X
2, X
3, X
4);
Scenery movement estimation in A2, actual scene, is tied to the transition matrix of displaing coordinate system, obtains the coordinate (Y under corresponding real coordinate system by computing reality coordinate
1, Y
2, Y
3, Y
4); After acquisition calibration point, calculate the normal vector obtaining conflict plane; Calculate the quality point C of conflict plane subsequently; And the motion vector of the actual scenery of particular moment is tried to achieve by the motion of quality point;
A3, set up the space geometric from of virtual scene, virtual scene is considered as spheroid, uses ball as space geometric from, space geometric from tree is set up to virtual scene;
A4, based on cloud computing, actual situation conflict to be detected, specifically comprises:
A41, quantity according to actual situation collision detection node, divide into groups to virtual scene space geometric from tree; The identifier of each space geometric from is designated as ID, its quality point G and radius r is designated as Gr, obtains raw data <ID, Gr>;
A42, each group of raw data <ID, Gr> are sent into actual situation collision detection node carry out actual situation collision detection;
Virtual scene is approximately a spheroid, virtual scene conflicts with actual scenery and regards conflicting of ball and plane as;
Namely the quality point G of virtual scene ball is vectorial Y to the distance of conflict plane
1the projection d of G on normal vector; The precondition of now collision detection is: d≤τ r, and wherein r is the radius of ball, the ratio of τ to be G to the distance of conflict plane account for r;
Namely actual situation conflict is not produced if do not satisfied condition.When meeting this condition, then calculate the projection G ' of quality point G in conflict plane; And judge whether that G ' is inner in the conflict area having calibration point to surround, if meet this condition, there occurs conflict; Quality point G is charged in Gr at the projection G ' of conflict plane, obtains new data <ID, Gr ' >;
A43, by <ID, Gr ' > sends into actual situation conflict response node, judge the information whether comprising conflict point in Gr ', namely whether there occurs actual situation conflict, if there is not actual situation conflict, then the motion vector of the motion vector of next frame virtual scene frame virtual scene is therewith identical; As there occurs actual situation conflict, then send collision detection result and upgrade regulon to actual situation.
According to the embodiment of the present invention, described virtual reality reliability judging unit specifically comprises according to the virtual detection performed the virtual image reliability obtained of drawing of virtual reality construction unit:
C1, using the virtual drawing of virtual reality construction unit as main with reference to collection, real scene image with reference to collection, builds the grouping of virtual reality scenario iterative data as auxiliary; And utilize the optical imagery difference of virtual objects and true picture, choose actual situation grouping feature; The described actual situation grouping feature chosen comprises partial statistics, plane of delineation gradient or second fundamental form etc., can extract the above-mentioned actual situation grouping feature obtaining this some correspondence at each pixel place of image;
C2, in iterative data grouping, utilize actual situation grouping feature, extract the region main body characteristic of virtual reality scenario and outdoor scene respectively, structure stage resolution ratio actual situation burster; During described structure stage resolution ratio burster, in iterative data grouping, to virtual reality scenario image, only choose virtual objects region as main with reference to collection region; And to real scene image, only choose and concentrate the akin region of virtual objects as auxiliary with reference to collecting region with main reference; For given image-region, calculate the actual situation grouping feature of every bit in region; Utilize used distance compression method to compress the given just auxiliary actual situation grouping feature with reference to collection region, obtain the region main body characteristic that this region is corresponding; The just auxiliary region main body characteristic set input support vector machine grouping with reference to collection is trained, obtains stage resolution ratio actual situation burster;
C3, in iterative data grouping, utilize actual situation grouping feature, extract the regional correlation feature of virtual reality scenario and outdoor scene respectively, structure region class actual situation burster; During described structure region class actual situation burster, in iterative data grouping, for just auxiliary with reference to collection region, itself is considered as subject area to be determined; And the homalographic rectangular area outside regional space solid is considered as the background area residing for object; Extract the actual situation grouping feature of every bit in subject area and background area respectively; The actual situation grouping feature that in objects of statistics region and background area, all-pair is answered, forms the joint distribution histogram of subject area feature and the joint distribution histogram of background area feature respectively; Normal state squared-distance between computing two histograms, is regarded as and weighs object and the feature of difference between background residing for it, be called regional correlation feature; The just auxiliary input of the regional correlation characteristic set with reference to the collection support vector machine grouping of extracting is trained, obtains region class actual situation burster;
C4, given test virtual reality scenario, utilize stage resolution ratio actual situation burster and small-bore detected envelope to detect, and obtains the Virtual matching figure reflecting each pixel actual situation group result; Described Virtual matching figure construction step comprises: for the virtual reality scenario image of input, utilizes small-bore detected envelope with less stride scanning entire image; The region main body characteristic of the little image block in each small-bore detected envelope of computing; The region main body characteristic of all little image blocks is input in stage resolution ratio actual situation burster, obtains the region main body characteristic evaluation of each little image block, evaluate the high degree of certainty representing that this image block is grouped into virtual region by stage resolution ratio burster high; Because the size of detected envelope is very little and densely distributed relative to entire image, therefore the region main body characteristic evaluation of each little image block can be mapped to the center pixel of this image block, and it can be used as the Virtual matching of this central pixel point; Thus constitute the Virtual matching figure of whole virtual reality scenario image;
C5, defining virtual reliability graph, and on the basis of Virtual matching figure, utilize Iamge Segmentation to obtain testing the virtual reliability graph of virtual reality scenario; Described virtual reliability graph builds and comprises: the Virtual matching figure for virtual reality scenario image carries out Iamge Segmentation process, and recording all Virtual matching is positive point; Arrange a fixing number percent N%, record the position that front N% and these that all Virtual matching are positive point are residing on original image, these points are called high Virtual Calibration point; Arrange one to fix and relatively little constant M, record front M point and these positions residing on original image that all Virtual matching are positive point, these points are called the highest Virtual Calibration point; Can be ensured by optimum configurations, the highest Virtual Calibration point is also contained in the set at high Virtual Calibration point place simultaneously, and namely the highest Virtual Calibration point is the part that in high Virtual Calibration point, Virtual matching value is the highest; Positional information on comprehensive high Virtual Calibration point, the highest Virtual Calibration point and place original image thereof, forms virtual reliability graph;
C6, based on virtual reliability graph, carry out the fuzzy location of virtual objects, obtain blurred shapes and the position of virtual object space solid; Described obtain virtual object space solid blurred shapes and position be specially: multiple subgraph is divided into the virtual reliability graph obtained, is tried to achieve the distribution center of the high Virtual Calibration point in each subgraph respectively; Subgraph center is considered as the virtual objects central point of candidate, searches for respectively to external expansion from each central point, obtain the region that high Virtual Calibration point is densely distributed; For the region that high Virtual Calibration point is densely distributed, Approximate Calculation goes out the candidate target shape in this region respectively, in conjunction with the positional information in this region, forms virtual objects preliminary candidate region; In virtual objects preliminary candidate region, according to the number of its each self-contained high Virtual Calibration point with the highest Virtual Calibration point, select one that weighted number is maximum, it can be used as virtual objects candidate region, namely this region contains virtual object space solid blurred shapes and positional information;
C7, on the basis of the fuzzy location of virtual objects, in test virtual reality scenario image, utilize region class actual situation burster and heavy caliber detected envelope to detect, obtain the final detection result of virtual objects; Described virtual objects detection method is specially: intensive sampling around the virtual objects candidate region in test virtual reality scenario, construct multiple overlapped detected envelope, and use region class actual situation burster to divide into groups, choose and evaluate the final detection result of best detected envelope as virtual objects.
According to the embodiment of the present invention, described actual situation upgrades the virtual drawing and virtual reality collision detection unit that regulon generates according to described virtual reality construction unit, carry out renewal with the testing result of virtual reality reliability judging unit to virtual drawing to regulate, specifically comprise to generate accurate virtual image: when actual situation collision detection result is actual situation conflict or virtual reality reliability judged result is unreliable, select the virtual graph picture frame abandoning current generation, and estimate the movement position of next frame virtual scene, again obtain real scene image and generate new virtual image.
The above; be only the present invention's preferably embodiment, but protection scope of the present invention is not limited thereto, is anyly familiar with those skilled in the art in the technical scope that the present invention discloses; the change that can expect easily or replacement, all should be encompassed within protection scope of the present invention.Therefore, protection scope of the present invention should described be as the criterion with the protection domain of claim.