CN103295258A - Novel multifunctional atmospheric science data graphic rendering engine - Google Patents

Abstract

The invention provides a multifunctional graphic rending engine integrating multiple graphic rendering techniques, in particular to a novel multifunctional atmospheric science data graphic rendering engine. The graphic rendering engine is characterized in that (1) the graphic rendering engine can provide speeds and efficiency of hardware accelerating graphic rendering techniques and meanwhile provide a vector graphic output function specific to software graphic rendering techniques; (2) when a software program of the graphic rendering engine is used, the graphic rendering techniques can be switched in operation, display properties of graphic images can keep unchanged, and a routine is needless to be restarted; (3) the multifunctional graphic rendering engine can be conveniently transplanted to operation system platforms and hardware equipment of different structures; (4) image file output of various coded formats is supported. The graphic rendering engine is integrated with a weather information analysis processing system or a geographical information system and the system is applied to practical services and scientific researches; interaction manufacturing of a weather map, real-time display of large quantities of scientific data, automatic generation of graphic weather forecast products and flexible obtaining of illustrations in scientific research papers can be achieved; accordingly requirements of weather forecast service and atmosphere scientific research are well met.

Description

A kind of novel multifunctional atmospheric science datagraphic render engine

One, technical field

Along with the atmospheric science technology rapid development, scientific research and operational staff have higher requirement to figure display speed and the efficient of related software, and requirement can be finished analysis and the visualization function of mass data in the short as far as possible time.Requirement simultaneously can be preserved into shown figure the graphic file of vector format, is convenient to service product or achievement in research are demonstrated and delivered.

Many powerful science data analyses, software for display system are arranged in the current atmospheric science field, and each system has the graphics rendering engine that oneself is realized.For example, widely used MICAPS[1 in China's meteorological station business]; The GrADS[2 that often uses when domestic worker in meteorology maps] and Surfer[3]; The Tao Zhu treasure professor of Peking University presides over the virtual reality visualization system LiveView[4 of exploitation]; The objective analysis diagnosis graphics system [5] of the Zheng Yong of China Meteorological Administration light etc. research and development, China Meteorological Administration wins honour for recklessly and waits the drafting system MeteoGS towards the meteorology application of exploitation.In addition, also have popular in the world general meteorological software systems to comprise the MetView[6 in European forecasting centre], NCL[7], Vis5D[8], IDV[9] and the Google earth [10] etc.

In general, above-mentioned software systems or be business platform (as MICAPS, MetView) and do not consider factors such as figure output quality, graphical element composing.At the data visualization tool (as GrADS, Surfer, NCL, IDV etc.) of scientific research, do not have to consider to requirement professional, service.Be a drafting system (as MeteoGS), function ratio is more single, can not satisfy all demands of user.

The problem that exists at present both at home and abroad science data visual software system and graphics rendering engine thereof, it is the graphics rendering engine of one that the present invention has designed and Implemented a kind of novel multiple graphics rendering technology of collection, it can either provide hardware graphics to accelerate to play up speed and the efficient that technology is enjoyed, and can provide the software graphics rendering technology peculiar vector graphics output function again.In addition, this graphics rendering engine can be transplanted on the operating system platform and hardware device of different frameworks more conveniently.For availability and the novelty that proves this graphics rendering engine in practice, on the basis based on this graphics rendering engine, realized that one makes things convenient for the atmospheric science worker to use, satisfies the vocational work needs, has the meteorological software systems of high quality graphics output function.Graphics rendering technology can be switched in this system in the middle of operation, and keeps the display properties of graph image constant, and need not restart routine.And support the image file output of multiple coded format.

The list of references that this section is quoted:

[1] Li Yuean, Cao Li, Gao Song, Luo Bing, " MICAPS forecast business platform current situation and development, " meteorology, vol.26, no.7,2010.

[2]Grid?Analysis?and?Display?System(GrADS).http://www.iges.org/grads.

[3]Surfer-contouring?and3D?surface?mapping?software.http://www.ssg-surfer.com.

[4] Wang Hongqing, Zhang Yan, the Tao Zhu treasure, Chen Shoujun, " the large complicated data computer of five dimensions is visual, " natural science progress, 1998.

[5] Zheng Yongguang, Wang Hongqing, Tao Zhu treasure, Chen Min, " two-dimentional meteorological mapping software under the Windows--objective analysis diagnosis graphics system, " meteorology, 2002.

[6]Metview.https://software.ecmwf.int/wiki/display/METV/Metview.

[7]CISL’s?NCAR?Command?Language(NCL).http://www.ncl.ucar.edu.

[8]Vis5D?Home?Page.http://www.ssec.wisc.edu/～billh/vis5d.html.

[9]Unidata|IDV.http://www.unidata.ucar.edu/software/idv.

[10] Google earth .http: //www.***.com/earth/index.html.

Two, background technology

On developer's angle, be used alone graphics rendering technology and do not have very big difficulty.For example in just part accelerates to play up, can select Direct3D technology [11] or OpenGL technology [12] among the DirectX.When adopting software to play up, can select GDI/GDI+.Even can consider these ins and outs and directly use integrated Development Framework (as J2EE .NET, Qt, MFC) graphics rendering technology of the inside encapsulation.Can also use more senior graph image function library, as Magics.These select respectively to have relative merits.Have best compatibility as the Direct3D technology under Windows operating system, the display card of Xiao Shouing is all supported Direct3D on the market; Shortcoming is only to support windows platform.The OpenGL technology then has an advantage cross-platform, and namely nearly all platform is all supported the OpenGL technology more or less; Shortcoming is that the function difference of realizing between each platform is bigger.The GDI/GDI+ technology then is the basis of vector graphics under the Windows.Graphics rendering technology abstraction hierarchy in the integrated Development Framework is higher, and more tight in combination with general frame.

Except having multiple popular graph rendering tool storage room, for a certain graph rendering tool storage room, at different computing platforms, under the different operating system, it realizes also having very big difference.Be example with OpenGL, Visual C++ only supports OpenGL v1.2 version under Windows operating system, and the developer need realize support to latest edition by third party's function library.On mobile devices such as mobile phone, only realized the subclass of OpenGL: OpenGL ES.In contrast, on graphics workstation, manufacturer often provides the expansion of various OpenGL, to strengthen the function of OpenGL.

The software systems of introducing in first segment are general uses single graph rendering tool storage room to realize the graph image Presentation Function, uses OpenGL as MICAPS, Vis5D, and MeteoView uses Magics, and MeteoGS uses GDI+, and IDV uses JAVA 3D.Have only the Google earth to use OpenGL and two kinds of technology of DirectX.Yet, when the user is switched between these two kinds of graphics rendering technology, must restart routine so that change come into force.

The problem that a certain graph rendering tool storage room of single use brings comprises:

At first, the realization of graph image Presentation Function is subject to employed graph rendering tool storage room.As can't provide function of hardware acceleration when using GDI/GDI+, and can't realize vector format output when using OpenGL.

The second, when needs arrive the unsupported operating system of employed graph rendering tool storage room or hardware device to application migration, need realize graph image Presentation Function module again.Be transplanted to Unix operating system following time as the application that will use DirectX, need to use other graph rendering tool storage rooms (as OpenGL) realize the graph image Presentation Function.

The list of references that this section is quoted:

[11]Frank?D?Luma，Introduction?to3D?Game?Programming?with?DirectX10.：Wordware?Publishing?Inc.，2008.

[12]Mason?Woo，Jackie?Neider，Tom?Davis，and?Dave?Shreiner，OpenGL?Programming?Guide：The?Office?Guide?to?Learning?OpenGL，Version1.2，3rd?ed.：Addison-Wesley?Professional，1999.

Three, summary of the invention

3.1 multifunction graphics render engine

3.1.1 the concept of graph rendering level of abstraction and design realize

At the problem that a certain graph rendering tool storage room of single use brings, consider present several operation systems simultaneously, computing equipment and graph rendering tool storage room and the fact of depositing, we have proposed following target when the design configuration render engine:

The first, with respect to computing platform, the independence of operating system and graph rendering tool storage room.The realization that is graphics rendering engine does not rely on a certain computing platform, operating system and graph rendering tool storage room.

The second, extensibility.Namely can use new graph rendering tool storage room to realize graphics rendering engine when needed, perhaps can more easily graphics rendering engine be transplanted on the new computing platform and operating system.

The 3rd, polymorphism.Namely according to user option or program needs, play up between the tool storage room at different graphic in real time and switch, the function to realize that graphics rendering engine was provided guarantees that simultaneously the display properties of graph image remains unchanged after switching.For example, the user changes the current graph rendering tool storage room that is using by option in program the time of running; Perhaps when user's telnet, program automatically switches under the software render mode.

According to above-mentioned design object, we have proposed this model of graph rendering level of abstraction (Graphics Rendering Abstract Layer abbreviates GRAL as).The logical organization of model is shown in Fig. 1 (seeing Figure of description).In the design of graph rendering level of abstraction, we have used abstract factory (Abstract Factory) this Design Mode [13].Abstract factory mode for create one group logically interrelated or complementary entity object interface is provided, hidden the details of these entity objects to be created simultaneously.This Design Mode of abstract factory provides possibility for creating the function library of crossing over different platform.

In realizing this model of graph rendering level of abstraction, we at first extract the common function of all entity objects, to each function abstractdesription in addition, as an interface of abstract base class.Then to the subclass of class of each entity object definition as abstract base class.In each subclass, realize the specified function of interface.We also need to define a customer-oriented interface class and call the interface function of abstract base class, thereby finish the function of customer requirements.Like this, the interface function that the client needs only the calling interface class to be provided is realized the function that oneself needs, and need not to know how concrete each entity object inside defines and realize.

Specify the graph rendering level of abstraction model that designs based on abstract factory mode below in conjunction with Fig. 1.At first, we define a customer-oriented functional interface class mgral, and it is the abstractdesription of graph rendering level of abstraction.Stated in such that all realize the function declaration of customer demand function.Function d rawString as display string; The function d rawSymbol of displaying symbol etc.In class mgral, we need define a pointer type member variable instance who has encapsulated the abstract base class mgral_toolkit of graph function tool storage room.Abstract base class mgral_toolkit has stated that equally all realize the function declaration of customer demand function, and these functions will provide concrete definition in its subclass.The following procedure example code relation between class mgral and the class mgral_toolkit:

In this example, when the client need be at the program window drafting character string, the interface function drawString that provides among the class mgral is provided for it, in the definition of this function, whether be initialised according to member variable instance, or call interface function among the abstract base class mgral_toolkit or printout miscue information.

After variable i nstance was instantiated, it always pointed to the instance objects of a certain subclass of abstract base class mgral_toolkit.Because a moment can only use a kind of graph function tool storage room to draw, therefore when defining variable instance, used the singlet in the Design Mode (singleton) pattern.The advantage that adopts the singlet pattern is to guarantee that any time has only the subclass example of mgral_toolkit to exist, and only exists a unique access point that this subclass example is carried out read-write operation.Like this, by client's designated parameters, perhaps program configuration in advance, class mgral generates an example of mgral_toolkit subclass, and variable i nstance is pointed to this example.For example, when the client specified current use OpenGL graph function tool storage room, class mgral generated the example of mgral_toolkit subclass mgral_opengl, and variable i nstance is pointed to this example.

The definition of interface when the subclass of abstract base class mgral_toolkit must realize using respective graphical function tool storage room.The definition of interface when for example, having realized among the entity class mgral_opengl using OpenGL graph function tool storage room.Like this, in the time need between different graph rendering function tool storage rooms, switching, only need at first delete variable i nstance, generate corresponding to the example of new graph rendering function tool storage room then and variable i nstance is pointed to this example and get final product.For example, when being transformed into Direct3D graph function tool storage room from OpenGL graph function tool storage room, the example pointer instance of deletion mgral_opengl creates the example of mgral_directx again and variable i nstance is pointed to this example earlier.

The advantage of graph rendering level of abstraction model has:

The first, make the establishment of CLIENT PROGRAM and entity object and function realize being separated, reduced contingency and dependence between CLIENT PROGRAM and the entity object.

Second, be convenient to switching entity object neatly, the establishment of a certain class entity object is selected according to user option or program parameter by customer-oriented interface class mgral and is realized, when transforming to another kind of entity object from a class entity object, as long as the example of the last class entity object of deletion, the example of creating back one class entity object then gets final product.The realization of all functions function is also switched thereupon together.For example, carry in the model in the present invention, when being transformed into GDI from OpenGL, only need the example of deletion mgral_opengl class, the example of creating mgral_gdi then gets final product.

The 3rd, be convenient to add new entity object.When needs add new entity object, only need the entity subclass of definition abstraction interface class mgral_toolkit to get final product.The realization of original entity object and CLIENT PROGRAM are all unaffected.For example when needing to use the function realization of Direct3D graph function tool storage room implementation model, only need the subclass mgral_directx of definition mgral_toolkit to get final product.

An advantage of abstract factory mode is the dependence that has retrained very naturally between each entity object.For example, when using OpenGL function tool storage room to play up, can only use the OpenGL function; And when using the Direct3D function library to play up, can only use the Direct3D function.

3.1.2 mixed graph render engine---the realization of graph rendering level of abstraction

We have selected the graph rendering tool storage room of the relative bottom of direct use to realize the mixed graph render engine in realizing graph rendering level of abstraction model, rather than use the graphics rendering technology that encapsulates in the integrated Development Framework.The benefit of doing like this has:

● do not rely on a certain integrated Development Framework; The development progress of integrated Development Framework and program error wherein can not influence the development of graphics rendering engine and the correctness of function.

● be convenient to system transplantation on the unsupported operating system of selected integrated Development Framework or computing platform.

At present, our multifunction graphics render engine has been realized the encapsulation to three kinds of graph function tool storage rooms: OpenGL, Direct3D and GDI.They are represented with three entity subclass mgral_opengl, mgral_directx and the mgral_gdi of abstract base class mgral_toolkit respectively.Stated the function declaration of following function among the abstract base class mgral_toolkit:

● the window view setting;

● the conversion between graph rendering tool storage room coordinate system and the device coordinate system;

● background color, foreground color setting;

● the geometric figure attribute arranges;

● the font attribute setting;

● light and material function;

● texture generates and shows;

● literal, geometric playing up;

● display list;

● polygon trigonometric ratio (tesselation);

Three entity subclass mgral_opengl, the mgral_directx of abstract base class mgral_toolkit and mgral_gdi realize the definition of above-mentioned interface respectively according to the api function of each self-corresponding tool storage room.

3.2 understand the figure coordinate system

By the discussion of a last joint, can be according to user's needs conversion graph rendering tool storage room neatly in the multifunction graphics render engine.In this conversion, we need keep the display properties of figure all in the program window, image object constant.The display properties here comprises magnification ratio, position, color and texture.Keeping display properties constant is principal character of the present invention.Yet the definition to coordinate system in the different graph rendering tool storage rooms has nothing in common with each other, and this coordinate figure conversion of playing up in the tool storage room at different graphic for same graphical object has brought difficulty.Therefore, as the first step of graph rendering, being necessary has one to understand clearly to the coordinate system in the graphic display system.

In a graphic display system, there are a plurality of character, coordinate system that purposes is different.Usually have following several:

3.2.1 view coordinate system

Window manager in each operating system (as the GDI under the Windows, the X Window under the Unix etc.) all is the coordinate figure that comes the presentation graphic object with the view coordinate system.In the view coordinate system, initial point is always in the upper left corner of indicator screen.X axle along continuous straight runs increases to the right; The y axle vertically increases downwards.Therefore be positioned at the lower right corner of indicator screen apart from initial point point farthest.The unit of view coordinate system is pixel, and the view coordinate system is also referred to as device coordinate system (device coordinate system) in the Windows graphics system.

3.2.2 the coordinate system that use in the graph function storehouse

In the graph function tool storage room of present main flow such as OpenGL, GDI and Direct3D, all defined the coordinate system of oneself.Here introduce the coordinate system that these three graph function tool storage rooms use respectively.

3.2.2.1OpenGL coordinate system

Under default configuration, the direction of coordinate axis is consistent in the Cartesian coordinates in the OpenGL coordinate system in the direction of coordinate axis and the mathematics, and namely x axle along continuous straight runs increases to the right; The y axle vertically upwards increases.The z axle becomes the right-handed coordinate system relation with the xy axle.

A three-dimensional scenic is presented on the two dimensional display plane, and in fact this process has comprised following coordinate transformation process:

Mould is looked conversion

Mould is looked conversion and is comprised model transferring (modeling transformation) and two processes of view transformation (viewing transformation).Because video camera and objective direction, therefore the duality between the distance puts together model transferring and view transformation usually, is referred to as mould and looks conversion (modelview transformation).Mould is looked conversion the coordinate of object point is transformed into the camera coordinate system from world coordinate system.OpenGL provides function gluLookAt to realize that mould looks conversion.It is claimed as:

void?gluLookAt(GLdouble?eyex，GLdouble?eyey，GLdouble?eyez，GLdouble?centerx，GLdouble?centery，GLdouble?centerz，GLdouble?upx，GLdouble?upy，GLdouble?upz)；

Parameter eyex wherein, eyey, eyez represent the position at video camera center, centerx, centery, centerz represent the center of observed object.Parameters u px, upy, upz have defined the direction of camera coordinate system z axle.

Projection and cutting conversion

In computational geometry, projective transformation (projection transformation) refers to the point on the three-dimensional body is mapped to process on the two-dimensional projection plane.Yet in OpenGL, projective transformation is not the projective transformation in the geometry, and in fact it comprised projection and two processes of cutting, and has defined visible space (being called the cutting pyramid).Be positioned at naming a person for a particular job of visible space outside and be dropped, be positioned at naming a person for a particular job of visible space inside and be retained.

Projective transformation transforms to the cutting coordinate system with the coordinate of three-dimensional body point from camera coordinate system.And then with the cutting coordinate conversion be normalized device coordinate (normalized device coordinates, NDC).

Projective transformation is divided into two kinds, and a kind of is perspective projection transformation, and another kind is the rectangular projection conversion.The sight line of having a few on the object in perspective projection intersects at the video camera center, and the imaging size of object far away apart from video camera like this is less than the imaging size of the near object of distance video camera.This is consistent with principle of eyes imaging.Therefore perspective projection is used to play up natural scene, to obtain authenticity.OpenGL provides two function glFrustrum and gluPerspetive to realize perspective projection transformation.Function glFrustum has defined the cutting pyramid that uses the cutting plane to form, and it is claimed as

void?glFrustum(GLdouble?left，GLdouble?right，GLdouble?bottom，GLdouble?top，GLdouble?near，?GLdouble?far)；

Parameter l eft wherein, bottom, near have provided the coordinate in the nearly projection plane lower left corner, and right, top, far have provided the coordinate in the upper right corner, far throw shadow plane.Function gluPerspetive has defined the cutting pyramid that uses visual angle and projection plane the ratio of width to height to form, and it is claimed as:

void?gluPerspetive(GLdouble?fovy，GLdouble?aspect，GLdouble?znear，?GLdouble?zfar)；

Wherein parameter f ovy has defined the visual angle on the vertical direction, and aspect has defined the ratio of projection plane width and height, and znear and zfar have provided nearly projection plane and far throw shadow plane respectively to the distance of video camera central point.

By comparison, in rectangular projection, the sight line of having a few on the object is parallel, and this is equivalent to video camera and is centered close to infinite distant place.The imaging size of object point and its range-independence apart from video camera.Rectangular projection often is used to computer-aided design (CAD) etc. and requires to keep in the constant application of object size.OpenGL provides two function glOrtho and gluOrtho2D to realize rectangular projection, and it is claimed as

void?glOrtho(GLdouble?left，GLdouble?right，GLdouble?bottom，GLdouble?top，GLdouble?near，GLdouble?far)；

void?gluOrtho2D(GLdouble?left，GLdouble?right，GLdouble?bottom，GLdouble?top)；

The parameter-definition of function glOrtho reduce pyramidal scope; Function gluOrtho2D is actually the reduced form of glOrtho, and it has removed two parameters on the z direction, is equivalent to that near is that-1, far is 1 among the glOrtho.

View transformation

View transformation is mapped to the point coordinate on the projection plane in the program view window.Provide function glViewport to realize this conversion among the OpenGL, it is claimed as:

void?glViewport(GLdouble?x，GLdouble?y，GLdouble?width，GLdouble?height)；

Parameter x wherein, y has provided the coordinate in the view window upper left corner, and parameter width and height have provided width and the height of view window.

Compare with the view coordinate system, the unit of OpenGL coordinate system is pixel equally, but has floating point precision, and the view coordinate system only has the integer type precision.

3.2.2.2Windows GDI coordinate system

GDI is the abbreviation of Graphics Device Interface, is the image display module in the Windows operating system nucleus.Coordinate system among the GDI is called logic (logical) coordinate system or form (window) coordinate system.The unit of logical coordinates system can be pixel, rice, foot, point (point) or user-defined unit, and its concrete selection is determined by second parameter among the function S etMapMode:

int?SetMapMode(HDC?hdc，int?fnMapMode)；

In addition, also provide function S etViewportOrgEx among the Windows SDK, SetViewportExtEx is provided with initial point and the scope of view coordinate system; Initial point and the scope of function S etWindowOrgEx and SetWindowExtEx logical coordinates system.

3.2.2.3Direct3D coordinate system

Different with OpenGL coordinate system and Cartesian coordinates, the Direct3D coordinate system is left-handed coordinate system under default configuration.Therefore, in the computational geometry object coordinates, should be noted that this point.

In this process of two dimensional surface that the three-dimensional point in the scene is presented at program window, Direct3D has defined the coordinate transformation process same with OpenGL.

Mould is looked conversion

Direct3D provides function D3DXMatrixLookAtLH to realize that mould looks conversion, namely the coordinate of object point is transformed into camera coordinate system from world coordinate system.Function declaration is

D3DXMATRIX*D3DXMatrixLookAtLH(D3DXMATRIX*pOut，CONST?D3DXVECTOR3*pEye，CONST?D3DXVECTOR3*pAt，CONST?D3DXVECTOR3*pUp)；

Wherein pOut is that mould is looked the matrix that transfer pair is answered; PEye represents the position at video camera center, and pAt represents the center of observed object.Parameter p Up has defined the direction of camera coordinate system z axle.

Projection and cutting conversion

Projection among the Direct3D is divided into two kinds of perspective projection and rectangular projections equally.The function corresponding with glFrustum is D3DXMatrixPerspectiveOffCenterLH (D3DXMATRIX*pOut, FLOAT left, FLOAT right, FLOAT bottom, FLOAT top, FLOAT znear, FLOAT zfar);

Parameter l eft wherein, bottom, znear have provided the coordinate in the nearly projection plane lower left corner, and right, top, zfar have provided the coordinate in the upper right corner, far throw shadow plane.The function corresponding with gluPerspective is

D3DXMATRIX*D3DXMatrixPerspectiveFovLH(D3DXMATRIX*pOut，FLOAT?fovY，FLOAT?Aspect，FLOAT?znear，FLOATzfar)；

Wherein parameter p Out is the matrix of projective transformation correspondence; FovY has defined the visual angle on the vertical direction, and Aspect has defined the ratio of projection plane width and height, and znear and zfar have provided nearly projection plane and far throw shadow plane respectively to the distance of video camera central point.

In rectangular projection, the function corresponding with glOrtho is

D3DXMATRIX*D3DXMatrixOrthoOffCenterLH(D3DXMATRIX*pOut，FLOAT?l，FLOAT?r，FLOAT?b，FLOAT?t，FLOAT?zn，FLOAT?zf)；

Wherein parameter p Out is the matrix of projective transformation correspondence.

View transformation

Direct3D provides function R SSetViewports to realize the view transformation function:

RSSetViewports(UINT?NumViewports，const?D3D10_VIEWPORT*pViewports)；

Wherein type is that the structure parameter p Viewports of D3D10_VIEWPORT has defined the coordinate in the view window upper left corner and width and the height of view window.

3.2.3 graphical object coordinate system

In the graph image edit routine, graphical object (perhaps image) is presented in the client area of program window.In order to describe the position of some unique point in the graphical object, we generally have two kinds of selections, and the first, use the coordinate system in the graph function tool storage room; The second, introduce the graphical object coordinate system.Practice shows that it is necessary introducing the graphical object coordinate system.Will so why, also be introduced separately into the graphical object coordinate system? reason is that the user need carry out operations such as convergent-divergent, movement to graphical object, and these application make thing become complicated.

Digital photograph with household's good fortune of the whole family is example, and the size that might as well establish this photo is 4000 * 3000 pixels, and mother and child's distance is 1000 pixels in the photo, and child is positioned at the center of photo.When watching this picture in the display that is 1024 * 768 pixels a resolution, if photo is amplified to original size, then have only wherein the part can be in sight, at this moment mother and child's distance still is 1000 pixels.Yet dwindling photo usually, the user can see wherein full content.Is mother and child's distance also 1000 pixels after photo is scaled? another method be the user by mobile photo to see full content, when mobile photo, is child's position also at the center of screen? be not difficult to find out, more than under two kinds of situations the answer of problem all negate.

Top example shows that because image is carried out operations such as convergent-divergent, translation, therefore the inherent characteristic that can't portray the graphical object coordinate figure when using the coordinate system in the graph function storehouse is necessary to introduce the graphical object coordinate system.Like this, in the graphical object coordinate system, the coordinate figure of a bit relative graphical object coordinate origin of certain on the graphical object is constant; Distance on the graphical object between any 2 is constant.This unchangeability is absolutely necessary in image is handled.For example in the above in the example of good fortune of the whole family digital photograph, under the graphical object coordinate system, child is positioned at the center of photo all the time; Mother and child's distance is 1000 pixels always, and this fact is not carried out operations such as convergent-divergent, translation with the user and changed to image.

3.2.4 world coordinate system

In Professional Meteorological software or Geographic Information System (GIS) software, the demonstration of map is necessary.So just introduce world coordinate system naturally with the position of expression arbitrary objects in the space.The locus of object uses longitude, latitude and height above sea level to represent usually.Yet in program window, represent the object in the three dimensions with two-dimensional coordinate, can bring many technical difficulties.The operation that wherein relates to comprises projection, reduction etc.Cheeringly be to have had the software function storehouse of geography information aspect of many maturations for people to use at present.

3.3 the coordinate transform of graph image object under different coordinates

Figure, the image object coordinate transform function under different coordinates is to guarantee render engine constant prerequisite of display properties when conversion graph function tool storage room proposed by the invention.

3.3.1 the conversion between graph function tool storage room coordinate system and the view coordinate system

Provide two function convertCoordDev2Toolkit and convertCoordToolkit2Dev to realize between the two coordinate conversion among the class mgral, namely function convertCoordDev2Toolkit is coordinate in the graph function tool storage room coordinate system with coordinate conversion in the view coordinate system; Function convertCoordToolkit2Dev is coordinate in the view coordinate system with the coordinate conversion in the graph function tool storage room coordinate system.Notice that because we introduce the concept of graph rendering level of abstraction, function convertCoordDev2Toolkit and convertCoordToolkit2Dev can realize the coordinate system of any graph function tool storage room and the coordinate conversion between the view coordinate system according to the graph rendering function tool storage room of current use automatically.

3.3.1.1OpenGL the conversion between coordinate system and the view coordinate system

Under rectangular projection, the coordinate conversion between OpenGL coordinate system and the view coordinate system is determined by formula (1).Parametric description among function glOrtho and the gluOrtho2D coordinate figure (x under the view coordinate system a bit on the object _Vp, y _Vp) with the OpenGL coordinate system under coordinate figure (x _GL, y _GL) between relation:

x _vp＝sgn(right-left)(x _GL-left)

(1)

y _vp＝sgn(top-bottom)(y _GL-bottom)

Wherein

$\mathrm{sgn}\left(x\right)=\left\{\begin{array}{cc}-1,& x<0\\ 0,& x=0\\ 1,& x>0\end{array}\right.$

By above equation as can be known, when calling the function of following form,

gluOrtho2D(0，w，0，h)；

The OpenGL coordinate system overlaps with the view coordinate system.

When calling the function of following form,

gluOrtho2D(-w/2，w/2，-h/2，h/2)；

The initial point of OpenGL coordinate system is positioned at screen center.

When calling the function of following form,

gluOrtho2D(0，w，h，0)；

The initial point of OpenGL coordinate system is positioned at the lower left corner of screen.

In above example, w and h are width and the height of client area (client area) in the program window, by function glViewport appointment.

Under perspective projection, the coordinate conversion between OpenGL coordinate system and the view coordinate system is determined by formula (2).

x _vp＝sgn(l-r)(nx _e/z _e+l)

(2)

y _vp＝sgn(b-t)(ny _e/z _e+b)

3.3.1.2GDI the conversion between coordinate system and the view coordinate system

Provide function LPtoDP and DPtoLP to realize the coordinate conversion between GDI coordinate system and the view coordinate system in the window application interface.Yet the rule of coordinate conversion is first parameter by these two functions---device context (Device Context) variable---decision.And this device context variable is the window that provides by in the window application interface---view transfer function SetMapMode, and SetViewportOrgEx, SetViewportExtEx, SetWindowOrgEx and SetWindowExtEx provide.Wherein comprise character string Viewport function setup initial point and the scope of view coordinate system; Comprise character string Window function setup initial point and the scope of GDI coordinate system (be called among the Windows logical coordinates system).If make (x _{O, vp}, y _{O, vp}), w _Vp, h _VpInitial point and scope for the view coordinate system; (x _{O, log}, y _{O, log}), w _Log, h _LogBe initial point and the scope of logical coordinates system, then have following formula to set up:

$x_{\log }=(x_{\mathrm{vp}}-x_{o,\mathrm{vp}})\frac{w_{\log }}{w_{\mathrm{vp}}}+x_{o,\log }$

$\left(3\right)$

$y_{\log }=(y_{\mathrm{vp}}-y_{o,\mathrm{vp}})\frac{h_{\log }}{h_{\mathrm{vp}}}+y_{o,\log }$

Can realize GDI coordinate system (x by (3) formula _Log, y _Log) and view coordinate system (x _Vp, y _Vp) between coordinate conversion.

3.3.1.3Direct3D the conversion between coordinate system and the view coordinate system

Under the orthogonal perspective projection, the coordinate transformation relation between Direct3D coordinate system and the view coordinate system is provided by formula (4).

x _vp＝sgn(right-left)(x _D3D-left)

(4)

y _vp＝sgn(top-bottom)(top-y _D3D)

(x wherein _Vp, y _Vp) be the coordinate figure under the view coordinate system, (x _D3D, y _D3D) be the coordinate figure under the Direct3D coordinate system.Other implications that meet are identical with formula (1).

Coordinate conversion formula under the perspective projection is seen formula (5):

x _vp＝sgn(r-1)(nx _e/z _e-1)

(5)

y _vp＝sgn(t-b)·(t-ny _e/z _e)

3.3.2 the conversion between graph function tool storage room coordinate system and the graphical object coordinate system

3.3.2.1OpenGL the conversion between coordinate system and the graphical object coordinate system

Although the graphical object coordinate system provides the unchangeability of describing the object feature point position, we will be transformed into it under the coordinate system in the graph function tool storage room, finish mapping operation with the function in the calling graph shape function tool storage room.In this transfer process, we introduce three variablees: preceding two variable (x _c, y _c), the position of whole graphical object (center of mass point of this object commonly used is described) in the OpenGL coordinate system described.When graphical object was moved, these two variablees also will upgrade thereupon.The 3rd variable z described the scale factor between OpenGL coordinate system and the graphical object coordinate system unit length ¹, when the convergent-divergent graphical object, variable z needs to upgrade thereupon.

When the program initial launch, these three variablees should be endowed more rational initial value.The coordinate figure of 1 A is on the graphical object:

$x_I=\frac{x_{\mathrm{GL}}-x_c}{z}$

$\left(6\right)$

$y_I=\frac{y_{\mathrm{GL}}-y_c}{z}$

(x wherein _GL, y _GL) and (x _I, y _I) respectively on the presentation graphic object 1 A at OpenGL coordinate system and the coordinate figure in the graphical object coordinate system.

When the mobile graphics object, the coordinate figure of graphical object barycenter in the OpenGL coordinate system is changed to:

x _C＝x _C+Δx _GL

(7)

y _C＝y _C+Δy _GL

¹Here only consider isotropic convergent-divergent situation.Under anisotropic convergent-divergent situation, should introduce two variable (z _x, z _y) represent the scale factor on level and the vertical both direction respectively.

(Δ x wherein _GL, Δ y _GL) represent the side-play amount of mouse in the OpenGL coordinate system.

When graphical object was carried out convergent-divergent, scale factor z need be updated to:

z′＝z*f (8)

F zoom factor wherein.

When graphical object is carried out convergent-divergent, the position that the user generally wishes corresponding cursor of mouse point on the graphical object corresponding cursor of mouse point still behind convergent-divergent.On the angle of function realization, this means the coordinate figure (x of point under the OpenGL coordinate system of corresponding cursor of mouse on the graphical object _GL, y _GL) constant.In addition, the coordinate figure (x in this dot pattern object coordinates system _I, y _I) also can not change because there being unchangeability.Therefore, by formula (6) translation variable (x _c, y _c) should move to (x _c', y _c')

x _c′＝x _GL-x _Iz′ (9)

y _c′＝y _GL-y _Iz′

The scale factor behind z ' convergent-divergent wherein.

Simultaneous formula (6) and formula (9), translation variable (x _c, y _c) displacement be

$\mathrm{\&Delta;}{x}_c={x_c}^{\&prime;}-x_c=(1-\frac{z^{\&prime;}}{z})(x_{\mathrm{GL}}-x_C)$

$\left(10\right)$

$\mathrm{\&Delta;}{y}_c={y_c}^{\&prime;}-y_c=(1-\frac{z^{\&prime;}}{z})(y_{\mathrm{GL}}-y_C)$

In our the superincumbent discussion graphical object coordinate system is put in the OpenGL coordinate system, this is from variable (x _c, y _c) be to describe whole graphical object (center of mass point of this object commonly used is described) this definition of position in the OpenGL coordinate system as can be seen.Certainly, we also can namely be put into the OpenGL coordinate system in the graphical object coordinate system this hypothesis conversely.At this moment, we use variable (x _o, y _o) replacement (x _c, y _c), represent the position of initial point in the graphical object coordinate system of OpenGL coordinate system.The meaning of variable z is constant.

Given above-mentioned three variablees, the coordinate figure of 1 A is on the graphical object:

$x_I=x_O+\frac{x_{\mathrm{GL}}}{z}$

$\left(11\right)$

$y_I=y_O+\frac{y_{\mathrm{GL}}}{z}$

Similar to the derivation of formula (6), think in the following formula that the origin value of OpenGL coordinate system is (0,0).

When the mobile graphics object, the coordinate figure of graphical object barycenter in the graphical object coordinate system is changed to:

$x_O=x_O-\frac{\mathrm{\&Delta;}{x}_{\mathrm{GL}}}{z}$

$\left(12\right)$

$y_O=y_O-\frac{{\mathrm{\&Delta;y}}_{\mathrm{GL}}}{z}$

Variable (Δ x wherein _GL, Δ y _GL) be the variable of the same name of formula (7).Notice minus sign in the formula (12) when graphical object is moved, the displacement vector of the relative graphical object coordinate origin of the initial point of OpenGL coordinate system is opposite with the displacement vector orientation of graphical object.

When the convergent-divergent graphical object, formula (8) is still set up.In addition, by formula (11), for the position that guarantees corresponding cursor of mouse point on the graphical object behind convergent-divergent still correspondence the cursor of mouse point, the position of the initial point of OpenGL coordinate system in the graphical object coordinate system is behind the convergent-divergent:

${x_O}^{\&prime;}=x_I-\frac{x_{\mathrm{GL}}}{z^{\&prime;}}$

$\left(13\right)$

${y_O}^{\&prime;}=y_I-\frac{y_{\mathrm{GL}}}{z^{\&prime;}}$

In like manner, simultaneous formula (11) and formula (13), the position displacement of the initial point of OpenGL coordinate system in the graphical object coordinate system is

$\mathrm{\&Delta;}{x}_O={x_O}^{\&prime;}-x_O=(1-\frac{z}{z^{\&prime;}})(x_I-x_O)$

$\left(14\right)$

$\mathrm{\&Delta;}{y}_O={y_O}^{\&prime;}-y_O=(1-\frac{z}{z^{\&prime;}})(y_I-y_O)$

In program realized, the coordinate figure conversion (formula (6)) between graphical object coordinate system and the OpenGL coordinate system was realized by the member function convert_coordinate_gl2im among the C++ class mCoordinateSystem and convert_coordinate_im2gl.Function input parameter variable x_origin wherein, y_origin and zoom_scale represent variable (x respectively _o, y _o) and z.Translation variable (x when carrying out the graphical object convergent-divergent _c, y _c) displacement calculate (formula (9)) and realized by function zoomin and zoomout.

3.3.2.2GDI the conversion between coordinate system and the graphical object coordinate system

Because GDI logical coordinates system all is the feature locations of describing object with the graphical object coordinate system, so we can merge both, namely with GDI logical coordinates system as the graphical object coordinate system.So just there is not the transfer problem between GDI logical coordinates system and the graphical object coordinate system.

Remaining only problem is to handle translation variable (x when carrying out the graphical object convergent-divergent _c, y _c) and the calculating of zoom variables z.This problem difficulty is bigger, because glScale* and the such function of glTranslate* among the similar OpenGL are not provided in the window application interface.By analysis, have two kinds of methods available, a kind of is the position coordinates attribute that changes graphical object.The drawback of doing like this is when changing graph rendering function tool storage room, also will change back the position coordinates attribute of graphical object come, and this operation easier is bigger, even can not accomplish in OpenGL and Direct3D.Destroyed constant this advantage of coordinate figure in the graphical object coordinate system simultaneously.Another kind method is the scope that changes window in the logical coordinates system.Like this, keeping under the constant situation of range of views, increasing or dwindle window ranges and be equivalent to and dwindle or amplify graphical object.The present invention has used back a kind of method.

3.3.2.3Direct3D the conversion between coordinate system and the graphical object coordinate system

Because Direct3D provides the coordinate transformation function identical with OpenGL, therefore the discussion result (formula (6) is to formula (13)) in the 3.3.2.1 joint stands good in this section institute main topic of discussion.Therefore no longer repeat here.

3.3.3 the conversion between graphical object coordinate system and the world coordinate system

In system realized, module maproj had encapsulated map projection transformation, common functions such as coordinate cutting.Provide two functions to realize conversion between graphical object coordinate system and the world coordinate system in this module: function project_image2world is converted to coordinate figure under the world coordinate system with the coordinate figure under the graphical object coordinate system; Function project_world2image is converted to coordinate figure under the graphical object coordinate system with the coordinate figure under the world coordinate system.Notice that the conversion of having only between world coordinate system and the graphical object coordinate system is just meaningful, and the conversion between world coordinate system and graph function tool storage room coordinate system and world coordinate system and the view coordinate system is nonsensical.This be because have only on the graphical object a bit corresponding to a geographic position.In convergent-divergent or translation, graph function tool storage room coordinate and the view coordinate of this point change, but its corresponding world coordinate is constant.

In function project_image2world and project_world2image, some image coordinate on the graphical object are passed through the Pan and Zoom conversion again:

${x_{\mathrm{im}}}^{,}=\frac{x_{\mathrm{im}}-t_x}{s}$

$\left(15\right)$

${y_{\mathrm{im}}}^{,}=\frac{t_y-y_{\mathrm{im}}}{s}$

(t wherein _x, t _y) being the translation variable, s is the convergent-divergent yardstick.

Why to introduce top coordinate transform? reason is that this conversion is to exist according to the actual requirements.The user often needs its interesting areas is presented at a suitable ratio central authorities of program window when watching graphical object or image.For example, the demestic user wishes the regional in the map is put into the center of program window when work, and size just in time is full of the whole procedure window.On program development personnel angle, this demand is actually the graph image object is presented at certain position in the graph function tool storage room coordinate system with the size that is fit to.Namely to variable (t _x, t _y) and s compose to a suitable numerical value.

Discussion above comprehensive, Fig. 2 (seeing Figure of description) has provided coordinate transformation relation between four coordinate systems and has realized the function of coordinate transform.

The list of references that this section is quoted:

[13]Erich?Gamma，Richard?Helm，Ralph?Johnson，and?John?Vlissides，Design?Patterns：Elements?of?Reusable?Object-Oriented?Software.Boston，Massachusetts：Addison?Wesley?Professional，1994.

Four, description of drawings

Fig. 1: graph rendering level of abstraction structure of models relation.Mutual relationship between the graph rendering level of abstraction model internal data structure.Wherein dotted arrow indicates adduction relationship; Solid arrow is represented the inheritance between the C++ class.Class mgral is the abstractdesription of graph rendering level of abstraction, as customer-oriented functional interface.Stated in such that all realize the function declaration of customer demand function.Class mgral has defined a pointer type member variable instance who has encapsulated the abstract base class mgral_toolkit of graph function tool storage room.The definition of interface when the subclass of abstract base class mgral_toolkit (as mgral_opengl) has realized using a certain graph function tool storage room (as OpenGL).

Fig. 2: the function of coordinate transformation relation and realization coordinate transform between four coordinate systems that the present invention discusses.

Fig. 3: in weather information analysis process system MeteoExplorer, the user can be easily copies to content displayed in the program window on the system clipboard with the form of polar plot or saves as the image file of EMF form, pastes then in the PowerPoint software.Paste then in word processor (as Word) and the demoware (as PowerPoint).

Fig. 4: the demonstration result of system under the Redhat (SuSE) Linux OS who uses mixed graph render engine proposed by the invention.

Fig. 5: mixed graph render engine proposed by the invention is transplanted to rendering result on the Windows RT operating system.

Five, embodiment

Graph rendering level of abstraction conceptual model and realization thereof proposed by the invention---mixed graph render engine, be included in the meteorological geography information interpretation software MeteoExplorer[14 of system as graph rendering module] in, this system architecture adopts stratification, modular design, the exploitation of use C Plus Plus, support Windows, three kinds of operating systems of Linux and SGI Irix.The user interface of system is succinctly friendly, and mode of operation is clicked and drag and drop based on mouse, meets demestic user's use habit.This system supports MICAPS, WMO GRIB1, and WMO GRIB2, GrADS, NetCDF, science data form commonly used such as HDF has been removed the worry of user's conversion, deal with data from.

5.1 support multiple format-pattern output

The MeteoExplorer system can save as content displayed in the present procedure window image file of multiple form, comprises bitmap BMP, JPG, PNG, EMF etc.Wherein EMF is called enhancement mode Windows metafile, be a kind of when image is scaled distortionless vector image form.As shown in Figure 3, the user can be easily the image file (perhaps copying on the system clipboard) that saves as to content displayed loyalty former state the EMF form in the program window, insert (perhaps pasting) then to word processing and demoware commonly used, in Word, PowerPoint.Like this, the user obtains the demonstration result that it is wanted in the mode of What You See Is What You Get with the vector image form.In this image output procedure, system transforms to GDI with current graph function tool storage room, to finish vector output.And then switch back original graph function tool storage room.

5.2 be applied to multiple computing platform

Mixed graph render engine proposed by the invention because it possesses the independence with graph rendering function library and Development Framework, therefore can be realized in several operation systems or computing platform easily.Fig. 4 has provided the demonstration result under the Redhat (SuSE) Linux OS.At this moment, system has used the OpenGL function library to realize graphics rendering engine.Fig. 5 has shown mixed graph render engine proposed by the invention has been transplanted to rendering result on the Windows RT operating system.In this realization, system has used the DirectX11 function library.

In a word, the multifunction graphics render engine of discussion of the present invention is realized and is integrated among the meteorological geography information analysis process system MeteoExplorer through computer program, this system can finish the dynamic switching of graph function tool storage room under the situation that need not to restart.The function of this novelty makes it learn in the research the professional gentle Elephantidae of actual weather forecast can provide following help.At first, realize synoptic mutual making; Secondly, the real-time Presentation Function of magnanimity science data is provided; The 3rd, have the high quality graphics output function, satisfy the demand that the user makes weather forecast figure product, scientific research paper illustration.

The list of references that this section is quoted:

[14]MeteoExploer，Eastmodelsoft?Software.http://www.eastmodelsoft/software/mexplorer.htm。

Claims (4)

1. multifunction graphics render engine that is used for the atmospheric science data visualization, it is all to use one or more graph function tool storage rooms (as OpenGL, DirectX etc.) to realize that it shows output function with the current common ground of popular graphics rendering engine both at home and abroad.

2. graphics rendering engine according to claim 1, it is characterized in that to switch according to setting in advance between different graph function tool storage rooms when the operation based on the computer software programs of this engine, and finish this handoff procedure and do not need to restart software program.

3. graphics rendering engine according to claim 1, it is characterized in that to be chosen in by the user between the different graph function tool storage rooms when the operation based on the computer software programs of this engine and switch, and finish this handoff procedure and do not need to restart software program.

4. graphics rendering engine according to claim 1 is characterized in that after switching the graph function tool storage room, and the figure that shows in the software program window, the display properties of image object are constant.Here the display properties of indication comprises the displaying ratio of graphical object and image, the position in program window, color and texture.

Images