CN102289846B - Tree simulation method based on generalized parametric modeling - Google Patents

Abstract

The invention provides a tree simulation method based on generalized parametric modeling, comprising the steps of: step 1, inputting nongeometric parameters relevant to tree growth; step 2, selecting and reading corresponding geometrical data according to the nongeometric parameters, wherein the geometric data file is generated by the following mode: calibrating a tree seed function structure model on the basis of real-measurement data to obtain the tree seed function structure model, then utilizing the tree seed function structure model to generate a geometric data file of the tree seeds corresponding to different nongeometric parameter values; step 3, establishing a rudimentary model according to the geometric data file; step 4, inputting geometric parameters of a tree to adjust the rudimentary model so as to obtain a tree model; and step 5, carrying out fast and realistic drawing on the tree model. In the tree simulation method, the appearance form and the inner biological mechanism of the tree can be fully unified together, and a geometric model for fully reflecting the tree growth rule can be established.

Description

A kind of tree simulation method based on Generalized parametering modeling

Technical field

The present invention relates to plant modeling and rendering technique, relate in particular to a kind of tree simulation method based on Generalized parametering modeling.

Background technology

Floristics is various, comes in every shape, and complex structure, people are to phytomorph structure, and finding out never of internal mechanism stops, and observes and be familiar with objective things by the mode of figure, is mankind's cognitive styles the most easily.Therefore, along with the development of computer technology, researchist has proposed theory and the method for various plants visual Simulation, for plant visual Simulation has been opened up wide road.

Plant visual Simulation, is mainly the visual of tree modelling, comprises trees modeling and tree rendering, and this is visual two different phases, but both have common target---show truly trees.Below respectively trees modeling and tree rendering are introduced.

One, trees modeling

Trees modeling is reflect the trees in real world and can simulate the process of growth, development and the variation of trees by constructing a naive model.

Common several tree modeling methods mainly comprise: various improved L-systems, fractal, modeling based on image, particIe system, modeling based on parameter control.But in above modeling method, major part is only focused on geometric shape, the moulding etc. of plant, seldom in conjunction with botany knowledge, therefore, can not simulate the advolution process of trees, in the research of agriculture and forestry and application, be subject to larger restriction.And this limitation of above method has promoted the development of structure functional mode, as Amap model (automaton model) and GreenLab model (functional architecture model) etc., these models had both been considered the morphosis of plant, consider again the physiological ecological mechanism of plant, can embody preferably the true rule of plant growth, the growth course of energy simulating plant, and can predict the upgrowth situation of plant.

Amap model is a kind of virtual plant model of the L-of being different from system, combine the knowledge of the structural botany aspect of botanist's proposition, adopt finte-state machine to control the axle growth course of plant, thereby the topologies change in simulate growth process, this technology be otherwise known as " axis of reference technology " or " Stochastic Process Method ".

GreenLab model is on the basis of Amap model, improve and expand, GreenLab model has adopted Dual-Scale Automaton to carry out the form generating process of simulating plant, and GreenLab model has formed the functional mode system that between more complete reflection plant organ, biomass distributes.In Dual-Scale Automaton, be provided with the state of two levels---Macro Flags and micro-state, Macro Flags represents different growing elements, and micro-state represents Ye Yuan, and Macro Flags is formed by micro-combinations of states.Dual-scale automaton, by the combination of this two states and the growth course of circulating analog plant, generates the topological structure of plant.Than the automaton model in Amap model, the expression-form of Dual-Scale Automaton and state parameter are often more succinct, are also more conducive to describe the plant growth feature relevant with growth cycle.GreenLab model is on the basis of Dual-Scale Automaton form generation model, and the biomass that carrys out the each organ of simulating plant according to source-remittance theory of organ distributes, and is distributed and extrapolated plant organ size over time by biomass.

The simulated technological process of GreenLab functional mode as shown in Figure 1, GreenLab model carrys out structure and the function feedback of simulating plant by a series of mathematical formulae iteration, compare other model, GreenLab model has clear superiority aspect embodiment biomechanism and model complexity, mainly contain following advantage: 1) represent respectively growing element and the Ye Yuan of plant with Macro Flags and micro-state, meet the periodically characteristic of growth of plant.And adopt semi-Markov chain to describe the relation between each state, the change of divergence of simulating plant morphosis and the growth of plant more truly.2) GreenLab model has carried out a large amount of simplification to phytomorph structure and growth rhythm, according to the difference of each organ physiological age, each organ or branch are classified, of a sort organ or branch adopt identical growth parameter(s), thereby greatly simplify plant growth parameter, and these growth parameter(s)s adopt the method for statistical fits to obtain, can with the good combination of measured data.3) GreenLab model has comparatively succinct form generation description form, and set up that complete simulation biomass distributes and the functional mode of organ size variation, the growth of simulating plant comparatively truly, has widely applied in model plant growth at present.

But GreenLab model still has many imperfection parts at present, does not also comprise the system model of considering various environmental factors, does not also further consider the physiology course in plant, this all needs improving and improving of researcher from now on.And because the geometric complexity of trees, the measurement to trees and GreenLab parameter fitting etc. has larger difficulty, the research that at present GreenLab is applied to trees is also more difficult.

Two, tree rendering

The geometric model of trees is the visual bases of trees, and realize the visual of trees also needs to use suitable algorithm that tree modelling is drawn out.Tree rendering utilizes computer graphics techniques to process trees three-dimensional geometrical structure exactly, generates the process of photorealism.

Along with tree modeling method enrich constantly and perfect, people more and more pay attention to the rendering technique of trees, for the feature such as tree structure complexity, data volume be huge, people have proposed more rendering algorithm, are mainly classified as three classes according to the difference of drawing element: the 1) method for drafting based on how much; 2) method for drafting based on image; 3) object plotting method.

1) method for drafting based on how much is to adopt sweep trace method for drafting, ray tracing method, radiosity method, Ph0ton mapping technology etc. directly to draw the geometric model of trees.Trees geometric model is more true to nature, careful to model illumination, texture, shade and anti-aliasing processing, can generate effect very true to nature.But often dough sheet quantity is very large due to trees geometric model true to nature, make the Realistic Rendering efficiency of this method very low, be difficult to meet in real time or the demand of interactive application.

2) method for drafting based on image.Drafting based on image utilizes texture image visually to simulate how much of trees, realizes simply, and dough sheet number is few, and drafting efficiency is high.Rendering technique based on image is from image angle, generate the image at different points of view place in certain limit by operations such as interpolation, mixing, distortion, this method can greatly reduce the quantity of drawing geometric graphic element, improves render speed, in interactive system, is widely used.

Wherein, the fidelity of drafting depends on the fineness of texture image, and because image is discrete data, but in the time that viewpoint is nearer, view data can not be segmented, and distortion phenomenon is obvious.And the method just represents trees with one or several dough sheet, does not set up real trees geometric model, can not be applied to the simulation of the complex effects such as swing with the wind of dynamic illumination, tree branches.

3) object plotting method.The geometry complexity of object plotting method and trees is irrelevant, and can obtain the support of graphics accelerator hardware, in recent years, along with updating and the high speed development of computer hardware of algorithm, Volume Rendering Techniques has become the important tool of complex structure natural scene, has had a large amount of application in the real-time rendering field of scale Forest Scene.But in volume drawing, whole trees, even a slice woods are all the entirety that cannot cut apart, make to be difficult to the each several part region of trees is controlled respectively and revised, also be difficult for expressing the complicated dynamic effect of trees, in dirigibility, seem not enough, and volume drawing memory data output be large, computing time is longer, in application, is subject to certain restrictions.

To sum up, existing plant visual Simulation technology also exists following deficiency:

1) for trees modeling, prior art could not fully be united Biological Mechanism inner with it mode of appearance of trees, could not set up the geometric model of abundant reflection arboreal growth rule.

2), for tree rendering, prior art can not improve render speed in improving the drafting sense of reality, there is no to provide the Fast Drawing technology with good visual effect.

Summary of the invention

Technical matters to be solved by this invention is to provide a kind of tree simulation method based on Generalized parametering modeling, Biological Mechanism inner with it mode of appearance of trees fully can be united, and can set up the fully geometric model of reflection arboreal growth rule.

In order to address the above problem, the invention provides a kind of tree simulation method based on Generalized parametering modeling, comprising:

Step 1, inputs the non-geometric parameter relevant to arboreal growth;

Step 2, selectes and reads corresponding geometric data file according to described non-geometric parameter; Wherein geometric data file generates in the following way: based on measured data, one seeds functional architecture model is calibrated, obtain the functional architecture model of described seeds, then utilize described functional architecture model to generate the geometric data file of the different non-geometric parameter values of correspondence of described seeds;

Step 3, according to described geometric data file set up rudimentary model;

Step 4, the geometric parameter of input trees is adjusted described rudimentary model, obtains tree modelling;

Step 5, carries out quick Realistic Rendering to described tree modelling.

Preferably, in above-mentioned tree simulation method, described non-geometric parameter comprises: seeds, age, season and/or water stress degree;

Described geometric parameter comprises: the diameter of a cross-section of a tree trunk 1.3 meters above the ground, base footpath, the height of tree, clear bole height, Branch angle and/or flexibility.

Preferably, in above-mentioned tree simulation method, described tree modelling is expressed geometry information and the topology information of trees based on binary tree.

Preferably, in above-mentioned tree simulation method, described binary tree is using trees trunk as root node, side shoot on each trunk is according to being sorted to the order on top the end of by, take first side shoot as left subtree, take sibling as right subtree, wherein said root node does not have right subtree.

Preferably, in above-mentioned tree simulation method, before described step 5, also comprise:

Branch is sorted to trees biomass with to the contribution degree of visual effect according to branch, then reject the branch of sequence after threshold value to simplify described tree modelling.

Preferably, in above-mentioned tree simulation method, in described step 5, draw the translucent effect of blade by the method for class scattering; Draw the shade of trees by echo method.

Preferably, in above-mentioned tree simulation method, also comprise:

According to arboreal growth conditional curve, between the tree modelling of various discrete state time point, carry out interpolation calculation, thereby generate the continuous animation of simulation trees growth and development process.

Preferably, in above-mentioned tree simulation method, in described step 5, by adopting vertex cache, the disposable first step that submits the data to rendering pipeline, thus reduce data transmission, realize drawing and accelerate.

In order better to realize above-mentioned purpose, the present invention also provides a kind of tree simulation device based on Generalized parametering modeling, comprising:

Load module, for: the non-geometric parameter relevant to arboreal growth inputted;

Data acquisition module, for: according to described non-geometric parameter, select and read corresponding geometric data file; Wherein geometric data file generates in the following way: based on measured data, one seeds functional architecture model is calibrated, obtain the functional architecture model of described seeds, then utilize described functional architecture model to generate the geometric data file of the different non-geometric parameter values of correspondence of described seeds;

MBM, for: according to described geometric data file set up rudimentary model;

Adjusting module, for: the geometric parameter of input trees is adjusted described rudimentary model, obtains tree modelling;

Drafting module, for: described tree modelling is carried out to quick Realistic Rendering.

Preferably, in above-mentioned tree simulation device, also comprise:

Simplify module, for: sort to trees biomass with to the contribution degree of visual effect according to branch, then reject the branch of sequence after threshold value to simplify described tree modelling;

Effects module, for: the translucent effect of drawing blade by the method for class scattering; Draw the shade of trees by echo method;

Interpolating module, for: according to arboreal growth conditional curve, between the tree modelling of various discrete state time point, carry out interpolation calculation, thereby generate the continuous animation of simulation trees growth and development process;

Accelerating module, for: by adopting vertex cache, the disposable first step of data being submitted to rendering pipeline, thus reduce data transmission, realize drawing and accelerate.

Compared with prior art, the embodiment of the present invention has following effect:

1) embodiment of the present invention proposes and has realized the method for Generalized parametering trees modelings, seeds, age, season, breeding condition etc. are set up to tree modelling as parameter, and can pass through the parameter adjustment tree modelling such as the diameter of a cross-section of a tree trunk 1.3 meters above the ground, base footpath, the height of tree, clear bole height, Branch angle and flexibility.This modeling method combines the geometry appearance of trees and its biological characteristics, is conducive to generate the geometric model of reflection arboreal growth rule, can the smooth and easy trees growth course of naturally simulating.

2) embodiment of the present invention proposes that trees topological structure based on binary tree is expressed and the method for trees geometric data storage, binary tree structure defines according to tree construction in similar nature, be important Information Organization form, be beneficial to trees modeling and growth simulation.The binary tree structure of embodiment of the present invention definition can clearly be expressed the topological structure relation of trees, and can clearly express the inheritance of branches at different levels and the corresponding relation of each branch all ages and classes.The tree structure based on binary tree of embodiment of the present invention definition is expressed and can be carried out tree structure simplification to trees biomass with to the contribution degree of visual effect according to branch.

3) embodiment of the present invention is expressed according to arboreal growth rule and the tree structure based on binary tree, set up the growth equation that is applicable to simulate arboreal growth, and the growth courses of trees is changed to the variation that is presented as branch length, radius, position, leaf position and quantity, to these variations interpolation calculation of growing, realize the level and smooth growth transition of trees.

4) in addition the embodiment of the present invention has realized the drafting of trees shades based on Shadow map (echo) technology, and adopts the method for approximate scattering to realize the translucent effect of leaf, has improved the sense of reality of tree rendering.In conjunction with the programmable features of graphic hardware, the drafting of branch and blade and growth course interpolation are all carried out in Programmable GPU, and adopt the automatic assembling function of Geometry Shader to generate the triangular mesh of trees, and improve render speed, accelerate the demonstration of model.

Accompanying drawing explanation

Fig. 1 is the modeling process chart of GreenLab functional mode;

Fig. 2 is the flow chart of steps of the method that provides of the embodiment of the present invention;

Fig. 3 is the definition schematic diagram of the geometric parameter that provides of the embodiment of the present invention;

Fig. 4 is the schematic diagram of the branch translation that provides of the embodiment of the present invention;

Fig. 5 is that the branch that the embodiment of the present invention provides branch angle variation schematic diagram;

Fig. 6 is that the branch bending degree that the embodiment of the present invention provides changes schematic diagram;

Fig. 7 is the trees ternary tree structure representation schematic diagram that the embodiment of the present invention provides;

Fig. 8 is that the trees binary tree structure that the embodiment of the present invention provides is expressed schematic diagram;

Fig. 9 is the arboreal growth conditional curve that the embodiment of the present invention provides;

The structural drawing of the tree simulation device that Figure 10 provides for the embodiment of the present invention.

Embodiment

Below in conjunction with the drawings and specific embodiments, the invention will be further described.

Fig. 2 is the flow chart of steps of the method that provides of the embodiment of the present invention, and as shown in Figure 2, the tree simulation method based on Generalized parametering modeling comprises:

Step 101, inputs the non-geometric parameter relevant to arboreal growth;

Step 102, according to described non-geometric parameter, selectes and reads corresponding geometric data file; Wherein geometric data file generates in the following way: based on measured data, one seeds functional architecture model is calibrated, obtain the functional architecture model of described seeds, then utilize described functional architecture model to generate the geometric data file of the different non-geometric parameter values of correspondence of described seeds;

Step 103, according to described geometric data file set up rudimentary model;

Step 104, the geometric parameter of input trees is adjusted described rudimentary model, obtains tree modelling;

Step 105, carries out quick Realistic Rendering to described tree modelling.

Wherein, described non-geometric parameter comprises: seeds, age, season and/or water stress degree; Described geometric parameter comprises: the diameter of a cross-section of a tree trunk 1.3 meters above the ground, base footpath, the height of tree, clear bole height, Branch angle and/or flexibility.

Visible, the embodiment of the present invention is set up rudimentary model according to non-geometric parameter and the structure function model (GreenLab) relevant to arboreal growth, and can obtain tree modelling by geometric parameter adjustment.Therefore, the embodiment of the present invention combines the geometry appearance of trees and its biological characteristics (non-geometric parameter), is conducive to generate the geometric model of reflection arboreal growth rule, can the smooth and easy trees growth course of naturally simulating.

Wherein, described tree modelling is expressed trees topology information and stores trees geometry information based on binary tree.Described binary tree is using trees trunk as root node, and the side shoot on each trunk is according to being sorted to the order on top the end of by, and take first side shoot as left subtree, take sibling as right subtree, wherein said root node does not have right subtree.

Binary tree structure defines according to tree construction in similar nature, is important Information Organization form, is beneficial to trees modeling and growth simulation.Therefore, the binary tree structure of embodiment of the present invention definition can clearly be expressed the topological structure relation of trees, and can clearly express the inheritance of branches at different levels and the corresponding relation of each branch all ages and classes, can also carry out tree structure simplification to trees biomass with to the contribution degree of visual effect according to branch.

Before described step 105, also comprise: sort to trees biomass with to the contribution degree of visual effect according to branch, then reject the branch of sequence after threshold value to simplify described tree modelling.

In the embodiment of the present invention, draw and adopt following technology: in described step 105, draw the translucent effect of blade by the method for class scattering; Draw the shade of trees by echo method; According to arboreal growth conditional curve, between the tree modelling of various discrete state time point, carry out interpolation calculation, thereby generate the continuous animation of simulation trees growth and development process; By adopting vertex cache, the disposable first step that submits the data to rendering pipeline, thus reduce data transmission, realize drawing and accelerate.

Below, Generalized parametering modeling, binary tree, GPU drafting etc. are carried out to careful explanation.

(1) Generalized parametering modeling

Parametric modeling concept, the earliest for industrial design, refers to that the planform of design object is relatively shaped, and can arrange size relationship by one group of parameter, by the modification of varying sized driving design result.The parametric modeling of plant is to keep, under the substantially constant prerequisite of plant topological law, changing size, form and the angle of Plant geometrical structure, completing the modeling method of the structure to plant display model with limited geometric graphic element.

The major parameter of the Generalized parametering modeling that the embodiment of the present invention proposes not only comprises the geometric parameters such as the diameter of a cross-section of a tree trunk 1.3 meters above the ground, base footpath, the height of tree, clear bole height, Branch angle and flexibility, also comprises the non-geometric parameters such as age, season, moisture and seeds.First call different geometric data files by non-geometric parameter, set up corresponding tree modelling, then can be according to the redjustment and modification tree modelling of geometric parameter.

The embodiment of the present invention has been carried out experiments of measuring to willow 107 and capital two willow seeds of 2 poplar, and measured data based on a large amount of calibrates the GreenLab model of these two seeds, has obtained its functional architecture model parameter.As previously mentioned, QINGYUAN software calculates according to the functional architecture model of these two seeds, exports its geometric data file.Like this, the system that the embodiment of the present invention realizes, can process willow 107 and capital two seeds of 2 poplar, and " seeds " parameter value scope is this two-value.In addition, " age " and " season " is time scale, calls different geometric data files according to the variation of Seasons and ages, sets up corresponding tree modelling; " season " parameter is used for controlling trees outward appearance Four seasons change, has spring, summer, autumn, winters four value; " moisture condition " parameter is for controlling the breeding conditions of trees, value have slightly coerce, moderate coerces and three kinds of normal water supplies.

It is more than the non-geometric parameter of tree modelling, and the definition of its geometric parameter as shown in Figure 3: the diameter of a cross-section of a tree trunk 1.3 meters above the ground is the trees trunk diameter of trees 1.3m eminence, base footpath is the diameter of trees trunk bottom, the height of tree is trunk total length, clear bole height be on trunk first side shoot apart from the distance of trunk bottom, Branch angle is side shoot and trunk angle, and flexibility is the angle of each internode of side shoot.The Branch angle of mentioning in the embodiment of the present invention and flexibility are Branch angle and the flexibility of secondary side shoot.

The embodiment of the present invention is adjusted tree modelling by above-mentioned geometric parameter, changes the mode of appearance of trees.Thereby drive the adjustment of trees mode of appearance by revising each geometric parameter, make tree modelling more true to nature.Geometric parameter variation is to the adjustment of tree modelling only for trunk and secondary branch, and the evolution of other branches or blade is according to the conversion of his father's branch.

Embodiment of the present invention trees trunk is made up of multiple internodes, on trunk, the length of each internode is identical, in the time revising the height of tree, in order to keep the constant of trees entirety relative position, revise the length of internode according to the variation of the height of tree, thereby reach the effect of revising the height of tree.

Clear bole height is that on trunk, first side shoot is apart from the distance of trunk base portion, and the clear bole height that the embodiment of the present invention is mentioned recently replaces with the percentage of actual clear bole height and the height of tree.When revising tree branches when high, first branch carries out translation according to specified value, but may occur as several situations in Fig. 4 when translation: the first kind normal condition, first branch does not reach the position of second branch through translation.The second situation, first branch is after translation, and second branch becomes clear bole height apart from trunk distance from bottom, and now the value of clear bole height changes.The third situation, first branch, after translation, separates with the branch of same type, is referred in another bunch of branch, has changed the one-piece construction of trees.

And in fact tree branches spacing has certain rule, concerning willow, the raw branch of same year is concentrated in a certain region, is called bunch.Therefore in the time revising clear bole height, can consider all branches of first bunch of translation, rather than first branch.When translation branch, according to sub-side shoot and all blades of the binary tree search to the first of definition bunch all branches and all branches, carry out overall translation according to the clear bole height of revising.

Tree branches is too many, and Branch angle differs, and for the ease of calculating, all secondary branches of the embodiment of the present invention adopt same Branch angle.In the time revising Branch angle, all secondary branches change thereupon.Branch angle is the angle between first internode of side shoot and trunk, and in the plane of trunk and current side shoot composition, current side shoot rotates Branch angle around trunk, is the final position of current side shoot.As shown in Figure 5, in the time that Branch angle changes, when the relative position between each internode on anterior branch constant, first internode position a of each branch ₀can not change, therefore work as the position a of other internodes on anterior branch _j' all obtain according to the position translation of first internode.

${a_j}^{\′}=a_0+\stackrel{\→}{a_0{a_j}^{\′}}$

Wherein

$\left\{\begin{array}{c}\left|\right|\stackrel{\→}{a_0{a_j}^{\′}}\left|\right|=\left|\right|\stackrel{\→}{a_0{a}_j}\left|\right|\\ \mathrm{dot}(\stackrel{\→}{a_{0}B},\stackrel{\→}{a_0{a_j}^{\′}})==\left|\right|\stackrel{\→}{a_{0}B}\left|\right|\×\left|\right|\stackrel{\→}{a_0{a_j}^{\′}}\left|\right|\×\cos ({\mathrm{\α}}^{\′}-\mathrm{\β})\end{array}\right.$

By above-mentioned, can obtain the translation vector when other internodes on anterior branch, then all subtree branches and all blades on this internode in binary tree traversal, carries out it

translation.

$\stackrel{\→}{a_j{a_j}^{\′}}={a_j}^{\′}-a_j=a_0+\stackrel{\→}{a_0{a_j}^{\′}}-a_j$

Branch flexibility refers to the skew number of degrees of the relative branch inceptive direction of branch end direction, and branch is made up of multiple internodes, and each internode has certain angle skew, finally form branch bending, therefore, revise branch bending degree the position of each internode can be changed, as shown in Figure 6.In the time that modification flexibility is β, first internode position a0 and the a1 of each branch remain unchanged, to the position a when other internodes on anterior branch _j' obtain according to the position of the first two internode.

${a_j}^{\′}={a_{j-1}}^{\′}+\stackrel{\→}{{a_{j-1}}^{\′}{a_j}^{\′}}$

Wherein

$\left\{\begin{array}{c}\left|\right|\stackrel{\→}{{a_{j-1}}^{\′}{a_j}^{\′}}\left|\right|=\left|\right|\stackrel{\→}{a_{j-1}{a}_j}\left|\right|\\ \mathrm{dot}(\stackrel{\→}{{a_{j-2}}^{\′}{a_{j-1}}^{\′}},\stackrel{\→}{{a_{j-1}}^{\′}{a_j}^{\′}})==\left|\right|\stackrel{\→}{{a_{j-2}}^{\′}{a_{j-1}}^{\′}}\left|\right|\×\left|\right|\stackrel{\→}{{a_{j-1}}^{\′}{a_j}^{\′}}\left|\right|\×\cos \left(\mathrm{\β}\right)\end{array}\right.$

By above-mentioned formula, can obtain the translation vector when other any internodes on anterior branch, then all subtree branches and all blades on this internode in binary tree traversal, carries out it

translation.

$\stackrel{\→}{a_j{a_j}^{\′}}={a_j}^{\′}-a_j={a_{j-1}}^{\′}+\stackrel{\→}{{a_{j-1}}^{\′}{a_j}^{\′}}-a_j$

(2) tree structure based on binary tree is expressed

One of core work of the Generalized parametering modeling that the embodiment of the present invention proposes is the data file in conjunction with GreenLab model, active data expression is proposed, geometry information and the topology information of expressing rationally and effectively GreenLab tree modelling.The embodiment of the present invention is by the document format data of labor GreenLab model, extract useful trees data, set up the structure representation of trees based on binary tree, geometry information and the topology information of storage trees, and carried out tree structure simplification based on binary tree.

1) express towards the tree structure of GreenLab

Tree structure complexity, branch is various, and geometric data amount is large, makes the storage of trees geometric data and the storage of topology information have larger difficulty, therefore adopts rational structure to express trees most important.The embodiment of the present invention is that the geometric data based on GreenLab model is exported the tree structure simulation of carrying out, existing multidigit researchist has carried out plant structure analog operation based on GreenLab functional architecture model above, but major part still concentrates on herbaceous plant, it is simple that herbaceous plant is compared tree structure, data volume is few, can directly adopt how much method for drafting to simulate, but tree structure complexity, geometric data amount is large, must adopt effective mode that these Organization of Datas are got up, effectively express the topological structure of trees and clearly express trees geometry, thank to luxuriant gold according to the principle of trees axle construction for this reason, ternary tree structure representation method has been proposed.The ternary tree of its definition is take main shaft base portion as root node, each node comprises three orderly and mutually disjoint ternary trees, wherein left subtree represents this axle first side shoot from the top to bottom, middle subtree is its brother tree, right subtree is next axle, and wherein other nodes of each subtree are equally according to sequential expression as above.As shown in Figure 7, left is the tree structure with the digital simplification that each entity is numbered, and the right side is the ternary tree structure representation of this tree.

Ternary tree can clearly be expressed the topological structure of plant naturally, embodies the axle construction feature of plant, and each node of ternary tree geological information and biological attribute information of having comprised entity, can clearly express the geometry information of trees.But tree structure complexity, its entity number is millions, the ternary tree complex structure of trees, its traversal and search all more difficult, therefore the embodiment of the present invention has proposed a kind of new expression, carry out expression tree timber structure based on binary tree, and each node canned data is improved, more simple and clear compared with ternary tree.

2) tree structure based on binary tree is expressed

Concerning complicated trees, the data volume of trident storage of data structure is still larger, as trunk originally can be expressed by data, and need to being divided into multiple axles, the expression of ternary tree based on axle construction express, data volume is larger, to trident traversal of tree with search more time-consumingly, greatly reduce drafting efficiency.The embodiment of the present invention is used the thinking of ternary tree, has proposed the tree structure expression based on binary tree, and the binary tree of definition is take axle construction as foundation, and take branch as base unit.Trees hierarchical structure is obvious, and has stronger self-similarity, and trees are generally divided into trunk and side shoot, and each side shoot is a similar stalk tree again, is still divided into trunk and side shoot.The binary tree of embodiment of the present invention definition, take branch as base unit, calls side shoot the child of trunk, and the side shoot on trunk is called brother between mutually, and the left node of binary tree is stored its child's information, and right node is stored its fraternal information.

As shown in Figure 8, left side is the trees geometry of simplifying, and for convenience of expressing, with digital, each side shoot is numbered.The binary tree semantical definition of expressing for tree structure is as follows: binary tree is using trees trunk as root node, side shoot on each trunk is according to being sorted to the order on top the end of by, take first side shoot as left subtree, take its sibling as right subtree (root node does not have right subtree), to each side shoot equally according to above definition.Fig. 8 right side is the binary tree abstract expression of tree structure as defined above.

By as above defining, from the some nodes of binary tree, can access as follows:

(step 1) first accesses its left node, the right subtree of its left node of depth-first traversal, until right subtree is empty, this left node and the node sequence of accessing form all side shoots of this node representative trunk, and sequence order is the sequence of positions of side shoot on trunk.As from { 5}, { 6}, depth-first access right subtree, until { 7}, left node { 6} and node sequence { 7} formation branch { all side shoots of 5} of access first to access its left node.

(step 2) is if this node side shoot rank is n, if n=1 represents that this node is root node, is n+1 by the side shoot rank of all nodes of access in (step 1).If n unequal to 1, its father node is accessed in depth-first, until the side shoot rank of node is n-1, is n by the side shoot rank of all nodes of access in (step 1), access the side shoot node of these nodes, the node sequence obtaining has represented all side shoots that side shoot rank is n+1.As from { 5}, its level=2, accesses its father node, until { its rank of 1} is 1, and { 2}, { 5}, { 8}, { 9}, { 10}, { the side shoot rank of 11} is 2 for all nodes of traveling through according to (step 1) mode.Then { 3}, { 4}, { 6}, { it is all side shoots of 3 that 7} is side shoot rank to access these ranks and be the side shoot node of all nodes of 2.

By above-mentioned traversal mode, can in fast access tree structure, specify side shoot, the binary tree of embodiment of the present invention definition can clear and definite be expressed the topological structure relation of trees equally as can be seen here, and data are expressed and had greatly minimizing than ternary tree.In the time that needs travel through or search for trees data, reduce a lot of workloads, greatly reduce the time.

3) tree structure based on binary tree is simplified

Tree structure complexity, along with trees age growth, the branch of trees and blade are more and more, the data volume of tree modelling can be increasing, and the drafting efficiency of trees can reduce greatly, and the certain applications of tree modelling often visually to trees details require be not very high, if each details of trees is all exported to drafting, can have a strong impact on drafting efficiency, therefore be necessary it to simplify to reduce complexity, improve and draw efficiency.

For geometric space information and the biomass spatial framework of leave tree, tree modelling is simplified should be in reducing model complexity as far as possible, keeps former shape of tree-crown and biomass to distribute as far as possible.Consider the biomass assignment problem of trees, more low-level limb has been collected the main biomass of trees, its space distribution can represent roughly the Spatial distribution of biomass of whole trees, and more low-level limb formed the crown skeleton of trees, is supporting other organ of trees.In addition, from low-level to high-level, the quantity of long shoot is fewer and feweri.Obviously, branch length is less, and its contribution to visual effect is less, less to the contribution of Spatial distribution of biomass.Can reduce tree branches by consideration and simplify tree structure, the contribution according to branch to trees biomass and the contribution of visual effect are sorted to branch, then reject the smaller branch of contribution degree to reach the effect of simplification.

By above analysis, the importance of branch depends primarily on the length of the weight of branch (can be determined by branch rank roughly) and branch.Definition branch weight metric function as:

w＝w _level×w _length

Wherein w _lengthfor the length weight of branch, count ratio by internode number and the maximum internode of tree branches and replace.W _levelfor the weight of branch different stage, determine according to statistical study, first count the internode number of each branch of one tree, the average internode that then statistical is separated out each side shoot rank is counted N _level, look for a N _levelin maximal value N _max.:

$w_{\mathrm{level}}=\frac{N_{\mathrm{level}}}{N_{\max }}$

To side shoot descending sort, obtain a side shoot numbered sequence A according to w, can be then k threshold value of w appointment as required, thereby A is divided into k discrete level of detail.

In actual application, can select suitable level of detail model to draw from the distance of viewpoint according to render target point, thereby kept to a certain extent the equivalence on vision and biological significance in improving drafting efficiency.

4) growth course simulation

The trees data file of QINGYUAN software output is all ages and classes discrete state, can only embody the trees difference of each all ages and classes, can not embody the annual interior growth course of trees and change.And be the emphasis problem of tree simulation research to the growth course simulation of trees, the embodiment of the present invention has been carried out interpolation calculation according to arboreal growth conditional curve to trees various discrete state, thereby has simulated the growth and development process of trees.

The new life that the growth course major embodiment of trees is branch and coming off, the new life of branch length and change in radius and blade and coming off, arranges seamlessly transitting of branch according to growth rhythm like this, can simulate growing of trees.And all ages and classes trees that binary tree is expressed are independently, between identical branch all ages and classes state, there is no corresponding relation, carry out growth simulation and need to first set up the corresponding relation of each branch.Tree structure based on binary tree is expressed and can clearly be embodied the topology information of trees, is conducive to extraction and the arboreal growth simulation of branch corresponding relation.

5) growth interpolation

Higher and meet the animation process of plant growth rule in order to obtain temporal resolution, can be by a series of models of QINGYUAN output as key frame, time scale between each key frame is 1 year, in the middle of each key frame, insert 365 frames, according to the growth rhythm in branch 1 year, key frame is carried out to interpolation, generate 365 frames in the middle of this.Interpolation algorithm carries out according to formula below:

N＝k×S+(1-k)×E，k∈[0，1]

Wherein N is current state, and S is initial state, and E is final state, and k is interpolation ratio, and reality is and time correlation function.

The growth of trees is slowly-vigorous-slowly-and finally stop, therefore the curve of all age group increment change procedure is a growth equation that is serpentine curve, curve is as shown in Figure 9.The feature of wood growth equation is:

(1) y (t)=0 in the time of t=0.This condition is the starting condition that wood growth equation should be satisfied.

(2) there is an asymptote y (t)=A in y (t), and A is this arboreal growth maximum value.

(3) growth of trees is irreversible, and making y (t) is the monotone nondecreasing function about age (t).

(4) y (t) is the continuous and smooth function curve about t.

The variation of biomass is presented as the variation of the each branch of trees, mainly comprises that the quantity of internode increases, the size variation of radius and the length variations of internode.The length variation of the quantity of internode, the size of radius and internode all meets Logistics curve, in order to determine the parameter of Logistics curve, need to know foliation time, growth characteristics and fallen leaves time that concrete trees are relevant.1 year 365 days are normalized, and willow germinates generally at late March, and foliation phase t0 is made as to 0.25, and the phase that stops growing, the phase t1 that will stop growing was made as 0.78, then becomes the leaf fall period generally in the middle ten days and the last ten days in September.Carry out approximate simulation logistics curvilinear equation with sin function.

First calculate the newborn internode of each branch in 1 year by S initial state and E final state and count Num, the equation that the newborn internode number of branch is y (t) is as follows:

$y\left(t\right)=\left\{\begin{array}{c}0,t<t0\\ \mathrm{Num}\&times;\sin (\frac{\mathrm{\&pi;}}{2}\&times;\frac{t-t0}{t1-t0}),t0\&le;t\&le;\mathrm{<figure-callout\; id="1"\; label="t"\; filenames="HDA0000089929370000031.png,HDA0000089929370000051.png"\; state="\{\{state\}\}">t</figure-callout>}1\\ \mathrm{Num},t>\mathrm{<figure-callout\; id="1"\; label="t"\; filenames="HDA0000089929370000031.png,HDA0000089929370000051.png"\; state="\{\{state\}\}">t</figure-callout>}1\end{array}\right.$ Can draw the each branch newborn internode number in each period by formula, thereby draw the leaf unit sum of each branch in each in period.The quantity of same blade also can change, and blade starts flavescence in late September, at the bottom of October, slowly comes off, and establishes the blade time t2=0.82 that starts to come off, and blade all fell in mid-December, t3=0.95, and this branch blade adds up to N _leaf, the newborn internode number of branch is y _leaf(t) equation is as follows:

$y_{\mathrm{leaf}}=\left\{\begin{array}{c}0,t<t0\\ N_{\mathrm{leaf}}\&times;\sin (\frac{\mathrm{\&pi;}}{2}\&times;\frac{t-t0}{t1-t0}),t0\&le;t\&le;t1\\ N_{\mathrm{leaf}},\mathrm{<figure-callout\; id="1"\; label="t"\; filenames="HDA0000089929370000031.png,HDA0000089929370000051.png"\; state="\{\{state\}\}">t</figure-callout>}1<t<t2\\ N_{\mathrm{leaf}}\&times;\frac{t-t2}{t3-t2},t2\&le;t\&le;t3\\ 0,t\&GreaterEqual;t3\end{array}\right.$

Except the variation of branch length and blade quantity, on branch, the position of each internode and radius also can temporal evolutions, and the radius of establishing internode primary state is r0, and internode final state radius is r1, and the equation of internode radius r (t) is as follows:

$r\left(t\right)=\left\{\begin{array}{c}r0,t<t0\\ r0+(r1-r0)\&times;\sin (\frac{\mathrm{\&pi;}}{2}\&times;\frac{t-t0}{t1-t0}),t0\&le;t\&le;t1\\ \mathrm{<figure-callout\; id="1"\; label="r"\; filenames="HDA0000089929370000031.png,HDA0000089929370000051.png"\; state="\{\{state\}\}">r</figure-callout>}1,t>\mathrm{<figure-callout\; id="1"\; label="t"\; filenames="HDA0000089929370000031.png,HDA0000089929370000051.png"\; state="\{\{state\}\}">t</figure-callout>}1\end{array}\right.$

If the position of internode original state is p0 (x, y, z), the position of internode final state is p1 (x, y, z), and internode position p (x, y, z) equation (t) is as follows:

$p\left(t\right)=\left\{\begin{array}{c}p0,t<t0\\ p0+(p1-p0)\&times;\sin (\frac{\mathrm{\&pi;}}{2}\&times;\frac{t-t0}{t1-t0}),t0\&le;t\&le;t1\\ \mathrm{<figure-callout\; id="1"\; label="p"\; filenames="HDA0000089929370000031.png,HDA0000089929370000051.png"\; state="\{\{state\}\}">p</figure-callout>}1,t>\mathrm{<figure-callout\; id="1"\; label="t"\; filenames="HDA0000089929370000031.png,HDA0000089929370000051.png"\; state="\{\{state\}\}">t</figure-callout>}1\end{array}\right.$

By above-mentioned equation, can from two growth age key frames, interpolation obtain 365 key frames, thereby realize the growth course interpolation of trees.

(3) Realistic rendering of trees method

Trees are important composition factors of natural scene, and the existence of height sense of reality trees can significantly promote the fidelity of a scene, and the authenticity of trees visual effect has directly affected the Flow experience of virtual reality system.Trees visual details is various, uneven with crackle or knot bark surface, the translucent attribute that the old and new's leaf is different, the hatching effect of complexity between leaf and trunk, all non-easy things of sense of reality simulation of these details, it is the problem of rich challenge in computer graphics that trees true to nature represent always.Translucent effect and the trees hatching effect of the Realistic rendering of trees of the embodiment of the present invention to blade simulated, its Leaf translucent effect adopts the method simulation of approximate scattering, and the shade drafting of trees adopts current more conventional echo (shadow map) method.

The drafting of blade

Draw the trees with the height sense of reality, the leaf that drafting has the height sense of reality is very important.Complicacy in blade geometry, the crossover interaction of complicated effect of light, especially blade interior, makes blade present semi-transparency property, thereby makes efficient sense of reality trees real-time rendering very challenging.

The leaf of embodiment of the present invention research is simulated with a quad patch that has shone upon blade transparent texture figure, then tests the place of rejecting beyond blade by transparency.The size of quad patch is determined by var1 in LigOrgan and var2, its initial position is all in XOY plane, centered by true origin, then by the matrix matrix in leaf structure, current tetragonal position is rotated to translation, just blade can be plotted to corresponding correct position.

The drafting of shade

In actual life, shade is ubiquitous, and shade is the very important illumination visual signature of occurring in nature, and the drafting of shade not only can improve the sense of reality of scene, can also provide much important space geometry information for scene.The shade of trees can reflect spatial relation and the mutual hiding relation etc. of trees and surrounding environment or trees self, and the drafting of trees shade has larger effect to the raising of the trees sense of reality.

The shade computing method that the embodiment of the present invention adopts are echo (Shadow Map) methods more widely of applying at present, its main thought is: in the time that first pass is played up scene, the depth information of scene is kept in texture maps, this texture maps is called depth map, then in the time drawing scene second time, information depth in depth map is taken out, compare with the distance depth2 of current summit and light source, determine hiding relation, then change the color of object according to hiding relation, thus simulation hatching effect.Because echo Method And Principle is simple, be easy to realize and expansion, and the powerful processing power that can make full use of existing graphic hardware carries out real-time rendering efficiently, therefore become the shadow algorithm being most widely used.

(4) drawing process based on GPU accelerates

The arboreal growth of embodiment of the present invention research be discrete model using each age as key frame, then calculate the tree modelling of current time according to the Geometrical change between two frames and growth rhythm, thereby generate continuous animation.Arboreal growth process is presented as increase or the minimizing of each branch internode and blade, tree branches and leaf amount are many, data volume is large, if each moment needs re invocation branch data while repainting, large batch of like this data call, can increase the burden of CPU and GPU, draw efficiency and can greatly reduce.Data volume be need to reduce on the one hand so improve drafting efficiency, date storage method or rendering algorithm needed on the other hand to improve.

The tree rendering of embodiment of the present invention research, data volume is mainly reflected on branch, the trees of general life in 7 years, the number of blade is 7747, formed by 7747 faces, internode number is 14969, if each internode replaces with an octahedral cylinder, need 14969 × 8 faces, between the glvertex of opengl, draw quadrilateral if want and simulate branch, need to call glvertex () function 14969 × 8 × 4 times, if normal is set equally, texture coordinate, need to call equally glnormal () and gltexcoord2f () 14969 × 8 × 4 times respectively, this situation is worthless, call in a large number opengl function and can greatly reduce drafting efficiency, affect render speed.Embodiment of the present invention employing VBO (vertex cache Vertex Buffer Objects) once submits the data to the first step of rendering pipeline, has reduced the efficiency of data transmission.

The tree branches data preparation at n age is become to n VBO, according to the age of current selection, select current VBO once to import rendering pipeline into, then according to the time, branch position, length and radius etc. are carried out to interpolation, generate continuous process.

In addition, the embodiment of the present invention also provides a kind of tree simulation device based on Generalized parametering modeling, the structural drawing of the tree simulation device that Figure 10 provides for the embodiment of the present invention, and as shown in figure 10, it comprises:

Load module 201, for: the non-geometric parameter relevant to arboreal growth inputted;

Data acquisition module 202, for: according to described non-geometric parameter, select and read corresponding geometric data file; Wherein geometric data file generates in the following way: based on measured data, one seeds functional architecture model is calibrated, obtain the functional architecture model of described seeds, then utilize described functional architecture model to generate the geometric data file of the different non-geometric parameter values of correspondence of described seeds;

MBM 203, for: according to described geometric data file set up rudimentary model;

Adjusting module 204, for: the geometric parameter of input trees is adjusted described rudimentary model, obtains tree modelling;

Drafting module 205, for: described tree modelling is carried out to quick Realistic Rendering.

Wherein, can also comprise:

Simplify module, for: sort to trees biomass with to the contribution degree of visual effect according to branch, then reject the branch of sequence after threshold value to simplify described tree modelling;

Effects module, for: the translucent effect of drawing blade by the method for class scattering; Draw the shade of trees by echo method;

Interpolating module, for: according to arboreal growth conditional curve, between the tree modelling of various discrete state time point, carry out interpolation calculation, thereby generate the continuous animation of simulation trees growth and development process;

Accelerating module, for: by adopting vertex cache, the first step of the disposable rendering pipeline that submits the data to graphic process unit GPU, thus reduce data transmission, realize drawing and accelerate.

As from the foregoing, the embodiment of the present invention has following advantage:

1) embodiment of the present invention proposes and has realized the method for Generalized parametering trees modelings, seeds, age, season, breeding condition etc. are set up to tree modelling as parameter, and can pass through the parameter adjustment tree modelling such as the diameter of a cross-section of a tree trunk 1.3 meters above the ground, base footpath, the height of tree, clear bole height, Branch angle and flexibility.This modeling method combines the geometry appearance of trees and its biological characteristics, is conducive to generate the geometric model of reflection arboreal growth rule, can the smooth and easy trees growth course of naturally simulating.

2) embodiment of the present invention proposes that trees topological structure based on binary tree is expressed and the method for trees geometric data storage, binary tree structure defines according to tree construction in similar nature, be important Information Organization form, be beneficial to trees modeling and growth simulation.The binary tree structure of embodiment of the present invention definition can clearly be expressed the topological structure relation of trees, and can clearly express the inheritance of branches at different levels and the corresponding relation of each branch all ages and classes.The tree structure based on binary tree of embodiment of the present invention definition is expressed and can be carried out tree structure simplification to trees biomass with to the contribution degree of visual effect according to branch.

3) embodiment of the present invention is expressed according to arboreal growth rule and the tree structure based on binary tree, set up the growth equation that is applicable to simulate arboreal growth, and the growth courses of trees is changed to the variation that is presented as branch length, radius, position, leaf position and quantity, to these variations interpolation calculation of growing, realize the level and smooth growth transition of trees.

4) embodiment of the present invention has realized the drafting of trees shades based on Shadow map technology in addition, and adopts approximate scattering method to realize the translucent effect of leaf, has improved the sense of reality of tree rendering.In conjunction with the programmable features of graphic hardware, the drafting of branch and blade and growth course interpolation are all carried out in Programmable GPU, and adopt the automatic assembling function of Geometry Shader to generate the triangular mesh of trees, and improve render speed, accelerate the demonstration of model.

The above is only the preferred embodiment of the present invention; it should be pointed out that for those skilled in the art, under the premise without departing from the principles of the invention; can also make some improvements and modifications, these improvements and modifications also should be considered as protection scope of the present invention.

Claims (8)

1. the tree simulation method based on Generalized parametering modeling, is characterized in that, comprising:

Step 1, inputs the non-geometric parameter relevant to arboreal growth;

Step 2, selectes and reads corresponding geometric data file according to described non-geometric parameter; Wherein geometric data file generates in the following way: based on measured data, one seeds functional architecture model is calibrated, obtain the functional architecture model of described seeds, then utilize described functional architecture model to generate the geometric data file of the different non-geometric parameter values of correspondence of described seeds;

Step 3, according to the preliminary parameterized model of described geometric data file set up;

Step 4, the geometry appearance parameter of input trees is adjusted described preliminary parameterized model, obtains trees display model, and wherein, described geometry appearance parameter comprises: the diameter of a cross-section of a tree trunk 1.3 meters above the ground, base footpath, the height of tree, clear bole height, Branch angle and/or flexibility;

Step 5, carries out quick Realistic Rendering to the display model of expressing its structural information based on binary tree of described trees; Wherein, described binary tree is using trees trunk as root node, and the side shoot on each trunk is according to being sorted to the order on top the end of by, and take first side shoot as left subtree, take sibling as right subtree, wherein said root node does not have right subtree.

2. tree simulation method according to claim 1, is characterized in that,

Described non-geometric parameter comprises: seeds, age, season and/or water stress degree.

3. tree simulation method according to claim 1, is characterized in that, before described step 5, also comprises:

Branch is sorted to trees biomass with to the contribution degree of visual effect according to branch, then reject the branch of sequence after threshold value to simplify described trees display model.

4. tree simulation method according to claim 1, is characterized in that, in described step 5, draws the translucent effect of blade by the method for class scattering; Draw the shade of trees by echo method.

5. tree simulation method according to claim 4, is characterized in that, also comprises:

According to arboreal growth conditional curve, between the trees display model of various discrete state time point, carry out interpolation calculation, thereby generate the continuous animation of simulation trees growth and development process.

6. tree simulation method according to claim 5, is characterized in that, in described step 5, and by adopting vertex cache, the disposable first step that submits the data to rendering pipeline, thus reduce data transmission, realize drawing and accelerate.

7. the tree simulation device based on Generalized parametering modeling, is characterized in that, comprising:

Load module, for: the non-geometric parameter relevant to arboreal growth inputted;

Data acquisition module, for: according to described non-geometric parameter, select and read corresponding geometric data file; Wherein geometric data file generates in the following way: based on measured data, one seeds functional architecture model is calibrated, obtain the functional architecture model of described seeds, then utilize described functional architecture model to generate the geometric data file of the different non-geometric parameter values of correspondence of described seeds;

MBM, for: according to the preliminary parameterized model of described geometric data file set up;

Adjusting module, for: the geometry appearance parameter of input trees is adjusted described preliminary parameterized model, obtains trees display model, and wherein, described geometry appearance parameter comprises: the diameter of a cross-section of a tree trunk 1.3 meters above the ground, base footpath, the height of tree, clear bole height, Branch angle and/or flexibility;

Drafting module, be used for: the display model of expressing its structural information based on binary tree to described trees is carried out quick Realistic Rendering, wherein, described binary tree is using trees trunk as root node, side shoot on each trunk is according to being sorted to the order on top the end of by, take first side shoot as left subtree, take sibling as right subtree, wherein said root node does not have right subtree.

8. tree simulation device according to claim 7, is characterized in that, also comprises:

Simplify module, for: sort to trees biomass with to the contribution degree of visual effect according to branch, then reject the branch of sequence after threshold value to simplify described trees display model;

Effects module, for: the translucent effect of drawing blade by the method for class scattering; Draw the shade of trees by echo method;

Interpolating module, for: according to arboreal growth conditional curve, between the trees display model of various discrete state time point, carry out interpolation calculation, thereby generate the continuous animation of simulation trees growth and development process;

Accelerating module, for: by adopting vertex cache, the disposable first step of data being submitted to rendering pipeline, thus reduce data transmission, realize drawing and accelerate.

Images