CN101661630A - Method for measuring significance of surface of three-dimensional mesh model - Google Patents

Abstract

The invention discloses a method for measuring the significance of a surface of a three-dimensional mesh model, which comprises the following steps: calculating the normal vector of each surface sheetof the surface of the three-dimensional mesh model; calculating the normal vector of each vertex; searching the R ring neighbor of the vertex v according to the set parameter R; calculating each vertex v' in a set N to obtain a set S of relative local height between the vertex v and the vertex v'; and traversing the elements in the set S, and calculating the significance of the vertex v. Becausethe significance of the invention is defined in a larger region, the robustness is better as compared with the curvature, and the size of the region is determined by the parameter R, wherein the larger the R value is, the larger the considered region is, and the better the robustness to the noise is.

Description

The measure of a kind of three-dimensional grid model surface conspicuousness

Technical field

The present invention relates to a kind of grid model character representation method, specifically relate to the measure of a kind of three-dimensional grid model surface conspicuousness.

Background technology

Along with the development of computer software and hardware, three-dimensional grid model has obtained using widely in a plurality of fields, has become the 4th kind of multi-medium data after image, sound and video.Therefore, the analyzing and processing to the three-dimensional grid model data becomes one of hot subject of Recent study field and industry member.In traditional image analysis processing technology, the marking area that detects in the image is an important techniques, but in the analyzing and processing technology of three-dimensional grid model, similarly the conspicuousness detection technique also comparatively lacks at present, and main conspicuousness measure is based on curvature.But the conspicuousness detection technique that is based on curvature exists two following shortcomings: at first, curvature itself is to be defined on the continuous smooth curved surface, but three-dimensional grid model just is similar to a kind of of smooth surface, and its error is often bigger in actual applications; Secondly, curvature is the result who obtains under a kind of limiting case, therefore, in order to calculate the curvature of certain point of three-dimensional model surface, often just considers geometric properties in minimum neighborhood of this point, and this result who causes is that curvature is very responsive to noise.Therefore, propose a kind ofly new can portray three-dimensional grid model surface conspicuousness and the measure that noise has a robustness is very important.

Summary of the invention

Goal of the invention: the objective of the invention is in order to solve the deficiencies in the prior art, the measure of a kind of three-dimensional grid model surface conspicuousness is provided.

Technical scheme: in order to realize above purpose, the measure of a kind of three-dimensional grid model of the present invention surface conspicuousness, it is as follows that this method comprises step:

(1) calculates the normal vector of each dough sheet f ∈ F of three-dimensional grid model surface;

(2) calculate the normal vector of each vertex v ∈ V of three-dimensional grid model surface;

(3) according to the parameters R that is provided with, the R ring neighbours of search vertex v obtain R ring neighbours and represent to gather N; Wherein parameters R is set arbitrarily.

(4) each vertex v among the pair set N ', by calculating, obtain the S set of the relative local height between vertex v and the v ';

(5) element among the pair set S travels through, and obtains among the S element number n greater than 0 ₁With their summation s ₁, element number n less than 0 ₂With their summation s ₂, utilize following formula to calculate the conspicuousness of vertex v _LR (v)

$L_R\left(v\right)=\frac{-8}{C_R{\left(v\right)}^2}\&CenterDot;(\frac{s_1}{{\mathrm{<figure-callout\; id="1"\; label="n"\; filenames="A2009100353330009H1.png"\; state="\{\{state\}\}">n</figure-callout>}}_1}+\frac{s_2}{n_2})$

C wherein _R(v) ²The area sum of the dough sheet of the relevant relation in summit among expression and the set N.Suppose that R-neighbours' set that vertex v constitutes is N _v, being F (, claiming that then a is relevant with dough sheet f) with the related dough sheet set in the summit in this set if certain summit a is one of summit that constitutes certain dough sheet f, the area sum of all dough sheets is exactly C among the set of computations F _R(v) ²

The computing formula of dough sheet normal vector is in the step (1):

${\stackrel{\&RightArrow;}{n}}_f=\left|\begin{array}{ccc}\stackrel{\&RightArrow;}{i}& \stackrel{\&RightArrow;}{j}& \stackrel{\&RightArrow;}{k}\\ a_x& a_y& a_z\\ b_x& b_y& b_z\end{array}\right|$

Wherein vectorial (a _x, a _y, a _z) and vector (b _x, b _y, b _z) representing two uneven vectors on the dough sheet f respectively, i, j, k are the expression symbols of coordinate vector.

The normal vector on summit represents that by the area weighted sum of the normal vector of its adjacent dough sheet computing formula is as follows in the step (2):

${\stackrel{\&RightArrow;}{n}}_v=\frac{1}{\underset{f\&Element;F\left(v\right)}{\mathrm{\&Sigma;}}S\left(f\right)}\underset{f\&Element;F\left(v\right)}{\mathrm{\&Sigma;}}{\stackrel{\&RightArrow;}{n}}_f\&CenterDot;S\left(f\right)$

In the step (3) R ring neighbours be meant by breadth First algorithm to the neighbours around the vertex v search for obtain be positioned at R layer and R with the set on summit; This R ring neighbor table is shown N={v ₁, v ₂..., v _t, wherein t is total number on summit in the set.

The computing formula of the relative local height in the step (4) between vertex v and the v ' is:

${\mathrm{RH}}_v\left(v^{\&prime;}\right)=\cos \mathrm{\&alpha;}\&CenterDot;\left|\stackrel{\&RightArrow;}{v{v}^{\&prime;}}\right|=\frac{{\stackrel{\&RightArrow;}{n}}_v\&CenterDot;\stackrel{\&RightArrow;}{v{v}^{\&prime;}}}{\left|{\stackrel{\&RightArrow;}{n}}_v\right|\&CenterDot;\left|\stackrel{\&RightArrow;}{v{v}^{\&prime;}}\right|}\&CenterDot;\left|\stackrel{\&RightArrow;}{v{v}^{\&prime;}}\right|={\stackrel{\&RightArrow;}{n}}_v\&CenterDot;\stackrel{\&RightArrow;}{v{v}^{\&prime;}},$

α representation vector wherein

With vector

Between angle.

Through calculating, all to a real number value should be arranged, the set that constitutes with the corresponding real number value of element among the set N is S={r corresponding to each summit among the set N ₁, r ₂..., r _t, r wherein _iExpression v and the v that gathers among the N _iThe local height value of using above-mentioned formula to obtain.

Beneficial effect: the measure of a kind of three-dimensional grid model provided by the invention surface conspicuousness, compared with prior art have the following advantages: because conspicuousness of the present invention is to be defined in a bigger zone, therefore compared better robustness with curvature, the size in this zone is determined by parameters R, the R value is big more, the zone of considering is big more, and is also good more to the robustness of noise.

Description of drawings

Fig. 1 is one-piece construction figure of the present invention;

Fig. 2 is a particular flow sheet of the present invention.

Embodiment:

Below in conjunction with the drawings and specific embodiments, further illustrate the present invention, should understand these embodiment only is used to the present invention is described and is not used in and limit the scope of the invention, after having read the present invention, those skilled in the art all fall within the application's claims institute restricted portion to the modification of the various equivalent form of values of the present invention.

Fig. 1 has provided one-piece construction figure of the present invention.The computing method of whole conspicuousness comprise the two large divisions: precomputation part and conspicuousness calculating section.A back part is a main part of the present invention.A preceding part is the precomputation part, mainly is to finish two work, promptly calculates the normal vector of each dough sheet of three-dimensional model surface and the normal vector on each summit.The normal vector that wherein calculates dough sheet uses following formula:

${\stackrel{\&RightArrow;}{n}}_f=\left|\begin{array}{ccc}\stackrel{\&RightArrow;}{i}& \stackrel{\&RightArrow;}{j}& \stackrel{\&RightArrow;}{k}\\ a_x& a_y& a_z\\ b_x& b_y& b_z\end{array}\right|,$

Wherein, vector (a _x, a _y, a _z) and vector (b _x, b _y, b _z) represent two uneven vectors on the dough sheet f respectively.

The normal vector on summit is then determined by the area weighted sum of the normal vector of the associated dough sheet in this summit, shown in following formula:

${\stackrel{\&RightArrow;}{n}}_v=\frac{1}{\underset{f\&Element;F\left(v\right)}{\mathrm{\&Sigma;}}S\left(f\right)}\underset{f\&Element;F\left(v\right)}{\mathrm{\&Sigma;}}{\stackrel{\&RightArrow;}{n}}_f\&CenterDot;S\left(f\right)$

In this formula, and F (v) be the dough sheet adjacent with vertex v, the area of S (f) expression dough sheet f.

Fig. 2 has provided the present invention after precomputation, calculates the process flow diagram of the conspicuousness of certain vertex v of three-dimensional model surface, and concrete steps are as follows: step 11 is according to set parameters R, and the R ring neighbours of search vertex v gather N={v ₁, v ₂..., v _t.Step 12 is provided with four variable s ₁, n ₁, s ₂, n ₂, initial value all is made as 0, these variablees be respectively applied for the storage two types of relative local height and and number.Step 13 is provided with the element that counter i is used for pair set N and travels through, and initial value is set to 1.Whether step 14 judges i smaller or equal to the element number of gathering N, if, execution in step 15, otherwise execution in step 16.Step 15 is calculated v ' and the relative height h that gathers among the N between i the element v ', and formula is as follows:

${h=\mathrm{RH}}_v\left(v^{\&prime;}\right)=\cos \mathrm{\&alpha;}\&CenterDot;\left|\stackrel{\&RightArrow;}{v{v}^{\&prime;}}\right|=\frac{{\stackrel{\&RightArrow;}{n}}_v\&CenterDot;\stackrel{\&RightArrow;}{v{v}^{\&prime;}}}{\left|{\stackrel{\&RightArrow;}{n}}_v\right|\&CenterDot;\left|\stackrel{\&RightArrow;}{v{v}^{\&prime;}}\right|}\&CenterDot;\left|\stackrel{\&RightArrow;}{v{v}^{\&prime;}}\right|={\stackrel{\&RightArrow;}{n}}_v\&CenterDot;\stackrel{\&RightArrow;}{v{v}^{\&prime;}}.$

Step 17 is selected next step according to the value of h, when h execution in step 18 greater than 0 time, otherwise execution in step 19.Step 18 is added to s1 with the value of h, and n1 is increased 1 certainly, and step 19 then is added to h on the s2, and n2 is increased 1 certainly.No matter be that step 18 or 19 equal execution in step 20 increase 1 certainly with counter i afterwards.Step 20 jumps to step 14 then.Formula below step 16 is utilized calculates the conspicuousness of last vertex v, thereby finishes whole computation process:

Claims (5)

1, the measure of the surperficial conspicuousness of a kind of three-dimensional grid model, it is characterized in that: this method comprises following steps:

(1) calculates the normal vector of each dough sheet f ∈ F of three-dimensional grid model surface;

(2) calculate the normal vector of each vertex v ∈ V of three-dimensional grid model surface;

(3) according to the parameters R that is provided with, the R ring neighbours of search vertex v obtain R ring neighbours and represent to gather N;

(4) each vertex v among the pair set N ', by calculating, obtain the S set of the relative local height between vertex v and the v ';

(5) element among the pair set S travels through, and obtains among the S element number n greater than 0 ₁With their summation s ₁, element number n less than 0 ₂With their summation s ₂, utilize following formula to calculate the conspicuousness L of vertex v _R(v)

$L_R\left(v\right)=\frac{-8}{C_R{\left(v\right)}^2}\&CenterDot;(\frac{s_1}{n_1}+\frac{s_2}{n_2})$

C wherein _R(v) ²The area sum of the dough sheet of the relevant relation in summit among expression and the set N.

2, the measure of the surperficial conspicuousness of a kind of three-dimensional grid model according to claim 1, it is characterized in that: the computing formula of dough sheet normal vector is in the step (1):

${\stackrel{\&RightArrow;}{n}}_f=\left|\begin{array}{ccc}\stackrel{\&RightArrow;}{i}& \stackrel{\&RightArrow;}{j}& \stackrel{\&RightArrow;}{k}\\ a_x& a_y& a_z\\ b_x& b_y& b_z\end{array}\right|$

Wherein vectorial (a _x, a _y, a _z) and vector (b _x, b _y, b _z) represent two uneven vectors on the dough sheet f respectively.

3, the measure of the surperficial conspicuousness of a kind of three-dimensional grid model according to claim 1, it is characterized in that: the normal vector on summit represents that by the area weighted sum of the normal vector of its adjacent dough sheet computing formula is as follows in the step (2):

${\stackrel{\&RightArrow;}{n}}_v=\frac{1}{\underset{f\&Element;F\left(v\right)}{\mathrm{\&Sigma;}}S\left(f\right)}\underset{f\&Element;F\left(v\right)}{\mathrm{\&Sigma;}}{\stackrel{\&RightArrow;}{n}}_f\&CenterDot;S\left(f\right).$

4, the measure of a kind of three-dimensional grid model according to claim 1 surface conspicuousness is characterized in that: in the step (3) R ring neighbours be meant by breadth First algorithm to the neighbours around the vertex v search for the resulting R of being positioned at layer and R with the set on summit; This R ring neighbor table is shown N={v ₁, v ₂..., v _t, wherein t is total number on summit in the set.

5, the measure of the surperficial conspicuousness of a kind of three-dimensional grid model according to claim 1, it is characterized in that: the expression mode of local height value set S is S={r in the step (4) ₁, r ₂..., r _t; R wherein _iRelative local height between expression vertex v and the v ', its computing formula is:

${\mathrm{RH}}_v\left(v^{\&prime;}\right)=\cos \mathrm{\&alpha;}\&CenterDot;\left|{\stackrel{\&RightArrow;}{\mathrm{vv}}}^{\&prime;}\right|=\frac{{\stackrel{\&RightArrow;}{n}}_v\&CenterDot;{\stackrel{\&RightArrow;}{\mathrm{vv}}}^{\&prime;}}{\left|{\stackrel{\&RightArrow;}{n}}_v\right|\&CenterDot;\left|{\stackrel{\&RightArrow;}{\mathrm{vv}}}^{\&prime;}\right|}\&CenterDot;\left|{\stackrel{\&RightArrow;}{\mathrm{vv}}}^{\&prime;}\right|={\stackrel{\&RightArrow;}{n}}_v\&CenterDot;{\stackrel{\&RightArrow;}{\mathrm{vv}}}^{\&prime;},$

α representation vector wherein With vector

Between angle.

Images