CN101482401B - Image-based measuring method for three-dimensional object incontinuous hollow volume - Google Patents

Abstract

The invention discloses a three dimensional object discontinuous cavity volume determining method based on image, which comprises the following steps: loading the original image into the image processing apparatus and preprocessing the original image; determining boundary according to the gray scale level of the preprocessed image based on the original image; capturing image target area data and loading into memory, establishing a spatial octree in the memory, and traversing the octree nodes to determine the discontinuous cavity volume. The invention applies the octree in the perforated entity pore volume measuring in three dimensional visualization post processing in MicroCT system, not only can position the cavity accurately, but also can realize the accurate measuring to the perforated entity pore volume in three dimensional visualization post processing in MicroCT system.

Description

Assay method based on the three-dimensional object incontinuous hollow volume of image

Technical field

The present invention relates to the method for Measuring Object volume in a kind of three-dimensional image, related to the method that a kind of three-dimensional image that obtains by the MicroCT system is measured three-dimensional object incontinuous hollow volume more specifically.

Background technology

MicroCT is a new technique, MicroCT at first has irreplaceable effect in industrial nondestructive testing, secondly the basis that relies on the perfect mathematical model in its system, the effect that can predict is exactly to bring into play critical role in extremely small microworld, even can survey nano level microscopic pattern.These depend on the accuracy of front-end collection equipment, and original data processing becomes the accuracy and the high efficiency of the mathematical model of layer data.The two performance for system plays important effect, but the operating platform of the aftertreatment of faultage image also is very important.Operating platform has easily been arranged, can realize the detection data of a lot of actual needs,, can under harmless situation, survey the detecting of object such as the reconstruction of 3-D view; Any extraction of Different Plane can be observed the situation of each aspect very easily.

Because the special requirement of industrial circle some have occurred and has pressed for the problem of solution after the visual efficient of improving 3-D view and effect.The built-in attribute that needs experimental material in the process of now a lot of materials, what research was more in the prior art is the inside detection of porosint.Certainly detection directly demonstrates the 3 d effect graph of required detection object, but theoretical and actual measurement can't reach.Accurately calculate the such data or the more complicated in position of the volume and the aluminium bubble of the aluminium bubble in the 3D solid by data.Existing method mainly contains by the variation between layer and the layer judges the frequency of occurrences in the hole of all size, thus accurate volume calculation and position obtain all will be difficult to carry out and also relative error bigger.The present invention comes therefrom.

Summary of the invention

The invention provides a kind of assay method of the three-dimensional object incontinuous hollow volume based on image, the measurement that has solved the volume of the porosint in the 3D solid in prior art problems such as relatively difficulty, relative error be bigger.

In order to solve these problems of the prior art, the solution that the inventor provides is as follows:

A kind of assay method of the three-dimensional object incontinuous hollow volume based on image may further comprise the steps:

Be written into original image in the image processing apparatus and carry out the pre-service original image;

Determine the border according to pretreated image according to the gray level of original image; The capturing image target area data is written into internal memory, and structure space Octree is determined discontinuous cavity volume by the accumulation of traversal octree nodes in internal memory.

Preferably, described original image is the three-dimensional visible image that obtains by behind the MicroCT system 3-D scanning.

Preferably, described original image pre-service is through the linear conversion process of gray scale with original image.

Preferably, the linear conversion process mode of described gray scale is to obtain by the intensity profile Function Mapping, and described intensity profile function is

$f^{\′}(x,y)=\left\{\begin{array}{c}\frac{c^{\′}-a^{\′}}{c-a}(f(x,y)-a)+a^{\′}\\ \frac{d^{\′}-c^{\′}}{d-c}(f(x,y)-c)+c^{\′}\\ \frac{b^{\′}-d^{\′}}{b-d}(f(x,y)-d)+d^{\′}\end{array}\right.$

Wherein, a, c, d, b represent four separations of original image gray level, and a ', c ', d ', b ' represent the separation of the image gray levels after the mapping respectively; (x y) is the gray distribution of image function after the conversion to f '; F (x, y) be the intensity profile function of original image and a≤f (x, y)＜c, c≤f (x, y)＜d, d≤f (x, y)≤b.

Preferably, described image purpose area data is to capture the original image rectangular block earlier, obtains by the coordinate of original image rectangular block and the number of plies of original image.

Preferably, described space Octree is selected from regular Octree, linear Octree or a pair of eight formula Octrees.

Preferably, described image purpose area data is to manage distribution by the space Octree, described space Octree basic structure comprises the data address pointer, point to the pointer of father unit, point to the array of pointers of subelement, the array of pointers of described sensing subelement comprises the zone bit whether expression traveled through, the zone bit whether expression satisfies empty condition, is written into the width of image, be written into the height of image, be written into the number of plies of image; According to above-mentioned structure distribution Octree each unit in space is set.

Preferably, Octree traversal in described space comprises the steps:

Begin traversal from the 0th layer of Octree, the 0th layer of first junior unit begins to search backward; Judge this unit whether traveled through with this unit whether be empty unit fully;

When this unit does not travel through and be empty unit fully, trace back to the father unit; Judge whether the father unit is whether empty unit and subelement can divide;

When but the father unit is not empty unit and subelement timesharing, move back level, check eight secondary subelements of subelement to subelement; Three of periphery that check each sub-secondary units are with the straton unit; Judgement is with the traversal situation of straton unit and whether satisfy the condition in cavity;

But when with non-cavity, straton unit and this unit timesharing will with the adjacent secondary subelement in straton unit eight minutes, check the adjacent of four eight subdivisions with the straton unit; Judge this unit whether traveled through with this unit whether be empty unit fully;

When adjacent with the non-traversal in straton unit and be that empty unit is to check whether subelement can divide, so circulation; When six unit at the same level on every side that are recycled to the inseparable unit of afterbody all do not belong to empty unit or traveled through out-of-date end traversal.

Preferably, during described space Octree traversal, when being empty unit, the father unit of subelement continues to trace back.

Preferably, described space Octree when traversal, when the unit does not satisfy empty condition and unit when inseparable, stop this unit and expand towards periphery, check six unit at the same level on every side.

The space octree structure that designs in the technical scheme of the present invention is a kind of volume data structure, and the benefit of this structure is to connect the relation of the upper and lower.By the lowermost layer of an individual data items at first is set, for convenient counting of each limit of calculating is arranged to N=2 ^m+ 1, m is that the minimum number of pixels that every limit of volume data can be divided subtracts 1.The elementary cell that minimum limit is formed is the 0th layer, and obviously, the 0th layer has 2 ^3mIndividual unit.Each unit is endowed a flag.Form the 1st layer with 8 the 0th layer of adjacent unit then, obviously, the 1st layer has 2 ^3m/ 8 unit, the rest may be inferred makes up the data structure of a such Octree.Utilize Octree like this, by the 0th layer of continuous judgement symbol position of passing through from bottom, be filled into higher level easily, thereby neglect the useless point useless to volume drawing, when reducing light by volume data to the calculating of sampled point, and then save time effectively, thereby the speed of accelerating volume drawing becomes a reality the real-time interactive of volume drawing.

The present invention is owing to be disconnected between the inner most hole of porosint, this just makes the method that plane domain is increased can be generalized to three dimensions, find out very approximate object space border by the dependence between Octree layer and the layer like this, measure the pore volume of object.

Adopt the beneficial effect of technical solution of the present invention being applied in the measurement of the porose entity pore in the three dimensional visualization post processing in the MicroCT system by the structure Octree, the position in accurate localization cavity very not only, and can realize accurate measurement very easily to the empty volume of the porose entity in the three dimensional visualization post processing in the MicroCT system.This method is simple, effectively.The present invention be directed in the three-dimensional data a kind of method that effectively calculate size, the position of closed irregular hole, can be applied in accurately measuring of 3D solid cavity volume.

Description of drawings

Below in conjunction with drawings and Examples the present invention is further described:

Fig. 1 is the process flow diagram of the embodiment of the invention based on the assay method of the three-dimensional object incontinuous hollow volume of image;

Fig. 2 is the original image that adopts in the application examples of the present invention, does not carry out pre-service;

Fig. 3 is a pretreated image in the application examples of the present invention, carries out the linear transformation pre-service;

Fig. 4 is that original image carries out pretreated basic skills synoptic diagram in the application examples of the present invention;

Fig. 5 is that the space of structure Octree in the application examples of the present invention spreads all over logical diagram.

Embodiment:

For the technical scheme of more detailed statement foregoing invention, the following inventor lists specific embodiment and comes bright technique effect; It is emphasized that these embodiment are used to the present invention is described and are not limited to limit the scope of the invention.

The process flow diagram of embodiment such as Fig. 1, this method step is as follows

1. step 1 at first is written into original image, and it is as follows to carry out the linear transformation preprocess method, and the segmentation grey scale change is followed following rule: $f^{\′}(x,y)=\left\{\begin{array}{c}\frac{c^{\′}-a^{\′}}{c-a}(f(x,y)-a)+a^{\′}\\ \frac{d^{\′}-c^{\′}}{d-c}(f(x,y)-c)+c^{\′}\\ \frac{b^{\′}-d^{\′}}{b-d}(f(x,y)-d)+d^{\′}\end{array}\right.$

Wherein, a, c, d, b represent four separations of original image gray level, and a ', c ', d ', b ' represent the separation of the image gray levels after the mapping respectively; (x y) is the gray distribution of image function after the conversion to f '; F (x, y) be the intensity profile function of original image and a≤f (x, y)＜c, c≤f (x, y)＜d, d≤f (x, y)≤b.

2. step 2,3 choose the purpose zone in the image, and being generally needs the useful data of observation to be written in the new internal memory, and all subsequently operations are all carried out this blocks of data district.Choosing the rectangular block of needs in the first tension fault image with mouse, is unit with the pixel, by four coordinate (x1 of rectangular block, y1) (x2, y1) (x1, y3) (x2, y3) represent the upper left of rectangle respectively, upper right, the lower-left, the coordinate in the lower right corner, if the wide of original image is width, height is height, and the number of plies of faultage image is pieces, former data field pointer is pData, makes i=0～pieces-1; J=y ¹～y ³K=x ¹～x ²The data that so newly are written into are pData[i*height*width+j*width+k];

In step 4 to the managing of the volume data in the new memory field according to the structure of Octree: the basic structure of each Octree unit comprises data address pointer pBit, point to the pointer pParent of father unit, point to the array of pointers pChildren[8 of subelement], pChildren[8] eight elements are arranged, two zone bits are respectively to represent the sign M that whether traveled through, whether expression satisfies the sign S of empty condition, be written into the width newwidth=x2-x1 of image, height newheight=y3-y1, the number of plies is pieces, gets newwidth, newheight, maximal value among the pieces is made as max; Make max≤2 ^m+ 1, solve the minimum value of m, volume data is made as every limit 2 ^mThe cube of+1 voxel, so every limit can be divided into 2 ^mSection, whole volume data can be divided into 8 ^mIndividual basic unit now.Theory according to Octree: at first the junior unit of the 0th layer of level0 of the bottom is operated, 8 promptly adjacent voxels form a bottom junior unit, and each unit has data address pBit, the address of the voxel in the upper left corner of sensing unit; Because the lowermost layer unit does not have subelement, so point to the array of pointers pChildren[8 of subelement] eight element pointer be sky, the pChildren[8 of other level unit] eight element pointer according to from left to bottom right 8 subelements of sensing; If it is that empty thresholding requires or all subelements of non-level0 level satisfy empty condition that 8 voxels of the junior unit of level0 layer satisfy, zone bit S=true is set, otherwise S=false is set; M=false represents that this unit does not travel through, and M=true represents to have traveled through; PParent represents to point to the parent pointer of father unit.

Distribute the bottom, also promptly begin to be provided with from the junior unit of level0 layer, the array of pointers pChildren[8 of the sensing subelement of level0 layer junior unit] eight element pointer be set to sky, zone bit M=false, whether the gray-scale value of the setting of zone bit S by judging 8 summits is 0 to be provided with entirely, if, S=true then, if not, then put S=false; Distribute the bottom, then carry out the setting of last layer (level1), with eight adjacent junior units is the synthetic father unit of subelement, be new square, the pChildren[8 of this father unit be set] value be respectively from pointing to the pointer of eight subelements left to bottom right, zone bit M=false, the setting of zone bit S is by judging whether eight subelements all satisfy empty condition, promptly whether eight subelements all satisfy S=true, if zone bit S=true then is set; If not, then establish S=false; The parent pointer pParent of eight subelements is pointed to this father unit; Carrying out the setting of level2, is father unit of subelement recomposition with eight adjacent father unit, and parameter setting method is with the setting in the step 4.2, and the rest may be inferred, until being set to levelm;

4. step 5-10, begin Octree is traveled through from the end level (level0) of Octree, begin to search backward from first unit: if unit is the unit that traveled through or is not empty unit then skips this element and continue downwards to seek, if the unit that finds is by his zone bit S, M judges that satisfying is the empty unit that does not travel through, then to the father unit of this element traversal, then check that whether the father unit is the unit of the satisfied empty condition that do not travel through, if then continue upwards to trace back, do not stop to trace back up to not satisfying one of " not traversal " or two conditions of " empty unit ".Concrete steps are: begin Octree is traveled through from the level0 layer of Octree; Begin to search backward from first junior unit of level0; Judge that this element is not travel through, if promptly M=true then gets back to step 6; If not, M=false then carry out step 8; Judge whether this element is empty unit fully, and if not, promptly S=false then gets back to step 6; If then this element satisfies M=false and S=true simultaneously, at first define volume variable volume=0; Carry out step 9; Father unit to this element traces back; Check whether the father unit is empty unit; If satisfy condition M=false, and S=true, then get back to step 9 and continue upwards to trace back; If not, promptly during S=false, stop to trace back, establishing stop element is the unit T of leveln level, and wherein 1≤n≤m (m is maximum progression) returns to level (n-1) level and finds that subelement Tu that passes up to it, and step 11 carry out in 0≤u≤7;

5. in step 11～23, check stop element, return the subelement that passes up to it.Check whether it is (the judging promptly whether he is the unit of the bottom) that can divide,

( ^*1) if it is the secondary subelement that continues to fall back on this subelement that can divide: 0,1,2,3,4,5,6,7 represent the position of secondary subelement in subelement respectively.Be that example is introduced following volume grows process (shown in Figure 5) wherein with secondary subelement 6:

(1) secondary subelement 6 has three adjacent peer external units, and step 14 judges at first whether these outside unit at the same level traveled through, and just traveled through to stop.

(2) otherwise judge that whether it satisfy is empty unit, (promptly upwards trace back and check whether the father unit satisfies empty condition and whether travel through) downwards if satisfy repeating step 9.If do not satisfy, suppose that 6c does not satisfy, check the at the same level unit 6 adjacent with 6c, observe it and whether can divide: (step 16-20)

＜1〉if just can divide it eight minutes, check four eight molecular cells adjacent (shown in the dotted line in the unit among Fig. 36, marking off four subelements) with 6c.Whether that checks these four unit respectively adjacently traveled through with a layer unit, if traveled through then stop the calculating of this unit.If no, check then whether it is empty unit, then wants repeating step 16-20 if not empty unit, segments this subelement once more.And if it is empty unit, repeating step 11 is downward: suppose that 6a satisfies condition, judge whether it can divide, can divide then 6a eight minutes, check adjacent 3 unit at the same level of each eight molecular cell.

＜2〉if inseparable, illustrate that then this unit is lowest element and does not satisfy the condition in cavity, so stop the calculating of this unit.

( ^*2) if not (shown in the step 21) that can divide, illustrate that then this unit is a lowest element, if six adjacent peer unit of this element all or traveled through or do not satisfy empty condition then stop, otherwise be delivered to step 9, upwards travel through the operation of repeating step below 9 with the unit of minimum.

6, so far first hole can obtain by traversal, and the volumetric errors in hole is the border of a minimum unit at edge.Travel through the volume of all bubbles that can obtain area-of-interest by unit to all bottoms of volume data.

Application examples

Porose entity pore measuring method in a kind of MicroCT system in the three dimensional visualization post processing, step is:

Step 1, the original image pre-service: be written into image volumetric data, by the linear transformation of gray scale image carried out pre-service, the linear conversion process mode of gray scale is to obtain by the intensity profile Function Mapping, and the intensity profile function is

$f^{\′}(x,y)=\left\{\begin{array}{c}\frac{c^{\′}-a^{\′}}{c-a}(f(x,y)-a)+a^{\′}\\ \frac{d^{\′}-c^{\′}}{d-c}(f(x,y)-c)+c^{\′}\\ \frac{b^{\′}-d^{\′}}{b-d}(f(x,y)-d)+d^{\′}\end{array}\right.$

Wherein, more than three formulas respectively when a≤f (x, y)＜c, c≤f (x, y)＜d, d≤f (x, y)≤b satisfies, a wherein, c, d, b represent four separations of original image gray level, a ', c ', d ', b ' represent the separation of the image gray levels after the mapping respectively, be a and be mapped to a ', c is mapped to c ', and d is mapped to d ', and b is mapped to b ', f (x, y) for available f in the intensity profile function formula of original image simply replaces, (x y) is the gray distribution of image function after the conversion to f ', the initial value a of linear transformation, a ' are 0; Second value c of conversion is the maximum gray scale in cavity, makes c '=0; Promptly gray level 0 of empty gray level boil down to; The general formula of above linear transformation is got b=b '=d=d '=255, and 3 formulas were omitted and were only had 1 formula and 2 formulas to carry out linear transformation this moment; Get c=d=90, c '=d '=0 in this aluminium bubble sample; B=b '=255, a=a '=0.

Step 2 is chosen the purpose zone in the image;

Step 3 is written into new region of memory with the purpose area data: choosing the rectangular block of needs in the first tension fault image with mouse, is unit with the pixel, four coordinates (x1, y1) (x2, y1) (x1 by rectangular block, y3) (x2 y3) represents the upper left of rectangle respectively, and is upper right, the lower-left, the coordinate in the lower right corner, that establishes original image widely is width, height is height, the number of plies of faultage image is pieces, and former data field pointer is pData, makes i=0～pieces-1; J=y ¹～y ³K=x ¹～x ²The data that so newly are written into are pData[i*height*width+j*width+k];

Step 4, use the structure of Octree to manage distribution to the data of new region of memory: the basic structure of each Octree unit comprises data address pointer pBit, point to the pointer pParent of father unit, point to the array of pointers pChildren[8 of subelement], pChildren[8] eight elements are arranged, two zone bits are respectively to represent the sign M that whether traveled through, whether expression satisfies the sign S of empty condition, be written into the wide newwidth=x2-x1 of image, height newheight=y3-y1, the number of plies is pieces, gets newwidth, newheight, maximal value among the pieces is made as max; Make max≤2 ^m+ 1, solve the minimum value of m; Volume data is made as every limit has 2 ^mThe cube of+1 voxel, the data that newly are written into internal memory are put from this individual upper left, and every limit of volume data is divided into 2 ^mSection, whole volume data is divided into 8 ^mIndividual unit, 8 promptly adjacent voxels form an elementary cell, 8 synthetic father unit of adjacent subelement; Each unit has data address pBit, the address of the voxel in the upper left corner of sensing unit; Because the lowermost layer unit does not have subelement, so point to the array of pointers pChildren[8 of subelement] eight element pointer be sky, the pChildren[8 of other level unit] eight element pointer according to from left to bottom right 8 subelements of sensing; If it is that empty thresholding requires or all subelements of non-level0 level satisfy empty condition that 8 voxels of the junior unit of level0 layer satisfy, zone bit S=true is set, otherwise S=false is set; M=false represents that this unit does not travel through, and M=true represents to have traveled through; PParent represents to point to the parent pointer of father unit.

Step 4.1: distribute the bottom, also promptly begin to be provided with from the junior unit of level0 layer, the array of pointers pChildren[8 of the sensing subelement of level0 layer junior unit] eight element pointer be set to sky, zone bit M=false, whether the gray-scale value of the setting of zone bit S by judging 8 summits is 0 to be provided with entirely, if, S=true then, if not, then put S=false;

Step 4.2: then carry out the setting that last layer is level1, with eight adjacent junior units is the synthetic father unit of subelement, it is new square, the pChildren[8 of this father unit is set] value be respectively from pointing to the pointer of eight subelements left to bottom right, zone bit M=false, the setting of zone bit S is by judging whether eight subelements all satisfy empty condition, and promptly whether eight subelements all satisfy S=true, if zone bit S=true then is set; If not, then establish S=false; The parent pointer pParent of eight subelements is pointed to this father unit;

Step 4.3 is carried out the setting of level2, is father unit of subelement recomposition with eight adjacent father unit, and parameter setting method is with the setting in the step 4.2, and the rest may be inferred, until being set to levelm;

Step 5 begins Octree is traveled through from the level0 layer of Octree;

Step 6 begins to search backward from first junior unit of level0;

Step 7 judges that this element is not travel through, if promptly M=true then gets back to step 6; If not, M=false then carry out step 8;

Step 8 judges whether this element is empty unit fully, and if not, promptly S=false then gets back to step 6; If then this element satisfies M=false and S=true simultaneously, at first define volume variable volume=0; Carry out step 9;

Step 9 is traced back to the father unit of this element;

Step 10 checks whether the father unit is empty unit;

Step 10.1, if satisfy condition M=false, and S=true, then get back to step 9 and continue upwards to trace back;

Step 10.2 if not, promptly during S=false, stops to trace back, and establishing stop element is the unit T of leveln level, and wherein 1≤n≤m (m is maximum progression) returns to level (n-1) level and finds that subelement Tu that passes up to it, and step 11 carry out in 0≤u≤7;

Step 11 checks whether Tu can divide;

Step 12, if 2≤n, then Tu can divide, and calculates volume, makes volume=volume+Tu; Make the zone bit M=true of all subelements of Tu up to bottom unit, the subordinate that falls back on Tu then is positioned at eight subelements of level (n-2) level;

Step 13 is checked each three adjacent external unit at the same level of eight subelement Tuv (0≤v≤7) of being positioned at level (n-2) respectively, and each unit has only the adjacent cells of father unit altogether of three peers herein.If three adjacent external of some subelements unit at the same level is Tuva, Tuvb, Tuvc;

Step 14 judges whether the traversal sign M of Tuva, Tuvb, Tuve is true;

Step 14.1 is if M=true then finishes Tuva, Tuvb, Tuvc are searched;

Step 14.2 is if M=false then carry out step 15;

Step 15 judges respectively whether the zone bit S of Tuva, Tuvb, Tuve is true;

Step 15.1, if, then get back to step 9, promptly Tuva, Tuvb, Tuvc upwards pass up to level (n-1) level separately and check the father unit;

Step 15.2, if not, promptly S=false then carry out step 16;

Step 16 judges whether this unit can divide;

Step 16.1 if can divide, then carry out step 17;

If step 16.2 inseparable, then carry out step 21;

Step 17, unit Tuv that will be adjacent with this unit eight minutes is designated as Tuvw (0≤w≤7) (" this unit " is among Tuva, Tuvb, the Tuvc);

Step 18, three adjacent external of checking four the eight subdivision Tuvws (w get 4 numbers in 0-7) adjacent with this unit (among Tuva, Tuvb, the Tuvc) respectively are with straton unit (being non-same layer unit with the father unit) Tuvwa, Tuvwb, Tuvwc;

Step 19, judge three adjacent with straton unit Tuvwa, Tuvwb, whether Tuvwc traveled through, if traveled through then stop the calculating of this unit; If not, then carry out step 20;

Step 20 is judged Tuvwa, Tuvwb, and whether Tuvwc is empty unit;

Step 20.1, if, make Tu=Tuvwa or Tuvwb or Tuvwc, forward step 11 to; This moment, if the Tu unit is inseparable, promptly n=1 then carry out step 21 through the judgement of step 11;

Step 20.2 if not, stops the calculating of this unit;

Step 21, volume=volume+Tu is changed to true to the traversal zone bit M of Tu, and expression traveled through, and checked Tu six unit at the same level on every side;

Step 22 judges whether do not travel through empty unit, the i.e. unit of M=false and S=true in the unit at the same level;

Step 22.1 is traced back to the father unit of these unit if the condition that has some or plurality of units to satisfy M=false and S=true then forwards step 9 to;

Step 22.2, six unit at the same level do not satisfy M=false and S=true around the Tu, then stop the continuation computing of Tu;

Step 23, end point obtains first independently volume in cavity;

Step 24, continuation is to the remaining element traversal of level0 layer, find the next one of level0 to satisfy the unit of M=false and S=true, make volume=0, repeating step calculates the step of empty volume below 6, unit until all level0 layers of volume data is traveled through, and obtains the volume of all bubbles in purpose zone.

The basis of the inventive method is to be based upon on the uncontinuity in cavity.Because the cavity is discontinuous, so the obstruction of the essence pixel of the expansion of each junior unit around will running into certainly just can structure goes out the volume and the position in complete cavity.

Above-mentioned example only is explanation technical conceive of the present invention and characteristics, and its purpose is to allow the people who is familiar with this technology can understand content of the present invention and enforcement according to this, can not limit protection scope of the present invention with this.All equivalent transformations that spirit is done according to the present invention or modification all should be encompassed within protection scope of the present invention.

Claims (4)

1. assay method based on the three-dimensional object incontinuous hollow volume of image may further comprise the steps:

Be written into original image in the image processing apparatus and carry out the pre-service original image;

Pretreated image is determined the border according to the gray level of original image; The capturing image target area data is written into internal memory, and structure space Octree is determined discontinuous cavity volume by the accumulation of traversal octree nodes in internal memory;

Described original image is the three-dimensional visible image that obtains by behind the MicroCT system 3-D scanning;

Described original image pre-service is through the linear conversion process of gray scale with original image;

The linear conversion process mode of described gray scale is to obtain by the intensity profile Function Mapping, and described intensity profile function is

$f^{\′}(x,y)=\left\{\begin{array}{c}\frac{c^{\′}-a^{\′}}{c-a}(f(x,y)-a)+a^{\′}\\ \frac{d^{\′}-c^{\′}}{d-c}(f(x,y)-c)+c^{\′}\\ \frac{b^{\′}-d^{\′}}{b-d}(f(x,y)-d)+d^{\′}\end{array}\right.$

Wherein, a, c, d, b represent four separations of original image gray level, and a ', c ', d ', b ' represent the separation of the image gray levels after the mapping respectively; (x y) is the gray distribution of image function after the conversion to f '; F (x, y) be the intensity profile function of original image and a≤f (x, y)＜c, c≤f (x, y)＜d, d≤f (x, y)≤b;

Described image purpose area data is to capture the original image rectangular block earlier, obtains by the coordinate of original image rectangular block and the number of plies of original image;

Described space Octree is selected from regular Octree, linear Octree or a pair of eight formula Octrees;

Described image purpose area data is to manage distribution by the space Octree, described space Octree basic structure comprises the data address pointer, point to the pointer of father unit, point to the array of pointers of subelement, the array of pointers of described sensing subelement comprises the zone bit whether expression traveled through, the zone bit whether expression satisfies empty condition, be written into the width of image, be written into the height of image, be written into the number of plies of image; According to above-mentioned structure distribution Octree each unit in space is set.

2. the assay method of the three-dimensional object incontinuous hollow volume based on image according to claim 1 is characterized in that described space Octree traversal comprises the steps:

Begin traversal from the 0th layer of Octree, the 0th layer of first junior unit begins to search backward; Judge this unit whether traveled through with this unit whether be empty unit fully;

When this unit does not travel through and be empty unit fully, trace back to the father unit; Judge whether the father unit is whether empty unit and subelement can divide;

When but the father unit is not empty unit and subelement timesharing, move back level to subelement, check eight secondary subelements of subelement, three of periphery that check each sub-secondary units again are with the straton unit; Judge these three with the traversal situation of straton unit with whether satisfy the condition in cavity;

But when with non-cavity, straton unit and this unit timesharing will with the adjacent secondary subelement in straton unit eight minutes, 3 adjacent external of checking four eight subdivisions are with the straton unit; Judge this unit whether traveled through with this unit whether be empty unit fully;

Cross and be not empty unit with the non-traversal in straton unit when adjacent, check whether subelement can divide, so circulation; When six unit at the same level on every side that are recycled to the inseparable unit of afterbody all do not belong to empty unit or traveled through out-of-date end traversal.

3. the assay method of the three-dimensional object incontinuous hollow volume based on image according to claim 2 when it is characterized in that described space Octree traversal, continues to trace back when the father unit of subelement is empty unit.

4. the assay method of the three-dimensional object incontinuous hollow volume based on image according to claim 2 when it is characterized in that described space Octree traversal, when the unit does not satisfy empty condition and unit when inseparable, stops this unit and expands towards periphery.

Images