CN102289779A - Device for obtaining width of Moire fringe by using image processing technology - Google Patents

Abstract

The invention relates to a device for obtaining the width of a Moire fringe by using an image processing technology. The device comprises a device for acquiring an image of the Moire fringe, a device for obtaining the direction of the Moire fringe and a device for calculating the width of the Moire fringe, wherein the device for obtaining the direction of the Moire fringe defines a group of straight lines aj which are intersected into the same point on the image of the Moire fringe, obtains image gray value root mean square errors sigma j in the direction of the straight lines alpha j by using pixel values xij of each point on the image of the Moire fringe at positions where the straight lines aj pass, and uses a straight line au which corresponds to the minimum image gray value root mean square error sigma u in the sigma j as the direction of the Moire fringe. According to gray values of pixel points on the image of the Moire fringe which is acquired by a computer, the device provided by the invention automatically identifies a direction angle of the Moire fringe, calculates the width of the Moire fringe, is high in calculation speed and high in accuracy, and can be widely applied to the fields of small-angle accurate measurement, photoelectric displacement accurate measurement and the like.

Description

Utilize image processing techniques to ask for the device of Moire fringe width

Technical field

The present invention relates to a kind of device of asking for the Moire fringe width, particularly a kind of device that utilizes image processing techniques to ask for the Moire fringe width.

Technical background

Moire fringe has the optical characteristics that high magnification amplifies, utilize the measurement of moire fringe displacement can realize high Precision Detection, therefore have the important application prospect in hull modification, machine tooling field, quick and precisely ask for the Moire fringe orientation angle and width has good practical significance.

Comprise that frequency asks for algorithm, Threshold Segmentation Algorithm etc. as long as ask for the method for Moire fringe at present, the thresholding method computing velocity is very fast but computational accuracy is low, and frequency domain Processing Algorithm calculated amount is difficult to satisfy the requirement on the speed greatly.Therefore how to seek the problem that a kind of algorithm of quick and precisely asking for the Moire fringe width becomes urgent solution.

Summary of the invention

The technical problem to be solved in the present invention provides a kind of device that utilizes image processing techniques to ask for the Moire fringe width fast and accurately.

In order to solve the problems of the technologies described above, the device that utilizes image processing techniques to ask for the Moire fringe width of the present invention comprises:

Moire fringe image harvester;

Ask for the device of Moire fringe direction;

Calculate the device of Moire fringe width;

The described device of asking for the Moire fringe direction: definition gang meets at the straight line a of same point o on the Moire fringe image _j(j=1,2 ... m), cut-off line a _jPixel value x through each point on the Moire fringe image of position _Ij(i=1,2 ... n); Utilize formula (1) calculated line a _jDirectional image gray-scale value root-mean-square error σ _j

${\mathrm{\σ}}_j=\frac{1}{n-1}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}\sqrt{{(x_{\mathrm{ij}}-\stackrel{\‾}{x_j})}^2}---\left(1\right)$

$\stackrel{\‾}{x_j}=\frac{1}{n}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{x}_{\mathrm{ij}}---\left(2\right)$

Get σ _jMiddle minimum gradation of image value root-mean-square error σ _uPairing straight line a _uDirection as Moire fringe.

The present invention according to computer acquisition to the Mohr's circle picture on gray values of pixel points, discern the deflection of Moire fringe automatically and calculate the width of striped, computing velocity is fast and precision is high.Can be widely used in fields such as low-angle precision measurement, photoelectric displacement precision measurement.

Described some o is positioned at the brightest position, picture centre zone or the darkest position.

Described straight line a _jEvenly distribute in 360 ° of scopes, adjacent straight wire spacing angle is in 0.25～0.5 ° of scope.

Described straight line a _j(j=1,2 ... m) comprise two groups of straight line a in _c(c=1,2 ... l) (l＜m) and a _d(d=1,2 ... p) (p=m-l); Straight line a _c(c=1,2 ... l) evenly distribute in 360 ° of scopes, adjacent straight wire spacing angle is in 1～2 ° of scope; Straight line a _d(d=1,2 ... p) be distributed in straight line a _V-1With straight line a _V+1Between, straight line a _d(d=1,2 ... p) adjacent straight wire spacing angle is in 0.05～0.1 ° of scope in; Its cathetus a _V-1, straight line a _V+1Be straight line a _c(c=1,2 ... l) in a _vTwo adjacent straight lines, a _vBe straight line a _c(c=1,2 ... l) in the gradation of image value root-mean-square error σ of minimum _vCorresponding straight line; Straight line a _d(d=1,2 ... p) in the gradation of image value root-mean-square error σ of minimum _wCorresponding straight line a _wDirection be the Moire fringe direction.

The device of described calculating Moire fringe width: define a straight line, obtain the period of change of this straight line, promptly obtain the width of Moire fringe through each point pixel value on the Moire fringe image of position perpendicular to the Moire fringe direction.

The device of described calculating Moire fringe width:

As true origin, set up coordinate system as horizontal ordinate, the v axle parallel as ordinate, utilize sciagraphy to obtain a sinusoidal curve with the o point with the Moire fringe direction with the vertical u axle of Moire fringe direction; Wherein the ordinate of each point is total gray-scale value that horizontal ordinate is identical on the Moire fringe each point grey scale pixel value adds up and obtains on the sinusoidal curve; Obtain this sinusoidal cycle, this cycle is the width of Moire fringe.

Description of drawings

Below in conjunction with the drawings and specific embodiments the present invention is described in further detail.

Fig. 1 is the apparatus structure synoptic diagram that utilizes image processing techniques to ask for the Moire fringe width automatically of the present invention.

Fig. 2 is a lowest mean square root error schematic diagram.

Fig. 3 a, 3b are the straight line a that meets at same point o _j(j=1,2 ... m) with Moire fringe image relative position synoptic diagram.

Fig. 4 is the figure of Moire fringe projection coordinate.

Fig. 5 is that Moire fringe is along the stripe direction perspective view.

Fig. 6 is a process flow diagram of asking for Moire fringe width program thereby.

Embodiment

As shown in Figure 1, the device that utilizes image processing techniques to ask for the Moire fringe width automatically of the present invention comprises

Moire fringe image harvester 101;

Ask for the device 102 of Moire fringe direction: definition gang meets at the straight line a of same point o on the Moire fringe image _j(j=1,2 ... m), cut-off line a _jPixel value x through each point on the Moire fringe image of position _Ij(i=1,2 ... n); Utilize formula (1) calculated line a _jDirectional image gray-scale value root-mean-square error σ _j

${\mathrm{\σ}}_j=\frac{1}{n-1}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}\sqrt{{(x_{\mathrm{ij}}-\stackrel{\‾}{x_j})}^2}---\left(1\right)$

$\stackrel{\‾}{x_j}=\frac{1}{n}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{x}_{\mathrm{ij}}---\left(2\right)$

Get σ _jMiddle minimum gradation of image value root-mean-square error σ _uPairing straight line a _uDirection as Moire fringe;

Calculate the device 103 of Moire fringe width.

One, Moire fringe angle calculation

For one group of data x ₁x ₂..x _i... x _n, its average as shown in Equation (3):

$\stackrel{\‾}{x}=\frac{1}{n}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{x}_i---\left(3\right)$

The data root-mean-square error is as shown in Equation (4):

$\mathrm{\σ}=\frac{1}{n-1}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}\sqrt{{(x_i-\stackrel{\‾}{x})}^2}---\left(4\right)$

By formula (4) as can be seen, if numerical value is approaching more between should group data, its root-mean-square error is more little, principle according to this, and we are the center with the Moire fringe picture centre, draw a straight line, and get Moire fringe image grey scale pixel value at this straight line through the position.As shown in Figure 2: straight line a direction is consistent with the Moire fringe direction, and then this rectilinear direction image grey scale pixel value is approximate, in the formula (4) in the radical sign data near 0, root-mean-square error value minimum; Straight line c direction is vertical with the Moire fringe direction, and this rectilinear direction image gradation data fluctuation is maximum, and its root-mean-square error is also maximum; And straight line b direction root-mean-square error falls between.

In Fig. 2, straight line a is parallel to the Moire fringe direction, find that by statistics lowest mean square root error error amount σ is relevant with the position of straight line a process, when straight line a process striped brightest area or dark areas, the σ value is minimum, and when straight line process Moire fringe light and shade transitional region, this location drawing causes the σ value bigger more greatly as the gray-value variation amplitude, this position is unfavorable for the calculating of Moire fringe deflection, so straight line is chosen and will be followed certain principle.Family of straight lines intersection point o should be chosen at the brightest position, picture centre zone or the darkest position (being gray-scale value maximum position place or gray-scale value minimum position place).

Each straight line a _j(j=1,2 ... m) in 360 ° of scopes, can be evenly distributed, also can be uneven distribution.

Two, calculate the Moire fringe width

After accurately obtaining the Moire fringe deflection, can obtain the Moire fringe width along the direction vertical with the Moire fringe direction.

Because there is noise in the Moire fringe imaging, it is very big to cause the Moire fringe width to calculate stochastic error.Sciagraphy is that a series of related datas are carried out data accumulation by fixed-direction since sciagraphy to have the good restraining effect to have calculated amount simultaneously to noise little, the present invention adopts sciagraphy to calculate the Moire fringe width.The tradition sciagraphy adopts x direction of principal axis projection or the projection of y direction of principal axis, because the Moire fringe deflection at random, the present invention adopts and carries out projection along the Moire fringe direction.As shown in Figure 4, set up coordinate system, the u axle is and the Moire fringe vertical direction, and the v axle is and the Moire fringe parallel direction, and view field is the white box enclosing region, with v axle permutation data projection at the u direction of principal axis.

Fig. 4 obtains the Moire fringe perspective view for adopting sciagraphy at the u axle, u coordinate representation Moire fringe image horizontal ordinate wherein, and ordinate is represented the gray-scale value that adds up of Moire fringe projection on the u coordinate, handles back Moire fringe data through sciagraphy and becomes level and smooth.Ask for the Moire fringe width and become and ask for the sinusoidal curve cycle (as shown in Figure 5), in Fig. 5, seek first crest location of sinusoidal curve u ₁With last crest location u ₂Utilize formula (5) to ask for width of fringe B.

$B=\frac{u_2-u_1}{k}---\left(5\right)$

In the formula (5): u ₁Represent first crest, u ₂Represent last crest, k represents crest quantity.

Embodiment 1

Shown in Fig. 3 a, straight line a _j(j=1,2 ... m) intersection point o is chosen at the darkest position, Moire fringe picture centre zone, and family of straight lines a _j(j=1,2 ... m) evenly distribute in 360 ° of scopes, adjacent straight wire spacing angle is 0.3 °, 0.4 ° or 0.5 °.

Cut-off line a _jPixel value x through each point on the Moire fringe image of position _Ij(i=1,2 ... n); Utilize formula (1) calculated line a _jDirectional image gray-scale value root-mean-square error σ _j

${\mathrm{\σ}}_j=\frac{1}{n-1}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}\sqrt{{(x_{\mathrm{ij}}-\stackrel{\‾}{x_j})}^2}---\left(1\right)$

$\stackrel{\‾}{x_j}=\frac{1}{n}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{x}_{\mathrm{ij}}---\left(2\right)$

Get σ _jMiddle minimum gradation of image value root-mean-square error σ _uPairing straight line a _uDirection as Moire fringe.

Define one by an o and perpendicular to straight line a _uStraight line b _u, straight line b _uMoire fringe image pixel gray-scale value through the position is the variation of sinusoidal curve cycle, obtains the width that this sinusoidal cycle is Moire fringe.

Embodiment 2

Family of straight lines a _j(j=1,2 ... m) intersection point o is chosen at the brightest position, Moire fringe picture centre zone, family of straight lines a _j(j=1,2 ... m) comprise two groups of straight line a _c(c=1,2 ... l) (l＜m) and a _d(d=1,2 ... p) (p=m-l).

Straight line group a _c(c=1,2 ... l) evenly distribute in 360 ° of scopes, adjacent straight wire spacing angle is 1 °, 1.5 ° or 2 °.

Cut-off line a _cPixel value x through each point on the Moire fringe image of position _Ic(i=1,2 ... n); Utilize formula (11) calculated line a _cDirectional image gray-scale value root-mean-square error σ _c

${\mathrm{\σ}}_c=\frac{1}{n-1}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}\sqrt{{(x_{\mathrm{ic}}-\stackrel{\‾}{x_c})}^2}---\left(11\right)$

$\stackrel{\‾}{x_c}=\frac{1}{n}\underset{i=1}{\overset{n}{\text{\Σ}}}{x}_{\mathrm{ic}}---\left(12\right)$

Get σ _cMiddle minimum gradation of image value root-mean-square error σ _vPairing straight line a _vApproximate direction as Moire fringe.

Straight line group a _d(d=1,2 ... p) be distributed in straight line a _V-1With straight line a _V+1Between (shown in Fig. 3 b), straight line a _V-1, straight line a _V+1Be straight line group a _c(c=1,2 ... l) in a _vTwo adjacent straight lines.Straight line group a _d(d=1,2 ... p) adjacent straight wire spacing angle is 0.05 °, 0.07 ° or 0.1 ° in.

According to asking straight line a _vSame method is obtained straight line group a _d(d=1,2 ... p) the pairing straight line a of gradation of image value root-mean-square error of minimum in _wAs the Moire fringe direction.

As shown in Figure 4, as true origin, set up coordinate system as horizontal ordinate, the v axle parallel as ordinate, utilize sciagraphy to obtain a sinusoidal curve with the o point with the Moire fringe direction with the vertical u axle of Moire fringe direction; Wherein the ordinate p of each point is total gray-scale value that horizontal ordinate is identical on the Moire fringe each point grey scale pixel value adds up and obtains on the sinusoidal curve; In Fig. 4, seek first crest location of sinusoidal curve u ₁With last crest location u ₂Utilize formula (5) can try to achieve width of fringe B.

$B=\frac{u_2-u_1}{k}---\left(5\right)$

In the formula 3: u ₁Represent first crest, u ₂Represent last crest, k represents crest quantity.

The invention is not restricted to the foregoing description, straight line a _j(j=1,2 ... m) can also select to comprise many group straight lines in.For example first group of adjacent straight wire spacing angle is 10 °, totally 18 straight lines; Obtaining the pairing straight line of gradation of image value root-mean-square error minimum in these 18 straight lines is a ₇, then second group of straight line is distributed in straight line a ₆With straight line a ₈Between, adjacent straight wire spacing angle is 1 °, totally 18 straight lines; Obtaining the pairing straight line of gradation of image value root-mean-square error minimum in these 18 straight lines is a ₃₀, then the 3rd component is distributed in straight line a ₂₉With straight line a ₃₁Between, adjacent straight wire spacing angle is 0.1 °, totally 18 straight lines, and obtaining the pairing straight line of gradation of image value root-mean-square error minimum in these 18 straight lines again is a ₄₀, with straight line a ₄₀The place direction is as the direction of Moire fringe.The calculating of Moire fringe width also is not limited to aforesaid way, embodiment 1 cathetus b _uAlso can be not by an o but perpendicular to straight line a _uAny straight line.Should be understood that every any simple deformation of making all the invention is intended within the protection domain on claim 1 of the present invention basis.

The present invention is exactly the Mohr's circle picture that arrives according to computer acquisition, discerns the deflection of Moire fringe and the width of striped automatically.As straight line a _j(j=1,2 ... m) intersection point o is chosen at the brightest position, picture centre zone, selects to be spaced apart between the adjacent straight line 1 ° of preliminary straight line a of determining _vDirection is the Moire fringe direction, then at straight line a _V-1With straight line a _V+1Between another group of definition straight line a _d(d=m-c), adjacent straight wire spacing angle is 0.1 °, gets minimum gradation of image value root-mean-square error σ in this group again _vPairing straight line a _vDirection as Moire fringe shows the result.Every two field picture is handled and is finished within 40ms, and the width of fringe error of calculation is less than 1%.

The present invention adopts the VC6.0 programming to realize that the program circuit of algorithm as shown in Figure 6 by at double-core Duo 2.4GHz, internal memory 2G computer.

Claims (6)

1. device that utilizes image processing techniques to ask for the Moire fringe width is characterized in that comprising:

Moire fringe image harvester;

Ask for the device of Moire fringe direction;

Calculate the device of Moire fringe width;

The described device of asking for the Moire fringe direction: definition gang meets at the straight line a of same point o on the Moire fringe image _j(j=1,2 ... m), cut-off line a _jPixel value x through each point on the Moire fringe image of position _Ij(i=1,2 ... n); Utilize formula (1) calculated line a _jDirectional image gray-scale value root-mean-square error σ _j

${\mathrm{\σ}}_j=\frac{1}{n-1}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}\sqrt{{(x_{\mathrm{ij}}-\stackrel{\‾}{x_j})}^2}---\left(1\right)$

$\stackrel{\‾}{x_j}=\frac{1}{n}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{x}_{\mathrm{ij}}---\left(2\right)$

Get σ _jMiddle minimum gradation of image value root-mean-square error σ _uPairing straight line a _uDirection as Moire fringe.

2. the device that utilizes image processing techniques to ask for the Moire fringe width according to claim 1 is characterized in that described some o is positioned at the brightest position, picture centre zone or the darkest position.

3. the device that utilizes image processing techniques to ask for the Moire fringe width according to claim 1 and 2 is characterized in that described straight line a _jEvenly distribute in 360 ° of scopes, adjacent straight wire spacing angle is in 0.25～0.5 ° of scope.

4. the device that utilizes image processing techniques to ask for the Moire fringe width according to claim 1 and 2 is characterized in that described straight line a _j(j=1,2 ... m) comprise two groups of straight line a in _c(c=1,2 ... l) (l＜m) and a _d(d=1,2 ... p) (p=m-l); Straight line a _c(c=1,2 ... l) evenly distribute in 360 ° of scopes, adjacent straight wire spacing angle is in 1～2 ° of scope; Straight line a _d(d=1,2 ... p) be distributed in straight line a _V-1With straight line a _V+1Between, straight line a _d(d=1,2 ... p) adjacent straight wire spacing angle is in 0.05～0.1 ° of scope in; Its cathetus a _V-1, straight line a _V+1Be straight line a _c(c=1,2 ... l) in a _vTwo adjacent straight lines, a _vBe straight line a _c(c=1,2 ... l) in the gradation of image value root-mean-square error σ of minimum _vCorresponding straight line; Straight line a _d(d=1,2 ... p) in the gradation of image value root-mean-square error σ of minimum _wCorresponding straight line a _wDirection be the Moire fringe direction.

5. the device that utilizes image processing techniques to ask for the Moire fringe width according to claim 1, the device that it is characterized in that described calculating Moire fringe width: define a straight line perpendicular to the Moire fringe direction, obtain the period of change of this straight line, promptly obtain the width of Moire fringe through each point pixel value on the Moire fringe image of position.

6. the device that utilizes image processing techniques to ask for the Moire fringe width according to claim 1 is characterized in that the device of described calculating Moire fringe width:

As true origin, set up coordinate system as horizontal ordinate, the v axle parallel as ordinate, utilize sciagraphy to obtain a sinusoidal curve with the o point with the Moire fringe direction with the vertical u axle of Moire fringe direction; Wherein the ordinate of each point is total gray-scale value that horizontal ordinate is identical on the Moire fringe each point grey scale pixel value adds up and obtains on the sinusoidal curve; Obtain this sinusoidal cycle, this cycle is the width of Moire fringe.

Images