CN105158892A - Interference measurement rapid focusing realization method - Google Patents

Abstract

The invention discloses an interference measurement rapid focusing realization method. In the focusing process, a sample is scanned in a Z-direction by using an interference microscope system to acquire an image, the image of which the sum of gray difference is maximum is found by applying a focusing function method, and the position corresponding to the image is a focusing point. Compared with conventional methods, operation time is reduced by times, and the operation flow of the interference microscope system is also simplified by the focusing method. Besides, a novel focusing evaluation function does not need to consider regional correlation, and interlaced calculation can be adopted in the data processing process so that calculation amount can be reduced without influencing effectiveness of the evaluation value.

Description

A kind of method realizing interferometry and focus on fast

Technical field

This invention relates to the automatic focus field of interferometry, specifically a kind of interference microscope quick focusing method based on Digital Image Processing be applicable in interferometry process.

Background technology

In modernization detection method, a kind of optical interference microscope of high precision Non-Destructive Testing is had to measure (including Michelson type, Mirau type, Linnik type etc.) technology.This kind of technology, for detecting micron, nanoscale object surface microscopic appearance, interferes the depth of field in micro-imaging process very little at high magnification, and manual focusing is real is inconvenience.In order to realize the automatic operating of system, one of gordian technique is exactly realize Techniques of Automatic Focusing.Be applied to microscopical automatic focus mode at present and mostly be focusing depth method based on Digital Image Processing.Focusing depth method is by building evaluation function, evaluating the sharpness of obstructed location drawing picture, and obtains focusing surface at the maximal value place of evaluation function.The correctness that autofocus evaluation function is chosen, directly has influence on precision and the focusing speed of focusing.

The autofocus evaluation function of domestic and international proposition of the prior art is broadly divided into:

1, frequency-domain function.As entropy function, wavelet transformation, Fourier transform etc.

2, gradient function.In image procossing, gradient function is often used to extract marginal information.Focus on good image, namely have the image at more sharp-pointed edge, the gradient function value that Ying Yougeng is large.Conventional gradient function has variance operator, grey scale difference absolute value sum operator, energy gradient operator, Brenner function, Tenengrad function, Roberts gradient operator, Laplace operator etc. based on Sobel gradient operator.

3, statistics function.As Range function, Enmay function, Masgrn function, AC power function, DC power function etc.

The value of shade of gray evaluation function adopts the gray scale difference (difference) between image neighbor to calculate, this approximate differential computing can eliminate the impact of ground unrest, can effectively suppress external condition to the entire effect of image, and meet unimodality, unbiasedness and sensitivity feature, therefore this class evaluation function is most widely used.Suppose that the gray-scale value at certain point (x, y) place in image is g (x, y), the scale of image is M × N (M row, N are capable) individual pixel, then shade of gray evaluation function has following several definition:

(1) average degree of grey scale change and variance are weighed.The gray variance operator definitions of image is as follows:

wherein, g ₀the mean value of gradation of image,

Be expressed as $g_0=\frac{1}{MN}\underset{x}{\overset{M}{\Σ}}\underset{y}{\overset{N}{\Σ}}g(x,y).$

(2) grey scale difference absolute value sum operator (SMD).Power and evolution is replaced by difference absolute value.Namely calculus of differences is done to the gray scale of point (x, y) and neighbor point thereof, extracts the change size of this gray-scale value, draw gradation of image difference absolute value sum operator: $F\left(k\right)=\underset{x}{\overset{M}{\Σ}}\underset{y}{\overset{N}{\Σ}}|g(x,y)-g(x+1,y)|+|g(x,y)-g(x,y+1)|$

(3) energy gradient operator: $F\left(k\right)=\underset{x}{\overset{M}{\Σ}}\underset{y}{\overset{N}{\Σ}}|g(x,y)-g(x+1,y){|}^2+|g(x,y)-g(x,y+1){|}^2$

(4) Brenner function, is also called gradient filter method, and it is when the gray scale difference of adjacent picture elements is greater than certain threshold value, and (the pixel gray scale difference be spaced apart k) to carry out square summation to adjoining.Its expression formula is as follows:

<math><math display = 'block'> <mrow> <mi>F</mi> <mrow> <mo>(</mo> <mi>k</mi> <mo>)</mo> </mrow> <mo>=</mo> <munder> <mi>&amp;Sigma;</mi> <mi>x</mi> </munder> <munder> <mi>&amp;Sigma;</mi> <mi>y</mi> </munder> <msup> <mrow> <mtext>&amp;nbsp;</mtext> <mo>&amp;lsqb;</mo> <mi>g</mi> <mrow> <mo>(</mo> <mi>x</mi> <mo>&amp;plus;</mo> <mi>k</mi> <mo>,</mo> <mi>y</mi> <mo>)</mo> </mrow> <mo>&amp;minus;</mo> <mi>g</mi> <mrow> <mo>(</mo> <mi>x</mi> <mo>,</mo> <mi>y</mi> <mo>)</mo> </mrow> <mo>&amp;rsqb;</mo> </mrow> <mn>2</mn> </msup> <mo>,</mo> </mrow></math> When | g (x+1, y)-g (x, y) | >G _threshold

In above formula, G _thresholdbe threshold condition, g (x, y) is the gray-scale value of picture dot (x, y), lower same.

(5) based on the Tenengrad function of Sobel gradient operator utilize Sobel operator come estimated image in the horizontal direction with the gradient of vertical direction, for making the gradient of image border amplify, carry out square operation to gradient, its expression formula is:

$F\left(k\right)=\underset{x}{\&Sigma;}\underset{y}{\&Sigma;}{\&lsqb;S(x,y)\&rsqb;}^2$

Wherein: g _x(x, y) and G _y(x, y) is respectively the difference value that the Sobel operator in X and Y-direction calculates.The Sobel operator expression formula of X-direction and Y-direction is as follows:

$\begin{array}{cc}{S}_x=\left(\begin{array}{ccc}-1& 0& 1\\ -2& 0& 2\\ -1& 0& 1\end{array}\right),& S_y=\left(\begin{array}{ccc}1& 2& 1\\ 0& 0& 0\\ -1& -2& -1\end{array}\right)\end{array}$

Tenengrad function is first weighted mean and then difference, carries out template computing respectively to horizontal and vertical direction, has certain rejection ability to noise.Along with the amplification of window, noise inhibiting ability is more obvious.

(6) Roberts gradient sum functions be actually by certain a bit centered by continuous gradient approximate, analyze theoretically, the process of its edge characteristic is better than Brenner function.

$F\left(k\right)=\underset{x}{\&Sigma;}\underset{y}{\&Sigma;}\left\{\right|g(x,y)-g(x+1,y+1)|+|g(x+1,y)-g(x,y+1)\left|\right\}$

(7) Laplace operator is a sharpening template, fuzzy by what strengthen that high fdrequency component reduces in image, is also called high-pass filtering.Common Laplace operator template have three kinds (left, in be 4 neighborhoods, the right side is 8 neighborhoods):

$\left[\begin{array}{ccc}0& 1& 0\\ 1& -4& 1\\ 0& 1& 0\end{array}\right]\left[\begin{array}{ccc}1& 0& 1\\ 0& -4& 0\\ 1& 0& 1\end{array}\right]\left[\begin{array}{ccc}1& 1& 1\\ 1& -8& 1\\ 1& 1& 1\end{array}\right]$

Improve Laplce's gradient function and add variable step size step to calculate second order difference.

ML(x,y)＝|2g(x,y)-g(x-step,y)-g(x+step,y)|+|2g(x,y)-g(x,y-step)-g(x,y+step)|

$F\left(k\right)=\underset{x}{\&Sigma;}\underset{y}{\&Sigma;}ML(x,y)$ Wherein ML (x, y) >=T

T is a threshold value, and the Laplacian values being greater than threshold value just participates in gathering.

In above shade of gray evaluation function, fastest is use gray variance operator.

And corresponding foregoing description, prior art Problems existing and shortcoming have:

Automatic focus process need gathers hundreds and thousands of width image, and shade of gray function is evaluated every piece image again, the camera that pixel is higher, and the data volume of process is doubled and redoubled especially.So in practical application, the focusing using shade of gray function still slowly, can not meet state-of-the-art facility requirement fast and efficiently.

Therefore, need a kind of new technical scheme to solve the problem.

Summary of the invention

The object of the invention is the deficiency existed for prior art, provide a kind of method realizing interferometry and focus on fast, have precision high, sensitivity is good, can reduce the searching times of focusing, accelerates the advantage of focusing speed.

For achieving the above object, the present invention can adopt following technical scheme:

Realize the method that interferometry focuses on fast, optical interference microscope is provided; The lateral coordinates arranging measured surface image is x, y coordinate, and close or away from measured surface image the coordinate of optical interference microscope is z coordinate;

Interference microscope scans acquisition image in z-direction;

Find gray scale difference and that maximum width image, the position that this image is corresponding is exactly focal position, and calculates gray scale difference and F (k) by following formula autofocus evaluation function,

$F\left(k\right)={\mathrm{\&Sigma;}}_x^M{\mathrm{\&Sigma;}}_y^N|g_{\left(k\right)}(x,y)-g_{(k+1)}(x,y)|$

Wherein, if the gray-scale value at certain point (x, y) place is g in kth width image _(k)(x, y), image is M × N number of pixel; The maximum figure position of evaluation of estimate is exactly the focal position found in this image scanning process.

Compared with background technology, because interference image is abundanter than common focusedimage at the high-frequency component information of focal position, so the autofocus evaluation function adopted in the present invention has stronger robustness and sensitivity.Meanwhile, autofocus evaluation function calculating process is by original square operation and repeatedly plus and minus calculation, and be simplified to single plus and minus calculation, operation time reduces at double, and simplifies the microscopical operating process of optical interference.And novel autofocus evaluation function is without the need to considering area coherence.Therefore, in process data procedures, interlacing can be adopted to calculate, while minimizing calculated amount, the also validity of not effect appraise value.

Accompanying drawing explanation

Fig. 1 is the process flow diagram using optical interference microscope to carry out interferometry in the present invention to focus on fast.

Fig. 2 is the functional image of the autofocus evaluation function adopted in the present invention.

Embodiment

Below in conjunction with the drawings and specific embodiments, illustrate the present invention further, these embodiments should be understood only be not used in for illustration of the present invention and limit the scope of the invention, after having read the present invention, the amendment of those skilled in the art to the various equivalent form of value of the present invention has all fallen within the application's claims limited range.

According to interference microscope characteristic, the quantity of information be distributed on a pixel is:

I(z,x,y)＝I ₀(z,x,y)+g(z(x,y)-z ₀)cos(k _c(z(x,y)-z ₀)(1+cosθ ₀)+α)(1)

Wherein, θ ₀n.A.=sin θ relevant to the numerical aperture of object lens ₀, be constant; Z is that the z of measuring surface is to scanning survey position; z ₀it is focal position; α is phase shift angle, is fixed constant; k _cit is ripple number (depending on light source center wavelength).Background light intensity I in any pixel ₀(z, x, y) slowly can change to the change of scanning position along with z.G (z (x, y)-z ₀) only have a peak value at z (x, y)=z ₀place, and along with larger with focal length, be worth less.

Equation (1) is by drawing z differentiate:

$\begin{array}{c}\frac{\&part;I(z,x,y)}{\&part;z}=\frac{\&part;g(z,x,y)}{\&part;z}cos\left(k\right(z(x,y)-z_0)+\mathrm{\&alpha;})-kg(z,x,y)sin\left(k\right(z(x,y)-z_0)+\mathrm{\&alpha;})\\ =\sqrt{{\left(\frac{\&part;g(z,x,y)}{\&part;z}\right)}^2+{\left(kg(z,x,y)\right)}^2}cos\left(\mathrm{\&Phi;}(z,x,y)\right))\end{array}---\left(2\right)$

Wherein k=k _c(1+cos θ ₀), $\mathrm{\&Phi;}(z,x,y)=k\left(z\right(x,y)-z_0)+\mathrm{\&alpha;}+arctan\frac{-kg(z,x,y)}{\frac{\&part;g(z,x,y)}{\&part;z}},$

Because above formula eliminates the background light intensity of slowly change, thus retain modulation signal.So the change function of image can be defined as:

$Q\left(z\right)=\&Integral;\&Integral;\left|\frac{\&part;I(z,x,y)}{\&part;z}\right|dxdy---\left(3\right)$

Due near focal position cos (Φ (z, x, y)) change is quicker than g (z, x, y), so equation (3) has unimodality, and when z is closest to z ₀, namely during focal position, Q (z) obtains maximal value.

According to above interference microscope characteristic, the present invention proposes a kind of new focus method as follows:

Optical interference microscope is provided; The lateral coordinates arranging measured surface image is x, y coordinate, and close or away from measured surface image the coordinate of optical interference microscope is z coordinate;

Interference microscope scans acquisition image in z-direction;

Find gray scale difference and that maximum width image, the position that this image is corresponding is exactly focal position, and calculates gray scale difference and F (k) by following formula autofocus evaluation function,

$F\left(k\right)={\mathrm{\&Sigma;}}_x^M{\mathrm{\&Sigma;}}_y^N|g_{\left(k\right)}(x,y)-g_{(k+1)}(x,y)|$

Wherein, if the gray-scale value at certain point (x, y) place is g in kth width image _(k)(x, y), image is M × N number of pixel; The maximum figure position of evaluation of estimate is exactly the focal position found in this image scanning process.

Because interference image is abundanter than common focusedimage at the high-frequency component information of focal position, so this evaluation function has stronger robustness and sensitivity.Meanwhile, functional operation process is by original square operation and repeatedly plus and minus calculation, and be simplified to single plus and minus calculation, operation time reduces at double.And novel autofocus evaluation function is without the need to considering area coherence.Therefore, in process data procedures, interlacing can be adopted to calculate, while minimizing calculated amount, the also validity of not effect appraise value.Through actual test, evaluate the digital picture of a width 1280x960, gray variance operator used time 70ms, and novel autofocus evaluation function only uses 12ms, when adopting interlacing to calculate, consuming timely shorten half especially.Image processing time greatly reduces, and whole focusing speed is accelerated, and the stand-by period of user reduces, and work efficiency is improved.This novel autofocus evaluation function more can embody fast in the application of high precision sample and high pixel camera, high efficiency, so applicability is stronger.

Shown in Fig. 1 and Fig. 2, the concrete steps of above-mentioned quick focusing method are:

(1), the motor of initialized optical interference microscope, make motor removing to system default position O1;

(2), arrange shooting step-length S1, motor moves down distance L1, and gathers storage figure picture in motor removing process;

(3), use autofocus evaluation function to calculate the evaluation of estimate of every piece image, as shown in Figure 2, the maximum figure position of evaluation of estimate is exactly focal position O2 for the first time to evaluation function;

(4), motor removing to first time focal position O2, amendment shooting step-length S2, make S2<S1, motor moves down distance L2, makes L2<L1, and gathers storage figure picture in motor removing process;

(5), use autofocus evaluation function to calculate the evaluation of estimate of the every piece image of process in displacement L2, judge second time focal position O3;

(6), motor removing to O3, focusing terminates.

Claims (2)

1. realize the method that interferometry focuses on fast, it is characterized in that:

Optical interference microscope is provided; The lateral coordinates arranging measured surface image is x, y coordinate, and close or away from measured surface image the coordinate of optical interference microscope is z coordinate;

Interference microscope scans acquisition image in z-direction;

Find gray scale difference and that maximum width image, the position that this image is corresponding is exactly focal position, and calculates gray scale difference and F (k) by following formula autofocus evaluation function,

$F\left(k\right)={\mathrm{\&Sigma;}}_x^M{\mathrm{\&Sigma;}}_y^N|g_{\left(k\right)}(x,y)-g_{(k+1)}(x,y)|$

Wherein, if the gray-scale value at certain point (x, y) place is g in kth width image _(k)(x, y), image is M × N number of pixel; The maximum figure position of evaluation of estimate is exactly the focal position found in this image scanning process.

2. realize the method that interferometry focuses on fast as claimed in claim 1, it is characterized in that: comprise searching twice focal position, concrete steps are:

(1), the motor of initialized optical interference microscope, make motor removing to system default position O1;

(2), arrange shooting step-length S1, motor moves down distance L1, and gathers storage figure picture in motor removing process;

(3), use autofocus evaluation function to calculate the evaluation of estimate of every piece image, the maximum figure position of evaluation of estimate is exactly focal position O2 for the first time;

(4), motor removing to first time focal position O2, amendment shooting step-length S2, make S2<S1, motor moves down distance L2, makes L2<L1, and gathers storage figure picture in motor removing process;

(5), use autofocus evaluation function to calculate the evaluation of estimate of the every piece image of process in displacement L2, judge second time focal position O3;

(6), motor removing to O3, focusing terminates.