CN102609915B - Grating automatic-splicing algorithm based on far-field spots - Google Patents

Abstract

The invention discloses a grating automatic-splicing algorithm based on far-field spots. The algorithm comprises the following steps of: collecting a current far-field spot image F1 through a CCD (Charge Coupled Device), preprocessing the image F1, and obtaining a filtered image F2; adopting a laser-spot optimal threshold segmentation algorithm for the image F2 to obtain an image F3; adopting a projection algorithm for the image F3, and respectively calculating characteristic parameters between a highest peak and a second peak in an x direction and a y direction, such as the distance, the distance ratio, the peak ratio and the like; and then controlling the motion of a motor and adjusting the gesture of grating according to feature extraction results, and finally achieving the purpose of good grating automatic-splicing.

Description

A kind of grating automatic Mosaic algorithm based on far-field spot

Technical field

The present invention relates to a kind of grating automatic Mosaic algorithm based on far-field spot.

Background technology

In chirped pulse amplification system, the damage threshold of ultrashort laser pulse energy constraint grating in pulse shortener of generation and bore, require that the size of grating is enough large.The grating higher to ultrashort pulse damage threshold is multilayer dielectric holographic grating at present, and it is very difficult to obtain this large-area grating.Therefore a large-area grating is become by several pieces little raster detect to be a good method.

Two blocks of gratings are spliced, main existence six kinds of basic stitching errors: two kinds of translation errors, three kinds of rotation errors and grating line width errors, translation error is respectively between not coplanar and two sub-gratings of two sub-gratings slit, and grid face and the grid line of three kinds of corresponding two sub-gratings of rotation error are not parallel each other; These errors will bring phase differential to light beam, make the Quality Down of pulse compression.It is three right to be divided into these six kinds of stitching errors, and be the rotation error on translation error, x and z direction, the rotation error on y direction and grating line width error respectively, often pair of error parameter can compensate mutually.

One piece of jointing grating of good performance must meet following three conditions: (1) each piece of sub-gratings must be coplanar; (2) grid line of sub-gratings must be parallel to each other; (3) gap between sub-gratings is the integral multiple of screen periods.From current bibliographical information, the means such as the far-field distribution of jointing grating, near field wavefront sensing are mainly utilized to carry out the error of analysis list group jointing grating existence.

Summary of the invention

Technical matters to be solved by this invention provides a kind of grating automatic Mosaic algorithm based on far-field spot for the deficiencies in the prior art.

Technical scheme of the present invention is as follows:

Based on a grating automatic Mosaic algorithm for far-field spot, comprise the following steps:

(1) CCD gathers current far-field spot image F1, carries out pre-service obtain filtered image F2 to image F1;

(2) laser facula Optimal-threshold segmentation algorithm is adopted to obtain image F3 to image F2;

(3) projection algorithm is adopted to image F3, calculate respectively the distance on x direction and y direction between top and secondary top, distance than with the characteristic parameter such as peakedness ratio;

(4) if the distance ratio in x direction is not less than threshold value D _0xor the distance ratio on y direction is not less than threshold value D _0y, then enter coarse tuning process, go to step (5); If the distance ratio in x direction is less than threshold value D _0xand the distance ratio on y direction is less than threshold value D _0y, then step (6) is entered;

(5) if front 2 two field pictures or occur reforming phenomena, motor movement direction is determined according to the spacing variation tendency between hot spot; If not front 2 two field pictures, then according to fixed motor movement direction, adjust grating attitude to the order of corresponding raster detect motor movement according to the spacing between hot spot, then go to step (1); If determining to occur concussion under the condition of motor movement direction, take motor to any one party to moving step sizes N ₁mode walks in advance and redefines motor movement direction, then goes to step (1);

(6) in fine-tuning process, if fine tuning first, then the relative position of first fixed reference hot spot and motion hot spot; If be not less than T at the peakedness ratio in x direction _0xor peakedness ratio is in y-direction not less than threshold value T _0y, according to the spacing between laser facula, (experimentally determine the moving step sizes N of motor to the order of corresponding raster detect motor ₄) carry out grating pose adjustment, then go to step (1), if be less than T at the peakedness ratio in x direction _0xand peakedness ratio is in y-direction less than threshold value T _0y, then raster detect terminates.

The described grating automatic Mosaic algorithm based on far-field spot, the partitioning algorithm described in step (2) is as follows:

$f(x,y)=\left\{\begin{array}{cc}255& (f(x,y)>\mathrm{\λ})\\ 0& f(x,y)\≤\mathrm{\λ}\end{array}\right.$

Wherein: $\mathrm{\λ}=\frac{{\mathrm{\μ}}_1+{\mathrm{\μ}}_2}{2}$

${\mathrm{\μ}}_1=\frac{\underset{i=1}{\overset{t}{\mathrm{\Σ}}}i*N_i}{\underset{i=0}{\overset{t}{\mathrm{\Σ}}}{N}_i},$

${\mathrm{\μ}}_2=\frac{\underset{i=t+1}{\overset{255}{\mathrm{\Σ}}}i*N_i}{\underset{i=t+1}{\overset{255}{\mathrm{\Σ}}}{N}_i}$

$t=\frac{f_{\max }(x,y)+f_{\min }(x,y)}{2}$

N _irepresent i-th grade of number of gray values in histogram, f _max(x, y) is max pixel value in image, f _min(x, y) is minimum pixel value in image.

The described grating automatic Mosaic algorithm based on far-field spot, in described step (3), described distance than computing method is:

D _0y＝D/S

Wherein D be back pitch between top and secondary top from, S is crest distance.In like manner can obtain distance ratio in the x direction.

Peakedness ratio is defined as: peakedness ratio T in y-direction _0y=T ₂/ T ₁, wherein T ₁for peak value, the T at top ₂for the peak value at secondary top; In like manner can in the peakedness ratio definition of T in x direction _0x.

By the far-field spot image obtained, image processing techniques is adopted to carry out Intelligent treatment to hot spot, be extracted in the distance on x and y direction between hot spot, distance than and the characteristic parameter such as peakedness ratio, then according to feature extraction output control motor movement, adjustment grating attitude, finally reaches the object that grating automatic Mosaic is good.

Accompanying drawing explanation

Fig. 1 is algorithm flow schematic diagram of the present invention;

Fig. 2 is that distance of the present invention is than schematic diagram;

Fig. 3 is peakedness ratio schematic diagram of the present invention;

Fig. 4 is grating automatic Mosaic process design sketch of the present invention, a, original state; B, the 20th frame; C, the 30th frame; D, the 50th frame; E, coarse adjustment result; F fine tuning result;

Fig. 5 is laser facula segmentation effect schematic diagram of the present invention, a, former figure; B, segmentation result.

Embodiment

Below in conjunction with specific embodiment, the present invention is described in detail.

With reference to figure 1, the grating automatic Mosaic algorithm based on far-field spot of the present invention comprises the following steps:

(1) CCD gathers current far-field spot image F1, carries out pre-service obtain filtered image F2 to image F1;

(2) adopt laser facula Optimal-threshold segmentation algorithm to obtain image F3 to image F2, partitioning algorithm is as follows:

$f(x,y)=\left\{\begin{array}{cc}255& (f(x,y)>\mathrm{\λ})\\ 0& f(x,y)\≤\mathrm{\λ}\end{array}\right.$

Wherein: $\mathrm{\λ}=\frac{{\mathrm{\μ}}_1+{\mathrm{\μ}}_2}{2}$

${\mathrm{\μ}}_1=\frac{\underset{i=1}{\overset{t}{\mathrm{\Σ}}}i*N_i}{\underset{i=0}{\overset{t}{\mathrm{\Σ}}}{N}_i},$

${\mathrm{\μ}}_2=\frac{\underset{i=t+1}{\overset{255}{\mathrm{\Σ}}}i*N_i}{\underset{i=t+1}{\overset{255}{\mathrm{\Σ}}}{N}_i}$

$t=\frac{f_{\max }(x,y)+f_{\min }(x,y)}{2}$

N _irepresent i-th grade of number of gray values in histogram, f _max(x, y) is max pixel value in image, f _min(x, y) is minimum pixel value in image.

(3) adopt projection algorithm (with reference to Lu Zongqi to image F3, " programming of C/C++ image procossing ", publishing house of Tsing-Hua University, 5.2.1 trifle), calculate respectively the distance on x direction and y direction between top and secondary top, distance than with the characteristic parameter such as peakedness ratio.

Distance than computing method is:

Be illustrated in figure 2 the projection of the strongest two laser faculas on y (vertically) direction, its distance is than orientating as:

D _0y＝D/S

Wherein D be back pitch between top and secondary top from, S is crest distance.In like manner can obtain distance ratio in the x direction.

Peakedness ratio is defined as:

As shown in Figure 3, the peakedness ratio T on y (vertically) direction _0y=T ₂/ T ₁, wherein T ₁for peak value, the T at top ₂for the peak value at secondary top; In like manner can in the peakedness ratio definition of T in x direction _0x.

(4) if the distance ratio in x direction is not less than threshold value D _0xor the distance ratio on y direction is not less than threshold value D _0y, then enter coarse tuning process, go to step (5).If be less than threshold value D at the distance ratio in x direction _0xand the distance ratio on y direction is less than threshold value D _0y, then step (6) is entered.

(5) if front 2 two field pictures or occur reforming phenomena, determine that (first time adjustment allows motor to any one party to motion in motor movement direction according to the spacing variation tendency between hot spot, if the spacing in the 2nd two field picture between hot spot reduces, then determine last motor movement direction be subsequently grating pose adjustment time motor movement direction, if the spacing in the 2nd two field picture between hot spot increases, then determine the reverse direction in last motor movement direction be subsequently grating pose adjustment time motor movement direction of motion, motor movement step-length N ₁experimentally determine).If not front 2 two field pictures, then according to fixed motor movement direction, (experimentally determine the moving step sizes N of motor to the order of corresponding raster detect motor movement according to the spacing between hot spot ₂) adjust grating attitude, then go to step 1; If determining to occur concussion (the spacing variation tendency between hot spot changes) under the condition of motor movement direction, take motor to any one party to moving step sizes N ₁mode walks in advance and redefines motor movement direction, then goes to step (1).

(6) in fine tuning, if fine tuning first, then (motion motor moves certain step number N to a certain fixed-direction to the relative position of first fixed reference hot spot and motion hot spot ₃, N ₃experimentally determine).If be not less than T at the peakedness ratio in x direction _0xor peakedness ratio is in y-direction not less than threshold value T _0y, then according to the spacing between laser facula, (the moving step sizes N of motor is experimentally determined to the order of corresponding raster detect motor ₄) carry out grating pose adjustment, then go to step (1).If be less than T at the peakedness ratio in x direction _0xand peakedness ratio is in y-direction less than threshold value T _0y, then raster detect terminates.

With reference to figure 4 and Fig. 5, it is an embody rule example of the inventive method.

Because the error of optical imaging system exists, CCD has multiple light spot image, as shown in Figure 4; In automatic adjustment grating attitude process, its distance than threshold value is: D _0x=0.1, D _0y=0.15; Peakedness ratio threshold value is: T _0x=0.5, T _0y=0.3.In experiment, moving step sizes gets N ₁=100, N ₂=30; N ₃=50, N ₄=1.

As shown in Figure 5, adopt laser facula Optimal-threshold segmentation, effectively can remove halation, retain laser facula.

Should be understood that, for those of ordinary skills, can be improved according to the above description or convert, and all these improve and convert the protection domain that all should belong to claims of the present invention.

Claims (2)

1., based on a grating method for automatically split-jointing for far-field spot, it is characterized in that, comprise the following steps:

(1) CCD gathers current far-field spot image F1, carries out pre-service obtain filtered image F2 to image F1;

(2) laser facula Optimal-threshold segmentation algorithm is adopted to obtain image F3 to image F2;

(3) projection algorithm is adopted to image F3, calculate the distance on x direction between top and secondary top, distance than and peakedness ratio, and distance on y direction between top and secondary top, distance than and peakedness ratio;

Described distance in y-direction than computing method is:

D _y＝D/S

Wherein D be back pitch between top and secondary top from, S is crest distance; In like manner can obtain distance ratio in the x direction;

Peakedness ratio is defined as: peakedness ratio T in y-direction _y=T ₂/ T ₁, wherein T ₁for peak value, the T at top ₂for the peak value at secondary top; In like manner can in the peakedness ratio definition of T in x direction _x;

(4) if the distance ratio in x direction is not less than threshold value D _0xor the distance ratio on y direction is not less than threshold value D _0y, then enter coarse tuning process, go to step (5); If the distance ratio in x direction is less than threshold value D _0xand the distance ratio on y direction is less than threshold value D _0y, then step (6) is entered;

(5) if front 2 two field pictures or occur reforming phenomena, motor movement direction is determined according to the spacing variation tendency between hot spot; If not front 2 two field pictures, then according to fixed motor movement direction, adjust grating attitude to the order of corresponding raster detect motor movement according to the spacing between hot spot, then go to step (1); If determining to occur concussion under the condition of motor movement direction, take motor to any one party to moving step sizes N ₁mode walks in advance and redefines motor movement direction, then goes to step (1);

(6) in fine-tuning process, if fine tuning first, then the relative position of first fixed reference hot spot and motion hot spot; If be not less than T at the peakedness ratio in x direction _0xor peakedness ratio is in y-direction not less than threshold value T _0y, according to the spacing between laser facula, carry out grating pose adjustment to the order of corresponding raster detect motor, then go to step (1), if be less than T at the peakedness ratio in x direction _0xand peakedness ratio is in y-direction less than threshold value T _0y, then raster detect terminates.

2. the grating method for automatically split-jointing based on far-field spot according to claim 1, is characterized in that, the partitioning algorithm described in step (2) is as follows:

$f(x,y)=\left\{\begin{array}{cc}255& (f(x,y)>\mathrm{\λ})\\ 0& f(x,y)\≤\mathrm{\λ}\end{array}\right.$

Wherein $\mathrm{\λ}=\frac{{\mathrm{\μ}}_1+{\mathrm{\μ}}_2}{2}$

${\mathrm{\μ}}_1=\frac{\underset{i=0}{\overset{t}{\mathrm{\Σ}}}i*N_i}{\underset{i=0}{\overset{t}{\mathrm{\Σ}}}{N}_i},$

${\mathrm{\μ}}_2=\frac{\underset{i=t+1}{\overset{255}{\mathrm{\Σ}}}i*N_i}{\underset{i=t+1}{\overset{255}{\mathrm{\Σ}}}{N}_i},$

$t=\frac{f_{\max }(x,y)+f_{\min }(x,y)}{2}$

N _irepresent i-th grade of number of gray values in histogram, f _max(x, y) is max pixel value in image, f _min(x, y) is minimum pixel value in image.