CN102297761A - Bolt faying face supersonic wave detection apparatus and data processing method thereof - Google Patents

Abstract

The invention discloses a bolt faying face supersonic wave detection apparatus and a data processing method thereof. The apparatus comprises a mobile platform which is composed of a mobile support and a micro-motion measuring mechanism, and is suitable for measurement of any large-scale bolt faying face. An accurate angular displacement bench in the mobile platform can adjust an angle of an ultrasonic transducer accurately, and detection precision and efficiency are raised. According to the data processing method, a problem of image processing difficulty caused by a measuring result when detecting the bolt faying face is overcome, and a problem of an unsatisfactory image restoration result by employing a traditional Wiener filtering method is solved. According to the method, lateral resolution of an ultrasonic c scanning image is greatly raised.

Description

Ultrasound examination bolt joint surface device and data processing method

Technical field

The present invention relates to the direct detection range of a kind of bolt joint surface characteristic, in particular to the method for ultrasound examination bolt joint surface performance and utilize the ultrasonic detection device of this method.

Background technology

The bolt joint surface performance has directly determined the performance of mechanical system, and particularly important bolt joint surface such as the bolt junction of knowing guide rail, can have a direct impact precision optical machinery precision, precision stabilities etc. such as lathes.Therefore, to the faying face performance, comprise that it is very important that faying face contact area, contact stiffness, contact pressure directly detect.In actual production, faying face is detected general the employing be coated with color method, membrane process etc., but these methods all are to add certain measuring media to measure faying face in the middle of faying face.Wherein, be coated with color method owing to mainly be that error ratio is bigger according to the judgement that the master worker of certain experience is arranged.Membrane process is present newer method, but it is to add a kind of new film in the middle of faying face, when between two faying faces interaction being arranged, will produce pressure like this, and film will change.But the present measuring accuracy of this membrane process is not high, and the adding film can change the contact condition of faying face in the middle of faying face, thereby makes experimental result can not reflect actual contact.And this film price is very high, and the film life-span is shorter.Utilizing ultrasonic c scanning that bolt joint surface is detected is the effective ways that directly obtain the faying face contact performance, it can directly detect the contact area of faying face by the ultrasonic reflections rate that detects faying face, obtain faying face rigidity, faying face pressure by updating formula, and then realize automatic detection faying face.Overcome traditional detection method of tinting and judge, can not detect the shortcoming of faying face rigidity, faying face pressure by experience.Ultrasound examination need not change the actual contact situation, overcomes the shortcoming that membrane process influences the faying face contact condition.

For realizing that faying face detects robotization, need to use the immersion type ultrasonic probe.But ultrasonic c scan image is generally fuzzyyer, and lateral resolution is lower, and is more accurate for making measurement result, must handle scan image, but adopt present data processing method, and the result is unsatisfactory.Simultaneously, because the existence of bolt in the bolt joint surface causes the discontinuous of ultrasonic c scan image, influence successive image and handle.And when adopting the immersion type ultrasonic probe to scan, require the upper surface of test specimen in water, to be difficult to satisfy this point in the actual production.The size of measured workpiece is bigger in the actual production simultaneously, and this has also proposed high requirement to ultrasonic detection device.Above-mentioned all multifactor having caused can't be detected faying face by the using ultrasound ripple in the actual production.

Summary of the invention

The object of the present invention is to provide a kind of data processing method that improves ultrasound wave c scan image lateral resolution, and the ultrasonic scanning device that can be used for the faying face detection of actual production.

For achieving the above object, according to ultrasound wave c scan image disposal route provided by the invention, described method comprises:

(1), utilize the immersion type ultrasonic probe, with ultrasonic c scan mode scanning faying face, obtain reflectivity matrix g (x, y), bolt place in the middle of the bolt web joint, in the g matrix with 0 representative;

(2), selecting tested spheroid, sphere diameter is 10 times of ultrasound wave wavelength;

Tested spheroid is put on the hyperacoustic focal plane of immersion type, vertical range is constant between maintenance probe and the spheroid, then in the acoustic beam scope of probe, utilize probe that spheroid is scanned, spheroid is in the reflecting acoustic pressure of every of xsect in the measurement acoustic beam, thereby obtain immersion type probe point spread function h (x, y);

(3), to reflectivity matrix g (x, y) point spread function, h (x, y) cycle of carrying out expands, it expanded be the identical matrix of dimension, and to the reflectivity matrix g after expanding (x, y), point spread function h (x, y) carry out Fast Fourier Transform (FFT), obtain reflectivity matrix Fourier transform G (u, v), point spread function Fourier transform H (u, v);

(4), utilize improved Wiener filtering method to reflectivity matrix g (x y) carries out deconvolution, and concrete steps are as follows:

A, a Wiener filtering:

$\widehat{H_w}(u,v)=\frac{H^{*}(u,v)}{{\left|H(u,v)\right|}^2+\mathrm{\ϵ}}---(3-2)$

$\hat{F}(u,v)={\hat{H}}_w(u,v)G(u,v)---(3-3)$

Formula (3-2) is to seek the best transmission function

(u v) is degenrate function H (u, a complex conjugate v) to H*; Constant ε is an empirical value, and formula 3-3 tries to achieve original image to estimate spectrum

Top formula (3-2) is brought in (3-3), carried out Wiener filtering one time, wherein: (u v) is degenrate function H (u, a complex conjugate v) to H*; Constant ε is an empirical value, and span is: 2.1 * 10 ^-5Between-0.1; Original image is estimated spectrum Carry out inverse Fourier transform, image f is restored ₁(x y), changes constant ε value and relatively recovers image f ₁(x y), finds out optimal constant ε value, determines final recovery image f ₁(x, y);

B, ask image power spectrum and noise mean square value;

According to recovering image f ₁(formula is as follows for x, y) estimating noise mean square value:

${\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="HDA0000068475660000011.png,HDA0000068475660000021.png"\; state="\{\{state\}\}">n</figure-callout>}}^2=\frac{1}{\mathrm{MN}}\mathrm{\&Sigma;}{[g(x,y)-f_1(x,y)*h(x,y)]}^2---(3-4)$

Formula (3-4) is in order to ask image power spectrum and noise mean square value, to ask for recovery image f ₁(x, power spectrum y) adopt following formula to calculate:

$p_f(u,v)=\frac{{\left|\hat{F}(u,v)\right|}^2}{M\&CenterDot;N}---(3-5)$

Formula (3-5) is in order to ask the recovery filter function;

C, secondary Wiener filtering

Blurred picture is carried out the secondary Wiener filtering, and wave filter is as follows:

$\widehat{H_{\mathrm{mw}}}(u,v)=\frac{H^{*}(u,v)}{{\left|H(u,v)\right|}^2+{\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="HDA0000068475660000011.png,HDA0000068475660000021.png"\; state="\{\{state\}\}">n</figure-callout>}}^2/p_f(u,v)}---(3-6)$

${\hat{F}}_m(u,v)={\hat{H}}_{\mathrm{mw}}(u,v)G(u,v)---(3-7)$

Formula (3-6) is to ask the best transmission function with this Wiener filtering method of advancing

Formula (3-7) is to ask original image to estimate spectrum

Formula (3-6) is brought in the formula (3-7), carried out the secondary Wiener filtering, original image is estimated spectrum Carry out Fourier inversion finally recovered image f (x, y).

Described ultrasonic detection device comprises mobile platform, ultrasonic emitting receiving mechanism and data acquisition and processing mechanism three parts composition;

Described mobile platform is made up of traversing carriage and micromotion mechanism, and traversing carriage comprises support, the mobile wheel on the support, the leveling nut on the support; Measurement of micromovements mechanism by three fine motion translation stages, accurate angular displacement platform, transducer liquid soak water injector, connecting rod is formed;

Described ultrasonic emitting receiving mechanism comprises immersion type ultrasonic transducer, ultrasonic pulse generation receiver;

Described data acquisition and processing mechanism comprise: oscillograph, computing machine and software.

Method of the present invention has overcome owing to when detecting bolt joint surface, owing to the Flame Image Process difficult problem that measurement result causes, solved the undesirable problem of image restoration result of the traditional Wiener filtering method of existing employing.Can significantly improve the lateral resolution of ultrasonic c scan image according to the inventive method.

Compare with other ultrasonic measuring devices, mobile platform of the present invention adopts traversing carriage, goes for the measurement of any large-scale bolt joint surface; Accurate angular displacement platform can accurately be adjusted the ultrasonic probe angle in the mobile platform of the present invention, improves accuracy of detection and efficient; Transducer liquid of the present invention soaks water injector to be compared with existing water injector, many one section very short rubber tubes, the easier water cushion that when measuring, forms, make measure convenient.

Description of drawings

Fig. 1 ultrasound examination bolt joint surface device.

Fig. 2 ultrasonic measurement faying face mobile platform.

Fig. 3 liquid soaks water injector.

Embodiment

Ultrasound examination bolt joint surface device:

Fig. 1 is the composition of ultrasound examination faying face device, and 4 is mobile platform among the figure; 2 accept transmitter for ultrasonic pulse among the figure, and 3 is ultrasonic transducer, and 2,3 form the ultrasonic emitting receiving traps; 1,5 are data acquisition and treating apparatus among the figure.Ultrasonic transducer shown in 3 is installed on the mobile platform among the figure, links to each other with 2 pulse transceivers by data line; 1 is oscillograph among the figure, links to each other with 2 by data line; 5 is computing machine among the figure, links to each other with oscillograph by netting twine, utilizes software labview to realize the collection and the storage of data.

Fig. 2 is a mobile platform, and the ultrasonic measurement faying face moves translation stage and mainly contains traversing carriage and measurement of micromovements device composition, and traversing carriage mainly comprises moving takes turns 9, and leveling nut 10 and support 7 are formed.Move wheel 9 and be installed on the support 7 by crossbeam, leveling nut 10 links to each other with threaded hole on the support 7.The measurement of micromovements device mainly comprises x axle fine motion translation stage 5, y axle translation stage line slideway 2, y axle translation stage ball-screw 3, y axle servomotor 15, rotation angle displacement platform 4, z axle fine motion translation stage 1, extension bar 14, universal intersection frame 13, liquid bath unit 11 compositions.Wherein x axle translation stage 5 is fixed on the support 7 by bolt, nut, rotation angle position platform 7 is installed on x axle translation stage 5 slide blocks, y axle fine motion translation stage is installed on the platform moving slider of rotation angle position, position, z axle fine motion angle platform 1 docks installation with y axle fine motion translation stage slide block, extension bar 14 is installed on the web joint of z axle fine motion translation stage 1 below, simultaneously, fixing universal intersection frame 13 on the other end of extension bar 14, ultrasonic probe 12 is being fixed in universal intersection frame 13 another holes, and liquid bath unit 11 is installed on the axle of universal intersection frame 13.

Fig. 3 is the liquid bath unit, and 1 is funnel among the figure, and 2 is rubber tube, and 3 is measured workpiece.Compare with the liquid bath unit of existing liquid-spraying type, added rubber tube among the present invention, after water flowing, make its easier below the stroke water cushion.

Measuring method:

After whole measuring system installs by diagram, send pulse by computer control pulse generation receiver (among Fig. 1 2) control ultrasonic probe (among Fig. 1 3), pulsating wave carries out transmission at workpiece then, reflection obtains reflection wave, and after handling through computer control pulse generation receiver (among Fig. 1 2), be transferred to oscillograph (among Fig. 1 1), and finally desired data is gathered out by oscillograph (among Fig. 1 1), deliver to computing machine (among Fig. 1 5) again, and adopt the method applied in the present invention to carry out data processing, obtain corresponding faying face contact area.

The course of work of this device is: after installing according to diagram, before carrying out the faying face measurement, earlier the leveling nut is adjusted to certain altitude, make mobile wheel land fully, by following mobile wheel transfer table is moved to measured workpiece top (using mobile wheel conveniently to move to somewhere) then, adjusting the leveling nut then makes after it is elongated to a certain degree, mobile wheel can be unsettled, so just fixing transfer table can not impact the result because of the resetting error.Then, in soaking into funnel, begin to inject certain liquid, wherein the speed of Zhu Ruing is greater than the speed that flows out from hopper base, so just can in funnel, form stable liquid level with certain altitude, do infiltrate if adopt lubricating oil, oil circulating system can be considered to add again, thereby resource can be saved.Adjust the focusing probe position by mobile xyz fine motion translation stage then,, can measure by rotation angle position platform if ultrasonic probe is not a level.Adjust after the focal length, just can carry out ultrasonic measurement, image data.

The scan-data disposal route:

The data processing method specific implementation process is as follows:

At first carry out data acquisition, obtain preliminary reflectivity matrix:

Mainly comprise following three parts:

1, obtain reference signal: when faying face does not contact, utilize the immersion type ultrasonic probe that the faying face upper plate is scanned, ultrasonic signal utilizes oscillograph to receive and preserves (oscillograph bandwidth must more than 200MHz), obtains first group of file.

2, keep upper plate motionless, allow faying face contact, and apply certain pretightning force.Track according to scanning for the first time rescans faying face, obtains second group of file.

3, at each point on the faying face, secondary reflection ripple in the corresponding data in second group and the first group of file is carried out Fast Fourier Transform (FFT) and is divided by, (x y) (to bolt place in the middle of the bolt web joint, represents with 0 in the G matrix) to obtain reflectivity matrix g.

Carry out the preliminary Wiener filtering of data then:

Concrete enforcement is as follows: the spatial domain expression formula of degeneration system is:

g(x，y)＝h(x，y)*f(x，y)+n(x，y) (2-1)

(x is that the spatial domain of degenrate function represents that * represents deconvolution y) to h;

(x y) belongs to noise matrix to n

4, for conveniently handling, must (x, y), h (x, y) expand by the cycle of carrying out to reflectivity matrix function g.Suppose former digital picture g (x y) is of a size of A * B, point spread function h (x, size y) is C * D, then all will expand to size is the cycle data of M * N, wherein:

M＝A+C-1，N＝B+D-1。

5, to the g after expanding (x, y), h (x y) carries out Fast Fourier Transform (FFT), obtain G (u, v), H (u, v).

6, utilize improved Wiener filtering method to g (x y) carries out deconvolution, and concrete steps are as follows:

(1) Wiener filtering:

$\widehat{H_w}(u,v)=\frac{H^{*}(u,v)}{{\left|H(u,v)\right|}^2+\mathrm{\&epsiv;}}---(2-2)$

$\hat{F}(u,v)={\hat{H}}_w(u,v)G(u,v)---(2-3)$

Formula (2-2) is to seek the best transmission function

(u v) is degenrate function H (u, a complex conjugate v) to H*; Constant ε is an empirical value, and formula (2-3) is to try to achieve original image to estimate spectrum (u v) brings top formula (2-2) in (2-3) into reflectivity Fourier transform G, carries out Wiener filtering one time, and wherein: (u v) is degenrate function H (u, a complex conjugate v) to H*; Constant ε is an empirical value, and span is: 2.1 * 10 ^-5Between-0.1; Original image is estimated spectrum

Carry out inverse Fourier transform, image f is restored ₁(x, y) (this recovers image effect and bad).Change constant ε value and relatively recover image f ₁(x y), finds out optimal constant ε value.

Use improved Wiener filtering at last, carry out data processing.

Mainly contain following two steps:

(1) asks image power spectrum and noise mean square value

According to recovering image f ₁(formula is as follows for x, y) estimating noise mean square value:

${\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="HDA0000068475660000011.png,HDA0000068475660000021.png"\; state="\{\{state\}\}">n</figure-callout>}}^2=\frac{1}{M\&CenterDot;N}\mathrm{\&Sigma;}{[g(x,y)-f_1(x,y)*h(x,y)]}^2---(2-4)$

Formula (2-4) is in order to ask image power spectrum and noise mean square value, to ask for recovery image f ₁(x, power spectrum y) adopt period map method to calculate, and adopt following formula to calculate:

$p_f(u,v)=\frac{{\left|\hat{F}(u,v)\right|}^2}{M\&CenterDot;N}---(2-5)$

Formula (2-5) is in order to ask the recovery filter function.

(2) improve Wiener filtering

Only adopt Wiener filtering one time, the result that can obtain being satisfied in the main in less demanding occasion, but this method fails to make full use of the information of image and noise itself, so be difficult to obtain high-quality image restoration effect, so blurred picture is carried out the secondary Wiener filtering, wave filter is as follows:

$\widehat{H_{\mathrm{mw}}}(u,v)=\frac{H^{*}(u,v)}{{\left|H(u,v)\right|}^2+{\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="HDA0000068475660000011.png,HDA0000068475660000021.png"\; state="\{\{state\}\}">n</figure-callout>}}^2/p_f(u,v)}---(2-6)$

${\hat{F}}_m(u,v)={\hat{H}}_{\mathrm{mw}}(u,v)G(u,v)---(2-7)$

Formula (2-6) is to ask the best transmission function with this Wiener filtering method of advancing

Formula (2-7) is to ask original image to estimate spectrum Formula (2-6) is brought in the formula (2-7), carried out the secondary Wiener filtering.Original image is estimated spectrum

Carry out Fourier inversion finally recovered image f (x, y).

Experiment showed, that this method is respond well.

Claims (2)

1. ultrasound examination bolt joint surface data processing method is characterized in that, described method comprises:

(1), utilize the immersion type ultrasonic probe, with ultrasonic c scan mode scanning faying face, obtain reflectivity matrix g (x, y), bolt place in the middle of the bolt web joint, in the g matrix with 0 representative;

(2), selecting tested spheroid, sphere diameter is 10 times of ultrasound wave wavelength;

Tested spheroid is put on the hyperacoustic focal plane of immersion type, vertical range is constant between maintenance probe and the spheroid, then in the acoustic beam scope of probe, utilize probe that spheroid is scanned, spheroid is in the reflecting acoustic pressure of every of xsect in the measurement acoustic beam, thereby obtain immersion type probe point spread function h (x, y);

(3), to reflectivity matrix g (x, y) point spread function, h (x, y) cycle of carrying out expands, it expanded be the identical matrix of dimension, and to the reflectivity matrix g after expanding (x, y), point spread function h (x, y) carry out Fast Fourier Transform (FFT), obtain reflectivity matrix Fourier transform G (u, v), point spread function Fourier transform H (u, v);

(4), utilize improved Wiener filtering method to reflectivity matrix g (x y) carries out deconvolution, and concrete steps are as follows:

A, a Wiener filtering:

$\widehat{H_w}(u,v)=\frac{H^{*}(u,v)}{{\left|H(u,v)\right|}^2+\mathrm{\&epsiv;}}---(3-2)$

$\hat{F}(u,v)={\hat{H}}_w(u,v)G(u,v)---(3-3)$

Formula (3-2) is to seek the best transmission function (u v) is degenrate function H (u, a complex conjugate v) to H*; Constant ε is an empirical value, and formula 3-3 tries to achieve original image to estimate spectrum

Top formula (3-2) is brought in (3-3), carried out Wiener filtering one time, wherein: (u v) is degenrate function H (u, a complex conjugate v) to H*; Constant ε is an empirical value, and span is: 2.1 * 10 ^-5Between-0.1; Original image is estimated spectrum

Carry out inverse Fourier transform, image f is restored ₁(x y), changes constant ε value and relatively recovers image f ₁(x y), finds out optimal constant ε value, determines final recovery image f ₁(x, y);

B, ask image power spectrum and noise mean square value;

According to recovering image f ₁(formula is as follows for x, y) estimating noise mean square value:

${\mathrm{\&sigma;}}_n^2=\frac{1}{\mathrm{MN}}\mathrm{\&Sigma;}{[g(x,y)-f_1(x,y)*h(x,y)]}^2---(3-4)$

Formula (3-4) is in order to ask image power spectrum and noise mean square value, to ask for recovery image f ₁(x, power spectrum y) adopt following formula to calculate:

$p_f(u,v)=\frac{{\left|\hat{F}(u,v)\right|}^2}{M\&CenterDot;N}---(3-5)$

Formula (3-5) is in order to ask the recovery filter function;

C, secondary Wiener filtering

Blurred picture is carried out the secondary Wiener filtering, and wave filter is as follows:

$\widehat{H_{\mathrm{mw}}}(u,v)=\frac{H^{*}(u,v)}{{\left|H(u,v)\right|}^2+{\mathrm{\&sigma;}}_n^2/p_f(u,v)}---(3-6)$

${\hat{F}}_m(u,v)={\hat{H}}_{\mathrm{mw}}(u,v)G(u,v)---(3-7)$

Formula (3-6) is to ask the best transmission function with this Wiener filtering method of advancing

Formula (3-7) is to ask original image to estimate spectrum

Formula (3-6) is brought in the formula (3-7), carried out the secondary Wiener filtering, original image is estimated spectrum

Carry out Fourier inversion finally recovered image f (x, y).

2. a ultrasound examination bolt joint surface device is characterized in that, described ultrasonic detection device comprises mobile platform, ultrasonic emitting receiving mechanism and data acquisition and processing mechanism three parts composition;

Described mobile platform is made up of traversing carriage and micromotion mechanism, and traversing carriage comprises support, the mobile wheel on the support, the leveling nut on the support; Measurement of micromovements mechanism by three fine motion translation stages, accurate angular displacement platform, transducer liquid soak water injector, connecting rod is formed;

Described ultrasonic emitting receiving mechanism comprises immersion type ultrasonic transducer, ultrasonic pulse generation receiver;

Described data acquisition and processing mechanism comprise: oscillograph, computing machine and software.

Images