CN104490365B - Method for rebuilding broad-band photo-acoustic image by using narrow-band sensor - Google Patents

Abstract

The invention discloses a method for rebuilding a broad-band photoacoustic image by using a narrow-band sensor. The method comprises the following steps: repeatedly illuminating a target by pulsed light; producing PA (photoacoustic) signals after the target absorbs the light; and using the narrow-band sensor to receiving multiple groups of PA signals. While the target is a high scattering medium, the PA signals are greatly attenuated before reaching the sensor, so the narrow-band PA signals received by the sensor are weak and the signal-to-noise ratio is very low, multiple groups of narrow-band PA signals need to be pre-processed, and the signal-to-noise ratio of the narrow-band PA signals is improved. Because the PA signal spectrum is distributed in 1-50 MHz, the PA signals in different frequency bands respond the different characteristics of the illuminated target, the actually used sensors are the narrow-band, so the received PA signals are narrow-band. In order to receive the board-band PA signals, multiple narrow-band sensors in the different frequency bands are used necessarily, but the actual operation is inconvenient.

Description

A kind of method using narrow sensor to rebuild broadband light acoustic image

Technical field

The present invention relates to Computer Image Processing field, particularly a kind of use narrow sensor rebuilds broadband light acoustic image Method.

Background technology

Photoacoustic imaging is a kind of noninvasive imaging technology, and the advantage combining optical imagery and acoustics imaging penetrates deeper, because of This is medically being widely used, such as, detect osteoporosis.General step is: pulsed light is radiated in target, target absorption light Produce PA signal, use sensor to receive PA (Photoacoustic optoacoustic) photoacoustic signal, measure PA signal reconstruction light sound spectrogram Picture.PA signal frequency is distributed in 1-50MHz, the different qualities of the illuminated target of PA signal reaction of different frequency bands, actually used Sensor mostly be narrow sensor, therefore want to absorb whole PA signals and to use the arrowband of multiple different-waveband to sense Device, it has not been convenient to the most unrealistic.

Summary of the invention

Goal of the invention: the technical problem to be solved is for the deficiencies in the prior art, it is provided that a kind of use is narrow Belt sensor rebuilds the method for broadband light acoustic image, the arrowband absorbed sensor according to the amplitude-frequency characteristic of narrow sensor PA (Photoacoustic) signal restoring becomes broadband signal, it is thus achieved that more target information amount, goes out with broadband PA signal reconstruction Photoacoustic image, resolution is high, and quality is good.

In order to solve above-mentioned technical problem, the invention discloses and a kind of use narrow sensor to rebuild broadband light acoustic image Method, comprises the following steps:

Step one, pulsed light is radiated in target repeatedly, and target absorption light produces PA (Photoacoustic) signal；

Step 2, uses narrow sensor to receive the PA signal organizing arrowband more；

Step 3, to the many groups arrowband PA Signal Pretreatment received, obtains high-quality arrowband PA signal；

Step 4, according to the amplitude-frequency characteristic of narrow sensor, carries out lost part frequency to pretreated arrowband PA signal The restoration disposal of spectrum, thus obtain the PA signal in broadband, it is thus achieved that more target information；

Step 5, broadband PA signal reconstruction photoacoustic image step 4 obtained by time delay summation.

The present invention only use single narrow sensor carry out data acquisition, due to the amplitude-frequency characteristic of single narrow sensor Identical, the PA signal collected is narrow band signal, and the target information amount comprised is limited, it is impossible to reconstruct preferable light sound spectrogram Picture.

In the present invention, it is preferable that in step one, target Stimulated Light irradiates generation optoacoustic effect repeatedly, meets equation:

${\▿}^{2}p(r,t)-\frac{1}{c^2}\frac{{\∂}^{2}p(r,t)}{{\∂t}^2}=-\frac{\mathrm{\β}}{C_p}\frac{\∂}{\∂t}H(r,t)---\left(1\right)$

Wherein r is the coordinate vector of target, and t is the time, and (r, t) is acoustic pressure to p, and (r is t) that incident laser is at imaging region to H The heat source function excited, (r, t)=A (r) I (t), A (r) are that the light of tissue absorbs distribution to H, and I (t) is for irradiating light intensity, and β is that heat is swollen Swollen coefficient, C_pFor specific heat capacity, c is the experience velocity of sound, and laser gets to target and sensor starts to receive signal and almost carries out simultaneously.

In the present invention, it is preferable that the sensor employed in step 2 is that the one of the same amplitude-frequency characteristic of linear array is Row narrow sensor, there are sound wave lens narrow sensor front end；

The arrowband PA signal that sensor absorbs with the relation of former PA signal is:

G (x, y)=f (x, y) * h (x, y)+η (x, y) (2)

((x, y) original PA signal, (x y) is narrow sensor to h to f to g for x, the arrowband PA signal that y) sensor receives Degenrate function, η (x, y) noise signal.

Sound wave lens in step 2 make the ultrasound wave produced through optoacoustic effect have certain convergence effect.Due to arrowband The amplitude-frequency characteristic of sensor, the signal received is all arrowband PA signal.But actually target is high scattering material, PA is made to believe Number the most significantly decay arriving before sensor, add the noise of system, the arrowband PA signal that final sensor receives The faintest, signal to noise ratio is the lowest, therefore to receive many groups, carry out subsequent treatment.

In the present invention, it is preferable that described pretreatment, it is each group arrowband PA weak output signal and letter received for sensor Make an uproar the low problem of ratio, utilize the time interval of each group arrowband PA signal that sensor receives to have small time delay to process.Small Each group arrowband PA signal that time delay causes cannot align, and utilizes the cross correlation between each group of arrowband PA signal to estimate and prolongs Time, the starting point of each group of arrowband PA signal to be alignd, is overlapped, interpolation processing, the sample value of the arrowband PA signal after process becomes Obtain intensive.The meaning of pretreatment is, removes noise, improves the quality of arrowband PA signal.

Specifically include:

Utilize two the signal p received₁(t)、p₂T () cross-correlation function estimates time delay t_s；

If p₁(t) and p₂T () is continuous signal, then cross-correlation function is:

$g\left(t\right)=p_1\left(t\right){*p}_1\left(t\right)=\underset{-\∞}{\overset{+\∞}{\∫}}{p}_1^{*}\left(\mathrm{\τ}\right)p_2(t+\mathrm{\τ})\mathrm{d\τ}---\left(3\right)$

It is p₁The complex conjugate of (t),For the convolution algorithm of continuous time, g (t) peak value institute is right The time coordinate answered is time delay t_s；

If p₁(t) and p₂T () is discrete signal, then cross-correlation function is:

${\hat{R}}_{p_1{p}_2}\left(m\right)=\left\{\begin{array}{cc}\frac{1}{L-\left|m\right|}\underset{s=0}{\overset{L-m-1}{\mathrm{\&Sigma;}}}{p}_1(s+m)p_2\left(s\right)& m\&GreaterEqual;0\\ \widehat{R_{p_1{p}_2}}(-m)& m<0\end{array}\right.---\left(4\right)$

L is signal p₁(t) and signal p₂The greatest length of (t), m is discrete sampling time point,For The convolution algorithm of discrete time,Time coordinate corresponding to peak value is signal p₁(t) and signal p₂T () exists Time accurate time delay t_s, f_sampleIt it is the sample frequency of reception system；

The arrowband PA signal received is alignd, carries out the arrowband PA signal after interpolation processingFor:

$\hat{p}(r,L_{\mathrm{int}})=\frac{1}{N_s}\underset{i=1}{\overset{N_s}{\mathrm{\&Sigma;}}}{p}_{i\_\mathrm{shift}}(r,L_{\mathrm{int}})---\left(5\right)$

L_intIt is the interpolation length of arrowband PA signal, N_sIt is the sampling frame number of arrowband PA signal, i value 1～N_s, p_{i_shift} (r,L_int) be displacement after arrowband PA signal, finally, N group signal after treatment, the PA signal obtained

$\hat{p}(r,n)=\hat{p}(r,L_{\mathrm{int}}){|}_{L_{\mathrm{int}}=\mathrm{nN}}---\left(6\right)$

N is discrete sampling time point.

In the present invention, it is preferable that described restoration disposal, be the target information amount contained for arrowband PA signal packet very little and The problem that cannot reconstruct desired light acoustic image, reverts to the PA signal of arrowband by the method for liftering the PA signal in broadband, Obtain more information about target.Have after pretreatment with reception sensor degradation function and two aspects of noise statistics feature The arrowband PA signal of better quality processes, and recovers the frequency spectrum of lost part, thus obtains the PA signal in broadband, it is thus achieved that more Many target information amount, prepare for reconstructing satisfied photoacoustic image.

In the present invention, it is preferable that described method for reconstructing is time delay summation algorithm.In optoacoustic is rebuild, time delay summation is calculated Method and Inverse Projection application are more, and backprojection reconstruction algorithm is that the time inverse to photoacoustic signal is weighted suing for peace, and time delay Summation algorithm is directly to sue for peace photoacoustic signal.Compared to backprojection reconstruction algorithm, low frequency and high frequency are believed by time delay summation algorithm Number having weighted sum, contain enough low-frequency information, the image of reconstruction is relatively smooth.

In the present invention, it is preferable that described in be reconstituted in pretreatment and restoration disposal after.Pretreated high-quality arrowband PA Signal, much noise is removed, the broadband PA signal after restoration disposal, and the target information amount comprised significantly promotes so that weight The photoacoustic image founded a capital out and increases significantly in picture quality in resolution.

Specifically, the process of reconstruction of image is the inverse process of optoacoustic effect described in formula (1), on an array of sensors The sound pressure signal detected is:

$p(r_0,t)={\mathrm{\&eta;}\&Integral;\&Integral;\&Integral;d}^3\mathrm{rA}\left(r\right)\frac{{\mathrm{\&delta;}}^{\&prime;}(t-\frac{|r_0-r|}{c})}{4\mathrm{\&pi;}|r_0-r|}---\left(7\right)$

Wherein η=β/C_p, r₀It is the coordinate vector of sensor, | r₀-r | being the reconstruction regions distance that arrives sensor, δ ' is sharp The derived function of the delta-function that light pulse produces, the inverse process of said process is i.e. process of reconstruction, and method for reconstructing is counted according to formula (8) Calculate:

For the effective aperture of sensor, when linear sensor, this absorption distribution approximation prolonging by formula (9) Shi Qiuhe algorithm is obtained:

$A\left(r\right)=\frac{{\mathrm{\&Sigma;}}_{k}w(k,r)p_k(r,t+T(k,r))}{{\mathrm{\&Sigma;}}_{k}w(k,r)}---\left(9\right)$

Wherein k represents sensor, and (k, r) represents weight to w, and (k is r) to the time of sensor k, p from position r to T_kFor passing Ultrasonic signal sequence received by sensor k；

Postpone summation to low frequency and high-frequency signal all weighted sums, the image smoothing obtained.

The present invention carries out the restoration disposal of lost part frequency spectrum to pretreated arrowband PA signal, recovers the PA in broadband Signal.Broadband light acoustic image is rebuild subsequently based on the PA signal of this broadband.A kind of use narrow sensor disclosed by the invention The method rebuilding broadband light acoustic image, a kind of method using narrow sensor to rebuild broadband light acoustic image disclosed by the invention, Not only can improve imaging resolution, and image quality can be effectively improved, widen the range of application of photoacoustic imaging.

Accompanying drawing explanation

With detailed description of the invention the present invention done below in conjunction with the accompanying drawings and further illustrates, the present invention above-mentioned and Otherwise advantage will become apparent.

Fig. 1 is the schematic diagram that sensor receives many group arrowbands PA signal.

Fig. 2 is present system schematic diagram.

Fig. 3 is flow chart of the present invention.

Detailed description of the invention

As it is shown on figure 3, the invention discloses a kind of use narrow sensor rebuild broadband light acoustic image method, including with Lower step:

Step one, pulsed light is radiated in target repeatedly, and target absorption light produces PA (Photoacoustic) signal；

Step 2, uses narrow sensor to receive the PA signal organizing arrowband more；

Step 3, to the many groups arrowband PA Signal Pretreatment received, obtains high-quality arrowband PA signal；

Step 4, according to the amplitude-frequency characteristic of narrow sensor, carries out lost part frequency to pretreated arrowband PA signal The restoration disposal of spectrum, thus obtain the PA signal in broadband, it is thus achieved that more target information；

Step 5, broadband PA signal reconstruction photoacoustic image step 4 obtained by time delay summation.

As in figure 2 it is shown, on the one hand functional generator sends laser trigger signal controls the laser output system of fixed wave length Output pulsed light, on the one hand sending signal informs gallery job initiation.Pulsed light, by fiber-optic illuminated on sample, occurs Optoacoustic effect, produces sound pressure signal.Sonac receives sound pressure signal, passes to gallery, and gallery is programmed, For computer work.

In the present invention, step one, when sharp target illuminated, produce optoacoustic effect.Target absorption light makes therein temperature Change thus cause some regional structure and change in volume, when using laser pulse repeatedly to irradiate target, target inside temperature Lifting can cause the harmomegathus of target volume, thus to external radiation PA signal.

In the present invention, step 2, narrow sensor is placed on the position parallel with target, receives PA signal, due to target It is the medium of high scattering, the most significantly decayed before PA signal arrives sensor, add the noise of system so that sensor connects The PA signal received is the faintest, and signal to noise ratio is low, thus to receive several groups more, prepares for follow-up pretreatment.Due to narrow The spectral characteristic of belt sensor self, can only receive narrow-band PA signal, and other frequency ranges are filtered, and is therefore narrow band signal, Quantity of information is limited.The relation of the arrowband PA signal that sensor absorbs and former PA signal is from (1)

G (x, y)=f (x, y) * h (x, y)+η (x, y) (1)

((x, y) original PA signal, (x y) is narrow sensor to h to f to g for x, the arrowband PA signal that y) sensor receives Degenrate function, η (x, y) noise signal.Be can be seen that by formula (1), due to the amplitude-frequency characteristic of narrow sensor, the signal received For arrowband PA signal.

In the present invention, by the reception of Fig. 1 it can be seen that the uncertainty of system, result in depositing of tiny time delay ?.The time interval of laser pulse very accurately (max{ | T₁-T₂| < 1ns), there's almost no time delay, but the connecing of sensor Receive time delay in time interval (max{ | T₃-T₄| > 0.1 μ s) can not ignore.Therefore the many groups arrowband PA signal received Cannot align.Utilize the cross correlation of each group of arrowband PA signal, time delay can be estimated, utilize the time delay obtained by each Group signal alignment, and be overlapped, interpolation operation, it is clear that make the sample value of arrowband PA signal become intensive, this operation, can have Effect removes noise, improves signal quality.The basic thought of correlation estimation delay algorithm is to utilize two signals received p₁(t)、p₂T () cross-correlation function estimates time delay.For continuous signal p₁(t) and p₂(t), cross-correlation function is:

$g\left(t\right)=p_1\left(t\right){*p}_1\left(t\right)=\underset{-\&infin;}{\overset{+\&infin;}{\&Integral;}}{p}_1^{*}\left(\mathrm{\&tau;}\right)p_2(t+\mathrm{\&tau;})\mathrm{d\&tau;}---\left(2\right)$

It is p₁The complex conjugate of (t), t is the time,Convolution algorithm for continuous time.According to Cross-correlation function, the time coordinate corresponding to g (t) peak value is time delay t_s.For discrete signal p₁(t) and p₂(t), Cross-correlation function is:

${\hat{R}}_{p_1{p}_2}\left(m\right)=\left\{\begin{array}{cc}\frac{1}{L-\left|m\right|}\underset{s=0}{\overset{L-m-1}{\mathrm{\&Sigma;}}}{p}_1(s+m)p_2\left(s\right)& m\&GreaterEqual;0\\ \widehat{R_{p_1{p}_2}}(-m)& m<0\end{array}\right.---\left(3\right)$

L is signal p₁(t) and signal p₂The greatest length of (t), m is discrete sampling time point,For The convolution algorithm of discrete time, but can only obtain signal p by formula (3)₁(t) and signal p₂T () existsIt is Reception system sample frequency) time accurate time delay t_sI.e.Time coordinate corresponding to peak value.

Utilizing the arrowband PA signal alignment that time delay makes to receive, be overlapped, interpolation, the arrowband PA signal after process is:

$\hat{p}(r,L_{\mathrm{int}})=\frac{1}{N_s}\underset{i=1}{\overset{N_s}{\mathrm{\&Sigma;}}}{p}_{i\_\mathrm{shift}}(r,L_{\mathrm{int}})---\left(4\right)$

L_intBeing the interpolation length of arrowband PA signal, r is the coordinate vector of target, N_sIt is the sampling frame number of arrowband PA signal, I value 1～N_s, p_{i_shift}(r,L_int) it is the arrowband PA signal after shifting.Finally, N group signal after treatment, the PA signal obtained

$\hat{p}(r,n)=\hat{p}(r,L_{\mathrm{int}}){|}_{L_{\mathrm{int}}=\mathrm{nN}}---\left(5\right)$

N is discrete sampling time point.

In the present invention, step 4, if with arrowband PA signal reconstruction photoacoustic image, obtain is fuzzy and the many figure of noise Picture, it is impossible to satisfactory.The fuzzy frequency component information deficiency, especially high fdrequency component of being because is lost, the method for available liftering Frequency information narrow sensor filtered out recovers, liftering be signal is carried out based on sensor pulse response anti- Convolution, can be close to primary signal, and in the method, the signal after liftering is close to the broadband signal of original PA signal, so And directly liftering is in the case of having noise jamming, recovery effect is undesirable, it is known that noise is to neglect in the entire system Slightly.

Therefore the method that this example uses Wiener filtering, on the one hand Wiener filtering considers degenrate function, on the one hand considers to make an uproar Sound system, compares direct liftering, can recover the broadband PA signal of better quality, and liftering is the special case of Wiener filtering.Dimension The broadband PA signal that nanofiltration ripple recovers, signal to noise ratio significantly improves, and the PA before the quantity of information comprised filters closer to sensor believes Number, carry out photoacoustic image reconstruction with the PA signal that this recovers, high-resolution, high-quality image can be obtained.Wiener filtering Target be to make the signal that recovers minimum with original signal mean square error.The Fourier transformation of the signal recovered obtained by under Formula (6) is given:

$\hat{F}(u,v)=\left[\frac{1}{H(u,v)}\frac{{\left|H(u,v)\right|}^2}{{\left|H(u,v)\right|}^2+K}\right]G(u,v)---\left(6\right)$

(u, v) for the conversion of transmission function, H for H^*(u v) is H (u, complex conjugate v)；

(u v) is the conversion of the image that step 3 obtains to G；

It it is the conversion recovering signal；

|H(u,v)|²=H^*(u,v)H(u,v)；

K is a constant；

In this example, Wiener filtering is that signal carries out deconvolution based on sensor amplitude-frequency characteristic, therefore wiener filter The transmission function of ripple device is exactly the transmission function of narrow sensor, and K takes 10dB.Noise factor is white noise, and white noise is respectively The noise that in individual frequency range, power is equal, because the bandwidth of noise is much larger than the frequency of the narrow band signal that narrow sensor receives Bandwidth, so processing system noise as white noise.With these two parameters, signal is carried out Wiener filtering through Wiener filtering Signal after process is the broadband signal that noise significantly reduces, and the target information amount that broadband signal is comprised is compared before treatment Signal is abundanter, can reconstruct gratifying high-resolution, high-quality photoacoustic image based on this.

In the present invention, step 5, the signal reconstruction photoacoustic image recovered by step Wiener filtering.This example have employed and prolongs Summation rebuilds photoacoustic image late.The process of reconstruction of image is the inverse process of optoacoustic effect, is detected on an array of sensors Sound pressure signal is

$p(r_0,t)={\mathrm{\&eta;}\&Integral;\&Integral;\&Integral;d}^3\mathrm{rA}\left(r\right)\frac{{\mathrm{\&delta;}}^{\&prime;}(t-\frac{|r_0-r|}{c})}{4\mathrm{\&pi;}|r_0-r|}---\left(7\right)$

Wherein η=β/C_p, β is thermal coefficient of expansion, C_pBeing specific heat capacity, A (r) is that reconstruction regions light absorbs distribution, r₀It it is sensing The coordinate vector of device, | r₀-r | being the reconstruction regions distance that arrives sensor, δ ' is the derived function of the delta-function that laser pulse produces. The inverse process of said process is i.e. process of reconstruction, and method for reconstructing is given by equation below

For the effective aperture of sensor, c is the experience velocity of sound being generally used to rebuild, value 1.48mm/ μ s in laboratory. When linear sensor, this absorption distribution can approximate to be obtained by time delay summation algorithm

$A\left(r\right)=\frac{{\mathrm{\&Sigma;}}_{k}w(k,r)p_k(r,t+T(k,r))}{{\mathrm{\&Sigma;}}_{k}w(k,r)}---\left(9\right)$

Wherein k represents sensor, and (k, r) represents weight to w, and (k is r) to the time of sensor k, p from position r to T_kFor passing Ultrasonic signal sequence received by sensor k.

Delay summation, to low frequency and high-frequency signal all weighted sums, the image smoothing obtained, can reveal that the interior of image Portion's structure.

In the present invention, whole flow process, comprises the steps:

Step one, pulsed light is radiated in target repeatedly, and target absorption light produces PA (Photoacoustic) signal；

Step 2, uses narrow sensor to receive the PA signal organizing arrowband more；

Step 3, to the many groups arrowband PA Signal Pretreatment received, obtains high-quality arrowband PA signal；

Step 4, according to the amplitude-frequency characteristic of narrow sensor, carries out lost part frequency to pretreated arrowband PA signal The restoration disposal of spectrum, thus obtain the PA signal in broadband, it is thus achieved that more target information；

Step 5, broadband PA signal reconstruction photoacoustic image step 4 obtained by time delay summation.

In whole flow process, step one, pulsed light irradiates target repeatedly, and the gradient of temperature of target can cause the harmomegathus of volume, Giving off PA signal, PA signal contains abundant target internal information, and frequency spectrum is at 1-50MHz.

In whole flow process, step 2, due to the amplitude-frequency characteristic of narrow sensor, PA signal filtering can only can be absorbed portion Frequency-division section, system noise can not be ignored, and sees formula (1), independent one group of weak output signal and poor signal to noise, and receiving many groups is later Process is prepared.

In whole flow process, step 3, the many groups PA signal absorbing sensor carries out pretreatment, utilizes sensor to have and prolong Time, according to the cross correlation of each group of signal, see formula (3), estimate concrete time delay t_s, then by the starting point pair of each group of signal Together, being overlapped, interpolation, the sample value of signal becomes intensive, sees that formula (4), (5) do so eliminate noise, improves signal Quality.

In whole flow process, step 4, pretreated signal Wiener filtering is restored, utilizes formula (6), recover Go out broadband signal, make the signal recovered closer to primary signal, comprise the target information of more horn of plenty.

In whole flow process, step 5, the time delay summation of the signal after restoring is rebuild photoacoustic image, available smooth Image, and the internal structure of image can be disclosed.Owing to step 4 has recovered broadband signal, the image after reconstruction and original image More closely, be a panel height resolution, the high quality graphic of high-contrast.

The invention provides a kind of method using narrow sensor to rebuild broadband light acoustic image, implement this technical side The method of case and approach are a lot, and the above is only the preferred embodiment of the present invention, it is noted that for the art For those of ordinary skill, under the premise without departing from the principles of the invention, it is also possible to make some improvements and modifications, these improve Also protection scope of the present invention is should be regarded as with retouching.Each ingredient the clearest and the most definite in the present embodiment all can use prior art in addition Realize.

Claims (4)

1. one kind uses the method that narrow sensor rebuilds broadband light acoustic image, it is characterised in that comprise the following steps:

Step one, pulsed light is radiated in target repeatedly, and target absorption light produces PA signal；

Step 2, uses narrow sensor to receive the PA signal of the N group arrowband that target produces；

Step 3, to the many groups arrowband PA Signal Pretreatment received, obtains high-quality arrowband PA signal；

Step 4, according to the amplitude-frequency characteristic of narrow sensor, carries out lost part frequency spectrum to pretreated arrowband PA signal Restoration disposal, thus obtain the PA signal in broadband, get more target information；

Step 5, broadband PA signal reconstruction photoacoustic image step 4 obtained by time delay summation；

In step one, target absorption light produces PA signal and produces optoacoustic effect, meets equation:

${\&dtri;}^{2}p(r,t)-\frac{1}{c^2}\frac{{\&part;}^{2}p(r,t)}{\&part;t^2}=-\frac{\mathrm{\&beta;}}{C_p}\frac{\&part;}{\&part;t}H(r,t)---\left(1\right)$

Wherein r is the coordinate vector of target, and t is the time, and (r, t) is acoustic pressure to p, and (r is t) that incident laser excites at imaging region to H Heat source function, H (r, t)=A (r) I (t), A (r) be tissue light absorb distribution, I (t) for irradiate light intensity, β is thermal expansion system Number, C_pFor specific heat capacity, c is the experience velocity of sound；

Sensor employed in step 2 is a series of narrow sensor of the same amplitude-frequency characteristic of linear array, and arrowband senses There are sound wave lens device front end；

The arrowband PA signal that sensor absorbs with the relation of former PA signal is:

G (x, y)=f (x, y) * h (x, y)+η (x, y) (2)

((x, y) original PA signal, (x y) is the degeneration of narrow sensor to h to f to g for x, the arrowband PA signal that y) sensor receives Function, η (x, y) noise signal.

A kind of method using narrow sensor to rebuild broadband light acoustic image the most according to claim 1, it is characterised in that Pretreatment in step 3 includes:

Utilize two the signal p received₁(t)、p₂T () cross-correlation function estimates time delay t_s；

If p₁(t) and p₂T () is continuous signal, then cross-correlation function is:

$g\left(t\right)=p_1\left(t\right)*p_2\left(t\right)=\underset{-\mathrm{\&infin;}}{\overset{+\mathrm{\&infin;}}{\&Integral;}}{p}_1^{*}\left(\mathrm{\&tau;}\right)p_2(t+\mathrm{\&tau;})d\mathrm{\&tau;}---\left(3\right)$

It is p₁The complex conjugate of (t),For the convolution algorithm of continuous time, corresponding to g (t) peak value Time coordinate is time delay t_s,；

If p₁(t) and p₂T () is discrete signal, then cross-correlation function is:

${\hat{R}}_{p_1{p}_2}\left(m\right)=\left\{\begin{array}{cc}\frac{1}{L-\left|m\right|}\underset{s=0}{\overset{L-m-1}{\&Sigma;}}{p}_1(s+m)p_2\left(s\right)& m\&GreaterEqual;0\\ \widehat{R_{p_1{p}_2}}(-m)& m<0\end{array}\right.---\left(4\right)$

L is signal p₁(t) and signal p₂The greatest length of (t), m is discrete sampling time point,For discrete The convolution algorithm of time,Time coordinate corresponding to peak value is signal p₁(t) and signal p₂T () existsTime Accurate time delay t_s, f_sampleIt it is the sample frequency of reception system；

The arrowband PA signal received is alignd, carries out the arrowband PA signal after interpolation processingFor:

$\hat{p}(r,L_{\mathrm{int}})=\frac{1}{N_s}\underset{i=1}{\overset{N_s}{\&Sigma;}}{p}_{i\_shift}(r,L_{\mathrm{int}})---\left(5\right)$

L_intIt is the interpolation length of arrowband PA signal, N_sIt is the sampling frame number of arrowband PA signal, i value 1～N_s, p_{i_shift}(r, L_int) be displacement after arrowband PA signal, finally, N group signal after treatment, the PA signal obtained

N is discrete sampling time point.

A kind of method using narrow sensor to rebuild broadband light acoustic image the most according to claim 2, it is characterised in that In step 4, use the method for Wiener filtering to restore, make the signal recovered minimum with original signal mean square error.

A kind of method using narrow sensor to rebuild broadband light acoustic image the most according to claim 3, it is characterised in that The method for reconstructing that step 5 uses is time delay summation algorithm, and the process of reconstruction of image is the inverse of optoacoustic effect described in formula (1) Process, the sound pressure signal p (r detected on an array of sensors₀, t) be:

$p(r_0,t)=\mathrm{\&eta;}\&Integral;\&Integral;\&Integral;d^{3}rA\left(r\right)\frac{{\mathrm{\&delta;}}^{\&prime;}(t-\frac{|r_0-r|}{c})}{4\mathrm{\&pi;}|r_0-r|}---\left(7\right)$

Wherein η=β/C_p, r₀It is the coordinate vector of sensor, | r₀-r | being the reconstruction regions distance that arrives sensor, δ ' is laser arteries and veins The derived function of the delta-function that punching produces, the inverse process of said process is i.e. process of reconstruction, and method for reconstructing calculates according to formula (8):

For the effective aperture of sensor, when linear sensor, this absorption distribution approximation is asked by the time delay of formula (9) Obtain with algorithm:

$A\left(r\right)=\frac{{\mathrm{\&Sigma;}}_{k}w(k,r)p_k(r,t+T(k,r\left)\right)}{{\mathrm{\&Sigma;}}_{k}w(k,r)}---\left(9\right)$

Wherein k represents sensor, and (k, r) represents weight to w, and (k is r) to the time of sensor k, p from position r to T_kFor sensor k Received ultrasonic signal sequence；

Postpone summation to low frequency and high-frequency signal all weighted sums, the image smoothing obtained.