JP2013128761A - Photoacoustic signal processing device and method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To accurately obtain absorption distribution from a detection signal even in such a case that a change is generated in an optical pulse waveform with a photoacoustic signal processing device.SOLUTION: A photoacoustic signal is generated in a subject by irradiating the subject with light. The photoacoustic signal is detected by a probe 11 to be sampled by an AD conversion means 22 through a reception circuit 21. An optical waveform measuring means 25 measures the time waveform of the light intensity of the pulse laser beam applied to the subject. A differential optical waveform deconvolution means 26 deconvolves the light pulse waveform measured by the optical waveform measuring means 25 from the detected photoacoustic signal. By this deconvolution, the absorption distribution is obtained.

Description

  The present invention relates to a photoacoustic signal processing apparatus and method, and more particularly to a photoacoustic signal processing apparatus and method for processing a photoacoustic signal generated in a subject by light irradiated on the subject.

  An ultrasonic inspection method is known as a kind of image inspection method capable of non-invasively examining the state inside a living body. In the ultrasonic inspection, an ultrasonic probe capable of transmitting and receiving ultrasonic waves is used. When ultrasonic waves are transmitted from the ultrasonic probe to the subject (living body), the ultrasonic waves travel inside the living body and are reflected at the tissue interface. By receiving the reflected sound wave with the ultrasonic probe and calculating the distance based on the time until the reflected ultrasonic wave returns to the ultrasonic probe, the internal state can be imaged.

  In addition, photoacoustic imaging is known in which the inside of a living body is imaged using a photoacoustic effect. In general, in photoacoustic imaging, a living body is irradiated with pulsed laser light. Inside the living body, for example, living tissue absorbs the energy of pulsed laser light, and ultrasonic waves (photoacoustic signals) are generated by adiabatic expansion due to the energy. By detecting this photoacoustic signal with an ultrasonic probe or the like and constructing a photoacoustic image based on the detection signal, in-vivo visualization based on the photoacoustic signal is possible.

  Here, a normal reconstruction method (for example, the Fourier domain method (FTA method) or the delay addition method (Delay & Sum method)) substantially visualizes the pressure distribution, and the generated photoacoustic image is the light absorber. It is not a distribution image. In the photoacoustic image obtained by imaging the pressure distribution, one blood vessel may be displayed twice. This phenomenon can be explained as follows. That is, if the pressure waveform generated by the microelement (microabsorber) is a micro waveform, the pressure waveform generated from the blood vessel as the macro structure can be regarded as a superposition of the micro waveforms. When the micro waveforms generated in the microabsorber are superposed, the micro pressure waveform generated inside the blood vessel is superposed on the adjacent micro waveforms in front and back and canceled. When all the micro pressure waveforms from the blood vessel are overlaid, part of the pressure waveform that finally occurred on the blood vessel surface (front end) and part of the pressure waveform that occurred on the back surface (rear end) are not canceled. The rest will be observed. Therefore, the blood vessels are displayed twice. When one blood vessel is displayed twice, there is a problem that it is difficult to confirm the position of the blood vessel for image determination. In addition, when calculating images acquired at a plurality of different wavelengths, there is a problem that the position is easily shifted and an appropriate result cannot be obtained.

  Several techniques for generating an absorption distribution image instead of a pressure distribution image have been known so far. For example, in Patent Document 1, an inverse filter that repairs resolution degradation of a photoacoustic image is obtained from a thermal impulse response of a sample, and an ideal filter is applied to the obtained photoacoustic image so that an ideal photoacoustic image is obtained. That is, it is described that thermal impedance information (= a set of point heat sources that are infinitely small) is obtained by being excited by a point light source on the sample surface and detected.

  In Patent Document 1, more specifically, first, a thermal impulse response h (x, y) of a sample is calculated, and then a photoacoustic image p (x, y) is constructed. The thermal impulse response is defined as a transfer function until a temperature change at an infinitely small point is converted into a minute displacement on the sample surface. Thereafter, the thermal impulse response h (x, y) and the photoacoustic image p (x, y) are Fourier transformed to obtain Fourier transformed images H (μ, υ) and P (μ, υ). Using 1 / H (μ, υ) as an inverse filter, Q (μ, υ) is calculated by Q (μ, υ) = P (μ, υ) · (1 / H (μ, υ)). An ideal photoacoustic image q (x, y) is obtained by inversely transforming Q (μ, υ) thus calculated.

  In addition to the above-mentioned patent document, non-patent document 1 logically describes that the derivative of the optical pulse η (t) having a finite time width obtained by Fourier transform η (k) is iη ( k) is considered to be considered. Experimentally, the object was irradiated with microwaves with a long optical pulse width so that the excitation light pulse waveform was within the detection band of PZT (lead zirconate titanate), which is an ultrasonic detection element. The photoacoustic signal is detected by the PZT probe, and the absorption distribution is reconstructed.

Non-Patent Document 2 describes associating an absorption distribution with a micro waveform obtained by combining a light pulse differential function and a device impulse response function as a pressure waveform from a microelement in a subject, and an absorption distribution. ing. Absorption image reconstruction measures p _d0 including optical differentiation and system response in an indivisible state, deconvolutes p _d0 from the pressure waveform of each element, and then performs a filtered back projection method (Filtered Backprojection method) To do. Experimentally, excitation is performed with a pulsed laser beam with a short optical pulse width, and the detection band of ultrasonic waves is expanded beyond that of normal ultrasonic diagnostic equipment. Configure.

Japanese Patent Laid-Open No. 3-156362

Yuan Xu, et al., IEEE Transactions on Medical Imaging, Volume 21 (2002), p.823-828. Yi Wang, et al., Physics in Medicine and Biology, Volume 49 (2004), p.3117-3124

  In Non-Patent Document 1, η (k) obtained by Fourier transforming an optical pulse is not treated as a position-dependent function η (r, k. Therefore, t = 0 can be reconstructed (time is determined based on the detected waveform). In the case of calculating the pressure distribution generated at the moment when light enters retrospectively (t = 0), an accurate absorption distribution can be obtained, but when t ≠ 0 is reconstructed ( When the pressure distribution at the moment when light is incident (t = 0) cannot be reconstructed, and the pressure distribution after a while has elapsed from time t = 0, the optical pulse width component is It cannot be removed, resulting in a pressure distribution.

In Non-Patent Documents 1 and 2, either the excitation laser or the ultrasonic detector is removed from the practical range, and the laser emission time and the ultrasonic detection time are reconfigured. For this reason, although the problem does not appear clearly in Non-Patent Documents 1 and 2, when a practical device configuration is considered, it is difficult to reconfigure t = 0 with the methods of Non-Patent Documents 1 and 2. . That is, for example, as a practical device configuration,
-Ultrasonic detection device using a narrowband probe using PZT or the like with a sampling frequency of 100 MHz or less-When using an excitation laser with an optical pulse width on the order of 1-100 ns that produces a strong photoacoustic signal, the laser emission is super Since the phenomenon is shorter than the sound wave detection time, it cannot be accurately reconstructed in a state corresponding to t = 0 (a time zone in which the absorption distribution and the pressure distribution are proportional).

  Here, since “pressure distribution at time t = 0” represents “absorption distribution”, the absorption distribution can be obtained if the pressure distribution at time t = 0 is obtained. However, the sampling interval of a general ultrasonic detection apparatus is about 25 ns, and even if the calculation is made with the intention of the time t = 0 at the moment when the light hits, the deviation actually takes place with a time width of about t = ± 12.5 ns. Arise. For example, when the optical pulse width is as long as 100 ns, the above deviation (± 12.5 ns) can be considered as an error, but when the optical pulse width is 10 ns, the above deviation cannot be called an error, Instead of the pressure distribution at the moment of hitting, the pressure wave is transferred to a pressure wave propagation process. The “pressure distribution of the propagation process of the pressure wave” does not coincide with the “absorption distribution”.

  Furthermore, it is difficult to define a pressure distribution at t = 0 in a sample such as an experimental living body. Assuming that the speed of sound in the living body is 1530 m / s, for example, and the difference between the detection time and the laser irradiation time is the propagation time, the propagation distance is obtained from the propagation time. If the sound speed in the living body is constant at 1530 m / s, the propagation distance obtained from the propagation time coincides with the actual propagation distance. However, in practice, the speed of sound is not uniform in the living body, and there is a difference between the calculated propagation distance and the actual propagation distance. Therefore, when the propagation distance is estimated from the detection signal, the ambiguity of the propagation distance due to the sound speed difference remains. If the ambiguity of the propagation distance in the living body is regarded as the ambiguity of the propagation time, the time t = 0 is also ambiguous, and the pressure distribution at t = 0 is also ambiguous, making it difficult to define. The distribution of t = 0 is an absorption distribution, whereas the distribution of t> 0 is a pressure distribution during propagation, and if these are mixed, it cannot be said to be an absorption distribution.

  From the above consideration, the present inventor has found that it is advantageous to deconvolute the differential waveform of the optical pulse waveform from the observed waveform or the pressure waveform after reconstruction. An absorption distribution can be obtained by deconvolution of the optical pulse differential waveform from the observed waveform. However, the time waveform of the pulsed laser light applied to the subject is not always constant. For example, when irradiating a subject with pulsed laser light a plurality of times, the time waveform of the pulsed laser light may be different for each laser emission. If the time waveform of the pulse laser beam deviates from the waveform assumed in advance, the optical pulse differential waveform used for deconvolution does not match the actual differential waveform, and the absorption distribution cannot be obtained correctly.

  In view of the above, an object of the present invention is to provide a photoacoustic signal processing apparatus and method that can correctly obtain an absorption distribution from a detection signal even when a change occurs in an optical pulse waveform.

  In order to solve the above-mentioned problems, the present invention measures a sampling means for sampling a photoacoustic signal generated in a subject by light emitted from the subject, and a time waveform of the light intensity of the light emitted to the subject. And an optical differential waveform deconvolution means for generating a signal obtained by deconvoluting the differential waveform of the time waveform measured by the optical waveform measurement means from the sampled photoacoustic signal. A photoacoustic signal processing apparatus is provided.

  In the photoacoustic signal processing apparatus of the present invention, a configuration in which the optical waveform measuring means includes a light intensity detector that detects the light intensity of the light emitted to the subject can be employed.

  The light intensity detector may detect the light intensity of the light branched from the optical path from the light emission end of the light source of the light emitted to the subject until the light is emitted to the subject. In the optical path, a branching section that branches a part of the light emitted to the subject may be provided.

  The sampling means samples a plurality of photoacoustic signals generated by a plurality of times of light emission, the light waveform measuring means measures a time waveform of the light intensity in each of the plurality of times of light, and a photodifferential deconvolution means However, the differential waveform of the average waveform of the plurality of measured time waveforms may be deconvolved from the signal obtained by averaging the plurality of photoacoustic signals.

  In the photoacoustic signal processing apparatus of the present invention, the photodifferential waveform deconvolution means has a first Fourier transform means for Fourier transforming the sampled photoacoustic signal, and a second Fourier transform for the signal indicating the photodifferential waveform. Fourier transform means, inverse filter operation means for obtaining the inverse of the Fourier-transformed optical differential waveform as an inverse filter, filter application means for applying an inverse filter to the Fourier-transformed photoacoustic signal, and light to which the inverse filter is applied A configuration having Fourier inverse transform means for inversely transforming the acoustic signal can be employed.

  In the present invention, the photoacoustic signal is sampled at the first sampling rate, the measured optical differential waveform is sampled at the second sampling rate higher than the first sampling rate, and sampling is performed at the first sampling rate. The re-sampling means for re-sampling the photoacoustic signal thus obtained at the second sampling rate may be further provided, and the first Fourier transforming means may employ a configuration for Fourier-transforming the photo-acoustic signal resampled by the re-sampling means. .

  Instead of the above, the photoacoustic signal is sampled at the first sampling rate, the measured optical differential waveform is sampled at the second sampling rate higher than the first sampling rate, and the first Fourier The transforming means performs Fourier transform with the first number of data points, and the second Fourier transforming means performs Fourier transform with the second number of data points larger than the first number of data points. The signal further comprises zero padding means for performing zero padding for adding 0 to the center by the difference between the first data point and the second data point, and the filter applying means is zero padding by the zero padding means. A configuration may be adopted in which an inverse filter is applied to a signal subjected to.

  Further, the photoacoustic signal is sampled at the first sampling rate, the measured optical differential waveform is sampled at the second sampling rate higher than the first sampling rate, and the first Fourier transform means is provided. Fourier transform is performed with the first number of data points, and the second Fourier transform means performs Fourier transform with a second number of data points greater than the first number of data points. From the Fourier-transformed optical differential waveform, High frequency component sample point removing means for removing the high frequency component sample points by the difference between the first data point and the second data point is further provided, and the inverse filter operation means is adapted to perform high frequency component from the Fourier-transformed optical pulse differential waveform. The reciprocal of the signal from which the sample points have been removed may be obtained as an inverse filter.

  The sampling interval of the photoacoustic signal may be longer than the pulse time width of the light emitted to the subject.

  The photoacoustic signal processing apparatus of the present invention further comprises photoacoustic signal reconstruction means for reconstructing the photoacoustic signal based on the photoacoustic signal detected by the plurality of detection elements and sampled by the sampling means, and a photodifferential waveform. A configuration in which the deconvolution means deconvolutes the optical differential waveform from the photoacoustic signal reconstructed by the photoacoustic signal reconstruction means can be employed.

  The optical differential waveform deconvolution means may deconvolute a photoacoustic signal obtained by multiplying the optical differential waveform by the device response function.

  The light emitted to the subject may be pulsed laser light.

  In the present invention, the emitted light to the subject includes light of a plurality of wavelengths, the sampling means samples a photoacoustic signal corresponding to the light of each wavelength, and the optical differential waveform deconvolution means corresponds to the light of each wavelength. A signal obtained by deconvolution of a photodifferential waveform from a photoacoustic signal may be generated. In this case, a signal obtained by deconvolution of a photodifferential waveform from a photoacoustic signal corresponding to light of each wavelength is calculated. The wavelength data calculation means may be further provided.

  It is good also as a structure further provided with the photoacoustic image generation means which produces | generates a photoacoustic image based on the signal which deconvoluted the photodifferential waveform from the photoacoustic signal.

  The sampling means further samples a reflected acoustic wave with respect to the acoustic wave transmitted to the subject, and a reflected acoustic wave image generating means for generating a reflected acoustic wave image based on the sampled reflected acoustic wave; You may employ | adopt the structure further provided with the image composition means to synthesize | combine a photoacoustic image and a reflected acoustic wave image.

  The image synthesis means may perform image synthesis by superimposing the photoacoustic image and the reflected acoustic wave image.

  The present invention also includes a step of detecting a photoacoustic signal generated in the subject due to light emission to the subject, a step of measuring a time waveform of the light intensity of the irradiated light, and the detected light And a step of deconvoluting a photodifferential waveform obtained by differentiating the time waveform of the measured light from the acoustic signal.

  In the photoacoustic signal processing apparatus and method of the present invention, the time waveform of the light intensity of the light irradiated to the subject is measured, and the measured time waveform is detected from the detection signal of the photoacoustic signal generated by the light irradiation. Deconvolve the differential waveform. In this way, by measuring the time waveform of the irradiated light and deconvoluting the differential waveform, the photodifferential waveform convoluted to the observed waveform is correctly derived from the photoacoustic signal detection signal in the deconvolution. Deconvolution can be performed, and an absorption distribution can be obtained correctly.

The block diagram which shows the basic algorithm of an optical pulse differential waveform deconvolution. The wave form diagram which shows the photoacoustic signal after a reconstruction. The wave form diagram which shows the photoacoustic signal FFT after FFT. The wave form diagram which shows an optical pulse differential waveform (h). The wave form diagram which shows the optical pulse differential waveform FFT (fft_h) after FFT. The wave form diagram which shows an optical pulse differential waveform FFT filter. The wave form diagram which shows the FFT waveform after deconvolution. The wave form diagram which shows the photoacoustic signal reversely converted. The figure which shows the photoacoustic image produced | generated based on the photoacoustic signal after a reconstruction. The figure which shows the photoacoustic image produced | generated based on the photoacoustic signal after a deconvolution. The block diagram which shows the photoacoustic image generation apparatus containing the photoacoustic signal processing apparatus of 1st Embodiment of this invention. The block diagram which shows an optical waveform measurement means. The block diagram which shows an optical pulse differential waveform reverse convolution means. The flowchart which shows the operation | movement procedure of photoacoustic image generation. The block diagram which shows the optical pulse differential waveform reverse convolution means in 2nd Embodiment of this invention. The wave form diagram which shows the optical pulse differential waveform of sampling rate 400MHz. The wave form diagram which shows the optical pulse differential waveform of sampling rate 40MHz. The block diagram which shows the optical pulse differential waveform reverse convolution means in 3rd Embodiment of this invention. The graph which shows a photoacoustic signal (frequency domain). The graph which shows the photoacoustic signal after zero padding. The block diagram which shows the optical pulse differential waveform reverse convolution means in 4th Embodiment of this invention. The graph which shows an optical pulse differential waveform (frequency domain). The graph which shows the optical pulse differential waveform from which the high frequency component sample point was removed. The block diagram which shows the photoacoustic image generating apparatus of 5th Embodiment of this invention. The block diagram which shows the photoacoustic image generating apparatus of 6th Embodiment of this invention.

Prior to the description of the embodiments of the present invention, an outline of the present invention will be described. Consider a micro-absorbing particle that is a light absorber, and consider that this micro-absorbing particle absorbs pulsed laser light to generate a pressure wave (photoacoustic pressure wave). The pressure waveform p _micro (R, t) when the photoacoustic pressure wave generated from the microabsorbent particle at the position r is observed at the position R, where time is t, is [Phys. Rev. Lett. 86 (2001 ) 3550.], the following spherical wave is obtained.
Here, I (t) is a time waveform of the light intensity of the excitation light, the coefficient k is a conversion coefficient when the particle absorbs light and outputs an acoustic wave, and v _s is the sound velocity of the subject. is there. Positions r and R are vectors indicating positions in space. The pressure generated from the microabsorbent particles is a spherical wave proportional to the optical pulse differential waveform, as shown in the above formula.

Since the pressure waveform actually obtained from the object to be imaged has a macroscopic absorber size, it is considered to be a waveform obtained by superimposing the above micro absorption waveforms (superposition principle). Here, the absorption distribution of particles emitting macrophotoacoustic waves is A (r-R), and the observation waveform of pressure from the macro absorber is p _macro (R, t). At the observation position R, since the photoacoustic wave from the absorbing particles located at the radius v _s t from the observation position R is observed at each time, the observation waveform p _macro (R, t) has the following pressure It is shown by the waveform formula.
As can be seen from the above equation (1), the observed waveform shows a convolution type of optical pulse differentiation. Therefore, in the present invention, in the deconvolution process, the deconvolution is performed in consideration of the optical pulse differential term.

Furthermore, in the present invention, after applying reconstruction (FTA method, DnS method, BP method, etc.) for obtaining a pressure distribution that is also used in conventional ultrasound systems, the pressure distribution of the image after reconstruction is t ≠ 0. In other words, after recognizing the pressure distribution in the propagation process of the pressure wave, we considered converting it into an absorption distribution. The basic concept of the pressure distribution reconstruction is that the pressure distribution p _rec (R, t) after the reconstruction at the detection position R = (x, y, 0) is the R detection axis (r -R) is obtained by adding the spherical waves generated from the absorber present at the position | r-R | on the position including the signals of the surrounding piezoelectric elements and calculating the pressure intensity at that position. Therefore, p _rec (R, t) generated by superimposing propagating photoacoustic waves generated from a micro absorber existing on the detection axis (r-R) can be expressed as follows.
As described above, the absorption distribution can be considered in a one-dimensional manner, whereby the pressure notation as described above can be performed. The above equation (2) can be expressed as follows, where the detection axis (r−R) is the z-axis and the distance | r−R | from the detection element is z.
Further, from the equation (3), x and y which are not related to integration are omitted, and the z-axis is represented by time, the above equation can be expressed as the following equation.
In this way, a convolution representation of one axis (time axis or z ′ axis) in the detection element located at (x, y, 0) becomes possible.

The optical pulse derivative can be deconvoluted by Fourier-transforming both sides of the above formula (4) and dividing the Fourier coefficient of the pressure distribution by the Fourier coefficient of the time derivative of the optical pulse on the frequency axis.
After deconvolution, the resulting equation, by inverse Fourier transform to obtain the A (x, y, v s t), can be imaged absorption distribution. Here, the A (x, y, v s t) determined, the detection element receives the angle-dependent D (x, y, z) and, the natural vibration of the probe band is likely to have been superimposed. For example, by obtaining the device function D (x, y, z) in advance and multiplying the reciprocal by A (x, y, v _s t), the influence of the detection element reception angle dependency can be removed. . Further, regarding the natural vibration of the band, if the intensity is imaged by Hilbert transform or orthogonal detection processing, the influence can be removed. Further, the light spatial distribution L (x, y, z) incident on the specimen is separately obtained by observation or simulation, and the pixel value μ (x, y, z) proportional to the absorption coefficient is determined by μ (x, y, z). _{= A (x, y, v} s t) / L (x, y, z) may be obtained by. In this case, an absorption coefficient distribution image, which is a physical quantity more closely related to the living tissue, can be obtained.

  FIG. 1 shows a basic algorithm of optical pulse differential waveform deconvolution. The reconstructed photoacoustic signal is input, and the reconstructed photoacoustic signal is Fourier transformed by FFT (Fast Fourier Transform) (step S101). FIG. 2A shows the photoacoustic signal after reconstruction, and FIG. 2B shows the photoacoustic signal FFT after FFT. By performing Fourier transform, the time domain signal shown in FIG. 2A is converted into a frequency domain signal as shown in FIG. 2B. In FIG. 2B, the absolute value of the photoacoustic signal FFT is shown, but in actual processing, it is processed as a complex number.

  The optical pulse differential waveform h is Fourier transformed by FFT (step S102). FIG. 2C shows an optical pulse differential waveform (h), and FIG. 2D shows an optical pulse differential waveform FFT (fft_h) after FFT. By performing the Fourier transform, the signal (waveform) in the time domain shown in FIG. 2C is converted into the signal in the frequency domain shown in FIG. 2D. Note that black circles in FIG. 2C represent sampling points in the optical pulse differential waveform. 2D shows the absolute value of the optical pulse differential waveform FFT, but in an actual process, it is processed as a complex number.

The inverse of the post-FFT optical pulse differential waveform FFT (fft_h) obtained in step S102 is obtained as an optical pulse differential waveform FFT filter (inverse filter) (step S103). Specifically, the optical pulse differential waveform FFT filter can be obtained by conj (fft_h) / abs (fft_h) ² . Here, conj (fft_h) represents the conjugate complex number of fft_h, and abs (fft_h) represents the absolute value of fft_h. FIG. 2E shows an optical pulse differential waveform FFT filter. By obtaining the inverse of the optical pulse differential waveform FFT shown in FIG. 2D, an optical pulse differential waveform FFT filter as shown in FIG. 2E can be obtained.

  The optical pulse differential FFT filter obtained as described above and the reconstructed photoacoustic signal FFT are multiplied for each element, and the optical pulse differential waveform is deconvolved from the photoacoustic signal FFT (step S104). FIG. 2F shows the FFT waveform after deconvolution. The FFT waveform shown in FIG. 2F is obtained by multiplying the photoacoustic signal FFT shown in FIG. 2B by the optical pulse differential waveform FFT filter shown in FIG. 2E.

  The FFT waveform obtained by deconvolution of the optical pulse differential waveform in step S104 is inversely Fourier transformed by inverse FFT to return the frequency domain signal to the time domain signal (step S105). FIG. 2G shows the inversely converted photoacoustic signal. By performing inverse FFT on the FFT waveform (frequency domain signal) shown in FIG. 2F, the deconvolution photoacoustic signal (time domain signal) shown in FIG. 2G is obtained. This deconvolved photoacoustic signal is an absorption distribution obtained by deconvolution of the optical pulse differential waveform from the reconstructed photoacoustic signal (FIG. 2A) in which the optical pulse differential waveform (FIG. 2C) is convolved with the optical absorption distribution. It corresponds to.

  3A shows a photoacoustic image generated based on the reconstructed photoacoustic signal (FIG. 2A), and FIG. 3B shows a photoacoustic image generated based on the deconvolved photoacoustic signal (FIG. 2G). Show. The photoacoustic image generated based on the reconstructed photoacoustic signal shown in FIG. 3A is substantially an image of the pressure distribution, and image determination such that one blood vessel is displayed twice. The blood vessel position is difficult to confirm. On the other hand, the photoacoustic image generated based on the deconvolved photoacoustic signal shown in FIG. 3B can visualize the distribution of the absorber by deconvolution of the optical pulse differential waveform. It is easy to confirm the position.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. FIG. 4 shows a photoacoustic image generation apparatus including the photoacoustic signal processing apparatus according to the first embodiment of the present invention. The photoacoustic image generation apparatus (photoacoustic image diagnostic apparatus) 10 includes an ultrasonic probe (probe) 11, an ultrasonic unit 12, and a light source unit (laser unit) 13.

  The laser unit 13 generates laser light that irradiates the subject. What is necessary is just to set the wavelength of a laser beam suitably according to an observation target object. Laser light emitted from the laser unit 13 is guided to the probe 11 using light guide means such as an optical fiber, and is irradiated from the probe 11 to the subject. The probe 11 detects the ultrasonic wave (photoacoustic signal) generated by the light absorber in the subject absorbing the laser light after the light emitted from the laser unit 13 is irradiated on the subject. The probe 11 has, for example, a plurality of ultrasonic transducers arranged one-dimensionally.

  The ultrasonic unit 12 includes a receiving circuit 21, an AD converting means 22, a photoacoustic image reconstruction means 24, an optical pulse differential waveform reverse convolution means 26, a detection / logarithmic conversion means 27, a photoacoustic image construction means 28, a trigger control circuit. 29 and control means 30. The ultrasonic unit 12 constitutes a photoacoustic signal processing device together with the optical waveform measuring means 25. In FIG. 4, the optical waveform measuring unit 25 is provided outside the ultrasonic unit 12, but a part or all of the optical waveform measuring unit 25 may be included in the ultrasonic unit 12.

  The receiving circuit 21 receives the photoacoustic signal detected by the probe 11. The AD conversion means 22 is a sampling means, which samples the photoacoustic signal received by the receiving circuit 21 and converts it into a digital signal. The AD conversion means 22 samples a photoacoustic signal with a predetermined sampling period based on, for example, an AD clock signal with a predetermined frequency input from the outside. The reception memory 23 stores the photoacoustic signal sampled by the AD conversion means 22.

  The photoacoustic image reconstruction unit 24 reads out the photoacoustic signal from the reception memory 23 and generates data of each line of the photoacoustic image based on the photoacoustic signals detected by the plurality of ultrasonic transducers of the probe 11. . The photoacoustic image reconstruction means 24 adds, for example, data from 64 ultrasonic transducers of the probe 11 with a delay time corresponding to the position of the ultrasonic transducer, and generates data for one line (delay). Addition method). The photoacoustic image reconstruction means 24 may perform reconstruction by the BP method (Back Projection) instead of the delay addition method. Alternatively, the photoacoustic image reconstruction unit 24 may perform reconstruction using the Hough transform method or the Fourier transform method.

  The optical waveform measuring means 25 measures the temporal waveform of the light intensity of light (for example, pulsed laser light) irradiated on the subject. The optical pulse differential waveform deconvolution means 26 generates a signal obtained by deconvolution of the differential waveform of the time waveform measured by the optical waveform measurement means 25 from the photoacoustic signal reconstructed by the photoacoustic image reconstruction means 24. . By deconvolution of the optical pulse differential waveform, the pressure distribution reconstructed at t = 0, that is, the absorption distribution can be obtained from the pressure distribution reconstructed at t ≠ 0. The optical pulse differential waveform deconvolution means 26 may perform deconvolution on the photoacoustic signal before reconstruction.

  The detection / logarithm conversion means 27 obtains an envelope of the data of each line after deconvolution, and logarithmically transforms the obtained envelope. Conventional detection methods such as Hilbert transform and quadrature detection can be used as detection means for obtaining the envelope. Thereby, the influence of the zone | band by the natural vibration of an ultrasonic transducer | vibrator can be removed. The photoacoustic image construction means 28 generates a photoacoustic image based on the data of each line subjected to logarithmic transformation. For example, the photoacoustic image construction unit 28 converts the position of the photoacoustic signal (peak portion) in the time axis direction into a position in the depth direction of the photoacoustic layer image to generate a photoacoustic image.

  The control means 30 controls each part in the ultrasonic unit 12. The trigger control circuit 29 sends a flash lamp trigger signal to the laser unit 13 when generating the photoacoustic image. Further, after outputting the flash lamp trigger signal, the Q switch trigger signal is sent. The laser unit 13 includes a flash lamp 31 and a Q switch 32. Upon receiving the flash lamp trigger signal, the laser unit 13 turns on the flash lamp 31 and starts laser excitation. When a Q switch trigger signal is input, the laser unit 13 turns on the Q switch and emits pulsed laser light. The trigger control circuit 29 sends a sampling trigger signal to the AD conversion means 22 in synchronization with laser light irradiation on the subject, and controls the sampling start timing of the photoacoustic signal in the AD conversion means 22.

  It should be noted that correction means is provided at the subsequent stage of the optical pulse differential waveform deconvolution means 26, and the correction means influences the reception angle dependence characteristics of the ultrasonic transducer in the probe 11 from the signal obtained by deconvolution of the optical pulse differential waveform. It is good also as a structure which removes. The correction means may remove the influence of the incident light distribution of the light on the subject from the signal in which the optical pulse differential waveform is deconvoluted in addition to or in place of the reception angle dependent characteristics.

  FIG. 5 shows the optical waveform measuring means 25. The optical waveform measuring unit 25 includes a light intensity detector 251, an amplifier 252, an AD converter 253, and a waveform memory 254. The light intensity detector 251 detects the light intensity of the pulsed laser light irradiated on the subject. The light intensity detection unit 251 outputs a voltage signal corresponding to the detected light intensity, for example. For example, a high-speed response photodiode can be used for the light intensity detection unit 251.

  For example, in the optical path (for example, an optical fiber) from the light emitting end of the laser unit 13 (FIG. 4) to the irradiation of the subject with the pulsed laser light, the pulsed laser light to be irradiated on the subject. Is provided with a branching section 255 that branches a part of the section. As the branching portion 255, for example, a mirror having a high transmittance, specifically, a mirror having a transmittance of 95% or more can be used. Such a mirror is disposed, for example, at an angle of 45 ° with respect to the laser main stream toward the subject, and the reflection component of the mirror is branched in the direction of the light intensity detection unit 251. The light intensity of the branched light corresponds to the light intensity of the pulsed laser light applied to the subject, and the light intensity detection unit 251 detects a part of the pulsed laser light branched by the mirror. A transparent glass may be used for the branching unit 255 instead of the mirror, and the reflection component of the transparent glass may be branched in the direction of the light intensity detecting unit 251. As another example, if a laser beam is guided using, for example, a bundle fiber in which a large number of optical fibers are bundled, some (for example, one) of a plurality of optical fibers are lighted. The light may be guided to the intensity detector 251 and the light detected by the light intensity detector 251.

  The amplifier 252 amplifies the signal output from the light intensity detector 251 with a predetermined amplification factor. The AD converter 253 samples the signal amplified by the amplifier 252 at a predetermined sampling rate. The AD converter 253 samples the intensity detected by the light intensity detector 251 over the light emission period of the pulse laser light, whereby the temporal waveform of the light intensity of the pulse laser light irradiated on the subject can be measured. The AD converter 253 stores light intensity sampling data (light pulse waveform) in the waveform memory 254.

  FIG. 6 shows the optical pulse differential waveform deconvolution means 26. The optical pulse differential waveform deconvolution means 26 includes differential waveform calculation means 40, Fourier transform means 41 and 42, inverse filter calculation means 43, filter application means 44, and Fourier inverse transform means 45. The differential waveform calculation means 40 generates an optical pulse partial waveform from the optical pulse waveform. For example, the differential waveform calculation means 40 reads the optical pulse waveform from the waveform memory 254 (FIG. 5), and generates an optical pulse differential waveform obtained by differentiating the read optical pulse waveform.

  The Fourier transform means (first Fourier transform means) 41 converts the reconstructed photoacoustic signal from a time domain signal to a frequency domain signal by discrete Fourier transform. The Fourier transform means (second Fourier transform means) 42 converts the optical pulse differential waveform generated by the differential waveform calculation means 40 from a time domain signal to a frequency domain signal by discrete Fourier transform. FFT can be used as an algorithm for Fourier transform.

  In the above description, the optical pulse differential waveform deconvolution means 26 generates the optical pulse differential waveform. However, the present invention is not limited to this. For example, the optical waveform differential means 25 may generate the optical pulse differential waveform. In that case, for example, a differential waveform calculating means may be provided between the AD converter 253 (FIG. 5) of the optical waveform measuring means 25 and the waveform memory 254. The differential waveform calculation means generates a differential waveform from the optical pulse waveform sampled by the AD converter 253, and stores the generated optical pulse differential waveform in the waveform memory 254. When the optical pulse differential waveform is generated before the waveform data is stored in the waveform memory 254, the Fourier transform means 42 reads the optical pulse differential waveform from the waveform memory 254, and performs a Fourier transform on the read optical pulse differential waveform. Just do it.

  In the present embodiment, it is assumed that the sampling rate of the photoacoustic signal is equal to the sampling rate of the optical pulse differential waveform. For example, the AD conversion means 22 (FIG. 4) which samples a photoacoustic signal samples a photoacoustic signal synchronizing with the sampling clock of Fs = 40MHz. On the other hand, the AD converter 253 (FIG. 5) of the optical waveform measuring means 25 samples the optical pulse waveform in synchronization with the sampling clock of Fs_h = 40 MHz. The Fourier transform means 41 performs a Fourier transform on the photoacoustic signal sampled at 40 MHz by, for example, a 1024-point Fourier transform. The Fourier transform means 42 Fourier transforms the optical pulse differential waveform obtained by differentiating the optical pulse waveform sampled at 40 MHz by 1024 points of Fourier transform.

The inverse filter calculation unit 43 obtains the inverse of the Fourier transformed optical pulse differential waveform as an inverse filter. For example, the inverse filter calculation means 43 obtains conj (fft_h) / abs (fft_h) ² as an inverse filter, where fft_h is a signal obtained by Fourier transforming the optical pulse differential waveform h. The filter applying unit 44 applies the inverse filter obtained by the inverse filter calculating unit 43 to the photoacoustic signal Fourier-transformed by the Fourier transform unit 41. For example, the filter application unit 44 multiplies the Fourier coefficient of the photoacoustic signal by the Fourier coefficient of the inverse filter for each element. By applying the inverse filter, the optical pulse differential waveform is deconvolved in the frequency domain signal. The Fourier inverse transform unit 45 transforms the photoacoustic signal to which the inverse filter is applied, from a frequency domain signal to a time domain signal by Fourier inverse transform. An absorption distribution signal in the time domain is obtained by inverse Fourier transform.

  FIG. 7 shows an operation procedure. The trigger control circuit 29 outputs a flash lamp trigger signal to the laser unit 13. The laser unit 13 turns on the flash lamp 31 in response to the flash lamp trigger signal. The trigger control circuit 29 outputs a Q switch trigger signal at a predetermined timing. When the Q switch trigger signal is input, the laser unit 13 turns on the Q switch 32 and emits a pulse laser beam. The emitted pulsed laser light is guided to the probe 11, for example, and irradiated from the probe 11 to the subject (step S1).

  The optical waveform measuring means 25 measures the time waveform of the pulse laser beam irradiated on the subject in step S1 (step S3). The probe 11 detects the photoacoustic signal generated in the subject by the laser beam irradiation after the laser beam irradiation (step S3). The receiving circuit 21 of the ultrasonic unit 12 receives the photoacoustic signal detected by the probe 11. The trigger control circuit 29 sends a sampling trigger signal to the AD conversion means 22 in accordance with the timing of light irradiation on the subject. The AD conversion means 22 receives the sampling trigger signal, starts sampling of the photoacoustic signal, and stores the sampling data of the photoacoustic signal in the reception memory 23.

  The photoacoustic image reconstruction unit 24 reads out the photoacoustic signal sampling data from the reception memory 23, and reconstructs the photoacoustic signal based on the read out photoacoustic signal sampling data (step S4). The optical pulse differential waveform deconvolution means 26 deconvolves the optical pulse differential waveform obtained by differentiating the optical pulse waveform measured in step S1 from the photoacoustic signal reconstructed in step S3 (step S5). By this deconvolution, a photoacoustic signal indicating an absorption distribution is obtained.

  The detection / logarithm conversion means 27 obtains the envelope of the photoacoustic signal subjected to the deconvolution, and logarithmically transforms the obtained envelope. The photoacoustic image construction means 28 generates a photoacoustic image based on the data of each line subjected to logarithmic transformation (step S6). This photoacoustic signal is an absorption distribution image obtained by imaging the absorption distribution. The image display unit 14 displays a photoacoustic image that is an absorption distribution image on the display screen (step S7).

  In this embodiment, the photoacoustic image (photoacoustic image) is first reconstructed by the photoacoustic image reconstruction means 24 as a pressure distribution at the light emission time (t = 0) by a normal reconstruction method. Next, since the light emission time actually has a finite length, the time when t = 0 at the time of reconstruction is considered as a finite time, and the optical differential waveform deconvolution means 26 The optical pulse differential waveform is deconvolved from the reconstructed photoacoustic image. By deconvolution of the optical pulse differential waveform, an absorption distribution can be obtained and an absorption distribution image can be generated. By adopting such a method, the absorption distribution can be imaged even when a practical light pulse width and a practical ultrasonic system or an actual living body is observed. This has the advantage that the current system detector bandwidth and AD sampling can be used. Further, in this embodiment, since the pressure distribution is once obtained by reconstructing the photoacoustic image, the compatibility with the existing ultrasonic algorithm and apparatus is high.

  In the present embodiment, the optical waveform measuring means 25 measures the time waveform of the light intensity of the pulsed laser light irradiated on the subject, and the optical differential waveform reverse convolution means 26 is measured by the optical waveform measuring means 25. Deconvolute the differential waveform of the optical pulse waveform. The time waveform of the light intensity of the pulsed laser light applied to the subject is not always constant, and may vary with each laser emission or change depending on the apparatus mode. It is also conceivable that the optical pulse waveform changes with the change with time, for example, as the number of flash lamps increases. In the present embodiment, the optical waveform measuring means 25 is used to measure the optical pulse waveform applied to the subject, and the differential waveform of the measured optical pulse waveform is deconvoluted. The absorption distribution can be obtained more accurately than when the waveform is deconvolved.

  Here, as another example, it is conceivable to approximate the optical pulse waveform or its differential waveform using a function. However, for example, the optical pulse waveform may be a waveform that is difficult to approximate with a function, such as a lid-shaped waveform or a long-tailed waveform. In the case of such a waveform, it is considered that a correct absorption distribution cannot be obtained if deconvolution is performed using a waveform that is forcibly approximated by a function. In this embodiment, the optical pulse waveform irradiated to the subject is measured, and deconvolution is performed using the differential waveform. Therefore, even if the optical pulse waveform is difficult to approximate with a function, accurate deconvolution correction is performed. Processing is possible.

  In addition, when generating a photoacoustic image, the photoacoustic signals generated by the multiple times of pulsed laser light irradiation may be added and averaged to generate one photoacoustic signal based on the multiple detection signals. In such a case, an average of optical pulse waveforms of a plurality of times of pulsed laser light may be obtained, and a differential waveform of the average waveform of the optical pulse waveform may be deconvolved from the added and averaged photoacoustic signal. For example, when the pulse laser beam is irradiated five times and the five photoacoustic signals are added and averaged, the photoacoustic image reconstruction means 24 reconstructs the photoacoustic signal every time the pulse laser beam is irradiated. The five photoacoustic signals thus obtained are added and averaged. The optical waveform measuring means 25 measures the optical pulse waveform of each of the five pulsed laser beams. The optical differential waveform deconvolution means 26 obtains an average optical pulse waveform obtained by averaging the five measured optical pulse waveforms, and deconvolutes the differential waveform of the average optical pulse waveform from the addition-averaged photoacoustic signal. . In the case of the child, the processing time can be shortened as compared with the case where the deconvolution is performed every time.

  Next, a second embodiment of the present invention will be described. In the first embodiment, the sampling rate of the photoacoustic signal and the sampling rate of the optical pulse differential waveform match, and both signals are Fourier-transformed with the same number of data points. In this embodiment, the photoacoustic signal is sampled at low speed, while the optical pulse differential waveform is sampled at high speed. That is, the sampling rate of the optical pulse waveform (optical pulse differential waveform) is set higher than the sampling rate of the photoacoustic signal. For example, the sampling interval of the photoacoustic signal (the reciprocal of the sampling rate) is set longer than the pulse time width of the light irradiated to the subject. In the Fourier transform, a low-sampling rate photoacoustic signal is resampled (upsampled) at the same sampling rate as the sampling rate of the optical pulse waveform, and then the Fourier transform is performed. Other points may be the same as in the first embodiment.

  FIG. 8 shows the optical pulse differential waveform reverse convolution means 26a in the present embodiment. The optical pulse differential waveform deconvolution means 26a in the present embodiment has resample means 46 and 47 in addition to the configuration of the optical pulse differential waveform deconvolution means 26 in the first embodiment shown in FIG. The re-sampling unit 46 is an up-sampling unit, and up-samples the sampling data of the photoacoustic signal sampled at a low sampling rate at the same sampling rate as the sampling rate of the optical pulse waveform. The resampling means 46 performs upsampling, for example, by adding zero between sample points of the photoacoustic signal sampled at a low sampling rate and applying a low-pass filter that cuts at the Nyquist frequency before upsampling.

  For example, the photoacoustic signal sampling rate (first sampling rate) in the AD converter 22 (FIG. 4) is 40 MHz, and the optical pulse waveform sampling rate (second sampling rate) in the AD converter 253 (FIG. 5). ) Is 400 MHz. In this case, the resampling means 46 upsamples the 40 MHz photoacoustic signal to a 400 MHz signal. The Fourier transform unit 41 performs a Fourier transform on the photoacoustic signal upsampled by the resample unit 46. The Fourier transform means 41 for Fourier transforming the photoacoustic signal and the Fourier transform means 42 for Fourier transforming the optical pulse differential waveform perform Fourier transform with the same number of data points. For example, the Fourier transform unit 41 converts the photoacoustic signal into a signal in the frequency region of 8192 points, and the Fourier transform unit 42 converts the optical pulse differential waveform into a signal in the frequency region of 8192 points.

  The filter application unit 44 applies an inverse filter to a signal obtained by Fourier transforming the upsampled photoacoustic signal. The Fourier inverse transform means 45 transforms the signal to which the inverse filter is applied from a frequency domain signal into a time domain signal (absorption distribution). The absorption distribution signal returned to the time domain signal is a signal in a state of being upsampled to, for example, 400 MHz. The resampling means 47 downsamples the absorption signal so that the absorption distribution signal becomes a signal sampled at the original sampling rate of the photoacoustic signal. The resampling unit 47 downsamples, for example, a 400 MHz absorption signal into a 40 MHz absorption signal. Downsampling is performed, for example, by thinning sample points after applying a low-pass filter that cuts at the Nyquist frequency after downsampling.

  FIG. 9A shows an optical pulse differential waveform corresponding to a sampling rate of 400 MHz, and FIG. 9B shows an optical pulse differential waveform corresponding to a sampling rate of 40 MHz. At a sampling rate of 400 MHz, as shown in FIG. 9A, a waveform obtained by differentiating the time waveform of the light intensity of the pulse laser beam can be accurately reproduced. On the other hand, if the sampling rate of the optical pulse differential waveform is matched to the sampling rate of the photoacoustic signal to obtain a signal corresponding to 40 MHz, the optical differential waveform cannot be accurately reproduced as shown in FIG. 9B.

  When the inverse filter is applied to the signal obtained by Fourier transforming the photoacoustic signal by the filter applying means 44, it is necessary to have both data points. When the sampling rate of the optical pulse differential waveform is set in accordance with the sampling rate of the photoacoustic signal, as shown in FIG. 9B, the sampling frequency is too low for the waveform change, and the optical differential waveform cannot be accurately reproduced. When an inverse filter obtained from such an optical pulse differential waveform is applied, the optical pulse differential term may not be accurately deconvolved, and the absorption distribution may not be obtained correctly.

  On the other hand, if the optical pulse differential waveform is a signal corresponding to 400 MHz in order to accurately reproduce the optical differential waveform and the photoacoustic signal sampling rate is set to 400 ef MHz, the optical pulse differential term can be accurately deconvolved and absorbed. The distribution can be obtained correctly. However, in that case, a high-speed AD converter is required for the AD conversion means 22, and the total number of sampling data increases, so that the memory capacity required for the reception memory 23 (FIG. 4) increases. Furthermore, since the data handled by the photoacoustic image reconstruction means 24 increases, the time required for reconstruction also increases.

  In the present embodiment, the resampler 46 resamples the sampling data of the photoacoustic signal afterwards. In this embodiment, since the photoacoustic signal after detection is upsampled by signal processing, the optical pulse differential term can be accurately deconvolved while performing low-speed sampling from photoacoustic detection to reconstruction. . In the present embodiment, a high-speed AD converter is not necessary for the AD conversion means 22 and the memory capacity required for the reception memory 23 does not increase. Further, the time required for reconstructing the photoacoustic signal does not increase, and the processing time can be shortened as compared with the case of sampling at a high sampling rate when detecting the photoacoustic signal.

  Subsequently, a third embodiment of the present invention will be described. Also in the present embodiment, as in the second embodiment, the sampling rate of the optical pulse waveform (optical pulse differential waveform) is set higher than the sampling rate of the photoacoustic signal. In the second embodiment, a photoacoustic signal sampled at a low sampling rate is upsampled, and both signals are Fourier transformed with the same number of data points. In the present embodiment, the Fourier transform of the optical pulse differential waveform is performed with more data points than the Fourier transform data points of the photoacoustic signal, and the Fourier transform photoacoustic signal is centered by the difference in the data points. A zero point is added to (high frequency component region). Other points may be the same as in the first embodiment.

  FIG. 10 shows the optical pulse differential waveform reverse convolution means 26b in the present embodiment. The optical pulse differential waveform deconvolution means 26b in this embodiment includes a zero padding means 48, a zero point removal means 49, in addition to the configuration of the optical pulse differential waveform deconvolution means 26 in the first embodiment shown in FIG. Have For example, it is assumed that the sampling rate (first sampling rate) of the photoacoustic signal is 40 MHz and the sampling rate (second sampling rate) of the optical pulse waveform (optical pulse differential waveform) is 320 MHz. For example, the Fourier transform means 41 converts a photoacoustic signal of 40 MHz into a signal in the frequency domain of 1024 points (first data point), and the Fourier transform means 42 converts the optical pulse differential waveform of 320 MHz to 8192 points (second Converted to a frequency domain signal. The second data score is equal to or greater than the data score obtained by multiplying the first data score by the ratio of the second sampling rate and the first sampling rate.

  The zero padding means 48 inputs the photoacoustic signal converted from the Fourier transform means 41 into a frequency domain signal. Zero padding means 48 adds a zero point (point of zero signal value) to the center by the difference between the number of data points of the photoacoustic signal after Fourier transform and the optical pulse differential waveform to the photoacoustic signal subjected to Fourier transform. To do. The zero padding means 48 divides, for example, a photoacoustic signal (frequency region) having 1024 data points into two at the center frequency of the frequency band, and zeros by the difference in the number of data points between the two divided frequency regions. A point is added to generate a photoacoustic signal having the same number of data points 8192 as the number of data points of the optical pulse differential waveform (frequency domain). The addition of the zero point corresponds to upsampling in the frequency domain.

  The filter applying unit 44 applies an inverse filter to the signal that has been subjected to zero padding by the zero padding unit 48. The zero point removing unit 49 removes the frequency band to which “0” is added by the zero padding unit 48 from the signal to which the inverse filter is applied. For example, when the photoacoustic signal (frequency domain) having 1024 data points is converted into a signal having 8192 data points by the zero padding unit 48, the zero point removing unit 49 uses the signal after applying the filter (data points 8192 points). ) To a signal having 1024 data points. Zero point removal corresponds to downsampling in the frequency domain. The inverse Fourier transform means 45 converts the signal returned to the number of data points of 1024 from a frequency domain signal to a time domain signal.

  FIG. 11A shows a photoacoustic signal subjected to Fourier transform, and FIG. 11B shows a photoacoustic signal after zero padding. For example, when the sampling rate of the photoacoustic signal in the AD conversion means 22 (FIG. 4) is 40 MHz, a signal obtained by Fourier transforming the photoacoustic signal is a signal in a frequency band from 0 MHz to 40 MHz as shown in FIG. 11A. It becomes. This signal is divided into two regions A and B, with 20 MHz being the Nyquist frequency (1/2 of the sampling frequency) as a boundary. As shown in FIG. 11B, the zero padding means 48 inserts 8192-1024 = 7168 zero points between two regions. As a result of adding the zero point, the signal in the region B becomes a signal corresponding to the frequency region from 300 MHz to 320 MHz.

  In the present embodiment, a photoacoustic signal sampled at a low sampling rate is converted into a frequency domain signal, and a zero point in the high frequency component area of the converted frequency domain signal is added. The difference between the present embodiment and the second embodiment is that, in the second embodiment, the photoacoustic signal is upsampled, whereas in the present embodiment, the photoacoustic signal is upsampled in the frequency domain. Even when resampling (upsampling) is performed so as to fill the band difference between both signals in the frequency domain instead of the time domain, low-speed sampling is performed from photoacoustic detection to reconstruction as in the second embodiment. However, the optical pulse differential term can be accurately deconvolved.

  Subsequently, a fourth embodiment of the present invention will be described. Also in the present embodiment, as in the second and third embodiments, the sampling rate of the optical pulse differential waveform is set higher than the sampling rate of the photoacoustic signal. In the present embodiment, the Fourier transform of the optical pulse differential waveform is performed with more data points than the Fourier transform data points of the photoacoustic signal, the high frequency component sample points are removed from the Fourier transformed optical pulse differential waveform, and the inverse number thereof. Is obtained as an inverse filter. Other points may be the same as in the first embodiment.

  FIG. 12 shows the optical pulse differential waveform reverse convolution means 26c in the present embodiment. The optical pulse differential waveform deconvolution means 26c in this embodiment has a high frequency component sample point removal means 50 in addition to the configuration of the optical pulse differential waveform deconvolution means 26 in the first embodiment shown in FIG. For example, it is assumed that the sampling rate (first sampling rate) of the photoacoustic signal is 40 MHz and the sampling rate (second sampling rate) of the optical pulse waveform (optical pulse differential waveform) is 320 MHz. For example, the Fourier transform unit 41 converts a photoacoustic signal of 40 MHz into a signal in the frequency domain of 1024 points (first data point), and the Fourier transform unit 42 converts the optical pulse differential waveform corresponding to 320 MHz to 8192 points (second). Frequency domain signal). The second data score is equal to or greater than the data score obtained by multiplying the first data score by the ratio of the second sampling rate and the first sampling rate.

  The high frequency component sample point removing means 50 receives the optical pulse differential waveform converted from the Fourier transform means 42 into a frequency domain signal. The high frequency component sample point removing means 50 removes the high frequency component sample points from the Fourier transformed optical pulse differential waveform by the difference between the number of data points of the photoacoustic signal after Fourier transformation and the optical pulse differential waveform. The high frequency component sample point removing means 50 deletes the central data point corresponding to the high frequency component from, for example, the optical pulse differential waveform (frequency region) having 8192 data points, and the same data as the number of data points of the photoacoustic signal (frequency region). An optical pulse differential waveform having 1024 points is generated. The removal of the high frequency component sample points corresponds to downsampling of the optical pulse differential waveform in the frequency domain.

  FIG. 13A shows an optical pulse differential waveform obtained by Fourier transform, and FIG. 13B shows an optical pulse differential waveform from which high-frequency component sample points have been removed. For example, when the sampling rate of the optical pulse differential waveform is 320 MHz, a signal obtained by Fourier transforming the optical pulse differential waveform (8192 data points) is a signal in a frequency band from 0 MHz to 320 MHz as shown in FIG. 13A. Become. This signal is divided into the region from the first data point to the 512th region (region A), the region from the 513th data point to the 7680th data point (region B), and the 8192nd from the 7681st data point. The area is divided into three areas up to the data point (area C), and the data points in area B are removed. As shown in FIG. 13B, by connecting region A and region C, an optical pulse differential waveform (frequency region) having 1024 data points corresponding to the frequency band from 0 MHz to 40 MHz can be obtained.

  The inverse filter calculation means 43 obtains the inverse of the optical pulse differential waveform (frequency domain) from which the high frequency component sample points have been removed as an inverse filter. The inverse filter calculation unit 43 obtains, as an inverse filter, the inverse of the optical pulse differential waveform in which the data points are reduced from 8192 points to 1024 points, for example. The filter application unit 44 multiplies, for example, a photoacoustic signal (frequency domain) having 1024 data points and an inverse filter for each element. The Fourier inverse transform means 45 transforms the signal to which the inverse filter is applied from a frequency domain signal to a time domain signal.

  Here, in the third embodiment, the filter application unit 44 includes a photoacoustic signal (frequency region) in which a zero point is added to the region of the high frequency component shown in FIG. 11B and an optical pulse differential waveform (frequency region) shown in FIG. 13A. ) And the inverse of. Since the value of the high frequency component region of the photoacoustic signal is “0”, the high frequency component of the optical pulse differential waveform (region B in FIG. 13A) does not affect the photoacoustic signal after application of the inverse filter. Therefore, as in this embodiment, the high frequency component sample points are removed from the frequency domain signal of the optical pulse differential waveform, an inverse filter is obtained from the optical pulse differential waveform from which the high frequency component has been removed, and the obtained inverse filter is used as the photoacoustic signal. Even when applied to (frequency domain), the obtained result is the same as that of the third embodiment. That is, also in this embodiment, the same effect as the third embodiment can be obtained.

  Subsequently, a fifth embodiment of the present invention will be described. FIG. 14 shows a photoacoustic image generation apparatus according to the fifth embodiment of the present invention. The ultrasonic unit 12a in the photoacoustic image generation apparatus 10a of the present embodiment includes a transmission control circuit 33 and data separation in addition to the configuration of the ultrasonic unit 12 in the photoacoustic image generation apparatus 10 of the first embodiment shown in FIG. Means 34, ultrasonic image reconstruction means 35, detection / logarithm conversion means 36, ultrasonic image construction means 37, and image composition means 38 are provided. The photoacoustic image generation apparatus 10a of the present embodiment is different from the photoacoustic image generation apparatus 10 of the first embodiment in that an ultrasonic image is generated in addition to the photoacoustic image. In the present embodiment, an ultrasonic wave is used as an acoustic wave. However, an acoustic wave having an audible frequency may be used by selecting an appropriate frequency according to a test object, measurement conditions, and the like. Further, this embodiment may be combined with any one of the second to fourth embodiments, and an ultrasonic image may be generated in those embodiments.

  In the present embodiment, in addition to detecting a photoacoustic signal, the probe 11 performs output (transmission) of ultrasonic waves to the subject and detection (reception) of reflected ultrasonic waves from the subject with respect to the transmitted ultrasonic waves. When generating an ultrasonic image (reflected acoustic wave image), the trigger control circuit 29 sends an ultrasonic transmission trigger signal to the transmission control circuit 33 to instruct ultrasonic transmission. When receiving the trigger signal, the transmission control circuit 33 transmits an ultrasonic wave from the probe 11. The probe 11 detects the reflected ultrasonic wave from the subject after transmitting the ultrasonic wave. Transmission and reception of ultrasonic waves may be separated. For example, ultrasonic waves may be transmitted from a position different from the probe 11, and reflected ultrasonic waves with respect to the transmitted ultrasonic waves may be received by the probe 11.

  The reflected ultrasonic wave detected by the probe 11 is input to the AD conversion means 22 via the receiving circuit 21. The trigger control circuit 29 sends a sampling trigger signal to the AD conversion means 22 in synchronization with the timing of ultrasonic transmission to start sampling of reflected ultrasonic waves. Here, the reflected ultrasonic waves reciprocate between the probe 11 and the ultrasonic reflection position, whereas the photoacoustic signal is one way from the generation position to the probe 11. Since the detection of the reflected ultrasonic wave takes twice as long as the detection of the photoacoustic signal generated at the same depth position, the sampling clock of the AD conversion means 22 is half the time when the photoacoustic signal is sampled, for example, It may be 20 MHz. The AD conversion means 22 stores the reflected ultrasound sampling data in the reception memory 23. Either detection (sampling) of the photoacoustic signal or detection (sampling) of the reflected ultrasonic wave may be performed first.

  The data separation unit 34 separates the photoacoustic signal sampling data and the reflected ultrasound sampling data stored in the reception memory 23. The data separation unit 34 inputs the sampling data of the separated photoacoustic signal to the photoacoustic image reconstruction unit 24. The measurement of the optical pulse waveform and the generation of the photoacoustic image (absorption distribution image) including the deconvolution of the optical pulse differential waveform are the same as in the first embodiment. The data separation unit 34 inputs the separated reflected ultrasound sampling data to the ultrasound image reconstruction unit 35.

  The ultrasonic image reconstruction unit 35 generates data of each line of the ultrasonic image based on the reflected ultrasonic waves (its sampling data) detected by the plural ultrasonic transducers of the probe 11. For the generation of the data of each line, a delay addition method or the like can be used as in the generation of the data of each line in the photoacoustic image reconstruction means 24. The detection / logarithm conversion means 36 obtains the envelope of the data of each line output from the ultrasonic image reconstruction means 35 and logarithmically transforms the obtained envelope.

  The ultrasonic image construction unit 37 generates an ultrasonic image based on the data of each line subjected to logarithmic transformation. The ultrasonic image reconstruction unit 35, the detection / logarithm conversion unit 36, and the ultrasonic image construction unit 37 include an ultrasonic image generation unit (reflection acoustic wave image generation unit) that generates an ultrasonic image based on the reflected ultrasonic wave. Configure. The image synthesizing unit 38 synthesizes the photoacoustic image and the ultrasonic image. The image composition unit 38 performs image composition by superimposing a photoacoustic image and an ultrasonic image, for example. The synthesized image is displayed on the image display means 14. It is also possible to display the photoacoustic image and the ultrasonic image side by side on the image display means 14 without performing image synthesis, or to switch between the photoacoustic image and the ultrasonic image.

  In the present embodiment, the photoacoustic image generation apparatus generates an ultrasonic image in addition to the photoacoustic image. By referring to the ultrasonic image, a portion that cannot be imaged in the photoacoustic image can be observed. Similar to the first embodiment, an absorption distribution can be imaged by generating an optical pulse differential waveform representing the differential waveform of the pulsed laser light irradiated to the subject and deconvoluting the optical pulse differential waveform from the observed waveform. It is. In addition, the generation of ultrasonic images and the generation of photoacoustic images can share most of the algorithms such as image reconstruction, detection and logarithmic conversion, and the FPGA circuit configuration and software can be simplified. Has the above advantages.

  Subsequently, a sixth embodiment of the present invention will be described. FIG. 15 shows a photoacoustic image generation apparatus according to the sixth embodiment of the present invention. The present embodiment is different from the first embodiment in that a subject is irradiated with light having a plurality of wavelengths. In addition to the configuration of the ultrasonic unit 12 in the photoacoustic image generation apparatus 10 of the first embodiment shown in FIG. 4, the ultrasonic unit 12b in the photoacoustic image generation apparatus 10b of the present embodiment is a photoacoustic for light of a plurality of wavelengths. Two-wavelength data calculation means 39 for calculating a signal (photoacoustic image) is provided. The present embodiment may be combined with any one of the second to fifth embodiments, and in those embodiments, light of a plurality of wavelengths may be irradiated and a photoacoustic signal (photoacoustic image) for the plurality of wavelengths may be calculated.

  In the present embodiment, the laser unit 13 is configured to be able to switch and emit light having a plurality of wavelengths. For example, the laser unit 13 switches and emits a pulse laser beam having a wavelength of 750 nm and a pulse laser beam having a wavelength of 800 nm. The probe 11 detects a photoacoustic signal from the subject after the emission of the pulsed laser beam of each wavelength, and the reception memory 23 stores sampling data of the photoacoustic signal corresponding to each wavelength. The stored photoacoustic signal corresponding to each wavelength is reconstructed by the photoacoustic image reconstruction means.

  After the reconstruction by the photoacoustic image reconstruction unit 24, the photodifferential waveform deconvolution unit 26, from the photoacoustic signal (photoacoustic image) corresponding to each wavelength, the light intensity of the light of each wavelength irradiated to the subject. Each of the differential waveforms (optical differential waveform) of the time waveform is deconvolved. The photoacoustic signal obtained by deconvolution of the optical differential waveform corresponding to each wavelength is processed by the two-wavelength data calculation means 39.

  Here, in many living tissues, the light absorption characteristics vary depending on the wavelength of light, and generally, the light absorption characteristics are also unique to each tissue. For example, the molecular absorption coefficient at a wavelength of 750 nm of oxygenated hemoglobin (hemoglobin combined with oxygen: oxy-Hb) contained in a large amount of human arteries is also low at a wavelength of 800 nm. Moreover, the molecular absorption coefficient at a wavelength of 750 nm of deoxygenated hemoglobin (hemoglobin deoxy-Hb not bound to oxygen) contained in a large amount in veins is higher than that at a wavelength of 800 nm. Using this property, by examining whether the photoacoustic signal obtained at a wavelength of 750 nm is relatively large or small with respect to the photoacoustic signal obtained at a wavelength of 800 nm, the photoacoustic signal from the artery and the vein From the photoacoustic signal.

  The two-wavelength data calculation means 39 compares the relative magnitude relationship between photoacoustic signals corresponding to a plurality of wavelengths, for example. Specifically, the two-wavelength data calculation means 39 compares the photoacoustic signal detected when light with a wavelength of 750 nm is irradiated with the photoacoustic signal detected when light with a wavelength of 800 nm is irradiated. And how big is it. When displaying an image, if the photoacoustic signal detected when the light having a wavelength of 750 nm is irradiated is large, it can be determined as a photoacoustic signal from a vein. In addition, if the photoacoustic signal detected when the light having the wavelength of 800 nm is irradiated is large, it can be determined as a photoacoustic signal from the artery, so that the portion may be displayed in red.

  In the present embodiment, the two-wavelength data calculation means 39 calculates a photoacoustic signal corresponding to a plurality of wavelengths after deconvolution of the optical differential waveform. When irradiating a subject with light of a plurality of wavelengths, for example, after detecting a photoacoustic signal that irradiates the subject with light of the first wavelength, a photoacoustic signal is emitted by irradiating the subject with the second light. When detecting, position shift may occur between wavelengths due to the influence of body movement. When comparing the photoacoustic signals between a plurality of wavelengths, it is preferable to compare the photoacoustic signals generated from the same place. When the optical differential waveform is not deconvolved, one blood vessel is displayed twice as shown in FIG. 3A, and it is difficult to confirm the position of the blood vessel for image determination, and it is difficult to correct the displacement. By deconvolution of the optical differential waveform, the light absorption distribution can be imaged as shown in FIG. 3B, the position of the blood vessel can be easily confirmed, and the positional deviation can be easily corrected.

  In addition, when one blood vessel is displayed twice as shown in FIG. 3A, there is no signal in the portion corresponding to the inside of the blood vessel (the signal level is lower than a predetermined level), and a plurality of wavelengths If there is a positional shift between the images, the portion where the signal exists, that is, the overlapping portion of the blood vessel is reduced. In this case, it is difficult to appropriately compare the relative magnitude relationships of the photoacoustic signals between a plurality of wavelengths. On the other hand, when the light absorption distribution is imaged as shown in FIG. 3B, many portions in the blood vessel are portions where signals are present, and even if there is a slight misalignment, there is a difference between images of a plurality of wavelengths. Many parts overlap. Therefore, in the present embodiment, even when both images are compared without being aligned, it is possible to reduce the influence of misalignment.

  In each of the above embodiments, the photoacoustic signal and the optical pulse differential waveform are converted into a frequency domain signal and returned to a time domain signal after deconvolution in the frequency domain. However, the present invention is not limited to this. It is also possible to perform deconvolution of the optical pulse differential waveform in the time domain. Further, the optical pulse differential waveform deconvolution means 26 may perform a process of applying some filter to the photoacoustic signal at the time of deconvolution. For example, the optical pulse differential waveform deconvolution means 26 may filter the noise amplification frequency band at the time of deconvolution. Or you may deconvolute what multiplied the device response function (a piezoelectric element or an electric frequency filter) to the optical pulse differential waveform.

  In each of the above embodiments, the photoacoustic image (absorption distribution image) is generated after deconvolution of the photodifferential waveform from the photoacoustic signal, but in addition to or instead of this, the photodifferential waveform is deconvolved. A photoacoustic image (pressure distribution image) may be generated without the volume. For example, the user can select whether or not to perform the deconvolution process by performing an operation on the switch or the display monitor, and when the user selects to perform the deconvolution process, the photodifferential waveform is deconvolved. When the photoacoustic image is generated above and the user selects not to perform the deconvolution process, the photoacoustic image may be generated without performing the deconvolution of the photodifferential waveform. For example, when deconvolution of the photodifferential waveform is performed, the photoacoustic signal is displayed in association with red and black colors. When there is no deconvolution, the photoacoustic signal is associated with blue and black colors. It may be displayed.

  Further, a photoacoustic image without deconvolution is generated, and the computer analyzes the photoacoustic image to determine whether or not the blood vessel portion is divided into two, and the blood vessel is divided into two. When it is determined that, the deconvolution processing of the optical differential waveform may be performed only on the blood vessel portion. At that time, the display color of the blood vessel part that has undergone the deconvolution process is different from the display color of the other unprocessed blood vessel part, and the blood vessel that has undergone the deconvolution process and the other unprocessed blood vessel It may be possible to easily discriminate.

  Although the present invention has been described based on the preferred embodiment, the photoacoustic signal processing apparatus and method of the present invention are not limited to the above embodiment, and various modifications can be made from the configuration of the above embodiment. Further, modifications and changes are also included in the scope of the present invention.

10: Photoacoustic image generation apparatus (photoacoustic image diagnostic apparatus)
11: Probe 12: Ultrasonic unit 13: Laser unit 14: Image display means 21: Reception circuit 22: AD conversion means 23: Reception memory 24: Photoacoustic image reconstruction means 25: Optical waveform measurement means 26: Optical pulse differential waveform Reverse convolution means 27: detection / logarithmic conversion means 28: photoacoustic image construction means 29: trigger control circuit 30: control means 31: flash lamp 32: Q switch 33: transmission control circuit 34: data separation means 35: ultrasonic image Reconstruction means 36: detection / logarithmic conversion means 37: ultrasonic image construction means 38: image synthesis means 39: two-wavelength data calculation means 40: differential waveform calculation means 41, 42: Fourier transform means 43: inverse filter calculation means 44: Filter applying means 45: Fourier inverse transform means 46, 47: Resample means 48: Zero padding means 49: Zero point removing means 0: high-frequency component sample point removal means 251: the light intensity detection section 252: amplifier 253: AD converter 254: waveform memory

Claims (18)

Sampling means for sampling a photoacoustic signal generated in the subject by the light irradiated on the subject;
An optical waveform measuring means for measuring a time waveform of the light intensity of the light irradiated on the subject;
A photoacoustic signal comprising photodifferential waveform deconvolution means for generating a signal obtained by deconvoluting a differential waveform of a time waveform measured by the optical waveform measurement means from the sampled photoacoustic signal. Processing equipment.   The photoacoustic signal processing apparatus according to claim 1, wherein the optical waveform measurement unit includes a light intensity detection unit that detects a light intensity of light applied to the subject.   The light intensity detection unit detects the light intensity of light branched from the optical path from the light emitting end of the light source of the light irradiated to the subject until the subject is irradiated with light. The photoacoustic signal processing apparatus according to claim 2, wherein:   The photoacoustic signal processing apparatus according to claim 3, further comprising a branching unit that branches a part of the light irradiated to the subject in the optical path. The sampling means samples a plurality of photoacoustic signals generated by a plurality of times of light irradiation,
The optical waveform measuring means measures a time waveform of light intensity in each of the plurality of times of light,
5. The optical differential deconvolution means deconvolutes a differential waveform of an average waveform of the plurality of measured time waveforms from a signal obtained by averaging the plurality of photoacoustic signals. The photoacoustic signal processing apparatus in any one. The optical differential waveform deconvolution means comprises:
First Fourier transform means for Fourier transforming the sampled photoacoustic signal;
Second Fourier transform means for Fourier transforming the signal indicating the optical differential waveform;
An inverse filter computing means for obtaining an inverse filter of the inverse of the Fourier transformed optical differential waveform;
Filter applying means for applying the inverse filter to the Fourier-transformed photoacoustic signal;
Fourier inverse transform means for performing Fourier inverse transform on the photoacoustic signal to which the inverse filter is applied;
The photoacoustic signal processing device according to claim 1, wherein the photoacoustic signal processing device is provided. The photoacoustic signal is sampled at a first sampling rate, and the measured optical differential waveform is sampled at a second sampling rate higher than the first sampling rate;
Re-sampling means for re-sampling the photoacoustic signal sampled at the first sampling rate at the second sampling rate;
The photoacoustic signal processing apparatus according to claim 6, wherein the first Fourier transform unit performs a Fourier transform on the photoacoustic signal resampled by the resample unit. The photoacoustic signal is sampled at a first sampling rate, the measured optical differential waveform is sampled at a second sampling rate higher than the first sampling rate, and the first Fourier transform is performed. Means performs a Fourier transform with a first number of data points, and the second Fourier transform means performs a Fourier transform with a second number of data points greater than the first number of data points;
Zero padding means for performing zero padding for adding zero to the center of the Fourier-transformed photoacoustic signal by the difference between the first data point and the second data point;
The photoacoustic signal processing apparatus according to claim 6, wherein the filter applying unit applies the inverse filter to a signal that has been subjected to zero padding by the zero padding unit. The photoacoustic signal is sampled at a first sampling rate, the measured optical differential waveform is sampled at a second sampling rate higher than the first sampling rate, and the first Fourier transform is performed. Means performs a Fourier transform with a first number of data points, and the second Fourier transform means performs a Fourier transform with a second number of data points greater than the first number of data points;
High frequency component sample point removing means for removing high frequency component sample points from the Fourier transformed optical differential waveform by the difference between the first data point and the second data point,
7. The photoacoustic signal processing according to claim 6, wherein the inverse filter calculation means obtains an inverse number of a signal obtained by removing high-frequency component sampling points from the Fourier-transformed optical pulse differential waveform as an inverse filter. apparatus.   The photoacoustic signal processing apparatus according to any one of claims 7 to 9, wherein a sampling interval of the photoacoustic signal is longer than a pulse time width of light irradiated on the subject. Further comprising photoacoustic signal reconstruction means for reconstructing the photoacoustic signal based on photoacoustic signals detected by a plurality of detection elements and sampled by the sampling means;
11. The optical differential waveform deconvolution means deconvolves the photo differential waveform from the photoacoustic signal reconstructed by the photoacoustic signal reconstruction means. The photoacoustic signal processing apparatus of description.   12. The photoacoustic according to claim 1, wherein the optical differential waveform deconvolution means deconvolutes the photoacoustic signal obtained by multiplying the optical differential waveform by a device response function. Signal processing device.   13. The photoacoustic signal processing apparatus according to claim 1, wherein the light irradiated on the subject is a pulsed laser beam. The light emitted to the subject includes light of a plurality of wavelengths, the sampling means samples a photoacoustic signal corresponding to the light of each wavelength, and the photodifferential waveform reverse convolution means corresponds to the light of each wavelength. The signal is generated by deconvolution of the optical differential waveform from the signal.
The photoacoustic according to any one of claims 1 to 13, further comprising two-wavelength data computing means for computing a signal obtained by deconvolution of a photodifferential waveform from a photoacoustic signal corresponding to light of each wavelength. Signal processing device.   15. The photoacoustic signal processing apparatus according to claim 1, further comprising photoacoustic image generation means for generating a photoacoustic image based on a signal obtained by deconvolution of the photodifferential waveform from the photoacoustic signal. . The sampling means further samples a reflected acoustic wave with respect to an acoustic wave transmitted to the subject,
Reflected acoustic wave image generation means for generating a reflected acoustic wave image based on the sampled reflected acoustic wave;
The photoacoustic signal processing apparatus according to claim 15, further comprising image synthesis means for synthesizing the photoacoustic image and the reflected acoustic wave image.   The photoacoustic signal processing apparatus according to claim 16, wherein the image synthesis unit performs image synthesis by superimposing the photoacoustic image and a reflected acoustic wave image. Detecting a photoacoustic signal generated in the subject due to light emission to the subject;
Measuring a time waveform of light intensity of the irradiated light;
And deconvoluting a photodifferential waveform obtained by differentiating the time waveform of the measured light from the detected photoacoustic signal.