Photoacoustic spectrum-confocal Raman spectrum multi-view-field high-dimensional imaging method and device

Abstract

The invention discloses a photoacoustic spectrum-confocal Raman spectrum multi-view-field high-dimensional imaging method and a photoacoustic spectrum-confocal Raman spectrum multi-view-field high-dimensional imaging device. Nanosecond pulse laser with adjustable wavelength is generated through a continuous spectrum nanosecond pulser and a narrow-band spectrum tuning module and cooperates with a microscopic detection module, a depth imaging module, a galvanometer system and an electric displacement table to obtain spatial three-dimensional data, a photoacoustic spectrum and a Raman spectrum under multiple view fields; shallow photoacoustic spectrum-Raman spectrum tomography and deep three-dimensional photoacoustic imaging are realized; correcting a photoacoustic spectrum quantitative error by using a spectrum decoupling algorithm, and fusing the multi-view-field data to generate multi-view-field high-dimensional photoacoustic spectrum-Raman spectrum data; high-dimensional data features are extracted based on a 3D convolution instance segmentation network, and quantitative analysis and spatial distribution visualization of blood oxygen, inflammatory factors and the like are achieved. According to the method, the limitation that substances are difficult to analyze in photoacoustic imaging is overcome, the contradiction between the imaging resolution and the penetration depth is solved, and an accurate analysis means is provided for diagnosis of diseases such as knee osteoarthritis and sepsis.

Description

Technical Field

[0001] The present invention belongs to the field of biomedical imaging, and particularly relates to a photoacoustic spectroscopy-confocal Raman spectroscopy multi-field high-dimensional imaging method and device. Background Art

[0002] In recent years, photoacoustic imaging technology has been widely used in the field of biomedical detection due to its excellent penetration imaging ability and optical resolution. However, traditional photoacoustic imaging technology still has the following technical bottlenecks: (1) Traditional photoacoustic imaging using single-wavelength lasers is difficult to analyze the material composition of complex biological tissues, especially in the identification and quantitative analysis of trace components such as inflammatory factors and oxidative substances; (2) There is an inherent contradiction between high-resolution microscopic imaging and deep tissue penetration ability, making it difficult to balance the analysis of the fine structure of shallow tissues and the three-dimensional distribution of deep lesions; (3) Existing multimodal imaging systems mostly adopt a time-sharing independent acquisition method, resulting in poor spatio-temporal matching of spectral data and making it difficult to construct a multi-parameter correlation analysis model.

[0003] Although confocal Raman spectroscopy technology can provide molecular fingerprint information, its penetration depth is usually limited to the sub-millimeter level, and a general high-power continuous laser light source is likely to cause photodamage to biological samples. Although there are attempts in the prior art to combine photoacoustic and Raman technologies, there are generally problems such as low system integration and lack of data fusion algorithms, and it is impossible to achieve comprehensive analysis of multiple scales and multiple parameters. Especially in the early diagnosis of knee osteoarthritis, existing methods are difficult to simultaneously obtain information on abnormal oxygen metabolism and inflammatory factor distribution related to cartilage degeneration; in the detection of sepsis, there is a lack of multi-parameter dynamic monitoring means for microcirculation disorders and oxidative stress status. Summary of the Invention

[0004] In order to overcome the problems in the prior art, the purpose of the present invention is to provide a photoacoustic spectroscopy-confocal Raman spectroscopy multi-field high-dimensional imaging method and device.

[0005] The technical solution for the present invention to achieve its purpose is as follows: A photoacoustic spectroscopy-confocal Raman spectroscopy multi-field high-dimensional imaging method, in which the wide-spectrum nanosecond pulsed laser emitted by a continuous-spectrum nanosecond pulsed laser is adjusted by a narrow-band spectral tuning module to generate a tunable-wavelength nanosecond pulsed laser; the nanosecond pulsed laser is transmitted through a series of optical elements, passes through a microscopic imaging detection module and a depth imaging detection module, and is focused on a biological sample placed in a water tank; the nanosecond pulsed laser is focused on the biological sample to generate Raman scattering, which is collected by a high-throughput Raman spectroscopy detection module to achieve pulsed Raman spectroscopy detection; the nanosecond pulsed laser causes the thermal expansion of the biological sample and then rapidly contracts to generate a photoacoustic signal; the nanosecond pulsed laser is tuned to different wavelengths by the narrow-band spectral tuning module to achieve photoacoustic spectroscopy detection.

[0006] The method described above includes the following steps: a) High-dimensional photoacoustic spectroscopy-Raman spectroscopy imaging; b) Multi-field photoacoustic spectroscopy imaging; c) According to the spectral decoupling method, use Raman spectroscopy to correct the errors in characterizing the types and contents of substances by photoacoustic spectroscopy, improve the accuracy of photoacoustic spectroscopy for substance analysis, and achieve multi-field high-dimensional data fusion.

[0007] The high-dimensional photoacoustic spectroscopy-Raman spectroscopy imaging described above: Through the microscopic imaging detection module, the nanosecond pulsed laser is focused on the position of the biological sample to excite the Raman spectroscopy signal at the millimeter-depth position of the biological sample, and pulsed Raman spectroscopy detection is achieved; Use the narrowband spectral tuning module to tune the nanosecond pulsed laser to different wavelengths to excite the photoacoustic signal at the millimeter-depth position of the biological sample, and photoacoustic spectroscopy detection of the shallow position is achieved; Photoacoustic spectroscopy detection realizes the detection of the content and spatial distribution of substances including water, blood oxygen, and fat substances according to the specific absorption differences of substances for different light wavelengths; Raman spectroscopy detection generates molecular fingerprint Raman spectral lines according to the differences in molecular vibration energy levels of substances, and realizes the detection of the content and spatial distribution of mitochondria, neutrophils, DNA genetic materials, and inflammatory factors; The photoacoustic spectroscopy detection and pulsed Raman spectroscopy detection share the excitation light path, and the excited photoacoustic spectroscopy signal and Raman signal correspond one by one; Combining the two-dimensional scanning of the galvanometer system and the depth tomography of the motorized Z-axis displacement stage, high-dimensional photoacoustic spectroscopy-Raman spectroscopy imaging including three-dimensional spatial information, Raman spectroscopy information, and photoacoustic spectroscopy information is achieved; The high-dimensional photoacoustic spectroscopy-Raman spectroscopy imaging combines three-dimensional spatial stereoscopic imaging, the absorption characteristics of photoacoustic spectroscopy and the molecular fingerprint characteristics of Raman spectroscopy, and performs high-dimensional model visualization display.

[0008] The multi-field photoacoustic spectroscopy imaging described above: Use the motorized XY displacement stage to move the sample detection area to the detection window position of the depth imaging module; Through the depth imaging detection module, the nanosecond pulsed laser is focused on the deep position at the centimeter depth of the sample; Use the narrowband spectral tuning module to tune the nanosecond pulsed laser to different wavelengths to excite the photoacoustic signal at the centimeter-depth position of the sample, and photoacoustic spectroscopy detection of the deep position is achieved; Calculate the positions of different sound source points excited within the excitation light path of the biological tissue according to the sound speed in the biological tissue; The computer controls the two-dimensional movement of the motorized XY displacement stage to achieve three-dimensional reconstruction of the centimeter depth of the biological tissue; The deep photoacoustic spectroscopy data detected by the depth imaging detection module is combined with the high-dimensional photoacoustic spectroscopy-Raman spectroscopy data detected by the microscopic imaging detection module to achieve imaging detection that takes into account both penetration ability and high resolution.

[0009] The multi-field high-dimensional data fusion described above: The computer system fuses the photoacoustic spectroscopy signal and the Raman spectroscopy signal at the shallow layer position to form the photoacoustic spectroscopy-Raman spectroscopy data at the shallow layer position; meanwhile, it fuses the photoacoustic spectroscopy-Raman spectroscopy data at the shallow layer position with the photoacoustic spectroscopy data at the deep layer position to form the multi-field high-dimensional photoacoustic spectroscopy-Raman spectroscopy data; according to the spectral decoupling method, the Raman spectroscopy is used to correct the errors in the photoacoustic spectroscopy for characterizing the types and contents of substances, and improve the accuracy of photoacoustic spectroscopy for substance analysis.

[0010] In the method described above, the continuous-spectrum nanosecond pulsed laser outputs tunable-wavelength nanosecond pulsed excitation light through the narrowband spectral tuning module. The nanosecond pulsed excitation light is absorbed by the substances in the biological tissue and generates photoacoustic signals. The computer controls the electric pitching adjustment mirror in the narrowband spectral tuning module to adjust the wavelength of the nanosecond pulsed excitation light to achieve the successive output of nanosecond pulsed excitation light with different wavelengths. The nanosecond pulsed excitation light with different wavelengths is incident on the biological sample in sequence to excite photoacoustic signals and realize photoacoustic spectroscopy detection.

[0011] In the spectral decoupling method described above, the Raman spectroscopy is used to correct the errors in the photoacoustic spectroscopy for characterizing the types and contents of substances; different biological tissues exhibit absorption characteristics at the same wavelength, and their absorption degrees and methods of light are different, resulting in interference in the measurement of photoacoustic spectroscopy. By introducing Raman spectroscopy signals for calibration and using the unique response of Raman spectroscopy to the vibration of different substance molecules, supplementary information is provided for the measurement results of photoacoustic spectroscopy; according to the Raman spectroscopy peak differences of water, oxyhemoglobin, and fat, the types of biological tissue contents are identified, or the errors in the quantitative analysis of substance contents by photoacoustic spectroscopy are corrected; ; Among them, is the content of the substance in the biological tissue m 1 ; is the content of the substance in the biological tissue m 2 ; is the content of the substance in the biological tissue m n ; is the empirical absorbance value of the substance m 1 in the biological tissue, is the empirical absorbance value of the substance m 2 in the biological tissue, is the empirical absorbance value of the biological tissue substance m n ; I photon (X, Y, Z, λ) is the absorbance value converted from the photoacoustic signalV(X, Y, Z, λ) is the intensity of the detected photoacoustic signal, V Input is the intensity of the monochromatic nanosecond pulsed excitation light, and ∂ is the efficiency of converting the optical signal into the ultrasonic signal; According to the absorption peak wavelengths of various substances in biological tissues λ 1 ~ λ n Substitute into the above formula, and the contents of various substances in biological tissues can be obtained according to the photoacoustic spectroscopy data, or the error of quantitatively analyzing the substance content by photoacoustic spectroscopy can be corrected.

[0012] The method described above uses a 3D convolutional instance segmentation deep network model to extract the features of multi-field high-dimensional photoacoustic spectroscopy-Raman spectroscopy data, realizes the detection and visual display of the content and spatial distribution of various physiological substances in biological samples, including blood oxygen, water, fat, inflammatory factors, and oxidative substances, and provides intuitive visual data for clinical physiological analysis; Based on the detection results of the content and spatial distribution of physiological substances, pixel segmentation of normal and diseased regions of biological tissues is realized; Using a deep neural network, the correlation between multi-field high-dimensional photoacoustic spectroscopy-Raman spectroscopy data and different development stages of knee osteoarthritis, sepsis, and colon cancer diseases is established, and the early identification and diagnosis of knee osteoarthritis, sepsis, and colon cancer diseases are realized.

[0013] A photoacoustic spectroscopy-confocal Raman spectroscopy multi-field high-dimensional imaging device adopting the method described above includes a continuous spectrum nanosecond pulsed laser, a narrowband spectrum tuning module, a second lens, a third lens, a second mirror, a beam splitter, a precision galvanometer system, a scanning lens, a telecentric lens, a microscopic imaging detection module, a depth imaging detection module, a water tank, an electric XY displacement stage, a high-throughput Raman detection module, a high-speed acquisition card, and a computer; The supercontinuum spectrum nanosecond pulsed laser is used to emit a wide-spectrum nanosecond pulsed laser in the wavelength range of 400nm - 2000nm. After being adjusted by the narrowband spectrum tuning module, a tunable wavelength nanosecond pulsed laser is generated; The nanosecond pulsed laser passes through the first mirror, the second lens, the third lens, and the second mirror, and is incident on the galvanometer system via the beam splitter. The galvanometer system adjusts the emission direction of the nanosecond pulsed laser, and the nanosecond pulsed laser is incident on the microscopic imaging detection module and the depth imaging detection module through the scanning lens and the telecentric lens respectively; The combination of the scanning lens and the telecentric lens is used to optimize the scanning image plane and realize the beam expansion of the nanosecond pulsed laser; The described narrowband spectral tuning module includes a first parabolic mirror, a second reflection grating, a second parabolic mirror, an electrically controlled pitch-adjustable mirror, a first pinhole, and a first lens. After the first parabolic mirror collimates the broadband nanosecond pulsed laser into parallel light, it is incident on the second reflection grating. Different wavelengths of light in the broadband nanosecond pulsed laser generate different diffraction angles after passing through the second reflection grating, producing diffracted light of different wavelengths. The diffracted light is focused onto the first pinhole after being adjusted by the second parabolic mirror and the electrically controlled pitch-adjustable mirror. The first lens is used to collimate the diffracted light passing through the first pinhole for subsequent use. The computer controls the pitch angle of the electrically controlled pitch-adjustable mirror to adjust the diffracted light passing through the first pinhole, so as to achieve the sequential output of the diffracted light of a specific wavelength and output a nanosecond pulsed laser with an adjustable wavelength. The described microscopic imaging detection module includes a first acoustic medium prism, an adjustable focus lens module, a microscopic objective lens, an electrically controlled Z-axis displacement stage, a first ultrasonic transducer, and a first low-noise signal amplifier. The nanosecond pulsed laser is focused on the shallow position of the biological sample in the water tank after passing through the first acoustic medium prism, the adjustable focus lens module, and the microscopic objective lens. The photoacoustic signal generated is propagated through the first acoustic medium prism and collected by the first ultrasonic transducer. At the same time, the Raman spectral signal generated returns along the original optical path, and after passing through the beam splitter, it is collected by the high-throughput Raman spectral detection module. The described depth imaging detection module includes a second acoustic medium prism, a second ultrasonic transducer, a second low-noise signal amplifier, a fourth lens, and a fifth lens. After the nanosecond pulsed laser passes through the first acoustic medium prism, a part of the light beam is incident on the second acoustic medium prism and focused on the deep position of the biological sample in the water tank after passing through the fourth lens and the fifth lens. The photoacoustic signal generated is propagated through the second acoustic medium prism and collected by the second ultrasonic transducer. The described high-throughput Raman spectral detection module includes a trap filter, a focusing lens, a coded aperture mask, two pinholes, a first free-form surface mirror, a first reflection grating, a second free-form surface mirror, and a high-speed camera. After the signal light returned by the microscopic imaging detection module passes through the trap filter, the trap filter filters out the excitation light. The coded aperture mask is used to overcome the problem of the mutual limitation of the Raman spectral resolution and the optical flux, and achieve a hyperspectral resolution of sub-nanometer level for the spectral resolution under a wide slit and high optical flux. The slit is used to block the stray light outside the focal plane of the microscopic objective lens, improve the signal contrast, and achieve confocal detection. The light emerging from the slit is collimated by the first free-form surface mirror, then diffracted by the reflection grating, and finally focused by the second free-form surface mirror to focus the Raman signal onto the high-speed camera, generating a clear Raman spectral image. The combination of the reflection surface parameters of the first free-form surface mirror and the second free-form surface mirror is designed to eliminate the phase difference and chromatic aberration in Raman spectral imaging.

[0014] The computer is connected to a high-speed acquisition card, a high-speed camera, a galvanometer system, an electric Z-axis displacement stage, an electric XY displacement stage, and an electric pitch-adjustable mirror through signal cables respectively; the computer controls the excitation, acquisition, and data processing of photoacoustic signals and Raman signals at different depth positions in the microscopic imaging detection module and the depth imaging detection module by controlling the galvanometer system, the electric Z-axis displacement stage, the electric XY displacement stage, and the high-speed acquisition card, and generates high-resolution multi-field high-dimensional photoacoustic spectrum-Raman spectrum data, so as to realize high-dimensional imaging of different fields of biological samples; The first ultrasonic transducer is connected to the first low-noise signal amplifier, and the second ultrasonic transducer is connected to the second low-noise signal amplifier to amplify the photoacoustic signal; the high-speed acquisition card is connected to the first low-noise signal amplifier and the second low-noise signal amplifier to collect the photoacoustic spectrum signal; the high-speed acquisition card transmits the photoacoustic spectrum signal to the computer for backend data processing.

[0015] Advantages of the present invention: A proposed multi-field high-dimensional imaging method for photoacoustic spectrum-confocal Raman spectrum overcomes the contradictions that traditional biological sample detection requires in vitro sampling for precise analysis, and photoacoustic imaging cannot analyze the species content and distribution of biological tissues, as well as the conflict between imaging resolution and depth. The present invention can non-destructively detect multi-field high-dimensional photoacoustic spectrum-Raman spectrum, and based on a 3D convolutional instance segmentation network, extract high-dimensional data features to realize quantitative analysis and spatial distribution visualization of blood oxygen and inflammatory factors, etc.; the present invention breaks through the limitations of traditional single-wavelength photoacoustic imaging. By combining a continuous spectrum nanosecond pulse generator and a narrowband spectrum tuning module, a nanosecond pulse laser with adjustable wavelength is generated, and the photoacoustic detection dimension is extended from single-structure imaging to spectral quantitative analysis. Combining with a spectral decoupling algorithm, the absorption contributions of biological components such as water, oxyhemoglobin, and fat can be accurately decoupled, improving the quantitative accuracy of photoacoustic spectra; the present invention combines multi-field photoacoustic spectrum imaging and confocal reflection Raman spectrum imaging to solve the contradiction that high resolution and deep penetration cannot be achieved simultaneously in traditional technologies, and realizes multi-field high-dimensional photoacoustic spectrum-Raman spectrum imaging including three-dimensional spatial information, Raman spectrum information, and photoacoustic spectrum information; multi-field high-dimensional photoacoustic spectrum-Raman spectrum imaging combines three-dimensional spatial imaging of different fields, the absorption characteristics of photoacoustic spectra, and the molecular fingerprint characteristics of Raman spectra, and performs high-dimensional model visualization display, giving full play to the complementary advantages of photoacoustic spectra and Raman spectra, and providing a more comprehensive analysis of the distribution of biological tissue information in three-dimensional space; through a 3D convolutional neural network, high-dimensional data feature extraction is realized, and an association model between the three-dimensional distribution of blood oxygen and microenvironment inflammatory factors, etc. is established by combining with a deep neural network, improving the recognition accuracy of early osteoarthritis cartilage degeneration detection and sepsis microcirculation disorder. Description of the drawings

[0016] Figure 1It is a schematic flow chart of a photoacoustic spectroscopy - confocal Raman spectroscopy multi - field high - dimensional imaging method.

[0017] Figure 2 It is a photoacoustic Raman spectroscopy detection device.

[0018] In the figure, continuous - light nanosecond pulse laser 1, narrow - band spectral tuning module 2, first parabolic mirror 3, second parabolic mirror 4, first pinhole 5, first lens 6, first mirror 7, second lens 8, third lens 9, second mirror 10, beam - splitting prism 11, galvanometer system 12, scanning lens 13, telecentric lens 14, adjustable - focus lens module 15, microscopic objective lens 16, high - throughput Raman spectroscopy detection module 17, focusing lens 18, second pinhole 19, first free - form surface mirror 20, first reflection grating 21, second free - form surface mirror 22, high - speed camera 23, first acoustic medium prism 24, first ultrasonic transducer 25, high - speed acquisition card 26, computer 27, water tank 28, electric Z - displacement stage 29, electric XY - displacement stage 30, second acoustic medium prism 31, second ultrasonic transducer 32, coded mask template 33, second reflection grating 34, electric pitch - adjustment mirror 35, microscopic detection and imaging module 36, depth - imaging detection module 37, trap filter 38, first low - noise signal amplifier 39, second low - noise signal amplifier 40, fourth lens 41, fifth lens 42. Specific embodiments

[0019] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0020] As Figure 1 shown, for a photoacoustic spectroscopy - confocal Raman spectroscopy multi - field high - dimensional imaging method, the broadband nanosecond pulsed laser emitted by the continuous - spectrum nanosecond pulsed laser, after being adjusted by the narrow - band spectral tuning module, generates a tunable - wavelength nanosecond pulsed laser; the nanosecond pulsed laser is transmitted through a series of optical elements, passes through the microscopic imaging detection module and the depth - imaging detection module, and is focused on the biological sample placed in the water tank; the nanosecond pulsed laser is focused on the biological sample to generate Raman scattering, which is collected by the high - throughput Raman spectroscopy detection module to achieve pulsed Raman spectroscopy detection; the nanosecond pulsed laser causes the thermal expansion and then rapid contraction of the biological sample to generate photoacoustic signals; the narrow - band spectral tuning module is used to tune the nanosecond pulsed laser to different wavelengths to achieve photoacoustic spectroscopy detection.

[0021] The method described above includes the following steps: a) High - dimensional photoacoustic spectroscopy - Raman spectroscopy imaging; b) Multi - field photoacoustic spectroscopy imaging; c) According to the spectral decoupling method, use the Raman spectrum to correct the error of the photoacoustic spectrum in characterizing the species and content of substances, improve the accuracy of photoacoustic spectrum material analysis, and achieve multi - field high - dimensional data fusion.

[0022] The described high-dimensional photoacoustic spectroscopy-Raman spectroscopy imaging uses a microscopic imaging detection module to focus a nanosecond pulsed laser on the position of a biological sample, exciting Raman spectroscopic signals at a millimeter-depth position in the biological sample to achieve pulsed Raman spectroscopy detection; uses a narrowband spectroscopic tuning module to tune the nanosecond pulsed laser to different wavelengths, exciting photoacoustic signals at a millimeter-depth position in the biological sample to achieve photoacoustic spectroscopy detection at a shallow position; photoacoustic spectroscopy detection realizes the detection of the content and spatial distribution of substances including water, blood oxygen, and fat substances based on the specific absorption differences of substances for different light wavelengths; Raman spectroscopy detection generates molecular fingerprint Raman spectral lines based on the differences in the molecular vibration energy levels of substances to realize the detection of the content and spatial distribution of mitochondria, neutrophils, DNA genetic materials, and inflammatory factors; the photoacoustic spectroscopy detection and the pulsed Raman spectroscopy detection share an excitation light path, and the excited photoacoustic spectroscopy signals and Raman signals correspond one by one; combining the two-dimensional scanning of a galvanometer system and the depth tomography of a motorized Z-axis displacement stage to achieve high-dimensional photoacoustic spectroscopy-Raman spectroscopy imaging including three-dimensional spatial information, Raman spectroscopic information, and photoacoustic spectroscopic information; the high-dimensional photoacoustic spectroscopy-Raman spectroscopy imaging combines three-dimensional spatial stereoscopic imaging, the absorption characteristics of photoacoustic spectroscopy, and the molecular fingerprint characteristics of Raman spectroscopy, and performs high-dimensional model visualization display.

[0023] The described multi-field photoacoustic spectroscopy imaging: uses a motorized XY displacement stage to move the sample detection area to the detection window position of the depth imaging module; through the depth imaging detection module, a nanosecond pulsed laser is focused on the deep position at the centimeter depth of the sample; uses a narrowband spectroscopic tuning module to tune the nanosecond pulsed laser to different wavelengths, exciting photoacoustic signals at the centimeter-depth position of the sample to achieve photoacoustic spectroscopy detection at a deep position; calculates the positions of different sound source points excited within the excitation light path of the biological tissue based on the sound speed in the biological tissue; the computer controls the two-dimensional movement of the motorized XY displacement stage to achieve three-dimensional reconstruction of the biological tissue at the centimeter depth; the deep photoacoustic spectroscopy data detected by the depth imaging detection module is combined with the high-dimensional photoacoustic spectroscopy-Raman spectroscopy data detected by the microscopic imaging detection module to achieve imaging detection that takes into account both penetration ability and high resolution.

[0024] The described multi-field high-dimensional data fusion: the computer system fuses the photoacoustic spectroscopy signals and Raman spectroscopy signals at the shallow position to form photoacoustic spectroscopy-Raman spectroscopy data at the shallow position; at the same time, fuses the photoacoustic spectroscopy-Raman spectroscopy data at the shallow position with the photoacoustic spectroscopy data at the deep position to form multi-field high-dimensional photoacoustic spectroscopy-Raman spectroscopy data; according to the spectral decoupling method, uses Raman spectroscopy to correct the errors in the characterization of substance types and contents by photoacoustic spectroscopy to improve the accuracy of photoacoustic spectroscopy substance analysis.

[0025] In the described method, a continuous-spectrum nanosecond pulsed laser outputs nanosecond pulsed excitation light with tunable wavelength through a narrow-band spectral tuning module. The nanosecond pulsed excitation light is absorbed by substances in biological tissue, generating photoacoustic signals. A computer controls the electric pitching adjustment mirror in the narrow-band spectral tuning module to adjust the wavelength of the nanosecond pulsed excitation light, so as to achieve successive output of nanosecond pulsed excitation light with different wavelengths. The nanosecond pulsed excitation light with different wavelengths is incident on a biological sample in sequence, exciting photoacoustic signals to achieve photoacoustic spectroscopy detection.

[0026] In the described spectral decoupling method, Raman spectroscopy is used to correct the errors in characterizing the types and contents of substances by photoacoustic spectroscopy. Different biological tissues exhibit absorption characteristics at the same wavelength, and their absorption degrees and modes of light are different, resulting in interference in the measurement of photoacoustic spectroscopy. By introducing Raman spectral signals for calibration and using the unique response of Raman spectroscopy to the vibration of different substance molecules, supplementary information is provided for the measurement results of photoacoustic spectroscopy. According to the Raman spectral peak differences of water, oxyhemoglobin, and fat, the types and contents of biological tissues are identified, or the errors in quantitatively analyzing the contents of substances by photoacoustic spectroscopy are corrected.

[0027] In the described method, a 3D convolutional instance segmentation deep network model is used to extract the features of multi-field high-dimensional photoacoustic spectroscopy-Raman spectroscopy data, realizing the detection and visual display of the contents and spatial distributions of various physiological substances in a biological sample, including blood oxygen, water, fat, inflammatory factors, and oxidative substances, providing intuitive visual data for clinical physiological analysis. Based on the detection results of the contents and spatial distributions of physiological substances, pixel segmentation of normal and diseased regions of biological tissue is realized, and the individual pixel information of the instance segmentation types can be used to analyze the severity of the progression of the disease. Using a deep neural network, the correlation between multi-field high-dimensional photoacoustic spectroscopy-Raman spectroscopy data and different development stages of knee osteoarthritis, sepsis, and colon cancer diseases is established, realizing the early identification and diagnosis of knee osteoarthritis, sepsis, and colon cancer diseases.

[0028] As Figure 2 shown, a multi-field high-dimensional imaging device for photoacoustic spectroscopy-confocal Raman spectroscopy includes a continuous-spectrum nanosecond pulsed laser 1, a narrow-band spectral tuning module 2, a second lens 8, a third lens 9, a second mirror 10, a beam splitter 11, a galvanometer system 12, a scanning lens 13, a telecentric lens 14, a microscopic imaging detection module 36, a depth imaging detection module 37, a water tank 28, an electric XY displacement stage 30, a high-throughput Raman detection module 17, a high-speed acquisition card 26, and a computer 27.

[0029] The described narrow-band spectral tuning module 2 includes a first parabolic mirror 3, a second reflection grating 34, a second parabolic mirror 4, an electric pitching adjustment mirror 35, a first pinhole 5, a first lens 6, and a first mirror 7.

[0030] The described microscopic imaging detection module 36 includes a first acoustic medium prism 24, an adjustable focus lens module 15, a microscopic objective lens 16, an electric Z-axis displacement stage 29, a first ultrasonic transducer 25, and a first low-noise signal amplifier 39.

[0031] The depth imaging detection module 37 includes a second acoustic medium prism 31, a second ultrasonic transducer 32, a second low-noise signal amplifier 40, a fourth lens 41, and a fifth lens 42.

[0032] The high-throughput Raman spectroscopy detection module includes a trap filter 38, a focusing speculum 18, a coded aperture mask 33, a second pinhole 19, a first free-form surface mirror 20, a first reflection grating 21, a second free-form surface mirror 22, and a high-speed camera 23.

[0033] The computer 27 is connected to a high-speed acquisition card 26, a high-speed camera 23, a galvanometer system 12, an electric Z-axis displacement stage 29, an electric XY displacement stage 30, and an electric pitch-adjustable mirror 35 through signal cables. The computer 27 controls the galvanometer system 12, the electric Z-axis displacement stage 29, and the electric XY displacement stage 30 to excite photoacoustic and Raman signals at different positions.

[0034] The first ultrasonic transducer 25 is connected to the first low-noise signal amplifier 39 through a BNC coaxial cable, and the second ultrasonic transducer 32 is connected to the second low-noise signal amplifier 40 through a BNC coaxial cable to amplify the weak photoacoustic signals. The high-speed acquisition card 26 is connected to the first low-noise signal amplifier 39 and the second low-noise signal amplifier 40 through BNC cables to collect photoacoustic signal data, and then transmit it to the computer for 4D reconstruction of photoacoustic Raman spectroscopy of biological tissues.

[0035] The first ultrasonic transducer 25 uses an ultrasonic transducer with a center frequency of 30 MHz. The higher the center frequency, the higher the resolution of photoacoustic imaging; the second ultrasonic transducer 32 uses an ultrasonic transducer with a center frequency of 5 MHz. The lower the center frequency, the deeper the detection depth of photoacoustic imaging.

[0036] The narrowband spectral tuning module 2: The pulsed white laser emitted by the continuous-wave nanosecond pulse laser 1 is collimated into parallel light by the first parabolic mirror 3 and incident on the second reflection grating 34. Lights of different wavelengths generate different diffraction angles after passing through the second reflection grating 34. After passing through the second parabolic mirror 4 and the motorized pitch-adjustable mirror 35, the focused light of the set wavelength will be focused on the light-transmitting position of the first pinhole 5, and the light not within the set wavelength will be blocked by the pinhole. Thus, the light emerging from the first pinhole 5 is collimated into monochromatic parallel light by the first lens 6 and incident on the backend system after passing through the first mirror 7. The computer 27 adjusts the angles of the motorized pitch-adjustable mirror 35 to adjust different set wavelengths to pass through the first pinhole 6, realizing different narrowband spectral tuning and gating.

[0037] The high-throughput Raman detection module 17: The signal light returned by the microscopic imaging detection module 36 and the depth imaging detection module is filtered by the trap filter 38 to remove the excitation light. The Raman signal light passes through the coded aperture mask 33 and is focused by the focusing lens 18 at the slit position. The coded aperture mask 33 can be used to overcome the problem of mutual limitation between the Raman spectral resolution and the light flux, achieving a hyperspectral resolution of sub-nanometer level with a wide slit and high light flux. The slit can block the light outside the focal plane of the microscopic objective 16, improving the signal contrast and realizing depth tomography. The light emerging from the slit is collimated into parallel light after passing through the first free-form mirror 20. After the parallel light passes through the first reflection grating 21, lights of different wavelengths generate different diffraction angles and are incident on the second free-form mirror 22, and then focused in the high-speed camera 23 to form a sharp Raman image. The Raman image data is transmitted to the computer 27 to form the Raman molecular data in the biological tissue. The design combination of the reflection surface parameters of the first free-form mirror 20 and the second free-form mirror 22 can be used to eliminate the phase difference in Raman spectral imaging.

[0038] The described microscopic imaging detection module 36: The monochromatic pulsed laser is incident on the scanning lens 13 and the telecentric lens 14 after passing through the galvanometer system 12. The combination of the scanning lens 13 and the telecentric lens 14 can optimize the scanning image plane and achieve excitation light beam expansion. After the excitation light passes through the first acoustic medium prism, half of the excitation light is reflected to the adjustable focus lens module and focused by the microscopic objective lens 16 inside the biological tissue in the water tank 28 (about millimeters deep). The nanosecond pulsed excitation light focused on the biological tissue will generate Raman scattering, and at the same time cause rapid contraction after thermal expansion of the biological tissue part, thereby generating ultrasonic signals. The ultrasonic signals pass through the water medium and the first acoustic medium prism and then propagate to the first ultrasonic transducer to form photoacoustic signals, which are collected by the high-speed acquisition card to form photoacoustic data. The Raman scattering returns along the original optical path, passes through the microscopic objective lens 16, the first acoustic medium prism 25, the telecentric lens 14, the scanning lens 13, and the beam splitting prism 11, and then the Raman spectral data is collected by the high-throughput Raman detection module 17. The computer 27 forms two-dimensional photoacoustic imaging and two-dimensional Raman spectral imaging at the focal plane position of the microscopic objective lens 16 by controlling the mirror swing of the galvanometer system 12. The computer 27 forms 3D microscopic photoacoustic and 3D Raman spectral tomography by controlling the electric Z-axis displacement stage to move the focal plane of the microscopic objective lens up and down.

[0039] The described depth imaging detection module 37: The nanosecond pulsed excitation light exits from the first acoustic medium prism, and half of the signal light is incident on the second acoustic medium prism 31 and then reflected into the fourth lens 41 and the fifth lens 42. The fourth lens 41 and the fifth lens 42 form a beam reduction module to reduce the beam diameter and then incident on the deep part (at the 4 - 5 cm position) inside the biological tissue in the water tank 28. In the path of the nanosecond pulsed excitation light from the body surface of the biological tissue to the deep part of the biological tissue, rapid contraction after thermal expansion will occur, thereby forming many sound source points in the excitation light path to emit ultrasonic signals, which pass through the second acoustic medium prism 31 and then propagate to the second ultrasonic transducer 32, and the photoacoustic data within the time sequence is collected by the high-speed acquisition card. The ultrasonic data generated by different sound source points excited in the excitation light path of the biological tissue reach the second ultrasonic transducer 32 at different times. According to the sound speed in the biological tissue, the specific positions of different sound source points excited in the excitation light path of the biological tissue can be calculated. The computer 27 can reconstruct the 3D photoacoustic data deep inside the biological tissue through the two-dimensional movement of the electric XY displacement stage 30.

[0040] The present invention takes into account the penetration ability and high-resolution imaging detection: it combines a depth imaging detection module 37 and a microscopic imaging detection module 36. The depth imaging detection module 37 is used to detect the 3D identification and classification, content distribution, etc. of various molecular substances in the deep part (at the 4-5 cm position) inside the biological tissue in the water tank 28, such as water, oxyhemoglobin, deoxyhemoglobin, fat, etc. The microscopic imaging detection module 36 is used to detect the 4D modeling detection of molecular substances at the shallow part (millimeter-level treatment) inside the biological tissue in the water tank 28. In addition to water, vascular fat, etc. detected by photoacoustic detection, Raman detection can also be combined, such as the identification and classification, content distribution, etc. of neutrophils, DNA genetic material, oxidized substances, etc. The combination of high-resolution imaging detection of the deep part (4-5 cm position) and the shallow part (millimeter-level position) of the biological tissue can provide a non-invasive and refined analysis method for AD / sepsis ophthalmology, osteoarthritis microenvironment detection, and colon cancer detection.

[0041] Regarding the photoacoustic spectroscopy imaging described above: The continuous-wave nanosecond pulsed laser 1 outputs a selected specific single-wavelength nanosecond pulsed excitation light through the narrowband spectral tuning module 2. The single-wavelength nanosecond pulsed excitation light can be absorbed by specific tissues in the deep and shallow parts of the living body, thereby generating photoacoustic signals with single-wavelength absorption in the biological tissue. The computer 27 controls the deflection of the electrically driven pitch-adjustable mirror of the narrowband spectral tuning module 2 to control the variation of the output excitation light wavelength. Thus, excitation lights of different wavelengths are sequentially incident on different substances in the biological tissue to absorb the photoacoustic signals, and photoacoustic spectroscopy imaging can be achieved.

[0042] Regarding the high-dimensional photoacoustic spectroscopy-Raman spectroscopy data model described above: Photoacoustic spectroscopy 3D imaging is based on the specific absorption differences of different wavelengths by physiological substances such as water, vascular distribution, oxyhemoglobin, deoxyhemoglobin, fat, etc. in the deep and shallow parts of the biological tissue to detect different physiological substances in the biological tissue. Raman spectroscopy 3D imaging mainly utilizes the fact that different substances in the biological tissue have different molecular vibration energy levels, thereby generating sharp molecular fingerprint Raman spectral lines. Raman spectroscopy can detect oxidized substances, DNA genetic material, inflammatory factors, etc. in the biological tissue, which cannot be detected by absorption spectroscopy. The excitation light paths for photoacoustic spectroscopy and Raman spectroscopy detection in the biological tissue are shared, and the excited ultrasonic data and Raman data correspond one by one. Through 3D tomography, a high-dimensional model of photoacoustic spectroscopy combined with Raman spectroscopy for detecting the biological tissue can be constructed. The high-dimensional model of photoacoustic spectroscopy combined with Raman spectroscopy can give full play to the complementary advantages of absorption spectroscopy and Raman molecular fingerprint spectroscopy, and detect and visually display the content and spatial distribution of various physiological substances in the biological tissue, including blood oxygen, moisture, fat, inflammatory factors, and oxidized substances.

[0043] In the described pulsed Raman spectroscopy detection, the Raman signal is weak, generally on the order of 10-9. Using continuous light detection will cause biological tissues to be irradiated by high-intensity excitation light for a long time, resulting in optical damage. The pulsed laser has a strong peak signal, and a strong change in the Raman signal will be generated within a short time of nanoseconds, thereby improving the sensitivity of Raman signal detection. At the same time, the nanosecond pulsed signal acts on the organism for a short time, and the phototoxicity to biological tissues is low.

[0044] The described spectral decoupling method: using Raman spectroscopy to correct the errors in the characterization of the types and contents of substances by photoacoustic spectroscopy; different biological tissues exhibit absorption characteristics at the same wavelength, and their absorption degrees and methods of light are different, resulting in interference in the measurement of photoacoustic spectroscopy. Especially when quantitative analysis is carried out on components such as water, hemoglobin, and fat, the error affects the accuracy of the results; by introducing Raman spectral signals for calibration, and using the unique response of Raman spectroscopy to the vibration of different substance molecules, supplementary information can be provided for the measurement results of photoacoustic spectroscopy; according to the Raman spectral peak differences of substances such as water, oxyhemoglobin, and fat, to identify the types of biological tissue contents: ; Among them, is the content of substance m 1 in biological tissue, is the content of substance m 2 in biological tissue, is the content of substance m n in biological tissue, is the empirical absorbance value of substance m 1 in biological tissue, is the empirical absorbance value of substance m 2 in biological tissue, is the empirical absorbance value of biological tissue substance m n ; I photon (X, Y, Z, λ) is the absorbance value converted from the photoacoustic signal, V(X, Y, Z, λ) is the intensity of the detected photoacoustic signal, V Input is the intensity of the monochromatic nanosecond pulsed excitation light, and ∂ is the efficiency of converting the optical signal into an ultrasonic signal. According to the absorption peak wavelengths of various substances in biological tissue ( λ 1 ~ λ n ), substituting into the above formula, for example, the Raman spectral peak differences of substances such as water, oxyhemoglobin, and fat ( λ 1 = 1600nm,λ λ2 = 850 nm, λ λ3 = 1600 nm), the content of various substances in biological tissues can be obtained from the photoacoustic spectroscopy data. For example, the water content in biological tissues C w , oxyhemoglobin C H , fat C f and other substances. For other biological parameters detected in other biological tissues, the error of quantitative analysis of substance content by photoacoustic spectroscopy can also be corrected by this method.

[0045] The 3D convolutional instance segmentation deep network model described above: Based on a 3D convolutional neural network, it extracts the spatial spectral features of multi-field high-dimensional photoacoustic spectroscopy-Raman spectroscopy data, and realizes the detection and visual display of the content and spatial distribution of various physiological substances in biological samples, including blood oxygen, water, fat, inflammatory factors, and oxidative substances; based on the detection results of the content and spatial distribution of physiological substances, it realizes the pixel segmentation of normal and diseased regions in biological tissues, and the individual pixel information of the instance segmentation types can be used to analyze the severity of disease progression.

[0046] The present invention is applied to AD / sepsis ophthalmology, osteoarthritis microenvironment detection, and colon cancer detection: The high-dimensional photoacoustic spectroscopy-Raman spectroscopy data model has a deep detection depth (4-5 cm) and high resolution (the three-dimensional tomographic resolution in XYZ can reach 200 nm), and can perform non-invasive high-resolution photoacoustic spectroscopy combined with Raman 4D imaging detection on biological tissues. It can perform non-invasive ophthalmic detection, construct the blood oxygen nerve distribution of the fundus retina, and the distribution of amyloid αβ enzyme in the surrounding microenvironment, so as to analyze the early recognition progress of AD or sepsis. The osteoarthritis microenvironment detection is characterized in that the high-dimensional photoacoustic spectroscopy-Raman spectroscopy data model can deeply detect the bone medium and the three-dimensional distribution of surrounding blood vessels in cartilage tissue, and combine photoacoustic spectroscopy and Raman spectroscopy to detect the inflammatory progress of bacteria, microorganisms, etc. in synovial fluid of joints.

[0047] The above embodiments can be further combined or replaced, and the embodiments are only descriptions of the preferred embodiments of the present invention, and do not limit the concept and scope of the present invention. Without departing from the design idea of the present invention, various changes and improvements made by those of ordinary skill in the art to the technical solutions of the present invention all belong to the protection scope of the present invention. The protection scope of the present invention is given by the appended claims and any equivalents thereof.

Claims

1. A multi-field high-dimensional imaging method for photoacoustic spectroscopy - confocal Raman spectroscopy, characterized in that The broadband nanosecond pulsed laser emitted by the continuous spectrum nanosecond pulsed laser generates a tunable wavelength nanosecond pulsed laser after being adjusted by the narrowband spectral tuning module; the nanosecond pulsed laser is transmitted through a series of optical elements, passes through the microscopic imaging detection module and the depth imaging detection module, and is focused on the biological sample placed in the water tank; the nanosecond pulsed laser is focused on the biological sample to generate Raman scattering, which is collected by the high-throughput Raman spectroscopy detection module to achieve pulsed Raman spectroscopy detection; the nanosecond pulsed laser causes the thermal expansion of the biological sample and then rapidly contracts to generate a photoacoustic signal. The narrowband spectral tuning module is used to tune the nanosecond pulsed laser to different wavelengths to achieve photoacoustic spectroscopy detection.

2. The method according to claim 1, characterized in that, It includes the following steps: a) High-dimensional photoacoustic spectroscopy-Raman spectroscopy imaging; b) Multi-field photoacoustic spectroscopy imaging; c) According to the spectral decoupling method, the Raman spectroscopy is used to correct the errors in the characterization of the species and content of substances by the photoacoustic spectroscopy, improve the accuracy of the photoacoustic spectroscopy for substance analysis, and achieve multi-field high-dimensional data fusion.

3. The method according to claim 2, wherein For the high-dimensional photoacoustic spectroscopy-Raman spectroscopy imaging described above, through the microscopic imaging detection module, the nanosecond pulsed laser is focused on the position of the biological sample to excite the Raman spectroscopy signal at the millimeter-level depth position of the biological sample, and pulsed Raman spectroscopy detection is achieved. The narrowband spectral tuning module is used to tune the nanosecond pulsed laser to different wavelengths to excite the photoacoustic signal at the millimeter-level depth position of the biological sample, and photoacoustic spectroscopy detection at the shallow position is achieved; the photoacoustic spectroscopy detection is based on the specific absorption differences of substances for different light wavelengths to achieve the detection of the content and spatial distribution of substances including water, blood oxygen, and fat; the Raman spectroscopy detection is based on the molecular vibration energy level differences of substances to generate molecular fingerprint Raman spectral lines to achieve the detection of the content and spatial distribution of mitochondria, neutrophils, DNA genetic materials, and inflammatory factors; the photoacoustic spectroscopy detection and the pulsed Raman spectroscopy detection share the excitation light path, and the excited photoacoustic spectroscopy signal and Raman signal correspond one by one; combined with the two-dimensional scanning of the galvanometer system and the depth tomography of the electric Z-axis displacement stage, high-dimensional photoacoustic spectroscopy-Raman spectroscopy imaging including three-dimensional spatial information, Raman spectroscopy information, and photoacoustic spectroscopy information is achieved; the high-dimensional photoacoustic spectroscopy-Raman spectroscopy imaging combines three-dimensional spatial stereoscopic imaging, the absorption characteristics of the photoacoustic spectroscopy and the molecular fingerprint characteristics of the Raman spectroscopy, and performs high-dimensional model visualization display.

4. The method according to claim 2, wherein For the multi-field photoacoustic spectroscopy imaging described above: the electric XY displacement stage is used to move the sample detection area to the detection window position of the depth imaging module; through the depth imaging detection module, the nanosecond pulsed laser is focused on the deep position at the centimeter level of the sample. The narrowband spectral tuning module is used to tune the nanosecond pulsed laser to different wavelengths to excite the photoacoustic signal at the centimeter-level depth position of the sample, and photoacoustic spectroscopy detection at the deep position is achieved; the positions of different sound source points excited in the excitation light path of the biological tissue are calculated according to the sound speed in the biological tissue; the computer controls the two-dimensional movement of the electric XY displacement stage to achieve three-dimensional reconstruction of the biological tissue at the centimeter level; the deep photoacoustic spectroscopy data detected by the depth imaging detection module is combined with the high-dimensional photoacoustic spectroscopy-Raman spectroscopy data detected by the microscopic imaging detection module to achieve imaging detection that takes into account both the penetration ability and high resolution.

5. The method according to claim 2, wherein The multi-field high-dimensional data fusion described above: The computer system fuses the photoacoustic spectroscopy signal and the Raman spectroscopy signal at the shallow position to form the photoacoustic spectroscopy-Raman spectroscopy data at the shallow position; meanwhile, it fuses the photoacoustic spectroscopy-Raman spectroscopy data at the shallow position with the photoacoustic spectroscopy data at the deep position to form the multi-field high-dimensional photoacoustic spectroscopy-Raman spectroscopy data; according to the spectral decoupling method, the Raman spectroscopy is used to correct the error of the photoacoustic spectroscopy in characterizing the species and content of substances, thereby improving the accuracy of the photoacoustic spectroscopy in substance analysis.

6. The method according to claim 1 or 2, characterized in that, The continuous-spectrum nanosecond pulsed laser outputs tunable-wavelength nanosecond pulsed excitation light through the narrowband spectral tuning module. The nanosecond pulsed excitation light is absorbed by the substances in the biological tissue and generates photoacoustic signals. The computer controls the electric pitching adjustment mirror in the narrowband spectral tuning module to adjust the wavelength of the nanosecond pulsed excitation light, so as to achieve the successive output of nanosecond pulsed excitation light with different wavelengths. The nanosecond pulsed excitation light with different wavelengths is incident on the biological sample in sequence to excite photoacoustic signals, thereby realizing photoacoustic spectroscopy detection.

7. The method according to claim 2, characterized in that The spectral decoupling method described above uses the Raman spectroscopy to correct the error of the photoacoustic spectroscopy in characterizing the species and content of substances; different biological tissues show absorption characteristics at the same wavelength, and their absorption degrees and ways of light are different, resulting in interference in the measurement of the photoacoustic spectroscopy. By introducing the Raman spectroscopy signal for calibration and using the unique response of the Raman spectroscopy to the vibration of different substance molecules, supplementary information is provided for the measurement result of the photoacoustic spectroscopy; according to the Raman spectroscopy peak differences of water, oxyhemoglobin, and fat, the species of the biological tissue content are identified, or the error of the photoacoustic spectroscopy in quantitatively analyzing the substance content is corrected. ； Wherein, is the content of the substance in the biological tissue m 1 , is the content of the substance in the biological tissue m 2 , is the content of the substance in the biological tissue m n , is the empirical absorbance value of the substance in the biological tissue m 1 , is the empirical absorbance value of the substance in the biological tissue m 2 , is the empirical absorbance value of the biological tissue substance m n , I photon (X, Y, Z, λ) is the absorbance value converted from the photoacoustic signal, V(X, Y, Z, λ) is the intensity of the detected photoacoustic signal, V Input is the intensity of the monochromatic nanosecond pulsed excitation light, ∂ is the efficiency of converting the optical signal into the ultrasonic signal; according to the absorption peak wavelengths of various substances in the biological tissue λ 1 ~ λ n Substituting into the above formula, the contents of various substances in the biological tissue can be obtained from the photoacoustic spectrum data, or the error of quantitatively analyzing the substance content by the photoacoustic spectrum can be corrected.

8. The method according to claim 1, characterized in that: Using the 3D convolutional instance segmentation deep network model to extract the features of the multi-field high-dimensional photoacoustic spectroscopy-Raman spectroscopy data, realizing the detection and visual display of the content and spatial distribution of various physiological substances in the biological sample, including blood oxygen, water, fat, inflammatory factors, and oxidative substances, and providing intuitive visual data for clinical physiological analysis; based on the detection results of the content and spatial distribution of physiological substances, realizing the pixel segmentation of the normal area and the diseased area of the biological tissue; using the deep neural network to establish the correlation between the multi-field high-dimensional photoacoustic spectroscopy-Raman spectroscopy data and different development stages of knee osteoarthritis, sepsis, and colon cancer diseases, and realizing the early identification and diagnosis of knee osteoarthritis, sepsis, and colon cancer diseases.

9. A photoacoustic spectroscopy - confocal Raman spectroscopy multi - field high - dimensional imaging device using the method according to claim 1, characterized in that, It includes a continuous-spectrum nanosecond pulsed laser (1), a narrowband spectral tuning module (2), a second lens (8), a third lens (9), a second mirror (10), a beam splitter (11), a precision galvanometer system (12), a scanning lens (13), a telecentric lens (14), a microscopic imaging detection module (36), a depth imaging detection module (37), a water tank (28), an electric XY displacement stage (30), a high-throughput Raman detection module (17), a high-speed acquisition card (26), and a computer (27); The supercontinuum nanosecond pulsed laser (1) is used to emit a wide-spectrum nanosecond pulsed laser in the wavelength range of 400 nm - 2000 nm. After being adjusted by the narrowband spectral tuning module (2), a nanosecond pulsed laser with adjustable wavelength is generated. The nanosecond pulsed laser passes through the first reflector (7), the second lens (8), the third lens (9), the second reflector (10), and then enters the galvanometer system (12) via the beam splitter (11). The galvanometer system (12) adjusts the output direction of the nanosecond pulsed laser. The nanosecond pulsed laser is incident on the microscopic imaging detection module (36) and the depth imaging detection module (37) through the scanning lens (13) and the telecentric lens (14) respectively. The combination of the scanning lens (13) and the telecentric lens (14) is used to optimize the scanning image plane and expand the beam of the nanosecond pulsed laser. The narrowband spectral tuning module (2) includes a first parabolic mirror (3), a second reflection grating (34), a second parabolic mirror (4), an electric pitch-adjustable mirror (35), a first pinhole (5), and a first lens (6). After the first parabolic mirror (3) collimates the wide-spectrum nanosecond pulsed laser into parallel light, it is incident on the second reflection grating (34). The light with different wavelengths in the wide-spectrum nanosecond pulsed laser generates different diffraction angles after passing through the second reflection grating (34), producing diffracted light with different wavelengths. The diffracted light is focused on the first pinhole (5) after being adjusted by the second parabolic mirror (4) and the electric pitch-adjustable mirror (35). The first lens (6) is used to collimate the diffracted light passing through the first pinhole (5) for subsequent use. The computer (27) controls the pitch angle of the electric pitch-adjustable mirror (35) to adjust the diffracted light passing through the first pinhole (5), so as to achieve the successive output of the diffracted light with a specific wavelength and output a nanosecond pulsed laser with adjustable wavelength. The microscopic imaging detection module (36) includes a first acoustic medium prism (24), an adjustable focus lens module (15), a microscopic objective lens (16), an electric Z-axis displacement stage (29), a first ultrasonic transducer (25), and a first low-noise signal amplifier (39). The nanosecond pulsed laser is focused on the shallow position of the biological sample in the water tank (28) through the first acoustic medium prism (24), the adjustable focus lens module (15), and the microscopic objective lens (16). The photoacoustic signal generated is propagated through the first acoustic medium prism (24) and collected by the first ultrasonic transducer (25). At the same time, the Raman spectral signal generated is returned along the original optical path and collected by the high-throughput Raman spectral detection module (17) after passing through the beam splitter (11). The described depth imaging detection module (37) includes a second acoustic medium prism (31), a second ultrasonic transducer (32), a second low-noise signal amplifier (40), a fourth lens (41), and a fifth lens (42); after the nanosecond pulsed laser passes through the first acoustic medium prism (24), part of the light beam is incident on the second acoustic medium prism (31), and is focused on the deep position of the biological sample in the water tank (28) through the fourth lens (41) and the fifth lens (42); the generated photoacoustic signal propagates through the second acoustic medium prism (31) and is collected by the second ultrasonic transducer (32). The described high-throughput Raman spectroscopy detection module (17) includes a trap filter (38), a focusing lens (18), a coded aperture mask (33), a double pinhole (19), a first free-form surface mirror (20), a first reflection grating (21), a second free-form surface mirror (22), and a high-speed camera (23); after the signal light returned by the microscopic imaging detection module (36) passes through the trap filter (38), the trap filter (38) filters out the excitation light; the coded aperture mask (33) is used to overcome the problem of mutual limitation between Raman spectroscopy resolution and optical flux, and achieve a hyperspectral resolution of sub-nanometer level for spectral resolution under a wide slit and high optical flux; the slit (14) is used to block the stray light outside the focal plane of the microscopic objective lens (16), improve the signal contrast, and achieve confocal detection; the light emerging from the slit is collimated by the first free-form surface mirror (20), diffracted by the reflection grating (21), and finally focused by the second free-form surface mirror (22) to focus the Raman signal on the high-speed camera (23) to generate a clear Raman spectrum image; the design combination of the reflection surface parameters of the first free-form surface mirror (20) and the second free-form surface mirror (22) is used to eliminate the phase difference and chromatic aberration in Raman spectroscopy imaging.

10. The device according to claim 9, characterized in that The described computer (27) is connected to the high-speed acquisition card (26), the high-speed camera (23), the galvanometer system (12), the motorized Z-axis displacement stage (29), the motorized XY displacement stage (30), and the motorized pitch-adjustable mirror (35) through signal cables respectively; the computer (27) controls the excitation, acquisition, and data processing of the photoacoustic signals and Raman signals at different depth positions in the microscopic imaging detection module (36) and the depth imaging detection module (37) by controlling the galvanometer system (12), the motorized Z-axis displacement stage (29), the motorized XY displacement stage (30), and the high-speed acquisition card (26), and generates high-resolution multi-field high-dimensional photoacoustic spectroscopy-Raman spectroscopy data, so as to achieve high-dimensional imaging of different fields of biological samples. The first ultrasonic transducer (25) is connected to the first low-noise signal amplifier (39), and the second ultrasonic transducer (32) is connected to the second low-noise signal amplifier (40) to amplify the photoacoustic signal; the high-speed acquisition card (26) is connected to the first low-noise signal amplifier (39) and the second low-noise signal amplifier (40) to collect the photoacoustic spectroscopy signal; the high-speed acquisition card (26) transmits the photoacoustic spectroscopy signal to the computer (27) for backend data processing.

Images