CN113109340B

CN113109340B - Imaging system and imaging method for signal processing of planar interferometer

Info

Publication number: CN113109340B
Application number: CN202110378092.7A
Authority: CN
Inventors: 袁进; 肖鹏; 胡晓东; 马可
Original assignee: Zhongshan Ophthalmic Center
Current assignee: Zhongshan Ophthalmic Center
Priority date: 2021-04-08
Filing date: 2021-04-08
Publication date: 2022-05-24
Anticipated expiration: 2041-04-08
Also published as: CN113109340A; CN114894794A; CN114894793A; CN114894794B; CN114894793B

Abstract

The present disclosure describes an imaging system for signal processing of a planar interferometer. The imaging system comprises an acquisition module for acquiring a plurality of interferograms acquired by a planar interferometer to form a first interference image set; the artifact eliminating module is used for performing singular value decomposition on the first interference image set to obtain a time eigenvector, and updating a singular value matrix based on the fluctuation degree of the time eigenvector to obtain a second interference image set through reconstruction; the signal-to-noise ratio enhancement module is used for carrying out signal-to-noise ratio enhancement processing on the second interference image set by utilizing the sliding window so as to obtain a third interference image set; and a dynamic imaging module for acquiring components of three-dimensional color space of each position based on time domain information and frequency domain information of a pixel sequence formed by a plurality of pixel values of the same position of the plurality of interference images to convert the third interference image into a color image reflecting internal dynamic signals of the fresh and live sample. Thereby, the internal dynamic signal of the biological tissue can be obtained in a non-invasive manner.

Description

Imaging system and imaging method for signal processing of planar interferometer

Technical Field

The present disclosure relates generally to an imaging system and imaging method for signal processing of planar interferometers.

Background

The internal dynamic signal of biological tissue can provide important reference for scientific research and clinical diagnosis and treatment in the biomedical field, for example, in the fields of histopathology or stem cell tissue engineering, the strength of the internal dynamic signal of biological tissue is often used for cell type differentiation or cell culture detection. Therefore, how to obtain the internal dynamic signals of biological tissues has become a popular research direction in the biomedical field. The internal dynamic signals of biological tissues can currently be acquired by invasive methods, such as methods based on fluorescent staining.

However, the invasive method is generally destructive and is not suitable for observing and measuring living biological tissues, while the structural imaging of biological tissues obtained by the non-invasive method (for example, measuring biological tissues by using a planar interferometer based on the principle of double-beam equal thickness interference) has the problem that dynamic signals cannot be reflected, and the measurement of internal dynamic signals of biological tissues cannot be realized.

Disclosure of Invention

The present disclosure has been made in view of the above circumstances, and an object thereof is to provide an imaging system and an imaging method for signal processing of a planar interferometer capable of obtaining an internal dynamic signal of a biological tissue in a non-invasive manner.

To this end, a first aspect of the present disclosure provides an imaging system for signal processing of a planar interferometer that performs measurement based on the principle of two-beam equal thickness interferometry, the imaging system comprising: the device comprises an acquisition module, an artifact eliminating module, a signal-to-noise ratio enhancing module and a dynamic imaging module; the acquisition module is used for acquiring a plurality of interferograms obtained by continuously acquiring a fresh sample by using the planar interferometer, and the interferograms form a first interferogram set according to acquisition time; the artifact eliminating module is used for performing singular value decomposition on the first interference image set to obtain a time characteristic matrix and a singular value matrix, wherein each column of the time characteristic matrix is a time characteristic vector, calculating the fluctuation degree of the time characteristic vector, screening the time characteristic vector meeting a threshold condition based on the fluctuation degree to update the singular value in the singular value matrix corresponding to the time characteristic vector to a preset singular value, and then reconstructing the first interference image set based on the updated singular value matrix and using the first interference image set as a second interference image set; the signal-to-noise ratio enhancement module utilizes a sliding window with a preset length to carry out signal-to-noise ratio enhancement processing on pixel values in each sliding window in the second interference image set so as to obtain a third interference image set, wherein the sliding window moves along the direction of the acquisition time according to a preset step length; and the dynamic imaging module is used for acquiring a plurality of pixel values of the same position of a plurality of interference images in the third interference image set to form a pixel sequence ordered according to the acquisition time, acquiring a component of a three-dimensional color space of each position based on time domain information and frequency domain information of the pixel sequence of each position, and converting the third interference image set into a color image based on the component of the three-dimensional color space, wherein the color image reflects an internal dynamic signal of the fresh and alive sample, the component of the three-dimensional color space comprises a hue component, a saturation component and a brightness component, and a target frequency is acquired based on the frequency domain information, and the hue component corresponds to the target frequency. In the disclosure, an imaging system obtains a time eigenvector based on singular value decomposition to eliminate artifacts of multiple interferograms through fluctuation degrees of the time eigenvector and performs signal-to-noise enhancement processing on the multiple interferograms by using a sliding window, and finally generates a color image based on time domain information and frequency domain information of the multiple interferograms. Thereby, the influence of artifacts or noise on the image quality of the color image can be reduced, which is beneficial for further research based on the color image, and the color image generated by the present disclosure can reflect the internal dynamic signal of the biological tissue and obtain the color image in a non-invasive manner.

In addition, in the imaging system according to the first aspect of the present disclosure, optionally, the imaging system further includes a calibration module, where the calibration module is configured to establish a linear correspondence between the target frequency and a color to calibrate the correspondence between the target frequency and the color. This makes it possible to intuitively obtain the frequency corresponding to the color in the color image.

In addition, in the imaging system according to the first aspect of the present disclosure, optionally, the artifact removing module represents a fluctuation degree of the temporal feature vector by using an accumulated zero-crossing rate of the temporal feature vector, and the accumulated zero-crossing rate of the ith column of temporal feature vectors is represented as: d _ ZRC_i＝|ZRC_i+1-ZRC_iWhere i is the column index of the time feature matrix, ZRC_iIs the zero-crossing rate of the ith column of temporal feature vectors. Thereby, the degree of fluctuation of the temporal feature vector can be expressed by the cumulative zero-crossing.

In addition, in the imaging system according to the first aspect of the present disclosure, optionally, the threshold condition is that an accumulated zero-crossing rate of each temporal feature vector is greater than a preset value, where the preset value is 3 times a standard deviation of the accumulated zero-crossing rate. Thus, the artifact can be eliminated by eliminating the signal in which the degree of dispersion of the cumulative zero-crossing rate reaches the preset value degree.

Further, in the imaging system according to the first aspect of the present disclosure, optionally, the preset length is greater than 1, and a dimension of the sliding window coincides with a dimension of the second interference image set.

Further, in the imaging system according to the first aspect of the present disclosure, optionally, in the signal-to-noise ratio enhancement process, an average value of pixel values in each sliding window is obtained, a cumulative sum is calculated for differences between the pixel values in each sliding window and the average value to obtain a cumulative value, an absolute value of the cumulative value is divided by the preset length to obtain an average cumulative value, and the average cumulative value is used as the pixel value of the third interference image set. Thus, the signal-to-noise ratio enhancement processing can be performed on the second interference image set based on the sliding window and cumulative sum algorithm.

Further, in the imaging system relating to the first aspect of the present disclosure, optionally, the dynamic imaging module acquires a power spectral density based on a pixel sequence of each position and takes an inverse of a bandwidth of the power spectral density as the saturation component; the dynamic imaging module performs Fourier transform on the pixel sequence of each position to obtain a frequency sequence, and obtains the target frequency as the hue component based on the frequency sequence and the power spectral density; the dynamic imaging module acquires a standard deviation or a variance of a pixel sequence based on the pixel sequence of each position and takes the standard deviation or the variance as the luminance component. Thereby, the saturation component, the hue component, and the luminance component can be obtained based on the pixel sequence of each position.

A second aspect of the present disclosure provides an imaging method for signal processing of a planar interferometer that performs measurement based on a principle of double-beam equal thickness interference, the imaging method including: acquiring a plurality of interferograms obtained by continuously acquiring a fresh sample by using the planar interferometer, wherein the interferograms form a first interferogram set according to acquisition time; performing singular value decomposition on the first interference image set to obtain a time characteristic matrix and a singular value matrix, wherein each column of the time characteristic matrix is a time characteristic vector, calculating the fluctuation degree of the time characteristic vector, screening the time characteristic vector meeting a threshold condition based on the fluctuation degree to update the singular value in the singular value matrix corresponding to the time characteristic vector to a preset singular value, and reconstructing the first interference image set based on the updated singular value matrix and using the first interference image set as a second interference image set; performing signal-to-noise ratio enhancement processing on pixel values in each sliding window in the second interference image set by using a sliding window with a preset length to obtain a third interference image set, wherein the sliding window moves along the direction of the acquisition time according to a preset step length; and acquiring a plurality of pixel values of the same position of a plurality of interference images in the third interference image set to form a pixel sequence ordered according to the acquisition time, acquiring a component of a three-dimensional color space of each position based on time domain information and frequency domain information of the pixel sequence of each position and converting the third interference image set into a color image based on the component of the three-dimensional color space, wherein the color image reflects an internal dynamic signal of the fresh sample, the component of the three-dimensional color space comprises a hue component, a saturation component and a brightness component, and a target frequency is acquired based on the frequency domain information, and the hue component corresponds to the target frequency. In the disclosure, the imaging method obtains a time eigenvector based on singular value decomposition to eliminate artifacts of multiple interferograms through fluctuation degrees of the time eigenvector and performs signal-to-noise enhancement processing on the multiple interferograms by using a sliding window, and finally generates a color image based on time domain information and frequency domain information of the multiple interferograms. Thereby, the influence of artifacts or noise on the image quality of the color image can be reduced, further research is facilitated based on the color image, and in addition, the color image generated by the present disclosure can reflect the internal dynamic signal of the biological tissue and the manner of obtaining the color image is non-invasive.

Further, in the imaging method according to the second aspect of the present disclosure, optionally, a linear correspondence relationship between the target frequency and a color is established to calibrate the correspondence relationship between the target frequency and the color. This makes it possible to intuitively obtain the frequency corresponding to the color in the color image.

In addition, in the imaging method according to the second aspect of the present disclosure, optionally, the degree of fluctuation of the temporal feature vector is represented by an accumulated zero-crossing rate of the temporal feature vector, and the accumulated zero-crossing rate of the ith column of temporal feature vectors is represented by: d _ ZRC_i＝|ZRC_i+1-ZRC_iWhere i is the column index of the time feature matrix, ZRC_iThe zero crossing rate of the ith column of time characteristic vectors; the threshold condition is that the cumulative zero-crossing rate of each time feature vector is greater than a preset value, wherein the preset value is 3 times of the standard deviation of the cumulative zero-crossing rate. Thus, the artifact can be eliminated by eliminating the signal in which the degree of dispersion of the cumulative zero-crossing rate reaches the preset value degree.

According to the present disclosure, it is possible to provide an imaging system and an imaging method for signal processing of a planar interferometer that obtain an internal dynamic signal of a biological tissue by a non-invasive manner.

Drawings

The disclosure will now be explained in further detail by way of example only with reference to the accompanying drawings, in which:

fig. 1 is an application scenario diagram illustrating an imaging method for signal processing of a planar interferometer according to an example of the present disclosure.

Fig. 2 is a block diagram of an imaging system showing signal processing for a planar interferometer to which examples of the present disclosure relate.

Fig. 3 is a schematic diagram illustrating an interferogram to which examples of the present disclosure relate.

Fig. 4(a) is a schematic diagram showing a color image generated without removing artifacts, to which the disclosed example pertains.

Fig. 4(b) is a schematic diagram showing a color image generated in a case where an artifact is removed, to which an example of the present disclosure relates.

Fig. 4(c) is a schematic diagram illustrating eliminated artifacts to which examples of the present disclosure relate.

Fig. 5 is a schematic diagram illustrating a sliding window movement according to an example of the present disclosure.

Fig. 6 is a schematic diagram illustrating a color image with calibration information according to an example of the present disclosure.

Fig. 7 is a flowchart illustrating an imaging method for signal processing of a planar interferometer according to an example of the present disclosure.

Detailed Description

Hereinafter, preferred embodiments of the present disclosure will be described in detail with reference to the accompanying drawings. In the following description, the same components are denoted by the same reference numerals, and redundant description thereof is omitted. The drawings are schematic and the ratio of the dimensions of the components and the shapes of the components may be different from the actual ones. It is noted that the terms "comprises," "comprising," and "having," and any variations thereof, in this disclosure, for example, a process, method, system, article, or apparatus that comprises or has a list of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include or have other steps or elements not expressly listed or inherent to such process, method, article, or apparatus. All methods described in this disclosure can be performed in any suitable order unless otherwise indicated herein or otherwise clearly contradicted by context.

The imaging system and the imaging method for signal processing of the planar interferometer can obtain a color image reflecting internal dynamic signals of biological tissues in a non-invasive mode based on a plurality of interferograms acquired by the planar interferometer. The imaging method to which the present disclosure relates is applied to an imaging system (described later). The present disclosure is described in detail below with reference to the attached drawings. In addition, the application scenarios described in the examples of the present disclosure are for more clearly illustrating the technical solutions of the present disclosure, and do not constitute a limitation on the technical solutions provided by the present disclosure.

Fig. 1 is an application scenario illustration showing an imaging method for signal processing of a planar interferometer according to an example of the present disclosure. As shown in fig. 1, a measurement of biological tissue 120 may be performed using a planar interferometer 110 to obtain multiple interferograms 130. After the acquisition of the plurality of interferograms 130 is completed, the plurality of interferograms 130 may be submitted to the server 140, and the server 140 may execute computer program instructions stored on the server 140 to implement an imaging method by which the plurality of interferograms 130 are received and the color image 150 is generated.

In some examples, the planar interferometer 110 may perform measurements based on the principle of two-beam equal thickness interferometry. In some examples, the planar interferometer 110 may be a two-dimensional planar interferometer, which may be the acquisition of the interferogram 130 by a 2D camera. In some examples, the device that obtains the plurality of interferograms 130 may be any device that performs measurements based on the principle of two-beam equal thickness interferometry. In some examples, the biological tissue 120 may be a cellular framework between cells and organs.

In some examples, server 140 may include one or more processors and one or more memories. Wherein the processor may include a central processing unit, a graphics processing unit, and any other electronic components capable of processing data, capable of executing computer program instructions. The memory may be used to store computer program instructions. In some examples, the server 140 may implement the imaging method by executing computer program instructions on a memory. In some examples, server 140 may also be a cloud server.

Hereinafter, the imaging system 1 according to the present disclosure is described in detail with reference to the drawings. The present disclosure relates to an imaging system 1 for implementing the imaging method described above. Fig. 2 is a block diagram of an imaging system 1 showing signal processing for a planar interferometer according to an example of the present disclosure. Fig. 3 is a schematic diagram illustrating an interferogram to which examples of the present disclosure relate.

As shown in fig. 2, in some examples, the imaging system 1 may include an acquisition module 10, and the acquisition module 10 may be configured to acquire a plurality of interferograms and to compose a first set of interferograms. The interferogram may be a pattern formed by interference of waves. In some examples, the plurality of interferograms may be uniform in size. For example, the size of the multiple interferograms may be 1440 x 1440 or 1024 x 1024. Examples of the disclosure are not limited thereto, and in other examples, the size of the interferogram may be determined by a camera in the planar interferometer used to acquire the interferogram. In some examples, the number of the plurality of interferograms may be greater than or equal to a preset number. For example, the predetermined number may be 16, and the number of the plurality of interferograms may be 16, 32, 48, 64, 100, 512, 1024, or the like. As an example of an interferogram, fig. 3 shows a schematic view of an interferogram obtained by planar interferometer measurement.

Additionally, in some examples, a sample may be acquired with a planar interferometer to obtain multiple interferograms. In some examples, the sample may be a live sample. For example, a fresh sample of mouse retina. In some examples, the fresh sample may include contents. In some examples, the contents may be cells. In some examples, multiple interferograms may be obtained by successive acquisitions. In some examples, the continuous acquisition may be based on a selected exposure frequency (i.e., number of exposures per second). For example, an exposure frequency of 100/sec can be selected for successive acquisitions to obtain 512 interferograms, in which case the acquisition of 512 interferograms takes 5.12 seconds.

Additionally, in some examples, the plurality of interferograms may be grouped into a first set of interferograms by acquisition time. Therefore, the subsequent simultaneous processing of a plurality of interference patterns can be facilitated.

Fig. 4(a) is a schematic diagram showing a color image generated without removing artifacts, to which the disclosed example pertains. Fig. 4(b) is a schematic diagram showing a color image generated in a case where an artifact is removed, to which an example of the present disclosure relates. Fig. 4(c) is a schematic diagram illustrating eliminated artifacts to which examples of the present disclosure relate.

As shown in fig. 2, in some examples, the imaging system 1 may include an artifact removal module 11, and the artifact removal module 11 may be configured to perform artifact removal processing on the first interference image set to obtain the second interference image set. In general, Artifacts (Artifacts) refer to images of various forms that appear in the final image, but in which the scanned object, e.g., a fresh sample, is not present. In general, there are many reasons for the artifacts. For example, a live sample may have motion in the active state that may cause artifacts, such motion being very small at a spatial level, e.g., on the order of microns. Also, for example, a disturbance of the ambient air may cause vibrations, which in turn produce artifacts. In this case, eliminating the artifacts in the first interference image set can improve the image quality of the subsequently generated color image.

As an example of the contrast of color images generated without and with artifacts removed. Fig. 4(a) and 4(b) show a color image generated without removing the artifact and a color image generated with removing the artifact, respectively, of a fresh sample of the mouse retina, and fig. 4(c) shows the artifact removed in the fresh sample of the mouse retina. As can be seen from fig. 4(a), 4(B) and 4(C), the area C from which the artifact shown in the area B is removed is significantly enhanced in contrast to the area a from which the artifact is not removed. This enables identification of small cells blocked by artifacts. Here, the area a, the area B, and the area C correspond to the same position in the generated color image.

In some examples, the deghost module 11 may deghost the first set of interference images based on singular value decomposition to obtain a second set of interference images. In some examples, in the artifact removal process, a Singular Value Decomposition (SVD) may be performed on the first interference image set to obtain a temporal feature matrix and a singular value matrix.

Specifically, the first interference image set may be converted into two-dimensional data S, which may be m × n, where m may be the resolution (size) of one interference image and n may be the number of multiple interference images. In some examples, each row in the two-dimensional data S may include all of the pixel values of the respective interferogram. For example, 512 interferograms are acquired in succession, each interferogram having a size of 1440 × 1440, then m may be 1440 × 1440 and n may be 512.

In some examples, the two-dimensional data S is subjected to singular value decomposition to represent the two-dimensional data S as:

S＝U∑V，

wherein U is a spatial characteristic matrix, sigma is a singular value matrix, and V is a time characteristic matrix. In some examples, the size of the temporal feature matrix may be n × n. For example, assuming 512 interferograms are acquired continuously, the size of the temporal feature matrix may be 512 × 512. The singular value matrix may be a diagonal matrix, with the element on the diagonal being a singular value. In some examples, the size of the matrix of singular values may take n × n. For example, assuming 512 interferograms are acquired continuously, the size of the matrix of singular values may be 512 × 512.

In some examples, after the temporal feature matrix and the singular value matrix are obtained, each column of the temporal feature matrix may be taken as a temporal feature vector. Because the time characteristic vector corresponding to the artifact is greatly different from the time characteristic vector corresponding to the normal interference signal, the artifact can be distinguished based on the fluctuation degree of the time characteristic vector. In some examples, the singular values in the singular value matrix may be updated based on a degree of fluctuation of the temporal eigenvector. Specifically, the fluctuation degree of the time eigenvector may be calculated, the time eigenvector meeting the threshold condition is screened based on the fluctuation degree, and then the singular value in the singular value matrix corresponding to the time eigenvector is updated to the preset singular value. Thereby, the singular values in the singular value matrix can be updated based on the fluctuation degree of the temporal eigenvector. In some examples, the preset singular value may be 0.

In some examples, the degree of fluctuation may be represented by an accumulated zero-crossing rate of the temporal feature vector. That is, the artifact removal module 11 may represent the fluctuation degree of the temporal feature vector by using the cumulative zero-crossing rate of the temporal feature vector. As described above, each column of the temporal feature matrix may be taken as a temporal feature vector. In some examples, the cumulative zero-crossing rate of the ith column of temporal feature vectors may be expressed as:

D_ZRC_i＝|ZRC_i+1-ZRC_i|，

where i is the column index of the time feature matrix, ZRC_iIs the zero-crossing rate of the ith column of temporal feature vectors. The zero crossing rate may refer to a rate of change of a symbol of one signal, and may refer to a number of times a fluctuation graph (which may also be referred to as a change profile) formed by a plurality of values in a temporal feature vector passes through zero in the present disclosure. Thus, a wave representing a temporal feature vector by accumulated zero crossings can be usedDegree of motion. In some examples, a curve for each temporal feature vector may be plotted and a zero-crossing rate may be determined based on the curve. In some examples, i can take on values of 1 to n-1, and n can be the number of the plurality of interferograms.

Examples of the disclosure are not limited thereto and in other examples, the degree of fluctuation may be represented in other ways.

In some examples, the threshold condition may be that the cumulative zero-crossing rate of the respective temporal feature vectors is greater than a preset value. In some examples, the preset value may be 3 times the standard deviation of the cumulative zero-crossing rate. That is, the threshold condition may be expressed as:

D_ZRC_i>3×std(D_ZRC)，

wherein D is_{_}ZRC_iIs the cumulative zero-crossing rate of the ith column of temporal feature vectors, D_{_}ZRC is a set of cumulative zero-crossing rates. Thus, the artifact can be eliminated by eliminating the signal in which the degree of dispersion of the cumulative zero-crossing rate reaches the preset value degree. However, examples of the present disclosure are not limited thereto, and in other examples, the preset value may be set according to experimental experience or experimental effect.

As described above, the singular values in the singular value matrix may be updated based on the degree of fluctuation of the temporal eigenvector. In some examples, the first interference image set may be reconstructed based on the updated matrix of singular values and used as the second interference image set. Thus, artifact removal processing can be performed on the first interference image set based on the fluctuation degree of the temporal feature vector. Specifically, the spatial feature matrix, the updated singular value matrix, and the temporal feature matrix described above may be multiplied to obtain updated two-dimensional data, and then the second interference image set corresponding to the first interference image set may be obtained based on the updated two-dimensional data.

However, examples of the present disclosure are not limited thereto, and in other examples, the imaging system 1 may not include the artifact removal module 11. In this case, subsequent modules, such as the signal-to-noise enhancement module 12 or the dynamic imaging module 13, may process the first interference image set.

Fig. 5 is a schematic diagram illustrating a sliding window movement according to an example of the present disclosure. As shown in fig. 2, in some examples, the imaging system 1 may include a signal-to-noise enhancement module 12, and the signal-to-noise enhancement module 12 may be configured to perform signal-to-noise enhancement processing on the second interference image set to obtain a third interference image set. In some examples, the snr enhancement module 12 can perform snr enhancement processing on pixel values in each sliding window in the second interference image set by using a sliding window with a preset length to obtain a third interference image set. In some examples, the preset length may be greater than 1. In some examples, the preset length may be 8.

In some examples, the second set of interference images may be represented as a three-dimensional matrix, where the first and second dimensions may correspond to the size of the interference images and the third dimension corresponds to the acquisition order (which may also be referred to as acquisition time). In some examples, the dimensions of the sliding window and the dimensions of the second set of interference images may coincide. In some examples, the first and second dimensions of the sliding window may coincide with the dimensions of the interferogram. In this case, the sliding window can be moved along the acquisition sequence based on the size of the interferograms.

In some examples, the sliding window may move in the direction of the acquisition time. As an example of the sliding window movement, fig. 5 is a schematic diagram showing that the sliding window H with the preset length of 2 moves along the second interference image set L, wherein the second interference image set L contains a plurality of interference images, for example, the number of the plurality of interference images may be n, and the plurality of interference images may include p1, p2, p3, … …, pn, and the like.

In some examples, the sliding window may be moved in the direction of the acquisition time by a preset step size. In some examples, the preset step size may be a preset proportion of the preset length. In some examples, the preset ratio may be less than 1. For example, the predetermined ratio may be 25%, 50%, 75%, 1/3, 2/3, or the like. Specifically, assuming that the preset ratio is 50%, the number of interferograms in the second set of interference images is 512, and the preset length is 8, the preset step size is 4, and the number of sliding windows is 512/4, that is, 128. In this case, there is an overlap between the sliding windows, and the latter sliding window can be subjected to the snr enhancement processing in combination with the partial interferogram in the former sliding window. However, the examples of the present disclosure are not limited thereto, and in other examples, the preset step size and the preset length may be set according to actual situations, for example, the preset length may be 6, and the preset step size may be 2.

In some examples, in the signal-to-noise ratio enhancement process, an average value of pixel values in each sliding window may be obtained. In some examples, the difference between the pixel values in the respective sliding windows and the average value may be summed to obtain a cumulative value. In some examples, the absolute value of the accumulated value may be divided by a preset length to obtain an average accumulated value, and the average accumulated value may be used as the pixel value of the third interference image set. Thus, the signal-to-noise ratio enhancement processing can be performed on the second interference image set based on the sliding window and cumulative sum algorithm.

In particular, the third interference image set S₃Can be expressed as:

where λ is a predetermined length (also referred to as a sliding window length), W_kFor the pixel values within the kth sliding window,

is the average of the pixel values in the kth sliding window, CumSum is the cumulative sum function, and K is the number of sliding windows.

Examples of the disclosure are not limited thereto and in other examples, the imaging system 1 may not include the signal-to-noise enhancement module 12. In this case, subsequent modules, such as the dynamic imaging module 13, may process the first interference image set or the second interference image set.

As shown in fig. 2, in some examples, imaging system 1 may include a dynamic imaging module 13, and dynamic imaging module 13 may generate a color image based on the third interference image set. The color image may reflect the internal dynamic signal of the fresh sample. In some examples, components of a three-dimensional color space may be obtained based on the third set of interference images and the third set of interference images may be turned into color images based on the components of the three-dimensional color space. In some examples, the components of the three-dimensional color space may include a hue component, a saturation component, and a brightness component (which may also be referred to as a luma component). In some examples, the three-dimensional color space may be an HSV (Hue: Saturation: Value: lightness) space or an HSI (Hue: Saturation: Intensity: lightness) space. Thereby, a color image can be generated by the HSV space or the HSI space. In some examples, the color image may be an image of an RGB (Red: Red, Green: Green, Blue: Blue) space.

In some examples, a plurality of pixel values of the same location of the plurality of interferograms in the third set of interference images may be acquired to form a sequence of pixels ordered by acquisition time, a component of a three-dimensional color space of each location is acquired based on time domain information and frequency domain information of the sequence of pixels of each location, and the third set of interference images is converted into a color image based on the component of the three-dimensional color space.

In general, frequency domain information may include frequency components of a signal and magnitude of amplitude of each frequency component, time domain information may include magnitude information of the signal, and how fast a waveform of the time domain signal changes with time may be measured. As described above, the pixel sequence is formed by a plurality of pixel values of respective positions in order of acquisition time, and thus, in the present disclosure, the time domain information may refer to a pixel value in the pixel sequence that varies with acquisition time. In some examples, the signal may be transformed from the time domain to the frequency domain by a fourier series or fourier transform to obtain frequency domain information.

As described above, the components of the three-dimensional color space may include a hue component, a saturation component, and a brightness component. In some examples, dynamic imaging module 13 may obtain the tonal components based on frequency domain information. In some examples, dynamic imaging module 13 may obtain the target frequency based on the frequency domain information and obtain the tonal component based on the target frequency.

In some examples, the tonal components may correspond to a target frequency. In some examples, the target frequency may be considered a tonal component. In some examples, the target frequency may be an average frequency or a median frequency. In some examples, the dynamic imaging module 13 may perform fourier transform on the pixel sequence of each position to obtain a frequency sequence (i.e., frequency domain information), and obtain an average frequency based on the frequency sequence and a power spectral density (described later) and as a hue component. In some examples, the frequency series and the power spectral density may be dot-product to obtain an average frequency. Thereby, the tone component can be obtained based on the pixel sequence of each position.

As described above, the target frequency may be the median frequency. In some examples, the median frequency may be the frequency that divides the power spectral density into upper and lower equal power areas. In some examples, the median frequency MDF may satisfy the formula:

where p (f) is the power spectral density and f is the frequency.

In some examples, dynamic imaging module 13 may obtain the saturation component based on the pixel sequence (i.e., temporal information) of each location. In some examples, dynamic imaging module 13 may obtain a power spectral density based on the sequence of pixels for each location and a saturation component based on the power spectral density. In some examples, the power spectral density may be normalized prior to acquiring the saturation component based on the normalized power spectral density. In some examples, the dynamic imaging module 13 may acquire the power spectral density based on the pixel sequence of each position and take the inverse of the bandwidth of the power spectral density as the saturation component. In some examples, the power spectral density may be obtained based on a sequence of pixels at various locations and using a Welch method. Thereby, the saturation component can be obtained based on the pixel sequence of each position.

In some examples, dynamic imaging module 13 may obtain a standard deviation or a variance of a pixel sequence (i.e., temporal information) based on the pixel sequence at each position and treat the standard deviation or the variance as a luminance component. Thereby, the luminance component can be obtained based on the pixel sequence of each position.

In some examples, dynamic imaging module 13 may normalize the plurality of interferograms in the third set of interferograms and acquire a sequence of pixels for each location based on the normalized interferograms prior to acquiring the sequence of pixels for each location. This makes it possible to normalize a plurality of interferograms. In some examples, the normalization method may include, but is not limited to, L1 normalization or L2 normalization.

Fig. 6 is a schematic diagram illustrating a color image with calibration information D according to an example of the present disclosure. In some examples, the imaging system 1 may further include a calibration module (not shown) that may be used to establish a linear correspondence of target frequency to color to calibrate the correspondence of target frequency to color. In some examples, the calibration module may be configured to obtain a frequency range of the target frequency, and calibrate the correspondence between the target frequency and the color based on a linear correspondence between the frequency range and a preset color range. This makes it possible to intuitively obtain the frequency corresponding to the color in the color image. In some examples, the linear correspondence may be expressed as:

wherein [ H ]_min，H_max]Is a frequency range, H_inIs the frequency of the input, H_minIs the minimum target frequency, H_maxAt the maximum target frequency, C_outIs the color of the output, [ C ]_min，C_max]In a predetermined color range, C_minIs the minimum value in a predetermined color range, C_maxIs the maximum value in the preset color range. In some examples, [ C_min，C_max]Can be [0, 4/3 pi]. In this case, C_minIs 0, C_maxIs 4/3 pi. Thus, a better visual effect can be obtained. Examples of the disclosure are not limited thereto, and in other examples, [ C_min，C_max]Can be [0, 2 pi ]]. As an example of calibration, fig. 6 shows a color image of a live sample of a mouse retina with calibration information D. As shown in FIG. 6, calibration informationThe message D may be a color scale. In this case, the frequency corresponding to the color in the color image can be intuitively obtained.

The imaging system 1 of the present disclosure obtains a time eigenvector based on singular value decomposition to eliminate artifacts of a plurality of interferograms through fluctuation degrees of the time eigenvector and performs signal-to-noise enhancement processing on the plurality of interferograms by using a sliding window, and finally generates a color image based on time domain information and frequency domain information of the plurality of interferograms. Thereby, the influence of artifacts or noise on the image quality of the color image can be reduced, further research is facilitated based on the color image, and in addition, the color image generated by the present disclosure can reflect the internal dynamic signal of the biological tissue and the manner of obtaining the color image is non-invasive.

Hereinafter, an imaging method of the present disclosure for signal processing of a planar interferometer is described in detail with reference to fig. 7. The imaging method for signal processing of a planar interferometer to which the present disclosure relates may sometimes be simply referred to as an imaging method. The imaging method is applied to the imaging system 1 described above. Fig. 7 is a flowchart illustrating an imaging method for signal processing of a planar interferometer according to an example of the present disclosure.

In some examples, as shown in fig. 7, the imaging method may include acquiring a plurality of interferograms and composing a first set of interferograms (step S110), and in step S110, the sample may be acquired using a planar interferometer to obtain the plurality of interferograms. In some examples, the sample may be a live sample. In some examples, multiple interferograms may be obtained by successive acquisitions. In some examples, the plurality of interferograms may be grouped into a first set of interferograms by acquisition time. Therefore, the subsequent simultaneous processing of a plurality of interference patterns can be facilitated. The detailed description may refer to the related description of the acquisition module 10.

In some examples, as shown in fig. 7, the imaging method may include deghosting the first interference image set to obtain a second interference image set (step S120). In this case, eliminating the artifacts in the first interference image set can improve the image quality of the subsequently generated color image. In some examples, in the artifact removal process of step S120, the first interference image set may be subjected to singular value decomposition to obtain a temporal feature matrix and a singular value matrix. In some examples, after the temporal feature matrix and the singular value matrix are obtained, each column of the temporal feature matrix may be taken as a temporal feature vector. In some examples, the singular values in the singular value matrix may be updated based on a degree of fluctuation of the temporal eigenvector. Specifically, the fluctuation degree of the time eigenvector may be calculated, the time eigenvector meeting the threshold condition is screened based on the fluctuation degree, and then the singular value in the singular value matrix corresponding to the time eigenvector is updated to the preset singular value. Thereby, the singular values in the singular value matrix can be updated based on the fluctuation degree of the temporal eigenvector. In some examples, the preset singular value may be 0. The detailed description may refer to the related description of the artifact removal module 11.

In some examples, in step S120, the degree of fluctuation may be represented by an accumulated zero-crossing rate of the temporal feature vector. That is, the degree of fluctuation of the temporal feature vector can be expressed by the cumulative zero-crossing rate of the temporal feature vector. As described above, each column of the temporal feature matrix may be taken as a temporal feature vector. In some examples, the cumulative zero-crossing rate of the ith column of temporal feature vectors may be expressed as: d _ ZRC_i＝|ZRC_i+1-ZRC_iWhere i is the column index of the time feature matrix, ZRC_iIs the zero-crossing rate of the ith column of temporal feature vectors. Thereby, the degree of fluctuation of the temporal feature vector can be expressed by the cumulative zero-crossing. In some examples, the threshold condition may be that the cumulative zero-crossing rate of the respective temporal feature vectors is greater than a preset value. In some examples, the preset value may be 3 times the standard deviation of the cumulative zero-crossing rate. Thus, the artifact can be eliminated by eliminating the signal in which the degree of dispersion of the cumulative zero-crossing rate reaches the preset value degree. A detailed description can be found in relation to the cumulative zero-crossing rate in the artifact removal module 11 of the imaging system 1. In some examples, in step S120, the first interference image set may be reconstructed based on the updated singular value matrix and used as the second interference image set. For a detailed description, reference may be made to the elimination artifact module 11 with respect to the reconstructionA correlated description of a set of interference images.

In some examples, as shown in fig. 7, the imaging method may include performing signal-to-noise enhancement processing on the second interference image set to obtain a third interference image set (step S130). In step S130, a sliding window with a preset length may be used to perform signal-to-noise ratio enhancement processing on pixel values in each sliding window in the second interference image set to obtain a third interference image set. In some examples, the preset length may be greater than 1. In some examples, the preset length may be 8. In some examples, the dimensions of the sliding window and the dimensions of the second set of interference images may coincide. In some examples, the sliding window may be moved in the direction of the acquisition time by a preset step size. In some examples, the preset step size may be a preset proportion of the preset length. For example, the predetermined ratio may be 25%, 50%, 75%, 1/3, 2/3, or the like. In some examples, in the signal-to-noise ratio enhancement process, an average value of pixel values in each sliding window may be obtained. In some examples, the difference between the pixel values in the respective sliding windows and the average value may be summed to obtain a cumulative value. In some examples, the absolute value of the accumulated value may be divided by a preset length to obtain an average accumulated value, and the average accumulated value may be used as the pixel value of the third interference image set. Thus, the signal-to-noise ratio enhancement processing can be performed on the second interference image set based on the sliding window and cumulative sum algorithm. The detailed description can be referred to the related description of the snr enhancement module 12.

In some examples, as shown in fig. 7, the imaging method may include generating a color image based on the third interference image set (step S140). The color image may reflect the internal dynamic signal of the fresh sample. In some examples, components of a three-dimensional color space may be obtained based on the third set of interference images and the third set of interference images may be turned into color images based on the components of the three-dimensional color space. In some examples, the components of the three-dimensional color space may include a hue component, a saturation component, and a brightness component (which may also be referred to as a luma component). In some examples, a plurality of pixel values of the same location of the plurality of interferograms in the third set of interference images may be acquired to form a sequence of pixels ordered by acquisition time, a component of a three-dimensional color space of each location is acquired based on time domain information and frequency domain information of the sequence of pixels of each location, and the third set of interference images is converted into a color image based on the component of the three-dimensional color space. The detailed description may refer to the related description of the dynamic imaging module 13.

In some examples, the tonal components may correspond to the target frequency in step S140. In some examples, the target frequency may be considered a tonal component. In some examples, the target frequency may be an average frequency or a median frequency. In some examples, the sequence of pixels for each location may be fourier transformed to obtain a sequence of frequencies, and the average frequency may be obtained as a tonal component based on the sequence of frequencies and the power spectral density. In some examples, in step S140, the power spectral density may be acquired based on the pixel sequence of each location and the inverse of the bandwidth of the power spectral density may be taken as the saturation component. In some examples, the power spectral density may be obtained based on a sequence of pixels at various locations and using a Welch method. In some examples, in step S140, a standard deviation or a variance of the pixel sequence may be acquired based on the pixel sequence of each position and taken as a luminance component. The detailed description may be referred to the relevant description of the hue component, the saturation component, and the luminance component in the dynamic imaging module 13.

However, examples of the present disclosure are not limited thereto, and in other examples, the imaging method may not include step S120 and/or step S130. That is, the multiple interferograms used to generate the color image may not be subjected to artifact removal processing and/or signal-to-noise enhancement processing.

In some examples, the imaging method further comprises establishing a linear correspondence of target frequency to color to calibrate the correspondence of target frequency to color (not shown). In some examples, the method may be used to obtain a frequency range of the target frequency, and calibrate the correspondence between the target frequency and the color based on a linear correspondence between the frequency range and a preset color range. This enables intuitive acquisition of the frequency corresponding to the color in the color image. The detailed description can refer to the relevant description of the calibration module.

The imaging method disclosed by the invention obtains the time characteristic vector based on singular value decomposition, eliminates the artifacts of a plurality of interference patterns through the fluctuation degree of the time characteristic vector, performs signal-to-noise ratio enhancement processing on the plurality of interference patterns by using the sliding window, and finally generates a color image based on the time domain information and the frequency domain information of the plurality of interference patterns. Thereby, the influence of artifacts or noise on the image quality of the color image can be reduced, further research is facilitated based on the color image, and in addition, the color image generated by the present disclosure can reflect the internal dynamic signal of the biological tissue and the manner of obtaining the color image is non-invasive.

While the invention has been described in detail in connection with the drawings and the embodiments, it is to be understood that the above description is not intended to limit the invention in any way. Those skilled in the art can make modifications and variations to the present invention as needed without departing from the true spirit and scope of the invention, and such modifications and variations are within the scope of the invention.

Claims

1. An imaging system for signal processing of a planar interferometer based on the principle of two-beam equal thickness interferometry for measurement, comprising: the device comprises an acquisition module, an artifact eliminating module, a signal-to-noise ratio enhancing module and a dynamic imaging module; the acquisition module is used for acquiring a plurality of interferograms obtained by continuously acquiring a fresh sample by using the planar interferometer, and the interferograms form a first interferogram set according to acquisition time; the artifact eliminating module is used for performing singular value decomposition on the first interference image set to obtain a time characteristic matrix and a singular value matrix, each column of the time characteristic matrix is a time characteristic vector, the fluctuation degree of the time characteristic vector is calculated, the time characteristic vector meeting a threshold value condition is screened based on the fluctuation degree, the singular value in the singular value matrix corresponding to the time characteristic vector is updated to be a preset singular value, then the first interference image set is reconstructed based on the updated singular value matrix and is used as a second interference image set, and the accumulated zero-crossing of the time characteristic vector is utilizedThe rate represents the fluctuation degree of the temporal feature vector, the first

The cumulative zero-crossing rate of the column temporal feature vector is expressed as:

，

is a column index of the temporal feature matrix,

is as follows

Zero-crossing rate of the column time eigenvector; the signal-to-noise ratio enhancement module performs signal-to-noise ratio enhancement processing on pixel values in each sliding window in the second interference image set by using a sliding window with a preset length to obtain a third interference image set, wherein the sliding window moves along the direction of the acquisition time according to a preset step length, an average value of the pixel values in each sliding window is obtained in the signal-to-noise ratio enhancement processing, an accumulated value is obtained by performing accumulated sum on differences between the pixel values in each sliding window and the average value, an absolute value of the accumulated value is divided by the preset length to obtain an average accumulated value, and the average accumulated value is used as the pixel value of the third interference image set; and the dynamic imaging module is used for acquiring a plurality of pixel values of the same position of a plurality of interference images in the third interference image set to form a pixel sequence ordered according to the acquisition time, acquiring a component of a three-dimensional color space of each position based on time domain information and frequency domain information of the pixel sequence of each position and converting the third interference image set into a color image based on the component of the three-dimensional color space, wherein the color image reflects an internal dynamic signal of the fresh and live sample and the manner of acquiring the color image is non-invasive, and the dynamic imaging module is used for acquiring a plurality of pixel values of the same position of the plurality of interference images in the third interference image set to form a pixel sequence ordered according to the acquisition time, acquiring a component of a three-dimensional color space of each position based on the time domain information and the frequency domain information of the pixel sequence of each position and converting the third interference image set into the color image based on the component of the three-dimensional color space, wherein the color image reflects an internal dynamic signal of the fresh and live sample and the manner of the fresh and the color image is acquired is non-invasive, and the dynamic imaging module is used for acquiring the color image setThe components of the three-dimensional color space include a hue component, a saturation component, and a brightness component, and a target frequency is acquired based on the frequency domain information, the hue component corresponding to the target frequency.

2. The imaging system of claim 1, wherein:

the imaging system further comprises a calibration module for establishing a linear correspondence of the target frequency and the color to calibrate the correspondence of the target frequency and the color.

3. The imaging system of claim 1, wherein:

the threshold condition is that the cumulative zero-crossing rate of each time feature vector is greater than a preset value, wherein the preset value is 3 times of the standard deviation of the cumulative zero-crossing rate.

4. The imaging system of claim 1, wherein:

the preset length is larger than 1, and the dimension of the sliding window is consistent with that of the second interference image set.

5. The imaging system of claim 1, wherein:

the dynamic imaging module acquires a power spectral density based on the pixel sequence of each position and takes the reciprocal of the bandwidth of the power spectral density as the saturation component;

the dynamic imaging module performs Fourier transform on the pixel sequence of each position to obtain a frequency sequence, and obtains the target frequency as the hue component based on the frequency sequence and the power spectral density;

the dynamic imaging module acquires a standard deviation or a variance of a pixel sequence based on the pixel sequence of each position and takes the standard deviation or the variance as the luminance component.

6. Signal for planar interferometerA method of imaging a process, said planar interferometer based on the principle of two-beam equal thickness interferometry for measurement, said method comprising: acquiring a plurality of interferograms obtained by continuously acquiring a fresh sample by using the planar interferometer, wherein the interferograms form a first interferogram set according to acquisition time; performing singular value decomposition on the first interference image set to obtain a time characteristic matrix and a singular value matrix, wherein each column of the time characteristic matrix is a time characteristic vector, calculating the fluctuation degree of the time characteristic vector, screening the time characteristic vector meeting a threshold condition based on the fluctuation degree to update the singular value in the singular value matrix corresponding to the time characteristic vector to a preset singular value, reconstructing the first interference image set based on the updated singular value matrix and using the first interference image set as a second interference image set, wherein the fluctuation degree of the time characteristic vector is represented by the cumulative zero-crossing rate of the time characteristic vector, and the second interference image set

，

is a column index of the temporal feature matrix,

is as follows

Zero-crossing rate of the column time eigenvector; utilizing sliding windows with preset lengths to carry out signal-to-noise ratio enhancement processing on pixel values in each sliding window in the second interference image set so as to obtain a third interference image set, wherein the sliding windows move along the direction of acquisition time according to preset step length, and in the signal-to-noise ratio enhancement processing, each sliding window is obtainedAveraging pixel values in the windows, performing cumulative summation on difference values between the pixel values in each sliding window and the average value to obtain a cumulative value, dividing an absolute value of the cumulative value by the preset length to obtain an average cumulative value, and taking the average cumulative value as a pixel value of the third interference image set; and acquiring a plurality of pixel values of the same position of a plurality of interference images in the third interference image set to form a pixel sequence ordered according to the acquisition time, acquiring a component of a three-dimensional color space of each position based on time domain information and frequency domain information of the pixel sequence of each position and converting the third interference image set into a color image based on the component of the three-dimensional color space, wherein the color image reflects an internal dynamic signal of the fresh sample and is non-invasive in a manner of acquiring the color image, the component of the three-dimensional color space comprises a hue component, a saturation component and a brightness component, and a target frequency is acquired based on the frequency domain information, and the hue component corresponds to the target frequency.

7. The imaging method according to claim 6, characterized in that:

and establishing a linear corresponding relation between the target frequency and the color so as to calibrate the corresponding relation between the target frequency and the color.

8. The imaging method according to claim 6, characterized in that: