CN110954481B - Optimized average polarization demodulation method for catheter polarization sensitive optical coherence tomography - Google Patents

Abstract

The invention relates to a method for optimizing the average polarization demodulation of catheter polarization-sensitive optical coherence tomography, comprising the following steps: respectively adjusting all three-ring polarization controllers in the system, so that reference light and input light are polarized in two balanced detectors ( H and V channels) to reach equilibrium; sample Mueller matrix acquisition; average the acquired Mueller matrix; select the special point of the sample as a reference point, and use the averaged Mueller matrix to calculate the measurement matrix; Matrix decomposition, using the matrix similarity relationship to calculate the phase retardation; coordinate interpolation transformation of the phase retardation in polar coordinates to convert from polar coordinates to Cartesian coordinates to obtain the phase retardation image of the catheter polarization-sensitive optical coherence tomography system sample , the polarized image.

Description

An optimized average polarization demodulation method for catheter polarization-sensitive optical coherence tomography

Technical field

The invention relates to a catheter optical coherence tomography method. In particular, it relates to a polarization demodulation averaging denoising method for catheter polarization-sensitive optical coherence tomography.

Background technique

Catheter OCT imaging technology is the vascular imaging method with the highest image resolution at present, especially the catheter PS-OCT imaging technology, which can solve the medical problem that the stability of atherosclerotic plaque is difficult to judge in vivo, real-time and quickly, and can improve the arterial stability. Prevention and treatment of atherosclerotic diseases. However, the resolution of the existing OCT system has reached the level that it is possible to judge the nature of tissue plaques, but it is still insufficient in terms of tissue penetration ability, image clarity and the accuracy of tissue plaque type judgment. PS-OCT technology is used. , improving the performance of related technologies is the key direction of OCT system development.

Although catheter OCT has made great progress in plaque identification and diagnosis, new technologies are still needed to further improve the accuracy of in vivo detection of plaque morphology and composition. Related studies have shown that tissues containing fibrous structures, such as interstitial collagen or lamellar arterial smooth muscle cells, exhibit birefringence effects. Lipid tissue exhibits a strong depolarization effect. At present, traditional OCT only provides the intensity information of tissue scattering, and cannot measure the polarization properties of tissue such as birefringence effect and depolarization effect. For example, the development of a catheter Polarization sensitive OCT (PS-OCT) system that can detect polarization characteristics will further improve the diagnostic accuracy of the nature and structure of atherosclerotic plaques, and correctly guide revascularization. Tianjin University Ding Zhenyang et al. proposed a similar Mueller matrix polarization solution method (201811088259.0), which can effectively demodulate the polarization information of biological tissues, but this method lacks the averaging process, and has problems such as depolarization and signal-to-noise ratio decline. difficult to overcome effectively.

SUMMARY OF THE INVENTION

The technical problem to be solved by the present invention is to provide a polarization demodulation average denoising method capable of realizing polarization demodulation of PS-OCT images of a catheter. The technical solution is as follows:

A method for optimizing average polarization demodulation for catheter polarization-sensitive optical coherence tomography, comprising the following steps:

(1) Adjust all three-ring polarization controllers in the system respectively, so that the polarization of the reference light and the input light are balanced on the two balanced detectors (H and V channels);

(2) The sample Mueller matrix is obtained, represented by S(z), and the method is as follows:

利用琼斯矩阵转换穆勒矩阵公式将之转换成穆勒矩阵S(z)；The sample is scanned, and the signals carrying the sample information directly collected by the two balanced detectors are respectively subjected to numerical dispersion compensation and interpolation Fourier transform to obtain the signals H and V. Parts H ₁ , H ₂ , V ₁ , V ₂ ; take the values of the four parts corresponding to the same Z position point of the sample to form the Jones matrix of the complex signal of the pixel point

Convert it to the Mueller matrix S(z) using the Jones matrix transformation Mueller matrix formula;

(3) The obtained Mueller matrix is averaged, and the method is as follows:

The Mueller matrix at the z position is expressed as:

Using the moving average window and averaging with surrounding points, the averaged Mueller matrix is expressed as:

I为所选的平均窗口包含的区域，t指代区域中所有的点，N为窗口所包含的点数，s为平均窗口的中心点,平均过程即窗口内的平均值代替中心位置处的值，移动平均窗口处理所有数据点得到所有数据点平均后的穆勒矩阵，代替原穆勒矩阵进行计算；in

I is the area contained in the selected average window, t refers to all points in the area, N is the number of points contained in the window, and s is the center point of the average window. The average process is that the average value in the window replaces the value at the center position. , the moving average window processes all data points to obtain the averaged Mueller matrix of all data points, and replaces the original Mueller matrix for calculation;

(4) Select the special point of the sample as the reference point, and use the averaged Mueller matrix to calculate the measurement matrix. The method is as follows:

将

和

做运算得到样品在z处的测量矩阵M(z_ref,z)：Determine the outer surface of the catheter or the surface of the sample as the reference point, and take the Mueller matrix after averaging these points to write

Will

and

Do the operation to get the measurement matrix M(z _ref ,z) of the sample at z:

M _ST is the sample round-trip Mueller matrix, M _out is the Mueller matrix of the system output optical path, Q _ref is the equivalent Mueller matrix of the reference light;

(5) Perform the Mueller matrix decomposition on the measurement matrix, and use the matrix similarity relationship to obtain the phase delay. The method is as follows:

The measurement matrix M(z _ref ,z) is decomposed by the Lu-Chipman method, and the depolarization and double attenuation effects are eliminated to obtain the Mueller matrix containing only birefringence:

和

是只包含双折射成分的穆勒矩阵，对应于M(z_ref,z),M_ST(z),M_out和Q_ref；where M ^R (z _ref ,z),

and

is the Mueller matrix containing only birefringent components, corresponding to M(z _ref ,z), M _ST (z), M _out and Q _ref ;

Using the matrix similarity relationship, the phase retardation R in polar coordinates is calculated from the measured Mueller matrix containing only birefringence. The formula is as follows:

tr() represents the trace of the matrix;

(6) Perform coordinate interpolation transformation on the phase retardation in polar coordinates, convert polar coordinates into Cartesian coordinates, and obtain the phase retardation image of the sample of the catheter polarization-sensitive optical coherence tomography system, that is, the polarization image.

In this method, the moving window averaging method is added to the measured Mueller matrix in the previous polarization demodulation method based on the similar Mueller matrix, which can significantly improve the signal-to-noise ratio of the polarization demodulated signal and improve the contrast of the polarization characteristic map. Compared with the global phase average and the complex domain average based on the similarity Jones matrix method, the average process of this method completely adopts the real number operation, and the operation is simple and efficient.

Description of drawings

1 is a schematic structural diagram of the catheter polarization-sensitive optical coherence tomography system in the present invention

FIG. 2 is a flow chart of a method of the present invention for polarization demodulation and averaging denoising for catheter polarization-sensitive optical coherence tomography.

Figure 3. Polarization demodulation results of chicken breast tissue

(a) Intensity image (b) Birefringence phase retardation processed by the unaveraged similar Mueller matrix method

(c) Birefringence phase retardation processed by the optimized average similar Mueller matrix method

Figure 4. Polarization demodulation results of porcine myocardial tissue

(a) Intensity image (b) Unaveraged birefringent phase retardation processed by the Mueller-like matrix method

(c) Birefringence phase retardation processed by the optimized average similar Mueller matrix method

Figure 5. The polarization demodulation results of human fingernail tissue

(a) Intensity image (b) Unaveraged birefringent phase retardation processed by the Mueller-like matrix method

(c) Birefringence phase retardation processed by the optimized average similar Mueller matrix method

Detailed ways

In the following, a detailed description will be given of a polarization demodulation averaging denoising method for catheter polarization-sensitive optical coherence tomography of the present invention with reference to the embodiments and the accompanying drawings.

The invention provides a polarization demodulation average denoising method for catheter polarization-sensitive optical coherence tomography, which utilizes the Mueller matrix to characterize the polarization characteristics of the system and the sample, and eliminates the depolarization and double attenuation effects of the system and the sample through matrix decomposition. By averaging the Mueller matrix, the noise is reduced, and by deriving the intrinsic relationship between the sample transmission matrix and the PS-OCT signal matrix, that is, the two matrices are similar and the traces of the similar matrices are equal, the birefringence phase delay of the sample is obtained. Polarization demodulation of PS-OCT images of catheters.

A polarization demodulation average denoising method for catheter polarization-sensitive optical coherence tomography of the present invention is used in the catheter polarization-sensitive optical coherence tomography (PS-OCT) system as shown in Figure 1, and its working principle is:

The outgoing light of the scanning light source 1 of the catheter PS-OCT system enters through the 1 port of the 1:99 fiber coupler 2, and is distributed from the 2 and 3 ports to the sample arm and the reference arm in a 1:99 ratio, respectively. The outgoing light from port 2 of the 1:99 fiber coupler 2 enters the sample arm, and the beam entering the sample arm enters the three-ring polarization controller 3 and then enters the polarization maintaining fiber 4 with a length of 18.5 meters, and enters the 1 of the circulator 6. port, the light exits from port 2 of the circulator 6, the outgoing light enters the imaging catheter 11 through the rotating mechanism 8, and the light reflected by the sample returns from the imaging catheter 11 to the circulator 6, and exits through the 3 port of the circulator 6 . The outgoing light from port 3 of the 1:99 fiber coupler 2 enters the reference arm, and the light entering the reference arm is incident on the single-mode fiber 5 with a length of 18.5 meters. type optical fiber delay line 10 , the reflected light is incident through the 2 port of the circulator 7 , and exits from the 3 port to the three-ring polarization controller 9 . The outgoing light of the sample arm passing through the 3 ports of the circulator 6 and the outgoing light of the reference arm passing through the three-ring polarization controller 9 respectively enter the 50:50 fiber coupler 12 from ports 1 and 2 of the fiber coupler 12 and interfere, And in the ratio of 50:50, it enters the three-ring polarization controller 13 and the three-ring polarization controller 14 from ports 3 and 4 respectively, and the outgoing light is incident on the polarization beam splitters 15 and 16 respectively. The outgoing light from the fiber beam splitter 15 Ports 1 and 2 are incident on the balanced detectors 17 and 18 respectively. The outgoing light of the polarization beam splitter 16 is incident on the balanced detectors 17 and 18 from ports 1 and 2 respectively. The electrical signals of the balanced detectors 17 and 18 are collected by the acquisition card. 19 is received and transmitted to the computer 20.

The light source adopts a fast scanning light source. The polarization maintaining fiber is used in the system to generate the retardation of the orthogonal polarization state, and the polarization diversity acquisition is performed through a polarization beam splitter. The length of the polarization maintaining fiber depends on its birefringence to generate a phase delay equal to half of the ordinary OCT imaging depth. This method ensures that the system can simultaneously present a polarization-divided integrated image of two orthogonal input polarization states in one image, which makes it possible to eliminate the birefringence change of the system caused by the rotation of the catheter subsequently.

As shown in FIG. 2 , a polarization demodulation averaging denoising method for catheter polarization-sensitive optical coherence tomography of the present invention includes the following steps:

1. Polarization adjustment of reference light and input light

First, adjust the system, adjust the three-ring polarization controller 9, the three-ring polarization controller 13 and the three-ring polarization controller 14 respectively so that the light intensities of the reference light on the two balanced detectors 17 and 18 (H and V channels) are equal, Then the three-ring polarization controller 3 is adjusted so that the light intensities of the input light on the two balanced detectors 17, 18 (H, V channels) are also equal.

2. Sample Mueller matrix acquisition

利用琼斯矩阵转换穆勒矩阵公式将之转换成穆勒矩阵S(z)。After adjusting the system, start to scan the sample. The signals carrying the sample information directly collected by the two balanced detectors 17 and 18 are respectively subjected to numerical dispersion compensation and interpolation Fourier transform to obtain the signals H and V, which are obtained through autocorrelation peak searching. The segmentation threshold divides the signal into four parts H ₁ , H ₂ , V ₁ , V ₂ . Consider the entire sample as composed of isolated points, take the values of the four parts corresponding to the same point of the sample to form the Jones matrix of the complex signal of the pixel point, and write the Jones matrix of the point at the z position as

Convert it to the Mueller matrix S(z) using the Jones matrix transformation Mueller matrix formula.

3. Average the obtained Mueller matrix

The Mueller matrix at the z position can be expressed as:

Using the moving average window and averaging with surrounding points, the averaged Mueller matrix is expressed as:

I为所选的平均窗口包含的区域，t指代区域中所有的点，N为窗口所包含的点数，s为平均窗口的中心点,平均过程即窗口内的平均值代替中心位置处的值。移动平均窗口处理所有数据点得到所有数据点平均后的穆勒矩阵，代替原穆勒矩阵进行计算。in

I is the area contained in the selected average window, t refers to all points in the area, N is the number of points contained in the window, and s is the center point of the average window. The average process is that the average value in the window replaces the value at the center position. . The moving average window processes all data points to obtain the averaged Mueller matrix of all data points, and replaces the original Mueller matrix for calculation.

4. Select the special point of the sample as the reference point, and use the averaged Mueller matrix to calculate the measurement matrix

Let M _ST be the sample round-trip Mueller matrix, M _in , M _out denote the Mueller matrix of the optical path of the system, Q _in , Q _ref denote the equivalent Mueller matrix of the input light and reference light, then the average Mueller matrix at the z position It can be expressed as:

则有Determine the outer surface of the catheter or the surface of the sample as the reference point, and take the Mueller matrix after averaging these points to write

then there are

和

分别做运算得到样品在z处的测量矩阵M(z_ref,z)：Will

and

Do the operations separately to get the measurement matrix M(z _ref , z) of the sample at z:

5. Perform Mueller matrix decomposition on the measurement matrix, and use the matrix similarity relationship to find the phase delay

Since the ultimate purpose of the present invention is to find the phase delay, and the Mueller matrix contains parts related to depolarization and double attenuation effects, matrix decomposition is performed to remove the parts related to depolarization and double attenuation effects.

The 4*4 Mueller matrix M is decomposed into the following form by the Lu-Chipman method:

M=M _Δ M _R M _D (13)

Among them, _{M Δ} is a depolarization matrix, MR is a birefringence matrix, and _MD is a double attenuation matrix, and only _MR is taken.

Perform matrix decomposition on the measurement matrix M(z _ref ,z) to eliminate the depolarization and double attenuation effects to obtain a Mueller matrix containing only birefringence:

和

是只包含双折射成分的穆勒矩阵，对应于M(z_ref,z),M_ST(z),M_out和Q_ref。where M ^R (z _ref ,z),

and

is the Mueller matrix containing only birefringent components, corresponding to M(z _ref ,z), M _ST (z), M _out and Q _ref .

Using the matrix similarity relationship, the phase retardation R in polar coordinates is calculated from the measured Mueller matrix containing only birefringence. The formula is as follows:

tr() represents the trace of the matrix.

6. Convert the phase delay from polar coordinates to Cartesian coordinates

The phase retardation in polar coordinates is transformed from polar coordinates to Cartesian coordinates by coordinate interpolation, and finally the phase retardation image of the sample of the catheter polarization-sensitive optical coherence tomography system is obtained, that is, the polarization image.

The coordinate interpolation transformation described is because in the data acquisition process of the PS-OCT system, the depth information A-Scan and the lateral information B-Scan are imaged, and the final imaging result outputs a polar coordinate image, but the actual demand is to Therefore, the processed polar coordinate image needs to be processed into a PS-OCT image in Cartesian coordinates.

As shown in Figure 3-5, it is the effect diagram of the polarization demodulation average denoising method for catheter polarization-sensitive optical coherence tomography used in the present invention. The left is the intensity image, the middle is the polarization image under the traditional Mueller matrix algorithm, and the right is the polarization image. is the polarization image under the average Mueller matrix algorithm. The first row is the image processing result of chicken breast, the second row is the image processing result of pig myocardium, and the third row is the image processing result of nail.

The invention enables the PS-OCT system to fully express the birefringence information of the sample, improves the analysis ability of microscopic lesions in blood vessels, and obtains more characteristic information of atherosclerotic plaques than traditional OCT intensity images. Interpretation, gaining additional capability for intravascular microscopic lesion analysis, while attenuating noise through averaging, resulting in higher definition of birefringent images. The invention utilizes the Mueller matrix to characterize the polarization characteristics of the system and the sample, and eliminates the depolarization and double attenuation effects of the system and the sample through matrix decomposition. By averaging the Mueller matrix, the noise is reduced, and by deriving the intrinsic relationship between the sample transmission matrix and the PS-OCT signal matrix, that is, the two matrices are similar and the traces of the similar matrices are equal, the birefringence phase delay of the sample is obtained. Polarization demodulation of PS-OCT images of catheters.

Although the present invention has been described above in conjunction with the accompanying drawings, the present invention is not limited to the above-mentioned specific embodiments. The above-mentioned specific embodiments are only illustrative rather than restrictive. Under the inspiration of the present invention, many modifications can be made without departing from the spirit of the present invention, which all belong to the protection of the present invention.

Claims (1)

1. A catheter polarization-sensitive optical coherence tomography method for optimizing average polarization demodulation, comprising the following steps: (1) Adjust all the three-ring polarization controllers in the system respectively, so that the polarization of the reference light and the input light are balanced on the H and V channels of the two balanced detectors; (2) The sample Mueller matrix is obtained, represented by S(z), and the method is as follows: Scan the sample, the signals carrying the sample information directly collected by the two balanced detectors are respectively subjected to numerical dispersion compensation and interpolation Fourier transform to obtain the signals H and V, and the autocorrelation peak search is used to obtain the segmentation threshold, and the signal is divided into four parts H ₁ , H ₂ , V ₁ , V ₂ ; take the values of the four parts corresponding to the same Z position point of the sample to form the Jones matrix of the complex signal of the pixel point

Convert it to the Mueller matrix S(z) using the Jones matrix transformation Mueller matrix formula; (3) The obtained Mueller matrix is averaged, and the method is as follows: The Mueller matrix at the z position is expressed as:

Using the averaging window, averaging with surrounding points, the averaged Mueller matrix is expressed as:

in

I is the area contained in the selected average window, t refers to all points in the area, N is the number of points contained in the window, s is the center point of the average window, and the average process is that the average value in the window replaces the value at the center position , the moving average window processes all data points, and obtains the Mueller matrix after the average of all data points, and replaces the original Mueller matrix for calculation; (4) Select the special point of the sample as the reference point, and use the averaged Mueller matrix to calculate the measurement matrix. The method is as follows: Determine the outer surface of the catheter or the surface of the sample as the reference point, and take the Mueller matrix after averaging these points to write

Will

and

Do the operation to get the measurement matrix M(z _ref ,z) of the sample at z:

M _ST is the sample round-trip Mueller matrix, M _out is the Mueller matrix of the system output optical path, Q _ref is the equivalent Mueller matrix of the reference light; (5) Perform the Mueller matrix decomposition on the measurement matrix, and use the matrix similarity relationship to obtain the phase delay. The method is as follows: The measurement matrix M(z _ref ,z) is decomposed by the Lu-Chipman method, and the depolarization and double attenuation effects are eliminated to obtain the Mueller matrix containing only birefringence:

where M ^R (z _ref ,z),

and

is the Mueller matrix containing only birefringent components, corresponding to M(z _ref ,z), M _ST (z), M _out and Q _ref ; Using the matrix similarity relationship, the phase retardation R in polar coordinates is calculated from the measured Mueller matrix containing only birefringence. The formula is as follows:

tr() represents the trace of the matrix; (6) Perform coordinate interpolation transformation on the phase retardation in polar coordinates, convert polar coordinates into Cartesian coordinates, and obtain the phase retardation image of the sample of the catheter polarization-sensitive optical coherence tomography system, that is, the polarization image.

Images