CN105232046B - A kind of single sweep Quantitative MRI Measurement T2 imaging methods based on overlapping echo - Google Patents

Abstract

The invention relates to a single-scan quantitative magnetic resonance T2 imaging method based on overlapping echoes, and relates to a magnetic resonance imaging method. Two echo signals with different evolution times are generated by adding two excitation pulses with the same deflection angle in a single scan, although the evolution times of the two echo signals are different, resulting in different T2 weighting of the two echo signals, The two echo signals come from the same imaging slice, and the prior knowledge between the two echo signals can be used to separate the two echo signals: the similar structure of the two echoes and the sparsity of the joint edge, and the sparse transformation is used to cooperate with the corresponding The separation algorithm separates the two echo signals; finally, the T2 calculation is performed on the two separated signals to obtain a quantitative T2 image. Quantitative T2 imaging of a single scan is obtained, and the time of quantitative T2 imaging is reduced from seconds or even minutes to milliseconds, and the quality of the obtained T2 images is comparable to that obtained by a conventional single-scan EPI sequence.

Description

A single-scan quantitative MRI T2 imaging method based on overlapping echoes

technical field

The invention relates to a magnetic resonance imaging method, in particular to a single-scan quantitative magnetic resonance T2 imaging method based on overlapping echoes.

Background technique

Magnetic resonance parametric imaging (T2 imaging, T2 ^* imaging, and diffusion imaging) has been widely used in clinical diagnosis because of its ability to provide rich quantitative information on tissue characteristic properties ([1] B. Zhao, F. Lam, and ZP Liang , "Model-based MR Parameter mapping with sparsity constraints: parameter estimation and performance bounds," IEEE Trans.Med.Imag., vol.33, no.9, pp.1832-1844, 2014), such as: the diagnosis of myocardial infarction, To measure excess iron in the liver, etc. In particular, the quantitative analysis of T2 relaxation time has attracted more and more attention in clinical medical magnetic resonance imaging such as psychiatry and neurology. However, MRI parametric imaging often needs to acquire a series of contrast-weighted images during its imaging process, and thus generally takes a long time to acquire data. Although there are many different imaging methods proposed to overcome the above problems, such as: downsampling spin echo magnetic resonance imaging (Spin-Echo MRI), gradient spin echo magnetic resonance imaging (Gradient Spin EchoMRI), parallel sensing based on compressed sensing Imaging (Parallel Imaging with CS), etc. At the same time, some model-based reconstruction methods and Bloch simulation-based reconstruction methods have been proposed to further speed up the imaging. However, the multi-shot magnetic resonance parametric imaging method still takes several seconds in the acquisition phase, so that real-time parametric imaging of non-repeatable neural activities becomes an almost impossible task. Therefore, the imaging method of echo-planar imaging (EPI) of single-scan multi-echo is proposed ([2] S.Posse, S.Wiese, D.Gembris, K.Mathiak, C.Kessler, ML Grosse-Ruyken, B. Elghawaghi, T. Richards, SRDager, and VG Kiselev, “Enhancement of BOLD-Contrast Sensitivity by Single-Shot Multi-Echo Functional MR Imaging,” Magn.Reson.Med., vol.42, pp.87– 97,1999), this method consists of acquiring a series of contrast-weighted images within multiple echoes acquired in one scan. However, this method has limitations. On the one hand, this method needs to extend the echo chain, which will inevitably lead to increased acquisition time and signal attenuation; At the expense of time (TR), it may be necessary to sacrifice the spatial resolution of the obtained echo image; and most importantly, this method can only be used for T2* quantitative imaging at present, and cannot be used for T2 quantitative imaging. In addition, although different fast T2 imaging methods have been proposed, including gradient spin echo sequences, these methods all use multiple excitation sequences for T2 imaging, which not only has unsatisfactory effects, but also needs to be improved in imaging efficiency.

Contents of the invention

The purpose of the present invention is to provide a single-scan quantitative magnetic resonance T2 imaging method based on overlapping echoes.

The present invention comprises the steps:

(1) Open the operating software in the magnetic resonance imager on the console of the magnetic resonance imager, first locate the region of interest of the imaging object, and then perform tuning, shimming, power correction, and frequency correction;

(2) Import the pre-compiled OLED imaging sequence; according to the specific experimental situation, set the parameters of the pulse sequence;

The structure of the OLED imaging sequence is as follows: chip select pulse with flip angle α, (TE ₂ −TE ₁ )/2, chip select pulse with flip angle α, TE ₁ /2, 180° refocusing pulse, sampling echo chain;

Combining two small-angle excitation pulses with two echo displacement gradients G ₁ and G ₂ makes the two echoes shift in the center of K space, the 180° refocusing pulse and the two small-angle excitation pulses are both Combined with the layer selection gradient G _ss to perform layer selection; echo delay (TE ₂ -TE ₁ )/2 and TE ₁ /2 are applied before and after the second small-angle excitation pulse, and the 180° refocusing pulse has The destruction gradient effect in the three directions of x, y, and z;

The sampling echo chain is composed of gradient chains acting in the x and y directions respectively; the gradient chain in the x direction is composed of a series of positive and negative gradients, and the area of each gradient is the first echo displacement gradient G ₁ three times of ; the gradient chain in the y direction is composed of a series of "blips" gradients of equal size, and the total area of the "blips" gradient is equal to four times the area of the shift gradient;

Before the sampling echo chain, refocusing gradients G _ror and _{Gar are applied in the x and y directions respectively, the area of the G ror is} half of the first gradient area in the x direction, and the direction is the same as that of the first gradient in the x direction The direction is opposite; the area of the G _ar is half of the total area of all the "blips" gradients, and the direction is opposite to the direction of the "blips"gradient;

(3) perform the described OLED imaging sequence set in step (2), and perform data sampling; obtain the K-space data of two echo signals after the data sampling is completed;

(4) Analyzing the K-space data obtained in step (3) and theoretically deriving the evolution of the echo signal magnetization vector M ₊ , after the _second echo displacement gradient G2, the following formula can be obtained:

In the formula is the spin density distribution, α is the flip angle of the excitation pulse, it is found through experiments that when α=45°, the intensity of the two echo signals are relatively high, δTE=(TE ₂ -TE ₁ )/2, where δ ₁ and δ ₂ correspond to the duration of the first and second echo shift gradients respectively, and γ is the magnetic gyro ratio; it can be seen from the above formula that there are actually three echo signals modulated by different phases, where The first item is generated by the second excitation pulse, and the latter two items are generated by the first excitation pulse; however, it is very complicated to separate these three signals compared to the signal acquired by a single scan, after analysis It can be seen that the echo center positions of the last two items are different, and the signal strength of the last item is relatively small compared with the second item, so the echo of the last item can be ignored through simple processing;

(5) Use the following separation algorithm to separate the echo signals obtained in step (4). According to Fourier transform theory, the linear phases of the two echo signals in the image domain are different. In addition, although the two echo signals The wave signals have different T2 weights due to different evolution times, but they come from the same image layer; therefore, the two echo signals can be jointly reconstructed by using the prior information of the similar image structure of the two, and the reconstruction algorithm is as follows:

Wherein, x ₁ and x ₂ are images reconstructed from the first and second echo signals respectively; is the scale factor, x ₁₀ , x ₂₀ are the initial images of the first and second echo signals respectively; λ ₁ , λ ₂ and λ ₃ are the adjustable constraint weights of the Lagrangian multiplier method; ▽ is the gradient operator; the first item is the fidelity item, the second and third items are the sparsity constraints on the first and second images, and the last item is the contour similarity constraint of the two images; the two images have The relationship is as follows:

in, are the linear phase shifts of the first and second images, respectively; x ₀ is the original signal, which is obtained by inverse Fourier transform of the original signal containing the first and second echo signals, and is solved by an iterative algorithm The above formula can obtain the images generated by the separated first and second echo signals;

(6) The image separated in step (5) is subjected to T2 imaging calculation; for the single-scan T2 imaging method, only two different echo time images are required, and the value of T2 is directly obtained through the T2 relaxation equation have to:

in is the correction factor, ΔTE=TE ₂ -TE ₁ ; S ₁ and S ₂ are the image intensities of the first echo signal and the second echo signal respectively, adding Total Variation (Total Variation) extrapolation method to enhance the image resolution, and set a threshold value, when the obtained value is lower than the threshold value, it will be considered as noise and ignored. Similarly, when the calculated T2 value is too large, it is also unreasonable and will be omitted; finally pass T2 imaging calculations obtained high-quality T2 images with better resolution.

The present invention provides a method of obtaining overlapping echo signals in the case of one scan, and then using a separation algorithm to separate the overlapping signals, and finally performing T2 calculations to obtain the acquisition time and A new imaging method with comparable resolution.

The present invention generates two echo signals with different evolution times by adding two excitation pulses with the same deflection angle in a single scan, although the evolution times of the two echo signals are different, resulting in T2 weighting of the two echo signals different, but these two echo signals come from the same imaging slice, so the prior knowledge between the two echo signals can be used to separate the two echo signals: the similar structure of the two, the sparsity of the joint edge, and thus The two echo signals are separated by sparse transformation and corresponding separation algorithm. Finally, the T2 calculation is performed on the two separated signals to obtain a quantitative T2 image. Using this method, quantitative T2 imaging of a single scan was first obtained, and the time of quantitative T2 imaging was reduced from seconds or even minutes to ms level, and the quality of the obtained T2 image was comparable to that obtained by a conventional single-scan EPI sequence The quality is comparable.

Description of drawings

FIG. 1 is a structural diagram of an OLED imaging sequence in the present invention.

Figure 2 shows a comparison of the results of a model experiment for an OLED imaging sequence. in:

(a) is an image including two echo signals reconstructed by the OLED sequence before separation;

(b) is the first echo signal image after separation from (a);

(c) is the second echo signal image after separation from (a);

(d) is the signal image reconstructed by the multi-scan single-echo spin-echo sequence (SE sequence);

(e) is the signal image reconstructed by the single-scan spin EPI sequence;

(g) is the T2 image reconstructed from (a);

(h) is the signal intensity value and T2 value of the horizontal section along the part corresponding to the dotted line in (d) and (e), respectively.

Figure 3 is the reconstructed T2 image from Figure 2(d).

detailed description

The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments.

Each step in the specific implementation process of the present invention is as follows:

(1) On the operating table of the magnetic resonance imager, open the corresponding operating software in the imager, first locate the region of interest of the imaging object, and then perform tuning, shimming, power and frequency correction;

(2) Import the pre-compiled OLED imaging sequence; according to the specific experimental situation, set the parameters of the pulse sequence;

The structure of the OLED imaging sequence is as follows: chip select pulse with flip angle α, (TE ₂ −TE ₁ )/2, chip select pulse with flip angle α, TE ₁ /2, 180° refocusing pulse, sampling echo chain;

Combining two small-angle excitation pulses with two echo shift gradients G ₁ and G ₂ , so that the two echoes are shifted in the center of K-space, the 180° refocusing pulse and the two small-angle excitation pulses Both are combined with the layer selection gradient G _ss to perform layer selection; echo delay (TE ₂ -TE ₁ )/2 and TE ₁ /2 are respectively applied before and after the second small-angle excitation pulse, and before and after the 180° refocusing pulse There are damage gradient effects in three directions of x, y, and z;

The sampling echo chain is composed of gradient chains acting in the x and y directions respectively; the gradient chain in the x direction is composed of a series of positive and negative gradients, and the area of each gradient is three times that of the displacement gradient _G1 ; The gradient chain in the y direction is composed of a series of "blips" gradients of equal size, and the total area of the "blips" gradient is equal to four times the area of the shift gradient;

Before the sampling echo chain, refocusing gradients G _ror and _{Gar are applied in the x and y directions respectively, the area of the G ror is} half of the first gradient area in the x direction, and the direction is the same as that of the first gradient in the x direction The direction is opposite; the area of the G _ar is half of the total area of all the "blips" gradients, and the direction is opposite to the direction of the "blips"gradient;

(3) Execute the OLED imaging sequence set in step (2), and perform data sampling; after the data sampling is completed, K-space data of two echo signals are obtained.

(4) Analyzing the K-space data obtained in step (3) and theoretically deriving the evolution of the echo signal magnetization vector M ₊ , after the _second displacement gradient G2, the following formula can be obtained:

In the formula is the spin density distribution, α is the flip angle of the excitation pulse, it is found through experiments that when α=45°, the intensity of the two echo signals is relatively high, δTE=(TE ₂ -TE ₁ )/2, Where δ ₁ and δ ₂ correspond to the duration of the first and second echo displacement gradients respectively, and γ is the magnetic gyro ratio. It can be seen from the above formula that there are actually three echo signals modulated by different phases, among which the first item is generated by the second excitation pulse, and the latter two items are generated by the first excitation pulse; however, it is necessary to separate the three Compared with the signal acquired by single scan, the first signal is very complicated. By analyzing the last two items, it can be seen that the echo center positions of the last two items are different, and the signal strength of the last item is relatively small compared with the second item. So the echo of the last item can be ignored by simple processing.

(5) Use the following separation algorithm to separate the echo signals obtained in step (4). According to Fourier transform theory, the linear phases of the two echo signals in the image domain are different. In addition, although the two echo signals Wave signals have different T2 weights due to different evolution times, but they come from the same image layer. Therefore, the two echo signals can be jointly reconstructed by using the prior information that the two image structures are similar. The reconstruction algorithm is as follows:

Where x ₁ and x ₂ are images reconstructed from the first and second echo signals respectively; is the scale factor, x ₁₀ , x ₂₀ are the initial images of the first and second echo signals respectively; λ ₁ , λ ₂ and λ ₃ are the adjustable constraint weights of the Lagrangian multiplier method; ▽ is the gradient operator. The first item is the fidelity item, the second and third items are sparsity constraints on the first and second images, and the last item is the contour similarity constraint of the two images. The two images are related as follows:

in are the linear phase shifts of the first and second images, respectively; x ₀ is the original signal, which is obtained by inverse Fourier transform of the original signal containing the first and second echo signals. The images generated by the separated first and second echo signals can be obtained by solving the above formula through an iterative algorithm.

(6) The image separated in step (5) is subjected to T2 imaging calculation. For the single-scan T2 imaging method, only two different echo time images are required, and the value of T2 is obtained directly through the T2 relaxation equation:

in is the correction factor, ΔTE=TE ₂ −TE ₁ ; S ₁ and S ₂ are the image intensities of the first echo signal and the second echo signal, respectively. Here we have added the Total Variation (Total Variation) extrapolation method to enhance the resolution of the image, and we have set a threshold. When the value obtained is lower than the threshold, it will be considered as noise and ignored. Similarly, when the calculated When the T2 value of is too large, it is also unreasonable and will be omitted. Finally, a high-quality T2 image with better resolution is obtained through T2 imaging calculation.

Provide specific embodiment below:

A water model experiment was carried out with a single-scan quantitative magnetic resonance T2 imaging method based on overlapping echoes to verify the feasibility of the present invention. Before the experiment, the agar gel with a concentration of 0.5% in seven small bottles was mixed with an aqueous solution containing different concentrations of manganese dichloride (MnCl2, 0.01-0.16mM) to generate a series of T2 close to human tissue. value, and the ratio of T1/T2 is about 10 in a 3T imager, which is equivalent to the ratio of T1 to T2 of human tissue under a 3T magnetic field. The time range of the experimentally measured T1 value is about 350-1500 ms. In addition, a T2 image generated by a multi-scan single-echo spin-echo sequence (SE sequence) was used as a reference with an imaging thickness of 2 mm. First, we import the compiled single-scan separation plane imaging sequence based on overlapping echoes as shown in Figure 1, and set the test parameters. The test parameters in this embodiment are set as follows: the excitation time of the 45° excitation pulse is 3 ms, and the sampling in the x direction The number of points N _x is 128, the number of sampling points N _y in the y direction is 64, and the sampling bandwidth sw is 91.4kHz. The imaging field of view FOV _x in the x direction is 20 cm, and the imaging field of view FOV _y in the y direction is 20 cm. After setting the above test parameters, the sampling time of running the whole sequence directly is about 160ms. After the sampling is finished, the sampling data of the overlapping echo signal is obtained. Then we use the separation algorithm in step (5) to separate the two overlapping echo signals. In the separation algorithm here, we set the regularization parameters as λ ₁ =1.4, λ ₂ =1, and λ ₃ =1. The results after separation are shown in Figures 2 and 3. The SE sequence uses 8 different echo delays (ranging from 8.8 to 120ms), the size of the sampling matrix is 128×128, TR=3.5s, and the total scanning time About 1h. It can be seen from Figure a and Figure e that there is obvious distortion, which is caused by the inhomogeneity of the background field. It is worth noting that in the part indicated by the arrow in Figure d and Figure e, the signal strength in each water pipe should be uniform, but the signal strength of the indicated part in the figure is not uniform. This may be caused by the combined effects of many factors, among which the uneven background field and the sensitivity difference of the detection coil are the main reasons for the unevenness of the spin echo sequence image, and the accumulation effect of the signal in the uneven field is also a single factor. The main reason for this phenomenon in scanning EPI sequences. From Figure 2h, it can be found that although the signal intensity of EPI is more heterogeneous, its T2 value is more uniform than the signal intensity. Therefore, the sequence used in the experiment is more robust to inhomogeneous fields and coil sensitivity differences than the spin echo EPI sequence. It can be proved that the OLED imaging method can obtain overlapping echo signals in the case of one excitation, and use the corresponding separation algorithm to separate them, which reduces the acquisition time and improves the spatial resolution of the image.

Table 1

See Table 1 for the description of each symbol.

Claims (1)

1. A single-scan quantitative magnetic resonance T2 imaging method based on overlapping echoes, characterized in that it comprises the steps: (1) Open the operating software in the magnetic resonance imager on the console of the magnetic resonance imager, first locate the region of interest of the imaging object, and then perform tuning, shimming, power correction, and frequency correction; (2) Import the pre-compiled OLED imaging sequence; according to the specific experimental situation, set the parameters of the pulse sequence; The structure of the OLED imaging sequence is as follows: chip select pulse with flip angle α, (TE ₂ −TE ₁ )/2, chip select pulse with flip angle α, TE ₁ /2, 180° refocusing pulse, sampling echo chain; Combining two small-angle excitation pulses with two echo displacement gradients G ₁ and G ₂ makes the two echoes shift in the center of K space, the 180° refocusing pulse and the two small-angle excitation pulses are both Combined with the layer selection gradient G _ss to perform layer selection; echo delay (TE ₂ -TE ₁ )/2 and TE ₁ /2 are applied before and after the second small-angle excitation pulse, and the 180° refocusing pulse has The destruction gradient effect in the three directions of x, y, and z; The sampling echo chain is composed of gradient chains acting in the x and y directions respectively; the gradient chain in the x direction is composed of a series of positive and negative gradients, and the area of each gradient is the first echo displacement gradient G ₁ three times of ; the gradient chain in the y direction is composed of a series of "blips" gradients of equal size, and the total area of the "blips" gradient is equal to four times the area of the shift gradient; Before the sampling echo chain, refocusing gradients G _ror and _{Gar are applied in the x and y directions respectively, the area of the G ror is} half of the first gradient area in the x direction, and the direction is the same as that of the first gradient in the x direction The direction is opposite; the area of the G _ar is half of the total area of all the "blips" gradients, and the direction is opposite to the direction of the "blips"gradient; (3) perform the described OLED imaging sequence set in step (2), and perform data sampling; obtain the K-space data of two echo signals after the data sampling is completed; (4) Analyzing the K-space data obtained in step (3) and theoretically deriving the evolution of the echo signal magnetization vector M ₊ , after the _second echo displacement gradient G2, the following formula can be obtained: <mrow><msub><mi>M</mi><mo>+</mo></msub><mo>=</mo><munder><mo>&amp;Integral;</mo><mover><mi>r</mi><mo>&amp;RightArrow;</mo></mover></munder><mi>&amp;rho;</mi><mrow><mo>(</mo><mover><mi>r</mi><mo>&amp;RightArrow;</mo></mover><mo>)</mo></mrow><mi>sin</mi><mi>&amp;alpha;</mi><mo>{</mo><mo>-</mo><mi>i</mi><mi></mi><msup><mi>cos&amp;alpha;e</mi><mrow><msub><mi>i&amp;theta;</mi><mn>2</mn></msub></mrow></msup><mo>+</mo><mfrac><mn>1</mn><mn>2</mn></mfrac><msup><mi>e</mi><mrow><mo>-</mo><mi>&amp;delta;</mi><mi>T</mi><mi>E</mi><mo>/</mo><msub><mi>T</mi><mn>2</mn></msub><mrow><mo>(</mo><mover><mi>r</mi><mo>&amp;RightArrow;</mo></mover><mo>)</mo></mrow></mrow></msup><mo>&amp;lsqb;</mo><mrow><mo>(</mo><mn>1</mn><mo>+</mo><mi>cos</mi><mi>&amp;alpha;</mi><mo>)</mo></mrow><msup><mi>e</mi><mrow><mi>i</mi><mrow><mo>(</mo><msub><mi>&amp;theta;</mi><mn>2</mn></msub><mo>-</mo><msub><mi>&amp;theta;</mi><mn>1</mn></msub><mo>)</mo></mrow></mrow></msup><mo>+</mo><mrow><mo>(</mo><mn>1</mn><mo>-</mo><mi>cos</mi><mi>&amp;alpha;</mi><mo>)</mo></mrow><msup><mi>e</mi><mrow><mi>i</mi><mrow><mo>(</mo><msub><mi>&amp;theta;</mi><mn>1</mn></msub><mo>+</mo><msub><mi>&amp;theta;</mi><mn>2</mn></msub><mo>)</mo></mrow></mrow></msup><mo>&amp;rsqb;</mo><mo>}</mo><mi>d</mi><mover><mi>r</mi><mo>&amp;RightArrow;</mo></mover></mrow> In the formula is the spin density distribution, α is the flip angle of the excitation pulse, it is found through experiments that when α=45°, the intensity of the two echo signals are relatively high, δTE=(TE ₂ -TE ₁ )/2, where δ ₁ and δ ₂ correspond to the duration of the first and second echo shift gradients respectively, and γ is the magnetic gyro ratio; it can be seen from the above formula that there are actually three echo signals modulated by different phases, where The first item is generated by the second excitation pulse, and the latter two items are generated by the first excitation pulse; however, it is very complicated to separate these three signals compared to the signal acquired by a single scan, after analysis It can be seen that the echo center positions of the last two items are different, and the signal strength of the last item is relatively small compared with the second item, so the echo of the last item can be ignored through simple processing; (5) Use the following separation algorithm to separate the echo signals obtained in step (4). According to Fourier transform theory, the linear phases of the two echo signals in the image domain are different. In addition, although the two echo signals The wave signals have different T2 weights due to different evolution times, but they come from the same image layer; therefore, the two echo signals can be jointly reconstructed by using the prior information of the similar image structure of the two, and the reconstruction algorithm is as follows: <mrow><mo>{</mo><msub><mi>x</mi><mn>1</mn></msub><mo>,</mo><msub><mi>x</mi><mn>2</mn></msub><mo>}</mo><mo>=</mo><munder><mi>argmin</mi><mrow><msub><mi>x</mi><mn>1</mn></msub><mo>,</mo><msub><mi>x</mi><mn>2</mn></msub></mrow></munder><mo>&amp;lsqb;</mo><mo>|</mo><mo>|</mo><msub><mi>x</mi><mn>1</mn></msub><mo>-</mo><msub><mi>x</mi><mn>10</mn></msub><mo>|</mo><msubsup><mo>|</mo><mn>2</mn><mn>2</mn></msubsup><mo>+</mo><msub><mi>&amp;lambda;</mi><mn>1</mn></msub><mo>|</mo><mo>|</mo><mo>&amp;dtri;</mo><msub><mi>x</mi><mn>1</mn></msub><mo>|</mo><msub><mo>|</mo><mn>1</mn></msub><mo>+</mo><msub><mi>&amp;lambda;</mi><mn>2</mn></msub><mo>|</mo><mo>|</mo><mo>&amp;dtri;</mo><msub><mi>x</mi><mn>2</mn></msub><mo>|</mo><msub><mo>|</mo><mn>1</mn></msub><mo>+</mo><msub><mi>&amp;lambda;</mi><mn>3</mn></msub><mo>|</mo><mo>|</mo><mo>&amp;dtri;</mo><mrow><mo>(</mo><msub><mi>x</mi><mn>1</mn></msub><mo>-</mo><msub><mi>&amp;beta;x</mi><mn>2</mn></msub><mo>)</mo></mrow><mo>|</mo><msub><mo>|</mo><mn>1</mn></msub><mo>&amp;rsqb;</mo></mrow> Wherein, x ₁ and x ₂ are images reconstructed from the first and second echo signals respectively; is the scale factor, x ₁₀ , x ₂₀ are the initial images of the first and second echo signals respectively; λ ₁ , λ ₂ and λ ₃ are the adjustable constraint weights of the Lagrangian multiplier method; is the gradient operator; the first item is the fidelity item, the second and third items are the sparsity constraints on the first and second images, and the last item is the similarity constraint on the contours of the two images; the two The images have the following relations: in, are the linear phase shifts of the first and second images, respectively; x ₀ is the original signal, which is obtained by inverse Fourier transform of the original signal containing the first and second echo signals, and is solved by an iterative algorithm The above formula can obtain the images generated by the separated first and second echo signals; (6) The image separated in step (5) is subjected to T2 imaging calculation; for the single-scan T2 imaging method, only two different echo time images are required, and the value of T2 is directly obtained through the T2 relaxation equation have to: <mrow><msub><mi>T</mi><mn>2</mn></msub><mrow><mo>(</mo><mi>r</mi><mo>)</mo></mrow><mo>=</mo><mfrac><mrow><mo>-</mo><mi>&amp;Delta;</mi><mi>T</mi><mi>E</mi></mrow><mrow><mi>l</mi><mi>n</mi><mrow><mo>(</mo><mi>&amp;mu;</mi><mfrac><mrow><msub><mi>S</mi><mn>2</mn></msub><mrow><mo>(</mo><mover><mi>r</mi><mo>&amp;RightArrow;</mo></mover><mo>)</mo></mrow></mrow><mrow><msub><mi>S</mi><mn>1</mn></msub><mrow><mo>(</mo><mover><mi>r</mi><mo>&amp;RightArrow;</mo></mover><mo>)</mo></mrow></mrow></mfrac><mo>)</mo></mrow></mrow></mfrac><mo>,</mo></mrow> in is the correction factor, ΔTE=TE ₂ -TE ₁ ; S ₁ and S ₂ are the image intensities of the first echo signal and the second echo signal respectively, adding Total Variation (Total Variation) extrapolation method to enhance the image resolution, and set a threshold value, when the obtained value is lower than the threshold value, it will be considered as noise and ignored, and when the calculated T2 value is too large, it will be omitted; finally, the T2 imaging calculation has a relatively high High quality T2 images with good resolution.