CN-121831643-B - Nuclear magnetic resonance passive shimming module design method and device based on Gaussian splatter

CN121831643BCN 121831643 BCN121831643 BCN 121831643BCN-121831643-B

Abstract

The invention discloses a design method and a device of a nuclear magnetic resonance passive shimming module based on Gaussian splatter, which relate to the technical field of precision electromagnetic field optimization and comprise the following steps of initializing in a design domain of the passive shimming module to obtain an initial parameter set of all Gaussian splatter particles; the method comprises the steps of constructing a mapping relation between equivalent magnetic moment and magnetization weight factor of each Gaussian splutter particle and intrinsic magnetization intensity and scaling vector of a magnetic material, constructing a magnetic field intensity calculation model, inputting an initial parameter set and real magnetic field intensity of all the Gaussian splutter particles into a Gaussian splutter particle parameter set iterative optimization algorithm, calculating predicted magnetic field intensity through the magnetic field intensity calculation model, constructing a multi-objective loss function based on the predicted magnetic field intensity, the real magnetic field intensity, the magnetization weight factor vector and a central position vector, carrying out iterative optimization, obtaining an optimized parameter set, and manufacturing a passive shimming module. The invention solves the problems of low accuracy and efficiency of compensating the magnetic field with complex morphology.

Inventors

YAO KAIWEN
ZHANG YUXING
YAO BIN
WU LIN
ZHANG YIZHUO

Assignees

华侨大学

Dates

Publication Date: 20260512
Application Date: 20260313

Claims (9)

1. The design method of the nuclear magnetic resonance passive shimming module based on Gaussian splatter is characterized by comprising the following steps of: randomly initializing a plurality of Gaussian splats in a design domain of a passive shimming module, and constructing an initial parameter set of all the Gaussian splats, wherein the parameter set comprises a magnetization weight factor vector, a central position vector, a rotation quaternion matrix and a scaling matrix; for each magnetic material filled in the design domain of the passive shimming module, constructing a mapping relation between the equivalent magnetic moment of each Gaussian splatter particle and a magnetization weight factor, and the intrinsic magnetization intensity and scaling vector corresponding to the magnetic material, wherein the construction process of the mapping relation is as follows: In the design domain, calculating the relative displacement of each sampling point in the target magnetic field area relative to the center position of each Gaussian splatter particle, and calculating the space duty ratio by combining the scaling vector of each Gaussian splatter particle, wherein the space duty ratio is shown in the following formula: ; Wherein, the Representing the location of a sample point of one of the gaussian splatter particles in the region of the target magnetic field The space occupation ratio at the point(s), Representing the position of any one of the sample points in the region of the target magnetic field, And Respectively representing the positions of sampling points in the target magnetic field region Relative to the center position of one of the gaussian splatter particles in the x-axis, y-axis and z-axis, 、 And Representing the scaling of one of the gaussian splatter particles in the x, y and z axes, respectively; linearly superposing the magnetization contributions of N Gaussian splats in the design domain to construct the magnetization intensity of one sampling point in the target magnetic field region, wherein the magnetization intensity is shown in the following formula: ; Wherein, the Representing the position of a sampling point in a region of a target magnetic field Is used for the magnetic field of the magnetic field sensor, Representing the peak magnetization of the ith gaussian splatter particle, N representing the total number of gaussian splatter particles; Integrating the volume by utilizing the analytic characteristic of the Gaussian function and dispersing the magnetization intensity which is continuously distributed to obtain the equivalent magnetic moment of each Gaussian splash particle, wherein the equivalent magnetic moment is shown in the following formula: ; Wherein, the Representing the equivalent magnetic moment of the ith gaussian splatter particle, V representing the volume, Representing the magnetization of the ith gaussian splatter particle in the region of the target magnetic field; decomposing the peak magnetization of the ith Gaussian splatter particle into the product of the magnetization weight factor of the ith Gaussian splatter particle and the intrinsic magnetization corresponding to the material, and defining the mapping relation as follows: ; Wherein, the Representing the magnetization weighting factor of the ith gaussian splatter particle, Constructing a magnetic field strength calculation model based on the micro-form of the Piaor-savart law; selecting magnetic materials to be filled in a design domain of a passive shimming module, inputting an initial parameter set and real magnetic field intensity of all Gaussian splats particles into an iterative optimization algorithm of the Gaussian splats particle parameter set and participating in a first iterative optimization process, determining equivalent magnetic moment vectors of all Gaussian splats particles according to a magnetization weight factor vector, a scaling matrix, intrinsic magnetization intensity corresponding to the magnetic materials and the mapping relation obtained in a previous iterative optimization process, inputting the equivalent magnetic moment vectors, the magnetization weight factor vector, a central position vector and a rotation quaternion matrix into a magnetic field intensity calculation model to obtain a predicted magnetic field intensity of each sampling point in a target magnetic field area, constructing a multi-target loss function based on the predicted magnetic field intensity, the real magnetic field intensity, the magnetization weight factor vector and the central position vector of each sampling point in the target magnetic field area, updating the parameter set obtained in the previous iterative optimization process based on the multi-target loss function to obtain the parameter set obtained in the current iterative optimization process, and repeating the iterative optimization process until the maximum times are reached to obtain the optimized parameter set; and carrying out materialization treatment based on the optimized parameter set and the magnetic material to be filled, and manufacturing a corresponding passive shimming module.
2. The gaussian splatter based nuclear magnetic resonance passive shimming module design method according to claim 1, wherein the rotation quaternion matrix is composed of rotation quaternions of all gaussian splatter particles, the scaling matrix is composed of scaling vectors of all gaussian splatter particles, the scaling vectors being scaling amounts of individual gaussian splatter particles in x-, y-, and z-axes, the magnetization weight factor vector is composed of magnetization weight factors of all gaussian splatter particles, and the center position vector is composed of center positions of all gaussian splatter particles.
3. The gaussian splatter based nuclear magnetic resonance passive shimming module design method of claim 2, wherein the magnetic field strength calculation model is: ; Wherein, the Representing the location of the sampling points of all gaussian splatter particles in the region of the target magnetic field The predicted magnetic field strength at the location, Representing the equivalent magnetic moment vectors of the equivalent magnetic moment composition of all gaussian splatter particles, Indicating the location of the center location of all gaussian splatter particles to the sampling point in the targeted magnetic field region Is used for the displacement vector of (a), Representing the magnetization weight factor vector, The center position vector is represented as such, A covariance matrix representing all gaussian splatter particles, expressed as: , Representing a rotation matrix mapped to the rotation quaternion matrix using the rondrigas transform logic, The scaling matrix is represented as such, Representing the transpose of the matrix.
4. A gaussian splatter based nuclear magnetic resonance passive shimming module design method according to claim 3, wherein said multi-objective loss function is expressed as follows: ; Wherein, the Representing a multi-objective loss function, Representing the magnetic field deviation loss function, The sparsity-loss function is represented as, Representing a spatial overlap loss function; And The weight coefficients of the sparsity loss function and the spatial overlap loss function are respectively; The expression of the magnetic field deviation loss function is as follows: ; Wherein, the Representing the position of a sampling point in a region of a target magnetic field The true magnetic field strength at which, N represents the total number of gaussian splatter particles, Represents an L2 norm; The sparsity loss function is expressed as follows: ; Wherein, the Represents L1 regularization; the expression of the spatial overlap loss function is as follows: ; where i and j represent the ith and jth gaussian splatter particles, respectively, For the distance between the center position of the i-th gaussian splatter particle and the center position of the j-th gaussian splatter particle, In order for the safety distance to be a minimum, Representing taking the maximum value of the two; And in the current iterative optimization process, taking the minimum multi-objective loss function as a target, and moving, splitting or aggregating the Gaussian splatter particles through an adaptive moment estimation algorithm to update a parameter set obtained in the previous iterative optimization process.
5. The gaussian splatter-based nuclear magnetic resonance passive shimming module design method according to claim 1, characterized in that based on the optimized parameter set and the magnetic material to be filled, a corresponding passive shimming module is fabricated, which specifically comprises: dividing a space voxel grid in the design domain, and linearly superposing magnetization weight factors of all Gaussian splats in the optimized parameter set and space occupation to obtain the position of a sampling point in the target magnetic field region A density field at the point, as shown in the following formula: ; searching Gaussian splats with the density field higher than a preset density threshold value from all Gaussian splats, converting the Gaussian splats into a topological structure model corresponding to the passive shimming module, and manufacturing the passive shimming module through an additive manufacturing process on the basis of the topological structure model corresponding to the passive shimming module.
6. A gaussian splatter based nuclear magnetic resonance passive shimming module design apparatus, comprising: The initialization module is configured to randomly initialize and generate a plurality of Gaussian splats in a design domain of the passive shimming module, and construct an initial parameter set of all the Gaussian splats, wherein the parameter set comprises a magnetization weight factor vector, a central position vector, a rotation quaternion matrix and a scaling matrix; The model building module is configured to build a mapping relation between the equivalent magnetic moment of each Gaussian splash particle and the magnetization weight factor, the intrinsic magnetization intensity corresponding to the magnetic material and the scaling vector for each magnetic material filled in the design domain of the passive shimming module, and the building process of the mapping relation is as follows: In the design domain, calculating the relative displacement of each sampling point in the target magnetic field area relative to the center position of each Gaussian splatter particle, and calculating the space duty ratio by combining the scaling vector of each Gaussian splatter particle, wherein the space duty ratio is shown in the following formula: ; Wherein, the Representing the location of a sample point of one of the gaussian splatter particles in the region of the target magnetic field The space occupation ratio at the point(s), Representing the position of any one of the sample points in the region of the target magnetic field, And Respectively representing the positions of sampling points in the target magnetic field region Relative to the center position of one of the gaussian splatter particles in the x-axis, y-axis and z-axis, 、 And Representing the scaling of one of the gaussian splatter particles in the x, y and z axes, respectively; linearly superposing the magnetization contributions of N Gaussian splats in the design domain to construct the magnetization intensity of one sampling point in the target magnetic field region, wherein the magnetization intensity is shown in the following formula: ; Wherein, the Representing the position of a sampling point in a region of a target magnetic field Is used for the magnetic field of the magnetic field sensor, Representing the peak magnetization of the ith gaussian splatter particle, N representing the total number of gaussian splatter particles; Integrating the volume by utilizing the analytic characteristic of the Gaussian function and dispersing the magnetization intensity which is continuously distributed to obtain the equivalent magnetic moment of each Gaussian splash particle, wherein the equivalent magnetic moment is shown in the following formula: ; Wherein, the Representing the equivalent magnetic moment of the ith gaussian splatter particle, V representing the volume, Representing the magnetization of the ith gaussian splatter particle in the region of the target magnetic field; decomposing the peak magnetization of the ith Gaussian splatter particle into the product of the magnetization weight factor of the ith Gaussian splatter particle and the intrinsic magnetization corresponding to the material, and defining the mapping relation as follows: ; Wherein, the Representing the magnetization weighting factor of the ith gaussian splatter particle, Constructing a magnetic field strength calculation model based on the micro-form of the Piaor-savart law; The parameter optimization module is configured to select magnetic materials to be filled in a design domain of the passive shimming module, input an initial parameter set and real magnetic field strength of all Gaussian splats to an iterative optimization algorithm of the Gaussian splats parameter set and participate in a first iterative optimization process, determine equivalent magnetic moment vectors of all Gaussian splats according to magnetization weight factor vectors, scaling matrixes, intrinsic magnetization strength corresponding to the magnetic materials and the mapping relation obtained in a previous iterative optimization process, input the equivalent magnetic moment vectors, magnetization weight factor vectors, a central position vector and a rotation quaternion matrix to the magnetic field strength calculation model to obtain a predicted magnetic field strength of each sampling point in a target magnetic field area, construct a multi-objective loss function based on the predicted magnetic field strength, the real magnetic field strength, the magnetization weight factor vector and the central position vector of each sampling point in the target magnetic field area, update the parameter set obtained in the previous iterative optimization process based on the multi-objective loss function to obtain the parameter set obtained in the current iterative optimization process, and repeat the above optimization process until the maximum iterative times are reached; And the materialization module is configured to carry out materialization processing based on the optimized parameter set and the magnetic material to be filled, so as to manufacture a corresponding passive shimming module.
7. An electronic device, comprising: one or more processors; storage means for storing one or more programs, When executed by the one or more processors, causes the one or more processors to implement the method of any of claims 1-5.
8. A computer readable storage medium, on which a computer program is stored, characterized in that the program, when being executed by a processor, implements the method according to any of claims 1-5.
9. A computer program product comprising a computer program which, when executed by a processor, implements the method according to any one of claims 1-5.

Description

Nuclear magnetic resonance passive shimming module design method and device based on Gaussian splatter Technical Field The invention relates to the technical field of precision electromagnetic field optimization, in particular to a method and a device for designing a nuclear magnetic resonance passive shimming module based on Gaussian splashing. Background In the nuclear magnetic resonance detection instrument system, the spatial uniformity of the main magnetic field directly determines the signal-to-noise ratio, the spatial resolution and the spectral line quality of the detection signal. To compensate for magnetic field inhomogeneities due to factors such as main magnet inherent errors, manufacturing assembly errors, and environmental disturbances, it is often necessary to design passive shimming modules to generate a compensating magnetic field within the target detection region. Most of the existing passive shimming module design methods are based on spherical harmonic function expansion, a target field method or a topological optimization method based on a regular grid. The publication nos. CN114636958B and CN119780810B disclose the use of linear target optimization methods to determine the thickness distribution of the iron sheet matrix. The paper titled A PASSIVE SHIMMING method for Halbach magnet based on MAGNETIC SHEET ARRAYS uses spherical harmonic decomposition to compensate for magnetic fields in multiple orders. The paper named An Improved PASSIVE SHIMMING DESIGN Method can quickly obtain the minimum iron sheet arrangement scheme under the lowest magnetic field non-uniformity by using An Improved target field Method, and simultaneously, the iteration time is greatly reduced. However, the method needs to fix the number of shimming trays, optimize the sizes and the intervals of iron sheets in the trays, and has certain limitations in terms of design freedom, local distortion magnetic field compensation capability and complex structural modeling. Although the paper titled "Additive Manufactured and Topology Optimized Passive Shimming Elements for Permanent Magnetic Systems" provides a new idea for shim element design using topology optimization and additive manufacturing to promote permanent magnet field uniformity, traditional topology optimization methods rely strongly on high frequency finite element analysis. Under the high-dimensional parameter space, the technical bottlenecks of exponentially increasing calculated amount, slow convergence speed, vanishing gradient, easy sinking into local optimal solution and the like exist, and the real-time and efficient precise shimming requirements are difficult to meet. And because of its inherent spatial discretization error, it is difficult to achieve high spatial resolution magnetic field compensation. Disclosure of Invention The invention provides a design method and a device of a passive nuclear magnetic resonance shimming module based on Gaussian splashing, which aim to make up for the defects of the existing passive shimming design method and aim to solve the problems of low magnetic field precision and low efficiency of the passive shimming module in the current nuclear magnetic resonance inspection instrument in compensating complex morphology. In a first aspect, the present invention provides a gaussian splatter-based nuclear magnetic resonance passive shimming module design method, comprising the steps of: Randomly initializing a design domain of a passive shimming module to generate a plurality of Gaussian splats, and constructing an initial parameter set of all the Gaussian splats, wherein the parameter set comprises a magnetization weight factor vector, a central position vector, a rotation quaternion matrix and a scaling matrix; Constructing a mapping relation between the equivalent magnetic moment of each Gaussian splash particle and a magnetization weight factor and between the intrinsic magnetization intensity corresponding to the magnetic material and a scaling vector according to each magnetic material filled in a design domain of the passive shimming module; Selecting magnetic materials to be filled in a design domain of a passive shimming module, inputting an initial parameter set and real magnetic field intensity of all Gaussian splats particles into an Gaussian splats particle parameter set iterative optimization algorithm and participating in a first iterative optimization process, determining equivalent magnetic moment vectors of all Gaussian splats particles according to a magnetization weight factor vector, a scaling matrix and intrinsic magnetization intensity and mapping relation corresponding to the magnetic materials obtained in a previous iterative optimization process, inputting the equivalent magnetic moment vectors, the magnetization weight factor vector, a central position vector and a rotation quaternion matrix into a magnetic field intensity calculation model to obtain a predicted magnetic field intensity of each s