See figure 1 The magnetic resonance scanner 10 includes a scanner housing 12, which includes a bore 14 or other receiving area for receiving a patient or other objects. The main magnet 20 placed in the scanner housing 12 is controlled by the main magnet controller 22 for generating the main magnetic field B in the inspection area in the bore 14 0. Generally, the main magnet 20 is a permanent superconducting magnet surrounded by cryoshrouding 24, but for lower B 0 The field strength can use a resistive or permanent main magnet.
 The magnetic field gradient coil 28 is provided in the housing 12 or on the housing 12 to superimpose the selected magnetic field gradient on the main magnetic field at least in the inspection area. Generally, magnetic field gradient coils include coils for generating three orthogonal magnetic field gradients (eg, x-gradient, y-gradient, and z-gradient). TEM transmit/receive (T/R) radio frequency coil 30 for injection B 1 RF excitation pulses and receive magnetic resonance signals. The illustrative coil 30 is a TEM coil and includes, for example, 8 wire rods 32, an optional endcap 34, and a surrounding radio frequency shield or spacer 36 (shown in dashed lines). The radio frequency coil 30 is placed near the human head 38, which is the relevant object.
 The scanner controller 42 controls the gradient amplifier 44, the multi-channel radio frequency amplifier 46, and the associated coil switching circuit 48 to excite, spatially localize, encode, or otherwise process magnetic resonance in the head object 38. During the transmission phase, the radio frequency amplifier 46 independently drives the amplitude and phase of the radio frequency power delivered to each wire rod of the 8-wire TEM coil 30 to use the TEM coil 30 as an 8-channel transmitter array. In other embodiments, the 8-wire TEM coil 30 can be used as a 4-channel transmitter array, which uses 4 interleaved wire rods for the 4 transmission channels, or 2 wire rods for each transmission channel Of a launch group. In other embodiments, the TEM coil may include multiple wire rods other than 8, such as 10 wire rods, 12 wire rods, 16 wire rods, etc., which are driven into a 4-channel transmitter and 5-channel transmission. Machine, 8-channel transmitter, 10-channel transmitter, 16-channel transmitter, etc. In other embodiments, other transmitter arrays may be used, such as a degenerate birdcage coil with decoupled meshes defining multiple transmit channels or a surface transmit coil array with multiple independent transmit channels. Wait.
 Each transmission channel works independently with the selected amplitude and phase of the input radio frequency power. B based on head 38 or other objects 1Mapping 58, the transmission configuration selector 54 selects the amplitude and phase for exciting magnetic resonance in the selected area for each transmission channel. The amplitude and phase selected for each transmission channel together define the selected transmission configuration 60, which generates B when applied to 8 wire rods 32 1 The field is substantially uniform or has another selected spatial distribution over the corresponding selected area of the head 38 or other object.
 In the phase of receiving the magnetic resonance sequence, the coil switching circuit 48 connects the TEM head coil 30 as a barrel resonator to the radio frequency receiver 64 to receive the excited and spatially encoded magnetic resonance signal. Depending on the type of magnetic resonance sequence implemented, the magnetic field gradient coil 28 can be operated during at least a part of the reception phase, for example to achieve frequency encoding or destruction of magnetic resonance. The data buffer 66 stores the received magnetic resonance signal, usually after the magnetic resonance signal is digitized and optionally undergoes other signal processing. In some embodiments, during the phase of receiving the magnetic resonance sequence, a separate receiving-only coil (not shown in the figure) is used instead of using the same coil 30 for both the transmitting and receiving phases.
 To perform imaging, the reconstruction processor 70 performs reconstruction processing on the collected magnetic resonance data, so as to thereby generate a reconstructed image or map. For example, the reconstruction processor 70 may use Fast Fourier Transform (FFT) or other reconstruction algorithms to process the spatially encoded magnetic resonance data to generate a spatial map or image of the object. For spectroscopy or other magnetic resonance applications, other types of post-processing can be used in combination with spatial image or map reconstruction, or other types of post-processing can replace spatial image or map reconstruction.
 The image memory 72 stores the reconstructed image or map. The user interface 74 displays the reconstructed image or graph to the associated user. in figure 1 In the illustrative embodiment shown, the user interface 74 also allows the user to interact with the scanner controller 42 to control the magnetic resonance scanner 10. In other embodiments, different scanner control interfaces may be provided. In some embodiments, the user interface 74 may be a computer or other digital electronic device. Optionally, the reconstruction processor 70, the memories 66, 72, and/or other components are combined with a computer or a digital electronic device as a software component, a hardware additional system, or the like.
 In some embodiments, some or all of the coil switching circuit 48 is located on the TEM coil 30 or other radio frequency coils. In some embodiments, the coil switching circuit 48 may selectively configure the radio frequency coil as a mono-tube receiving coil or as a receiving coil array. For example, each wire rod of the TEM coil 30 or the selected wire rod group is optionally used as a SENSE (inductive) receiving element in the magnetic resonance sequence receiving stage. In some embodiments, the coil 30 can be used as an 8-channel transmitter and 8-channel receiving array with suitable switching circuits. In some embodiments, different transmit and receive coils or coil arrays are provided.
 B can be determined in various ways 1 Mapping 58. In one way, according to the B of the object obtained by using the TEM coil 30 and the magnetic resonance scanner 10 1 Mapping the measured value, determine B 1 Mapping 58. Alternatively, the B 1 Mapping 58. Suitable anatomical models for various parts of human anatomy and the anatomy of Sprague-Dawley rats, dwarfs, and macaques can be obtained from the United States Air Force Research Laboratory (http://www.brooks.af. mil/AFRL/HED/hedr, last interview on August 30, 2005).
 The inventors have found that a substantially uniform B is provided on a selected area of the head 38 1 The launch configuration of the field may provide highly uneven B on another selected area 1 field. For example, to provide a substantially uniform B for axial slices near the head 38 1 The launch configuration of the field may provide highly uneven B for the axial slices closer to the center. 1 field. Therefore, the transmission configuration selector 54 repeatedly determines the selected transmission configuration 60 for multiple slices, multiple adjacent slice groups, or multiple other selection regions. In some embodiments, the selection area corresponds to the acquisition area. For example, the selected launch configuration 60 can be re-determined for each axial slice acquired. In other embodiments, each selection area corresponds to multiple adjacent acquisition areas. For example, the selected emission configuration 60 can be re-determined for the top slice group, for one, two or more middle adjacent slice groups, and for slice groups close to the neck region.
 Combine below figure 2 A suitable embodiment of the transmission configuration selector 54 is described. The transmission configuration selector 54 sets the B provided by the scanner controller 42 for example 1 The map 58 and the selection area 90 are received as input. Optionally, the transmission configuration selector 54 also receives the previously selected transmission configuration 92 (if any) and uses it as the starting transmission configuration to be considered. Or (ie, if there is no previously selected transmission configuration), the default transmission configuration 94 can be used as the starting transmission configuration to be considered. For example, the default emission configuration may be known to provide a substantially uniform B for the selected area 90 of a typical head 1 The emission configuration of the field, alternatively, can be known to provide a substantially uniform B when the TEM coil 30 is not loaded 1 The launch configuration of the field.
 The transmission configuration 96 to be considered is initially the previously selected transmission configuration 92, the default transmission configuration 94, and so on. B 1 The field mapping 58 determines B according to the position of the emission configuration to be considered at least within the selection area 90 1 field. B 1 The mapping 58 can directly measure B using the radio frequency coil 30 and the magnetic resonance scanner 10 1 Field, alternatively, can also be estimated by modeling or calculation B 1 field. In order to proceed B 1 Field modeling or calculation, using phantom data 80 or anatomical model 82 ( figure 1 Shown) together with the model of the radio frequency coil 30. B 1 The mapping 58 can be isotropic or anisotropic, it can be the same as the imaging resolution, or, in order to speed up the B 1 The calculation of the field, it can also be coarser than the imaging resolution. The coarse resolution is suitable for B 1 Field is modeled because B 1 The uneven pattern of the field is expected to exhibit a very low spatial frequency.
 In some embodiments, B 1 Field mapping 58 uses XFDTD full wave 3D electromagnetic parser software (available from Remcom, State College, PA). The electromagnetic field generated by each wire rod 32 is calculated according to the amplitude and phase of the wire rod given by the emission configuration to be considered, and the combined electromagnetic field generated by all the wire rods 32 depends on each wire rod 32 alone. B produced in object 38 1 Superposition of radio frequency fields. Calculate B for all cells or pixels in at least selected area 90 1 + Magnetic resonance excitation field. The described FDTD method is just an example, other techniques can also be used to calculate B 1 Field or pair B 1 The field is modeled.
 B 1 The field quality evaluator 102 evaluates the calculated B for the considered launch configuration 1 The quality of the field. Various measures can be used to evaluate B 1 Field quality. The range measure expressed as "r" in this article is appropriately given by the following equation:
 r = | B 1 + | max | B 1 + | min - - - ( 1 )
 Among them, the corresponding range "r" is determined for the selection area 90. The term "range" and the corresponding symbol "r" are intended to cover obvious variations of equation (1), including, for example, linear scaling or normalization, reciprocal ratios, and the like. The statistical deviation measure denoted as "s" herein is appropriately given by the variance, standard deviation, root mean square (rms) value, etc. applied to the selected region 90. It is also possible to use specific absorption rate (SAR) measures, such as local SAR values (maximum SAR on a local barrel unit, such as the average value on a 10 g local barrel unit), or head SAR (maximum average in head 38). SAR).
 The inventors have found that the evaluation of B based on a single measure 1 The field, for example, based only on the range "r" or only on the statistical deviation "s", or only on the local SAR or only on the head SAR, usually does not get a satisfactory selected emission configuration. For example, selecting a transmission configuration by individually minimizing the statistical deviation "s" may produce a mostly uniform B across the selection area 90 including one or more areas. 1 Field, here | B 1 Significantly deviates from the average|B 1 | avg , Resulting in an undesirable large range value "r" and high local SAR. Similarly, selecting the launch configuration so that the range "r" is closest to one, regardless of other quality measures, may produce B without a spatial location 1 Field, here B 1 The field becomes very large or very small. However, because B 1 The apparently smaller amplitude change of the field, so that the selected B 1 The field may exhibit unexpectedly large statistical deviations, or the power requirements may be relatively high, etc.
 Accordingly, B 1 The field quality evaluator 102 uses at least two different quality measures to evaluate the calculated B 1 Field quality. In one embodiment, the evaluator 102 uses the range "r" measure and the statistical deviation "s" measure to evaluate together, for example, by minimizing the statistical deviation "s" while making the range "r" smaller than the threshold value to evaluate B 1 Field quality:
 s—>0 + And r 0 , Where r 0 = Threshold (2)
 Other evaluation methods can also be used. For example, in the evaluation of formula (2), the threshold value of the local or head SAR can be used to replace the range threshold standard, and the following evaluation can be obtained:
 s—>0 + And SAR 0 , Where SAR 0 = Threshold (3)
 Here SAR can refer to local SAR, head SAR, or the maximum specific absorption rate or average specific absorption rate on another selected tube.
 The non-exhaustive searcher 110 applies B to the different launch configurations considered 1Field map 58 and evaluator 102 to determine the selected transmission configuration 60. The search is non-exhaustive. As discussed in the background art, for "N" transmitting elements, each has "A" amplitude steps or setting values that span the reachable radio frequency power amplitude range, and a phase range of 0°-360°. "P" phase orders or set values, the total number of possible emission configurations of N emitting elements is (A×P) N. For the illustrative case of N=8, A=10, P=36, the total number of possible transmission configurations is approximately 2.8×10 20. Because the image size used to calculate "r", "s", "SAR" or other evaluation values is quite large (for example, in some embodiments, the image size is 100×100×the number of slices), the number of possible transmission configurations Represents an impractical exhaustive search.
 The non-exhaustive searcher 110 cannot perform an exhaustive search. The non-exhaustive searcher 110 searches for a portion of possible transmission configurations, and for each such considered transmission configuration 96, applies B 1 Field map 58 and evaluator 102.
 image 3 A possible non-exhaustive search suitable for execution by the non-exhaustive searcher 110 is shown. The single-channel amplitude search step 112 is performed on the current transmission channel without changing the amplitude or phase of other channels. Consider "A" amplitude steps for the current channel, and update the amplitude of the current channel with the considered amplitude setting value or order evaluated by the evaluator 102 to generate the best or highest quality B 1 field. For each of the "N" channels, the single-channel amplitude search/update step 112 is repeated to update the amplitude of each channel. This process is repeated "R" times, thereby considering A×N×R configurations.
 Similarly, the single-channel phase search step 114 is performed on the current transmission channel without changing the amplitude or phase of other channels. Consider "P" phase orders for the current channel, and update the phase of the current channel with the considered phase setting value or order evaluated by the evaluator 102 to produce the best or highest quality B 1 field. For each of the "N" channels, the single-channel phase search/update step 114 is repeated to update the phase of each channel. This process is repeated "R" times, thereby considering P×N×R configurations.
 The amplitude search/update step and the phase search/update step are repeated "M" times, resulting in a total of (A×N×R+P×N×R)×M considered transmission configurations, or expressed in a simplified form as (A+ P)×N×R×M considered transmission configurations. For the exemplary case of N=8, A=10, P=36, and R=50, M=25, the total number of transmission configurations under consideration is 460,000.
 It should be understood that, because the single-channel searchers 112 and 114 update the amplitude and phase of the current channel respectively, the transmission configuration considered subsequently is based on the transmission configuration considered previously. The inventors have found that for a 4-channel transmitter (exhaustive search is feasible here, but the amount of calculation is large), according to image 3 The method of non-exhaustive search to quickly determine the selected launch configuration, the launch configuration and the implementation of all (10×36) 4 =1.6×10 10 The best transmission configuration identified by an exhaustive search of the next possible transmission configuration is basically the same.
 In general, the number of launch configurations to be considered can be reduced by using prior knowledge to ensure that the initially considered configurations are close to satisfaction. For example, using the previously selected emission configuration 92, which is obtained by selecting adjacent slices that have been imaged, can generally provide a near starting point for the search. In this case, the repetition factors "R" and "M" can be reduced.
 Once the non-exhaustive searcher 110 finds that the transmission configuration evaluated by the evaluator 102 is suitable for selection, the converter 116 optionally adjusts the channel amplitude of the transmission configuration in proportion to the average field|B 1 |Set as target value|B 1 | T. Appropriate scaling factor A S It is given by the following equation:
 A s = | B 1 | T | B 1 | avg - - - ( 4 )
 Where|B 1 | avg By B 1 B calculated by field mapping 58 1 The average value of the field. The amplitude of each transmit channel evaluated as suitable for the selected transmit configuration multiplied by the scaling factor A S To convert the average field|B 1 |Converted to target value|B 1 | T , Resulting in the selected launch configuration 60.
 apart from image 3 In addition to the example shown, other search/update algorithms can also be used. For example, in another conceivable method, the searcher/updater 110 randomly modifies the amplitude or phase of a randomly selected channel. For example, a randomly selected channel can have its amplitude or phase randomly increased or decreased by 1 step. If random modification can be improved to be like B 1 B evaluated by the field quality evaluator 102 1 Field quality, keep the random modification; otherwise, discard it. In addition, in this method, the subsequent considered transmission configuration is derived from the previously considered transmission configuration, so the search is not random, but the search is directed by the evaluator 102 to better satisfy the B 1 The field quality evaluator 102 uses the emission configuration of the evaluation criterion to be driven.
 In another conceivable method, the non-exhaustive searcher 110 executes a genetic algorithm for calculating the total number of chromosomes, each chromosome representing the considered launch configuration. Chromosome genes correspond to the amplitude and phase of each channel—each chromosome includes at least 2×N genes. For the 8-channel transmitter example, a chromosome of 16 genes is suitable. B 1 The field quality evaluator 102 defines the chromosome fitness for determining the chromosome population for breeding the next generation. By randomly or pseudo-randomly changing the gene value, the offspring chromosomes are appropriately mutated to produce a new considered launch configuration, and optionally crossover operators or algorithms are used, and appropriate operations such as gene copy, gene mixing or exchange, gene mutation, etc. , Combine contemporary parent chromosomes to generate offspring chromosomes. In some conceivable methods based on genetic algorithms, soft restart or other techniques for expanding the range of the chromosome population are used to reduce the possibility of premature convergence.
 Regardless of the specific search/update algorithm used, the evaluation value used by the evaluator 102 should be selected to provide a substantially uniform B 1 Field or other desired B 1 Field distribution. However, the evaluation value should not be selected to be so aggressive that none of the considered transmission configurations may be evaluated as satisfactory. For example, using r with close to unity 0 The evaluation of equation (2) may be impractical because there may not be any of the considered launch configurations that can meet this demanding evaluation. On the other hand, the inventors have found that setting r 0 =2.5 A reasonable uniform B can be obtained 1 Field, while the limited number of considered emission configurations searched by the non-exhaustive searcher 110 easily satisfy this condition.
 For head imaging using axial slices as the selection area and using a head coil transmitter with a cover at the end, the inventors have achieved the transmission configuration selection as described above. The head coil transmitter has 4 transmissions. Channel, 8 transmit channel or 16 transmit channel. It is found that when the number of channels is increased from 4 to 8, B can be obtained 1 The field uniformity is significantly improved; however, a further increase to 16 channels yields less improvement, and involves a significantly longer search time.
 The present invention has been described above in conjunction with preferred embodiments. Obviously, after reading and understanding the foregoing detailed description, a person of ordinary skill in the art can make various modifications and changes. The present invention should be interpreted as including all these modifications and changes, because they still fall within the protection scope of the claims or their equivalents.