JPH05168621A - X-ray ct system - Google Patents

Abstract

PURPOSE:To constitute a measuring instrument for handling a minute signal such as an X-ray CT system, etc., so that it is prevented that a minute disturbance noise is mixed into measuring data, and this noise appears as a beat noise on the measuring data by synchronizing with a measuring sampling period, and a moire pattern artifact generated on an image due to this noise is reduced. CONSTITUTION:A weighted mean processing between measuring data of every ch and the view direction being adjacent to this ch or ch direction data is executed, and by replacing this value with the measuring data and forming it in the same shape as the original measuring data, all the same image processing as a conventional one is executed. Accordingly, a mixed noise appears as an oblique line-like noise (beat noise) on the measuring data by synchronizing a measuring sampling period. When a suitable weighted mean processing of the measuring data and the adjacent data is executed, this beat noise is scattered suitably and subjected to smoothing, and a moire pattern on an image comes not to be generated.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an X-ray CT apparatus for removing a beat pattern from measurement data.

[0002]

2. Description of the Related Art Mechanical vibration noise of an X-ray tube device or X-ray scanner, switching noise of an X-ray generator, and induction noise from an AC voltage source are measured data D of an X-ray CT device.
_May be mixed in _ij . Here, i is the channel number of the X-ray detector, and j is the view direction number in the circumferential direction of the subject. An example of mixing vibration noise and induction noise is shown in FIG. In FIG. 3A, the horizontal axis represents the transition of the measurement time, and the vertical axis represents the signal amplitude. Black circle marks on the time axis indicate sample points.
FIG. 3A shows that the amplitude portion N ₁ of the minute high-frequency component is external noise including vibration noise and N ₁ induction noise, and the disturbance noise 4 is superimposed on the regular measurement signal. According to FIG. 3A, the disturbance noise N ₁
It can be seen that is uniformly applied to the time transition.

On the other hand, FIG. 3 (b) is an example in which the disturbance noise period matches each sample point. In the figure, all sample points match, but there may be matching at every 2 sample points or every 3 sample points. The disturbance noise N ₂ does not have a large amplitude on the average of the time transition, but when the peak of the disturbance noise coincides with the sample point, the ratio of the disturbance noise N ₂ to the measurement signal increases because the sample point is the measurement point. Therefore, even if the disturbance noise N ₂ is small on average,
When the noise peak value matches the sample point, it cannot be ignored.

FIG. 4A shows a state of the measurement data D _ij defined by the view direction number j and the channel number i in the measurement as shown in FIG. 3A, where D _ij is stored in the buffer memory. Has been). According to FIG. 4A, the random noises are randomly arranged like black circle points. Figure 4 (b)
FIG. 4 is a diagram showing a state of measurement data D _ij in the measurement shown in FIG. FIG. 4B shows an example in which the sampling period and the disturbance noise period coincide with each other at a specific period, which causes noise (beat pattern noise) on the diagonal line in the measurement data D _ij .

FIG. 4C is a CT image when the measurement is performed by the measurement system having random noise of FIG. 4A and the data at this time is reconstructed. In this CT image, if the random noise is large, the noise is increased, but it does not appear clearly on the figure. Figure 4 (d) shows
It is a CT image when the data measured by the measurement system having the beat pattern of FIG. 4B is reconstructed. Figure 4
According to (d), a spiral artifact (called a moire pattern artifact) occurs.

Conventionally, in order to remove beat pattern noise, the setting of each frequency is shifted so that the resonance frequency of each unit, the measurement cycle, and the X-ray inverter cycle are not synchronized.
In addition to mounting a device for preventing vibration of the detector, increasing the distance between the detection circuit and the noise source, and mechanical processing such as shield processing, electrical processing such as I / F filter between units has been used together.

[0007]

Due to the downsizing of the apparatus and the high output of the X-ray apparatus, which are the above-mentioned conventional measures, it is difficult to completely remove the noise only by the usual method. SUMMARY OF THE INVENTION It is an object of the present invention to provide an X-ray CT apparatus that eliminates moire pattern artifacts generated by image reconstruction from measurement data mixed with beat noise that cannot be removed by the conventional technique and obtains a good CT image.

The present invention performs weighted averaging between measurement data along the view direction number (or channel direction number) so as not to be synchronized with the disturbance noise period mixed into the measurement data D _ij from the outside, and the weighted average is calculated. It was decided to replace the value as the measurement data along the view direction number (or channel direction number).

[0009]

According to the present invention, a weighted average value is taken along the view direction number (or channel direction number) between measurement data, and this is replaced in place of the measurement data. Thus, the beat pattern noise due to the disturbance noise is smoothed and is effectively removed.

[0010]

EXAMPLE FIG. 1 shows an example diagram in which a view direction weighted average of measurement data D _ij is performed. The data group 1 is the measurement data D _ij with the view number j and the channel number i, and the view section is 0 ° to 360 ° (or 0 ° to 180 ° in half scan), which is the view pitch. It is classified and numbered by j. The channel number i indicates the detector number of the multi-channel X-ray detector. In the figure, the channel number i is i = 1, 2, ..., N, and the view number j is j = 1, 2 ,.

The data group 2 is a weighted average data group S _ij obtained by weighting the data group 1 in the view direction. The weighted average processing and the replacement with the measured data are as follows. (1), when i = 1 ..., N-1.

## EQU1 ## S _ij = a × D _ij + b × D _{i + 1, j} where a + b = 1.0 a and b are weighted average weights. (2) When i = n.

## EQU2 ## S _ij = a × D _nj + b × _D1j where a + b = 1.0 The weights a and b are selected according to the following concept. That is, the conventional example in which the weighted average is not performed is an example in which a = 1 and b = 0, and it is understood that the disturbance noise is reduced by decreasing a and increasing b. However, if b is set too large, there is a drawback that the blur of the image becomes large. From the two points of the purpose of reducing such noise and how much blur is allowed
And b are determined. In the above embodiment, the number of measurement data is the same in the data groups 1 and 2.

When the weighted average processing in the channel direction is performed, in the range of j = 2 ...

## _EQU3 ## The process of S _ij = a × D _ij + b × D _{i, j + 1} may be performed. The data of j = 1 and i = n may not be processed, or may be corrected to such an extent that the difference from the processed channel data does not become large. In addition, although an example of processing between adjacent data is shown in this example, processing between data over several channels or several views is also possible, and the weighted average processing value in matrix form over the ch direction can be replaced.

Also, in this example, the original measurement data and the replacement data have the same number of channels and the same number of views, but in the case of measurement with a large number of measurement views, as shown in FIG. It is also effective to add the processing of (Equation 1) ..., (Equation 3) after compressing the measurement data by replacing with. Further, the measurement data may be after offset correction and sensitivity correction, or before correction. Here, the weighting of a and b can be arbitrarily changed, and the blur of the image becomes larger as the value of b becomes larger. However, since this blur becomes more remarkable in the peripheral ch data portion, FIG. As shown in (b), it is also possible to reduce the blur of the image by reducing the weight of b for the peripheral ch. In this example, the shape of the hyperbola is shown, but it is needless to say that this shape may be gradually changed from the periphery to the central portion, and the weight of the central portion H can be arbitrarily changed. After performing the above correction, an ordinary image reconstruction is performed to obtain an image, but details of the reconstruction method by the X-ray CT apparatus and the like will be omitted.

The above various weighted average processing is executed by a computer. In this case, the computer accesses the buffer storing the measurement data D _ij , gives a and b, performs the weighted averaging process, and sets the resulting S _ij as the measurement data at i, j. , Store it in another buffer one after another.

When the weighted averaging process is performed, the peak noise of the measurement data is also reduced, which has the advantage that streak artifact noise can be reduced.

In this embodiment, the X-ray CT apparatus is used as an example.
It can be defined two-dimensionally and is applicable to all measurement data of a measurement system that measures data in time series.

[0017]

According to the present invention, it is possible to suppress beat noise that is mixed with measurement data and that is synchronized with measurement sampling, and eliminate moire pattern-like artifacts generated on an image. Further, since the peak noise of the measurement data is reduced by the weighted averaging process, streak artifact noise is also reduced and a good image can be obtained.

[Brief description of drawings]

FIG. 1 is a diagram illustrating an example of replacement of measurement data by a view direction weighted average processing according to the present invention.

FIG. 2 is a diagram illustrating another embodiment of replacement of measurement data by the view direction weighted average processing according to the present invention.

FIG. 3 is a diagram showing mixing of disturbance noise.

FIG. 4 is an explanatory diagram of an influence on a CT image due to mixing of disturbance noise.

FIG. 5 is a diagram showing a relationship between weighted average weighting factors.

[Explanation of symbols]

1 Measurement data group D _ij 2 Replacement measurement data group S _ij after weighted average processing

Claims (1)

[Claims] 1. In an X-ray CT apparatus for reconstructing using measurement data D _ij defined by an X-ray detector channel number i and a view direction number j, a disturbance noise period mixed into the measurement data D _ij from the outside. In order not to synchronize with, the weighted average is performed between the measurement data along the view direction number (or channel direction number), and this weighted average value is replaced as the measurement data along the view direction number (or channel direction number). An X-ray CT device that reconstructs from the measurement data after this replacement.