JP2014081203A - Radiographic apparatus - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a radiographic apparatus which is capable of preventing a false image of a tomographic image by reducing statistical noise included in a poorly permeable area of an analyte image or the like.SOLUTION: In a radiographic apparatus, a logarithmic transformation unit 33 acquires X-ray transmission path length data Gc by logarithmic transformation of transmission data Ga. A difference data generation unit 35 generates difference data Gd by calculating a difference between two pieces of X-ray transmission path length data Gc acquired by the logarithmic transformation unit 33. A noise map image generation unit 37 generates a noise map Nm by, if each pixel value of the difference data Gd is outside a preliminarily set range, determining the pixel value to be statistical noise. In accordance with the noise map image Nm, statistical noise pixels in two pieces of X-ray transmission path length data Gc used for generating the difference data Gd are subjected to noise reduction processing. Thus statistical noise included in a poorly permeable area of an analyte image or the like can be reduced more surely than prior arts.

Description

  The present invention is used for industrial and medical purposes such as foreign body inspection, and obtains transmission data obtained by projecting radiation onto a subject, and obtains radiation transmission path length data obtained by logarithmically converting the acquired transmission data. About.

  Conventionally, there is an X-ray imaging apparatus that acquires a tomographic image as this type of apparatus (for example, see Patent Document 1). The X-ray imaging apparatus includes an X-ray tube that irradiates an X-ray that is a kind of radiation toward a subject, and an X-ray detector that detects X-rays transmitted through the subject. The X-ray detector detects X-rays irradiated from the X-ray tube and transmitted through the subject, and acquires transmission data. The transmission data is acquired from a plurality of different directions with respect to the subject. The acquired plurality of transmission data is logarithmically converted, and the tomographic image is acquired by reconstructing the plurality of logarithmically converted transmission data.

  When logarithmically converting the acquired transmission data, the following two types of image data are prepared in advance. The first is an air image Gair in which the subject is not shown, and the second is an offset image (dark current data) Goff. The air image Gair is an image obtained by irradiating an X-ray without placing a subject between the X-ray tube and the X-ray detector, and the offset image Goff is obtained without irradiating the X-ray. It is the image that was made.

Based on the air image and the offset image, logarithmic conversion processing is performed on each of transmission data Ga photographed from a plurality of directions according to the following equation (1), and converted into X-ray (radiation) transmission path length data Gc.
Gc = log ((Gair-Goff) / (Ga-Goff))
= Log (Gair-Goff) -log (Ga-Goff) (1)

JP 2004-286759 A

  It is generally known that transmission data contains noise. The noise included in the transmission data includes statistical noise caused by the statistical fluctuation of radiation. If the influence of statistical noise is large, a false image appears on the tomographic image by the action of a high-frequency emphasis filter, which is one of the internal processes of the FBP (filtered back projection) method generally used in the tomographic image calculation method. (Artifact) will occur. The statistical noise is a portion where the radiation having a reduced S / N ratio in the transmission data in which the subject image or the like is reflected during the logarithmic transformation of the above-described equation (1) (hereinafter referred to as “hard transmission region”). And the effect appears remarkably.

  In addition, in the conventional statistical noise reduction processing, for example, 10 pieces of transmission data are averaged to obtain one piece of transmission data before logarithmic conversion of the above-described equation (1). Thereby, statistical noise is reduced. That is, by using a plurality of transmission data, the detection amount of radiation is increased, and statistical noise is smoothed and reduced. However, even if such an averaging process is performed, the statistical noise is not sufficiently reduced, and a false image is generated on the tomographic image by the statistical noise amplified by logarithmic transformation. Further, although it is possible to increase the transmission data used for the average processing, the number of the transmission data is also limited.

  The present invention has been made in view of such circumstances, and a radiation imaging apparatus capable of suppressing a false image of a tomographic image by reducing statistical noise included in a hardly transmissive region such as a subject image. The purpose is to provide.

In order to achieve such an object, the present invention has the following configuration.
That is, the radiation imaging apparatus according to the present invention is (A1) a logarithmic conversion unit that obtains radiation transmission path length data by logarithmically converting a plurality of transmission data acquired by detecting radiation transmitted through the subject from the same imaging direction. (B1) a difference data creation unit that creates a difference data by calculating a difference between two radiation transmission path length data acquired by the logarithmic conversion unit, and (C) each pixel value of the difference data is preset. A noise map image creation unit that creates a noise map image by determining as statistical noise when outside the specified range, and (D) the two used when creating the difference data according to the noise map image A noise reduction processing unit that performs noise reduction processing on statistical noise pixels of each of the radiation transmission path length data is provided.

  The radiation imaging apparatus according to the present invention has each configuration for creating statistical noise included in the amplified transmission data as a noise map image and reducing the statistical noise on the transmission data using the noise noise. The logarithm conversion unit logarithmically converts the transmission data to obtain radiation transmission path length data, thereby amplifying the statistical noise included in the transmission data. The difference data creation unit calculates the difference between the two radiation transmission path length data acquired by the logarithmic conversion unit and creates the difference data, thereby removing the subject image and extracting statistical noise. Then, the noise map image creation unit creates a noise map image by determining as statistical noise when each pixel value of the difference data is outside the preset range. That is, a pixel having a large variation amount of the logarithmically converted pixel value is determined as statistical noise. In accordance with the noise map image obtained in this way, noise reduction processing is performed on each statistical noise pixel of the two radiation transmission path length data used when the difference data is created. As a result, noise reduction processing can be performed from a state in which the position where statistical noise appears in advance, so that noise reduction processing can be performed targeting the statistical noise portion included in the radiation transmission path length data. By doing in this way, the statistical noise contained in the difficult-to-transmit regions such as the subject image can be reduced more reliably than before, and the false image of the tomographic image can be suppressed.

  Further, since a plurality of transmission data are acquired from the same direction, noise reduction processing can be performed without depending on the shape of the contour of the subject M.

  Further, in the radiation imaging apparatus according to the present invention, a plurality of transmission data acquired by detecting radiation transmitted through the subject from the same imaging direction is grouped by a preset number of transmission data, and a plurality of transmission data is provided for each group. It is preferable that an average transmission data creation unit that averages transmission data to create average transmission data is provided, and the logarithmic conversion unit logarithmically converts the average transmission data. Since a plurality of transmission data is averaged for each group to create one average transmission data, statistical noise can be further suppressed.

  The radiation imaging apparatus according to the present invention preferably includes a noise reduction data averaging unit that averages two pieces of radiation transmission path length data subjected to noise reduction processing by the noise reduction processing unit. Thereby, statistical noise can be further reduced between two pieces of radiation transmission path length data. For example, conventionally, 10 transmission data are averaged, and the averaged transmission data is logarithmically converted to generate radiation transmission path length data. The same 10 pieces of transmission data are divided into two groups of 5 and 5, and the noise reduction processing of the present invention is performed, and the two groups subjected to the noise reduction processing are averaged. As a result, the present invention can be applied using the same number of transmission data as the 10 transmission data that have conventionally created one radiation transmission path length data.

  In addition, the radiation imaging apparatus according to the present invention is (A2) logarithmic conversion that obtains radiation transmission path length data by logarithmically converting a plurality of transmission data acquired by detecting radiation transmitted through the subject while changing the imaging direction. A difference data creation unit that creates difference data by subtracting two radiation transmission path length data obtained by the logarithmic conversion unit and acquired in succession, and (C) each pixel value of the difference data A noise map image creating unit that creates a noise map image by determining as statistical noise when is outside a preset range, and (D) used when creating the difference data according to the noise map image And a noise reduction processing unit that performs noise reduction processing on the statistical noise pixels of the two pieces of radiation transmission path length data obtained.

  The radiation imaging apparatus according to the present invention has each configuration for creating statistical noise included in the amplified transmission data as a noise map image and reducing the statistical noise on the transmission data using the noise noise. The logarithm conversion unit logarithmically converts the transmission data to obtain radiation transmission path length data, thereby amplifying the statistical noise included in the transmission data. The difference data creation unit calculates the difference between the two radiation transmission path length data acquired by the logarithmic conversion unit and creates the difference data, thereby removing the subject image and extracting statistical noise. Then, the noise map image creation unit creates a noise map image by determining that each pixel value of the difference data is out of a preset range as noise. That is, a pixel having a large variation amount of the logarithmically converted pixel value is determined as statistical noise. In accordance with the noise map image obtained in this way, noise reduction processing is performed on each noise pixel of the two radiation transmission path length data used when the difference data is created. As a result, noise reduction processing can be performed from a state in which the position where statistical noise appears in advance, so that noise reduction processing can be performed targeting the statistical noise portion included in the radiation transmission path length data. By doing in this way, the statistical noise contained in the difficult-to-transmit regions such as the subject image can be reduced more reliably than before, and the false image of the tomographic image can be suppressed.

  In addition, since a plurality of transmission data are acquired while changing the direction, it is possible to perform noise reduction processing suitable for a subject constituted by a curve whose contour is relatively unclear.

  Further, in the radiographic apparatus according to the present invention, a plurality of transmission data acquired by detecting radiation transmitted through the subject while changing the imaging direction is grouped into a predetermined number of transmission data continuous in time. It is preferable that an average transmission data generation unit that averages a plurality of transmission data for each group to generate average transmission data, and the logarithmic conversion processing unit logarithmically converts the average transmission data. Since a plurality of transmission data is averaged for each group to create one average transmission data, statistical noise can be further suppressed.

  Also, in an example of the radiographic apparatus according to the present invention, the range is dynamically set based on the transmission data so that the range is wider as the S / N ratio is larger, and is narrower as the S / N ratio is smaller. It is. Thereby, a range for automatically discriminating noise can be set according to the S / N ratio of the transmission data. Further, when the S / N ratio fluctuates, the noise can be discriminated within the automatically set range each time.

  According to the radiation imaging apparatus of the present invention, the logarithmic conversion unit logarithmically converts the transmission data to acquire radiation transmission path length data, thereby amplifying statistical noise included in the transmission data. The difference data creation unit calculates the difference between the two radiation transmission path length data acquired by the logarithmic conversion unit and creates the difference data, thereby removing the subject image and extracting statistical noise. Then, the noise map image creation unit creates a noise map image by determining that each pixel value of the difference data is outside the preset range as statistical noise. That is, a pixel having a large variation amount of the logarithmically converted pixel value is determined as statistical noise. In accordance with the noise map image obtained in this way, noise reduction processing is performed on each statistical noise pixel of the two radiation transmission path length data used when the difference data is created. As a result, noise reduction processing can be performed from a state in which the position where statistical noise appears in advance, so that noise reduction processing can be performed targeting the statistical noise portion included in the radiation transmission path length data. By doing in this way, the statistical noise contained in the difficult-to-transmit regions such as the subject image can be reduced more reliably than before, and the false image of the tomographic image can be suppressed.

1 is a schematic configuration diagram of an X-ray imaging apparatus according to a first embodiment. It is a figure which shows the structure of an image process part. (A) is a figure with which it uses for description of acquiring fluoroscopy data continuously along time, (b) is a figure with which it uses for description of carrying out an average process for every group. (A) is a photograph showing an example of X-ray transmission path length data, and (b) is a photograph showing an example of X-ray transmission path length data acquired next in the same direction as (a). (A) is a photograph showing the difference data in FIG. 4 (a) and FIG. 4 (b), and (b) is an enlarged photograph of the rectangular frame in (a). It is a figure where it uses for description of noise map image creation. It is a figure which shows an example of a noise map image. It is a flowchart with which it uses for operation | movement description of the X-ray imaging apparatus which concerns on 1st Embodiment. It is a top view with which it uses for operation | movement description of the X-ray imaging apparatus which concerns on 2nd Embodiment. It is a figure which shows the relationship between the magnitude | size of the range for determining noise, and S / N ratio.

<First Embodiment>
Hereinafter, a first embodiment of the present invention will be described with reference to the drawings. In this embodiment, an X-ray imaging apparatus will be described as a radiation imaging apparatus.

  1 is a schematic configuration diagram of the X-ray imaging apparatus according to the first embodiment, and FIG. 2 is a diagram illustrating a configuration of the image processing unit. FIG. 3A is a diagram for explaining that fluoroscopic data is continuously acquired over time, and FIG. 3B is a diagram for explaining that an average process is performed for each group. FIG. 4A is a photograph showing an example of X-ray transmission path length data, and FIG. 4B is an example of X-ray transmission path length data acquired next in the same direction as FIG. It is a photograph shown.

  Please refer to FIG. The X-ray imaging apparatus 1 includes an X-ray tube 3 that emits X-rays toward the subject M, and an X-ray detector 5 that detects X-rays transmitted through the subject M. A stage 7 on which the subject M is placed is provided between the X-ray tube 3 and the X-ray detector 5.

  The X-ray tube 3 is controlled by the X-ray tube control unit 9. The X-ray tube controller 9 has a high voltage generator 11 that generates the tube voltage and tube current of the X-ray tube 3. The X-ray tube control unit 9 irradiates X-rays from the X-ray tube 3 according to X-ray irradiation conditions such as tube voltage, tube current, and irradiation time. Further, a collimator (not shown) for narrowing the X-rays emitted from the X-ray tube 3 is provided on the X-ray irradiation side of the X-ray tube 3, and the X-rays are emitted in a cone beam shape. If no collimator is required, it may be omitted.

  The X-ray detector 5 is constituted by, for example, a flat panel X-ray detector (FPD). The FPD has a horizontal and vertical two-dimensional matrix shape in which a large number of X-ray detection elements that detect X-rays by converting them into electrical signals on an X-ray detection surface on which a transmission X-ray image of a detection target (subject M) is projected Is arranged. Examples of the array matrix of the X-ray detection elements include horizontal: several thousand × vertical: several thousand. The X-ray detection element is configured as a direct conversion type in which X-rays are directly converted into electric signals, or an indirect conversion type in which X-rays are once converted into light and then converted into electric signals. Further, the X-ray detector 5 is not limited to the FPD, and may be configured with an image intensifier and a camera.

  The stage 7 is configured to be moved about the three axes in the X, Y, and Z directions by the stage driving mechanism 13 and rotated about the axis z. Further, the X-ray tube 3 and the X-ray detector 5 are configured to be attached to both ends of a C-shaped arm (not shown), for example, and in FIG. However, the X-ray tube 3 and the X-ray detector 5 may be arranged so that the central axis Q is inclined with respect to the horizontal.

  An A / D converter 15, an image processing unit 17, and a main control unit 19 are provided in order after the X-ray detector 5. The A / D converter 15 converts an analog X-ray detection signal, which is a kind of transmission data Ga output from the X-ray detector 5, into a digital X-ray detection signal and outputs it as transmission data Ga. The image processing unit 17 performs necessary processing such as a tone reduction process and the like on the digitally converted transmission data Ga, in addition to a noise reduction process described later. The main control unit 19 comprehensively controls each component of the X-ray imaging apparatus 1 and includes a central processing unit (CPU) and the like. For example, the main control unit 19 controls the stage driving mechanism 13 to rotate the stage 7 around the axis z.

  The X-ray imaging apparatus 1 includes a display unit 21, an input unit 23, and a storage unit 25. The display unit 21 includes a monitor or the like. The input unit 23 includes a keyboard, a mouse, and the like. The storage unit 25 includes a removable storage medium such as a ROM (Read-only Memory), a RAM (Random-Access Memory), or a hard disk. The X-ray imaging apparatus 1 includes an image reconstruction unit 27 that reconstructs an image by, for example, the FBP method and creates a tomographic image.

  Next, the detailed configuration of the image processing unit 17 will be described with reference to FIG. The image processing unit 17 performs noise reduction processing that is a characteristic part of the present invention.

  An outline of noise reduction processing performed by the image processing unit 17 will be described. The average transmission data creation unit 31 creates the average transmission data Gb by averaging the acquired transmission data Ga for each group. The logarithmic conversion unit 33 logarithmically converts the average transmission data Gb to obtain X-ray transmission path length data Gc. The difference data creation unit 35 calculates the difference between the two X-ray transmission path length data Gc acquired by the logarithmic conversion unit 33 and creates difference data Gd. The noise map image creation unit 37 creates a noise map image Nm indicating a statistical noise distribution based on the difference data Gd. The noise reduction processing unit 39 performs noise reduction processing on each noise pixel of the two X-ray transmission path length data Gc used when the difference data Gd is created according to the noise map image Nm. Thereby, noise reduction data Ge is acquired. The noise reduction data averaging unit 41 averages two noise reduction data Ge to create average noise reduction data Gf.

  In the present embodiment, a plurality of transmission data Ga (X-ray detection signals) acquired by the X-ray detector 5 are taken from the same direction. The transmission data Ga is continuously photographed along the time. For example, as shown in FIG. 3A, the transmission data Ga is in the order of time t1, t2, t3,..., T (n-1), t (n). Taken.

  The average transmission data creation unit 31 groups the acquired plurality of transmission data Ga into a predetermined number of transmission data Ga, and averages the plurality of transmission data Ga for each group to create one average transmission data Gb. To do. For example, as shown in FIG. 3B, average transmission data Gb1, which is one transmission data, is created by averaging five transmission data Ga from time t1 to time t5, and from time t6 to time t6. Average transmission data Gb2 is generated by averaging the five transmission data Ga at time t10.

  The reason for grouping in this embodiment includes the following cases, for example. That is, there are cases where 10 pieces of transmission data Ga are acquired and one image is formed with 10 pieces of transmission data Ga as in the prior art. In this case, if the average transmission data Gb is created with 10 pieces of transmission data Ga, the difference data Gd cannot be created from one average transmission data Gb. By dividing the acquired 10 transmission data Ga into 5 and 5 groups, two average transmission data Gb1 and Gb2 for obtaining the difference data Gd can be acquired. Further, by dividing into two equal parts, the two average transmission data Gb1 and Gb2 can be handled equally.

  It is assumed that 20 pieces of transmission data Ga are acquired and two images are formed with 10 pieces of transmission data Ga as in the prior art. In this case, the acquired 20 pieces of transmission data Ga are divided into 10 groups and 10 groups. Thereby, two average transmission data Gb1 and Gb2 for obtaining the difference data Gd may be acquired.

  In the average transmission data Gb1, Gb2, X-ray transmission path length data Gc1, Gc2, which will be described later, and noise reduction data Ge1, Ge2, etc., unless otherwise specified, the average transmission data Gb, X-ray transmission path length data Gc and noise reduction Data Ge and the like are shown.

  The logarithmic conversion unit 33 obtains X-ray transmission path length data Gc (also referred to as projection data) by logarithmically converting the average transmission data Gb. That is, the average transmission data Gb1 is logarithmically converted to obtain the X-ray transmission path length data Gc1 and the average transmission data Gb2 is logarithmically converted to obtain the X-ray transmission path length data Gc2 by the above-described equation (1). The X-ray transmission path length data Gc is data indicating a product of a path length (path length) and a linear absorption coefficient when X-rays pass through the subject M. During the logarithmic conversion, statistical noise included in each average transmission data Gb1 and Gb2 is amplified. The X-ray transmission path length data Gc corresponds to the radiation transmission path length data of the present invention. The two X-ray transmission path length data Gc1 and Gc2 are in a state in which the imaging order is continuous.

  The difference data creation unit 35 calculates the difference between the two X-ray transmission path length data Gc1 and Gc2 acquired by the logarithmic conversion unit 33 and creates difference data Gd. That is, the difference data Gd shown in FIG. 5A is calculated by calculating the difference between the X-ray transmission path length data Gc1 shown in FIG. 4A and the X-ray transmission path length data Gc2 shown in FIG. create. In each of the X-ray transmission path length data Gc1 and Gc2 shown in FIG. 4A and FIG. 4B, the statistical noise is difficult to recognize in the difficult transmission region (white portion) where the subject M is thick. However, in the difference data Gd shown in FIG. 5A, statistical noise is extracted in which the reflected image of the subject M is removed by the difference. That is, the object image or the like is removed by making a difference, whereas the statistical noise fluctuates with time, so it appears by making a difference. FIG. 5B shows an enlarged photograph of the rectangular frame in FIG.

  The noise map image creation unit 37 creates a noise map image Nm based on the difference data Gd. The noise map image Nm shows the distribution of statistical noise reflected in the two X-ray transmission path length data Gc1 and Gc2. FIG. 6 is a diagram for explaining the creation of the noise map image Nm. The noise map image creation unit 37 creates a noise map image Nm by determining that each pixel value of the difference data Gd is outside the range indicated by the oblique line in FIG. 6 of the preset range R as statistical noise. The range R for determining statistical noise includes an upper threshold Ra and a lower threshold Rb.

  That is, the noise map image creation unit 37 performs threshold processing using the upper threshold Ra, and determines pixel values in an area equal to or greater than (or larger than) the threshold Ra as statistical noise. In addition, threshold processing is performed using the lower threshold value Rb, and pixel values in an area below (or smaller than) the threshold value Rb are determined as statistical noise. The noise map image creation unit 37 creates a noise map image Nm indicating the position of the pixel determined as statistical noise. The noise map image Nm is, for example, a noise pixel represented by “1” and the other pixels represented by “0”. FIG. 7 shows an example of the noise map image Nm. In FIG. 7, the black dot N indicates the statistical noise reflected in the two X-ray transmission path length data Gc1 and Gc2. The reason why the differential data Gd is calculated after logarithmic conversion and statistical noise is extracted is that statistical noise is amplified by logarithmic conversion and appears prominently. Then, the noise map image creation unit 37 determines a pixel having a large variation amount of the logarithmically converted pixel value as statistical noise.

  The noise reduction processing unit 39 performs noise reduction processing on the noise pixels of the two X-ray transmission path length data Gc1 and Gc2 used when the difference data Gd is created according to the noise map image Nm shown in FIG. . That is, the noise reduction processing unit 39 reduces sesame salt-like statistical noise that appears in each of the X-ray transmission path length data Gc1 and Gc2. The noise reduction processing unit 39 performs noise reduction processing on the pixels of the X-ray transmission path length data Gc1 and Gc2 corresponding to the noise pixel position of the noise map image Nm. As a specific noise reduction process, for example, a median filter is used. For example, the median filter arranges the pixel values of 8 surrounding pixels (8 neighboring pixels) in the pixel j of the X-ray transmission path length data Gc1 corresponding to the statistical noise in order of size. At this time, the pixel value in the middle order is selected and set as the pixel value of the pixel j.

  In addition, although the median filter was mentioned as an example as a noise reduction process, it is not limited to this. For example, a known method such as a Gaussian filter, a bilateral filter, or a mean filter may be used.

  The noise reduction processing unit 39 outputs two noise reduction data Ge1 and Ge2 as X-ray transmission path length data subjected to noise reduction processing, and transmits the data to the noise reduction data averaging unit 41. The noise reduction data averaging unit 41 averages the noise reduction data Ge1 and Ge2 to create average noise reduction data Gf. Thereby, the statistical noise between the two noise reduction data Ge1 and Ge2 is reduced. The noise reduction data Ge1 and Ge2 are acquired based on the five pieces of transmission data Ga. Therefore, the noise reduction data averaging unit 41 averages them, and the average noise reduction data Gf is obtained based on 10 pieces of transmission data.

  The image processing unit 17 performs necessary image processing such as gradation processing in addition to the noise reduction processing.

  Next, the operation of the X-ray imaging apparatus 1 will be described along the flowchart of FIG. First, the positions of the X-ray tube 3, the X-ray detector 5, and the stage 7 are fixed. In this fixed state, X-rays are irradiated from the X-ray tube 3 toward the subject M. X-rays that have passed through the subject M enter the X-ray detector 5, and the X-ray detector 5 outputs an X-ray detection signal that is a type of transmission data Ga corresponding to the incident X-ray intensity. The A / D converter 15 converts an analog X-ray detection signal into a digital X-ray detection signal and outputs it as transmission data Ga. The transmission data Ga is stored in a storage unit (not shown). In addition, the X-ray tube 3 and the X-ray detector 5 are continuously fixed, for example, X-ray imaging is performed ten times in total, and ten pieces of transmission data Ga are acquired and stored.

[Step S01] Creation of Average Transmission Data The average transmission data creation unit 31 groups the acquired 10 transmission data Ga into 5 transmission data Ga, and averages the 5 transmission data Ga for each group. Average transmission data Gb1 and Gb2 are created (see FIGS. 3A and 3B).

[Step S02] Logarithmic Conversion The logarithmic conversion unit 33 logarithmically converts the average transmission data Gb1 and Gb2 by the above-described equation (1) to obtain X-ray transmission path length data Gc1 and Gc2 (FIG. 4A and FIG. 4 (b)).

[Step S03] Creation of Difference Data The difference data creation unit 35 calculates difference between the two X-ray transmission path length data Gc1 and Gc2 acquired by the logarithmic conversion unit 33 and creates difference data Gd (see FIG. 5). .

[Step S04] Creation of Noise Map Image The noise map image creation unit 37 creates a noise map image Nm by determining that each pixel value of the difference data Gd is outside the preset range R as noise. (See FIGS. 6 and 7).

[Step S05] Noise Reduction Processing The noise reduction processing unit 39 uses the statistical noise pixels of the two X-ray transmission path length data Gc1 and Gc2 used when creating the difference data Gd according to the noise map image Nm shown in FIG. Noise reduction processing is performed on That is, noise reduction processing by the median filter is performed on pixels corresponding to all statistical noise pixels in the two X-ray transmission path length data Gc1 and Gc2. The noise reduction processing unit 39 outputs two pieces of noise reduction data Ge1 and Ge2 as X-ray transmission path length data subjected to noise reduction processing.

[Step S06] Creation of Average Noise Reduction Data The noise reduction data averaging unit 41 averages the two noise reduction data Ge1 and Ge2 to create average noise reduction data Gf.

  The average noise reduction data Gf created as described above is displayed on the display unit 21 or stored in the storage unit 25 after other necessary image processing is performed.

  Further, the average noise reduction data Gf may be generated again by slightly rotating the stage 7 around the axis z to change the shooting direction. Then, the tomographic image S may be created by reconstructing a plurality of average noise reduction data Gf photographed from different directions by the image processing unit 17.

  According to the present embodiment, the logarithmic conversion unit 33 logarithmically converts the transmission data Ga to acquire X-ray transmission path length data, thereby amplifying the statistical noise included in the transmission data Ga. The difference data creation unit 35 calculates the difference between the two X-ray transmission path length data Gc acquired by the logarithmic conversion unit 33 and creates difference data Gd, thereby removing the subject image and extracting statistical noise. Then, the noise map image creation unit 37 creates a noise map image Nm by determining as noise when each pixel value of the difference data Gd is outside the range R set in advance. A pixel having a large variation amount of the logarithmically converted pixel value is determined as statistical noise. In accordance with the noise map image Nm obtained in this way, noise reduction processing is performed on each noise pixel of the two X-ray transmission path length data Gc used when the difference data Gd is created. Thereby, noise reduction processing can be performed from a state in which the position where statistical noise appears in advance, so that noise reduction processing can be performed targeting the statistical noise portion included in the X-ray transmission path length data Gc. By doing in this way, the statistical noise contained in the difficult-to-transmit regions such as the subject image can be reduced more reliably than before, and the false image of the tomographic image can be suppressed.

  Further, since a plurality of transmission data are acquired from the same direction, noise reduction processing can be performed without depending on the shape of the contour of the subject M.

  In addition, the X-ray imaging apparatus 1 groups a plurality of transmission data Ga by a predetermined number of transmission data Ga, and averages the plurality of transmission data Ga for each group to create one average transmission data Gb. An average transmission data creation unit 31 is provided, and a logarithmic conversion unit 33 logarithmically converts the average transmission data Gb. Since a plurality of transmission data Ga is averaged for each group to create one average transmission data Gb, statistical noise can be further suppressed. That is, by using a plurality of transmission data Ga, the amount of X-ray detection is increased, and statistical noise is smoothed and reduced.

  Further, the X-ray imaging apparatus 1 averages the two X-ray transmission path length data Gc (noise reduction data Ge) subjected to noise reduction processing by the noise reduction processing unit 39 to generate one average noise reduction data Gf. A reduced data averaging unit 41 is provided. Thereby, statistical noise can be further reduced between two pieces of radiation transmission path length data.

  For example, conventionally, ten pieces of transmission data are averaged, and the averaged transmission data is logarithmically converted to generate X-ray transmission path length data. The same 10 pieces of transmission data are divided into two groups of 5 and 5, and the noise reduction processing of the present invention is performed, and the two groups subjected to the noise reduction processing are averaged. As a result, the present invention can be applied using the same number of transmission data Ga as the 10 transmission data Ga for which one X-ray transmission path length data has conventionally been created.

<Second Embodiment>
Next, a second embodiment of the present invention will be described with reference to the drawings. In addition, the description which overlaps with 1st Embodiment is abbreviate | omitted. In the first embodiment, the shooting direction is fixed, and transmission data is acquired from the same shooting direction. In this regard, in the second embodiment, the transmission data is acquired while changing the shooting direction without fixing the shooting direction. Moreover, in 2nd Embodiment, the noise reduction data average part 41 shown in FIG. 2 is not provided.

  That is, in the second embodiment, X-ray imaging is performed while the stage 7 is rotated about the axis z by the stage driving mechanism 13. FIG. 9 is a plan view for explaining the operation of the X-ray imaging apparatus 1 when imaging is performed while the X-ray tube 3 and the X-ray detector 5 are relatively rotated with the stage 7. As shown in FIG. 9, after imaging at time t1, X-ray imaging is performed in the imaging direction at time t2 where the stage 7 is slightly rotated (not shown). Thereafter, while the X-ray tube 3 and the X-ray detector 5 and the stage 7 are relatively rotated, for example, images are taken in the order of time t3, t4, t5,. Different consecutive fluoroscopic data are acquired (time t3 to t5 are not shown). In the present embodiment, 40 pieces of transmission data are acquired and described. However, actually, the periphery of the subject M is divided into, for example, 1000, and 1000 pieces of transmission data are acquired.

  The X-ray detector 5 acquires a plurality of X-ray detection signals which are a kind of transmission data Ga while changing the imaging direction. The X-ray detection signal digitally converted by the A / D converter 15 is output as transmission data Ga.

  The average transmission data creation unit 31 groups the acquired plurality of transmission data Ga into a predetermined number of transmission data Ga whose imaging order is continuous, and averages the plurality of transmission data Ga for each group to obtain the average transmission data. Gb is created. That is, the average transmission data creation unit 31 averages 10 pieces of transmission data Ga X-rayed at time t1 to time t10 to create one average transmission data Gb1. Further, the average transmission data Gb2 is created by averaging the 10 transmission data Ga obtained by X-ray imaging at the time t11 to the time t20, in which the imaging order is continuous with the time t1 to the time t10.

  Similarly, the average transmission data creation unit 31 averages 10 pieces of transmission data Ga X-rayed at time t21 to time t30 to create one average transmission data Gb3. In addition, ten pieces of transmission data Ga obtained by X-ray imaging at time t31 to time t40 are averaged to create one average transmission data Gb4.

  The four average transmission data Gb1 to Gb4 are logarithmically converted by the above-described equation (1), respectively, as in the first embodiment, to obtain X-ray transmission path length data Gc1 to Gc4. The difference data creation unit 35 creates difference data Gd1 by subtracting two pieces of X-ray transmission path length data Gc1 and Gc2 in which the imaging order is continuous. Further, difference data Gd2 is generated by subtracting two pieces of X-ray transmission path length data Gc3 and Gc4 in which the imaging order is continuous. When calculating the difference, it is desirable to use two X-ray transmission path length data in which the imaging order is continuous. If the two X-ray transmission path length data Gc are acquired based on transmission data Ga whose imaging directions are greatly different, a part of the image of the subject M remains when the difference is calculated. Thereby, when creating a noise map image, there is a possibility that a part of the remaining image of the subject M is erroneously determined as noise. When this influence is small, the imaging order of the two different X-ray transmission path length data Gc1 and Gc2 may not be continuous.

  The noise map image creation unit 37 determines that the pixel values of the difference data Gd1 and Gd2 are outside the preset range R as statistical noise, and creates noise map images Nm1 and Nm2, respectively. The noise reduction processing unit 39 performs noise reduction processing on the noise pixels of the two X-ray transmission path length data Gc1 and Gc2 used when the difference data Gd1 is created according to the noise map image Nm1, thereby reducing the noise reduction data. Ge1 is output. Further, noise reduction processing is performed on the noise pixels of the two X-ray transmission path length data Gc3 and Gc4 used when the difference data Gd2 is created according to the noise map image Nm2, and noise reduction data Ge2 is output.

  The noise reduction data Ge1 and Ge2 output from the noise reduction processing unit 39 (image processing unit 17) are temporarily stored in the storage unit 25 and transmitted to the image reconstruction unit 27. The image reconstruction unit 27 creates a tomographic image S by performing image reconstruction based on the noise reduction data Ge1 and Ge2 and other noise reduction data in the imaging direction.

  According to the present embodiment, there are substantially the same effects as in the first embodiment. In addition, since a plurality of transmission data Ga are acquired while changing the direction, it is possible to perform noise reduction processing suitable for the subject M configured by a curve in which the contour of the subject M is relatively unclear.

<Third Embodiment>
Next, a third embodiment of the present invention will be described with reference to the drawings. In addition, the description which overlaps with 1st and 2nd embodiment is abbreviate | omitted. In the present embodiment, the range R for determining statistical noise is made variable according to the S / N ratio.

  That is, in the first and second embodiments described above, when the noise map image is created, the pixel values of the difference data Gd are determined as noise when they are outside the preset range R. . That is, fixed threshold values Ra and Rb are provided on the upper side and the lower side, and if the threshold value Ra is equal to or higher than (or larger than) the upper threshold value Ra or equal to or lower than (or smaller than) the lower threshold value Rb, it is determined as noise. It was. However, the range R formed by the upper and lower thresholds Ra and Rb may be automatically set according to the features such as the S / N ratio of the original transmission data. FIG. 10 is a diagram showing the relationship between the size of the range R for determining noise and the S / N ratio.

  For example, any one transmission data of the average transmission data Gb and the transmission data Ga used as the basis for creating the difference data Gd is used. First, the S / N ratio of the transmission data is obtained. As shown in FIG. 10, the range R formed by the upper and lower thresholds Ra and Rb is wider as the S / N ratio is larger based on the transmission data. It is dynamically set so as to become narrower as the S / N ratio is smaller. Specifically, for example, a fixed center value T (see FIG. 6) of the range R formed by the upper and lower thresholds Ra and Rb is set in advance, and the center is determined according to the S / N ratio. The range R may be varied based on the value T.

  According to the present embodiment, the range R is dynamically set so as to be wider as the S / N ratio is larger and narrower as the S / N ratio is smaller based on the transmission data. Thereby, the range R for automatically discriminating noise can be set according to the S / N ratio of the transmission data Ga. Further, when the S / N ratio varies, the noise can be discriminated in the automatically set range R each time.

  The present invention is not limited to the above embodiment, and can be modified as follows.

  (1) In each embodiment described above, the X-ray tube 3 and the X-ray detector 5 are fixed, and the imaging direction is changed by rotating the stage 7 around the axis z. However, the imaging direction may be changed by fixing the stage 7 and rotating the X-ray tube 3 and the X-ray detector 5 around the axis z. Further, the imaging direction may be changed by fixing the stage 7 and moving the X-ray tube 3 and the X-ray detector 5 in parallel in the opposite directions.

  (2) In each of the above-described embodiments and modification (1), the X-ray tube 3 and the X-ray detector 5 are arranged in the left-right direction with the stage 7 interposed therebetween as shown in FIG. It may be arranged in the up-and-down direction across.

  (3) In each of the above-described embodiments and modifications, the X-ray detector 5 detects the X-ray and acquires the transmission data Ga. However, the transmission data Ga is, for example, a γ-ray detector. You may make it detect and acquire.

  (4) In each embodiment and each modification described above, as illustrated in FIG. 2, the image processing unit 17 includes an average transmission data creation unit 31 and a noise reduction data averaging unit 41 (first embodiment). Moreover, although it was the structure (2nd Embodiment) provided with the average transmission data preparation part 31, it is not limited to this. The configuration may be such that at least one of the average transmission data creation unit 31 and the noise reduction data averaging unit 41 is excluded.

  (5) In each embodiment and each modification described above, the noise reduction process is performed by the image processing unit 17, but the present invention is not limited to this. For example, the transmission data Ga may be temporarily stored in the storage unit 25 and the main control unit 19 may perform noise reduction processing. Further, noise reduction processing may be performed by a personal computer (not shown) connected to the main control unit 19.

  (6) The X-ray imaging apparatus 1 of the first embodiment and each modification described above may be a tomography apparatus that acquires a tomographic image S like a CT apparatus, or may be noise reduction data Ge or average noise. It may be an imaging device that acquires the reduced data Gf.

DESCRIPTION OF SYMBOLS 1 ... X-ray imaging apparatus 3 ... X-ray tube 5 ... X-ray detector 17 ... Image processing part 19 ... Main control part 27 ... Image reconstruction part 31 ... Average transmission data preparation part 33 ... Logarithmic conversion part 35 ... Difference data preparation Section 37 ... Noise map image creation section 39 ... Noise reduction processing section 41 ... Noise reduction data averaging section Ga ... Transmission data Gb ... Average transmission data Gc ... X-ray transmission path length data Gd ... Difference data Ge ... Noise reduction data Gf ... Average Noise reduction data S ... Tomographic image Nm ... Noise map image R ... Range

Claims (6)

(A1) a logarithmic conversion unit for logarithmically converting a plurality of transmission data acquired by detecting radiation transmitted through the subject from the same imaging direction to acquire radiation transmission path length data;
(B1) a difference data creation unit that creates a difference data by calculating a difference between two radiation transmission path length data acquired by the logarithmic conversion unit;
(C) a noise map image creation unit that creates a noise map image by determining as statistical noise when each pixel value of the difference data is outside a preset range;
(D) a noise reduction processing unit that performs noise reduction processing on the statistical noise pixels of each of the two radiation transmission path length data used when the difference data is created according to the noise map image A radiographic apparatus characterized by. The radiographic apparatus according to claim 1,
Multiple transmission data acquired by detecting radiation transmitted through the subject from the same imaging direction are grouped by a preset number of transmission data, and average transmission data is created by averaging multiple transmission data for each group. An average transmission data creation unit
A logarithmic conversion unit logarithmically converts the average transmission data. The radiographic apparatus according to claim 1 or 2,
A radiation imaging apparatus comprising: a noise reduction data averaging unit that averages two pieces of radiation transmission path length data subjected to noise reduction processing by the noise reduction processing unit. (A2) a logarithmic conversion unit for logarithmically converting a plurality of transmission data acquired by detecting radiation transmitted through the subject while changing the imaging direction to acquire radiation transmission path length data;
(B2) a difference data creation unit that creates difference data by subtracting two pieces of radiation transmission path length data in which the imaging order acquired by the logarithmic conversion unit is continuous;
(C) a noise map image creation unit that creates a noise map image by determining as statistical noise when each pixel value of the difference data is outside a preset range;
(D) A noise reduction processing unit that performs noise reduction processing on statistical noise pixels of the two radiation transmission path length data used when the difference data is created according to the noise map image. A characteristic radiographic apparatus. The radiation imaging apparatus according to claim 4,
Multiple transmission data acquired by detecting radiation that has passed through the subject while changing the imaging direction are grouped by a preset number of transmission data that are temporally continuous, and the multiple transmission data is averaged for each group. Equipped with an average transmission data creation unit for creating average transmission data
A logarithmic conversion processing unit logarithmically converts the average transmission data. In the radiography apparatus in any one of Claim 1 to 5,
The radiographic apparatus is characterized in that the range is dynamically set so as to be wider as the S / N ratio is larger and narrower as the S / N ratio is smaller based on the transmission data.