JPH06181918A - Transmission type three-dimenisonal tomograph - Google Patents

Abstract

PURPOSE:To shorten the time for measurement and data processing by adapting the inverse projection method to the projection image of the transmission factor distribution of the radioactive rays transmitting an object. CONSTITUTION:In the wide-angle data analysis, the maximum visual field angle theta1 is set for the whole object 10 by collimators 16, 21, the object 10 on a sample base 11 is rotated by the preset rotation angle phi, and the projection data of two-dimensional static images photographed by a camera 20 for each rotation angle phi are analyzed and processed. In the narrow-angle data analysis, an X-ray cone beam is radiated to only the slice cross section of the object 10 for each preset thickness DELTAh, and the narrow visual field angle theta2 capable of sufficiently ignoring the low-frequency distortion by the cross talk of the tomographic image is set. The object 10 is rotated by the preset rotation angle phieach time the object 10 is moved along the coordinate axis (z) by the preset pitch interval DELTAk, and the projection data of the static images photographed by the camera 20 at each rotation angle phi are analyzed and processed. The data of the high-frequency constituent and the low-frequency constituent of the two-dimensional tomographic image of each slice fault of the object 10 are finally synthesized to obtain the three-dimensional tomographic image.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention applies a back projection method to a three-dimensional slice of a subject by irradiating the subject with radiation from a radiation source and applying a back projection method to a projected image of the transmittance distribution of the radiation transmitted through the subject. The present invention relates to a transmission-type three-dimensional tomography apparatus for reconstructing an image, and particularly to a transmission-type three-dimensional tomography apparatus in which measurement time and data processing time are significantly shortened and processing accuracy is improved.

[0002]

2. Description of the Related Art As the most typical conventional transmission type two-dimensional tomography apparatus, an X-ray tomography apparatus (hereinafter referred to as X-ray CT apparatus) which is applied to the field of clinical medicine is known. The basic principle of such an X-ray CT apparatus will be described with reference to FIG. 14. A linear or arc-shaped detector 2 and an X-ray tube 3 are arranged to face each other so as to sandwich a subject 1 such as a patient. The detector 2 detects the intensity of the transmitted X-ray that has been emitted and transmitted through the subject 1. That is, an imaginary cross section in the xy orthogonal coordinates including the X-ray emitting portion (focus) of the X-ray tube 3 and the detecting portion provided along the longitudinal direction s of the detector 2 at the same time is represented by AR and xy coordinates. Assuming that the z-axis is an axial direction that is orthogonal to and that passes through the center of the subject 1, the intensity of the X-ray measured by the detection unit of the detector 2 is transmitted through each portion inside the subject 1 along the virtual cross section AR. It corresponds to the X-ray transmittance of transmitted X-rays. An appropriate logarithmic conversion is applied to the measurement data group of the X-ray transmittance to convert it into the sum (integral value) of the X-ray absorption coefficients of the subject along the X-ray beam. The measurement data regarding the integral value of the X-ray absorption coefficient is called projection data. Furthermore, the detector 2 and the X-ray tube 3
While keeping the facing relationship constant, the projection data group in each rotation direction φ is measured by rotating one of the devices 2 and 3 or the subject 1 relatively around the z coordinate axis. It is designed to measure projection data in azimuth. Then, a two-dimensional tomographic image regarding the X-ray absorption coefficient of the subject 1 is reconstructed by performing a predetermined computer calculation process on the projection data group regarding all the azimuths.

Here, in order to perform such reconstruction, a superposition integral backprojection method or a filter backprojection method is generally applied. That is, as shown in FIG. 15, the distribution of the projection data when the X-ray is transmitted from the azimuth φ1 with respect to the xy coordinate axis in the virtual cross section AR at a certain position on the z coordinate axis is P _AR (s, φ1). ), First, as shown in the following equation (1), a filter function g called a reconstruction filter g
₀ (s) is superimposed and integrated on P _AR (s, φ1), and new projection data Q _AR (s, φ1) obtained by this is obtained.

[0004]

[Equation 1]

Incidentally, * in the above equation (1) means a superposition integral operator. As the filter function g ₀ (s), a Shepp-Logan filter is generally applied. Then, new projection data Q _AR (s, φ) is generated for each of the remaining azimuths φ2, φ3, ... φn in the virtual section AR.
2), Q _AR (s, φ3) ... Q _AR (s, φn) are obtained in the same manner, and these data are back-projected along the direction of the X-ray beam to obtain a two-dimensional tomographic image at the virtual cross section AR. Constitute.

Although the above is to obtain a tomographic image for one virtual cross section AR, in order to obtain the three-dimensional distribution of the subject 1 using this apparatus, an apparatus (detector 2 and X-ray) is used. By moving the tube 3) or the bed supporting the subject 1 relatively, a two-dimensional tomographic image is obtained for the multiple layers.

[0007] Further, the helicopter scan method, that is, the apparatus (detector 2 and X-ray tube 3) is moved around the subject 1 along a spiral orbit, or conversely,
By rotating the subject 1 along the z-axis while rotating it in the fixed device (the detector 2 and the X-ray tube 3), the device is controlled to move substantially along the spiral orbit. A two-dimensional tomographic image is obtained for multiple layers of the subject 1.

However, according to this conventional technique, if a high-definition three-dimensional tomographic image is to be reconstructed, the detector 2 and the X-ray tube 3 or the subject 1 are finely spaced along the z coordinate axis.
Since it is necessary to measure the projection data P (s, φ) while moving relative to, there is a problem that the measurement takes a long time.

Therefore, a further improved and advanced X-ray CT apparatus has been proposed. Such an X-ray CT apparatus is called a circular orbit cone beam CT, and as shown in FIG. 16, a two-dimensional image detector 5 such as an X-ray image intensifier (X-ray II) and a cone are arranged so as to sandwich the subject 4. And an X-ray tube 6 that radiates X-rays in a circular shape, and is configured so that the subject 4 is always present within this cone-shaped X-ray irradiation range. That is, the two-dimensional image detector 5 two-dimensionally detects the transmittance of the transmitted X-rays that have been emitted from the X-ray tube 6 in a cone shape and transmitted through the subject 4. Then, the X-ray tube 6 and the two-dimensional image detector 5 are relatively rotated with respect to the subject 4 along a circular orbit centered on the z coordinate axis passing through the center of the subject 4, thereby obtaining the two-dimensional projection data in all directions. A three-dimensional tomographic image of the subject 4 is reconstructed by measuring and further applying the above-mentioned backprojection method to the three-dimensional computer processing.

The Feldkamp method is typical as a three-dimensional image reconstruction method of such a circular orbit cone beam CT, and the two-dimensional projection data P (s, h, φ) is corrected by multiplying it by a two-dimensional load, the same superposition integral as in the above equation (1) is performed in the direction s (direction orthogonal to the z coordinate axis), and new projection data Q ( s, h, φ) is three-dimensionally back-projected along the direction opposite to the X-ray radiation direction.

According to this method, if the maximum viewing angle θ of the cone-shaped X-rays (hereinafter referred to as X-ray cone beam) in the z-coordinate axis direction is relatively small, the X-ray CT apparatus shown in FIG. By approximating and extending the two-dimensional tomographic image by 3D reconstruction, the reconstruction of the three-dimensional tomographic image can be realized relatively easily.

[0012]

However, the above circular orbit cone beam CT has the following problems. That is, when the maximum viewing angle θ of the X-ray cone beam in the z-coordinate axis direction is large, the above-mentioned approximation accuracy is lowered and a fine tomographic image cannot be reconstructed. In particular, when the maximum viewing angle θ is set large, the angle between the transmission direction of X-rays transmitted through the virtual cross section AR near both ends of the z coordinate axis and the virtual cross section AR becomes large, and therefore the projection data P ( s, h, φ) is not obtained based on only the transmitted X-rays transmitted through one virtual cross section AR, but is obtained based on the transmitted X-rays transmitted through a plurality of virtual cross sections obliquely. Become. As a result, there has been a problem that interference (crosstalk) occurs in each tomographic image and the reconstructed image is distorted.

In general terms, circular orbit cone beam CT
It is known that an accurate three-dimensional tomographic image cannot be reconstructed when the apex of the X-ray cone beam (focal point of the X-ray tube) rotates along one circular orbit as shown in FIG. . For example, Smith's literature (Smith BD: Image reconstr
auction from cone-beam projections: Necessary and s
ufficient conditions and reconstruction methods. I
According to EEE Trans.Med.Image. MI-4: 14-25, 1985.)
"The sufficient condition for determining a three-dimensional tomographic image is that when the arbitrary plane that passes through all points in the subject is considered, the vertex of the cone beam must exist on that plane." This is a condition for. Therefore, as shown in FIG. 16, accurate one-dimensional reconstruction of a three-dimensional tomographic image cannot be realized only by rotating the two-dimensional image detector 5 and the X-ray tube 6 once along one circular orbit. .

Therefore, in order to solve this problem, FIG.
7 (a) or 7 (b), a method of measuring two-dimensional projection data in each direction while rotating the two-dimensional image detector 5 and the X-ray tube 6 along two circular orbits (see FIG. 1).
7 shows the orbit of the focal point of X storage), while moving the two-dimensional image detector 5 and the X-ray tube 6 along one circular orbit and the z coordinate axis as shown in FIG. Method for measuring the two-dimensional projection data, or the two-dimensional projection data in each direction while moving the two-dimensional image detector 5 and the X-ray tube 6 along a spiral trajectory as shown in FIG. A method of measurement is proposed, and based on the two-dimensional projection data obtained by these measurement methods, for example, by applying a three-dimensional Radon transform, it will be possible to reconstruct a three-dimensional tomographic image. However, there are many problems in putting it to practical use because it requires complicated and huge calculation processing.

In particular, as shown in FIG. 18, when a part of the subject 1 protrudes from the area of the X-ray cone beam, X-rays are emitted depending on the scanning positions of the detector 5 and the X-ray tube 6 with respect to the subject 1. Since a part that is only partially illuminated (hatched part U in the figure) occurs and the projection data of such part U causes crosstalk between images, it is extremely difficult to realize accurate image reconstruction of the measurement part. Met. Such a case usually occurs when measuring a tomographic image of a human body and is not a special case. Therefore, a practical solution has been desired.

The present invention has been made in view of such conventional problems, and effectively expands a relatively simple projection data measuring method and a two-dimensional tomographic image reconstruction algorithm,
An object of the present invention is to provide a transmission type three-dimensional tomography apparatus capable of realizing high-definition three-dimensional tomographic image reconstruction with little image distortion and crosstalk.

[0017]

In order to achieve such an object, the present invention relates to a radiation source that radiates radiation to a subject in a cone shape at an arbitrary viewing angle, and a transmission transmitted through the subject. Image detecting means for generating two-dimensional projection data by detecting radiation two-dimensionally, and one of the radiation and the image detecting means or the object is arranged along a spiral orbit or a plurality of circular orbits. Driving means capable of relatively moving in a direction orthogonal to the orbital plane while rotating relative to each other, and the driving means relative to each other in a state where the radiation source emits cone-shaped radiation having a wide viewing angle including the entire subject. When the rotational drive is performed, the first two-dimensional projection data detected by the image detecting means is input for each predetermined rotation angle, and a predetermined reconstruction filter is superposed and integrated on the first two-dimensional projection data. Two new Based on the original projection data, a backprojection image corresponding to a plurality of preset slice sections is formed, and a two-dimensional tomographic image corresponding to each slice section is formed based on a predetermined high frequency component of the backprojection image. When the drive means performs the relative rotation drive and the relative movement in a state in which the wide-angle data analysis means to be reconstructed and the radiation source irradiate a part of the subject, at every predetermined rotation angle. The second two-dimensional projection data detected by the image detecting means is input, and preset based on new two-dimensional projection data obtained by superimposing and integrating a predetermined reconstruction filter on the second two-dimensional projection data. A backprojection image corresponding to a plurality of slice sections is formed, and a two-dimensional tomographic image corresponding to each slice section is reconstructed based on a predetermined low-frequency component of the backprojection image. Degree data analyzing means, and configuration processing means for adding and synthesizing the two-dimensional tomographic images reconstructed by the wide-angle data analyzing means and the narrow-angle data analyzing means to form a three-dimensional tomographic image. .

Further, a radiation source for radiating a cone-shaped radiation having a wide viewing angle and covering the entire subject,
An image detecting unit that generates two-dimensional projection data by two-dimensionally detecting the transmitted radiation that has passed through the subject, and the radiation source and the image detecting unit are opposed to each other in a predetermined direction around a predetermined origin of the subject. And a rotation means for rotating the radiation source and the image detecting means relative to the object along a spiral orbit or a plurality of circular orbits, and a predetermined rotation detected by the image detecting means by the relative rotation. Based on new two-dimensional projection data obtained by inputting the first two-dimensional projection data obtained for each angle and superimposing and integrating a predetermined reconstruction filter on the first two-dimensional projection data. Wide-angle data analysis that forms backprojection images corresponding to multiple slice sections and reconstructs two-dimensional tomographic images corresponding to each slice section based on predetermined high-frequency components of the backprojection images A second step, and second 2D projection data corresponding to each slice section that is transmitted substantially horizontally with respect to a plurality of predetermined slice sections among the 2D projection data. Based on new two-dimensional projection data obtained by superimposing and integrating a predetermined reconstruction filter on the data, a backprojection image corresponding to a plurality of preset slice sections is formed, and a predetermined backprojection image Narrow angle data analyzing means for reconstructing a two-dimensional tomographic image corresponding to each slice section based on low frequency components, and the two-dimensional tomographic images reconstructed by the wide angle data analyzing means and the narrow angle data analyzing means. A configuration processing unit configured to reconstruct a three-dimensional tomographic image by adding and synthesizing images is configured.

Further, the high-low frequency components of the two-dimensional projection data are subjected to the superposition integration of the reconstruction filter, and then the backprojected image is obtained.

Further, a filter function for simultaneously performing the discrimination processing of high and low frequency components and the superposition integration processing of the reconstruction filter is applied to the above two-dimensional projection data, and then the backprojected image is obtained.

Further, in the backprojection operation in the wide-angle data analysis means, the writing density distribution function used when backprojecting new two-dimensional projection data onto a preset slice section is superimposed.

[0022]

According to the present invention having such a configuration, the high-frequency component and the low-frequency component of the reconstructed image are processed separately, and the projection data obtained when the radiation with a wide viewing angle passes through the object. The high-frequency component is reconstructed by reconstructing the low-frequency component by the projection data obtained when radiation with a narrow viewing angle passes through the subject, and then these are combined to reconstruct a three-dimensional tomographic image. Like the Feldkamp method in cone beam CT, image distortion due to image crosstalk can be significantly reduced.

Further, this method is an extension of the conventional two-dimensional tomographic image reconstruction method to reconstruct a three-dimensional tomographic image similarly to the Feldkamp method, and is based on a relatively simple algorithm. As a result, it is possible to reduce the scale of the apparatus and improve the processing speed when applying a computer system.

Further, the projection data obtained by transmitting radiation with a wide viewing angle is backprojected along a tilt direction with respect to a plurality of preset slice sections. Since the projection data is corrected by the writing density distribution function that intersects, it is possible to reconstruct a three-dimensional tomographic image with higher accuracy in actuality.

[0025]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment according to the present invention will be described below with reference to the drawings. In this embodiment, an X-ray cone beam is emitted from an X-ray source to a subject, and transmitted X-rays transmitted through the subject are measured by a two-dimensional detection device to obtain projection data in two dimensions, and this two-dimensional projection is performed. The present invention relates to a transmission type three-dimensional tomography apparatus of a type that reconstructs a three-dimensional tomographic image from data.

First, referring to FIG. 1, the overall structure of the apparatus will be described. A sample table 11 for supporting a subject 10 such as a patient.
Then, by moving the sample table 11 in the vertical direction (vertical direction) z, the subject 10 is also moved vertically along the coordinate axis z, and by rotating the sample table 11 about the coordinate axis z, the object 10 is moved. And a drive device 12 for rotating the same. The sample table drive unit 13 has a power source and the like for supplying drive power to the drive unit 12, and the position detection unit 14 detects the position (height) and the rotational position of the sample table 11 on the coordinate axis z. And a sensor for the above are built in, and both are controlled by the system control unit 24.

An X-ray emission mechanism A and an image pickup mechanism B are provided so as to face each other so as to sandwich the subject 10. The X-ray emission mechanism A includes an X-ray tube 15 that emits X-rays toward the subject 10 in a cone shape, and a collimator 16.
By changing the aperture amount of 6, the viewing angle θ of the X-ray cone beam in the z coordinate axis direction can be adjusted. The high-voltage generator 17 supplies a high-voltage power source for generating X-rays to the X-ray tube 15, and the collimator driver 18 is for adjusting the diaphragm of the collimator 16, both of which are controlled by the system controller 24. There is. The image pickup mechanism B uses an X-ray image intensifier (X-ray II) 19 that injects and magnifies the transmitted X-rays that have passed through the subject 10, and outputs the X-ray image intensifier 19 by a two-dimensional CCD or the like. The camera 20 is provided with a two-dimensional image, and the collimator 21 is provided to prevent incident X-rays outside the range of the viewing angle θ from entering the X-ray image intensifier 19 by adjusting the diaphragm amount. . The collimator drive unit 22 includes the collimator 2
It is for adjusting and driving the aperture amount of 1 and is controlled by the system control unit 24.

The arithmetic unit 23 inputs the data of the two-dimensional image output from the camera 20, that is, the two-dimensional projection data, and performs the data processing described below to reconstruct a three-dimensional tomographic image of the subject 10. The timing of data processing is controlled by the system controller 24.

The system control unit 24 is composed of a computer system, and in accordance with a command issued by an operator via a keyboard or the like provided in the input unit 25, the respective units 13, 14, 17, 18, 22, 23. In addition to controlling, the data of the three-dimensional tomographic image finally obtained by the calculation unit 23 is transferred to the display unit 26 for graphic display and the like.

Further, the X-ray emission mechanism A for the subject 10
The relative positional relationship between the image pickup mechanism B and the image pickup mechanism B will be described in detail with reference to FIG.
That is, the maximum viewing angle θ of the X-ray cone beam emitted from the X-ray tube 15 is set to an angle that can cover the entire subject 10 to be measured. On the other hand, when the collimator 16 is narrowed down, the visual field angle θ is narrowed, and a part of the subject 10 is irradiated with X-rays. However, although the visual field angle θ in the direction of the coordinate axis z is adjusted in this way, the visual field angle α in the direction of the coordinate axis x orthogonal to the coordinate axis z (that is, the horizontal direction when the axis z is the vertical direction) is always It is fixed so as to include the subject 10. Three-dimensional coordinates based on these coordinate axes x, y, and z set with the subject 10 as the center are assumed to be fixed in advance with respect to the subject 10.
Therefore, when the subject 10 is rotated with respect to the measurement system, the coordinate systems x, y, z also rotate around the z axis during measurement.

An imaginary section parallel to the (xy) coordinate plane, which is divided by a minute interval Δz along the coordinate axis z, is called a slice section, and many slice sections are always coordinate axes x ,.
It is fixed in the three-dimensional coordinate system by y and z.
The coordinates of the X-ray incident surface of the X-ray image intensifier 19 are represented by s and h, and the (s−h) plane is always parallel to the z axis. When the object 10 moves together with the sample table 11 in the coordinate axis z direction by a predetermined pitch Δk while the facing relationship between the X-ray tube 15 and the X-ray incident surface of the X-ray image intensifier 19 is fixed, the coordinate axis for the object 10 is changed. If the facing position in the z direction changes relatively, and at the same time, the subject 10 rotates together with the sample stage 11, the facing relationship between the X-ray tube 15 and the X-ray incident surface of the X-ray image intensifier 19 will be changed. A circular orbit is drawn with respect to the subject 10 while being fixed, and the positional relationship between these is freely adjusted by the rotational / moving position of the sample table 11. A pixel array such as a CCD built in the camera 20 is also set in a predetermined positional relationship with respect to the coordinate systems s and h. The horizontal scanning direction s is parallel to the slice section and the vertical scanning direction h is the coordinate axis. It is parallel to z.

Next, the configuration of the arithmetic unit 23 will be described in detail with reference to FIGS. The calculation unit 23 includes a wide angle data analysis unit 27.
And a narrow angle data analysis unit 28 and a synthesis processing unit 29.

First, the wide angle data analysis unit 27 is composed of a wide angle data collection unit 27a, a superposition integration unit 27b, a back projection processing unit 27c, and a high frequency filtering unit 27d, and the wide angle data analysis from the system control unit 24 is performed. Predetermined analysis processing is performed in response to the command. In the wide angle data analysis, as shown in the vertical sectional view of FIG.
The maximum field angle θ1 is set by the collimators 16 and 21 so as to irradiate the X-ray cone beam on the whole 0, and the object 10 on the sample table 11 is fixed at every predetermined rotation angle φ while being fixed at this field angle θ1. The projection data P _H (s, h, φ) of each two-dimensional still image captured by the camera 20 is rotated and analyzed for each rotation angle φ.
Details will be described later.

The narrow angle data analysis unit 28 includes a narrow angle data collection unit 28a, a superposition integration unit 28b, and a back projection processing unit 28.
c and a low-frequency filtering unit 28d, and performs a predetermined analysis process in response to a narrow angle data analysis command from the system control unit 24. In the narrow angle data analysis, as shown in the vertical sectional view of FIG.
A predetermined thickness Δh of 0 (however, in general, Δh > Δk > Δz
The X-ray cone beam is irradiated only on each slice cross section (set to the relationship of), and a narrow viewing angle θ2 is set so that low frequency distortion due to crosstalk of the tomographic image can be sufficiently ignored. In the fixed state, every time the subject 10 is moved along the coordinate axis z by a predetermined pitch interval Δk as shown in FIG. 5B, the subject 10 is rotated by a predetermined rotation angle φ.
Projection data P _L (s, h, φ, z) of each still image captured by the camera 20 for each rotation angle φ.
Is analyzed. That is, each projection data P _L (s, h, φ, z) obtained when the coordinate axis z and the rotation angle φ are used as parameters is analyzed, and the details will be described later. Note that the same imaging effect as in FIG. 5B may be obtained by rotating the subject 10 in a spiral manner as shown in FIG. 5C.

The synthesizing unit 29 is a wide angle data analyzing unit 2
7 and high-frequency component data H (x, y, z) and low-frequency component data L of the two-dimensional tomographic image of each slice slice of the subject 10 reconstructed by the narrow angle data analysis unit 28 and the narrow angle data analysis unit 28, respectively.
Finally, a high-definition three-dimensional tomographic image is reconstructed by synthesizing (x, y, z), and the data I (x, y, z) of this three-dimensional tomographic image is transferred to the system control unit 24. .

Next, the operation of the embodiment having such a configuration will be described.

When the subject 10 is supported on the sample table 11 and measurement is instructed, wide angle data analysis is first performed according to a command from the system controller 24. That is, as shown in FIG. 4, the wide field angle θ1 is set by adjusting the diaphragm amounts of the collimators 16 and 21 so that the entire subject 10 is irradiated with the X-ray cone beam. Then, the object 10 on the sample table 11 is rotated once by a predetermined rotation angle while being fixed at the viewing angle θ1, and the camera 20 captures an image for each rotation angle, whereby the two-dimensional projection data P _H (s, h, φ) is obtained. ) Is output,
The wide-angle data collection unit 27a stores the respective projection data P _H (s, h, φ).

Next, the superposition integrator 27b calculates a predetermined reconstruction filter function g (s) for each of all projection data P _H (s, h, φ) as shown in the following equation (2). By superposing and integrating in the s direction, new projection data Q
Form _H (s, h, φ).

[0039]

[Equation 2]

Here, the reconstruction filter function g (s) is
As shown in the following equation (3), it is a function obtained by the product of the conventionally used Shepp-Logan reconstruction filter function g ₀ (s) and the like and the writing function w (s, θ ′) peculiar to the present invention. .

[0041]

[Equation 3]

The Shepp-Logan reconstruction filter function g ₀ (s) has the real space characteristic shown in FIG.
On the other hand, the write function w (s, θ ′) is obtained based on the principle diagram shown in FIG. That is, when the new projection data Q _H (s, h, φ) is backprojected onto a certain slice section, the angle θ ′ along the backprojection beam of width b with respect to the width a (= Δz) of the slice section. If it is written as, the writing density distribution can be approximated by a triangular writing function w (s, θ ′) shown by the following equation (4).

[0043]

[Equation 4]

Therefore, in reality, the X-ray beam obliquely passes through each slice section, so that the conventionally known Shepp-Logan reconstruction filter function g ₀ (s) is used.
By multiplying the write function w (s, θ ′) by the above, a more realistic reconstruction filter function g (s) can be obtained. More specifically, the reconstruction filter function g (s) has the real space characteristic shown in FIG. 8. Compared with FIG. 6, the Shepplogan reconstruction filter function g ₀ (s) is s. While having a value in the infinite range of, the reconstruction filter function g (s) has a value only within the range regulated by the write function w (s, θ ′) as shown in FIG. As a result, when the Shepp-Logan reconstruction filter function g ₀ (s) is applied, excessive filtering is performed outside the regulation range, and the error increases, whereas the reconstruction filter function g (s) New projection data Q with reduced errors that is realistic
_H (s, h, φ) can be formed.

Next, the back-projection processing unit 27c performs back-projection processing on the new projection data Q _H (s, h, φ) for all angles φ, so that two slices for all slice sections are obtained. A three-dimensional backprojection image r _H (x, y, z) is created.

Next, the high frequency filtering section 27d
Filtering processing is performed to remove low-frequency region components included in the two-dimensional backprojected image r _H (x, y, z). That is, as shown in FIG. 9, the two-dimensional back projection image r
_{Since H} (x, y, z) is reconstructed from the transmittance distribution of the transmitted X-rays that have been transmitted through a plurality of slice sections in an oblique direction,
It includes image distortion resulting from the interference of image information of a plurality of slice sections of the subject 10. Most of this image distortion exists in the low frequency region, and the low frequency region removal type filter removes it.

Specifically, such filtering is realized by a predetermined arithmetic processing. That is, each two-dimensional backprojection image r _H (x, y, z) is converted into two-dimensional data R _H (X, Y, z) in the spatial frequency domain (X-Y) by a two-dimensional fast Fourier transform (FFT). After conversion, a Gaussian function type two-dimensional filter F _H (X, Y, z) corresponding to a low frequency band elimination type filter is integrated. Where F _H (X,
Y, z) is a function of the following equation (5), X and Y are spatial frequencies corresponding to the slice cross section, and σ is a coefficient that determines the cutoff frequency.

[0048]

[Equation 5]

Here, a is the slice width, b is the width of the backprojection beam, and θ ″ is the backprojection angle with respect to the slice plane (x = 0, y = 0, z).

Further, the coefficient σ is σ = ξ | (a + b) /
(2tan θ ″) | /2.35, where ξ is an arbitrary value, and in this embodiment,
Values in the range of 0.25 to 0.5 have been applied.

Then, the integration result R _H (X, Y, z) · F _H (X, Y, z) corresponding to each slice section is subjected to a two-dimensional inverse Fourier transform to obtain a filtered two-dimensional backprojection. The image H (x, y, z) is created, transferred to the synthesis processing unit 29, temporarily stored therein, and the wide angle data analysis ends.

Next, the narrow angle data analysis is performed according to the instruction from the system control unit 24. That is, as shown in FIG. 5A, by adjusting the diaphragm amounts of the collimators 16 and 21, it is set to irradiate the X-ray only with respect to the preset narrow visual field angle θ2. The camera 20 is rotated for each rotation angle φ while rotating the subject 10 on the sample table 11 once for each predetermined rotation angle φ while keeping the viewing angle θ2 fixed.
The process of capturing a still image is repeatedly performed while moving the sample table 11 at an appropriate pitch along the coordinate axis z. Therefore, this movement pitch in the z direction and each rotation φ
The two-dimensional projection data P _L (s, h, φ, z) is obtained for each and stored in the frame memory of the narrow angle data collection unit 28a.

Next, the superimposing and integrating unit 28b superimposes and reconstructs a reconstruction filter g ₀ (s) such as a Shepp-Logan filter for each two-dimensional projection data P _L (s, h, φ, z). New two-dimensional projection data Q _L (s, h,
φ, z) is formed.

Next, the backprojection processing unit 28c performs backprojection processing based on the new two-dimensional projection data Q _L (s, h, φ, z) to generate a two-dimensional backprojection image r _L (x, y,
z) is created.

Next, the low frequency filtering unit 28d
The high frequency component is removed from the two-dimensional backprojected image r _L (x, y, z). That is, specifically, each two-dimensional backprojection image r _L (x, y, z) is subjected to a two-dimensional fast Fourier transform (FFT) to generate two-dimensional data R _L (X, Y in the spatial frequency domain (XY). Y, z), and further integrated with a Gaussian function type two-dimensional filter _FL (X, Y, z) corresponding to a high frequency band elimination type filter. Where _FL (X,
Y, z) is a function of the following equation (6), X and Y are spatial frequencies corresponding to the slice cross section, and σ is a coefficient that determines the cutoff frequency, and is obtained in the same manner as the above equation (5).

[0056]

[Equation 6]

Then, the integrated result R _L (X, Y, z) · F
By performing a two-dimensional inverse Fourier transform on _L (X, Y, z), the filtered two-dimensional backprojection data L (x, y, z) is obtained, transferred to the synthesis processing unit 29, and temporarily stored. , Narrow angle data analysis ends.

Next, the addition processing unit 29 uses the high frequency component image data H (x, y, z) from the high frequency filtering unit 27d and the low frequency from the low frequency filtering unit 28d as shown in the following equation (7). Image data L (x,
y, z) are multiplied by appropriate loads ω _H , ω _L and added and combined to obtain final data of the three-dimensional tomographic image of the subject 10, that is, data I (x, y, z) including high and low frequency components.
And output. The loads ω _H and ω _L are determined so that the image densities of the high and low frequency components become equal in consideration of the measurement time in each measurement, the viewing angle, and the detection efficiency of the detector.

[0059]

[Equation 7]

As described above, in this embodiment, the high-frequency two-dimensional tomographic image is obtained by the wide-angle data analysis, the low-frequency two-dimensional tomographic image is obtained by the narrow-angle data analysis, and the respective two-dimensional tomographic images are obtained. Is finally synthesized to synthesize a three-dimensional tomographic image in a predetermined frequency range.

The effect of this embodiment will be described in comparison with the conventional technique. In order to obtain an accurate three-dimensional tomographic image in the conventional technique, in addition to the simple circular orbit scanning as shown in FIG. When performing scanning in the direction at right angles, it was necessary to measure with sufficient resolution also in this scanning direction. On the other hand, in this embodiment, high-resolution measurement may be performed only when obtaining a high-frequency two-dimensional tomographic image, and when obtaining a low-frequency tomographic image, high resolution is not required, so the amount of data is small. It may be small, and processing time for calculation and memory may be small.

Generally, in two-dimensional CT image reconstruction, the sampling interval of projection data and the angular interval in the measurement direction are related to the resolution to be obtained. If the resolution (half-value width) is d, the sampling interval of the projection needs to be d / 2 or less. If the diameter of the object is D, the projection data in one direction has at least 2D / d data. is necessary. Also, the required number of measurement directions is almost equal to this, and it is (2D / d) ² per slice section in high-frequency images.
Data is required. That is, the number of data per slice section is inversely proportional to the square of the resolution. Since the low-frequency cross-sectional image may have a low resolution, the number of data is much smaller than that in the case of forming a high-frequency tomographic image. For example, if the resolution of the high-frequency tomographic image is 1 mm and the resolution of the low-frequency tomographic image is 5 mm, the total number of data per slice slice of the low-frequency tomographic image may be 1/25 of that of the high-frequency tomographic image. The interval is 5 times lower resolution. Therefore, in this embodiment, the viewing angle of the X-ray cone beam when obtaining the low-frequency tomographic image is made considerably smaller than the viewing angle when obtaining the high-frequency tomographic image, and the number of circular orbits is increased accordingly. Even so, the measurement time of the projection data for obtaining the low-frequency tomographic image can be made equal to or shorter than the measurement time of the projection data for obtaining the high-frequency tomographic image. Further, the calculation time required for reconstruction of a tomographic image also largely depends on the amount of data, and particularly in reconstruction of a low-frequency tomographic image, when the viewing angle θ2 is sufficiently small, two-dimensional backprojection is used instead of three-dimensional backprojection. May be used, the calculation time is greatly shortened.

The required measuring time is also related to the quantum noise (statistical fluctuation) of the detected radiation. In general, the quantum noise in a two-dimensional CT image is inversely proportional to the cube of the resolving power (half-width), so in the above example, the amount of radiation required per slice section of a low-frequency tomographic image is 1 /
It may be 5 ³ , that is, 1/125. Therefore, the influence of quantum noise on the low-frequency image can be reduced to a negligible level.

From the above points, according to this embodiment, it is possible to provide a cone beam CT apparatus with less image distortion and crosstalk by a relatively simple scanning and reconstruction algorithm.

Next, a second embodiment will be described with reference to FIG. The overall structure of the apparatus of this embodiment is the same as that shown in FIG. The difference lies in the configuration of the calculation unit 23,
In the previous embodiment shown in FIG. 3, the wide angle data analyzing unit 27 and the narrow angle data analyzing unit 28 are the back projection processing units 27c and 28, respectively.
New two-dimensional projection data r _H (x, y, z) and r _{L in} c
After (x, y, z) is created, the high-frequency filtering unit 27d and the low-frequency filtering unit 28d perform a filtering process to remove unnecessary frequency components and generate image data H (x, y, z) and L. The configuration is such that (x, y, z) is obtained, but in the second embodiment, as shown in FIG. 10, high-low frequency filtering is first performed and then backprojection processing is performed.

That is, as shown in FIG. 10, in the wide-angle data analysis unit 27, the projection data Q is calculated by the superimposition integration unit 27b by the calculation processing based on the equations (2) to (4).
_H (s, h, φ) is obtained, and then the high frequency filtering unit 27d performs a filtering process for removing the low frequency region component included in the projection data Q _H (s, h, φ). Specifically, the high-frequency filtering unit 27d superimposes the high-frequency filter function f _H (s) shown in the following equation (8) on all projection data Q _H (s, h, φ) in the s direction. New high frequency component projection data Q ′ _H (s, H, φ) is formed by the integration.

[0067]

[Equation 8]

Σ is σ = ξ · | (a + b) / (2t
an θ ′) | /2.35. However, ξ
Is an arbitrary value to some extent. Next, the back projection processing unit 27c
Performs backprojection processing on the high-frequency component projection data Q ′ _H (s, h, φ) to create two-dimensional backprojection images for all slice planes. When processed in this way, the two-dimensional backprojection image is the image H shown in FIG.
This matches (x, y, z) and is input to the synthesis processing unit 29.

On the other hand, in the narrow angle data analysis unit 28, first, the low frequency filtering unit 28d is replaced with the superposition integration unit 28b.
The filtering process for removing the low frequency region component included in the projection data Q _L (s, h, φ) output from Specifically, the low frequency filtering unit 28d
But the low-frequency filter function f _L (s) all projection data Q _L shown in the following equation (9) (s, h, φ, z) for each of, by convolving the s direction, the new Low frequency component projection data Q ′ _L (s, h, φ, z).
Note that σ in the equation (9) is determined so that the low frequency filter has a frequency characteristic complementary to the high frequency filter.

[0070]

[Equation 9]

Next, the backprojection processor 28c determines that all angles φ
New low-frequency component of the projection data Q _'L (s,
By performing the backprojection process on (h, φ, z), a two-dimensional backprojection image for all slice sections is created. When processed in this way, the two-dimensional backprojected image matches the image L (x, y, z) shown in FIG. 3, and this is input to the synthesis processing unit 29.

Next, a third embodiment will be described with reference to FIG. The overall structure of the apparatus of this embodiment is the same as that shown in FIG. Explaining the differences from the first and second embodiments, the arithmetic unit 23 shown in FIGS. 3 and 10 includes a superposition integration unit 27b, 28b and a high frequency filtering unit 27.
d and the low-frequency filtering unit 28d are provided separately and independently, but in the present embodiment, as shown in FIG. 11, the superimposing and integrating unit 27e causes the reconstructing filter function g of the superimposing and integrating unit 27b.
(S) and the filter function f _H (s) of the high-frequency filtering unit 27 d, the projection function P _H (s,
h, φ) and the removal of low-frequency components are collectively performed, and the superposition integration unit 28e causes the reconstruction filter function g (s) of the superposition integration unit 28b and the filter function f _L (of the high-frequency filtering unit 28d). By the filter function g _L (s) shown in the following equation (11) having both characteristics of
The configuration is such that the superposition integration on the projection data P _L (s, h, φ, z) and the removal of high frequency components are collectively processed.

[0073]

[Equation 10]

[0074]

[Equation 11]

Note that * indicates a convolution operator.

Therefore, according to this third embodiment,
The high frequency filtering unit 27d and the low frequency filtering unit 28d in FIGS. 3 and 10 can be omitted, and the processing speed can be improved.

Next, a fourth embodiment will be described with reference to FIGS. The device configuration is almost the same as in FIG. However, the difference lies in the irradiation method of the X-ray cone beam on the subject 10, that is, the X-ray scanning method. Figure 12 (a)
, The X-ray cone beam is fixed at a wide viewing angle so that the entire subject 10 is irradiated with X-rays.
The X-ray emission central axis connecting the center of the incident surface of the image intensifier 19 with the center O of the subject 10 (that is,
x, y, z coordinate origin). Then, the X-ray tube 1 is configured so that the angle β formed by the X-ray emission central axis and the (xy) coordinate plane is changed by a predetermined step angle Δβ.
5 and the image intensifier 19 are tilted, and the X-ray tube 15 and the image intensifier 19 are rotated along a circular orbit centered on the coordinate axis z for each change of the angle β, and each of the predetermined rotation angles φ The camera 20 captures the projected image. That is, as shown in the cross-sectional view of FIG. 12B, if the X-ray tube 15 is displaced upward along the coordinate axis z while drawing a circular orbit, it is at a position facing the center O of the subject 10. The image intensifier 19 performs X-ray scanning while displacing downward while drawing a circular orbit. Therefore, in this case, the X-ray incident surface sh of the imaging intensifier 19 is inclined with respect to the z axis.

Further, the projection data P (s, h, φ, β) obtained by this X-ray scanning is input to the arithmetic unit 23. In this embodiment, as shown in FIG. 13, as shown in FIG. A frame memory 30 is further provided in the input section, and the camera 20
All projection data P (s, h, φ, β) from are temporarily held.

When the X-ray scanning of this embodiment is performed, the direction of the X-ray beam in all the projection data P (s, h, φ, β) seems to be substantially parallel to all the slice sections. Two-dimensional projection data P _L (s, h, φ, β) is included. Therefore, the two-dimensional projection data P _L (s, h, φ, β) of the low frequency component is selectively read out from all the projection data P (s, h, φ, β) stored in the frame memory 30 and narrowed. It is transferred to the frame memory of the angle data collection unit 28a. On the other hand, all the projection data P (s, h, φ, β) are set as the two-dimensional projection data P _H (s, h, φ, β) for each rotation angle φ as the frame memory of the wide-angle data collection unit 27a. And are stored in the memory. Then, these two-dimensional projection data P _H (s, h, φ, β), P _L (s, h, φ, β)
Performs the same processing as in the first embodiment described above, and the combining processing unit 2
9 outputs image data I (x, y, z) of an image reconstructed from a three-dimensional tomographic image of the subject 10.

That is, according to the fourth embodiment, the view angle of the X-ray cone beam is switched between the narrow angle and the wide angle in the first embodiment, while the high angle is achieved only in the wide angle mode. Since the projection data in the frequency band and the low frequency band can be measured, a simple device can be provided.

However, the arithmetic unit 23 shown in FIG. 13 is a modification of the arithmetic unit 23 shown in FIG. 1, but the arithmetic unit 23 shown in FIG. 10 or 11 may be added with a frame memory 30.

In the first to fourth embodiments described above, the filtering function is of the Gaussian function type as shown in Expressions (5) and (6) and Expressions (8) and (9). However, these functions may be arbitrary to the extent that these low-frequency filters and high-frequency filters are complementary to each other with respect to the spatial frequency (which becomes 1 when combined).

As shown in FIGS. 3, 10, 11, and 13, the frequency band covered by the measurement system was divided into two. The crosstalk frequency increased as the maximum viewing angle of the X-ray cone beam increased. Since the band becomes higher, it is necessary to narrow the frequency band covered by the measurement system. In this case,
The number of frequency bands to be divided is not limited to two and may be three or more. Also in this case, each projection data P _H (s,
Filtering is performed so that the crosstalk component included in (h, φ) is limited to a frequency band that can be ignored, and when they are combined, uniform frequency characteristics can be obtained as a whole.

Further, as described above, not only X-ray scanning along a circular orbit but also X-ray scanning along a spiral orbit can be performed, but X-ray scanning along a spiral orbit is possible. In this case, the measurement of each frequency band scans the region of the subject to be measured as uniformly as possible. Further, instead of one spiral orbit, a combination of a plurality of spiral orbits having different rotation angle phases may be adopted.

Further, in scanning, the measurement system may be moved with respect to a stationary subject, or the subject may be moved with respect to a stationary measurement system.

Further, a plurality of frequency band images (projection data)
In order to measure, the same image detector may be used to measure different conditions, for example, by changing the effective range of the image detector, or multiple image detectors with different effective ranges may be applied. You may.

Further, in the cone beam CT, radiation is scattered after being detected by the object and then detected, so that the contrast of the image tends to be lowered. The larger the viewing angle of the cone beam, the more scattered rays are measured from a wide angle, so the image quality is significantly degraded by this effect. In such a case, since the contribution of this scattered radiation to the image is mainly in the low spatial frequency region, most of the high frequency image of the present invention can be removed by high frequency filtering. On the other hand, in the measurement of low-frequency images, the influence of scattered rays can be reduced by using a slit collimator that limits the visual field in the axial direction of the image detector.

Although the first to fourth embodiments are X-ray CT apparatuses in which an X-ray tube is used as a radiation source and an image intensifier is used as an image detecting means, the present invention is not limited to this. For example, an isotope ray source may be applied to the radiation source, a gamma camera, an imaging plate, or an image detector using a scintillator and a photomultiplier, a two-dimensional semiconductor position sensor, or a solid-state imaging device may be applied to the image detecting means. Further, a light generator may be applied to the radiation source, a two-dimensional semiconductor position sensor, a solid-state imaging device, or a position detection type photomultiplier tube may be applied to the image detecting means. Furthermore, a neutron generator may be applied to the radiation source, and an image detector including a scintillator and a photomultiplier tube may be applied to the image detecting means.

[0089]

As described above, according to the present invention, the high-frequency component and the low-frequency component of a desired three-dimensional image are projected data obtained when radiation with a wide viewing angle passes through a subject, and radiation with a narrow viewing angle. The projection data obtained when the object passes through the object is reconstructed separately and then combined to obtain a third-order tomographic image. It is based on a simple algorithm. As a result, it is possible to reduce the scale of the apparatus and improve the processing speed when applying a computer system.

Further, even when the viewing angle of the X-ray cone beam is large, low-frequency components that cause image crosstalk in the projection data with a wide viewing angle are removed by high-frequency filtering. The original tomographic image can be reconstructed.

Further, the projection data obtained by the transmission of the radiation with the wide viewing angle is caused by the radiation transmitted from the inclination direction with respect to a plurality of preset slice sections, but the projection data is inverted after the superposition integration. Since the projection data is corrected by the writing density distribution function at the intersection of the slice cross section and the back projection beam upon projection, it is possible to reconstruct a more accurate three-dimensional tomographic image in actuality.

Further, as shown in FIG. 18, in actual measurement, it is difficult to completely include an object in the X-ray cone beam and perform X-ray scanning with a wide viewing angle. For example, if a three-dimensional tomographic image of a human head is to be reconstructed, the X-ray will be uniformly irradiated because the neck and the like are always inside or outside the X-ray cone beam range. There is a portion that is not covered (such as the shaded portion U in FIG. 16). In such a case, the conventional cone beam CT cannot reconstruct an accurate three-dimensional tomographic image with the shaded portion U,
According to the present invention, since a two-dimensional tomographic image for each slice section is constructed and then combined to reconstruct a three-dimensional tomographic image, crosstalk of the image between the slice sections is eliminated, and diagonal lines are drawn. Even when there is the portion U, a three-dimensional tomographic image with little distortion can be reconstructed.

By applying the present invention to, for example, an X-ray CT apparatus, it is extremely excellent in basic research in the fields of medicine and biology, diagnosis of diseases in the field of clinical medicine, non-destructive inspection in the field of industry, and the like. It is effective.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a first exemplary embodiment of the present invention.

FIG. 2 is an explanatory diagram illustrating a main configuration of FIG.

FIG. 3 is a block diagram showing a configuration of a calculation unit in FIG.

FIG. 4 is an explanatory diagram for explaining an operation during wide angle data analysis according to the first embodiment.

FIG. 5 is an explanatory diagram for explaining an operation during narrow angle data analysis according to the first embodiment.

FIG. 6 is an explanatory diagram showing a real space characteristic of a reconstruction filter of Shepp-Logan.

FIG. 7 is an explanatory diagram illustrating a function of a write function.

FIG. 8 is an explanatory diagram showing real space characteristics of a reconstruction filter applied in the first embodiment.

FIG. 9 is an explanatory diagram for explaining the principle of backprojection processing according to the first embodiment.

FIG. 10 is a block diagram showing a main configuration of a second embodiment.

FIG. 11 is a block diagram showing a main configuration of a third embodiment.

FIG. 12 is an explanatory diagram for explaining the principle of X-ray scanning in the fourth embodiment.

FIG. 13 is a block diagram showing a main configuration of a fourth embodiment.

FIG. 14 is a principle explanatory view showing a principle of a conventional transmission type two-dimensional tomography apparatus.

FIG. 15 is a principle explanatory diagram showing a generation principle of projection data in a conventional transmission type two-dimensional tomography apparatus.

FIG. 16 is a principle explanatory view showing the principle of a conventional cone-beam CT apparatus.

FIG. 17 is an explanatory diagram showing a trajectory of a focus of an X-ray tube required for a conventional cone-beam CT apparatus.

FIG. 18 is an explanatory diagram for explaining a problem in the conventional cone-beam CT apparatus.

[Explanation of symbols]

10 ... Subject, 11 ... Sample stand, 12 ... Drive device, 13 ...
Sample stage drive unit, 14 ... Position detection unit, 15 ... X-ray tube, 1
6, 21 ... Collimator, 17 ... High-voltage generator, 18, 22
... Collimator drive section, 23 ... Calculation section, 24 ... System control section, 25 ... Input section, 26 ... Display section, 27 ... Wide angle data analysis section, 27a ... Wide angle data collection section, 27b ... Superposed integration section, 27c ... Backprojection processing unit, 27d ... High-frequency filtering unit, 27e ... Superimposition integration unit having high-frequency filtering function, 28 ... Narrow angle data analysis unit, 28a ... Narrow angle data collection unit, 28b ... Superposition integration unit, 28c ... Backprojection processing Section, 28d ... Low-frequency filtering section, 28e ...
Superimposed integrator with low frequency filtering function, 29 ...
Composition processing unit, 30 ... Frame memory.

Claims (11)

[Claims] 1. A two-dimensional projection data is generated by two-dimensionally detecting a radiation source that radiates a radiation in a cone shape at an arbitrary viewing angle to a subject and a transmitted radiation transmitted through the subject. Image detecting means, and a driving means capable of relatively moving either the radiation and the image detecting means or the subject along a spiral orbit or a plurality of circular orbits and relatively moving in a direction orthogonal to the orbital plane. And when the drive means performs the relative rotation drive in a state where the radiation source emits cone-shaped radiation with a wide viewing angle that covers the entire subject, the image detection means for each predetermined rotation angle. Input the first 2D projection data detected by
Based on the new two-dimensional projection data obtained by superimposing and integrating a predetermined reconstruction filter on the first two-dimensional projection data, while forming a back projection image corresponding to a plurality of preset slice sections, Wide-angle data analysis means for reconstructing a two-dimensional tomographic image corresponding to each slice section based on a predetermined high-frequency component of the backprojection image, and in a state in which the radiation source irradiates a part of the subject, When the drive means performs the relative rotation drive and the relative movement, the second two-dimensional projection data detected by the image detection means is input for each predetermined rotation angle, and the second two-dimensional projection data is predetermined. Based on the new two-dimensional projection data obtained by superimposing and integrating the reconstruction filter of, a backprojection image corresponding to a plurality of preset slice sections is formed, and a predetermined backprojection image Narrow angle data analysis means for reconstructing a two-dimensional tomographic image corresponding to each slice section based on low frequency components, and the two-dimensional tomographic images reconstructed by the wide angle data analysis means and the narrow angle data analysis means A transmission type three-dimensional tomography apparatus, comprising: a configuration processing unit that synthesizes images to reconstruct a three-dimensional tomographic image. 2. A two-dimensional projection data is generated by two-dimensionally detecting a radiation source that radiates radiation to a subject in a cone shape at an arbitrary viewing angle and transmitted radiation that has passed through the subject. Image detecting means, and a driving means capable of relatively moving either the radiation and the image detecting means or the subject along a spiral orbit or a plurality of circular orbits and relatively moving in a direction orthogonal to the orbital plane. And when the drive means performs the relative rotation drive in a state where the radiation source emits cone-shaped radiation with a wide viewing angle that covers the entire subject, the image detection means for each predetermined rotation angle. Input the first 2D projection data detected by
A backprojection image corresponding to a plurality of preset slice sections based on a predetermined high-frequency component of new two-dimensional projection data obtained by superposing and integrating a predetermined reconstruction filter on the first two-dimensional projection data By forming a wide angle data analysis means for reconstructing a two-dimensional tomographic image corresponding to each slice cross section, and in a state where the radiation source irradiates a part of the subject, the drive means is When performing the relative rotation drive and the relative movement, the second two-dimensional projection data detected by the image detecting means is input for each predetermined rotation angle, and a predetermined reconstruction filter is input to the second two-dimensional projection data. By forming a backprojection image corresponding to a plurality of preset slice sections based on a predetermined low-frequency component of new two-dimensional projection data obtained by superimposing integration, A narrow-angle data analysis means for reconstructing a two-dimensional tomographic image corresponding to a chair cross section, a three-dimensional tomographic image by combining the two-dimensional tomographic images reconstructed by the wide-angle data analysis means and the narrow-angle data analysis means. A transmission type three-dimensional tomography apparatus, comprising: a configuration processing unit that reconstructs an image. 3. A two-dimensional projection data is generated by two-dimensionally detecting a radiation source that radiates a radiation in a cone shape at an arbitrary viewing angle to a subject and a transmitted radiation transmitted through the subject. Image detecting means, and a driving means capable of relatively moving either the radiation and the image detecting means or the subject along a spiral orbit or a plurality of circular orbits and relatively moving in a direction orthogonal to the orbital plane. And when the drive means performs the relative rotation drive in a state where the radiation source emits cone-shaped radiation with a wide viewing angle that covers the entire subject, the image detection means for each predetermined rotation angle. Input the first 2D projection data detected by
New two-dimensional projection data is formed by performing predetermined reconstruction filtering and predetermined low-frequency component removal filtering processing on the first two-dimensional projection data, and is preset based on the new two-dimensional projection data. Wide-angle data analysis means for reconstructing a two-dimensional tomographic image corresponding to each slice section by forming a backprojection image corresponding to a plurality of slice sections, and the radiation source for a part of the subject. The second two-dimensional projection data detected by the image detecting means is input at each predetermined rotation angle when the driving means performs the relative rotation driving and the relative movement in the state of being emitted by the second driving means. When new 2D projection data is formed by performing predetermined reconstruction filtering and predetermined high frequency component removal filtering processing on the 2D projection data of Together, based on the new two-dimensional projection data, a backprojection image corresponding to a plurality of preset slice sections is formed to reconstruct a two-dimensional tomographic image corresponding to each slice section. An angle data analysis means; and a configuration processing means for reconstructing a three-dimensional tomographic image by synthesizing the two-dimensional tomographic images reconstructed by the wide-angle data analyzing means and the narrow-angle data analyzing means. A transmission type three-dimensional tomography apparatus characterized by the following. 4. A two-dimensional object, comprising: a radiation source that radiates a cone-shaped radiation having a wide viewing angle that covers the entire subject, and two-dimensionally detecting transmitted radiation that has passed through the subject. Image detecting means for generating projection data, and the radiation source and the image detecting means are opposed to each other in a predetermined direction around a predetermined origin of the subject, and the radiation source and the image detecting means are spirally orbited with respect to the subject. Alternatively, the driving means capable of relative rotation along a plurality of circular orbits, and the first two-dimensional projection data obtained at each predetermined rotation angle detected by the image detection means by the relative rotation are input, and the first second Based on the new two-dimensional projection data obtained by superimposing and integrating a predetermined reconstruction filter on the three-dimensional projection data, a backprojection image corresponding to a plurality of preset slice sections is formed and the backprojection image is formed. Wide-angle data analysis means for reconstructing a two-dimensional tomographic image corresponding to each slice section based on a predetermined high-frequency component of the image, and among the two-dimensional projection data, a plurality of predetermined slice sections Based on new two-dimensional projection data obtained by inputting the second two-dimensional projection data transmitted substantially horizontally and superimposing and integrating a predetermined reconstruction filter on the second two-dimensional projection data, Forming a backprojection image corresponding to a plurality of slice sections, narrow angle data analysis means for reconstructing a two-dimensional tomographic image corresponding to each slice section based on a predetermined low-frequency component of the backprojection image, And a configuration processing unit configured to combine the two-dimensional tomographic images reconstructed by the wide-angle data analyzing unit and the narrow-angle data analyzing unit to reconstruct a three-dimensional tomographic image. Over-type three-dimensional tomography apparatus. 5. A two-dimensional object is obtained by two-dimensionally detecting a radiation source that radiates a cone-shaped radiation having a wide viewing angle that covers the entire subject and two-dimensionally the transmitted radiation that has passed through the subject. Image detecting means for generating projection data, and the radiation source and the image detecting means are opposed to each other in a predetermined direction around a predetermined origin of the subject, and the radiation source and the image detecting means are spirally orbited with respect to the subject. Alternatively, the driving means capable of relative rotation along a plurality of circular orbits, and the first two-dimensional projection data obtained at each predetermined rotation angle detected by the image detection means by the relative rotation are input, and the first second A backprojection image corresponding to a plurality of preset slice sections is formed based on a predetermined high-frequency component of new two-dimensional projection data obtained by superimposing and integrating a predetermined reconstruction filter on the three-dimensional projection data. By doing so, wide-angle data analysis means for reconstructing a two-dimensional tomographic image corresponding to each slice section, and of the above-mentioned two-dimensional projection data, is transmitted substantially horizontally to a plurality of predetermined slice sections. The second two-dimensional projection data is input, and preset based on a predetermined low-frequency component of new two-dimensional projection data obtained by superimposing and integrating a predetermined reconstruction filter on the second two-dimensional projection data. Narrow angle data analyzing means for reconstructing a two-dimensional tomographic image corresponding to each slice section by forming a backprojection image corresponding to a plurality of slice sections, and the wide angle data analyzing means and narrow angle data A transmission type three-dimensional tomography apparatus comprising: a configuration processing unit configured to reconstruct a three-dimensional tomographic image by synthesizing the two two-dimensional tomographic images reconstructed by an analyzing unit. 6. A two-dimensional object, comprising: a radiation source that emits cone-shaped radiation having a wide viewing angle that covers the entire subject and two-dimensionally detecting transmitted radiation that has passed through the subject. Image detecting means for generating projection data, and the radiation source and the image detecting means are opposed to each other in a predetermined direction around a predetermined origin of the subject, and the radiation source and the image detecting means are spirally orbited with respect to the subject. Alternatively, the driving means capable of relative rotation along a plurality of circular orbits, and the first two-dimensional projection data obtained at each predetermined rotation angle detected by the image detection means by the relative rotation are input, and the first second Forming new two-dimensional projection data by performing predetermined reconstruction filtering and predetermined low frequency component removal filtering processing on the two-dimensional projection data, and based on the new two-dimensional projection data, Wide-angle data analysis means for reconstructing a two-dimensional tomographic image corresponding to each slice section by forming a backprojection image corresponding to a plurality of preset slice sections; , Inputting the second two-dimensional projection data transmitted substantially horizontally with respect to a plurality of predetermined slice sections, and performing predetermined reconstruction filtering and predetermined high-frequency component removal filtering processing on the second two-dimensional projection data By forming new two-dimensional projection data by, by forming a back-projection image corresponding to a plurality of preset slice sections based on the new two-dimensional projection data, to each slice section Narrow angle data analysis means for reconstructing the corresponding two-dimensional tomographic image, and reconstructed by the wide angle data analysis means and the narrow angle data analysis means. Transmission type three-dimensional tomography apparatus characterized by comprising a structure processing means for reconstructing a three-dimensional tomographic image by combining the two two-dimensional tomographic image. 7. The wide-angle data analysis means determines a writing density of a new two-dimensional projection data at a portion where a beam for writing projection data at the time of back projection intersects with the preset slice cross section. The transmission type three-dimensional tomography apparatus according to claim 1, wherein a distribution function corresponding to the distribution is superimposed. 8. The transmission type three-dimensional structure according to claim 1, wherein the radiation source is an X-ray generator, and the image detection means is an X-ray imaging intensifier. Tomography equipment. 9. The radiation source is an isotope source, the image detecting means is a gamma camera, or an imaging plate, or an image detector using a scintillator and a photomultiplier tube.
Alternatively, the transmission type three-dimensional tomography apparatus according to any one of claims 1 to 6, comprising a two-dimensional semiconductor position sensor or a solid-state imaging device. 10. The radiation source is a light generator, and the image detecting means is a two-dimensional semiconductor position sensor, a solid-state imaging device, or a position detection type photomultiplier tube. Item 7. The transmission type three-dimensional tomography apparatus according to any one of items 6. 11. The radiation source is a neutron generator, and the image detecting means is an image detector including a scintillator and a photomultiplier tube, according to any one of claims 1 to 6. Transmission type three-dimensional tomography device.