JP2014113207A - Tomographic apparatus and tomographic image processing method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a tomographic apparatus capable of forming and displaying a horizontal sectional image by which an operator can check a detailed structure to a peripheral part in a prescribed site of a specimen.SOLUTION: The tomographic apparatus for forming three-dimensional tomographic images comprises: first selection means that selects one boundary surface (layer) in a plurality of boundary surfaces (layers) contained in a B-scan image; second selection means that selects a position in the depth direction (z direction) of the B-scan image; and horizontal sectional image generation means that generates a horizontal sectional image in the depth position selected by the second selection means on the basis of the three-dimensional position information of the boundary surface selected by the first selection means.

Description

The present invention relates to a tomographic imaging apparatus and a tomographic image processing method.

In recent years, an optical coherence tomography apparatus for photographing a tomographic image of a subject's eye (eyeball) by optical coherence tomography (OCT) has been provided as an inspection apparatus used for ophthalmic examination. Yes.

The optical coherence tomography apparatus uses a time domain method called a time domain method, which uses a time domain OCT that acquires a tomographic image while moving a mirror and mechanically changes the optical path length of reference light, and a spectroscope called a Fourier domain method. There is a spectral domain OCT that detects spectral information and acquires a tomographic image, or an optical frequency sweep OCT that detects a spectral interference signal using a wavelength scanning light source and acquires a tomographic image.

In general, in OCT, a one-dimensional signal in the depth direction of the eye to be examined is acquired (A-scan), and a two-dimensional tomographic image is acquired by scanning the measurement light with respect to the eye to be examined one-dimensionally (B-). Scanning), and further obtaining a three-dimensional image by repeatedly obtaining a two-dimensional tomographic image while shifting the position with respect to the eye to be examined (C-scan).

  Therefore, from the obtained three-dimensional image data (also referred to as volume data), a horizontal cross-sectional image (X− Y plane image) can be obtained. Such a horizontal cross-sectional image is particularly useful in the choroid, which has a complicated vascular structure, because the vascular structure can be confirmed in detail.

However, the fundus is curved and an OCT tomogram is generally extracted with a certain curvature. For this reason, when a horizontal cross-sectional image is simply extracted on the XY plane perpendicular to the Z direction with respect to the three-dimensional image data (volume data), there is a problem that a detailed cross-sectional image cannot be obtained up to the peripheral portion.

Patent Document 1 discloses a technique for forming a tomographic image representing the form of the optic nerve head based on a three-dimensional image of the fundus and forming horizontal cross-sectional images for xy slice planes at three different depth positions.

In a narrow range such as the optic papilla, a useful tomographic image can be obtained even if a horizontal cross-sectional image at an arbitrary depth position (Z-direction position) is acquired on the xy slice plane with respect to the acquired three-dimensional image of the fundus. However, when acquiring a horizontal cross-sectional image of a wide range in the choroid, even if a similar xy slice plane is acquired, detailed tomographic images are obtained up to the periphery due to the curvature of the fundus as described above. There are things you can't get.

In other words, in the horizontal cross-sectional image extracted in the XY plane perpendicular to the Z direction, the images in the peripheral portion are combined with images of a plurality of different layers, and the image may be blurred. This is because it cannot be done.

Patent Document 2 discloses a method of acquiring a horizontal cross-sectional image (three-dimensional plane image) of an arbitrary boundary surface, for example, a retinal pigment epithelium layer. According to this method, the three-dimensional of an arbitrary layer is obtained from the two-dimensional arrangement data of each arbitrary layer (such as a retinal pigment epithelium layer) existing in two or more B-scan images that are substantially orthogonal to or substantially parallel to each other. Position information is calculated, and based on the obtained three-dimensional position information, the optical path length in the reference mirror and the scanning of the scanning optical system are controlled to obtain a horizontal sectional image (three-dimensional plane image) of the retinal pigment epithelium layer of the eye to be examined. get. According to this method, since the horizontal cross-sectional image is extracted based on the three-dimensional position information, it is considered that a detailed image can be acquired up to the peripheral part.

JP 2012-100811 A JP 2010-151713 A

However, the horizontal tomographic image (three-dimensional planar image) obtained based on the three-dimensional position information in Patent Document 2 is more than the case of acquiring a horizontal cross-sectional image with a simple xy slice plane perpendicular to the Z direction according to Patent Document 1. Is considered to be able to obtain a detailed horizontal cross-sectional image over the periphery, but the three-dimensional position information, which is a reference for obtaining the horizontal cross-sectional image, is estimated from two or more B-scan images. Since it is only a value, it is considered that an accurate horizontal cross-sectional image (three-dimensional plane image) cannot be obtained.

In addition, the method of acquiring a horizontal cross-sectional image (three-dimensional plane image) by controlling the optical path length in the reference mirror and scanning of the scanning optical system based on the three-dimensional position information requires complicated control processing. In some cases, acquisition may take time.

Further, when acquiring a B-scan image for calculating three-dimensional position information and when acquiring a horizontal cross-sectional image (three-dimensional plane image) based on the three-dimensional position information calculated from the B-scan image, a time difference is obtained. Therefore, the eye to be examined moves during that time, and as a result, there is a possibility that an accurate horizontal cross-sectional image cannot be obtained.

The present invention is not limited to a means for calculating three-dimensional position information from data of two or more B-scan images or a control processing means for a complicated optical system as disclosed in Japanese Patent Application Laid-Open No. 2003-260688. An object of the present invention is to provide a tomographic imaging apparatus and an image processing method capable of forming and displaying a horizontal cross-sectional image in which a detailed structure can be confirmed.

  In order to achieve the above object, in the tomographic imaging apparatus for creating a three-dimensional tomographic image by superimposing a plurality of B-scan images (tomographic images), one B-scan image in the plurality of B-scan images is obtained. First selection means for selecting one boundary surface (layer) from a plurality of boundary surfaces (layers) included in the scan image, and a depth direction (Z direction) with respect to the one B-scan image Second selection means for selecting a position, and three-dimensional position information of a boundary surface selected by the first selection means for a horizontal cross-sectional image at the depth position selected by the second selection means. Horizontal cross-sectional image generating means for extracting from the image data and generating a horizontal cross-sectional image at the position selected by the second selecting means by re-processing the three-dimensional tomographic image data based on the extracted position information; It is characterized by that.

  According to the above configuration, the three-dimensional position information on the boundary surface selected by the first selection unit used as a reference when generating the horizontal cross-sectional image at the target depth position is the acquired three-dimensional tomographic image data ( Since the boundary surface data selected by the first selection means is extracted from the volume data) and the three-dimensional position information is acquired, the reference three-dimensional position information of the boundary surface can be accurately acquired.

  When generating a horizontal cross-sectional image at the target depth position, a horizontal cross-sectional image is generated by re-processing the acquired three-dimensional tomographic image data (volume data) based on the acquired three-dimensional position information. Therefore, it is not necessary to perform complicated control on the optical system as in Patent Document 2.

  Furthermore, since the three-dimensional position information and the generated horizontal cross-sectional image use the same three-dimensional tomographic image data (volume data), an image defect due to the movement of the eye to be inspected due to a time difference as feared in Patent Document 2 There is no fear of it.

As described above, according to the tomographic imaging apparatus and the tomographic image processing method of the present invention, the horizontal cross-sectional image at the target depth position has a boundary surface (with a curvature close to the curvature of the fundus). Since it is generated based on the three-dimensional position information obtained from the layer), an excellent effect is obtained that a horizontal cross-sectional image in which a detailed structure can be confirmed up to the peripheral part can be obtained.

It is the figure which showed the detail of the tomogram acquisition part. It is the figure which showed the structure of the tomography apparatus. It is a figure explaining the acquisition flow of a three-dimensional tomogram, and the operation flow of the 1st selection means in the Example of this invention, and a 2nd selection means. It is a figure explaining the flow until acquisition of a three-dimensional tomogram. It is the figure which showed an example of the method of selecting one B-scan image from a fundus front image. It is a figure explaining the 1st selection means of this invention. It is a figure explaining the 2nd selection means of this invention. It is a figure explaining the method to acquire the three-dimensional position information of the horizontal slice image produced | generated using three-dimensional tomogram data (volume data). It is the figure which showed the operation flow which produces | generates the horizontal tomogram in the depth position of the target fundus. It is the figure which showed the horizontal cross-sectional image in the position shown with the B-scan image and the B-scan image. It is a figure explaining the example of a display of the monitor screen based on the Example of this invention. It is the figure which showed the arbitrary A-scan data in B-scan data.

Hereinafter, a tomographic imaging apparatus according to an embodiment of the present invention will be described with reference to the drawings. First Embodiment FIG. 1 shows a detailed configuration of a tomographic image acquisition unit 100.

  As shown in FIG. 1, the tomographic image acquisition unit 100 shoots a three-dimensional tomographic image of the fundus oculi Er by irradiating measurement light onto the fundus oculi (fundus retina) Er of the eye E. In this embodiment, a Fourier domain (optical frequency sweep) method using a wavelength scanning light source 101 that scans while changing the wavelength with time is employed.

  That is, the light output from the wavelength scanning light source 101 is input to the polarization controller 102 and the isolator 103 through the optical fiber, and then input to the first fiber coupler 104 through the optical fiber. In the first fiber coupler 104, for example, The reference light and the measurement light are demultiplexed and output at a ratio of 10:90. Of these, the reference light passes through the optical fiber and enters the collimator lens 112 and enters the delay line unit 113. The delay line unit 113 is a unit for adjusting the optical path length that aligns the reference optical path on the retina of the fundus and adjusts the measurement optical path length and the reference optical path length before measuring the OCT tomographic image.

The reference light radiated from the delay line unit 113 is input from the collimator lens 114 through the optical fiber to the polarization controller 115 and then input to the first input unit of the second fiber coupler 116 through the optical fiber.

On the other hand, the measurement light output from the first fiber coupler 104 is input to the collimator lens 105 through the optical fiber and input to the galvanometer mirror unit 106. The galvano mirror unit 106 is for scanning the measurement light, and the galvano driver 107 scans the measurement light in the horizontal direction and the vertical direction on the fundus of the eye to be inspected. .

The measurement light output from the galvanometer mirror unit 106 passes through the lens 108, exits from an inspection window (not shown) through the objective lens 109, and enters the eye E to be examined. The measurement light incident on the eye E is reflected by each tissue part (retina, choroid, etc.) of the fundus Er, and the reflected light is incident from the examination window. 108, the light is input to the collimator lens 105 through the galvanometer mirror unit 106. Then, the reflected light passes through the first fiber coupler 104 through the optical fiber, and then is input to the second input unit of the second fiber coupler 116 through the optical fiber.

In the second fiber coupler 116, the reflected light from the fundus Er and the reference light input through the optical fiber are combined at a ratio of, for example, 50:50, and the signal is transmitted through the optical fiber. Are input to the differential amplification detector 117. In the detector 117, interference for each wavelength is measured, and the measured interference signal is input to the AD board 201 provided in the control device 200. Further, the arithmetic unit 202 provided in the control device 200 performs processing such as Fourier transform on the interference signal, and thus obtains a tomographic image of the fundus retina Er along the scanning line. (Figure 2)

  At this time, as will be described in detail later, the scan pattern of the measurement light by the galvano mirror unit 106, in other words, the direction of the scanning line (B-scan) is set in the control device 200. The galvano driver 107 controls the galvanometer mirror unit 106 based on a command signal from the control device 200 (arithmetic unit 202). The obtained tomographic image data of the fundus oculi Er is stored in the storage unit 203. (Figure 2)

  Next, a method for acquiring a three-dimensional tomographic image will be described with reference to FIGS. FIG. 3A shows a series of processing flow of the tomographic imaging apparatus of the present invention from B-scan image capturing to 3D image construction.

  Before starting imaging with OCT, the reference mirror in the delay line unit 113 is moved to match the measurement optical path length with the reference optical path length (step 301). Thereafter, the tomographic image acquisition unit 100 acquires a tomographic image (B-scan image) (step 302).

  FIG. 4 shows a state in which a tomographic image (B-scan image) is acquired by the tomographic image acquisition unit 100. 4A shows an example of the fundus retina of the eye E, and FIG. 4B shows a plurality of two-dimensional tomograms (B-scan images) of the fundus retina 401 obtained from the tomogram acquisition unit 100. An example is shown. FIG. 4C shows an example of a three-dimensional tomographic image of the fundus generated in the present embodiment. 4A to 4C, the x-axis indicates the B-scan scan direction, and the y-axis indicates the C-scan direction. Furthermore, the z-axis in FIGS. 4B and 4C indicates the depth direction of the A-scan signal, that is, the depth direction of the fundus.

  Reference numeral 404 in FIG. 4B denotes an acquired two-dimensional tomogram, which is created by the calculation unit 202 reconstructing the A-scan signal 403 while scanning the galvanometer mirror unit 106 in the x direction. This two-dimensional tomographic image is a B-scan image, and is defined by a two-dimensional cross section in the x direction perpendicular to the depth direction (z direction) with respect to the fundus retina 401, that is, the x axis and z axis in FIG. It is a two-dimensional tomographic image in a plane. Reference numeral 402 in FIG. 4A indicates the photographing position of the two-dimensional tomographic image 404.

  Further, the B-scan image at the shooting position 402 is acquired by successively shifting the shooting position 402 of the B-scan image in the y-axis direction (step 303). When acquisition of a predetermined number of B-scan images is completed (step 304), the B-scan image is subjected to addition averaging processing with one or more adjacent B-scan images (step 305), and thereafter , B-scan images are superimposed and reconstructed into a three-dimensional tomographic image (step 306). FIG. 4B shows a state in which a plurality of B-scan images at each imaging position are acquired. FIG. 4C shows a three-dimensional tomogram (volume) generated from a plurality of B-scan images. Data).

  The three-dimensional tomogram (volume data) generated in the tomogram acquisition unit 100 is stored in the storage unit 203 (step 307).

  Next, an operation method of the first selection means and the second selection means, which is the main configuration of the present invention, will be described with reference to FIGS.

  First, a fundus front image (fundus retina image) is displayed on the monitor (step 311). Reference numeral 501 in FIG. 5A denotes a fundus front image. The fundus front image may be an image taken by a separately provided fundus camera or SLO (Scanning Laser Ophthalmoscope), or may be an image generated from a three-dimensional tomogram (volume data) generated by the tomogram acquisition unit 100.

  An arrow mark 502 and a straight line 503 are displays for operation for selecting a B-scan image to be displayed on the monitor, and the arrow mark 502 is moved in the Y direction (vertical direction on the monitor) on the monitor using a mouse or the like. As it moves, the straight line 503 also moves in the Y direction, and the position of the B-scan image to be displayed on the monitor is selected (step 312).

  When the arrow mark 502 is set, a B-scan image 504 at the set position is displayed as shown in FIG. 5B (step 313). The B-scan image 504 may be displayed by setting the position with the arrow mark 502 and then performing another operation (for example, clicking a mouse button). First, in conjunction with the position indicated by the arrow mark 502 and the straight line 503, the B-scan image 504 at that position is automatically extracted from the three-dimensional tomographic image (volume data) and is always displayed on the monitor. Good.

  The B-scan image 504 is a schematic diagram of the B-scan image. For example, the boundary surface A indicates the outer boundary film, the boundary surface B indicates the IS-OS line, and the boundary surface C indicates the retinal pigment epithelium. In addition, the terminal end of the cone outer segment, CSI (boundary of the choroid and sclera), and the like exist, but are omitted in the schematic diagram.

  When the B-scan image 504 selected in step 312 is displayed, a cross mark 505 is displayed so as to overlap the B-scan image 504. The cross mark 505 is moved up and down and left and right with a mouse or the like, and the center of the cross mark 505 is selected by selecting the boundary surface to be used as a reference (step 314). For the selection operation, the mouse button may be clicked at the position where the cross mark 505 is set, or another operation button may be pressed. Further, as in step 312, the selection may be made using an arrow mark and a straight line. In addition, when the monitor adopts a touch panel method, a method of touching and selecting the reference boundary surface with a touch pen or an examiner's finger without displaying the cross mark 505 as described above is also adopted. it can.

  FIG. 6 shows a case where the boundary surface C (retinal pigment epithelium) is selected, for example. When the boundary surface C is selected by the cross mark 505 (FIG. 6A), the selected boundary surface C is traced on the B-scan image, and the trace line 601 is overlaid as shown in FIG. 6B. Are displayed (step 315). It is desirable to devise the trace line 601 so that the examiner can recognize it by changing the color, changing the thickness, or displaying it with a broken line. In this embodiment, the color is changed to display with double lines.

  When the boundary surface C (retinal pigment epithelium) is selected on the B-scan image 504, the process proceeds to a second selection for selecting an arbitrary depth position.

  FIG. 3C shows the operation flow of the second selection means of the present invention. When the first selection means selects a reference boundary surface, an arrow mark 701 and a straight line 702 are displayed again on the B-scan image 504 as shown in FIG. 7A (step 321). When the arrow mark 701 is moved in the vertical direction (Z direction) with a mouse or the like, the straight line 702 moves in conjunction with it.

  The arrow mark 701 is moved to the depth position of the fundus of the horizontal cross-sectional image to be observed, and the depth position is selected by clicking the mouse button (step 322). When selected, a trace line 704 having the same shape as the trace line 601 of the boundary surface selected in the first step is displayed at the selected depth position as shown in FIG. 7B, and the position of the horizontal sectional image to be observed is selected. The process ends (step 323).

  The operation method may not be the configuration of the arrow mark 701 and the straight line 702 as described in the first selection means. For example, a method of displaying and moving only the straight line 702 by a mouse operation or a keyboard operation may be used, and if the monitor adopts a touch panel method, it is displayed on the monitor with a touch pen or an examiner's finger. A method of touching and selecting a position to be observed on the B-scan image 504 can also be adopted.

  When the position of the horizontal sectional image to be observed is selected by the second selection means, the process proceeds to generation of a horizontal tomographic image. FIG. 9 shows a flow of a method for generating a horizontal sectional image according to the present invention.

  First, the data of the boundary surface selected by the first selection means, in this embodiment, the boundary surface C (retinal pigment epithelium) is extracted from the three-dimensional tomogram (volume data) using image processing (step 901). FIG. 8A shows the extraction of data of the boundary surface C selected from the volume data 801, and shows a three-dimensional plane image 802 of the boundary surface C generated based on the extracted data.

Three-dimensional position information P (n) = (Xn, Yn, Zn) in the volume data 801 of the three-dimensional plane image 802 is acquired (step 903).

Next, a difference ΔZ in coordinates in the Z direction between the boundary surface C selected by the first selection means and the trace line 704 selected and traced by the second selection means is calculated (step 904).

Then, the three-dimensional position information Q (n) of the horizontal cross-sectional image at the position selected by the second selection means is calculated by the following formula (step 905). Q (n) = (Xn, Yn, Zn-ΔZ)

FIG. 8B shows the three-dimensional plane image 802 of the boundary surface C selected by the first selection means that is the reference plane and the three-dimensional horizontal cross-sectional image at the position selected by the second selection means calculated in step 905. A planar image 803 is shown. That is, based on the three-dimensional position information calculated in step 905, data is extracted from the volume data 801 to generate a horizontal cross-sectional image of the target fundus depth position (step 906) and stored in the storage unit 203. To do. The stored horizontal sectional view is displayed on the monitor (step 907).

FIG. 10A shows a B-scan image (left figure) and a B-scan image when a horizontal cross-sectional image is extracted on the XY plane perpendicular to the Z direction as described in Patent Document 1. The horizontal cross-sectional image (right figure) in the line 1001 which selected the choroid position is shown. As can be seen from the image of the horizontal cross-sectional image of the choroid (right figure), the vascular structure of the choroid is displayed in the central part, but the vascular structure is not displayed in the peripheral part.

FIG. 10B shows a horizontal cross-sectional image (right figure) generated based on the embodiment of the present invention. A trace line 1002 shown in the B-scan image (left figure) is obtained by selecting the retinal pigment epithelium with the first selection means and tracing based on the selected retinal pigment epithelium. The right figure shows the horizontal cross-sectional image of the choroid in the trace line 1002 produced | generated according to the horizontal cross-sectional image production | generation method mentioned above based on the trace line 1002. FIG. It can be seen that the vascular structure of the choroid is displayed in detail up to the periphery.

As described above, according to the present embodiment, a horizontal cross-sectional image at an arbitrary depth position of the fundus can be acquired in detail up to the peripheral portion.

FIG. 11 is a display example of a monitor screen based on the embodiment of the present invention. Frontal fundus image (upper left), vertical B-scan image (upper right) in vertical line 1101 shown in the frontal fundus image, horizontal B-scan image (lower left) in horizontal line 1102 shown in the frontal fundus image, retinal pigment epithelium Four images of horizontal cross-sectional images (lower right) generated on the respective trace lines 1103 and 1104 at an arbitrary position of the choroid part with respect to the reference (position selected by the second selection means of the present invention) Is displayed.

Each B-scan image (upper right and lower left) can be selected by moving the positions of the vertical line 1101 and the horizontal line 1102 shown in the fundus front image.

In the above embodiment, the first selection means and the second selection means are manually selected by the examiner using a mouse or the like, but can also be automatically selected as described below.

FIG. 12A shows an arbitrary B-scan image 1201 selected by the first selection means, and FIG. 12B shows A-scan data in a broken line 1202 on the B-scan image 1201. .

A-scan data (for example, data in FIG. 12B) at a predetermined position of a preset B-scan image is measured, and the peak value at each boundary surface in each B-scan image is stored in the storage means. Remember. Then, the boundary surface can be selected by measuring A-scan data at the same place at the time of the first selection and detecting each peak value. For example, in the case of FIG. 12B, the apparatus can automatically select the boundary surface C if the peak value of the boundary surface to be selected is set to V4 or more.

The means for automating the first selection means is not limited to the method described above, and a method using image processing such as smoothing processing or binarization processing is also conceivable. A boundary surface existing in the B-scan image is extracted in advance by smoothing processing or binarization processing, an average value or total value of pixel values of each boundary surface is calculated, and the value is stored in a storage unit. Then, a B-scan image at the same position is selected at the time of the first selection, and an average value or total value of pixel values of each boundary surface is calculated by a smoothing process or a binarization process, thereby selecting a boundary surface. Is possible.

Also, if the second selecting means also sets the position from the boundary surface selected by the first selecting means in advance, the examiner selects the position using a mouse or the like as described above. Clearly, it is possible to automatically select the position and generate a horizontal cross-sectional image of the target depth of the fundus.

In the above embodiment, straight lines, arrow marks, and cross marks are employed on the monitor during the first selection means and the second selection means. Of course, other methods may be employed. For example, it is possible to directly move a straight line with a mouse without using an arrow mark, and it goes without saying that if the monitor is a touch panel, selection with a touch pen or an examiner's finger is also possible.

The embodiments of the present invention have been described in detail above. However, these are merely examples, and the present invention is not construed as being limited by specific descriptions in the embodiments. The present invention can be carried out in a mode in which various changes, modifications, improvements, etc. are added based on the knowledge, and such a mode is within the scope of the present invention as long as it does not depart from the gist of the present invention. It should be understood that it is included in.

As described above, according to the present embodiment, it is possible to take a single three-dimensional tomographic image at a desired depth position of the fundus in a short time without controlling a complicated optical system as in Patent Document 2. The horizontal cross-sectional image can be acquired in detail up to the peripheral portion.

100 .. Tomographic image acquisition unit 101. Wavelength scanning light source 104. First fiber coupler 106. Galvano mirror unit 113. Delay line unit 116. Second fiber coupler 201. AD board 202. Calculation Part 203 .. storage part

Claims (10)

In a tomographic imaging apparatus that creates a three-dimensional tomographic image by superimposing a plurality of B-scan images (tomographic images),
First selection means for selecting one boundary surface (layer) from a plurality of boundary surfaces (layers) included in one B-scan image of the plurality of B-scan images;
Second selection means for selecting a position in the depth direction (Z direction) for the one B-scan image;
A horizontal cross-sectional image generating means for generating a horizontal cross-sectional image at a depth selected by the second selecting means on the basis of a three-dimensional planar shape of the boundary surface selected by the first selecting means; Tomography device. The tomography apparatus according to claim 1, further comprising display means for displaying a horizontal cross-sectional image generated by the horizontal cross-sectional image generating means. 2. The horizontal cross-sectional image (three-dimensional plane image) of the reference boundary surface selected by the first selection unit is generated by reprocessing the data of the three-dimensional tomographic image. The tomography apparatus according to 2. The tomography apparatus according to claim 3, wherein the horizontal cross-sectional image generating unit reprocesses and generates the data of the three-dimensional tomographic image. The tomography apparatus according to any one of claims 1 to 4, wherein the three-dimensional tomographic image is a tomographic image of a fundus including a retina of an eye to be examined. The boundary surface selected by the first selection means is one of an outer boundary membrane, an IS-OS line, a terminal end of a cone outer segment, a CSI (boundary of the choroid and sclera), and a retinal pigment epithelium layer. The tomography apparatus according to claim 5, wherein the tomography apparatus is provided. The tomography apparatus according to claim 6, wherein the position selected by the second selection unit is an arbitrary depth position in the choroid. The first selection means selects one boundary surface based on a predetermined condition from data of peak values of a plurality of boundary surfaces in A-scan data at a predetermined position of the one B-scan image. The tomography apparatus according to claim 1, wherein: The first selection means processes the one B-scan image by a smoothing process or the like, traces a plurality of boundary surfaces from the result, and based on a predetermined condition from the traced boundary surfaces The tomography apparatus according to claim 1, wherein one boundary surface is selected. The second selection unit selects a position determined based on a preset condition between another boundary surface adjacent to the boundary surface selected by the first selection unit. Item 10. The tomography apparatus according to any one of Items 1 to 9.