JP6179942B2 - Scanning confocal endoscope system - Google Patents

Description

  The present invention relates to an image distortion correction apparatus and an image distortion correction method for correcting distortion of an image of a subject scanned in a spiral shape.

  A scanning image system that images a signal obtained by periodically scanning a subject is known (for example, Patent Document 1). This type of scanning image system includes a scanning probe having a single-mode optical fiber and a biaxial actuator. Inside the scanning probe, the optical fiber is held in a cantilever shape by a biaxial actuator at the base end. The biaxial actuator modulates and amplifies the amplitude of vibration while vibrating the tip of the optical fiber two-dimensionally at a predetermined frequency (for example, two-dimensionally resonating at the natural frequency), thereby Is spirally moved on a predetermined surface. Thereby, the scanning light transmitted from the light source by the optical fiber scans the subject in a spiral shape. The scanning image system detects return light from a subject scanned with scanning light, generates an image of a scanning area from the detected return light, and displays the generated image on a display screen of a monitor.

  More specifically, this type of scanning image system detects return light from a subject scanned with scanning light as a signal that forms subject information. The scanning image system assigns a pixel position in the two-dimensional pixel array having a positional relationship corresponding to the scanning position in the scanning region in accordance with the detection timing of the signal to the detected signal. In the scanning image system, subject information by each signal is arranged and imaged at assigned pixel positions, and the imaged subject is displayed on a display screen of a monitor.

  In this type of scanning image system, unlike general electronic scopes equipped with a CCD (Charge Coupled Device) image sensor or the like, image distortion due to scanning errors becomes a problem. In order to suppress the distortion of the image due to the scanning error, for example, a method of accurately matching the position of the scanning light corresponding to each pixel position with the display coordinate system of the monitor can be considered. Patent Document 1 describes a specific example of this method.

  In Patent Document 1, the locus of scanning light is calibrated using a calibration jig. In Patent Document 1, a scanning light locus is detected by causing a PSD (Position Sensitive Detector) of a calibration jig to receive scanning light emitted from an optical fiber, and the difference between the detected locus and an ideal locus is detected. Is converged within a predetermined tolerance by adjusting various parameters such as the amplitude, phase, and frequency of the voltage applied to the biaxial actuator. When the difference between the measured trajectory and the ideal trajectory converges within a predetermined tolerance, the position of the scanning light corresponding to each pixel position accurately matches the display coordinate system of the monitor.

Special table 2008-514342 gazette

  The scanning image system exemplified in Patent Document 1 does not perform simple scanning such as raster scanning, but rotates a point light source (tip of an optical fiber) on a predetermined surface to make a subject spiral. It has a complicated configuration of scanning. Therefore, the image distortion due to the scanning error is not a simple distortion but a complicated distortion caused by a combination of a plurality of factors. In Patent Document 1, it is considered that various parameters are adjusted to a dark cloud for a scanning image system having such a characteristic that an image is complicatedly distorted during calibration. Therefore, in the method for correcting image distortion described in Patent Document 1, a considerable number of parameter adjustments and scanning trajectory detections are performed until the difference between the actual trajectory and the ideal trajectory converges within a predetermined tolerance. Need to repeat. Therefore, the method described in Patent Document 1 points out a problem that it takes time for calibration.

  The present invention has been made in view of the above circumstances, and an object of the present invention is to provide an image distortion correction apparatus and an image distortion correction method capable of correcting image distortion by processing with a light temporal load. It is to be.

  In the image distortion correction apparatus according to the present embodiment, a plurality of indexes scanned in a spiral shape with a predetermined rotation period from one of the center and the periphery to the other within a substantially circular scanning region by a predetermined scanning light. The image signal acquisition means for acquiring the image signal of the object detected by receiving the return light from the object arranged in the position corresponding to the scanning position in the scanning area by the scanning light with respect to the acquired image signal Image generation means for generating a subject image by assigning pixel positions in a two-dimensional pixel array in a positional relationship according to the detection timing of an image signal, and arranging each image signal at the assigned pixel position; An index position detecting means for detecting a pixel position of each index in the generated image, and an index for creating an index image array in which a plurality of indices are arranged based on the detected pixel position of each index It compares the image sequence creation means, an index image array created with predetermined ideal sequence, and a distortion correction means for correcting the distortion of each index in the image based on the comparison result.

  By using the image distortion correction apparatus according to the present embodiment, the distortion of the subject image is corrected in units of fine areas corresponding to at least the index arrangement interval. Therefore, the subject image is corrected with high accuracy even if it has complicated distortion. Further, the distortion is corrected based on a comparison result between the index image array and the ideal array. In the present embodiment, unlike the method described in Patent Document 1, it is not necessary to repeat the calibration work, so the time load required for distortion correction is light.

  Further, the image distortion correction apparatus defines, for each index in the image, a straight line having the index as a start point and having at least one adjacent index positioned in a predetermined direction with respect to the index as an end point, and the length of the defined straight line Length slope calculating means for calculating the length and slope, and the distribution of the number of appearances of the length peak and the slope of the straight line when the number of appearances of the straight line length calculated for each index is distributed Peak detecting means for detecting an inclination peak, and an ideal arrangement creating means for creating an ideal arrangement by arranging a plurality of indexes at an arrangement interval having a length corresponding to the length peak in the direction of the inclination corresponding to the inclination peak. It is good also as a structure.

The index image array creation means
(1a) An attention index is determined from the index image array,
(2a) assigning array position information in the index image array to the determined attention index,
(3a) Define at least one reference line having a predetermined length and a predetermined inclination starting from the attention index to which the array position information is assigned, and position within a predetermined range including the end point of the defined reference line Assign the sequence position information adjacent to the target index to the index to be
(4a) Repeat (1a) to (3a) until sequence position information is assigned to all indices.
It is good also as a structure.

The ideal array creation means is
(1b) determining an attention index from within the ideal sequence;
(2b) assigning array position information corresponding to the index image array to the determined attention index,
(3b) Define at least one reference line having a predetermined length and a predetermined inclination starting from the attention index to which the array position information is assigned, and position within a predetermined range including the end point of the defined reference line Assign the sequence position information adjacent to the target index to the index to be
(4b) Repeat (1b) to (3b) until array position information corresponding to the index image array is assigned to all indices.
It is good also as a structure.

  The distortion correction means calculates a correction vector for each index in the index image array based on the difference in pixel position from the index in the ideal array to which the corresponding array position information is assigned, and based on the calculated correction vector The distortion of each index may be corrected.

  The subject is, for example, a white square cell arranged in a matrix on a black background.

  In the image distortion correction method according to the present embodiment, a plurality of indexes scanned in a spiral shape with a predetermined rotation period from one of the center and the periphery to the other within a substantially circular scanning region by a predetermined scanning light are regularly arranged. An image signal acquisition step for acquiring an image signal of the subject detected by receiving the return light from the subject arranged in the array, and the acquired image signal corresponds to the scanning position in the scanning region by the scanning light. An image generation step for generating a subject image by allocating pixel positions in a two-dimensional pixel array in a positional relationship according to the detection timing of an image signal, and arranging each image signal at the assigned pixel position, and generation An index position detecting step for detecting a pixel position of each index in the image, and an index image array in which a plurality of indices are arranged based on the detected pixel position of each index And index image array generating step of forming, by comparing the index image array created with predetermined ideal sequence, and a distortion correction step of correcting the distortion of each index in the image based on the comparison result.

  According to the present embodiment, an image distortion correction apparatus and an image distortion correction method capable of correcting image distortion by processing with a light time load are provided.

It is a block diagram which shows the structure of the scanning confocal endoscope system of embodiment of this invention. FIG. 2A schematically shows the configuration of the confocal optical unit according to the embodiment of the present invention (FIG. 2A), and FIG. 2B schematically shows the configuration of the biaxial actuator provided in the confocal optical unit. )). The rotation locus | trajectory of the front-end | tip of the optical fiber of a confocal optical unit on an XY approximate surface is shown. FIG. 4A is a side sectional view showing the internal configuration of the jig according to the embodiment of the present invention (FIG. 4A), a diagram showing a chart (FIG. 4B), and a diagram showing a photographic chart image (FIG. 4C). is there. It is a figure which shows the flowchart of the cell center coordinate collection process which collects the center coordinate of the cell reflected on the imaging | photography chart image. FIG. 6 is an explanatory auxiliary diagram for assisting the description of the cell center coordinate collection processing of FIG. 5. It is a figure which shows the flowchart of the grid labeling process which labels a grid with which the center coordinate was extract | collected. FIG. 8 is an explanatory auxiliary diagram for assisting the description of the grid labeling process of FIG. 7. It is a figure which shows the flowchart of the distortion correction table creation process which produces the distortion correction table corresponding to all the pixels in the chart image containing the pixel of the labeled center coordinate. FIG. 10 is an explanatory auxiliary diagram for assisting the description of the distortion correction table creation processing of FIG. 9. It is a conceptual diagram of a distortion correction vector. It is a figure which shows the flowchart of the image distortion correction process performed using a distortion correction table. It is a comparison figure which compares the distortion in an imaging | photography chart image.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. In the following, a scanning confocal endoscope system will be described as an example of an embodiment of the present invention.

[Configuration of Scanning Confocal Endoscopy System 1]
FIG. 1 is a block diagram illustrating a configuration of a scanning confocal endoscope system 1 according to the present embodiment. The scanning confocal endoscope system 1 according to the present embodiment is a system designed by applying the principle of a confocal microscope, and is preferably configured to observe a subject with high magnification and high resolution. As shown in FIG. 1, the scanning confocal endoscope system 1 includes a system main body 100, a confocal probe 200, and a monitor 300. The confocal observation using the scanning confocal endoscope system 1 is performed in a state where the distal end surface of the flexible tubular confocal probe 200 is applied to the subject.

  The system body 100 includes a light source 102, an optical demultiplexer / multiplexer (photocoupler) 104, a damper 106, a CPU 108, a CPU memory 110, an optical fiber 112, a light receiver 114, a video signal processing circuit 116, an image memory 118, and a video signal output. A circuit 120, a laser control circuit 122, a remapping table memory 124, an operation panel 126, and an A / D conversion circuit 128 are included. The confocal probe 200 includes an optical fiber 202, a confocal optical unit 204, a sub CPU 206, a sub memory 208, and a scanning driver 210.

  The light source 102 emits excitation light (continuous light or pulsed light) that excites the fluorescent substance contained in the medicine administered into the body cavity of the patient in accordance with drive control by the laser control circuit 122. The excitation light emitted from the light source 102 enters the optical demultiplexer / multiplexer 104. An optical connector 152 is coupled to one of the ports of the optical demultiplexer / multiplexer 104. The unnecessary port of the optical demultiplexer-multiplexer 104 is coupled to a damper 106 that terminates the excitation light emitted from the light source 102 without reflection. The excitation light incident on the former port is incident on the optical system disposed in the confocal probe 200 via the optical connector 152.

  The proximal end of the optical fiber 202 is optically coupled to the optical demultiplexer / multiplexer 104 through the optical connector 152. The distal end portion of the optical fiber 202 is housed in a confocal optical unit 204 incorporated in the distal end portion of the confocal probe 200. When the excitation light emitted from the optical demultiplexer / multiplexer 104 is incident on the proximal end of the optical fiber 202 via the optical connector 152, it is transmitted through the optical fiber 202.

  FIG. 2A is a diagram schematically showing the configuration of the confocal optical unit 204. Hereinafter, for convenience of describing the confocal optical unit 204, the axial direction (longitudinal direction) of the confocal optical unit 204 is defined as the Z direction, and two directions orthogonal to the Z direction and orthogonal to each other are respectively defined as the X direction and the Y direction. It is defined as

  As shown in FIG. 2A, the confocal optical unit 204 has a metal outer cylinder 204A that houses various components. The outer cylinder 204A holds an inner cylinder 204B having an outer wall surface shape corresponding to the inner wall surface shape of the outer cylinder 204A so as to be slidable coaxially (Z direction). The distal end 202a of the optical fiber 202 is accommodated and supported by the inner cylinder 204B through openings formed in the base end surfaces of the outer cylinder 204A and the inner cylinder 204B. The distal end 202 a functions as a secondary point light source of the scanning confocal endoscope system 1 that emits excitation light transmitted through the optical fiber 202. The position of the tip 202a, which is a point light source, periodically changes based on control by the CPU. In FIG. 2A, the central axis AX indicates the central axis of the confocal optical unit 204. When the optical fiber 202 is in the initial position (when the optical fiber 202 is not vibrated), the center axis AX and the axis of the optical fiber 202 coincide.

  The sub memory 208 stores probe information such as identification information and various properties of the confocal probe 200. The sub CPU 206 reads the probe information from the sub memory 208 when the system is activated, and transmits the read probe information to the CPU 108 via the electrical connector 154 that electrically connects the system main body 100 and the confocal probe 200. The CPU 108 stores the probe information received from the sub CPU 206 in the CPU memory 110. The CPU 108 reads the probe information stored in the CPU memory 110 when necessary, generates a control signal necessary for controlling the confocal probe 200, and transmits the generated control signal to the sub CPU 206. The sub CPU 206 designates necessary setting values for the scan driver 210 in accordance with the control signal received from the CPU 108.

  A biaxial actuator 204C is bonded and fixed to the outer peripheral surface of the optical fiber 202 near the tip 202a. FIG. 2B is a diagram schematically showing the configuration of the biaxial actuator 204C. As shown in FIG. 2B, the biaxial actuator 204C includes a pair of X-axis electrodes (“X” and “X ′” in the figure) and Y-axis electrodes (in the figure) connected to the scanning driver 210. “Y”, “Y ′”) are piezoelectric actuators formed on a piezoelectric body.

  The scan driver 210 generates a drive signal corresponding to the setting value designated by the sub CPU 206, and drives and controls the biaxial actuator 204C with the generated drive signal. More specifically, the scanning driver 210 applies the AC voltage X between the X-axis electrodes of the biaxial actuator 204C to vibrate the piezoelectric body in the X direction, and has the same frequency as the AC voltage X and the phase is orthogonal. An alternating voltage Y is applied between the Y-axis electrodes to vibrate the piezoelectric body in the Y direction. The AC voltages X and Y are respectively defined as voltages that increase linearly in proportion to time and reach effective values (X) and (Y) over time (X) and (Y). The tip 202a of the optical fiber 202 is on a surface that approximates the XY plane (hereinafter referred to as "XY approximate surface") by combining the kinetic energy in the X and Y directions by the biaxial actuator 204C. In FIG. 5, the coil moves so as to draw a spiral pattern around the central axis AX. The movement trajectory of the tip 202a increases in proportion to the applied voltage, and draws a circular trajectory having the largest diameter when an AC voltage having effective values (X) and (Y) is applied. FIG. 3 shows the rotation locus of the tip 202a on the XY approximate plane.

  The excitation light is emitted from the tip 202a of the optical fiber 202 during the period from the start of application of the alternating voltage to the biaxial actuator 204C to the stop of application. Hereinafter, for convenience of explanation, this period is referred to as a “sampling period”. The excitation light emitted from the tip 202a scans a predetermined substantially circular scanning region centered on the central axis AX in a spiral shape by moving the tip 202a spirally on the XY approximate plane during the sampling period. .

  When the application of the AC voltage to the biaxial actuator 204C is stopped after the sampling period has elapsed, the vibration of the optical fiber 202 is attenuated. The movement of the tip 202a of the optical fiber 202 on the approximate XY plane converges as the vibration of the optical fiber 202 is attenuated, and becomes substantially zero after a predetermined time (that is, the tip 202a almost stops on the central axis AX). ). Hereinafter, for convenience of explanation, a period from the end of the sampling period until the tip 202a substantially stops on the central axis AX is referred to as a “braking period”. After the braking period, the next sampling period is started after waiting for a predetermined time. Hereinafter, for convenience of description, a period from the end of the braking period to the start of the next sampling period is referred to as a “settling period”. The settling period is a waiting time for completely stopping the tip 202a on the central axis AX. By setting the settling time, the movement locus of the tip 202a can be stabilized. By stabilizing the movement trajectory of the tip 202a, it is possible to guarantee the scanning accuracy for the subject. In addition, the period corresponding to one frame includes a sampling period, a braking period, and a settling period. The frame rate can be flexibly changed by adjusting the settling period. Thus, the settling period can be set as appropriate from the relationship between the time until the tip 202a completely stops and the frame rate. In order to shorten the braking period, a braking torque may be positively applied by applying a reverse phase voltage to the biaxial actuator 204C in the initial stage of the braking period.

  An objective optical system 204D is installed in front of the tip 202a of the optical fiber 202. The objective optical system 204D is composed of a plurality of optical lenses, and is held by the outer cylinder 204A via a lens frame (not shown). The optical axis of the objective optical system 204A coincides with the central axis AX. The lens frame is fixed and supported relative to the inner cylinder 204B inside the outer cylinder 204A. Therefore, the optical lens group held by the lens frame slides in the Z direction integrally with the inner cylinder 204B inside the outer cylinder 204A.

  A compression coil spring 204E and a shape memory alloy 204F are attached between the base end surface of the inner cylinder 204B and the inner wall surface of the outer cylinder 204A. The compression coil spring 204E is initially compressed and clamped in the Z direction from the natural length. The shape memory alloy 204F has a long bar shape in the Z direction, deforms when an external force is applied at room temperature, and has a property of restoring to a predetermined shape by a shape memory effect when heated to a certain temperature or higher. ing. The shape memory alloy 204F is designed such that the restoring force due to the shape memory effect is larger than the restoring force of the compression coil spring 204E. The scan driver 210 generates a drive signal corresponding to the set value specified by the sub CPU 206, and controls the amount of expansion / contraction of the shape memory alloy 204F by energizing and heating the shape memory alloy 204F with the generated drive signal. . The shape memory alloy 204F advances and retracts the inner tube 204B in the Z direction together with the optical fiber 202 according to the amount of expansion and contraction. Specifically, the shape memory alloy 204F is heated and extends (restores) in the Z direction, thereby pushing the inner tube 204B forward (Z direction) together with the optical fiber 202. The shape memory alloy 204F is also compressed in the Z direction by the compression coil spring 204E as the restoring force due to the shape memory effect decreases as the slow cooling progresses, and the inner tube 204B is moved backward along the optical fiber 202 (Z direction). Withdraw.

  The excitation light emitted from the tip 202a of the optical fiber 202 passes through the objective optical system 204D and forms a spot on the surface or surface layer of the subject. The spot forming position is displaced in the Z direction in accordance with the advance / retreat of the tip 202a which is a point light source. That is, the confocal optical unit 204 scans the subject three-dimensionally by combining the periodic circular motion of the tip 202a on the XY approximate plane by the biaxial actuator 204C and the advance and retreat in the Z direction.

  Since the tip 202a of the optical fiber 202 is disposed at the front focal position of the objective optical system 204D, it functions as a confocal pinhole. Of the fluorescence emitted from the subject excited by the excitation light, only the fluorescence from the condensing point optically conjugate with the tip 202a is incident on the tip 202a. Fluorescence incident on the optical fiber 202 from the tip 202 a is transmitted through the optical fiber 202 and then incident on the optical demultiplexer-multiplexer 104 via the optical connector 152. The optical demultiplexer / multiplexer 104 separates the incident fluorescence from the excitation light emitted from the light source 102 and guides it to the optical fiber 112. The fluorescence is transmitted through the optical fiber 112 and detected by the light receiver 114. The photodetector 114 is a high-sensitivity photodetector such as a photomultiplier tube, for example, in order to detect weak light with low noise. The detection signal from the light receiver 114 is sampled and AD converted at a constant rate by a circuit not shown. The digital detection signal after AD conversion is image information forming a confocal image of the subject, and is hereinafter referred to as a “confocal image detection signal”. The confocal image detection signal is input to the video signal processing circuit 116.

  Here, when the position (trajectory) of the tip 202a of the optical fiber 202 during the sampling period is determined, the position (spot formation position) in the scanning region by the excitation light emitted when the tip 202a reaches a certain position, The signal acquisition timing (hereinafter referred to as “sampling point”) for obtaining the confocal image detection signal by receiving the return light (fluorescence) from the spot formation position by the light receiver 114 is almost uniquely determined. Therefore, in this embodiment, the correspondence relationship between the spot formation position and the sampling point is defined based on the design data, and then the position of the pixel corresponding to each sampling point (pixel address of each pixel forming the confocal image) ) Is defined. The remap table memory 124 stores the correspondence (remap table) between sampling points and pixel addresses. In the remapping table, corresponding sampling points are associated one-to-one with respect to each of all the pixel addresses forming the confocal image. The remap table converts temporal elements (sampling points) into spatial elements (pixel addresses). According to another expression, the first spatial element (spot formation position) is the first. This is converted into a second spatial element (pixel address).

  The video signal processing circuit 116 refers to the remapping table and assigns a pixel address to the confocal image detection signal obtained at each sampling point. Hereinafter, this assignment work is referred to as “remapping”. By performing the remapping, it is possible to generate a point image represented by each confocal image detection signal in a two-dimensional arrangement (confocal image) according to the pixel address.

  The confocal image detection signal after remapping (more precisely, after remapping described later) is buffered in the image memory 118 in units of frames. The buffered confocal image detection signal is swept from the image memory 118 to the video signal output circuit 120 at a predetermined timing, and conforms to a predetermined standard such as NTSC (National Television System Committee) or PAL (Phase Alternating Line). It is converted into a compliant video signal and output to the monitor 300. Thereby, a confocal image of the subject with high magnification and high resolution is displayed on the display screen of the monitor 300.

[Distortion correction for confocal images]
As described above, the scanning confocal endoscope system 1 according to the present embodiment spirals the subject by periodically moving the point light source (tip 202a of the optical fiber 202) in the circumferential direction on the XY approximate plane. It has a complicated configuration of scanning. Therefore, the image distortion due to the scanning error is not a simple distortion but a complicated distortion caused by a combination of a plurality of factors. In this embodiment, such a complicated distortion can be corrected by a process with a light time load.

[Shooting Chart]
The scanning confocal endoscope system 1 is set on a jig 400 to correct distortion of a confocal image. 4A is a side sectional view showing the internal configuration of the jig 400, and shows a state where the confocal probe 200 is set on the jig 400. FIG. As shown in FIG. 4A, the jig 400 has a box-like (for example, rectangular parallelepiped) case 402. The housing 402 has a plurality of wall portions including opposing wall portions 402a and 402b, and the plurality of wall portions define an internal space that is substantially shielded from the outside.

  A chart CHT is printed on the surface of the wall portion 402a on the inner space side. FIG. 4B shows a chart CHT. In the chart CHT, as shown in FIG. 4B, black grid lines (vertical lines and horizontal lines orthogonal to each other) are arranged at equal intervals on a white base so that white squares are in a matrix form. It has become a lineup. Note that a fluorescent agent corresponding to excitation light emitted from the light source 102 may be applied to the chart CHT.

  The housing 402 is formed with a holding portion 404 at a position substantially in the center of the wall portion 402b and facing the chart CHT. The holding portion 404 is a concave portion formed on the outer wall surface of the wall portion 402b, and an opening A is formed on a surface (bottom surface forming the concave portion) that faces the chart CHT. Further, the concave shape of the holding portion 404 corresponds to the outer shape of the distal end portion of the confocal probe 200. Therefore, the tip of the confocal probe 200 can be inserted into the holding unit 404 as shown in FIG. The holding unit 404 holds the tip of the inserted confocal probe 200. The tip of the confocal probe 200 held by the holding unit 404 has a posture in which the central axis AX (and the optical axis of the objective optical system 132D) is perpendicular to the chart CHT, and scans the scanning light on the chart CHT. It depends on where you can do it.

  In this imaging step, the relative position between the scanning confocal endoscope system 1 and the chart CHT and the imaging in the dark room are not necessarily required. In this imaging step, it is sufficient that the chart CHT can be imaged by the scanning confocal endoscope system 1 at a minimum. For example, the operator may simply attach a sheet on which the chart CHT is printed on the wall surface in the work chamber, and photograph it with the scanning confocal endoscope system 1. Thus, a large jig is not essential in the main photographing step.

  FIG. 4C shows an example of the chart CHT photographed by the scanning confocal endoscope system 1. As shown in FIG. 4C, the image of the captured chart CHT (hereinafter referred to as “photographed chart image”) is complicatedly distorted.

[Flow of collecting center coordinates of grid]
The system main body 100 collects the coordinates (pixel position) of the center of each cell appearing in the photographing chart image as information necessary for correcting the distortion of the photographing chart image. The collected center coordinates are handled as information indicating the position of the mesh. FIG. 5 shows a flowchart of the cell center coordinate collection process executed by the CPU 108 of the system main body 100. FIGS. 6A to 6D are explanatory assistance diagrams for assisting the explanation of the cell center coordinate collection process.

[S11 in FIG. 5 (binarization of photographing chart image)]
In this processing step S11, noise in the photographic chart image is removed by binarizing the photographic chart image. The threshold value for the binarization process is set to an appropriate value for determining the cell as white and determining the grid line as black. The set value is a fixed value determined empirically, for example. Also, an appropriate threshold value for separating the two from the luminance value of the grid and the luminance value of the grid line may be automatically determined.

[S12 in FIG. 5 (shrinkage of the cell image)]
In this processing step S12, by applying an erosion to all the cells in the photographic chart image, the interval between the adjacent cells is increased (see FIG. 6A). Specifically, when there is even a black pixel around the pixel of interest, the process of replacing it with black (Erosion) is repeated a predetermined number of times. By expanding the interval between adjacent cells, the center coordinates of the cells can be detected without being affected by the adjacent cells.

[S13 in FIG. 5 (search for cells)]
In this processing step S13, a square window W having a predetermined size is arranged at a predetermined position in the photographing chart image. The window W defines a calculation area of the center of gravity G of the photographing chart image, and is larger than the cell after the erosion (Erosion). When the window W is arranged at a predetermined position, it is determined whether or not the window W includes a cell (that is, a white pixel). In this flowchart, when a cell (white pixel) is included in the window W (S13 in FIG. 5: YES), the process proceeds to processing step S16 (detection of the cell center). When the grid (white pixel) is not included in the window W (S13 in FIG. 5: NO), the process proceeds to the processing step S14 (scanning completion determination) in FIG.

[S14 in FIG. 5 (scanning completion determination)]
In this processing step S14, it is determined whether or not all areas in the photographic chart image have been scanned by the window W. This flowchart ends when all the areas in the photographing chart image are scanned by the window W (S14 in FIG. 5: YES). When the area not scanned by the window W remains (S14 in FIG. 5: NO), the process proceeds to the processing step S15 (scanning of the photographing chart image by the window W).

[S15 in FIG. 5 (scanning of photographing chart image by window W)]
In this processing step S15, the window W is scanned by one pixel (that is, the position of the window W is shifted by one pixel). In this flowchart, after the window W is scanned, the processing returns to the processing step S13 (search for cell) in FIG. In process step S13 (search for cell), it is determined whether or not the cell (white pixel) is included in the window W after one pixel scan.

[S16 in FIG. 5 (Detection of center coordinates of cell)]
In this processing step S16, the center coordinates of the cells are detected using the average value shift method (Mean-Shift). First, the center of gravity G of the area in the photographing chart image surrounded by the window W is calculated. The calculation of the center of gravity G is repeated until the center C and the center of gravity G of the window W coincide with each other, as shown in FIGS. Here, the squares are white squares surrounded by black grid lines. Therefore, the center of gravity G coincides with the center of the cell when the center of gravity calculation is repeated and converges to a position that coincides with the center C (see FIG. 6D). For this reason, the coordinates of the center C when the center of gravity G converges to a position that coincides with the center C are stored in the CPU memory 110 as the center coordinates of the cells.

[S17 in FIG. 5 (Pseudo Erase)]
In this processing step S17, for the sake of convenience of processing, the cell in which the center coordinates are stored is erased in a pseudo manner.

  The flowchart then proceeds to processing step S14 (scanning completion determination) in the figure, and when an area not scanned by the window W remains (S14: NO in FIG. 5), processing step S15 (in FIG. 5) ( After execution of the scanning of the captured chart image by the window W), the processing returns to the processing step S13 (retrieval search) in FIG. In process step S13 (search for cell), it is determined whether or not the cell (white pixel) is included in the window W after one pixel scan. Since the cells surrounded by the window W are deleted, the cells (white pixels) are not included in the window W after one pixel scan. Therefore, the processing step S14 (scanning completion determination), the processing step S15 (scanning of the photographed chart image by the window W) and the processing step S13 (searching for the cell) are repeated until the next cell is detected.

  By executing the processing of the flowchart of FIG. 5, the center coordinates of all the cells included in the photographing chart image are collected.

[Masame Labeling Flow]
The system main body 100 performs labeling on the grid as a process necessary for correcting the distortion of the photographing chart image. FIG. 7 shows a flowchart of the cell labeling process executed by the CPU 108 of the system main body 100. FIGS. 8A to 8D are explanatory assistance diagrams for assisting the explanation of the cell labeling process.

[S21 in FIG. 7 (Determination of Attention Center Coordinates)]
In this processing step S21, an array in which the center coordinates (pixel positions) of all cells stored in the CPU memory 110 are arrayed (hereinafter referred to as “center coordinate array”) is scanned, and the cells detected by the scanning are scanned. Is determined as the center coordinate of interest.

[S22 in FIG. 7 (Calculation of positional relationship of center coordinates)]
In this processing step S22, a straight line connecting the neighboring center coordinates close to the attention center coordinates and the attention center coordinates in the center coordinate array is defined, and the length and inclination of the defined straight line are calculated. The neighborhood center coordinates are typically the center coordinates of a grid adjacent to the grid of the target center coordinates. Therefore, the calculated straight line length and inclination represent the positional relationship (arrangement interval, alignment direction) between adjacent cells. Here, the straight line has the center point of interest as the start point and the center point of the neighborhood as the end point. Since the slope of the straight line is the rate of change from the start point to the end point, it is calculated based on the center coordinate of interest that is the start point.

  The straight line may be drawn between one neighboring center coordinate located in a prescribed direction with respect to the center coordinate of interest. As an example of the position in the specified direction, there is a position in the −Y direction with respect to the center coordinate of interest, where the X coordinate is not changed and only the Y coordinate is changed. Hereinafter, the position in this direction is referred to as “−Y direction position”. The same applies to positions in other directions. In the present embodiment, in order to improve the accuracy of distortion correction, the straight line is between four neighboring center coordinates at the + Y direction position, the −Y direction position, the + X direction position, and the −X direction position with respect to the center coordinate of interest. Be drawn to. In order to further improve the accuracy of distortion correction, a straight line is drawn between the eight neighboring central coordinates including the diagonal position (position in the direction of 45 ° or −45 ° with respect to the X direction or the Y direction). Also good.

  Taking the case of specifying the near center coordinates of the position in the -Y direction as an example, the description of this processing step S22 will be supplemented. Since the photographed chart image has distortion, as shown in FIG. 8A, there is almost no case where there is a neighborhood center coordinate at the position in the −Y direction. Therefore, the neighborhood closest to the −Y direction position among a plurality of neighboring center coordinates on the Y direction position side (a position on the −Y direction side with respect to the center coordinate of interest and a position where not only the Y coordinate but also the X coordinate changes) The center coordinate is specified as the vicinity center coordinate of the position in the -Y direction.

[S23 in FIG. 7 (scan completion determination)]
In this processing step S23, it is determined whether or not all center coordinates in the center coordinate array have been detected as the target center coordinates. This flowchart repeats the loop from processing step S21 (determination of attention center coordinates) to this processing step S23 in FIG. 7 until all center coordinates in the center coordinate array are detected as the attention center coordinates (FIG. 7). S21, S22, S23: NO). In this flowchart, when all the center coordinates in the center coordinate array are detected as the center coordinates of interest (S23 in FIG. 7: YES), the process goes out of the loop and processing step S24 (determination of the reference length L) in FIG. Proceed to

[S24 in FIG. 7 (Determination of Reference Length L)]
In this processing step S24, the lengths of all the straight lines calculated in the processing step S22 (calculation of the positional relationship of the center coordinates) in the figure are histogrammed. That is, a statistical graph is created in which the length of a straight line is taken on the horizontal axis and the number of appearances of the length of the straight line is taken on the vertical axis. Next, in this processing step S24, by detecting the peak from the created histogram, the length of the straight line corresponding to the peak (that is, having the highest number of appearances) is specified, and the specified length is the reference length L. Defined.

  Although the distortion of the photographic chart image due to the scanning error is inevitable, the ratio to the entire image is not high. For this reason, in the present embodiment, the reference length L having the highest number of appearances is treated as approximating the ideal arrangement interval (an arrangement interval in a state where there is no distortion of the captured chart image due to a scanning error).

[(Determination of the reference angle theta _h and theta _v) S25 of FIG. 7
Since the slope takes a value of ∞ at the maximum, it is often inconvenient in processing. In this processing step S25, for the convenience of the processing, the inclinations of all straight lines calculated in the processing step S22 (calculation of the positional relationship of the center coordinates) in the figure are converted into angles (−π to π). In this embodiment, the angle is based on the axis in the + X direction (= 0), takes a value of 0 to π in the range of the first quadrant and the second quadrant, and in the third quadrant and the fourth quadrant. Take a value between 0 and -π in the range. The angle π and the angle −π are the same for convenience of processing. The angle of the straight line after conversion is converted into a histogram. That is, a statistical graph is created in which the angle of a straight line is taken on the horizontal axis and the number of appearances of the angle of the straight line is taken on the vertical axis.

As shown in FIG. 8A, the angle of the straight line is close to the horizontal direction connecting the center coordinates of interest and each of the vicinity center coordinates of the ± X direction positions, and Many things close to the vertical direction connecting each of the neighboring center coordinates appear. For this reason, as shown in FIG. 8B, the peak corresponding to the former and the peak corresponding to the latter appear in the histogram. Angle corresponding to the former peak is defined as a reference angle theta _h, an angle corresponding to the latter peak is defined as the reference angle theta _v.

The reference angle θ _h and the reference angle θ _v are ideally zero and π / 2, respectively. However, since the photographic chart image has a larger or smaller distortion, it does not match the ideal value. Therefore, in the simple condition such that the angle the larger the reference angle theta _h Toshikatsu a smaller angle with the reference angle theta _v, is a concern that decision errors may occur. Therefore, in the present embodiment, the following conditions (1) and (2) are introduced to improve the determination accuracy.
(1) The angle corresponding to two peaks is | π / 4 | or more apart.
(2) the angle | π / 4 | and than the smaller reference angle theta _h, to the others with reference angle theta _v.

In the present embodiment, for the same reason as the reference length L, the inclination direction corresponding to the reference angle θ _h is the ideal alignment direction of the grids adjacent to the horizontal direction (X direction) (the distortion of the photographic chart image due to the scanning error is It is treated as approximating the direction of alignment in the absence). Also treated as tilt directions corresponding to the reference angle theta _v approximates the ideal arrangement direction squares adjacent in the vertical direction (Y direction).

[S26 in FIG. 7 (resetting of the center coordinates of interest)]
In this processing step S26, the center coordinates of interest are newly set in order to execute the subsequent processing steps.

  In this flowchart, the processing step group from this processing step S26 to the processing step S30 (updating the reference straight line) in the figure loops. Therefore, this processing step S26 is executed a plurality of times until exiting the loop. At the first execution, a label (i (= 0), j (= 0)) is assigned to the center coordinate closest to the center of the photographing chart image. Center coordinates to which labels (i (= 0), j (= 0)) are assigned are set as attention center coordinates. When this processing step S26 is executed again, the center coordinates adjacent in the + X direction from the center coordinates (i (= 0), j (= 0)) are sequentially set as the center coordinates of interest, and then the center coordinates (i Center coordinates adjacent in the −X direction from (= 0), j (= 0)) are sequentially set as target center coordinates. When this is completed, the center coordinates adjacent in the + X direction from the center coordinates (i (= 0), j (= 1)) are sequentially set as the attention center coordinates, and then the center coordinates (i (= 0), j From (= 1)), the center coordinates adjacent in the −X direction are set in order as the center coordinates of interest. When this is completed, the center coordinates adjacent in the + X direction from the center coordinates (i (= 0), j (= 2)) are sequentially set as the attention center coordinates, and then the center coordinates (i (= 0), j From (= 2)), the center coordinates adjacent in the −X direction are sequentially set as the center coordinates of interest. The above processing is repeatedly executed until all center coordinates are reset as the center coordinates of interest.

[S27 in FIG. 7 (Set Search Range)]
In this processing step S27, as shown in FIG. 8C, straight lines starting from the attention center coordinates set in the immediately preceding processing step S26 (resetting attention center coordinates) are drawn in four directions. Hereinafter, “the attention center coordinates set in the immediately preceding processing step S26 (resetting of the attention center coordinates)” will be referred to as “latest attention center coordinates”. Specifically, as a starting point the latest attention center coordinates, it has the reference length L and with a straight line having a slope corresponding to the reference angle theta _h is pulled in the + X direction position side and the -X direction position side, the reference length L have and straight line having a slope corresponding to the reference angle theta _v is pulled in the + Y direction position side and the -Y direction position. Next, in this processing step S27, a predetermined range centering on the end point of each of the four straight lines (hereinafter referred to as “reference straight line”) is set as a search range (dotted line in FIG. 8C). .

[S28 in FIG. 7 (labeling to adjacent center coordinates)]
In this processing step S28, the respective center coordinates belonging to the four search ranges set in the processing step S27 (search range setting) in the figure are labeled as adjacent center coordinates adjacent to the latest attention center coordinate. The label (i + 1, j) is assigned to the adjacent center coordinate on the + X direction position side, and the label (i−1, j) is assigned to the adjacent center coordinate on the −X direction position side. Further, a label (i, j + 1) is assigned to the adjacent center coordinate on the + Y direction position side, and a label (i, j−1) is assigned to the adjacent center coordinate on the −Y direction position side. Note that no labeling is performed when there is no center coordinate in the search range. Alternatively, labeling may be performed by regarding the end point of the reference straight line as the adjacent center coordinates. Further, when the center coordinates belonging to the search range are already labeled, the labeling is not performed again.

[S29 in FIG. 7 (scan completion determination)]
In this processing step S29, it is determined whether or not labeling has been performed on all center coordinates in the center coordinate array. This flowchart is ended when the labeling is performed on all the center coordinates in the center coordinate array (S29 in FIG. 7: YES). If center coordinates that have not been labeled remain (S29 in FIG. 7: NO), the process proceeds to processing step S30 (update of the reference straight line) in FIG.

[S30 in FIG. 7 (Update of Reference Line)]
The difference in pixel position between the adjacent center coordinates belonging to the search range and the end point of the reference line corresponds to the distortion difference on the photographing chart image between the cell corresponding to the latest attention center coordinate and the cell corresponding to the adjacent center coordinate. . The operation of the optical fiber 202 is controlled at a constant resonance frequency. For this reason, the scanning speed is higher at the periphery of the scanning region. As a result of repeated experiments and examinations based on these points, the present inventor has found that the distortion gradually increases as it approaches the periphery of the scanning region where the scanning speed is high.

  Based on the above knowledge, in this processing step S30, as shown in FIG. 8D, the adjacent center coordinates where the end points of the respective reference straight lines are labeled in the processing step S28 (labeling to the adjacent center coordinates) in the same figure. To be corrected. Each reference line is updated by correcting the end point position, so that the reference line has a length corresponding to the difference in distortion on the captured chart image between the cell corresponding to the latest attention center coordinate and the cell corresponding to the adjacent center coordinate. And the tilt changes. The reference straight line is gradually corrected and updated according to the difference in distortion every time this processing step S30 is executed.

  By repeating the loop from processing step S26 (resetting of the center coordinates of interest) to this processing step S30 in FIG. 7, labels (i, j) are assigned to all the center coordinates in the center coordinate array.

[Flow for creating a distortion correction table for shooting chart images]
The system main body 100 creates a distortion correction table for correcting distortion of the photographing chart image in each pixel. FIG. 9 shows a flowchart of distortion correction table creation processing executed by the CPU 108 of the system main body 100. Also, FIGS. 10A to 10C are explanatory assistance diagrams for assisting the explanation of the distortion correction table creation processing.

[S41 in FIG. 9 (Creation of ideal center coordinate array)]
In this processing step S41, a center coordinate array corresponding to the ideal array of cells is created. Specifically, one center coordinate is arranged at the center of a predetermined virtual area corresponding to the photographing chart image. Next, as shown in FIG. 10A, center coordinates are arranged at respective positions in the predetermined four directions and by a reference length L from the center coordinates arranged at the center of the virtual region. Specifically, the four predetermined directions correspond to a direction corresponding to the reference angle θ _h (+ X direction position side), a direction corresponding to the reference angle θ _h (−X direction position side), and a reference angle θ _v . direction (+ Y direction position side), a direction corresponding to the reference angle theta _v (-Y direction position side). Next, as shown in FIG. 10B, the center coordinates are arranged at a position that is separated from each of the four center coordinates arranged at each position by a reference length L in at least one predetermined direction. . The predetermined at least one direction is a direction excluding the direction in which the center coordinates are already arranged among the four directions. In the example of FIG. 10B, the center coordinates are arranged at positions separated from each of the four center coordinates by the reference length L in three directions excluding the direction in which the center coordinates are already arranged.

  In this processing step S41, as shown in FIG. 10C, the center coordinate arrangement process is repeated until the center coordinates are arranged in the entire virtual area. Hereinafter, for convenience of explanation, an array of center coordinates over the entire virtual region is referred to as an “ideal center coordinate array”.

[S42 in FIG. 9 (labeling to the ideal center coordinate array)]
In this processing step 42, a label (i, j) corresponding to the center coordinate array of the photographing chart image is assigned to each center coordinate in the ideal center coordinate array. Until the label (i, j) is assigned to all the center coordinates in the ideal center coordinate array, the processing step S26 in FIG. 7 (resetting the attention center coordinates) to the processing step S29 in FIG. 7 (scanning completion determination). A processing step group similar to the previous processing step group is loop-processed.

[S43 in FIG. 9 (Calculation of Distortion Correction Vector)]
In this processing step 43, the attention label is determined. Next, a vector is calculated that starts from the center coordinates in the center coordinate array of the photographic chart image to which the determined attention label is assigned and starts from the center coordinates of the ideal center coordinate array to which the attention label is assigned. Hereinafter, this vector is referred to as a “distortion correction vector”. The distortion correction vector is stored in the CPU memory 110.

  FIG. 11 is a conceptual diagram showing a distortion correction vector. In FIG. 11, a circle (solid line) indicates the center coordinate in the center coordinate array of the imaging chart image, and a circle (dotted line) indicates the center coordinate of the ideal center coordinate array. An arrow indicates a distortion correction vector between center coordinates to which a corresponding label (same label) is assigned.

[S44 in FIG. 9 (Distortion Correction Vector Calculation Completion Determination)]
In this processing step S44, it is determined whether or not distortion correction vectors have been calculated for all center coordinates to which labels have been assigned. This flowchart repeats the loop from S43 (calculation of distortion correction vector) to this processing step S44 until the distortion correction vector is calculated for all center coordinates to which labels are assigned (S43, FIG. 9). S44: NO). In this flowchart, when the distortion correction vectors are calculated for all the center coordinates to which the labels are assigned (S44 in FIG. 9: YES), the process exits from the loop and the processing step S45 in FIG. Go to (Calculation).

[S45 in FIG. 9 (calculation of distortion correction vector by interpolation)]
At the time of exiting the above loop, although the distortion correction vectors are calculated for all the pixels of the central coordinates to which the labels are assigned, the distortion correction vectors are calculated for the pixels between the pixels of the respective central coordinates. Absent. Therefore, in this processing step S45, a distortion correction vector is obtained by a known interpolation method such as spline interpolation for the pixels arranged between the pixels of the respective center coordinates.

[S46 in FIG. 9 (Create Distortion Correction Table)]
In this processing step S46, the distortion correction vectors and pixel addresses of all the pixels obtained in the processing step S43 (calculation of distortion correction vector) and the processing step S45 (calculation of distortion correction vector by interpolation) shown in FIG. A distortion correction table is created. The created distortion correction table is stored in the CPU memory 110.

  By executing the processing of the flowchart of FIG. 9, a distortion correction table that is a set of correction data in pixel units for image distortion is created.

[Chart correction flow of chart image]
The system main body 100 corrects the distortion of the confocal image using the distortion correction table. FIG. 12 shows a flowchart of image distortion correction processing executed by the CPU 108 of the system main body 100.

[S51 (Remapping) in FIG. 12]
In this processing step S51, remapping is performed using a remapping table. As a result, it is possible to generate a confocal image in which point images expressed by the respective confocal image detection signals are two-dimensionally arranged according to pixel addresses. However, at this stage, a distortion remains due to a scanning error or the like.

[S52 in FIG. 12 (Re-Mapping)]
Therefore, in this processing step S52, remapping is performed using the distortion correction table. Specifically, the position of the point image expressed by the confocal image detection signal remapped in the processing step S51 (remapping) in FIG. The position is converted from a position corresponding to the start point of the distortion correction vector to a position without distortion (a position corresponding to the ideal center coordinate array and corresponding to the end point of the distortion correction vector). This pixel address conversion is performed for all confocal image detection signals.

[S53 in FIG. 12 (Output of Confocal Image)]
In this processing step S53, the confocal image detection signal after the remapping processing is buffered in the image memory 118 in units of frames. The buffered confocal image detection signal is swept from the image memory 118 to the video signal output circuit 120 at a predetermined timing, and converted into a video signal conforming to a predetermined standard such as NTSC or PA. By outputting the converted video signal to the monitor 300, a confocal image whose distortion has been corrected is displayed on the display screen of the monitor 300.

  FIG. 13A shows a captured chart image generated when the processing step S52 (re-remapping) of FIG. 12 is not executed, and FIG. 13B shows that the processing step S52 (re-remapping) is executed. The imaging | photography chart image produced | generated in a case is shown. Comparing FIG. 13 (a) and FIG. 13 (b), it can be seen that the distortion in the captured chart image is corrected by executing the processing step S52 (re-remapping).

  Distortion correction is performed in units of fine areas such as pixels. Therefore, the confocal image is corrected with high accuracy even if it has complicated distortion. The distortion correction value can be easily calculated as a distortion correction vector. In the present embodiment, unlike the method described in Patent Document 1, it is not necessary to repeat the calibration work, so the time load required for distortion correction is light. Moreover, since it is not necessary to provide PSD with a jig, the structure of the jig is simplified as compared with Patent Document 1.

  The above is the description of the exemplary embodiments of the present invention. Embodiments of the present invention are not limited to those described above, and various modifications are possible within the scope of the technical idea of the present invention. For example, the embodiment of the present application also includes contents appropriately combined with examples and the like clearly shown in the specification or obvious examples.

1 Scanning Confocal Endoscope System 100 System Main Body 200 Confocal Probe 300 Monitor 400 Jig

Claims (5)

Receives return light from a subject in which a plurality of indicators are regularly arranged, scanned in a spiral shape with a predetermined rotation period from one of the center and the periphery to the other within a substantially circular scanning area by a predetermined scanning light. Image signal acquisition means for acquiring an image signal of the subject detected by
A pixel position in a two-dimensional pixel array having a positional relationship corresponding to a scanning position in a scanning area by the scanning light is assigned to the acquired image signal according to the detection timing of the image signal, and the assigned Image generating means for generating an image of the subject by arranging the image signals at pixel positions;
Index position detection means for detecting a pixel position of each index in the generated image;
Index image array creating means for creating an index image array in which the plurality of indices are arranged based on the detected pixel positions of the indices;
For each index in the image, a straight line starting from the index and having at least one adjacent index positioned in a predetermined direction with respect to the index is defined, and the length and inclination of the defined straight line are calculated. A length inclination calculating means to perform,
Peak detection means for detecting the length peak when distributing the number of appearances of the length of the straight line calculated for each index and the slope peak when distributing the number of appearances of the slope of the straight line;
Ideal array creating means for creating an ideal array by arranging the plurality of indices at an arrangement interval of a length corresponding to the length peak in the direction of the slope corresponding to the tilt peak;
Distortion correcting means for comparing the index image array created by the index image array creating means with the ideal array created by the ideal array creating means and correcting distortion of each index in the image based on the comparison result When,
Comprising
Scanning confocal endoscope system . The index image array creation means includes
(1a) determining a target index from within the index image array;
(2a) assigning array position information in the index image array to the determined target index,
(3a) Define at least one reference line having a predetermined length and a predetermined inclination starting from the attention index to which the array position information is assigned, and within a predetermined range including the end point of the defined reference line Assigning sequence position information adjacent to the target index to the index that is located,
(4a) Repeat (1a) to (3a) until the sequence position information is assigned to all indexes.
The scanning confocal endoscope system according to claim 1. The ideal array creating means includes
(1b) determining an attention index from within the ideal sequence;
(2b) assigning array position information corresponding to the index image array to the determined target index,
(3b) Define at least one reference line having a predetermined length and a predetermined inclination starting from the attention index to which the array position information is assigned, and within a predetermined range including the end point of the defined reference line Assigning sequence position information adjacent to the target index to the index that is located,
(4b) Repeat (1b) to (3b) until the array position information corresponding to the index image array is assigned to all indices.
The scanning confocal endoscope system according to claim 2 . The distortion correction means includes
For each index in the index image array, calculate a correction vector based on the difference in pixel position from the index in the ideal array to which the corresponding array position information is assigned,
Correcting distortion of each index based on the calculated correction vector;
The scanning confocal endoscope system according to claim 3 . The subject is
The white squares are arranged in a matrix on a black background.
The scanning confocal endoscope system according to any one of claims 1 to 4 .