JP4968057B2 - Method for detecting line-shaped abnormal image element of radiation detector and radiation imaging apparatus - Google Patents

Description

  The present invention relates to a radiation detector of a radiation imaging apparatus used in the medical field, industrial fields such as non-destructive inspection, RI (Radio Isotope) inspection and optical inspection, and nuclear power field. The present invention relates to a technique for detecting a line-like abnormal image element among dimensionally arranged image elements (detection elements).

  In a medical X-ray fluoroscopic apparatus, which is one of the representative apparatuses of radiation imaging apparatuses, in recent years, a flat panel X-ray detector (or “two-dimensional X-ray sensor”) is also used as a radiation detector. Many are equipped with “FPD” as appropriate.

  On the detection surface of the FPD on which the X-ray image transmitted through the subject is projected, an X-ray sensitive semiconductor film and a plurality of (for example, 1536 vertical × 1536 horizontal) arranged in a two-dimensional matrix form. And a detection element (image element). Specifically, a thin film transistor (TFT), carrier collection electrode, etc. constituting the detection element are separately arranged on a glass substrate, and an amorphous selenium (a-Se) film as a semiconductor film is deposited thereon. Is done.

  When an X-ray image is projected on the FPD, charges proportional to the density of the image are generated in the semiconductor film. This charge is collected and stored as charge information from the carrier collection electrode for a predetermined time. Thereafter, the thin film transistor is turned on, and the accumulated charge information is read out.

  Thus, the charge information obtained from the detection elements is amplified by an amplifier provided individually for each column of the detection elements. Then, it is digitized by an A / D converter. In this specification, digitized charge information is referred to as “X-ray detection signal”. One X-ray image is generated by performing digital image processing on the plurality of X-ray detection signals. Note that the pixels constituting the X-ray image have a corresponding relationship with the detection element.

  Among the plurality of detection elements in the FPD described above, detection elements having an abnormality in outputting an X-ray detection signal corresponding to the irradiated X-ray image are continuously arranged in a line (vertical direction or horizontal direction). Thus, a plurality of (for example, 720) line-shaped abnormality detection elements (or called line-shaped abnormality pixel elements) may be formed. These line-shaped abnormality detection elements (a plurality of detection elements that are continuous in a line shape) are not completely broken, but do not perform their normal functions and can be said to be defective elements (missing elements). This line-like abnormality detecting element is considered to be caused by the influence of noise or the like due to reading of the thin film transistor of each detecting element. If an X-ray image is acquired as it is based on the X-ray detection signal obtained from the line-like abnormality detection element, the pixel corresponding to the line-like abnormality detection element cannot correctly display the X-ray image, but may be white. , It will turn black and cause inconvenience.

For this reason, in digital image processing, a line-shaped abnormality detection element is first detected, and an X-ray detection signal obtained from the detected line-shaped abnormality detection element is corrected. Specifically, the line-shaped abnormality detection element is determined using a threshold or the like as to whether or not the obtained X-ray detection signal is within a normal range. If it is determined that the detection element is not normal, the detection element that outputs the detection element is regarded as a defective element. Then, the X-ray detection signal obtained from the detection element adjacent to the defective element is used for restoration such as replacement processing and interpolation processing. In addition, the detection element regarded as a defective element is stored, and in the subsequent digital image processing, the X-ray detection signal is corrected while referring to the stored information. This avoids degradation of the image quality of the X-ray image (see, for example, Patent Document 1).
Special table 2002-538639 gazette

However, the conventional example having such a configuration has the following problems.
That is, according to the conventional method, the value (pixel value) of the X-ray detection signal obtained from the defective element is larger than the value (pixel value) of the X-ray detection signal of any other normal detection element. When taking a protruding value, it is possible to set a threshold value for extracting only defective elements and not normal detecting elements, so that it can be easily regarded as defective elements. However, the X-ray detection signal obtained from the defective element does not necessarily take a protruding value.

  That is, a line shape in which the value (pixel value) of the X-ray detection signal of the defective element has only a minute change compared to the value (pixel value) of the X-ray detection signal of another normal detection element. There is a problem that it is difficult to extract (detect) a defect.

  For example, if the horizontal axis represents each detection element and the vertical axis represents the value (pixel value) of the X-ray detection signal, the profile of the X-ray detection signal spatially developed according to the position of each detection element is shown. The value (pixel value) of the X-ray detection signal of the element may appear within a normal range apparently as a whole of the X-ray detection signal.

  In this case, the X-ray detection signal obtained from the defective element is erroneously determined to be “normal” by the conventional method, and the defective element is mistaken for a normal detection element. If an X-ray image is acquired based on this determination, deterioration in image quality cannot be avoided.

  That being said, if it is strictly determined whether or not the obtained X-ray detection signal is within a normal range (set to a threshold value that would cause a normal detection element to be erroneously detected), the normal detection element May be mistaken for a defective element, and similarly the image quality is degraded.

  This invention is made in view of such a situation, Comprising: The detection method of the line-shaped abnormal image element of a radiation detector which can detect the line-shaped abnormal image element in a radiation detector correctly, and An object of the present invention is to provide a radiation imaging apparatus capable of acquiring a high-quality radiation image.

In order to achieve such an object, the present invention has the following configuration.
That is, the invention according to claim 1 is a linear detector of a radiation detector for detecting a line-like abnormal image element in a radiation detector in which a plurality of image elements for detecting radiation irradiated by a radiation source are two-dimensionally arranged. An abnormal image element detection method for detecting initial image data created based on a detection output signal from a radiation detector when the detection surface of the radiation detector is irradiated with radiation. A smoothing step for generating image data after smoothing filter processing by performing smoothing filter processing with a filter size that is longer in the direction orthogonal to the line direction than the line direction of the line-like abnormal image element; A difference processing step for generating difference image data obtained by taking a difference between the image data and the image data after the smoothing filter processing; A determination processing step of determining a plurality of line-shaped image elements in which image elements having pixel values equal to or higher than the threshold value are set as line abnormal image elements by setting a predetermined threshold value for the image data. It is characterized by that.

  [Operation / Effect] According to the invention described in claim 1, in the smoothing processing step, the line-like abnormality to be detected is detected with respect to the initial image data generated based on the detection output signal from the radiation detector. Since smoothing filter processing with a filter size that is longer in the direction perpendicular to the line direction than the line direction of the image element is performed, the line-like abnormal image is compared with smoothing filter processing using a square (n × n) box filter. The element can be blurred accurately. In other words, since the square (n × n) box filter performs smoothing processing in a state in which many line-like abnormal image elements are included in the line direction, it cannot accurately blur the line-like abnormal image elements. The smoothing filter processing with a filter size longer in the direction perpendicular to the line direction than the line direction of the line-like abnormal image element can reduce the inclusion of line-like abnormal image elements in the line direction. It is possible to generate image data after smoothing filter processing in which the image element is accurately blurred accordingly. In the difference processing step, difference image data is generated by taking a difference between the initial image data and the image data after the smoothing filter processing of the long filter size, so that the line-like abnormal image element is changed to a normal image element. The difference image data can be generated more accurately, and the threshold value is set for the difference image data in which the line-like abnormal image element is raised more accurately than the normal image element in the determination processing step. A plurality of line-shaped image elements in which image elements having pixel values equal to or greater than the threshold value can be determined as line-shaped abnormal image elements, and line-shaped abnormal image elements (line-shaped (Defect) can be extracted (detected). Then, by correcting the radiation detection signal obtained from the line-like abnormal image element, a high-quality radiation image can be acquired. The correction of the radiation signal includes known techniques such as so-called replacement processing and interpolation processing.

  The invention described in claim 2 further comprises an edge enhancement processing step of performing edge enhancement processing on the difference image data in the detection method of a line-shaped abnormal image element of the radiation detector according to claim 1, In the determination processing step, a predetermined threshold is set for the difference image data after the edge enhancement processing, whereby a plurality of line-shaped image elements in which image elements having pixel values equal to or higher than the threshold are continuous are formed in a line shape. It is determined that the image element is abnormal.

  [Operation / Effect] According to the invention described in claim 2, since the edge emphasis processing is performed on the difference image data in the edge emphasis processing step, the line-like abnormal image element is further emphasized and emerged. The difference image data after the edge enhancement processing can be generated, and the threshold value is set for the difference image data after the edge enhancement processing in which the line-like abnormal image element is further emphasized and raised in the determination processing step. It is possible to more accurately extract (detect) a line-like abnormal image element (line-like defect) that has only one. In other words, the line-shaped abnormal image element (line-shaped defect) is more accurately extracted (detected) by setting the threshold value for the differential image data after the edge enhancement process than when the threshold value is set for the differential image data without the edge enhancement process. can do. Then, by correcting the radiation detection signal obtained from the line-like abnormal image element, a high-quality radiation image can be acquired. The correction of the radiation signal includes known techniques such as so-called replacement processing and interpolation processing.

According to a third aspect of the present invention, in the method for detecting a line-like abnormal image element of the radiation detector according to the second aspect, the image elements are arranged in the vertical and horizontal directions in the radiation detector, and the detection is attempted. When the line-like abnormal image element is a vertical line-like abnormal image element, the smoothing processing step smoothes the initial image data by performing a smoothing filter process with a long filter size in the horizontal direction. Image data after filtering is generated, and the determination processing step generates a profile in the horizontal direction for the difference image data after the edge enhancement processing, and an image having a pixel value equal to or greater than a vertical line threshold value in each profile. The element is determined to be a vertical line-shaped abnormal image element.

  [Operation / Effect] According to the invention described in claim 3, in the case of detecting the vertical line-shaped abnormal image element, in the smoothing processing step, the smoothing of the filter size which is long in the horizontal direction with respect to the initial image data. Image data after smoothing filter processing is generated by performing the filtering processing, and edge enhancement processing is performed on the difference image data in the edge enhancement processing step, and the determination processing step is performed on the difference image data after the edge enhancement processing. A profile in the horizontal direction is generated, and a plurality of image elements in the vertical direction in which pixel values equal to or greater than the threshold value for the vertical line in the profile are determined as vertical line-like abnormal image elements. Even a vertical line-shaped abnormal image element can be extracted (detected).

According to a fourth aspect of the present invention, in the method for detecting a line-shaped abnormal image element of the radiation detector according to the second aspect, the image elements are arranged in the radiation detector in the vertical and horizontal directions, and the detection is attempted. When the line-like abnormal image element is a horizontal line-like abnormal image element, the smoothing process step performs smoothing by performing a smoothing filter process with a filter size that is long in the vertical direction on the initial image data. Image data after filtering is generated, and the determination processing step generates a horizontal profile for the difference image data after the edge enhancement processing, and the image element having a pixel value equal to or greater than a horizontal line threshold in the profile Is determined as a horizontal line-shaped abnormal image element.

  [Operation / Effect] According to the invention described in claim 4, in the case of detecting a horizontal line-like abnormal image element, in the smoothing processing step, smoothing of a filter size which is long in the vertical direction with respect to the initial image data. The image data after the smoothing filter processing is generated by performing the smoothing filter processing, the image data after the smoothing filter processing is subjected to edge enhancement processing in the edge enhancement processing step, and the difference after the edge enhancement processing is determined in the determination processing step A horizontal profile for image data is generated, and an image element having a pixel value greater than or equal to the threshold for the horizontal line in the profile is determined as a horizontal line abnormal image element. Even a line-like abnormal image element can be extracted (detected).

The invention described in Claim 5, in the detection method of the linear abnormal image elements of the radiation detector according to any one of claims 1 to 4, further tone reversal processes the images data comprises a black and white reversal processing step, the smoothing process step, the difference processing step, the initial image data in the determination processing step is characterized in the this black and white reversal processing has been made.

  [Operation / Effect] According to the invention described in claim 5, in the black and white reversal processing step, the initial image data is subjected to black and white reversal processing, and in the smoothing processing step, the initial image data after reversal which is the initial image data subjected to black and white reversal processing. Smoothing filter processing is performed on the image data. In the difference processing step, the difference between the inverted initial image data and the inverted image data after smoothing filter processing obtained by performing smoothing filter processing on the inverted initial image data is obtained. In the determination processing step, a predetermined threshold value is set for the inverted image data after the smoothing filter processing, so that a plurality of line-shaped image elements in which image elements having pixel values equal to or higher than the threshold value are continuous are formed in a line shape. Since it is determined as an abnormal image element, a line-like abnormal image element having a lower pixel value than that of a group of normal image elements in the initial image data (that is, differential processing) Pixel value the linear abnormal image element takes a negative value) can be extracted. Further, since such a black-and-white reversal processing step is provided, a line-like abnormal image element having a negative pixel value in the difference processing can be handled as a positive value.

Further, the invention according to claim 6, in the detection method of the linear abnormal image elements of the radiation detector according to any one of claims 1 to 5, further negative pixels in the images data with a certain process step for a constant the value zero or a positive value, the difference image data in the determination processing step, those characterized by the this were subjected to constant process on the difference image data in the differential processing step It is.

  [Operation and Effect] According to the invention described in claim 6, in the stabilization processing step, the negative pixel value in the differential image data after the edge enhancement processing is fixed to zero or a positive value, and in the determination processing step By setting a predetermined threshold for the difference image data after the edge enhancement processing after the stabilization processing, a plurality of line-shaped image elements having continuous pixel elements having a pixel value equal to or greater than the threshold are detected as line abnormalities. Since it is determined as an image element, an image element having a negative pixel value in the edge enhancement process can be handled as a positive value.

  The invention according to claim 7 is a radiation imaging apparatus comprising a radiation detector in which a plurality of image elements for detecting radiation emitted from a radiation source are two-dimensionally arranged. Storage means for storing position information data of the line abnormal image element of the radiation detector used in the apparatus detected by the detection method of the line abnormal image element of the radiation detector according to any one of And correction processing means for correcting a radiation detection signal of the line abnormal image element of the radiation detector using position information data of the line abnormal image element stored in the storage means. To do.

  [Operation / Effect] According to the invention described in claim 7, in the radiation imaging apparatus comprising a radiation detector in which a plurality of image elements for detecting radiation irradiated by a radiation source are two-dimensionally arranged. The positional information data of the line abnormal image element of the radiation detector used in the apparatus detected by the detection method of the line abnormal image element of the radiation detector according to any one of claims 1 to 6 is stored. And a correction processing means for correcting the radiation detection signal of the line abnormal image element of the radiation detector using the positional information data of the line abnormal image element stored in the storage means. The correction of the radiation detection signal of the line-shaped abnormal image element of the radiation detector can be suitably performed.

  According to the detection method of the line-shaped abnormal image element of the radiation detector according to the present invention, in the smoothing processing step, it is attempted to detect the initial image data created based on the detection output signal from the radiation detector. Compared with the smoothing filter processing with a square (n × n) box filter, the smoothing filter processing is performed with a filter size that is longer in the direction orthogonal to the line direction than the line direction of the line-shaped abnormal image element. The line-like abnormal image element can be accurately blurred. In other words, since the square (n × n) box filter performs smoothing processing in a state in which many line-like abnormal image elements are included in the line direction, it cannot accurately blur the line-like abnormal image elements. In the smoothing filter process having a filter size that is long in the direction perpendicular to the line direction of the line-like abnormal image element, the inclusion of the line-like abnormal image element in the line direction can be reduced. It is possible to generate image data after smoothing filter processing that has been accurately blurred. In the difference processing step, difference image data is generated by taking a difference between the initial image data and the image data after the smoothing filter processing of the long filter size, so that the line-like abnormal image element is changed to a normal image element. The difference image data can be generated more accurately, and the threshold value is set for the difference image data in which the line-like abnormal image element is raised more accurately than the normal image element in the determination processing step. An image element having a pixel value equal to or higher than the threshold value can be determined as a line-like abnormal image element, and a line-like abnormal image element (line-like defect) having only a minute change can be extracted (detected). it can. Then, by correcting the radiation detection signal obtained from the line-like abnormal image element, a high-quality radiation image can be acquired.

  The radiation imaging apparatus according to the present invention is used in the apparatus detected by the detection method for a line-like abnormal image element of the radiation detector according to any one of claims 1 to 6. The storage means for storing the position information data of the line-shaped abnormal image element of the radiation detector and the position information data of the line-shaped abnormal image element stored in the storage means are used for the line-shaped abnormal image element of the radiation detector. Since the correction processing means for correcting the radiation detection signal is provided, it is possible to suitably correct the radiation detection signal of the line-like abnormal image element of the radiation detector.

Embodiment 1 of the present invention will be described below with reference to the drawings.
FIG. 1 is a block diagram illustrating an overall configuration of the X-ray fluoroscopic imaging apparatus according to the first embodiment. In this embodiment, a medical X-ray fluoroscopic imaging apparatus will be described as an example of the radiation imaging apparatus.

  The imaging unit 1 of the X-ray fluoroscopic imaging apparatus includes a top 2 on which a subject M as a subject is placed, an X-ray tube 3 that irradiates the subject M with X-rays, and a plurality of detection elements (images). And a flat panel X-ray detector (hereinafter referred to as “FPD” as appropriate) 4 for detecting X-rays transmitted through the subject M.

 The X-ray tube 3 described above corresponds to the radiation source in the present invention. The FPD 4 described above corresponds to the radiation detector in the present invention.

  The X-ray fluoroscopic imaging apparatus includes a movement control unit 5 that moves the top plate 2, the X-ray tube 3, and the FPD 4, an X-ray tube control unit 6 that controls the tube voltage and tube current of the X-ray tube 3, and the FPD 4. FPD control unit 7 that reads and controls the charge information of A, D / A converter 9 that digitizes the charge information read from each detection element and converts it into an X-ray detection signal, and an X-ray image from the X-ray detection signal And a digital image processing unit 10 to be acquired.

  The digital image processing unit 10 is further connected to the output side of the A / D converter 9 to store an X-ray detection signal and a line from the X-ray detection signal stored in the signal storage unit 11. A line-shaped abnormality detecting element extracting unit 13 for extracting a line-shaped abnormality detecting element, a defect information storing unit 17 for storing position information of the line-shaped abnormality detecting element extracted by the line-shaped abnormality detecting element extracting unit 13, and an A / A correction unit 21 that is connected to the output side of the D converter 9 and corrects / restores the X-ray detection signal while referring to the defect information, and an X-ray detection signal (correction is made for the X-ray detection signal obtained from the defective element). And an image generation unit 23 that generates an X-ray image based on a later X-ray detection signal).

  The X-ray tube 3 and the FPD 4 are disposed to face each other with the subject M interposed therebetween. The movement control unit 5 horizontally moves or rotates the X-ray tube 3 and the FPD 4 so that this state is maintained. The X-ray tube 3 irradiates the subject M with a predetermined dose of X-rays based on the control of the X-ray tube control unit 6. In FIG. 1, the irradiated X-ray is schematically shown by a one-dot chain line.

  FIG. 2 is a vertical cross-sectional view of the main part of the FPD 4, and FIG. 3 is a plan view of the main part of the FPD 4. As shown in the figure, an application electrode 31, an X-ray sensitive semiconductor film 33, a carrier collection electrode 35, and an active matrix substrate 37 are stacked in this order from the X-ray incident side. The carrier collecting electrodes 35 are separately formed in a two-dimensional matrix in plan view.

  On the active matrix substrate 37, a capacitor Ca that accumulates charge information for each carrier collection electrode 35 and thin film transistors (Thin Film Transistors) Tr that are switch elements for extracting the charge information are separately formed. The carrier collection electrode 35 and the capacitor Ca are connected to the source S of the thin film transistor Tr. Also, a plurality of gate bus lines 41 commonly connected to the gates of the thin film transistors Tr in each row and a plurality of data bus lines 43 commonly connected to the drains of the thin film transistors Tr in each column are formed.

  Each data bus line 43 drawn from the active matrix substrate 37 is connected to a separate amplifier 47. The output side of the amplifier 47 is collected in the A / D converter 9.

  When X-rays enter the FPD 4 with a bias voltage applied to the application electrode 31, charges are generated in the semiconductor film 33, and the charges are accumulated in the capacitor Ca via the carrier collection electrodes 35. The gate bus line 41 transmits a scanning signal from the gate driver 45 and applies it to the gate of the thin film transistor Tr. As a result, the charge information stored in the capacitor Ca is read out to the data bus line 43 via the thin film transistor Tr that has been turned on. The FPD control unit 7 performs the charge information read operation control corresponding to the X-ray irradiation. The time during which charge information is collected and accumulated (corresponding to the time during which the thin film transistor Tr is in the off state) is referred to as “accumulation time”.

  The charge information read through each data bus line 43 is amplified by an amplifier 47. Thereafter, it is digitized by the A / D converter 9 to obtain an X-ray detection signal.

  Thus, one set of carrier collection electrode 35, capacitor Ca, and thin film transistor Tr constitute one detection element d that outputs charge information.

  Therefore, it can be seen that a large number of detection elements d are arranged in a two-dimensional matrix on the detection surface of the FPD 4 as shown in FIG. For example, 2880 vertical × 2880 horizontal detection elements d are arranged on a detection surface having a width of about 50 cm × 50 cm. Each detection element d has a correspondence relationship with each pixel constituting the generated X-ray image. In the following description, it is assumed that (n × m) detection elements d are arranged in n rows × m columns on the detection surface of the FPD 4, and each detection element d has two-dimensional coordinates (i, j). ) (I is an integer from 1 to n and j is an integer from 1 to m).

  The signal storage unit 11 is configured by a storage medium represented by a RAM (Random Access Memory) or an HDD (Hard Disk Drive), and has a capacity for storing at least one frame of X-ray detection signals for each detection element d. .

  Here, the configuration of the line abnormality detection element extraction unit 13 will be described with reference to FIG. FIG. 4 is a block diagram illustrating a configuration of the line-like abnormality detection element extraction unit 13.

  As shown in FIG. 4, the line-shaped abnormality detection element extraction unit 13 includes a vertical line defect extraction processing unit 14 that extracts line-shaped abnormality detection elements in the vertical direction (column direction) of the FPD 4, and a horizontal direction (row direction) of the FPD 4. And a horizontal line defect extraction processing unit 15 for extracting the line-shaped abnormality detection element 15). Next, the configuration of the vertical line defect extraction processing unit 14 and the horizontal line defect extraction processing unit 15 will be described in order below.

  First, as shown in FIG. 4, the vertical line defect extraction processing unit 14 performs, for example, a gentle fluctuation on the X-ray detection signal of each detection element d for one frame of the FPD 4 read from the signal storage unit 11. A pre-processing unit 51 that generates and stores trend-processed initial image data for removing inclination (gradient) is provided. In addition, if there is no uniform gradient etc. about these X-ray detection signals, it is not necessary to perform a trend process.

  As shown in FIG. 4, the vertical line defect extraction processing unit 14 uses the initial image data read from the preprocessing unit 51 to detect the line of the vertical (column direction) line-like abnormal image element to be detected by the FPD 4. Performs smoothing filter processing with a filter size (for example, 1 × vertical 33 horizontal filter size) longer in the direction (horizontal direction: row direction) orthogonal to the line direction (vertical direction) than the direction (vertical direction). Thus, a horizontal median processing unit 52 is provided as a smoothing processing unit that generates image data after the smoothing filter processing.

  Further, as shown in FIG. 4, the vertical line defect extraction processing unit 14 generates difference image data obtained by taking the difference between the initial image data and the image data after the smoothing filter processing by the horizontal median processing unit 52. An image subtraction processing unit 53 serving as a difference processing unit, an edge enhancement processing unit 54 that applies edge enhancement processing (for example, 3 × 3 Laplacian processing) to the difference image data, and difference image data after edge enhancement processing. A vertical line loss determination processing unit 55 as a determination processing unit for determining an image element having a pixel value equal to or higher than the threshold value as a vertical (column direction) line-shaped abnormal image element by setting a predetermined threshold for the It has.

  The vertical (column direction) line-like abnormal image element extracted by the vertical line defect determination processing unit 55 has a pixel value that slightly changes in the positive direction compared to other normal image elements (in the positive direction). The position information of the line-shaped abnormal image element in the vertical direction (column direction) is stored in the defect information storage unit 17.

  Further, as shown in FIG. 4, the vertical line defect extraction processing unit 14 includes a black and white reversal processing unit 56 that performs black and white reversal processing on the initial image data read from the preprocessing unit 51. Then, the horizontal median processing unit 52 performs the smoothing filter processing of the above-described horizontal filter size on the initial image data after inversion which is the initial image data subjected to the black and white inversion processing, and the image subtraction processing unit 53 performs the inversion The edge enhancement processing unit 54 calculates a difference between the post-initial image data and the reverse image data after the smoothing filter process obtained by performing the smoothing filter process on the post-inversion initial image data. The vertical line loss determination processing unit 55 sets a predetermined threshold value for the inverted image data after the smoothing filter process that has been subjected to the edge enhancement process, so that an image element having a pixel value equal to or higher than the threshold value is set in the vertical direction. The direction (column direction) is determined to be a line-like abnormal image element. That is, the determined line-like abnormal image element has a pixel value that slightly changes in the minus direction (those that have a minute fluctuation range in the minus direction) compared to other normal image elements. The position information of the vertical (column direction) line-shaped abnormal image element is also stored in the defect information storage unit 17.

  Next, as shown in FIG. 4, the horizontal line defect extraction processing unit 15 performs, for example, a gentle fluctuation on the X-ray detection signal of each detection element d for one frame of the FPD 4 read from the signal storage unit 11. , A pre-processing unit 61 that generates and stores trend-processed initial image data for removing inclination (gradient) is provided. In addition, if there is no uniform gradient etc. about these X-ray detection signals, it is not necessary to perform a trend process.

  As shown in FIG. 4, the horizontal line defect extraction processing unit 15 applies the line of the horizontal (row direction) abnormal line image element to be detected by the FPD 4 to the initial image data read from the preprocessing unit 61. Performs smoothing filter processing with a filter size (for example, vertical 33 × horizontal 1 vertical filter size) longer in the direction (vertical direction: column direction) orthogonal to the line direction (horizontal direction) than the direction (horizontal direction). Thus, a vertical median processing unit 62 is provided as a smoothing processing unit that generates image data after the smoothing filter processing.

  Further, as shown in FIG. 4, the horizontal line defect extraction processing unit 15 generates difference image data obtained by taking the difference between the initial image data and the image data after the smoothing filter processing by the vertical median processing unit 62. An image subtraction processing unit 63 as a difference processing unit, an edge enhancement processing unit 64 that performs edge enhancement processing (for example, 3 × 3 Laplacian processing) on the difference image data, and difference image data after the edge enhancement processing. By setting a predetermined threshold for the horizontal line loss determination processing unit 65 as a determination processing unit for determining an image element having a pixel value equal to or higher than the threshold as a horizontal (row direction) line-shaped abnormal image element; It has.

  The lateral (row direction) line-like abnormal image element extracted by the lateral line defect determination processing unit 65 has a pixel value that slightly changes in the positive direction compared to other normal image elements (in the positive direction). The position information of the line-like abnormal image element in the lateral direction (row direction) is stored in the defect information storage unit 17.

  Further, as shown in FIG. 4, the horizontal line defect extraction processing unit 15 includes a black and white reversal processing unit 66 that performs black and white reversal processing on the initial image data read from the preprocessing unit 61. The vertical median processing unit 62 performs smoothing filter processing of the above-described vertical filter size on the inverted initial image data that is the initial image data that has been subjected to black and white inversion processing, and the image subtraction processing unit 63 performs inversion The edge enhancement processing unit 64 takes a difference between the post-initial image data and the reverse image data after the smoothing filter process obtained by performing the smoothing filter process on the post-inversion initial image data. The horizontal line loss determination processing unit 65 sets a predetermined threshold value for the inverted image data after the smoothing filter process subjected to the edge enhancement process, so that image elements having pixel values equal to or higher than the threshold value are continuous. A plurality of lateral (row direction) line-shaped image elements are determined as lateral (row direction) line-like abnormal image elements. That is, the determined line-like abnormal image element has a pixel value that slightly changes in the minus direction (those that have a minute fluctuation range in the minus direction) compared to other normal image elements. The positional information of the horizontal (row direction) line-shaped abnormal image element is also stored in the defect information storage unit 17.

  The defect information storage unit 17 is configured by a storage medium represented by a RAM (Random Access Memory) or an HDD (Hard Disk Drive). And the defect information which is the positional information etc. of the line-like abnormal image element extracted in the vertical direction (column direction) and the horizontal direction (row direction) extracted by the line-like abnormality detecting element extraction unit 13 is stored. The defect information is position information based on two-dimensional coordinates (i, j) for identifying the line-like abnormal image element.

  The defect information storage unit 17 described above corresponds to the storage unit in the present invention, and the correction unit 21 described above corresponds to the correction processing unit in the present invention.

  The correction unit 21 includes a median filter or the like. Then, with reference to the defect information (position information in the vertical direction (column direction) and the horizontal direction (row direction) line-shaped abnormal image element) stored in the defect information storage unit 17, the X-ray detection signal output by the defect information The X-ray detection signal obtained from the normal detection element d adjacent to the defective element is replaced with the median value.

  The correction unit 21 is configured by a median filter or the like and configured to replace the X-ray detection signal. However, the correction unit 21 is not limited to this, and the X-ray obtained from a normal detection element d adjacent to the defective element. A known correction method such as interpolation processing (for example, spline interpolation, Lagrangian interpolation, SINC function) may be employed based on the detection signal.

  Next, a method for detecting a line-like abnormal image element of the FPD 4 according to the first embodiment will be described with reference to the drawings. FIG. 5 is a flowchart illustrating a process of detecting a line-like abnormal image element of the FPD 4 according to the first embodiment. FIG. 6 is a diagram showing correction data for making various vertical line-shaped abnormal image elements. 7A is initial image data spatially developed according to the position of each detection element, FIG. 7B is image data obtained by performing smoothing filter processing, FIG. 7C is difference data, and FIG. (D) is the figure which showed typically each profile of the difference data after edge emphasis. FIG. 8A shows initial image data having a negative line loss, FIG. 8B shows data obtained by performing black-and-white inversion processing on the initial image data shown in FIG. 8A, and FIG. FIG. 8D is a diagram schematically showing the profile of the difference data after edge enhancement.

  Here, as the correction data of the X-ray fluoroscopic imaging apparatus, data of various vertical line-like abnormal image elements as shown in FIG. 6 are created in advance. That is, the image size (detection surface size) of the FPD 4 is 2880 [pixel: pix] × 2880 [pixel: pix], and the pixel values are “+4”, “+5”, “+6”, “+7”, “+10”. ”And“ +30 ”are created by arranging approximately 24 sets of the horizontal lines arranged at intervals of 20 [pixel: pix] side by side (hereinafter referred to as first correction data as appropriate). Specifically, this line extends from the upper end to the lower end of the FPD 4, and there are 24 for “+4” to “+10” and 23 for “+30”. In addition, six types of vertical lines with pixel values “−4”, “−5”, “−6”, “−7”, “−10”, and “−30” are arranged at intervals of 20 [pixel: pix]. A set of approximately 24 sets arranged side by side (hereinafter referred to as second correction data as appropriate) is also created separately.

  Similarly, the data of various horizontal line-like abnormal image elements are created in advance as correction data of the X-ray fluoroscopic imaging apparatus. In other words, six types of horizontal lines with pixel values “+4”, “+5”, “+6”, “+7”, “+10”, and “+30” are vertically arranged at intervals of 20 [pixel: pix]. Approximately 24 sets vertically arranged (hereinafter, appropriately referred to as third correction data) are created. Specifically, this line extends from the left end of the FPD 4 to the right end, with 24 for “+4” to “+10” and 23 for “+30”. In addition, six horizontal lines with pixel values “−4”, “−5”, “−6”, “−7”, “−10”, and “−30” are arranged at intervals of 20 [pixel: pix]. A set of approximately 24 sets vertically arranged (hereinafter referred to as fourth correction data as appropriate) is also created separately.

<Step S1> X-ray irradiation to the FPD 4 to obtain initial image data As shown in FIG. 5, with the correction function of the X-ray fluoroscopic imaging device turned on, it is equivalent to 0.03 mR (millientgen) (pixel value) A uniform irradiation image of about 100) is taken. That is, by shooting with the correction function turned on, the image includes the various lines shown in FIG.

  Specifically, the FPD 4 does not have vertical (column direction) and horizontal (row direction) line-like abnormal image elements as the apparatus itself. For example, vertical line correction data (first correction data) ) Is turned on, initial image data obtained by adding correction data (first correction data) of various vertical line-shaped abnormal image elements shown in FIG. 6 is obtained (FIG. 7A). The correction function using the horizontal line correction data (third correction data) is turned on to add correction data (third correction data) of various horizontal line-like abnormal image elements (not shown). Initial image data is obtained. As described above, the initial image data may be subjected to trend processing as necessary.

<Step S2> Vertical line-shaped defect extraction?
As shown in FIG. 5, when a vertical line-shaped abnormal image element is extracted, the process proceeds to step S3. When a horizontal line-shaped abnormal image element is extracted, the process proceeds to step S11.

<Step S3> Horizontal Median Processing The horizontal median processing unit 52 performs FPD 4 on the initial image data added with correction data (first correction data) of various vertical line-like abnormal image elements shown in FIG. Compared to the line direction (vertical direction) of the vertical (column direction) line-like abnormal image element to be detected, the filter size is longer in the direction (horizontal direction: row direction) perpendicular to the line direction (vertical direction) (for example, The image data after the smoothing filter processing is generated by performing the smoothing filter processing of (vertical 1 × horizontal 33 horizontal filter size) (see FIG. 7B). As shown in FIG. 7B, the pixel values of various vertical line-shaped abnormal image elements are, for example, “0” by the smoothing process.

<Step S4> Image Subtraction Processing The image subtraction processing unit 53 includes initial image data obtained by adding the first correction data in step S1, and an image after smoothing filter processing by the horizontal median processing unit 52 in step S3. Difference image data obtained by taking a difference from the data is generated (see FIG. 7C). By this smoothing filter process, the vertical (column direction) line-shaped abnormal image element can be raised to some extent.

<Step S5> Edge Enhancement Processing The edge enhancement processing unit 54 performs edge enhancement processing (for example, 3 × 3 Laplacian processing) on the difference image data in step S4. By this edge emphasis processing, the vertical (column direction) line-shaped abnormal image element can be made to emerge more (see FIG. 7D).

<Step S6> Vertical Line Loss Determination Processing The vertical line loss determination processing unit 55 sets a predetermined threshold value for the difference image data after the edge enhancement processing in step S5, thereby having a pixel value equal to or greater than the threshold value. The image element is determined to be a vertical (column direction) line-shaped abnormal image element (see FIG. 7D).

<Step S <b>7> The vertical line defect information is stored. The defect information storage unit 17 stores, as defect information, position information for identifying the detection element d determined as a vertical (column direction) line-like abnormal image element. That is, the vertical (column direction) line-shaped abnormal image element extracted in step S6 has a pixel value that slightly changes in the positive direction compared to other normal image elements (changes slightly in the positive direction). The position information of the line-shaped abnormal image element in the vertical direction (column direction) is stored in the defect information storage unit 17 in step S7.

<Step S8> Has the black-and-white reversal process been performed?
If the initial image data has not been subjected to black and white reversal processing, the process proceeds to step S9, and if black and white reversal processing has been performed, the process proceeds to step S10. The next black-and-white inversion processing is to treat and treat various vertical line-like abnormal image elements that have negative pixel values when the pixel values of other normal image elements are zero as positive data. Here, initial image data (see FIG. 8A) obtained by adding correction data (second correction data) of various vertical line-like abnormal image elements having a negative pixel value described above is used. To do.

<Step S9> Monochrome Reversal Processing The black and white reversal processing unit 56 performs monochrome reversal processing on the initial image data added with the correction data (second correction data) of various vertical line-like abnormal image elements whose pixel values are negative. Then (see FIG. 8B), the process returns to step S3. In step S3, the horizontal median processing unit 52 performs the above-described smoothing filter processing of the horizontal filter size on the initial image data after reversal that is the initial image data subjected to the black and white reversal processing (FIG. 8 ( In step S4, the image subtraction processing unit 53 obtains a difference between the initial image data after inversion and the inverted image data after smoothing filter processing obtained by performing smoothing filter processing on the initial image data after inversion. In step S5, the edge enhancement processing unit 54 performs edge enhancement processing on the difference image data (see FIG. 8D), and in step S6, the vertical line loss determination processing unit 55 performs edge enhancement. By setting a predetermined threshold value for the processed difference data, an image element having a pixel value equal to or higher than the threshold value is identified as a vertical (column direction) line-shaped abnormal image element. (Refer to FIG. 8 (d)). That is, the determined line-like abnormal image element has a pixel value that slightly changes in the minus direction (those that have a minute fluctuation range in the minus direction) compared to other normal image elements. The positional information of the vertical (column direction) line-like abnormal image element is also stored in the defect information storage unit 17 in step S7.

<Step S10> Is the horizontal line-shaped defect extracted?
If the horizontal (row direction) line-like abnormal image element has not been extracted, the process proceeds to step S11. If the horizontal (row direction) line-like abnormal image element has been extracted, this process ends.

<Step S11> Longitudinal Median Processing The vertical median processing unit 62 adds the correction data (third correction data) of the various horizontal line-shaped abnormal image elements having positive pixel values described above to the initial image data. Thus, it is longer in the direction (vertical direction: column direction) perpendicular to the line direction (horizontal direction) than the line direction (horizontal direction) of the horizontal direction (row direction) line-like abnormal image element to be detected by the FPD 4. Image data after the smoothing filter process is generated by performing a smoothing filter process with a filter size (for example, a vertical filter size of 33 × 1).

<Step S12> Image Subtraction Processing The image subtraction processing unit 63 calculates the difference between the initial image data added with the third correction data and the image data after the smoothing filter processing by the vertical median processing unit 62. Generate image data. By this smoothing filter processing, the horizontal (row direction) line-shaped abnormal image element can be raised to some extent.

<Step S13> Edge Enhancement Processing The edge enhancement processing unit 64 performs edge enhancement processing (for example, 3 × 3 Laplacian processing) on the difference image data. By this edge emphasis processing, the horizontal (row direction) line-shaped abnormal image element can be further highlighted.

<Step S14> Horizontal Line Loss Determination Processing The horizontal line loss determination processing unit 65 sets a predetermined threshold value for the difference image data after the edge enhancement processing, thereby horizontally processing an image element having a pixel value equal to or higher than the threshold value. The direction (row direction) is determined to be a line-like abnormal image element.

<Step S15> Storing horizontal line defect information The defect information storage unit 17 stores, as defect information, positional information that identifies the detection element d determined as a lateral (row direction) line-like abnormal image element. That is, the lateral (row direction) line-like abnormal image element extracted in step S14 has a pixel value that slightly changes in the plus direction (small fluctuation in the plus direction) compared to other normal image elements. The position information of the horizontal (row direction) line-shaped abnormal image element is stored in the defect information storage unit 17 in step S15.

<Step S16> Has the black-and-white reversal processing been performed?
If the initial image data has not been subjected to the monochrome inversion process, the process proceeds to step S17, and if the initial image data has been subjected to the monochrome inversion process, the process proceeds to step S18. The next black-and-white inversion processing is to treat and treat various horizontal line-like abnormal image elements that have negative pixel values when the pixel values of other normal image elements are zero as positive data. Here, it is assumed that initial image data obtained by adding correction data (fourth correction data) of various horizontal line-like abnormal image elements having a negative pixel value described above is used.

<Step S17> Black and White Reversal Processing The black and white reversal processing unit 66 performs black and white reversal processing on the initial image data added with the fourth correction data, and returns to step S11. In step S11, the vertical median processing unit 62 performs the above-described smoothing filter processing of the vertical filter size on the inverted initial image data that is the initial image data subjected to the black and white inversion processing, and then proceeds to step S12. Then, the image subtraction processing unit 63 takes the difference between the inverted initial image data and the inverted image data after smoothing filter processing obtained by performing smoothing filter processing on the inverted initial image data, and in step S13, edge enhancement is performed. The processing unit 64 performs edge enhancement processing on the difference image data. In step S14, the horizontal line loss determination processing unit 65 performs predetermined processing on the inverted image data after the smoothing filter processing subjected to the edge enhancement processing. By setting a threshold value, an image element having a pixel value equal to or greater than the threshold value is identified as a horizontal (row direction) line-like abnormal image element. To. That is, the determined line-like abnormal image element has a pixel value that slightly changes in the minus direction (those that have a minute fluctuation range in the minus direction) compared to other normal image elements. The positional information of the horizontal (row direction) line-like abnormal image element is also stored in the defect information storage unit 17 in step S15.

<Step S18> Has vertical line-shaped defects been extracted?
If the vertical (column direction) line-like abnormal image element has not been extracted, the process returns to step S3, and if the vertical direction (column direction) line-like abnormal image element has been extracted, this process ends.

  Here, FIG. 9A shows the vertical line defect extraction result in the initial image data added with the first correction data. As shown in FIG. 9A, for the vertical line defect, when edge enhancement is not performed, 14 pixels can be extracted with the pixel value “+4”, the extraction rate is 58%, and the pixel value is “ 23 are extracted at “+5”, the extraction rate is 96%, and all the pixel values are extracted from “+6” to “+30”. Further, in the case of edge enhancement, all the pixel values “+4” to “+30” can be extracted, and it can be seen that the edge enhancement is superior to that without edge enhancement.

  In addition, the vertical line defect extraction result in the initial image data to which the second correction data described above is added is also the same as that in FIG.

  Further, FIG. 9B shows a horizontal line defect extraction result in the initial image data added with the third correction data. As shown in FIG. 9B, for the horizontal line defect, when there is no edge enhancement, all pixel values cannot be extracted with the pixel values “+4” to “+30”. Two pixels can be extracted with the value “+5”, the extraction rate is 8%, three pixels can be extracted with the pixel values “+6” and “+7”, and the extraction rate is 13%. 19 values can be extracted with a value of “+10”, the extraction rate is 79%, all pixel values of “+30” can be extracted, and edge enhancement is better than without edge enhancement. I understand that.

  Note that the vertical line defect extraction result in the initial image data to which the above-described fourth correction data is added is also the same as in FIG. 9B.

  The reason why the horizontal line extraction rate is lower than that of the vertical line is that the evaluation image has horizontal line-shaped low-frequency noise. The smaller the line pixel value, the more the noise is buried and the extraction is impossible. It has become.

  From the vertical line loss result shown in FIG. 9A, in the future FPD4, when vertical line loss is extracted with edge enhancement, the threshold value in the vertical line loss determination processing unit 55 is set to “less than 4”. (For example, “3”) may be set uniformly, and the vertical line defect extraction of the future FPD 4 may be performed. Similarly, from the horizontal line loss result shown in FIG. 9B, in the future FPD4, when horizontal line loss is extracted with edge enhancement, the threshold in the horizontal line loss determination processing unit 65 is set to “30”. Less than "(for example," 29 ") may be set uniformly, and lateral line defect extraction of future FPD4 may be performed.

  As described above, the detection method of the line-shaped abnormal image element of the radiation detector according to the first embodiment, that is, a plurality of detection elements d (image elements) for detecting X-rays irradiated by the X-ray tube 3 are two-dimensionally arranged. According to the detection method of the line-like abnormal image element of the radiation detector that detects the line-like abnormal image element in the FPD 4 (radiation detector) that has been detected, detection by the FPD 4 when the detection surface of the FPD 4 is irradiated with radiation By performing smoothing filter processing for the initial image data created based on the output signal, the filter size is longer in the direction perpendicular to the line direction than the line direction of the line-shaped abnormal image element to be detected. A smoothing process step (steps S3 and S11) for generating image data after the smoothing filter process, an initial image data and an image after the smoothing filter process A difference processing step (steps S4 and S12) for generating difference image data obtained by taking a difference from the data, an edge enhancement processing step (steps S5 and S13) for performing edge enhancement processing on the difference image data, and edge enhancement processing And a determination processing step (steps S6 and S14) for determining an image element having a pixel value equal to or higher than the threshold value by setting a predetermined threshold value for the subsequent difference image data as a line-like abnormal image element. Yes. Therefore, in the edge enhancement processing step, since the edge enhancement processing is performed on the difference image data, it is possible to generate the difference image data after the edge enhancement processing in which the line-like abnormal image element is further enhanced and emerged. In the step, the threshold value is set for the differential image data after the edge enhancement process that further highlights and highlights the line-like abnormal image element, so that the line-like abnormal image element (line-like defect) having only a minute change is set. ) Can be extracted (detected) more precisely. In other words, the line-shaped abnormal image element (line-shaped defect) is more accurately extracted (detected) by setting the threshold value for the differential image data after the edge enhancement process than when the threshold value is set for the differential image data without the edge enhancement process. can do. Then, by correcting the radiation detection signal obtained from the line-like abnormal image element, a high-quality radiation image can be acquired.

  In addition, when detecting a vertical line-shaped abnormal image element, in the smoothing processing step, the image after smoothing filter processing is performed by performing smoothing filter processing with a long filter size in the horizontal direction on the initial image data. In the edge emphasis processing step, the image data after the smoothing filter processing is subjected to edge emphasis processing, and the determination processing step generates a horizontal profile for the difference image data after the edge emphasis processing. Since an image element having a pixel value equal to or greater than the vertical line threshold value is determined as a vertical line-shaped abnormal image element, even a vertical line-shaped abnormal image element having only a minute change is extracted (detected). Can do.

  Further, when detecting a horizontal line-shaped abnormal image element, in the smoothing processing step, the smoothed filter processing is performed by performing smoothing filter processing with a filter size that is long in the vertical direction on the initial image data. In the edge enhancement processing step, the image data after the smoothing filter processing is subjected to edge enhancement processing, and the determination processing step generates a vertical profile for the difference image data after the edge enhancement processing. An image element having a pixel value equal to or greater than the threshold value for the horizontal line is determined as a horizontal line-shaped abnormal image element, so that even a horizontal line-shaped abnormal image element having only a minute change is extracted (detected). Can do.

  In the black and white reversal processing steps (steps S9 and S17), the initial image data is subjected to black and white reversal processing. In the difference processing step, the difference between the inverted initial image data and the inverted image data after the smoothing filter process obtained by performing the smoothing filter processing on the inverted initial image data is obtained. By setting a predetermined threshold value for the inverted image data after the filter processing, an image element having a pixel value equal to or higher than the threshold value is determined as a line-like abnormal image element. Therefore, a group of normal images in the initial image data Line-like abnormal image element having a pixel value lower than that of the element (that is, a line-like abnormal image element having a negative pixel value in the difference processing) ) It can also be extracted.

  As described above, according to the fluoroscopic imaging apparatus according to the first embodiment, the FPD 4 (radiation detector) in which a plurality of detection elements (image elements) for detecting X-rays emitted from the X-ray source 3 are two-dimensionally arranged. ) And a defect information storage unit 17 for storing position information data of the line-shaped abnormal image element of the radiation detector used in the apparatus detected by the above-described line-shaped abnormal image element detection method, and the defect information Since the correction unit 21 for correcting the radiation detection signal of the line-like abnormal image element of the FPD 4 using the positional information data of the line-like abnormal image element stored in the storage unit 17 is provided, the line-like abnormal image of the FPD 4 Correction of the radiation detection signal of the element can be suitably performed.

  In the above-described embodiment, as the correction data of the fluoroscopic imaging apparatus, data of various vertical line-like abnormal image elements as shown in FIG. 6 is created in advance, and the initial image data including the correction data is added. However, it goes without saying that the vertical line and horizontal line abnormal image elements can be extracted from the initial image data obtained by photographing the uniform irradiation image with the actual FPD 4 without using such correction data. .

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

  (1) In the above-described embodiment, a predetermined threshold is set for the differential image data after the edge enhancement processing, and the vertical (column direction) and horizontal (row direction) line-shaped abnormal image elements are extracted (detected). Although the extraction accuracy of line-like abnormal image elements is improved, a predetermined threshold is set for differential image data that is not subjected to edge enhancement, and vertical (column direction) and horizontal (row direction) lines are set. The abnormal state image element may be extracted (detected). In this case, the extraction accuracy of the line-like abnormal image element is inferior to that in the case of edge enhancement processing, but after the square box filter processing in which the initial image data is subjected to the smoothing filter processing by the square (n × n) box filter. Compared with the case where a threshold value is set for difference image data obtained by taking the difference between the first image data and the initial image data, the extraction accuracy of the line-like abnormal image elements is high.

  (2) In the above-described embodiment, as shown in FIG. 1, the vertical line defect extraction processing unit 14 and the horizontal line defect extraction processing unit 15 are provided, and the vertical (column direction) and horizontal (row direction) line shapes are provided. Although abnormal image elements are extracted (detected), when only vertical (column direction) line-shaped abnormal image elements need to be extracted, only the vertical line defect extraction processing unit 14 may be provided. When only the extraction of the direction (row direction) line-like abnormal image element is required, only the horizontal line defect extraction processing unit 15 may be provided.

  (3) In the above-described embodiment apparatus, as shown in FIG. 1, the line-like abnormality detection element extraction unit is provided, but as shown in FIG. 10, the line-like abnormality detection element extraction unit is not provided. The position information data of the line-shaped abnormal image element of the radiation detector used in the apparatus detected by the detection method of the line-shaped abnormal image element may be input and stored in the defect information storage unit 17. .

  (4) In the above-described embodiment, the FPD 4 is taken as an example of the radiation detection means. However, the present invention can be applied to any radiation detector having a plurality of detection elements on the detection surface.

  (5) In the above-described embodiment, as shown in FIG. 11, the stabilization processing units 57 and 67 (constant stabilization) that stabilize the negative pixel value in the difference image data after the edge enhancement processing to zero or a positive value. The vertical line loss determination processing unit 55 and the horizontal line loss determination processing unit 65 (determination processing step) set a predetermined threshold value for the difference image data after the edge enhancement processing that has been made constant. Thus, a plurality of line-shaped image elements in which image elements having pixel values equal to or greater than the threshold value are continuous may be determined as line-like abnormal image elements.

  In this case, in the stabilization processing units 57 and 67, the negative pixel value in the difference image data after the edge enhancement processing is fixed to zero or a positive value, and the vertical line loss determination processing unit 55 and the horizontal line loss determination are performed. The processing unit 65 sets a predetermined threshold value for the difference image data after the edge enhancement processing that has been subjected to the stabilization processing, thereby a plurality of line-shaped image elements in which image elements having pixel values equal to or greater than the threshold value are continuous. Is determined as a line-like abnormal image element, an image element having a negative pixel value in the edge enhancement process can be handled as a positive value.

  (6) In the above-described embodiment, in order to extract the vertical (column direction) line-like abnormal image elements, the horizontal median processing unit 52 has a filter size (for example, vertical 1 × horizontal) that is long in the horizontal direction (row direction). 33 horizontal median filter processing), and the vertical median processing unit 62 extracts a filter size that is long in the vertical direction (column direction) in order to extract horizontal abnormal image elements in the horizontal direction (row direction). The vertical median filter processing (vertical 33 × horizontal 1 vertical filter size) is performed, but a long moving average processing or the like may be performed.

  (7) In the above-described embodiment, the medical X-ray fluoroscopic imaging apparatus is used. However, the present invention is not limited to this. In other words, the present invention can be applied to apparatuses using radiation other than X-rays, and is also applicable to radiation imaging apparatuses used in industrial fields such as non-destructive inspection, RI (Radio Isotope) inspection, and optical inspection, and in the nuclear field. Is also applicable.

1 is a block diagram illustrating an overall configuration of an X-ray fluoroscopic imaging apparatus according to Embodiment 1. FIG. It is a vertical sectional view of the principal part of FPD. It is a top view of the principal part of FPD. It is a block diagram which shows the structure of the line-shaped abnormality detection element extraction part of Example 1. FIG. 6 is a flowchart illustrating processing for detecting a line-like abnormal image element of an FPD according to the first embodiment. It is a figure which shows the correction data for making various vertical line-shaped abnormal image elements. (A) is initial image data spatially developed according to the position of each detection element, (b) is image data obtained by performing smoothing filter processing, (c) is difference data, and (d) is after edge enhancement. It is the figure which showed each profile of difference data typically. (A) is initial image data when there is a negative line loss, (b) is data obtained by performing black-and-white reversal processing on the initial image data of (a), (c) is image data obtained by subjecting the initial image data to smoothing filter processing, ( FIG. 6D is a diagram schematically showing the profiles of difference data after edge enhancement. (A) is a figure which shows a vertical line defect | deletion extraction result, (b) is a figure which shows a horizontal line defect | deletion extraction result. It is a block diagram which shows the whole structure of the X-ray fluoroscopic imaging apparatus of a modification. It is a block diagram which shows the structure of the line-shaped abnormality detection element extraction part of a modification.

Explanation of symbols

3 ... X-ray tube (radiation source)
4 ... FPD (radiation detector)
DESCRIPTION OF SYMBOLS 13 ... Line-shaped abnormality detection element extraction part 14 ... Vertical line defect | deletion extraction process part 15 ... Horizontal line defect | deletion extraction process part 17 ... Defect information storage part (memory | storage means)
21 ... Correction unit (correction processing means)
52. Horizontal median processing unit (smoothing processing unit)
53 ... Image subtraction processing unit (difference processing unit)
54 ... Edge enhancement processing unit 55 ... Vertical line loss determination processing unit (determination processing unit)
56... Black and white inversion processing unit 62... Vertical median processing unit (smoothing processing unit)
63 ... Image subtraction processing unit (difference processing unit)
64 ... Edge enhancement processing unit 65 ... Horizontal line loss determination processing unit (determination processing unit)
66 ... black and white inversion processing section

Claims (7)

A detection method of a line-shaped abnormal image element of a radiation detector, detecting a line-shaped abnormal image element in a radiation detector in which a plurality of image elements for detecting radiation irradiated by a radiation source are two-dimensionally arranged,
Compared to the line direction of the line-shaped abnormal image element to be detected with respect to the initial image data created based on the detection output signal at the radiation detector when the detection surface of the radiation detector is irradiated with radiation Smoothing processing step of generating image data after smoothing filter processing by performing smoothing filter processing with a long filter size in a direction orthogonal to the line direction;
A difference processing step for generating difference image data obtained by taking a difference between the initial image data and the image data after the smoothing filter processing;
A determination processing step of determining a plurality of line-shaped image elements in which image elements having pixel values equal to or greater than the threshold value are continuous as line-shaped abnormal image elements by setting a predetermined threshold value for the difference image data;
A method of detecting a line-shaped abnormal image element of a radiation detector, comprising: In the detection method of the line-shaped abnormal image element of the radiation detector according to claim 1,
An edge enhancement processing step of performing edge enhancement processing on the difference image data;
In the determination processing step, a predetermined threshold is set for the difference image data after the edge enhancement processing, whereby a plurality of line-shaped image elements in which image elements having pixel values equal to or higher than the threshold are continuous are formed in a line shape. A method for detecting a line-shaped abnormal image element of a radiation detector, characterized by determining an abnormal image element. In the detection method of the line-shaped abnormal image element of the radiation detector according to claim 2,
When the image elements are arranged vertically and horizontally in the radiation detector, and the line-like abnormal image element to be detected is a vertical line-like abnormal image element, the smoothing processing step includes the initial image data. The image data after smoothing filter processing is generated by performing smoothing filter processing with a filter size that is long in the horizontal direction,
The determination processing step generates a horizontal profile for the difference image data after the edge enhancement processing, and determines an image element having a pixel value equal to or greater than a vertical line threshold in the profile as a vertical line-shaped abnormal image element. A method for detecting a line-shaped abnormal image element of a radiation detector. In the detection method of the line-shaped abnormal image element of the radiation detector according to claim 2,
When the image elements are arranged vertically and horizontally in the radiation detector, and the line-shaped abnormal image element to be detected is a horizontal line-shaped abnormal image element, the smoothing processing step includes the initial image data. The image data after smoothing filter processing is generated by performing smoothing filter processing with a long filter size in the vertical direction,
The determination processing step generates a vertical profile for the difference image data after the edge enhancement processing, and determines an image element having a pixel value equal to or greater than a horizontal line threshold in the profile as a horizontal line abnormal image element. A method for detecting a line-shaped abnormal image element of a radiation detector. In the detection method of the line-shaped abnormal image element of the radiation detector according to any one of claims 1 to 4,
Further comprising a black-and-white reversal processing step of black-and-white reversal processing the images data,
A detection method for a line-like abnormal image element of a radiation detector , wherein the initial image data in the smoothing step, the difference processing step, and the determination processing step is subjected to the black-and-white reversal processing . In the detection method of the line-shaped abnormal image element of the radiation detector according to any one of claims 1 to 5,
Further comprising a constant process step for constant the negative pixel values in images data zero or a positive value,
The method for detecting a line-like abnormal image element of a radiation detector, wherein the difference image data in the determination processing step is obtained by performing a stabilization process on the difference image data in the difference processing step . In a radiation imaging apparatus comprising a radiation detector in which a plurality of image elements for detecting radiation emitted by a radiation source are two-dimensionally arranged,
Position information of the line abnormal image element of the radiation detector used in the apparatus detected by the method of detecting a line abnormal image element of the radiation detector according to any one of claims 1 to 6. Storage means for storing data;
Correction processing means for correcting the radiation detection signal of the line-like abnormal image element of the radiation detector using the positional information data of the line-like abnormal image element stored in the storage means;
A radiation imaging apparatus comprising:

Images