Device for evaluating quality of radiographic testing image and method therefor

Abstract

To provide a device and method for counting the number of wires of Image quality Indicators (IQI) by image processing in order to evaluate the quality of a radiographic testing film image.SOLUTION: A method for evaluating quality of a radiographic testing image includes: when a wire area image is input which is an image of an area where wires of image quality indicators exist in a radiographic image of a target object with one or more image quality indicators attached thereto through radiation, a step of preprocessing, by a preprocessing unit, the wire area image; a step of binarizing, by a binarization unit, the wire area image; a step of detecting, by a line detection unit, lines in the wire area image; a step of counting, by a counting unit, the number of wires of the image quality indicators according to the detected lines; and a step of evaluating, by an evaluation unit, the quality of the radiographic image according to the number of counted wires.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present invention relates to image quality assessment techniques, and more particularly to an apparatus and method for assessing the quality of radiographic inspection film images. [Background technology]

[0002] Image Quality Indicator (IQI) testing is a procedure used to evaluate the quality of images taken by non-destructive radiographic testing and to check the image resolution and clarity. It involves a skilled inspector inspecting the quality of analog radiographic images taken directly with the naked eye. Summary of the Invention [Problem to be solved by the invention]

[0003] SUMMARY OF THE INVENTION It is an object of the present invention to provide an apparatus and method for assessing the quality of radiographic inspection film images. [Means for solving the problem]

[0004] To achieve the above object, a method for evaluating the quality of a radiographic inspection film image according to a preferred embodiment of the present invention includes the steps of: when a wire area image, which is an image of an area where one or more image quality indicators (IQIs) are present, is input from a radiographic image of an object having one or more image quality indicators (IQIs) attached thereto, a pre-processing unit pre-processing the wire area image; a binarization unit binarizing the wire area image; a line detection unit detecting lines from the wire area image; a counting unit counting the number of wires of the radiographic inspection film along the detected lines; and an evaluation unit evaluating the quality of the radiographic image according to the number of wires counted.

[0005] The step of detecting the wire lines includes the steps of: the line detection unit converting pixels of the wire area image into a polar coordinate system to generate a plurality of curves; configuring points where the generated curves intersect as an accumulation array; and detecting a plurality of lines from a plurality of points whose accumulation array values ​​are equal to or greater than a threshold; and the line detection unit clustering the detected plurality of lines based on the density of the detected plurality of lines, and clustering the detected plurality of lines into one or more line clusters.

[0006] The step of counting the number of wires of the penetrometer may be characterized in that the counting unit counts the number of wires of the penetrometer in accordance with the number of the line clusters.

[0007] The step of preprocessing the wire area image includes a step of normalizing the wire area image by the preprocessing unit so that the wire area image is converted into a predetermined format, and a step of performing image processing on the wire area image by the preprocessing unit so that a difference between pixels constituting a line in the wire area image and the remaining pixels becomes more noticeable.

[0008] The step of normalizing the wire region image includes the steps of: rotating the wire region image in a preset direction by the pre-processing unit; and adjusting the wire region image to a preset size by the pre-processing unit.

[0009] The image processing step includes the steps of: performing contrast limited adaptive histogram equalization (CLAHE) processing on the wire region image by the pre-processing unit to remove noise from the radiation image and increase brightness and contrast so as to increase the clarity of lines present in the wire region image; filtering the wire region image using one or more filters in the frequency domain by the pre-processing unit; and filtering the wire region image using a Sobel filter in the upward and downward directions based on the Y-axis by the pre-processing unit so as to detect boundaries of vertical components of lines present in the wire region image.

[0010] The step of filtering using one or more filters includes the steps of: converting the wire-domain image into a frequency domain by the pre-processing unit to generate a frequency-domain image; performing frequency-domain filtering using a composite filter that combines a low-pass filter that passes only low-band components of the frequency-domain image; a vertical weighting filter that multiplies vertical components in the frequency-domain image by a predetermined weight; a sandclock-shaped filter that eliminates upper and lower regions of a region divided by two different line segments that pass through a reference point vertically symmetrically with respect to a reference point in the frequency-domain image; and a Butterworth filter that minimizes phase distortion in a boundary region of a frequency to be filtered in the frequency-domain image; and the step of inversely converting the frequency-domain image from the frequency domain to a time domain by the pre-processing unit.

[0011] The Butterworth filter is Formula

number

[0012] The binarization unit binarizing the wire region image includes binarizing the wire region image by using a predetermined ratio of upper pixel values ​​in a distribution of pixel values ​​of the wire region image as a threshold.

[0013] The step of binarizing the wire region image by the binarization unit may further include a step of performing a morphology operation on the wire region image by the binarization unit after the erasing step, and the morphology operation may include at least one of an erosion operation for removing noise from the wire region image and a dilation operation for dilating lines in the wire region image.

[0014] The step of evaluating the quality of the radiation image is characterized in that the evaluation unit determines that the quality of the radiation image is suitable for defect inspection if the number of counted wires is equal to or greater than a reference value of a predefined standard, and determines that the quality of the radiation image is unsuitable for defect inspection if the number of counted wires is less than the reference value of a predefined standard.

[0015] To achieve the above object, according to a preferred embodiment of the present invention, an apparatus for evaluating the quality of a radiographic inspection film image includes: a pre-processing unit that pre-processes a wire area image, which is an image of an area where one or more image quality indicators (IQIs) are present, among a radiographic image of an object having one or more image quality indicators (IQIs) attached thereto; a binarization unit that binarizes the wire area image; a line detection unit that detects lines from the wire area image; a counting unit that counts the number of wires of the radiographic inspection film along the detected lines; and an evaluation unit that evaluates the quality of the radiographic image according to the number of wires counted.

[0016] The line detection unit converts pixels of the wire area image into a polar coordinate system to generate a plurality of curves, configures points where the generated curves intersect as an accumulation array, detects a plurality of lines from a plurality of points where the accumulation array value is equal to or greater than a threshold, and clusters the detected plurality of lines into one or more line clusters by clustering the detected plurality of lines based on their density.

[0017] The counting unit counts the number of wires of the penetrometer according to the number of the line clusters.

[0018] The pre-processing unit normalizes the wire area image so that the wire area image is converted into a predetermined format, and performs image processing on the wire area image so that a difference between pixels constituting a line and the remaining pixels in the wire area image becomes more noticeable.

[0019] The pre-processing unit may rotate the wire area image in a predetermined direction and adjust the wire area image to a predetermined size.

[0020] The pre-processing unit performs a contrast limited adaptive histogram equalization (CLAHE) process on the wire region image to remove noise from the radiation image and increase brightness and contrast so as to increase the clarity of lines present in the wire region image, performs filtering using one or more filters in the frequency domain of the wire region image, and performs filtering using a Sobel filter in the upward and downward directions based on the Y-axis on the wire region image so as to detect boundaries of vertical components of lines present in the wire region image.

[0021] The pre-processing unit converts the wire-domain image into a frequency domain to generate a frequency-domain image, performs frequency-domain filtering using a composite filter that combines a low-pass filter that passes only low-band components of the frequency-domain image, a vertical weighting filter that multiplies vertical components in the frequency-domain image by a predetermined weight, a sandclock-shaped filter that eliminates areas included within a predetermined angle symmetrically above and below a reference point in the frequency-domain image, and a Butterworth filter that minimizes phase distortion in boundary areas of frequencies filtered in the frequency-domain image, and then inversely converts the frequency-domain image from the frequency domain to a time domain.

[0022] The Butterworth filter is Formula

number

[0023] The binarization unit binarizes the wire area image using a predetermined ratio of upper pixel values ​​in a distribution of pixel values ​​of the wire area image as a threshold.

[0024] The binarization unit performs a morphology operation on the wire region image, and the morphology operation includes at least one of an erosion operation for additionally removing noise from the wire region image and a dilation operation for dilating lines in the wire region image.

[0025] The evaluation unit may determine that the quality of the radiation image is suitable for defect inspection if the number of counted wires is equal to or greater than a predefined reference value, and may determine that the quality of the radiation image is unsuitable for defect inspection if the number of counted wires is less than the predefined reference value. [Effects of the Invention]

[0026] According to the present invention, while an inspector evaluates by the naked eye, the number of wires can be counted by an image quality indicator (IQI) using image processing, thereby shortening the inspection time and obtaining objective results. [Brief explanation of the drawings]

[0027] [Figure 1] 1 is a diagram illustrating the configuration of an apparatus for evaluating the quality of a radiographic inspection film image according to an embodiment of the present invention. [Figure 2] 1 is a diagram illustrating a detailed configuration of an apparatus for evaluating the quality of a radiographic inspection film image according to an embodiment of the present invention; [Figure 3] 1 is a diagram illustrating a detection model for evaluating the quality of a radiographic inspection film image according to an embodiment of the present invention. [Figure 4] 1 is a flowchart illustrating a method for generating a detection model according to an embodiment of the present invention. [Figure 5] 1 is a flowchart illustrating a method for digitizing a radiographic inspection film according to an embodiment of the present invention. [Figure 6] 1 is a flowchart illustrating a method for assessing the quality of radiographic inspection film images based on a learning model according to an embodiment of the present invention. [Figure 7] 10 is an example of a screen for explaining a method for identifying a wire region according to an embodiment of the present invention. [Figure 8]1 is a flowchart illustrating a method for counting the number of wires included in a wire area image according to an embodiment of the present invention; [Figure 9] 10 is an example of a screen illustrating a method for counting the number of wires included in a wire area image according to an embodiment of the present invention; [Figure 10] 10 is an example of a screen illustrating a method for counting the number of wires included in a wire area image according to an embodiment of the present invention; [Figure 11] 10 is an example of a screen illustrating a method for counting the number of wires included in a wire area image according to an embodiment of the present invention; [Figure 12] 10 is an example of a screen illustrating a method for counting the number of wires included in a wire area image according to an embodiment of the present invention; [Figure 13] 10 is an example of a screen illustrating a method for counting the number of wires included in a wire area image according to an embodiment of the present invention; [Figure 14] 10 is an example of a screen illustrating a method for counting the number of wires included in a wire area image according to an embodiment of the present invention; [Figure 15] 10 is an example of a screen illustrating a method for counting the number of wires included in a wire area image according to an embodiment of the present invention; [Figure 16] 10 is an example of a screen illustrating a method for counting the number of wires included in a wire area image according to an embodiment of the present invention; [Figure 17] 10 is an example of a screen illustrating a method for counting the number of wires included in a wire area image according to an embodiment of the present invention; [Figure 18] FIG. 1 illustrates a computing device according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0028] Although the present invention can be implemented in various forms by adding various modifications, specific embodiments will be illustrated and described in detail in the detailed description, but it should be understood that this is not intended to limit the present invention to the specific embodiments, but rather to include all modifications, equivalents, or alternatives that fall within the spirit and technical scope of the present invention.

[0029] The terms used in the present invention are merely used to describe specific embodiments and are not intended to limit the present invention. A singular expression includes a plural expression unless the context clearly indicates otherwise. It should be understood that, in the present invention, terms such as "comprise" or "have" are intended to specify the presence of features, numbers, steps, operations, components, parts, or combinations thereof described in the specification, and do not preclude the possibility of the presence or addition of one or more other features, numbers, steps, operations, components, parts, or combinations thereof.

[0030] In particular, the terms and words used in the following specification and claims should not be interpreted as being limited to their ordinary or dictionary meanings, but should be interpreted as meanings and concepts that correspond to the technical idea of ​​the present invention, in accordance with the principle that the inventor can appropriately define the concepts of terms to best describe his / her invention. In particular, in the embodiments of the present invention, estimation means deriving a calculation result based on what a machine learning model (LM: Machine Learning Model / Deep Learning Model) has learned.

[0031] First, an apparatus for evaluating the quality of a radiographic inspection film image according to an embodiment of the present invention will be described. Fig. 1 is a diagram illustrating the configuration of an apparatus for evaluating the quality of a radiographic inspection film image according to an embodiment of the present invention. Fig. 2 is a diagram illustrating the detailed configuration of an apparatus for evaluating the quality of a radiographic inspection film image according to an embodiment of the present invention. Fig. 3 is a diagram illustrating a detection model for evaluating the quality of a radiographic inspection film image according to an embodiment of the present invention.

[0032] An apparatus (10, hereinafter referred to as the “evaluation apparatus”) for evaluating the quality of a radiographic inspection film image according to an embodiment of the present invention includes a radiographic test (RT) device 20 that performs radiography on an object having an image quality indicator (IQI) attached thereto, extracts a radiographic inspection film on which the object having the IQI attached is imaged, and a scanner 30 that scans the radiographic inspection film to generate a digitized radiographic image. The apparatus then analyzes the generated radiographic image to evaluate whether the quality of the radiographic image is suitable for performing a defect inspection on the object imaged in the radiographic image. To this end, the evaluation apparatus 10 includes a learning unit 100, a data processing unit 200, a detection unit 300, a calculation unit 400, and an evaluation unit 500.

[0033] The learning unit 100 is for generating a detection model DM through learning (deep learning or machine learning). Examples of the detection model DM include CNN, YOLO, RCNN, and Faster RCNN. The detection model DM includes multiple layers (or modules) connected to each other, and the multiple layers (or modules) are performed with multiple calculations. The multiple layers (or modules) are also connected with weights (W). That is, the output resulting from the calculation of any one layer (or module) is input to the calculation of the next layer after a weight is applied. The detection model DM derives an output by performing multiple calculations on input data to which weights are applied between multiple layers (or modules). That is, multiple calculations are performed that are connected by weights between multiple layers (or modules). Such multiple calculations that are connected by weights between multiple layers (or modules) of the detection model DM are referred to as "weight calculations."

[0034] 3, the transmittance meter has a text region consisting of text and a wire region consisting of a plurality of wires. When a radiation image RI is input to the detection model DM, the training unit 100 trains the detection model DM so that the detection model DM detects the text region, which is the region occupied by the transmittance meter text, in the input radiation image RI via a bounding box BB. When training of the detection model DM is completed, the training unit 100 provides the trained detection model DM to the detection unit 300.

[0035] The data processing unit 200 receives a radiation image, which is a digitized radiation image of a radiographic inspection film on which an object having a radiometer attached thereto is photographed, and information about the object corresponding to the radiation image, maps the input radiation image and information about the object, and stores the mapped information in a database (DB). The data processing unit 200 can also load the radiation image and information about the mapped object from the database (DB) for quality evaluation of the radiation image.

[0036] The detection unit 300 can detect a translucent text region through a bounding box in the radiation image RI using the detection model DM. Referring to FIG. 3, when the detection unit 300 inputs the radiation image RI to the detection model DM, the detection model DM performs a weighting operation to which learned weights are applied to derive a detection vector. The detection vector predicts an area occupied by the translucent text through a bounding box (e.g., BB) in the training radiation image. Thus, the detection unit 300 can detect the text region through the bounding box (e.g., BB). The detection unit 300 can identify an area having a predetermined width and height below the bounding box BB indicating the text region based on the direction of the detected translucent text as the wire region ROI.

[0037] The calculation unit 400 detects a wire region ROI, generates a wire region image, and counts the number of wires included in the generated wire region image. To this end, the calculation unit 400 includes a pre-processing unit 410, a binarization unit 420, a line detection unit 430, and a counting unit 440, as shown in Fig. 2. Specific operations of the calculation unit 400 including the pre-processing unit 410, the binarization unit 420, the line detection unit 430, and the counting unit 440 will be described in more detail below.

[0038] The evaluation unit 500 determines the quality of the radiation image according to the number of wires counted by the calculation unit 400. If the number of counted wires is equal to or greater than a predetermined standard, the evaluation unit 500 determines that the quality of the radiation image is suitable for defect inspection. On the other hand, if the number of counted wires is less than the predetermined standard, the evaluation unit 500 may determine that the quality of the radiation image is unsuitable for defect inspection. That is, the better the quality of the radiation image, the more wires are identified as genuine, and the worse the quality of the radiation image, the fewer wires are identified as genuine. Thus, the quality of the radiation image can be determined by comparing the number of counted wires with a predetermined standard and determining whether the number of wires identified as genuine is equal to or greater than the predetermined standard. After determining the quality of the radiation image as described above, if the evaluation unit 500 determines that the quality of the radiation image is suitable for defect inspection, the evaluation unit 500 analyzes the radiation image and outputs an inspection command to start defect inspection. On the other hand, if the quality of the radiation image is determined to be inappropriate for defect inspection, the evaluation unit 500 can output a re-imaging command to perform re-imaging.

[0039] Next, a method for generating a detection model according to an embodiment of the present invention will be described below. Fig. 4 is a flowchart illustrating a method for generating a detection model according to an embodiment of the present invention.

[0040] 3 and 4, the learning unit 100 loads pre-defined learning data in step S110. The learning data includes a learning radiation image RI and a target vector. The learning radiation image RI is an image of an object with at least one transmissometer attached. The target vector represents a target bounding box (e.g., GT) that indicates an area occupied by text on the transmissometer in the learning radiation image. For example, the target vector may represent the bounding box using the center coordinates, width, and height of the bounding box.

[0041] Next, in step S120, the training unit 100 inputs the training radiation image to the detection model DM having untrained weights. Then, in step S130, the detection model DM performs a weighting operation to apply the untrained weights to the training radiation image to derive a detection vector. The detection vector is a prediction of the area occupied by the translucency meter text through a bounding box (e.g., BB) in the training radiation image. As shown in FIG. 3, the detection vector may be a bounding box (e.g., BB) indicating the area occupied by the text in the training radiation image (e.g., RI). For example, the detection vector includes the center coordinates, width, and height of the bounding box indicating the area occupied by the translucency meter text.

[0042] Next, in step S140, the learning unit 100 calculates a loss indicating the difference between the target vector and the detection vector using a loss function. After that, in step S150, the learning unit 100 performs optimization to correct the weights of the detection model DM so that the loss derived by the loss function is minimized.

[0043] Next, in step S160, the learning unit 100 determines whether a learning completion condition is satisfied. According to one embodiment, the learning completion condition may be when a predetermined learning rate is exceeded and the loss calculated previously (S140) converges to a preset target value or less. If the learning completion condition is not satisfied as a result of the determination in step S160, the learning unit 100 repeats the above-described steps S120 to S150 using multiple different training data. On the other hand, if the learning completion condition is satisfied as a result of the determination in step S160, the learning unit 100 completes learning of the detection model DM in step S170.

[0044] Next, a method for digitizing a radiographic test (RT) film according to an embodiment of the present invention will be described. Fig. 5 is a flowchart illustrating a method for digitizing a radiographic test film according to an embodiment of the present invention.

[0045] 5, in step S210, a radiographic test (RT) device 20 performs radiography on an object having an image quality indicator (IQI) attached thereto, and extracts a radiographic test film capturing an image of the object having the IQI attached thereto. The object may be a pipe, a tube, or the like.

[0046] Then, in step S220, the scanner 30 scans the radiographic inspection film to generate a digitized radiographic image.

[0047] After the radiation image is generated, the data processing unit 200 may input the digitized radiation image and information about the object that is the basis of the radiation image in step S230. Here, the information about the object may include facility information, the size of the object (e.g., pipe, tube), the imaging position, imaging radiation conditions, etc. In particular, the information about the object may include at least the size of the object (e.g., pipe, tube), and the size of the object may include the shape or thickness of the object.

[0048] At this time, the digitized radiation image may be input directly from the scanner 30 or input via other media. Also, information about the object of the radiation image may be input by the user. When the digitized radiation image and information about the object of the radiation image are input in this manner, the data processing unit 200 can map the radiation image and information about the object corresponding to the radiation image and store the mapped information in the database DB in step S240.

[0049] Next, a method for evaluating the quality of a radiographic inspection film image according to an embodiment of the present invention will be described. Fig. 6 is a flowchart for explaining a method for evaluating the quality of a radiographic inspection film image based on a learning model according to an embodiment of the present invention. Fig. 7 is an example of a screen for explaining a method for identifying a wire region according to an embodiment of the present invention.

[0050] 6, the data processing unit 200 loads a radiation image from the database DB in step S310. The radiation image RI is a digitized image of a radiographic inspection film on which an object having an image quality indicator (IQI) attached is photographed. Therefore, the radiation image is also an image of an object having an image quality indicator (IQI) attached.

[0051] Next, in step S320, the detection unit 300 detects the text region of the translucent meter through a bounding box in the radiation image RI using the detection model DM. That is, when the detection unit 300 inputs the radiation image RI to the detection model DM, the detection model DM performs a weighting operation to which the learned weights are applied to derive a detection vector. As shown in FIG. 3, the detection vector is a prediction of the region occupied by the text of the translucent meter through a bounding box (e.g., BB) in the training radiation image. As a result, the detection unit 300 can detect the text region through the bounding box (e.g., BB).

[0052] Next, in step S330, the detection unit 300 determines a region having a predetermined width and height below a bounding box indicating the text region based on the direction of the text on the penetrometer selected as the inspection target as a wire region, and extracts the determined wire region to generate a wire region image. As shown in FIG. 7, based on the text "ASTM 11F" on the penetrometer selected as the inspection target, the up direction corresponds to the down direction of the radiation image in FIG. 7, and the down direction corresponds to the up direction of the radiation image in FIG. 7. The detection unit 300 may determine a predetermined region from the bounding box BB indicating the text region, for example, a region having the same width (W) as the width of the bounding box BB indicating the text and a height (3H) three times the height (H) of the bounding box BB indicating the text region, as the wire region ROI.

[0053] Once the wire region is identified and the wire region image is generated, the calculation unit 400 counts the number of wires included in the wire region image in step S340, which will be described in more detail below.

[0054] Next, in step S350, the evaluation unit 500 determines the quality of the radiation image according to the number of wires previously counted (S340). At this time, if the number of wires previously counted (S340) is equal to or greater than a predetermined standard reference value, the evaluation unit 500 determines that the quality of the radiation image is suitable for defect inspection. On the other hand, if the number of wires previously counted (S340) is less than the predetermined standard reference value, the evaluation unit 500 may determine that the quality of the radiation image is unsuitable for defect inspection. That is, the better the quality of the radiation image, the more counted wires there are, and the worse the quality of the radiation image, the fewer counted wires there are. Thus, the quality of the radiation image can be determined by comparing the number of counted wires with a predetermined standard and determining whether the number of counted wires is equal to or greater than the predetermined standard reference value.

[0055] If the evaluation unit 500 determines that the quality of the radiation image is suitable for defect inspection in step S360, it analyzes the radiation image and outputs an inspection command to start defect inspection. On the other hand, if the evaluation unit 500 determines that the quality of the radiation image is not suitable for defect inspection, it outputs a re-photographing command to perform re-photography.

[0056] Next, a method for counting the number of wires included in the wire area image in step S340 will be described. Figure 8 is a flowchart illustrating a method for counting the number of wires included in the wire area image according to an embodiment of the present invention. Figures 9 to 17 are examples of screens illustrating a method for counting the number of wires included in the wire area image according to an embodiment of the present invention.

[0057] As described above, the wire region can be identified and a wire region image can be generated. The wire region image is an image of the region where the wires of one or more image quality indicators (IQIs) are present in a radiation image of an object attached with one or more IQIs.

[0058] 8 to 17, a wire area image is generated, and in step S410, the wire area image is input to the calculation unit 400. Then, the pre-processing unit 410 of the calculation unit 400 performs pre-processing on the input wire area image.

[0059] To this end, the pre-processing unit 410 normalizes the wire area image so that the wire area image is converted into a predetermined format in step S420. Hereinafter, step S420 will be described in more detail with reference to FIG.

[0060] As shown in FIG. 9, the pre-processing unit 410 first rotates the wire region image W01 in a predetermined direction, thereby deriving a rotated wire region image W02. That is, the pre-processing unit 410 rotates the wire region image W01 so that the wire in the wire region image W01 is oriented horizontally. Then, the pre-processing unit 410 resizes the rotated wire region image W02 to a predetermined size, thereby deriving a resized wire region image W03.

[0061] After normalization by rotation and resizing, the pre-processing unit 410 performs image processing on the wire area image in step S430 so that the difference between the pixels that make up the lines in the wire area image and the remaining pixels becomes more noticeable.

[0062] Step S440 will be described in more detail below with reference to FIGS.

[0063] First, the pre-processing unit 410 performs a contrast limited adaptive histogram equalization (CLAHE) process on the wire region image W03 to increase the clarity of the lines present in the resized wire region image W03, thereby removing noise from the radiation image and deriving a wire region image W04 with increased brightness and contrast.

[0064] 11 and 12, the pre-processing unit 410 then performs filtering using one or more filters in the frequency domain of the wire-region image W04.

[0065] At this time, the pre-processing unit 410 converts the wire domain image W04 into the frequency domain to generate a frequency domain image W05.

[0066] Then, as shown in FIG. 12, the pre-processing unit 410 generates a composite filter CF by combining a low-pass filter F01 that passes only the low-band components of the frequency domain image W05, a vertical weighting filter F02 that multiplies the vertical components of the frequency domain image W05 by a predetermined weight, a sandclock-shaped filter F03 that eliminates the upper and lower regions of a region divided by two different line segments that pass through a reference point (S: for example, the midpoint of the frequency domain image W05) in a vertically symmetrical manner with respect to a reference point in the frequency domain image W05, and a Butterworth filter F04 that minimizes phase distortion in boundary regions of frequencies to be filtered in the frequency domain image W05. The pre-processing unit 410 then performs frequency domain filtering on the frequency domain image W05 using the composite filter CF to derive a frequency domain image W06 that has been filtered in the frequency domain.

[0067] At this time, the Butterworth filter is configured by the following Equation 1.

[0068]

number

[0069] As mentioned above, once the frequency domain filtering is completed, the pre-processing unit 410 inversely transforms the frequency domain video W06 from the frequency domain to the time domain to derive the frequency domain filtered wire domain video W07.

[0070] As mentioned before, once the frequency domain filtering is complete,

[0071] 13, the pre-processing unit 410 performs filtering using a Sobel filter in both directions including the upward direction U and downward direction D based on the Y axis in the wire region image W07 so as to detect the boundaries of the vertical components of the lines present in the wire region image W07. This derives a wire region image W08 in which the boundaries of the vertical components of the lines have been detected.

[0072] Next, the binarization unit 420 of the calculation unit 400 binarizes the wire region image W08 to derive a binarized wire region image in step S440. Hereinafter, the binarization in step S440 will be described in more detail with reference to FIG. 14.

[0073] First, the binarization unit 420 binarizes the wire region image W08 using a predetermined ratio of upper pixel values ​​in the distribution of pixel values ​​of the wire region image W08 as a threshold to derive a binarized wire region image W09, as shown in Fig. 14. That is, the binarization unit 420 derives the wire region image W09 by binarizing pixels having pixel values ​​equal to or greater than the threshold and the remaining pixels.

[0074] 14, the binarization unit 420 performs a morphology operation on the binarized wire region image W09, thereby deriving a wire region image W10 to which the morphology operation has been applied. For the morphology operation, the binarization unit 420 may perform at least one of an erosion operation to remove noise from the wire region image W09 and a dilation operation to dilate the lines of the wire region image W09, or may perform the erosion operation and the dilation operation sequentially.

[0075] Next, the line detection unit 430 detects lines from the wire area image in step S450. Step S450 will be described in more detail below with reference to FIGS.

[0076] First, as shown in FIG. 15, the line detection unit 430 performs a Hough transform to detect multiple lines. The line detection unit 430 generates multiple curves in a polar coordinate system image PC, which is obtained by converting a wire area image W10 in a two-dimensional Cartesian coordinate system into a polar coordinate system. Then, the line detection unit 430 forms an accumulation array of points where the multiple curves generated in the polar coordinate system image PC intersect. Next, the polar coordinate system image PC detects multiple lines by again mapping multiple points whose accumulation array values ​​are equal to or greater than a threshold onto the wire area image W11 in the rectangular coordinate system.

[0077] 16, the line detection unit 430 performs clustering CLU based on the density of the detected lines in the wire area image W11, clustering the detected lines into one or more line clusters LC. Preferably, the clustering is performed using density-based spatial clustering of application with noise (DBSCAN). This clustering results in a wire area image W12 containing one or more line clusters LC.

[0078] Also, as shown in FIG. 17, the line detection unit 430 can additionally and selectively remove line clusters LC located within a predetermined percentage (e.g., 2%) of the upper and lower ends of the wire area image.

[0079] As described above, when a line cluster is detected, the counting unit 440 counts the number of wires of the penetrameter according to the number of line clusters in step S460.

[0080] 18 is a diagram illustrating a computing device according to an embodiment of the present invention. The computing device TN100 of FIG. 18 may be a device described herein (e.g., evaluation device 10, etc.).

[0081] 18, computing device TN100 may include at least one processor TN110, a transceiver device TN120, and a memory TN130. Computing device TN100 may further include a storage device TN140, an input interface device TN150, an output interface device TN160, etc. The components included in computing device TN100 may be connected by a bus TN170 to communicate with each other.

[0082] Processor TN110 can execute program commands stored in at least one of memory TN130 and storage device TN140. Processor TN110 can refer to a central processing unit (CPU), a graphics processing unit (GPU), or a dedicated processor on which methods according to embodiments of the present invention are performed. Processor TN110 can be configured to implement procedures, functions, methods, etc. described in connection with embodiments of the present invention. Processor TN110 can control each component of computing device TN100.

[0083] Memory TN130 and storage device TN140 can each store various information related to the operation of processor TN110. Memory TN130 and storage device TN140 can each be configured with at least one of a volatile storage medium and a non-volatile storage medium. For example, memory TN130 may be configured with at least one of a read-only memory (ROM) and a random access memory (RAM).

[0084] The transceiver TN120 can transmit or receive wired or wireless signals and can communicate with a network.

[0085] In particular, the learning unit 100, data processing unit 200, detection unit 300, calculation unit 400, and evaluation unit 500 according to an embodiment of the present invention may be realized in the form of a program readable via a computer means, stored in memory TN130, and executed by processor TN110, or may become a lower module of processor TN110.

[0086] Meanwhile, the various methods according to the above-described embodiments of the present invention can be realized in the form of a program readable by various computer means and recorded on a computer-readable recording medium. Here, the recording medium may include program instructions, data files, data structures, and the like, alone or in combination. The program instructions recorded on the recording medium may be specially designed and constructed for the present invention, or may be known and available to those skilled in the art of computer software. For example, recording media include magnetic media such as hard disks, floppy disks, and magnetic tapes, optical media such as CD-ROMs and DVDs, magneto-optical media such as floptical disks, and hardware devices specially configured to store and execute program instructions, such as ROM, RAM, and flash memory. Examples of program instructions include not only machine language, such as that produced by a compiler, but also high-level language codes executable by a computer using an interpreter, etc. Such hardware devices can be configured to operate as one or more software modules to perform the operations of the present invention, or vice versa.

[0087] Although one embodiment of the present invention has been described above, a person having ordinary skill in the art may modify and change the present invention in various ways by adding, changing, deleting or adding components within the scope of the concept of the present invention as set forth in the claims, and this also falls within the scope of the present invention.

Claims

1. When a wire area image is input, which is an image of an area where a wire of one or more image quality indicators (IQIs) is present, among radiation images taken of an object having one or more image quality indicators (IQIs) attached thereto by radiation, a preprocessing unit preprocessing the wire region image; a binarization unit binarizing the wire region image; a line detection unit detecting lines from the wire area image; a counting unit counting the number of wires of the penetrameter along the detected line; an evaluation unit evaluating quality of the radiation image according to the number of counted wires; Including, Methods for assessing quality.

2. The step of detecting the line of wire comprises: the line detection unit converts the pixels of the wire area image into a polar coordinate system to generate a plurality of curves, composes points where the generated curves intersect in an accumulation array, and detects a plurality of lines from a plurality of points where the accumulation array value is equal to or greater than a threshold value; the line detection unit clustering the detected lines into one or more line clusters based on a density of the detected lines; Including, 2. A method for assessing quality according to claim 1.

3. The step of counting the number of wires of the penetrometer comprises: the counting unit counts the number of wires of the penetrometer according to the number of the line clusters; A method for assessing quality according to claim 2.

4. The step of pre-processing the wire region image includes: the pre-processing unit normalizing the wire area image so that the wire area image is converted into a predetermined format; the pre-processing unit performs image processing on the wire area image so that a difference between pixels forming a line and the remaining pixels in the wire area image becomes noticeable; Including, 2. A method for assessing quality according to claim 1.

5. The step of normalizing the wire region image includes: the pre-processing unit rotates the wire region image in a preset direction; the pre-processing unit adjusts the wire area image to a preset size; Including, A method for assessing quality according to claim 4.

6. The step of performing image processing includes: the pre-processing unit performs a contrast limited adaptive histogram equalization (CLAHE) process on the wire area image to remove noise from the radiation image and increase brightness and contrast, so as to increase the clarity of lines present in the wire area image; the pre-processing unit filtering the wire-domain image in the frequency domain using one or more filters; the pre-processing unit performs filtering using a Sobel filter in the wire area image in an upward and downward direction based on a Y axis so as to detect boundaries of vertical components of lines present in the wire area image; Including, A method for assessing quality according to claim 4.

7. The step of filtering using one or more filters includes: the pre-processing unit transforming the wire-domain image into a frequency domain to generate a frequency-domain image; The pre-treatment unit a low-pass filter that passes only low-band components of the frequency domain image; a vertical weighting filter for multiplying a vertical component of the frequency domain image by a predetermined weight; an hourglass-shaped filter for eliminating upper and lower regions of a region divided by two different line segments passing through a reference point in a vertically symmetrical manner in the frequency domain image; performing frequency domain filtering using a composite filter combining a Butterworth filter and a frequency domain filter that minimizes phase distortion in a boundary region of a frequency to be filtered in the frequency domain image; the pre-processing unit inversely transforming the frequency domain image from the frequency domain to the time domain; Including, A method for assessing quality according to claim 6.

8. The Butterworth filter is Formula [Equation 4] It is composed of where w is the frequency, where n is the filtering order. A method for assessing quality according to claim 7.

9. The step of binarizing the wire region image by the binarization unit includes: the binarization unit binarizing the wire region image using a predetermined ratio of upper pixel values ​​in a distribution of pixel values ​​of the wire region image as a threshold; Including, A method for assessing quality according to any one of claims 1 to 8.

10. The step of binarizing the wire region image by the binarization unit includes: After the erasure step, the binarization unit performing a morphology operation on the wire region image; further comprising The morphology operation is an erosion operation to remove noise from the wire region image; and a dilation operation causing lines of the wire region image to be dilated; A method for assessing quality according to claim 9.

11. When a wire area image is input, which is an image of an area where a wire of one or more image quality indicators (IQIs) is present, among radiation images taken of an object having one or more image quality indicators (IQIs) attached thereto by radiation, a preprocessing unit that preprocesses the wire region image; a binarization unit for binarizing the wire region image; a line detection unit for detecting lines from the wire area image; a counting unit that counts the number of wires of the penetrameter along the detected line; an evaluation unit for evaluating the quality of the radiation image according to the number of counted wires; Including, A device for assessing quality.

12. The line detection unit converting the pixels of the wire area image into a polar coordinate system to generate a plurality of curves, configuring points where the generated plurality of curves intersect as an accumulation array, and detecting a plurality of lines from a plurality of points where the accumulation array value is equal to or greater than a threshold; performing clustering based on a density of the detected lines to cluster the detected lines into one or more line clusters; An apparatus for assessing quality according to claim 11.

13. The counting unit Counting the number of wires of the penetrometer according to the number of the line clusters; An apparatus for assessing quality according to claim 12.

14. The pre-treatment unit normalizing the wire area image so that the wire area image is converted into a predetermined format; performing image processing on the wire region image so that differences between pixels forming a line and remaining pixels in the wire region image become more noticeable; An apparatus for assessing quality according to claim 11.

15. The pre-treatment unit Rotating the wire area image in a preset direction; adjusting the wire area image to a preset size; 15. An apparatus for assessing quality according to claim 14.

16. The pre-treatment unit performing contrast limited adaptive histogram equalization (CLAHE) processing on the wire area image to increase the clarity of lines present in the wire area image, thereby removing noise from the radiation image and increasing brightness and contrast; filtering the wire-region image in a frequency domain using one or more filters; filtering the wire area image using a Sobel filter in an upward and downward direction based on a Y axis so that boundaries of vertical components of lines present in the wire area image are detected; 15. An apparatus for assessing quality according to claim 14.

17. The pre-treatment unit transforming the wire-domain image into a frequency domain to generate a frequency-domain image; a low-pass filter that passes only low-band components of the frequency domain image; a vertical weighting filter for multiplying a vertical component of the frequency domain image by a predetermined weight; an hourglass-shaped filter for eliminating upper and lower regions of a region divided by two different line segments passing through a reference point in a vertically symmetrical manner in the frequency domain image; and performing frequency domain filtering using a composite filter combining a Butterworth filter and a frequency domain filter, which minimizes phase distortion in a boundary region of a frequency to be filtered in the frequency domain image; transforming the frequency domain image back from the frequency domain to the time domain; 17. An apparatus for assessing quality according to claim 16.

18. The Butterworth filter is Formula [Equation 5] It is composed of where w is the frequency, where n is the filtering order.

18. An apparatus for assessing quality according to claim 17.

19. The binarization unit is binarizing the wire region image using a predetermined ratio of upper pixel values ​​in the distribution of pixel values ​​of the wire region image as a threshold; A device for assessing quality according to any one of claims 11 to 18.

20. The binarization unit is performing a morphology operation on the wire region image; The morphology operation is an erosion operation to additionally remove noise from the wire region image; and a dilation operation causing lines of the wire region image to be dilated; 20. An apparatus for assessing quality according to claim 19.

Images