JP2007298386A - Ultrasonicflaw detection data processing device, method, and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an ultrasonic flaw detection data processing device, a method, and a program for discriminating a shape echo at high speed. <P>SOLUTION: This ultrasonic flaw detection data processing device includes a shape data storage part for an ultrasonic flaw detection object, a flaw candidate domain extraction part for extracting a domain wherein an ultrasonic flaw detection echo satisfies a prescribed standard, a shape candidate domain extraction part for extracting a domain overlapped with a shape gate domain from among flaw candidate domains, a generation part of a shape candidate domain to be divided for performing dividing processing of the shape candidate domain by using a part having the minimal echo height as a boundary, a characteristic position setting part for setting a characteristic position in the shape candidate domain to be divided, and a shape echo determination part for determining that the shape candidate domain to be divided is a shape echo when the characteristic position is included in a set domain of shape data. The dividing processing includes A-scope separation processing for dividing the shape candidate domain in the path length direction of a beam for ultrasonic flaw detection, and a focal row direction separation processing for re-dividing the shape candidate domain in the vertical direction to the path length direction in B-scope after the A-scope separation processing. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a technique for processing data obtained by ultrasonic flaw detection.

  Ultrasonic flaw inspection is used to inspect the internal structure of objects such as welds of metallic materials. In the ultrasonic flaw detection inspection, if a flaw exists inside the object, the position and shape of the flaw are obtained as an ultrasonic echo image.

  In the ultrasonic flaw detection inspection, not only an echo image derived from a wound that is an original inspection object but also an echo image derived from the shape of the object is detected. If this shape echo can be automatically removed from the inspection object, the inspection efficiency, accuracy, and homogeneity (no matter who inspects can obtain the same result) are improved.

  In high-accuracy ultrasonic flaw detection, a huge amount of inspection data is generated. There is a demand for a technique that makes it possible to remove shape echoes from such inspection data in a short time.

Patent Document 1 describes a quality inspection method for a welded portion of a welded steel pipe. In this document, it is an object to provide a quality inspection method for welded steel pipe welds that can reliably capture only harmful flaws, including thick materials, during ultrasonic flaw detection of welded steel pipe welds. ing.
JP 2006-30218 A

An object of the present invention is to provide an ultrasonic flaw detection data processing apparatus, method and program for identifying shape echoes at high speed in ultrasonic flaw detection.
Another object of the present invention is to provide an ultrasonic flaw detection data processing apparatus, method, and program capable of identifying shape echoes with high accuracy in ultrasonic flaw detection.
Still another object of the present invention is to provide an ultrasonic flaw detection data processing apparatus, method, and program for suppressing variation in identification results depending on the skill of an inspector in ultrasonic flaw detection.

  In the following, means for solving the problem will be described using the numbers used in [Best Mode for Carrying Out the Invention] in parentheses. These numbers are added to clarify the correspondence between the description of [Claims] and [Best Mode for Carrying Out the Invention]. However, these numbers should not be used to interpret the technical scope of the invention described in [Claims].

  An ultrasonic flaw detection data processing apparatus (52) according to the present invention includes a shape data storage unit (54) for storing shape (18) data indicating the shape of an object (2) to be inspected by ultrasonic flaw detection, and an ultrasonic flaw detection. In the obtained target data, a scratch candidate region extraction unit (56) that extracts a region where an echo satisfies a predetermined criterion as a scratch candidate region (33), and the scratch candidate region are set based on shape data. A shape candidate region (34) extraction unit (58) that extracts a region that overlaps the shape gate region (30) as a shape candidate region (34), and a shape candidate region (34) are divided using a portion having a minimum echo height as a boundary. A segmented shape candidate region generation unit (59) that generates a segmented shape candidate region (34) by the segmenting process, and a representative point that sets a representative point that represents the position of the segmented shape candidate region (34) And a shape echo determination unit (70) for determining that the divided shape candidate region (34) is a shape echo when the representative point is included in the setting region (30) set based on the shape data. With.

  In the ultrasonic flaw detection data processing apparatus (52) according to the present invention, the division process includes an A scope separation process (S10) for dividing the shape candidate region (34) in the direction of the path of the beam used for ultrasonic flaw detection.

  In the ultrasonic flaw detection data processing apparatus (52) according to the present invention, the division processing further includes focal row direction separation that further divides the shape candidate region (34) in the direction perpendicular to the path length direction in the B scope after the A scope separation processing. A process (S14) is included. Here, the direction perpendicular to the path direction which is the traveling direction of the ultrasonic flaw detection beam in the B scope is called the focal low direction.

  In the ultrasonic flaw detection data processing apparatus (52) according to the present invention, the focal low direction separation processing corresponds to the echo height in the path direction of the shape candidate region (34) corresponding to each position in the focal low direction perpendicular to the path length direction. A maximum value extraction process for extracting the maximum value of the shape, and a valley portion that generates the divided shape candidate region (34) by dividing the shape candidate region (34) with the valley position in the vertical distribution of the maximum value as a boundary. Extraction processing.

  In the ultrasonic flaw detection data processing apparatus (52) according to the present invention, the focal low direction separation processing is performed by ultrasonic flaw detection of effective value data excluding data equal to or less than a predetermined value from the echo height data inside the shape candidate region. A route direction average value calculation process for calculating a route direction average value distribution, which is an average value of the data used in the route direction, and a valley of the effective value data in a vertical direction perpendicular to the route direction. And a valley extraction process for generating the divided shape candidate area by dividing the position of the position as a boundary.

  In the ultrasonic flaw detection data processing method according to the present invention, in the step (S1) of obtaining shape data (18) relating to the shape of the object to be inspected by the ultrasonic flaw detection, an echo is detected in the target data obtained by the ultrasonic flaw detection. A scratch candidate region extraction step (S4) for extracting a region that satisfies a predetermined criterion as a scratch candidate region (33), and a region that overlaps the shape gate region (30) set based on the shape data among the scratch candidate regions. Extracting the shape candidate region (34) as a shape candidate region (34), and generating the divided shape candidate region (34) by dividing the shape candidate region (34) with a portion having a minimum echo height as a boundary; A representative point setting step (S18) for setting representative points representing the positions of the divided shape candidate regions (34-1 to 34-4); And a determining that an object to be split shape candidate regions (34-1 to 34-4) is shaped echo (S20) when included in the set set area (30) based on.

  In the ultrasonic flaw detection data processing method according to the present invention, the representative point setting step (S18) extracts a plurality of echo height peaks from the divided shape candidate regions (34-1 to 34-4) (S32). And a step (S38, S40) of determining a representative point by a predetermined method from an area surrounded by a plurality of peaks.

  In the ultrasonic flaw detection data processing apparatus (52) according to the present invention, the division process includes an A scope separation process (S10) for dividing the shape candidate region (34) in the direction of the path of the beam used for ultrasonic flaw detection.

  In the ultrasonic flaw detection data processing method according to the present invention, the division process further includes a focal row direction separation process (S14) for further dividing the shape candidate region (34) in the direction perpendicular to the path direction in the B scope after the A scope separation process. )including.

  In the ultrasonic flaw detection data processing method according to the present invention, the focal low direction separation process (S14) corresponds to the echo height in the path direction of the shape candidate region (34) corresponding to each position in the focal low direction perpendicular to the path direction. And a shape candidate region (34-1 to 34-4) by dividing the shape candidate region (34) with the valley position in the vertical distribution of the maximum value as a boundary. ) To generate a valley extraction process.

  In the ultrasonic flaw detection data processing method according to the present invention, the focal low direction separation process is used for ultrasonic flaw detection of effective value data excluding data equal to or less than a predetermined value from echo height data inside the shape candidate region. The route direction average value calculation process for calculating the distribution of the route direction average value, which is the average value for the route direction of the data, and the valley position of the distribution of the effective value data in the vertical direction perpendicular to the route direction. And valley extraction processing for generating a divided shape candidate region by dividing.

  The ultrasonic flaw detection data processing program according to the present invention causes a computer to execute the ultrasonic flaw detection data processing method according to the present invention.

According to the present invention, an ultrasonic flaw detection data processing apparatus, method, and program for identifying shape echoes at high speed in ultrasonic flaw detection are provided.
Furthermore, according to the present invention, there is provided an ultrasonic flaw detection data processing apparatus, method, and program capable of identifying shape echoes with high accuracy in ultrasonic flaw detection.
Furthermore, according to the present invention, there is provided an ultrasonic flaw detection data processing apparatus, method and program for suppressing variation in identification results depending on the skill of an inspector in ultrasonic flaw detection.

  The best mode for carrying out the present invention will be described below with reference to the drawings. FIG. 1 is a diagram for explaining an ultrasonic flaw detection test. In the ultrasonic flaw detection inspection, the probe 6 is directed to the inspection object 2. As shown in the left diagram of FIG. 1, the probe 6 includes a large number of excitation elements 8 that generate ultrasonic waves. The probe 6 performs phased array ultrasonic testing by controlling the excitation element 8.

  The probe 6 excites the excitation element 8 with a simultaneous excitation element 10 including a plurality of adjacent excitation elements 8 as a unit. With respect to the nth simultaneous excitation element 10, the (n + 1) th simultaneous excitation element 10 is set as a group of excitation elements 8 that are shifted by i number of excitation elements 8 in the first direction in the arrangement direction of the excitation elements 8. Is done. The simultaneous excitation element 10 performs phased array ultrasonic flaw detection using a synthetic wavefront generated at an electronically controlled angle.

  As shown in the right diagram of FIG. 1, the ultrasonic flaw detector includes a first probe 6-1, a second probe 6-2, and a third probe 6-3 which are probes 6 having such a configuration. The first probe 6-1 performs vertical flaw detection from above the weld line 4. The weld line 4 extends in a direction perpendicular to the paper surface. The second probe 6-2 is arranged at a position shifted from the welding line 4 in the first direction with respect to the first probe 6-1 and is arranged obliquely toward the welding line 4 to perform oblique flaw detection. . The third probe 6-3 is disposed at a position shifted from the weld line 4 in the direction opposite to the first direction, and is disposed obliquely toward the weld line 4 to perform oblique flaw detection.

  FIG. 2 shows an A scope waveform of an echo generated by the ultrasonic flaw detector performing ultrasonic flaw detection on the inspection object 2. A plurality of A scope waveforms 14-1 to 14 -N are generated corresponding to the plurality of simultaneous excitation elements 10, respectively. The horizontal axis indicates the beam path, that is, the distance of the ultrasonic beam in the direction perpendicular to the combined wavefront 12. The vertical axis indicates the height of the echo standardized in common for the plurality of A scope waveforms 14-1 to 14-N.

  FIG. 3 is a diagram in which a plurality of A scope waveforms 14-1 to 14-N are arranged and the echo height in the direction perpendicular to the beam path is visualized. The beam path is tilted by a refraction angle (that is, an incident angle) with respect to the surface of the inspection object 2 and arranged so that the path becomes longer from the upper left to the lower right. In such an arrangement, the distance in the vertical direction corresponds to the depth from the side surface to which the probe 6 of the inspection object 2 is applied.

  The horizontal axis of FIG. 3 indicates that the A scope waveforms 14-1 to 14-N indicate the pitch of the flaw detection step, that is, the position between one simultaneous excitation element 10 and the adjacent simultaneous excitation element 10 that has the largest overlap. It is shown that the difference is arranged as a pitch. Assuming that the beam path is the x-axis and the pitch of the flaw detection step is the y-axis, the z-axis direction indicates the echo height. The upper peak set in FIG. 3 shows an example of the shape echo 15, that is, the echo generated by the shape of the material of the inspection object 2, not the scratch. A set of peaks in the lower part of FIG. 3 shows an example of the defect indication echo 17 indicating a flaw.

  FIG. 4 shows the B scope waveform 16. The B scope waveform 16 corresponds to an image using luminance instead of the echo height in the arrangement of the A scope waveforms 14-1 to 14-N shown in FIG. The B scope waveform 16 visually checks the result of ultrasonic flaw detection in a cross section perpendicular to the weld line 4 in FIG.

  The A scope waveform 14 and the B scope waveform 16 are generated as inspection data of the inspection object 2 by the ultrasonic flaw detector. The ultrasonic flaw detection data processing apparatus, method and program in the present embodiment are used to automatically remove a part derived from the shape of the inspection object 2 from the echo waveform in such inspection data.

  FIG. 15 shows the configuration of an ultrasonic flaw detection data processing apparatus. The ultrasonic flaw detection data processing device 52 includes a shape data storage unit 54, a flaw candidate region extraction unit 56, a shape candidate region extraction unit 58, a divided shape candidate region generation unit 59, a region representative point setting unit 62, and a shape echo determination unit 70. including. The divided shape candidate region generation unit 59 includes an A scope separation unit 60 and a focal row direction separation unit 61. The region representative point setting unit 62 includes a geometric center extraction unit 63, a peak echo extraction unit 64, a peak region generation unit 66, and a peak echo center generation unit 68. The ultrasonic flaw detection data processing device 52 is preferably realized by a computer such as a personal computer. In that case, the shape data storage unit 54 is realized by a reading device included in the computer reading shape data from the storage device and writing it into a storage medium exemplified by a hard disk. The scratch candidate area 56 to the shape echo determination unit 70 are realized as software written in a storage device included in the computer, and each function is realized by being read and executed by an arithmetic control device included in the computer.

FIG. 5 is a flowchart showing an ultrasonic flaw detection data processing method executed by the ultrasonic flaw detection data processing apparatus.
Step S1: The ultrasonic flaw detection data processing device 52 reads the inspection data of the inspection object 2. The ultrasonic flaw detection data processing device 52 further reads shape data indicating the shape of the inspection object 2 (a region in which an echo image not derived from a flaw is generated, which is exemplified by a surface or welded portion where different materials come into contact). Image data superimposed in correspondence with the data is created and stored in the shape data storage unit 54.

  Step S2: The ultrasonic flaw detection data processing apparatus 52 sets an evaluation gate and a shape recognition gate for the work image. This setting is performed automatically by a predetermined algorithm or in response to an input operation from the input device. 6A to 6D are diagrams for explaining this setting. FIG. 6A shows the positional relationship between the probe and the inspection object 2. For simplicity, only the first probe 6-1 is depicted.

  FIG. 6B is a diagram for describing setting of the evaluation gate 20. The evaluation gate 20 indicates a region to be inspected that is a part of the inspection target 2. The first probe 6-1 generates a beam whose path direction is a direction perpendicular to the surface of the inspection object 2. The evaluation gate 20 is set to have a width with the shape 18 as one end and the other end at a position where the shape 18 has entered the inspection object 2 by an offset amount 22 set in the beam path direction. The offset amount 22 is set depending on the position of the simultaneous excitation element 10.

  FIG. 6C is a diagram for explaining the shape recognition gate. The shape recognition gate indicates a region to be subjected to a shape recognition process, which is a process of recognizing an echo image indicating the shape 18 in the inspection object 2 by distinguishing it from an echo image indicating a flaw. The shape recognition gate is set as follows. The shape recognition gate upper end 24 is set at a position shifted in the direction close to the surface of the inspection object 2, that is, the first probe 6-1 by the amount α set in the beam path direction with respect to the shape 18. The shape recognition gate lower end 26 is set at a position where the shape 18 has entered the inspection object 2 by an amount β set in the beam path direction with respect to the shape 18. A region between the shape recognition gate upper end 24 and the shape recognition gate lower end 26 is a shape recognition gate.

  FIG. 6D shows a data structure of the evaluation gate 20 and the shape recognition gate 30. For a certain simultaneous excitation element 10 (numbered F1 in FIG. 6D), an evaluation gate 28 and a shape recognition gate 30 are set as values corresponding to the beam path length W. The shape recognition gate 30 is a partial region of the evaluation gate 28. The evaluation gate 20 and the shape recognition gate 30 are set for each of the plurality of simultaneous excitation elements 10 (indicated by F1, F2, F3, etc. in FIG. 6D).

  Step S4: The scratch candidate area extraction unit 56 executes the intra-section area recognition process. The intra-section region recognition processing is processing for extracting a region where an echo image of a predetermined intensity exists in the B scope waveform 16. FIG. 7A shows image data that has undergone cross-sectional area recognition processing. The wound candidate region extraction unit 56 extracts a portion satisfying a predetermined criterion from the echo group inside the evaluation gate 20 as a wound candidate region 33 that is a candidate for a region indicating a wound echo. The predetermined reference is, for example, a reference that the echo height is higher than a predetermined reference value.

In the present embodiment, the scratch candidate region 33 is defined to be a rectangle having two sides perpendicular to the path direction of the ultrasonic beam and two parallel sides in the B scope image. Instead, the contour of the echo image 32 can be extracted, and the region inside the contour can be defined as the scratch candidate region 33.

  As illustrated in FIG. 7A, the wound candidate region 33 may include a plurality of peaks that can be distinguished if a human views the B-scope image. Of these multiple peaks, one may show a flaw echo and the other may show a shape echo.

  Step S6: The ultrasonic flaw detection data processing apparatus 52 determines whether the area range of the flaw candidate area 33 is inside or outside the shape recognition gate 30. The scratch candidate region 33 is determined to be inside the shape recognition gate 30 when a part thereof overlaps the shape recognition gate 30. If the scratch candidate region 33 is outside the shape recognition gate 30, it is determined that the scratch candidate region 33 indicates a scratch (step S8). The shape candidate region extraction unit 58 extracts the scratch candidate region 33 inside the shape recognition gate 30 as the shape candidate region 34 that is a region that may be an echo due to the shape, and executes the processes after step S10. (Step S6 YES).

  Step S10: The A scope separation unit 60 performs A scope separation, which is a process of separating the shape candidate region in the beam path direction. FIG. 7B is a diagram for explaining A scope separation. In the example of FIG. 7B, the A scope waveform has two peaks in the vicinity of the center in the focal law direction perpendicular to the path. The A scope separation unit 60 divides shape candidate regions that satisfy these conditions in the path length direction.

  8A and 8B are diagrams for explaining the A scope separation in more detail. FIG. 8A shows an A scope waveform 38i for a certain focal law i. In the A scope waveform 38i, two peaks, a shape candidate peak 40 and a flaw peak 42, are recognized. These two peaks are connected at the base.

  The A scope separation unit 60 extracts a shape candidate peak 41 having a maximum echo height inside the shape recognition gate 30. Then, a predetermined ratio of the height of the shape candidate peak 41 (for example, the height of the shape candidate peak 41—A decibel, A is a predetermined value) is set as the cut threshold. Then, the part of the beam path whose echo height is smaller than the cut threshold is discarded.

  FIG. 8B shows the A scope waveform 38'i after the truncation process has been performed. The shape candidate peak 41 and the flaw peak 42 are separated. However, at this stage, it is not determined whether the shape candidate peak 41 and the flaw peak 42 are flaw echoes or shape echoes.

  Step S12: After the shape A scope separation is performed, the scratch candidate region extraction unit 56 executes the in-section region recognition process again. As a result, region recognition is achieved in which peaks adjacent to each other in the path direction included in the same shape candidate region 34 in FIG. 7A are separated. This is illustrated in FIG. 7C. The shape candidate region 34 is recognized as being separated into a first shape candidate region 34-1 and a second shape candidate region 34-2.

  Step S14: The focal row direction separation unit 61 performs focal row direction separation, which is a process of separating the flaw candidate area subjected to shape A scope separation in the focal row direction. FIG. 7D is a diagram for explaining focal-row direction separation. In the example of FIG. 7D, the B scope waveform of the shape candidate region 34-2 has a shape having two peaks in the focal low direction. The focal law direction separation unit 61 divides shape candidate regions that satisfy these conditions in the focal law direction.

  The operation of the focal row direction separation unit 61 will be described in more detail with reference to FIGS. 9A, 9B, and 9C. FIG. 9A shows a shape candidate region 34-2 to be processed. When the cross section in the focal low direction is viewed in Line 1 where the path of the shape candidate region 34-2 has a certain value, one echo peak shown in the echo height distribution 40-L1 in the upper diagram of FIG. 9B is observed. When the cross section in the focal low direction is viewed in Line 2 where the path of the shape candidate region 34-2 has another constant value, one echo peak shown in the echo height distribution 40-L2 in the lower diagram of FIG. It is observed to be located on the right side.

The focal row direction separation unit 61 performs the following operation.
(A) Maximum value extraction process: The maximum value of the echo height in the path direction is extracted corresponding to each position in the focal law direction of the shape candidate region 34-2. That is, a maximum value curve that is a line connecting the maximum values when the echo height of the shape candidate region 34-2 is projected in the path direction is extracted. Thus, an image including two echo peaks is generated as shown in the echo height distribution 40-L3 in FIG. 9C.
(B) A shape candidate region to be divided is generated by dividing the shape candidate region 34-2 with the minimum portion (the valley portion in FIG. 9C) of the maximum value curve as a boundary. FIG. 7E shows the generated divided shape candidate regions 34-3 and 34-4.

  With reference to FIG. 10, the minimum part recognition process performed by the focal row direction separation unit 61 in the process (b) will be described. The focal row direction separation unit 61 recognizes the valley portion of the maximum value curve (the portion near the position where the maximum value curve in FIG. 9C becomes a minimum) using any one of the following three pattern recognition methods. Process.

Condition (1):
i-1> i <i + 1
Is satisfied, the shape scope region 34 is separated in the focal low direction by setting the A scope waveform 14 at the position indicated by i in the focal low direction to a zero value. I-1, i, i + 1 in the above formulas indicate the echo heights of the maximum value curves of positions i-1, i, i + 1 in the focal low direction, respectively. This condition indicates that the maximum value curve is a valley (minimum) at the position i in the focal law direction, as in the condition (1) indicated by the solid line in FIG. Thereby, the shape candidate area | region 34 is isolate | separated by the shallow trough of the maximum value curve. This condition is suitable for high-speed processing because the amount of calculation required for determination is small.

Condition (2):
i-1> i <i + 1 and i-2> i <i + 2
Is satisfied, the shape candidate region 34 is separated in the focal law direction by setting the A scope waveform 14 at the position indicated by i to zero as the focal law. As shown in the condition (2) in FIG. 10, this condition is such that the maximum value curve is located at positions i−1, i + 1 in the focal law direction and focal points adjacent to the outside at the position i in the focal law direction. It is smaller than the positions i-2 and i + 2 in the row direction. As a result, the shape candidate region 34 is separated by a slightly shallow valley of the maximum value curve. This condition is more suitable for reducing the influence of noise than in the case of condition (1).

Condition (3):
i-1> i <i + 1 and i-2> i-1 and i + 1 <i + 2
Is satisfied, the shape candidate region 34 is separated in the focal law direction by setting the A scope waveform 14 at the position indicated by i to zero as the focal law. In this condition (3), as shown in the condition (3) in FIG. 10, the maximum value curve is the smallest value in the range of the two positions in the surrounding focal low direction at the position i in the focal low direction. Indicates that Thereby, the shape candidate region 34 is separated in the deep valley of the maximum value curve. This condition is suitable when the boundary between the shape echo and the flaw echo seems to be clear.

  Step S16: After the focal row direction separation is performed, the flaw candidate region extraction unit 56 executes the intra-section region recognition processing again. As a result, region recognition is achieved in which peaks adjacent to each other in the focal law direction included in the same shape candidate region 34 in FIG. 7D are separated. This is shown in FIG. 7E. The shape candidate region 34-2 is recognized as being separated into a shape candidate region 34-3 and a shape candidate region 34-4.

Step S18: The area representative point setting unit 62 sets area representative points for each of the shape candidate areas 34-1, 34-3, and 34-4. As a setting method, one of the two patterns described below is selected.
In the first setting method, the region representative point setting unit 62 determines the region representative point 49 according to the position where the echo height of the shape candidate region 34 takes the maximum value.

A second method for setting region representative points will be described with reference to FIG.
Step S32: The peak echo extraction unit 64 calculates the peak position of the echo level within the shape candidate region 34.

  Step S34: If there is one peak point of the calculated echo level (NO in step S34), the peak echo extraction unit 64 sets that peak point as a region representative point (step S36). When there are two or more peak points of the calculated echo level, the peak echo extraction unit 64 executes the processing after step S38.

  Step S38: The peak area generation unit 66 extracts a peak echo area 47 from the shape candidate area 34 (area i) as shown in FIG. The peak echo region 47 is a continuous region in which the echo height is saturated and takes an upper limit value at all positions inside the peak echo region 47. When there are two or more peak points, there is a continuous region where the echo height is saturated to the upper limit value. The peak area generation unit 66 extracts this area as the peak echo area 47.

  Step S40: The peak echo center generator 68 defines the center of the peak echo region 47 as the region representative point 49. In one embodiment, the center of gravity of the peak echo region 47 may be defined as the region representative point 48. In another embodiment, as shown in FIG. 12, a peak rectangular region 48 circumscribing the peak echo region 47 is defined, and the intersection of the diagonal points can be determined as the region representative point 49.

  In the second setting method, as shown in FIG. 16, the peak echo center generation unit 68 performs the distribution of the average value of the path direction in the effective value data and the vertical direction of the effective value data for each of the shape candidate regions 34. The average value distribution is calculated, and the region representative points 49 are determined from the calculated path direction average value distribution and vertical direction average value distribution. Here, the effective value data is data that is equal to or smaller than a predetermined value among the echo height data at each position in the shape candidate region 34. In the following description, the predetermined value is set to “0”. When the process of discarding echoes smaller than the cut threshold is being executed in step S10, the predetermined value is preferably set to “0”.

ここで、Ｎ_１は、フォーカルロー”ｉ”の各路程のうち、形状候補領域３４の内部にあり、且つ、エコー高さが”０”でない路程の数であり、Ｅｃｈｏ（ｐ）は、路程ｐのエコー高さであり、Σ_ｐは、形状候補領域３４の内部にあり、且つ、エコー高さが”０”でない路程についての和を表している。有効値データの路程方向平均値が形状候補領域３４に関与する全てのフォーカルローについて算出され、これにより、有効値データの路程方向平均値の分布が得られる。 The average value in the path direction of the effective value data is a value defined for each focal law. The average value in the path direction of the effective value data of a certain focal row for a certain shape candidate area 34 is the average value of the effective value data of a path corresponding to the position inside the shape candidate area 34 of the focal law. It should be noted that the data whose echo height is “0” is not used in the calculation of the average value in the path direction of the effective value data. When the shape candidate area 34 is defined as the contour of the echo image 32, the path height direction average value of the echo height is the road length direction average value of the effective value data as it is. For example, the path direction average value Ave ₁ (i) of valid value data of a certain focal law “i” is defined by the following formula:

Here, N ₁ is the number of paths within the shape candidate region 34 and the echo height is not “0” among the respective paths of the focal law “i”, and Echo (p) is the path length. p is the echo height, and Σ _p represents the sum of the path lengths within the shape candidate region 34 and the echo height is not “0”. The average value in the path direction of the effective value data is calculated for all focal rows related to the shape candidate region 34, and thereby the distribution of the average value in the path direction of the effective value data is obtained.

ここで、Ｎ_２は、ラインＬｉｎｅｊ上の点のうち、形状候補領域３４の内部にあり、且つ、エコー高さが”０”でない点の数であり、Ｅｃｈｏ（ｑ）は、ラインＬｉｎｅｊ上の点ｑのエコー高さであり、Σ_ｑは、形状候補領域３４の内部にあり、且つ、エコー高さが”０”でない点についての和を表している。有効値データの垂直方向平均値が形状候補領域３４に関与する全てのラインについて算出され、これにより、有効値データの垂直方向平均値の分布が得られる。 On the other hand, the average value in the vertical direction of the effective value data is a value defined for each line defined perpendicular to the path direction. The vertical average value of effective value data of a certain line for a certain shape candidate region 34 is the average value of effective value data at each point corresponding to the position inside the candidate shape region 34 on that line. It should be noted that even in the calculation of the average value in the vertical direction of the effective value data, data having an echo height of “0” is not used. When the shape candidate region 34 is defined as the contour of the echo image 32, the vertical average value of the echo height is the same as the average value in the path direction of the effective value data. For example, the path direction average value Ave ₂ (j) of the effective value data of a certain line Linej is defined by the following formula:

Here, N ₂ is the number of points on the line Linej that are inside the shape candidate region 34 and the echo height is not “0”, and Echo (q) is on the line Linej. an echo height of the point q, the sigma _q, is inside the shape candidate area 34, and the echo height represents the sum of the points is not "0". The vertical average value of the effective value data is calculated for all the lines involved in the shape candidate region 34, thereby obtaining the distribution of the vertical average value of the effective value data.

  An area representative point 49 is determined from the distribution of the calculated average value in the path direction and the distribution of the average value in the vertical direction. The area representative point 49 is defined as an intersection of a straight line extending in the path direction at the peak position of the path direction average value distribution and a line at the peak position of the vertical direction average value distribution in the shape candidate area 34. Is done.

The operation after the region representative points are set will be described with reference to the flowchart of FIG.
Step S20: The shape echo determination unit 70 extracts a shape candidate region 34 in which the region representative point is located outside the shape recognition gate 30 (step S20 NO), and the echo of the shape candidate region 34 is a flaw echo derived from a defect. It is determined that there is (step S22). In the example of FIG. 14, the shape candidate region 50-2 is determined as a flaw echo because the region representative point (rectangular geometric center) 44-2 is located outside the shape recognition gate 30.

  Step S24: The shape echo determination unit 70 extracts the shape candidate region 34 in which the region representative point is located inside the shape recognition gate 30 (step S20 YES), and determines that the echo of the shape candidate region 34 is a shape echo. . In the example of FIG. 14, the shape candidate regions 50-1 and 50-3 are determined as shape echoes because the region representative points 44-1 and 44-3 are located inside the shape recognition gate 30. The shape echo determination unit 70 creates an A scope waveform and a B scope waveform from which the shape echo has been deleted, and displays it on the display. By referring to this display screen, it is possible to obtain an easy-to-see ultrasonic flaw detection result in which the shape echo is erased.

  In the shape identification process described above, as shown in FIG. 5, after the process of separating the echo image in the path direction (A scope separation process) is performed in step S10, the echo image is converted in the focal low direction in step S14. A process of separating in the (direction perpendicular to the path length direction) (focal row direction separation process) is performed. Instead, as shown in FIG. 17, after the process of separating the echo image in the focal law direction (step S10 ′) is performed, the process of separating the echo image in the path length direction (step S14 ′) is performed. Can be done.

  In this case, in step S10 ', the echo image is separated in the focal low direction by the same process as the A scope separation process except that the direction in which the echo image is separated is different. That is, in the process in step S10 ′, as shown in FIG. 18A, instead of the A scope data 38i, the echo heights along the lines LINE1, 2,. Processing similar to the above-described A scope separation processing is performed using the distribution data. As a result, as shown in FIG. 18B, the echo image is separated in the focal low direction at the valley portion of the echo height distribution. In the example of FIG. 18B, the echo image 32-1 is separated into two echo images 32-4 and 32-5 adjacent in the focal law direction. The echo image 32-5 is composed of two echo images 32-1 and 32-3 that are overlapped side by side in the path direction. However, the echo image 32-5 is not separated in the process of separating the echo image in the focal low direction in step S10 '. After the process of separating the echo image in step S10 'in the focal law direction, the intra-section area recognition process is executed again (step S12).

  After the cross-sectional area recognition process in step S12, a process (step S14 ') for separating the echo image in the path direction is performed. The echo image is separated in the path direction by the same process as the focal law separation process except that the direction in which the echo image is separated is different. That is, in the process in step S14 ', the same process as the above-described focal row separation process is performed using A scope data instead of the echo height distribution data along each line. More specifically, for each line defined in the shape candidate area 34, the maximum value of the echo height inside the shape candidate area 34 is extracted, and the correspondence between the line and the echo height inside the shape candidate area 34 is extracted. The maximum value curve representing is extracted. In other words, the maximum value curve is a line connecting the maximum values when the echo height at each position of the shape candidate region 34 is projected in the focal low direction. Further, as shown in FIG. 18C, the shape candidate region 34 is divided with a valley portion of the maximum value curve (a portion near the position where the maximum value curve in FIG. 9C becomes a minimum) as a boundary. Is newly defined. In the example of FIG. 18C, the shape candidate region 34-5 is divided in the path direction, and the shape candidate regions 34-1 and 34-3 are newly defined. Further, for each of the shape candidate regions 34 newly defined by the process of separating the echo image in step S14 ′ in the path direction, a peak having the maximum echo height is extracted inside, and the height of the extracted peak is extracted. A predetermined percentage of the height is set as the cut threshold. Furthermore, the A scope separation unit 60 performs a truncation process of replacing the echo height data of each path whose echo height is smaller than the cut threshold of the A scope data related to the shape candidate region 34 with zero. As a result, the echo images are separated in the path direction. For example, in the example of FIG. 18D, the echo image 32-5 is separated into echo images 32-1 and 32-4.

  After step S14 'is performed, the intra-section area recognition process is executed again (step S16). As a result, as shown in FIG. 18D, the echo images 32-1 and 32-3 that are overlapped while being shifted in the path direction are separated and recognized, and each of the echo images 32-1 and 32-3 is further recognized. The shape candidate region 34-1 and the shape candidate region 34-3 are defined separately. The subsequent processing is the same as the shape identification processing described above.

  The waveform of the A scope has many variations in peak values at the rise of defects and shape echoes. If this variation is large, as shown in FIG. 17, it is possible to perform appropriate division by separating the focal row direction first and then separating the path direction.

FIG. 1 is a diagram for explaining an ultrasonic flaw detection test. FIG. 2 shows an A scope waveform. FIG. 3 is a diagram in which a plurality of A scope waveforms are arranged. FIG. 4 shows a B scope waveform. FIG. 5 shows the operation of the ultrasonic flaw detection data processing apparatus. FIG. 6A shows the positional relationship between the probe and the inspection object. FIG. 6B shows the evaluation gate. FIG. 6C shows a shape recognition gate. FIG. 6D shows an evaluation gate and a shape recognition gate. FIG. 7A shows an image that has undergone region recognition processing. FIG. 7B shows an image during A-scope separation. FIG. 7C shows an image that has undergone A-scope separation. FIG. 7D shows an image during focal row direction separation. FIG. 7E shows an image that has undergone focal row direction separation. FIG. 7F shows an image in which shape echoes have been identified. FIG. 8A shows the A scope waveform before A scope separation is performed. FIG. 8B shows the A scope waveform after A scope separation has been performed. FIG. 9A is a diagram for explaining focal-low direction separation. FIG. 9B is a diagram for explaining focal-low direction separation. FIG. 9C is a diagram for explaining focal row direction separation. FIG. 10 shows conditions for separation in the focal row direction. FIG. 11 is a diagram for explaining setting of region representative points. FIG. 12 is a diagram for explaining setting of region representative points. FIG. 13 shows a method for setting region representative points. FIG. 14 shows the shape identification process. FIG. 15 shows the configuration of an ultrasonic flaw detection data processing apparatus. FIG. 16 is a diagram for explaining a method for setting region representative points. FIG. 17 shows the operation of the ultrasonic flaw detection data processing apparatus. FIG. 18A shows an image being subjected to focal row direction separation. FIG. 18B shows an image that has undergone focal row direction separation. FIG. 18C shows an image during A-scope separation. FIG. 18D shows an image that has undergone A-scope separation.

Explanation of symbols

2 ... Inspection object 4 ... Welding line 6 ... Probe 8 ... Excitation element 10 ... Simultaneous excitation element 12 ... Composite wavefront 14 ... A scope waveform 16 ... B scope waveform 18 ... Shape 20 ... Evaluation gate 22 ... Offset amount 24 ... Shape recognition gate Upper end 26 ... Shape recognition gate lower end 28 ... Evaluation gate 30 ... Shape recognition gate 32 ... Echo image 33 ... Scratch candidate area 34 ... Shape candidate area 36 ... Focal low direction peak 38 ... A scope waveform 40 ... Focal low direction echo height distribution 41 ... Shape candidate peak 42 ... Scratch peak 44 ... Geometric center 46 ... Peak echo region 48 ... Peak echo region geometric center 50 ... Shape candidate region 52 ... Ultrasonic flaw detection data processing apparatus

Claims (21)

A shape data storage unit for storing shape data indicating the shape of the object to be inspected;
In the target data obtained by ultrasonic flaw detection, a wound candidate region extraction unit that extracts a region where an echo satisfies a predetermined criterion as a wound candidate region;
A shape candidate region extraction unit that extracts a region that overlaps a shape gate region set based on the shape data among the scratch candidate regions, as a shape candidate region;
A divided shape candidate region generating unit that generates a divided shape candidate region by a dividing process of dividing the shape candidate region with a portion having a minimum echo height as a boundary;
A representative point setting unit for setting a representative point representing the position of the divided shape candidate region;
An ultrasonic flaw detection data processing apparatus comprising: a shape echo determination unit that determines that the divided shape candidate region is a shape echo when the representative point is included in a setting region set based on the shape data. The ultrasonic flaw detection data processing apparatus according to claim 1,
The representative point setting unit
A peak echo extraction unit for extracting a plurality of echo height peaks from the divided shape candidate regions;
An ultrasonic flaw detection data processing apparatus comprising: a peak echo center generation unit that determines the representative point from a region surrounded by the plurality of peaks by a predetermined method. The ultrasonic flaw detection data processing apparatus according to claim 2,
The representative point setting unit
A path direction average value which is an average value of the path value direction of a beam used for the ultrasonic flaw detection of effective value data excluding data equal to or less than a predetermined value among the echo height data inside the divided shape candidate area A path direction average value calculation unit for calculating the distribution of
A vertical direction average value calculating unit that calculates a distribution of vertical average values that are average values in the vertical direction perpendicular to the path direction of the effective value data;
The representative point is set as an intersection of a straight line extending in the path direction at the peak position of the path direction average value distribution and a straight line extending in the vertical direction at the peak position of the vertical direction average value. Sonic flaw detection data processing device. The ultrasonic flaw detection data processing device according to any one of claims 1 to 3,
The said division | segmentation process is an ultrasonic flaw detection data processing apparatus containing A scope separation process which divides | segments the said shape candidate area | region in the path | route direction of the beam used for the said ultrasonic flaw detection. The ultrasonic flaw detection data processing apparatus according to claim 4,
The ultrasonic flaw detection data processing apparatus further includes a focal row direction separation process for further dividing the shape candidate region in a direction perpendicular to the path direction in the B scope after the A scope separation process. The ultrasonic flaw detection data processing device according to claim 5,
The focal row direction separation process is:
A maximum value extraction process for extracting the maximum value of the echo height in the path direction of the shape candidate region corresponding to each position in the vertical direction perpendicular to the path direction,
An ultrasonic flaw detection data processing apparatus comprising: a valley extraction process for generating the divided shape candidate area by dividing the shape candidate area by using a valley position in the vertical distribution of the maximum value as a boundary. The ultrasonic flaw detection data processing device according to claim 5,
The focal row direction separation process is:
Distribution of path direction average value which is an average value of path value directions of beams used for ultrasonic flaw detection of effective value data excluding data below a predetermined value from echo height data inside the shape candidate region A path direction average value calculating process for calculating
Ultrasonic flaw detection data processing including: valley extraction processing for generating the divided shape candidate region by dividing the effective value data by dividing a valley position in the vertical distribution perpendicular to the path direction as a boundary apparatus. Obtaining shape data relating to the shape of the object to be inspected;
In the target data obtained by ultrasonic flaw detection, a scratch candidate region extraction step for extracting a region where an echo satisfies a predetermined criterion as a scratch candidate region;
Extracting a region overlapping with a shape gate region set based on the shape data, as a shape candidate region, from among the scratch candidate regions;
Generating a shape candidate region to be divided by a division process of dividing the shape candidate region with the echo height as a boundary, and
A representative point setting step for setting a representative point representing the position of the divided shape candidate region;
An ultrasonic flaw detection data processing method comprising: determining that the shape candidate region to be divided is a shape echo when the representative point is included in a set region set based on the shape data. The ultrasonic flaw detection data processing method according to claim 8,
The representative point setting step includes:
Extracting a plurality of echo height peaks from the divided shape candidate regions;
An ultrasonic flaw detection data processing method comprising: determining the representative point by a predetermined method from a region surrounded by the plurality of peaks. The ultrasonic flaw detection data processing method according to claim 8,
The representative point setting step includes:
A path direction average value which is an average value of the path value direction of a beam used for the ultrasonic flaw detection of effective value data excluding data equal to or less than a predetermined value among the echo height data inside the divided shape candidate area Calculating the distribution of
Calculating a distribution of vertical average values that are average values in a vertical direction perpendicular to the path direction of the effective value data;
Setting the representative point as an intersection of a straight line extending in the path direction at the peak position of the path direction average value distribution and a straight line extending in the vertical direction at the peak position of the vertical direction average value An ultrasonic flaw detection data processing method comprising: The ultrasonic flaw detection data processing method according to any one of claims 8 to 10,
The said division | segmentation process is an ultrasonic flaw detection data processing method including the A scope separation | segmentation process which divides | segments the said shape candidate area | region in the path | route direction of the beam used for the said ultrasonic flaw detection. The ultrasonic flaw detection data processing method according to claim 11,
The ultrasonic flaw detection data processing method further includes a focal low direction separation process in which the shape candidate area is further divided in a direction perpendicular to the path direction in the B scope after the A scope separation process. The ultrasonic flaw detection data processing method according to claim 11,
The division processing further includes a focal flaw direction separation process in which the shape candidate region is divided in a direction perpendicular to the path direction in the B scope before the A scope separation process. The ultrasonic flaw detection data processing method according to claim 12,
The focal row direction separation process is:
A maximum value extraction process for extracting the maximum value of the echo height in the path direction of the shape candidate region corresponding to each position in the vertical direction perpendicular to the path direction,
An ultrasonic flaw detection data processing method comprising: a valley extraction process for generating the divided shape candidate area by dividing the shape candidate area by using a valley position in the vertical distribution of the maximum value as a boundary. Obtaining shape data relating to the shape of the object to be inspected;
In the target data obtained by ultrasonic flaw detection, a scratch candidate region extraction step for extracting a region where an echo satisfies a predetermined criterion as a scratch candidate region;
Extracting a region overlapping with a shape gate region set based on the shape data, as a shape candidate region, from among the scratch candidate regions;
Generating a shape candidate region to be divided by a dividing process of dividing the shape candidate region with a portion having a minimum echo height as a boundary;
A representative point setting step for setting a representative point representing the position of the divided shape candidate region;
Ultrasonic flaw detection data for causing a computer to execute a method comprising: determining that the divided shape candidate region is a shape echo when the representative point is included in a set region set based on the shape data Processing program. An ultrasonic flaw detection data processing program according to claim 15,
The representative point setting step includes:
Extracting a plurality of echo height peaks from the divided shape candidate regions;
An ultrasonic flaw detection data processing program comprising: determining the representative point by a predetermined method from a region surrounded by the plurality of peaks. An ultrasonic flaw detection data processing program according to claim 15 or 16,
The division process is an ultrasonic flaw detection data processing program including an A scope separation process that divides the shape candidate region in a path direction of a beam used for the ultrasonic flaw detection. An ultrasonic flaw detection data processing program according to claim 17,
The division processing further includes an ultrasonic flaw detection data processing program including, after the A scope separation processing, a focal row direction separation processing for further dividing the shape candidate region in a direction perpendicular to the path length direction in the B scope. An ultrasonic flaw detection data processing program according to claim 17,
The division processing further includes an ultrasonic flaw detection data processing program including a focal row direction separation process for dividing the shape candidate region in a direction perpendicular to the path direction in the B scope before the A scope separation process. An ultrasonic flaw detection data processing program according to claim 18 or 19,
The focal row direction separation process is:
A maximum value extraction process for extracting the maximum value of the echo height in the path direction of the shape candidate region corresponding to each position in the vertical direction perpendicular to the path direction,
An ultrasonic flaw detection data processing program comprising: a valley extraction process for generating the divided shape candidate area by dividing the shape candidate area by using a valley position in the vertical distribution of the maximum value as a boundary. An ultrasonic flaw detection data processing program according to claim 18 or 19,
The focal row direction separation process is:
Distribution of mean value of path direction which is an average value of path value direction of a beam used for ultrasonic flaw detection of effective value data excluding data below a predetermined value among data of echo height inside the shape candidate region A path direction average value calculating process for calculating
Ultrasonic flaw detection data processing including: valley extraction processing for generating the divided shape candidate region by dividing the effective value data by dividing a valley position in the vertical distribution perpendicular to the path direction as a boundary program.

Images