JP3616193B2 - Method and apparatus for determining wounds to be inspected - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a method and apparatus for inspecting an object to be inspected, which is applied to a rail flaw determination device or the like in an ultrasonic rail flaw detection vehicle and determines a flaw on a railroad rail.
[0002]
[Prior art]
FIG. 5 is a perspective view showing a measurement state in which a conventional railroad rail or the like is flaw-detected with ultrasonic waves. In FIG. 5, in this example, the rail 1 is flawed by moving the ultrasonic probe 4 provided at the tip of the cable 3 connected to the ultrasonic flaw detector 2 to the rail 1 to be flawed. . At this time, a transmission signal output at a constant period is radiated to the rail 1 as an ultrasonic pulse from the transducer in the ultrasonic probe 4. The ultrasonic pulse is reflected by a flaw on the rail 1 and the reflected wave is received by the transducer of the ultrasonic probe 4.
[0003]
This received signal is extracted by the ultrasonic flaw detector 2 through a gate circuit and subjected to digital signal processing. Furthermore, the received signal level is compared with the judgment level to detect a flaw in the rail 1, and the position of the rail 1 based on the flaw data and the amount of movement from a travel distance sensor (not shown) is displayed on a cathode ray tube (CRT) 6 or the like. it's shown. Further, the flaw detection information is stored and stored in a memory or the like of the recording device, and further printed on a recording sheet and output as necessary.
[0004]
In such a rail flaw detection, the ultrasonic probe 4 emits ultrasonic waves in consideration of the site and direction of the flaws through a plurality of channels, and flaw detection is performed from the reflected echoes. As multiple channels, to accommodate important rail wounds, for example, vertical channels for horizontal cracks from the head to the bottom of the rail 1, and ± 45 ° channels for lateral fissures from the abdomen to the bottom, For transverse head cracks, ultrasonic waves are emitted from a plurality of channels such as ± 70 ° channels in consideration of the site and direction of the wound, and flaw detection is performed focusing on the reflected echo.
[0005]
In recent rail flaw detection, automation of reflection echo processing has progressed, and some automatic determination of flaws has also been performed.
[0006]
[Problems to be solved by the invention]
However, with such conventional rail flaw detection, it is difficult to perform highly reliable automatic determination. For example, in the reflected echo, a reflected echo from an artificial structure such as a bolt hole is generated in addition to the original scratch in the rail. Furthermore, since there is a delayed echo generated by the reflection path such as the rail side surface instead of the shortest path to the rail flaw, the automatic determination with high reliability could not be performed.
[0007]
For this reason, in order to identify the reflected echo and the delayed echo from the artificial structure and perform the rail scratch determination with high reliability, it is drawing attention to perform the determination from the B scope image of the reflected echo. For example, it has been found that the B-scope image shown in FIG. 6 is obtained by the vertical channel and the ± 45 ° channel, and the bolt hole can be identified from this image.
[0008]
However, the data amount of the B scope image is enormous, and when the ultrasonic rail flaw detection vehicle travels at a high speed, the determination process in real time becomes difficult. Moreover, even if it is a known artificial structure, for example, the B-scope image may change depending on the acoustic coupling state between the ultrasonic probe and the rail, so that a highly reliable automatic determination cannot be performed. There is also.
[0009]
The present invention solves such a conventional problem, and automatically recognizes an artificial structure and a delayed echo from a B-scope image of a reflected echo at high speed in real time to automatically perform highly reliable damage determination. It is an object of the present invention to provide a method and apparatus for determining a wound to be inspected that can be performed automatically.
[0010]
[Means for Solving the Problems]
In order to achieve this object, the method for determining an object to be inspected according to the present invention uses a B-scope image of a reflected echo obtained by radiating a plurality of channels of ultrasonic pulses to the object to be inspected. A connected region on a B-scope image for each channel of a plurality of channels is extracted, then one or a plurality of feature amounts are calculated for each extracted connected region, and the calculated feature amount The scratch on the object to be inspected is determined based on the above.
[0011]
In addition, the inspection object damage determination method of the present invention calculates the center of gravity as the feature amount of each connected area, and then identifies the artificial structure in the object to be inspected using the calculated center of gravity of each connected area. doing.
In addition, the method for determining a test object wound according to the present invention identifies delayed echoes as a condition that all the reflected echoes constituting the connected areas in the extracted connected areas are not the first reflected echoes after the emission of ultrasonic pulses. doing.
[0012]
In addition, the inspected object wound determination apparatus of the present invention determines an object to be inspected from a B-scope image of a reflected echo obtained by radiating a plurality of channels of ultrasonic pulses to the inspected object. A channel ultrasonic transmission / reception unit radiates a plurality of channels of ultrasonic pulses to the object to be inspected, and outputs received reception data. Further, the connected region extraction processing unit extracts a connected region on the B-scope image for each channel of the plurality of channels from the reception data of the multi-channel ultrasonic transmission / reception unit. Further, the feature amount calculation processing unit calculates one or a plurality of feature amounts for each connected region extracted by the connected region extraction processing unit. Then, the determination processing unit determines a scratch on the object to be inspected based on the feature amount calculated by the feature amount calculation processing unit.
[0013]
Furthermore, the inspected object wound determination apparatus of the present invention is provided with an artificial structure identification processing unit in the inspected object wound determination apparatus, and the artificial structure identification processing unit is calculated by the feature amount calculation processing unit. The artificial structure is identified based on the center of gravity of the connection area.
The inspected object wound determination apparatus according to the present invention includes a delayed echo identification processing unit provided in the inspected object wound determination apparatus, and each connected region extracted by the connected region extraction processing unit by the delayed echo identification processing unit. The delayed echo is identified as a condition that the reflected echo constituting the connected region in the inside is not the first reflected echo after the ultrasonic pulse radiation.
[0014]
Furthermore, the inspection object wound determination method and apparatus of the present invention uses an inspection object as a rail.
In addition, the method and apparatus for determining the inspection object flaw of the present invention is applied to a rail flaw determination apparatus in an ultrasonic rail flaw detection vehicle.
In such a method and apparatus for inspecting an object to be inspected according to the present invention, one connected region on the B scope image of the reflected echoes of each of a plurality of channels corresponds to one reflection source. Therefore, a connected region on the B-scope image is extracted for each of a plurality of channels, and one or a plurality of feature amounts are calculated for each of the extracted connected regions. The data is greatly compressed into several feature quantities. By performing subsequent data processing using this feature amount, it is possible to determine a scratch at high speed.
[0015]
In the inspection object wound determination method and apparatus of the present invention, the size of the connection area on the B-scope image changes depending on the acoustic coupling state between the probe and the object to be inspected. The artificial structure of the object to be inspected is surely identified by paying attention to the center of gravity of each connection region, which is hardly affected by the influence.
Furthermore, in the inspection object wound determination method and apparatus of the present invention, since the reflected echo of the shortest path exists before the delayed echo, all echoes constituting the connected region are the first after the ultrasonic pulse radiation. A delayed echo is reliably identified on the condition that it is not an echo of As a result, the artificial structure and the delayed echo are reliably identified from the B-scope image of the reflected echo at high speed in real time, the damage determination is automated, and high reliability is obtained.
[0016]
DETAILED DESCRIPTION OF THE INVENTION
Next, an embodiment of the method and apparatus for determining a wound to be inspected according to the present invention will be described in detail with reference to the drawings.
FIG. 1 is a block diagram showing a configuration of an embodiment of an inspection subject wound determination apparatus of the present invention. In FIG. 1, this inspected object damage determination apparatus is mounted on an ultrasonic rail flaw detection vehicle that travels at a high speed, so that a multi-channel ultrasonic probe 11 moves in contact with a rail 10 to be inspected. It has become. The multi-channel ultrasonic probe 11 includes a vertical channel ultrasonic probe 11a and ± 45 ° channel ultrasonic probes 11b and 11c.
[0017]
Further, the multi-channel ultrasonic probe 11 is radiated with ultrasonic waves, and the reflected signal of the reflected echo received by the multi-channel ultrasonic probe 11 is amplified, and further processing such as digital signal processing is performed. A multi-channel ultrasonic flaw detector 12 is provided. This multi-channel ultrasonic flaw detector 12 is composed of a transmission unit (not shown) that transmits a multi-channel ultrasonic signal and a reception unit that amplifies the reflection signal of the multi-channel reflection echo and performs digital signal reception processing. Yes.
[0018]
Furthermore, this inspected object wound determination apparatus generates a B scope image from the reception data output from the multi-channel ultrasonic flaw detection apparatus 12, identifies artificial structures and delayed echoes, and automatically obtains high reliability. A scratch determination processing device 13 for performing scratch determination at high speed is provided.
FIG. 2 is a block diagram showing a detailed configuration of the scratch determination processing device 13 shown in FIG. In FIG. 2, the flaw determination processing device 13 includes a CPU 21 that controls each unit of the device, an image memory 22 that stores B-scope image data generated by the arithmetic processing of the CPU 21, and a ROM 23 that stores a control program. A working RAM 24 is provided.
[0019]
Further, the scratch determination processing device 13 transmits / receives storage data to / from the storage medium, an I / O circuit 25 that captures reception data output from the multi-channel ultrasonic inspection device 12, a CRT controller 26 that controls screen display. An I / O circuit 27 to perform is provided. A device adapter 28 for connecting devices, a bus line 29 to which each unit is connected, and a mouse 30 for performing coordinate input operations are provided.
[0020]
Further, a keyboard 31 for performing an input operation such as function selection, an external storage device 32 for storing the processing data through the I / O circuit 27 and sending the storage data, and a CRT monitor 33 for displaying the processing data on the screen are provided. It has been.
FIG. 3 is a block diagram illustrating the function of the flaw determination process performed by the CPU 21. In FIG. 3, this flaw determination processing function includes a B scope image generation processing unit 40 that generates a B scope image, a connected region extraction processing unit 41 that processes extraction of a connected region, and a feature amount calculation process that processes calculation of a feature amount. A portion 42 is provided.
[0021]
Further, the function of the scratch determination process includes a delayed echo identification processing unit 44 for processing identification of delayed echo, an artificial structure identification processing unit 45 for processing identification of artificial structures such as bolt holes, and the like in FIG. A scratch determination unit 46 that determines a scratch on the rail 10 and an output display processing unit 47 for displaying the determination result are provided.
The operation in the configuration of this embodiment will be described.
[0022]
In FIG. 1, this inspected object damage determination apparatus is mounted on an ultrasonic rail flaw detection vehicle that travels at a high speed, and the multi-channel ultrasonic probe 11 to the rail 10 generates an ultrasonic wave in consideration of the generation site and directionality. It emits sound waves. That is, ultrasonic waves are emitted from the ultrasonic probe 11a of the vertical channel for horizontal cracks from the head to the bottom of the rail 10, and ± 45 ° channels are used for lateral fissures from the abdomen to the bottom. Ultrasonic waves are emitted from the ultrasonic probes 11b and 11c, and the reflected echoes are received.
[0023]
Ultrasound is input to the multi-channel ultrasonic probe 11 from the transmission unit of the multi-channel ultrasonic flaw detector 12. In addition, the reception unit amplifies the reflected signal of the reflected echo received by the multi-channel ultrasonic probe 11, further performs processing such as digitization, and the received data is input to the flaw determination processing device 13.
In the flaw determination processing device 13 shown in FIG. 2, the CPU 21 fetches the received data through the I / O circuit 25 and the bus line 29, and based on the control program read from the ROM 23, the flaw determination processing function block shown in FIG. Data processing is performed. That is, a B-scope image of a reflected echo is generated from reception data acquired from the multi-channel ultrasonic flaw detector 12, and an artificial structure and a delayed echo are identified from the B-scope image, and thereafter, the reliability is automatically high. Wound judgment is performed. This data processing is performed at high speed in real time.
[0024]
The processing data in this case is temporarily stored in the working RAM 24, and the processing data is displayed on the CRT monitor 33 on the screen. Further, the processing data is stored in the external storage device 32 through the I / O circuit 27. This process instruction operation is performed by the mouse 30 and the keyboard 31 connected to the device adapter 28.
[0025]
FIG. 4 is a flowchart showing a processing procedure of the operation of the scratch determination processing. 1 to 4, in the wound determination processing device 13, the B scope image generation processing unit 40 performs the B scope image generation processing in step S <b> 10. In this B-scope image generation processing, as shown in FIG. 1, ultrasonic pulses from a plurality of vertical channels and ± 45 ° channels are radiated to the rail 10 to be inspected, and the reflected echoes of the respective channels obtained by this radiation are reflected. Processing to generate a B scope image is performed.
[0026]
In this B scope image, the incident position of the ultrasonic pulse in the longitudinal direction of the rail 10 and the beam path length (propagation time) are set vertically and horizontally, and the pixel value at the position where the reflected echo exists is a constant value (for example, “1”). ) Or a value corresponding to the intensity of the reflected echo, and the pixel value at a position where the reflected echo does not exist is, for example, “0”. Although such a B-scope image may be used as it is, it is converted into a B-scope image with the position in the longitudinal direction of the rail 10 as the horizontal axis and the depth from the surface of the rail 10 as the vertical axis for easy understanding. As shown in FIG. 1, the multi-channel ultrasonic probe 11 is composed of a vertical channel ultrasonic probe 11a and ± 45 ° channel ultrasonic probes 11b and 11c. When the three B scope images of the channel are overlaid, the bolt hole of the rail 10 is obtained as the image shown in FIG.
[0027]
In this conversion, the longitudinal position of the rail 10 is x, the depth from the surface of the rail 10 is y, the incident position of the ultrasonic pulse in the longitudinal direction of the rail 10 is p, the beam path length is q, and the incident angle of the ultrasonic pulse. Is represented by the following equations (1) and (2).
x = p + q · sin θ (1)
y = q · cos θ (2)
Next, in step S11, the connected region extraction processing unit 41 performs a connected region extraction process. In this connected region extraction process, a process of extracting a connected region from the B scope image from the reflected echoes of the vertical channel and ± 45 ° channel is performed. That is, a set of reflection echo pixels that are mutually connected on the B-scope image is extracted as one connected region. This can be done, for example, using an image processing technique called labeling. This labeling is described in, for example, “Introduction to Computer Image Processing” (Supervised by Tamura, Research Institute Publishing, 1985), and performs a process of assigning a label (number) to each connected area.
[0028]
In addition to the case where the pixels with the reflective echo are adjacent to each other, the connection condition may be, for example, that the number of pixels without the reflective echo existing between the pixels with the reflective echo is a predetermined number or less. In this case, even if the reflected echo is somewhat missing, it can be regarded as one connected region.
Further, in step S12, the feature amount calculation processing unit 42 performs a feature amount calculation process. This feature amount calculation process is a process of calculating the feature amount of each connected region extracted by the connected region extraction processing. As the feature amount, in addition to the center of gravity, for example, the size (number of pixels), the length (Distance between end points) and the like can be considered.
[0029]
Although the center of gravity is defined as a moment feature, it can be simply calculated as the center of both end points of the connected region.
Next, in step S13, the delayed echo identification processing unit 44 performs delayed echo identification processing. This delayed echo identification process is performed under the condition that, among the connected areas extracted by the connected area extraction process, all the pixels constituting the connected area are pixels of the second and subsequent reflected echoes after the ultrasonic pulse emission. Is a process for identifying a delayed echo. Whether each pixel is the first echo after the ultrasonic pulse radiation is determined by, for example, examining whether there is a pixel with a reflected echo in the direction of a short beam path from each pixel on the B scope image. .
[0030]
In step S14, the artificial structure identification processing unit 45 performs an artificial structure identification process. In the artificial structure identifying process, the artificial structure is identified based on the center of gravity of each connected region calculated in the feature amount calculating process. In this process, a bolt hole is assumed as an artificial structure of the rail 10, and identification of the bolt hole will be described below. It has been found that when the three B scope images of the vertical channel and the ± 45 ° channel are superimposed, the bolt hole of the rail 10 is obtained as the image shown in FIG. That is, the positional relationship of the connection regions using the bolt holes of the three channels as the reflection source is determined, and it is also found that the bolt holes exist in the abdomen of the rail 10. Therefore, the bolt hole can be identified by the following conditions.
[0031]
This condition is that “the center of gravity of the connection region of the vertical channel is a predetermined vertical position (depth from the surface of the rail 10) corresponding to the abdomen of the rail 10 and a predetermined position relative to the center of gravity position. There are both a + 45 ° channel connection region and a −45 ° channel connection region having a center of gravity within the range.
As other artificial structures, bond holes and Meihan holes smaller than bolt holes can be similarly identified.
[0032]
Further, in step S15, the scratch determination unit 46 performs a scratch determination process on the rail 10 shown in FIG. In the determination of the scratch, the scratch on the rail 10 is determined with respect to the connection region of the vertical channel and the ± 45 ° channel that is not identified as a delayed echo or an artificial structure. For example, first, the vertical channel is determined as a horizontal fissure, and the ± 45 ° channel is determined as a lateral fissure. And the head part, abdominal part, and bottom part of a generation | occurrence | production site | part are identified by the position (depth from the surface of the rail 10) on the vertical axis | shaft of the gravity center of a connection area | region. Further, the weight of the wound is determined based on the length of the connection region. In step S16, the output display processing unit 47 performs processing for displaying the scratches determined by the CRT monitor 33 shown in FIG.
[0033]
In this embodiment, the rail 10 is described as the object to be inspected and applied to the ultrasonic rail flaw detection vehicle. However, the present invention can also be applied to other objects to be inspected in which the scratch should be identified at high speed. . The effect is also the same.
Thus, in this embodiment, a B-scope image of the reflected echo is generated from the reception data output from the multi-channel ultrasonic flaw detector 12, and the artificial structure and the delayed echo of the rail 10 are equal to the real time from this B-scope image. It can be identified at high speed and reliably, and the damage can be judged automatically and with high reliability.
[0034]
【The invention's effect】
As is clear from the above description, according to the inspected object wound determination method and apparatus of the present invention, one or a plurality of feature amounts are calculated for a connected region on a B-scope image for each channel of a plurality of channels, Based on the calculated feature amount, the scratch on the object to be inspected is determined, so that real-time high-speed processing is possible, and artificial structures and delayed echoes can be reliably identified. In addition, it is possible to perform highly reliable scratch determination.
[Brief description of the drawings]
FIG. 1 is a block diagram showing a configuration of an embodiment of an inspection subject wound determination apparatus according to the present invention. FIG. 2 is a block diagram showing a detailed configuration of a wound determination processing apparatus of FIG. FIG. 4 is a functional block diagram of a flaw determination process performed by a CPU in the determination processing apparatus. FIG. 5 is a flowchart showing a processing procedure of flaw determination processing in the embodiment. FIG. FIG. 6 is a perspective view showing a state. FIG. 6 is a diagram showing a B-scope image of a conventional example.
10: Rail 11: Multichannel ultrasonic probe 11a-11c: Ultrasonic probe 12: Multichannel ultrasonic flaw detector 13: Flaw determination processing device 21: CPU
22: Image memory 23: ROM
24: RAM
25, 27: I / O
26: CRT controller 28: device adapter 29: bus line 30: mouse 31: keyboard 32: external storage device 33: CRT monitor 40: B scope image generation processing unit 41: connected region extraction processing unit 42: feature amount calculation processing unit 43 : File 44: Delayed echo identification processing unit 45: Artificial structure identification processing unit 46: Scratch determination unit 47: Output display processing unit

Claims (10)

In the inspection object wound determination method for determining a wound of the inspection object from a B scope image of a reflected echo obtained by radiating a plurality of channels of ultrasonic pulses to the inspection object,
A connected region on the B-scope image for each channel of the plurality of channels is extracted, and then one or a plurality of feature amounts are calculated for each extracted connected region, and the coverage is further calculated based on the calculated feature amount. A method for determining a wound to be inspected, comprising: determining a wound on an inspection object. In the inspected object wound determination method according to claim 1,
Inspected object wound determination method characterized by calculating a center of gravity as a feature value of each connected area and then identifying an artificial structure in the inspected object using the calculated center of gravity of each connected area . In the inspected object wound determination method according to claim 1 or 2,
Inspected object wound determination method characterized in that delayed echoes are identified on condition that all the reflected echoes constituting the connected region in each extracted connected region are not the first reflected echo after the ultrasonic pulse radiation. . In the inspected object wound determination apparatus for determining a wound of the inspected object from a B scope image of a reflected echo obtained by radiating a plurality of channels of ultrasonic pulses to the inspected object,
A multi-channel ultrasonic transmission / reception unit that radiates a plurality of channels of ultrasonic pulses to an object to be inspected and outputs received reception data;
A connected region extraction processing unit that extracts a connected region on a B-scope image for each channel of the plurality of channels from the reception data of the multi-channel ultrasonic transmission / reception unit;
A feature amount calculation processing unit that calculates one or a plurality of feature amounts for each connected region extracted by the connected region extraction processing unit;
A determination processing unit that determines a scratch on the object to be inspected based on the feature amount calculated by the feature amount calculation processing unit;
A device for determining an object to be inspected is provided. An artificial structure identification processing unit is provided in the inspected object wound determination apparatus according to claim 4, and the artificial structure identification processing unit is configured to perform artificial determination based on the center of gravity of each connected region calculated by the feature amount calculation processing unit. An inspection object wound determination apparatus characterized by identifying a structure. 5. A reflection echo identification processing unit is provided in the inspected wound determination apparatus according to claim 4, and a reflection that constitutes a connected region in each connected region extracted by the connected region extraction processing unit by the delayed echo identification processing unit. echo, on condition that not the first echo after ultrasonic pulsed radiation, the test subject wound determination apparatus characterized by identifying the delay echo. Claims 1 to In test subject wound determination how according to any of 3, the inspection object wound determination how, wherein the object to be inspected is rail. Claims 1 to In test subject wound determination how according to any of 3, the inspection object wound determination how, characterized in that applied to the rail flaw determination unit in the ultrasonic rail flaw detection vehicle. 7. The inspected object wound determination apparatus according to claim 4, wherein the inspected object is a rail. 7. The inspection object damage determination apparatus according to claim 4, wherein the inspection object damage determination apparatus is applied to a rail damage determination apparatus in an ultrasonic rail flaw detection vehicle.

Images