JPH0915217A - Plate wave ultrasonic flaw detection method and apparatus - Google Patents

Abstract

(57) [Summary] [Object] To provide a plate wave ultrasonic flaw detection method capable of easily detecting a flaw without discriminating between a flaw and noise and increasing a flaw detection speed, and an apparatus used for the implementation. To do. [Structure] The smoothing processing unit 91 of the CPU 9, for each signal corresponding to a plurality of pixels forming a two-dimensional flaw detection image read from the frame memory 7, for example, in a pixel range near eight arbitrary pixels in the two-dimensional flaw detection image. The gray level of the signal of each pixel is smoothed, and the level of the signal of the pixel is smoothed to obtain a smoothed image. The binarization processing unit 92 binarizes the smoothed image using 128 gradations as a threshold. The binarized defect detection unit 93 obtains a defect extraction image in which a defect image is extracted from the binarized image, and the defect detection unit 94 superimposes the defect extraction image on the two-dimensional flaw detection image or the smoothing processed image to overlap the overlapped portion. Is extracted as a defect image from the two-dimensional flaw detection image or the smoothed image, and the defect determination unit 95 determines the degree of the defect based on the extracted defect image.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for injecting a plate wave ultrasonic wave into a material to be inspected and receiving a reflected wave thereof to detect defects in the material to be rolled, and an apparatus used for implementing the method. .

[0002]

2. Description of the Related Art In order to perform online non-destructive inspection of defects on the surface or inside of a material to be inspected having a relatively small thickness, such as hot-rolled steel sheet and cold-rolled steel sheet, the tire probe is used to perform flaw detection. Plate wave ultrasonic flaw detection is carried out to detect defects in the material to be inspected by propagating plate wave ultrasonic waves to the material and receiving the reflected waves, and whether or not a signal based on the defects is contained in it. It is being appreciated.

FIG. 18 is a schematic cross-sectional view showing a usage mode of a tire probe, and S in the figure is a belt-shaped flaw-to-be-detected material conveyed in its longitudinal direction. The surface of the material S to be inspected has a contact medium 15
Is uniformly applied to a predetermined thickness. A support rod 13 is vertically arranged above the flaw detection material S, and a fixed shaft 16 is supported in the width direction of the flaw detection material S near the lower end of the support rod 13. The fixed shaft 16 has a tire probe 22 that rolls on the material S to be inspected.
Is rotatably mounted.

The tire probe 22 has grooves 18 and 18 on its periphery.
And a belt-like tire portion 14 made of rubber or the like that surrounds the wheels 17 and 17, and both edges of the tire portion 14 are grooves 18 and 18 of the wheels 17 and 17, respectively. It is fixed to. A plate wave probe 20 that transmits and receives ultrasonic waves at a predetermined cycle is fixed to a fixed shaft 16 of the tire probe 22 while being inclined at a predetermined angle in the direction of the edge of the material S to be detected. Further, the tire probe 22 is filled with the contact medium 15, and the ultrasonic wave generated by the plate wave probe 20 passes through the contact medium 15, the tire portion 14 and the contact medium 15 in the tire probe 22. S to be inspected
Is incident at a predetermined incident angle in parallel with the width direction of the flaw-detecting material, and there is a plate-wave ultrasonic wave in a vibration mode corresponding to the incident angle of the ultrasonic wave, the thickness of the flaw-detecting material S, and the frequency of the ultrasonic wave. It is converted and propagates in the flaw detection material S.

The plate wave ultrasonic waves propagated through the material S to be detected are
Defects generated on the surface or inside of the flaw-detecting material S, and reflected by the edge portion of the flaw-detecting material S, reflected waves are contacted in the contact medium 15, the tire portion 14, and the tire probe 22 on the surface of the flaw-detecting material S. medium
A flaw detection signal is obtained by being received by the plate wave probe 20 via 15.

FIG. 19 is a graph showing an example of a flaw detection signal by plate wave ultrasonic flaw detection. In the figure, the vertical axis represents the strength of the flaw detection signal, and the horizontal axis represents the time after the ultrasonic wave is transmitted. There is. As shown in FIG. 19, within a predetermined time A immediately after the transmission of ultrasonic waves, a plurality of signals R _E received by diffuse reflection near the probe appear. Then, within the time B which is the propagation time of the plate wave ultrasonic wave propagated in the width direction of the flaw detection material, the signal T _E received by the reflection of the defect of the flaw detection material and the signal N _E generated by noise appear. After that, the signal T _E received by the reflection of the edge portion of the flaw detection material appears for a predetermined time C. The width of the reflected signal from the edge portion is wide as described above because the plate wave ultrasonic waves propagating in the flaw-detecting material are waves in which plate waves of a plurality of vibration modes having different propagation velocities are superposed. This is because the reflected wave from the edge portion is received for each plate wave.

[0007] In the conventional plate wave ultrasonic flaw detector, the gate is opened at the timing when the ultrasonic wave is transmitted and time A elapses.
By opening the gate only for time B, only the reflected wave due to the defect is received, and when the flaw detection signal includes a signal equal to or higher than a predetermined threshold value,
Judge that there is a defect. Then, the size of the defect is evaluated based on the intensity of the flaw detection signal, and half of the time that the reflected wave due to the defect is received after the ultrasonic wave is transmitted, The existing position of the defect in the width direction of the flaw detection target material was obtained from the product of the propagation velocity of the flaw and the propagation velocity of the flaw.
The timing of opening the gate and the width of the gate region described above are determined in advance based on the material, plate width, plate thickness, etc. of the material to be inspected.

However, in such a plate wave ultrasonic flaw detector, as shown in FIG. 19, the noise signal may also exceed the threshold value, in which case there is a problem that a defect is erroneously detected. Therefore, a plurality of flaw detection signals obtained by receiving the plate wave ultrasonic waves propagating in the width direction of the flaw detection material are given as A (analog) /
It is considered to be a defect based on the shapes of a plurality of images in the image by performing a D (digital) conversion to obtain a grayscale signal, arranging the grayscale signal in the length direction of the flaw detection material to form a two-dimensional image. A plate wave ultrasonic flaw detection apparatus has been developed which extracts an image, determines whether each extracted image is a defect or noise, and detects only the defect.

[0009]

However, in the conventional plate wave ultrasonic flaw detector for performing image processing, a high-performance processing device for discriminating between defects and noises based on the area of each image. Therefore, there is a problem that the device cost is high. Further, since such a determination process requires time, there is a problem in that the flaw detection speed cannot be increased by increasing the transport speed of the flaw detection target material. Further, as the width of the material to be inspected increases, the processing amount of the image processing increases, so that there is a limit to the width of the material to be inspected that can efficiently detect defects by the image processing.

The present invention has been made in view of the above circumstances, and an object of the present invention is to smooth a two-dimensional image formed by a plurality of flaw detection signals and then binarize it to generate noise from the two-dimensional image. By adopting a configuration in which the image relating to is removed, it is possible to easily detect a defect without discriminating whether the image is an image due to a defect or an image due to noise and to increase the flaw detection speed. An object of the present invention is to provide an ultrasonic flaw detection method and an apparatus used for the implementation.

Another object of the present invention is as follows.
By compressing while holding the information of the maximum value of each signal in a predetermined range of a two-dimensional image or a smoothed image, the processing amount of image processing is reduced, the flaw detection speed is high, and a wide-width flaw-detecting material is obtained. Another object of the present invention is to provide a plate wave ultrasonic flaw detection method applicable to the above and an apparatus used for the implementation.

[0012]

SUMMARY OF THE INVENTION A plate wave ultrasonic flaw detection method according to a first aspect of the present invention is a method for performing ultrasonic wave detection while relatively moving a flaw detection material and an ultrasound probe arranged to face the flaw detection material. Two-dimensional flaw detection image is formed from a plurality of flaw detection signals obtained by transmitting ultrasonic waves from a probe in a predetermined cycle, propagating them as plate wave ultrasonic waves in a direction orthogonal to the moving direction, and receiving each reflected wave. Then, in the method of detecting a defect of a material to be inspected based on the two-dimensional flaw detection image, the two-dimensional flaw detection image is smoothed to obtain a smoothed image, and the obtained smoothed image is preset. It is characterized in that it is binarized based on the threshold value and a binarized image relating to the defect is extracted from the binarized image.

In addition to the first invention, the plate wave ultrasonic flaw detection method according to the second invention further comprises the above-mentioned two methods based on the binarized image.
Extract a defect image from a three-dimensional flaw detection image or a smoothed image,
The feature is that the degree of defect is determined based on the extracted defect image.

In the plate wave ultrasonic flaw detection method according to the third aspect of the invention, the ultrasonic flaw is detected from the ultrasonic probe while the relative movement of the flaw detection material and the ultrasonic probe arranged to face the flaw detection material. Is transmitted in a predetermined cycle, propagates in the direction orthogonal to the moving direction as plate wave ultrasonic waves, forms a two-dimensional flaw detection image from a plurality of flaw detection signals obtained by receiving each reflected wave, and the two-dimensional flaw detection is performed. In the method of detecting defects in the material to be inspected based on the image,
The maximum value processed image was created by replacing the signal level corresponding to each pixel of the two-dimensional flaw detection image with the maximum value of the signal levels of the pixels within a predetermined range including the pixel. It is possible to compress the maximum-value-processed image to obtain a compressed image, binarize the obtained compressed image based on a preset threshold value, and extract a binarized image related to a defect from the binarized image. Characterize.

A plate wave ultrasonic flaw detection method according to a fourth aspect of the present invention is a method for ultrasonically detecting ultrasonic waves from the ultrasonic probe while relatively moving the flaw detection material and the ultrasound probe arranged to face the flaw detection material. Is transmitted in a predetermined cycle, propagates in the direction orthogonal to the moving direction as plate wave ultrasonic waves, forms a two-dimensional flaw detection image from a plurality of flaw detection signals obtained by receiving each reflected wave, and the two-dimensional flaw detection is performed. In the method of detecting defects in the material to be inspected based on the image,
The two-dimensional flaw detection image is smoothed to obtain a smoothed image, and the signal level corresponding to each pixel of the obtained smoothed image is the signal level of a pixel within a predetermined range including the pixel. The maximum value processed image is created by substituting the maximum value of the above, the created maximum value processed image is compressed to obtain the compressed image, and the obtained compressed image is binarized based on a preset threshold value. It is characterized in that a binarized image relating to a defect is extracted from the binarized image.

The plate wave ultrasonic flaw detection method according to a fifth aspect of the present invention is, in addition to the third or fourth aspect of the present invention, further, a defect image is extracted from the compressed image based on the binarized image, and the extracted defect image. It is characterized in that the degree of the defect is judged based on.

A plate wave ultrasonic flaw detector according to a sixth aspect of the present invention is such that the ultrasonic flaw is detected from the ultrasonic probe while relatively moving the material to be detected and the ultrasonic probe arranged to face the material to be detected. Is transmitted in a predetermined cycle, propagates in the direction orthogonal to the moving direction as plate wave ultrasonic waves, forms a two-dimensional flaw detection image from a plurality of flaw detection signals obtained by receiving each reflected wave, and the two-dimensional flaw detection is performed. In the device for detecting defects in the material to be inspected based on the image,
Means for smoothing the two-dimensional flaw detection image to obtain a smoothed image, means for binarizing the obtained smoothed image based on a preset threshold, and a defect from the binarized image And a means for extracting a binarized image according to.

A plate wave ultrasonic flaw detector according to a seventh aspect of the present invention is, in addition to the sixth aspect of the present invention, further the above-mentioned two components based on the binarized image.
It is characterized by comprising means for extracting a defect image from the three-dimensional flaw detection image or the smoothed image, and means for determining the degree of the defect based on the extracted defect image.

A plate wave ultrasonic flaw detector according to an eighth aspect of the present invention is such that the ultrasonic flaw is detected from the ultrasonic probe while relatively moving the material to be detected and the ultrasonic probe arranged to face the material to be detected. Is transmitted in a predetermined cycle, propagates in the direction orthogonal to the moving direction as plate wave ultrasonic waves, forms a two-dimensional flaw detection image from a plurality of flaw detection signals obtained by receiving each reflected wave, and the two-dimensional flaw detection is performed. In the device for detecting defects in the material to be inspected based on the image,
Means for replacing a signal level corresponding to each pixel of the two-dimensional flaw detection image with a maximum value of signal levels of pixels within a predetermined range including the pixel to create a maximum-value-processed image; A means for compressing the created maximum-value-processed image to obtain a compressed image, a means for binarizing the obtained compressed image based on a preset threshold, and a binary value relating to a defect from the binarized image. And a means for extracting the transformed image.

A plate wave ultrasonic flaw detector according to a ninth aspect of the present invention is such that an ultrasonic wave is ultrasonically emitted from the ultrasonic probe while relatively moving the material to be detected and the ultrasonic probe arranged to face the material to be detected. Is transmitted in a predetermined cycle, propagates in the direction orthogonal to the moving direction as plate wave ultrasonic waves, forms a two-dimensional flaw detection image from a plurality of flaw detection signals obtained by receiving each reflected wave, and the two-dimensional flaw detection is performed. In the device for detecting defects in the material to be inspected based on the image,
Means for obtaining a smoothed image by performing a smoothing process on the two-dimensional flaw detection image, and a signal level corresponding to each pixel of the obtained smoothed image, a signal of a pixel within a predetermined range including the pixel Of the levels, the means for creating a maximum-value processed image by replacing the maximum value, a means for compressing the created maximum-value processed image to obtain a compressed image, and the obtained compressed image are preset. It is characterized by comprising means for binarizing based on a threshold value and means for extracting a binarized image relating to a defect from the binarized image.

A plate wave ultrasonic flaw detector according to the tenth invention is
In addition to the eighth or ninth invention, further comprising means for extracting a defect image from the compressed image based on the binarized image, and means for determining the degree of defect based on the extracted defect image. Is characterized by.

[0022]

In the first and sixth aspects of the invention, the ultrasonic probe is moved, for example, while moving the ultrasonic probe disposed near one edge of the sample to be detected or facing the sample to be detected. Ultrasonic waves are transmitted from the probe at a predetermined cycle, propagated as a plate wave ultrasonic wave in the direction orthogonal to the moving direction, each reflected wave is received to obtain a flaw detection signal, and it is obtained at each ultrasonic wave transmission cycle. Further, a plurality of flaw detection signals are arranged on the time axis, and a two-dimensional flaw detection image is formed by gradation of gradation according to the signal level.

With respect to each of the signals corresponding to a plurality of pixels forming the two-dimensional flaw detection image, for example, the signal density of each pixel in a predetermined range including an arbitrary pixel in the two-dimensional flaw detection image, for example, a range of 3 × 3 pixels The level is smoothed and the level of the signal of the pixel is smoothed to obtain a smoothed image. This dilutes the signal level of the pixel forming the image due to the noise, which has only a small area, but the defect image has a larger area than the image due to the noise, so the signal corresponding to the pixel forming the defect image is diluted. The level of does not change.

The obtained smoothed image is binarized based on a preset threshold value. As a result, the image due to the noise diluted by the smoothing process is erased, and the defect image and the image related to the edge of the flaw detection target remain. Then, only the binarized image related to the defect is extracted from the binarized image based on the shape or area of each remaining image.

In the second and seventh inventions, the binarized image extracted as described above is superimposed on the two-dimensional flaw detection image or the smoothed image, and the overlapped portion is used as a defect image to form the two-dimensional flaw detection image. Alternatively, by extracting from the smoothed image, information on the gray level relating to the defect is obtained. Then, the degree of defect is determined by the degree of gray level of the extracted defect image.

In the third, fourth, eighth and ninth inventions, the two-dimensional flaw detection is carried out for each signal corresponding to a plurality of pixels constituting the two-dimensional flaw detection image or the smoothed image obtained as described above. Maximum value conversion processing for replacing the maximum value of the signal levels corresponding to each pixel in a predetermined range including an arbitrary pixel in the image or the smoothed image, for example, a range of 3 × 3 pixels with the signal level of the pixel Go to and create a maximization image. Then, the created maximum-value-processed image is compressed to obtain a compressed image.

In this compression, a signal of at least one pixel in the processing target range in the maximization processing is selected. In the case described above, the signal of the pixel at which the remaining matrix intersects is selected by thinning out the maximum of 2 rows for every 3 rows and the maximum of 2 rows for every 3 columns.
Then, a compressed image is obtained by forming an image composed of a plurality of selected signals. As a result, the amount of image processing is reduced without reducing the level of the signal corresponding to the pixels forming the smoothed image or the two-dimensional flaw detection image.

By binarizing the obtained compressed image based on a preset threshold value, a noise-related image is eliminated in the compressed image based on the smoothed image.
In the compressed image based on the three-dimensional flaw detection image, the image of noise having an intensity lower than the threshold is erased, and the processing amount of the subsequent image processing is reduced. And from the binarized image,
A binary image related to the defect is extracted based on the shape or area of each image remaining in the image.

In the fifth and tenth aspects of the invention, the binary image extracted as described above is superimposed on the compressed image, and the defective image is extracted from the compressed image by using the overlapping portion as a defect image. Obtaining information on the gray level relating to the defect.
Then, the degree of defect is determined by the degree of gray level of the extracted defect image.

[0030]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be specifically described below with reference to the drawings showing the embodiments. 1 is a block diagram showing a configuration of a plate wave ultrasonic flaw detector according to the present invention, and FIG. 2 is a block diagram showing functions of a central processing unit (CPU) 9 shown in FIG. The function of each block shown in FIG. 2 is realized by software or hardware. Tire probe 11
Is rolled on one edge portion E of the strip-shaped flaw detection material S conveyed in the arrow direction. A voltage is applied from the pulser 2 to the tire probe 11, and the pulser 2 controls the cycle of applying the voltage by the pulse signal from the pulse timing controller 3. Then, the tire probe 11 is excited by the applied voltage, transmits an ultrasonic wave to the other edge portion E of the material S to be detected, and receives a reflected wave thereof.

The signal received by the tire probe 11 for flaw detection is amplified by the signal amplifier 4 and then input to the bandpass filter 5 having a predetermined pass frequency band to remove noise components. The signal passed through the bandpass filter 5 is A /
The digital signal is converted into an 8-bit digital signal having 0 to 255 gradations by the D converter 6 and given to the frame memory 7.
A pulse signal is also given to the frame memory 7 from the pulse timing controller 3 described above, and based on the pulse signal, the flaw detection signal given from the A / D converter 6 is two-dimensionalized to be two-dimensional. It is stored in the frame memory 7 as a flaw detection image.

FIG. 3 is a waveform diagram of the signal supplied to the frame memory 7, and FIG. 4 is a three-dimensional waveform diagram of the flaw detection signal stored in the frame memory 7. In the frame memory 7, as shown in FIG. 3A, a pulse signal from a pulse timing controller and, as shown in FIG. 3B, a plate wave ultrasonic wave propagated in the width direction of the flaw detection material at the timing of each pulse signal. The detected flaw detection signal is given. When this flaw detection signal is divided into pulse signals, flaw detection signals for each width direction of the flaw detection target material are obtained in the order of the conveyance direction of the flaw detection target material. Then, each flaw detection signal is placed on a coordinate axis in which the x axis is the plate width direction of the flaw detection material, the y axis is the conveyance direction of the flaw detection material, and the z axis is the signal intensity.
FIG. 4 shows a case in which the flaw detection materials are arranged in the conveying direction. In both figures, R _E is a signal due to diffused reflection in the vicinity of the tire probe, K _E is a signal due to a defect, P _E is a signal due to noise, and T _E is a signal due to the other edge portion of the material to be detected. Is.

FIG. 5 is an explanatory diagram for explaining the two-dimensional flaw detection image stored in the frame memory 7. In FIG. 5, the vertical axis represents the conveyance direction of the flaw detection material, and the horizontal axis represents the width direction of the flaw detection material. As shown in FIG. 5, an image R _P due to diffused reflection near the tire probe and an image T _P due to the other edge portion of the flaw detection material are vertically formed in strips at both ends of the two-dimensional flaw detection image. A plurality of island-shaped defect images K _P and noise images P _P are formed between the two images. These images R _P , T _P , K
_P and the image P _P are represented by light and shade that becomes darker as the signal level becomes higher.

The smoothing processing unit 91 of the CPU 9 smoothes the two-dimensional flaw detection image stored in the frame memory 7 as follows to obtain a smoothed image, and the smoothing image is processed by the binarization processing unit 92, for example. Binarize 0 or 255 using 128 as the threshold
Use a digitized image.

6A and 6B are explanatory views for explaining the smoothing processing by the smoothing processing unit 91. FIG. 6A shows a partial image of a two-dimensional flaw detection image, and FIG. 6B shows an image after the smoothing processing. Shows. The smoothing processing unit 91 selects any one pixel among the plurality of pixels P ₁₁ , P ₁₂ , ... Forming the two-dimensional flaw detection image, as shown in FIG.
In FIG. 6A, attention is focused on the pixel P ₂₂ (e gradation), and the smoothing target pixel region of m × n pixels including the focused pixel P ₂₂ is shown in FIG.
In (a), a 3 × 3 pixel region of pixels P _{11 to} P ₃₃ is set.

The set pixels P _{11 to} P ₃₃ have gradations a to i, respectively, as shown in FIG. 6A, and these pixels P _{11 to} P ₃₃ are based on the following equation (1). A smoothing calculation is performed, and the calculation result t is set as the gradation of the pixel P ₂₂ of interest as shown in FIG. 6B. _{t = (a × w a +} b × w b + c × w c + d × w d + e × w e + f × w f + g × w g + h × w h + i × w i) / (w a + w b + w c + w d _{+ w e + w f + w} g + w h + w i) ... (1) However, w _a ~w _i: the weight of the integer

Then, the smoothing processing section 91 performs such smoothing processing on all the pixels of the two-dimensional flaw detection image to obtain a smoothed image. As a result, the pixels forming the defect image P _P in the two-dimensional flaw detection image are subjected to the smoothing process together with the surrounding pixels of 0 gradation, so that the gradation of the defect image P _P is diluted and the smoothing process is continued. By the binarization process, 0 gradation is obtained.

FIG. 7 is an explanatory diagram for explaining a binarized image.
You. This binary image is the same as the two-dimensional flaw detection image of FIG.
After smoothing as described above, 0 is set with 128 gradations as a threshold.
It is binarized into (light) or 255 gradations (dark).
As is clear from FIG. 7, a vertically long strip image R_PAnd image T
_PA plurality of island-shaped defect images K between_PAre all displayed in 255 gradations
And the noise image P that was present in the two-dimensional flaw detection image
_P(See Figure 5) have been removed. Like this
It is possible to determine what the multiple images in the original flaw detection image are.
Somehow, the noise image P_POnly can be easily removed
Therefore, the noise image P _PTakes a short time to remove.

The binarized defect detection unit 93 of the CPU 9 erases an image in the binarized image, for example, one whose vertical length is longer than a preset length, as a non-defective image, so that the image shown in FIG. As shown in, only the defect images K _P and K _P are extracted to obtain a binarized defect extraction image, which is detected by the defect detection unit.
Give to 94. The defect detection unit 94 includes the frame memories 7 to 2
A two-dimensional flaw detection image is also given, and the defect detection unit 94 superimposes the binarized defect extraction image on the two-dimensional flaw detection image shown in FIG. 5 and extracts the overlapped portion from the two-dimensional flaw detection image to obtain the image. A defect extraction image as shown in 9 is obtained. Then, the defect determination unit 95 of the CPU 9 finds the maximum gradation for each of the defect images K _P and K _{P of the} defect extraction image, and if the found gradation exceeds, for example, 200 gradations which is an evaluation threshold, it is a serious defect. If it is less than that, a two-stage evaluation that the defect is a minor defect is performed, and the output unit 96 reports the result to the CRT.
Alternatively, the output device 10 such as a printer outputs the alarm and an alarm device (not shown) is activated. In this evaluation method, two or four evaluation thresholds are set according to the type of the material to be inspected, and evaluation is performed in three or five steps.

FIG. 10 is a block diagram showing another function of the CPU 9 shown in FIG. 1, which corresponds to speeding up of image processing. The 2D flaw detection image of the frame memory 7 is the CPU
The smoothing processing unit 91 of 9 performs smoothing processing as before, and the smoothed image is given to the maximum value processing unit 97. On the other hand, the maximum value processing unit 97 is adapted to be provided with a two-dimensional flaw detection image from the frame memory 7, and the maximum value processing unit 97 uses the smoothed image or the two-dimensional flaw detection image to determine the level of any pixel. Maximum value processing is performed to convert the key to the maximum gradation within a predetermined pixel range.

FIG. 11 is an explanatory view for explaining the maximum value processing. In FIG. 11, (a) shows a partial image of the smoothed image or the two-dimensional flaw detection image, and (b) shows the image after the maximum value processing. Shows. The maximum value processing unit 97 pays attention to an arbitrary pixel among the plurality of pixels P ₁₁ , P ₁₂ , ... Forming the smoothed image, that is, the pixel P ₂₂ (n gradation) in FIG. 11A. A maximum value processing target area of m × n pixels including the pixel P _22, that is, a 3 × 3 pixel area of pixels P _{11 to} P _{33 in} FIG. 11A is set.

The set pixels P _{11 to} P ₃₃ have gradations j to r, respectively, as shown in FIG. 11A, and these pixels P _{11 to} P ₃₃ are based on the following equation (2). The maximum value u is obtained, and the gradation of the pixel P ₂₂ of interest is replaced with u as shown in FIG. u = max (j, k, l, m, n, o, p, q, r) (2)

Then, the maximum value processing unit 97 performs the maximum value processing on all the pixels forming the smoothed image or the two-dimensional flaw detection image in the same manner as before, and the maximum value processing as shown in FIG. Obtain the processed image. The compression processing unit 98 performs the following compression processing on the maximum value processed image to obtain the compressed image as shown in FIG.

12A and 12B are explanatory views for explaining the image compression processing by the compression processing unit 98, in which FIG. 12A shows the maximum value processed image and FIG. 12B shows the compressed image. FIG.
As indicated by a cross mark in (a), a plurality of rows and columns of pixels arranged in a matrix and forming a maximum-value processed image are set to be thinned out at predetermined intervals. This setting is defined according to the maximum value processing target area set in the maximum value processing, and the maximum value processing target area is 3 × 3.
In the case of the pixel region, as shown in FIG. 12A, up to 2 rows or 2 columns can be thinned out of 3 rows or 3 columns adjacent to each other. Then, a compressed image as shown in FIG. 12 (b) is obtained by forming an image composed of the gradation of each pixel in the portion where the rows and columns intersect with each other shown by the circles in FIG. 12 (a). The compressed image shown in FIG. 12 is compressed to 1/9 of the original image.

The binarization processing unit 92 of the CPU 9 converts the above-mentioned compressed image into 0 or 255 by using, for example, 128 as a threshold value.
It is binarized to form a binarized image as shown in FIG. C
The binarized defect detection unit 93 of the PU 9 erases, of the images in this binarized image, for example, ones whose vertical length is longer than a preset length as non-defect-related images, and presets the width. By deleting those smaller than the predetermined size as not the image related to the defect, as shown in FIG. 16, only the defect images K _P and K _P are extracted to obtain the binarized defect extraction image, and the defect detection is performed. Give to part 94. The defect detection unit 94 is also given a compressed image as shown in FIG. 14 from the compression processing unit 98,
The defect detection unit 94 obtains the defect extraction image as shown in FIG. 17 by superimposing the binarized defect extraction image on the compression image and extracting the overlapped portion from the compression image. And C
The defect determination unit 95 of the PU 9 detects each defect image K of this defect extraction image.
The degree of defect is determined for each of _P and K _P as before, and the output unit 96 causes the output device 10 such as a CRT or a printer to output the determination result and an alarm device (not shown).
Activate

[0046]

As described above in detail, in the first and sixth inventions, it is possible to judge whether the image is due to noise or a defect by performing smoothing processing and binarization processing on the image. Without this, it is possible to easily erase the image related to noise from the image. Therefore, a high-performance judgment device is not required and the device cost is low.

In the second and seventh inventions, only the grayscale signal of the image relating to the defect is reliably extracted, and erroneous determination is prevented.

In the third, fourth, eighth and ninth inventions, by forming a compressed image, the processing amount of the subsequent image processing is reduced and the processing speed is improved. Therefore, it is possible to improve the flaw detection speed of the defect, and it is also possible to perform the defect flaw detection by the image processing even on the wide flaw-detected material.

In the fifth and tenth aspects of the invention, only the grayscale signal of the image relating to the defect is reliably and rapidly extracted,
The present invention has an excellent effect such that a defect can be accurately determined in a short time.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of a plate wave ultrasonic flaw detector according to the present invention.

FIG. 2 is a block diagram showing functions of the central processing unit shown in FIG.

FIG. 3 is a waveform diagram of a signal applied to a frame memory.

FIG. 4 is a three-dimensional waveform diagram of a flaw detection signal stored in a frame memory.

FIG. 5 is an explanatory diagram illustrating a two-dimensional flaw detection image stored in a frame memory.

FIG. 6 is an explanatory diagram illustrating smoothing processing by a smoothing processing unit.

FIG. 7 is an explanatory diagram illustrating a binarized image.

FIG. 8 is an explanatory diagram illustrating a binarized defect extraction image.

FIG. 9 is an explanatory diagram illustrating a defect image.

10 is a block diagram showing another function of the CPU shown in FIG.

FIG. 11 is an explanatory diagram illustrating maximum value processing.

FIG. 12 is an explanatory diagram illustrating image compression processing by a compression processing unit.

FIG. 13 is an explanatory diagram illustrating a maximum value processed image.

FIG. 14 is an explanatory diagram illustrating a compressed image.

FIG. 15 is an explanatory diagram illustrating a binarized image.

FIG. 16 is an explanatory diagram illustrating a binarized defect extraction image.

FIG. 17 is an explanatory diagram illustrating a defect extraction image.

FIG. 18 is a schematic cross-sectional view showing a usage mode of the tire probe.

FIG. 19 is a graph showing an example of a flaw detection signal by plate wave ultrasonic flaw detection.

[Explanation of symbols]

 7 frame memory 9 central processing unit 11 tire probe 91 smoothing processing unit 92 binarization processing unit 93 binarization defect detection unit 94 defect detection unit 95 defect determination unit 97 maximum value processing unit 98 compression processing unit S flaw detection Material E Edge part

Claims (10)

[Claims] 1. An ultrasonic wave is transmitted from the ultrasonic probe at a predetermined cycle while relatively moving the material to be detected and the ultrasonic probe arranged to face the material to be detected, and the ultrasonic wave is transmitted as a plate wave. A two-dimensional flaw detection image is formed from a plurality of flaw detection signals obtained by receiving each reflected wave as an ultrasonic wave propagating in a direction orthogonal to the moving direction, and flaws in the flaw detection material are flaw-detected based on the two-dimensional flaw detection image. In the method, the two-dimensional flaw detection image is smoothed to obtain a smoothed image, the obtained smoothed image is binarized based on a preset threshold value, and the binarized image is converted into a defect. A plate wave ultrasonic flaw detection method characterized by extracting the binarized image. 2. The defect image is extracted from the two-dimensional flaw detection image or the smoothed image based on the binarized image, and the degree of defect is determined based on the extracted defect image.
The described plate wave ultrasonic flaw detection method. 3. The ultrasonic wave is transmitted from the ultrasonic probe at a predetermined cycle while relatively moving the material to be detected and the ultrasonic probe arranged to face the material to be detected, and the ultrasonic wave is transmitted as a plate wave. A two-dimensional flaw detection image is formed from a plurality of flaw detection signals obtained by receiving each reflected wave as an ultrasonic wave propagating in a direction orthogonal to the moving direction, and flaws in the flaw detection material are flaw-detected based on the two-dimensional flaw detection image. In the method described above, the level of the signal corresponding to each pixel of the two-dimensional flaw detection image is replaced with the maximum value of the signal levels of the pixels within a predetermined range including the pixel to create a maximum-value-processed image. Then, the created maximum-value-processed image is compressed to obtain a compressed image, the obtained compressed image is binarized based on a preset threshold value, and the binarized image related to the defect is generated from the binarized image. A plate wave ultrasonic flaw detection method characterized by extraction. 4. The ultrasonic wave is transmitted from the ultrasonic probe at a predetermined cycle while relatively moving the material to be detected and the ultrasonic probe arranged to face the material to be detected, and the ultrasonic wave is transmitted as a plate wave. A two-dimensional flaw detection image is formed from a plurality of flaw detection signals obtained by receiving each reflected wave as an ultrasonic wave propagating in a direction orthogonal to the moving direction, and flaws in the flaw detection material are flaw-detected based on the two-dimensional flaw detection image. In the method described above, the two-dimensional flaw detection image is smoothed to obtain a smoothed image, and the level of the signal corresponding to each pixel of the obtained smoothed image is set to a pixel within a predetermined range including the pixel. The maximum value processed image is created by substituting the maximum value of the signal levels of the above, the created maximum value processed image is compressed to obtain a compressed image, and the obtained compressed image is set to a preset threshold value. Binary image based on the defect from the image binarized based on Lamb wave ultrasonic testing method and extracts. 5. The plate wave ultrasonic flaw detection according to claim 3, further comprising extracting a defect image from the compressed image based on the binarized image and determining a degree of the defect based on the extracted defect image. Method. 6. The ultrasonic wave is transmitted from the ultrasonic probe in a predetermined cycle while relatively moving the material to be detected and the ultrasonic probe arranged to face the material to be detected, and the ultrasonic wave is transmitted as a plate wave. A two-dimensional flaw detection image is formed from a plurality of flaw detection signals obtained by receiving each reflected wave as an ultrasonic wave propagating in a direction orthogonal to the moving direction, and flaws in the flaw detection material are flaw-detected based on the two-dimensional flaw detection image. In the device, a means for smoothing the two-dimensional flaw detection image to obtain a smoothed image, a means for binarizing the obtained smoothed image based on a preset threshold value, and a binarization And a means for extracting a binarized image related to a defect from the obtained image. 7. A means for extracting a defect image from the two-dimensional flaw detection image or a smoothed image based on the binarized image, and a means for determining the degree of defect based on the extracted defect image. The plate wave ultrasonic flaw detector according to claim 6. 8. The ultrasonic wave is transmitted from the ultrasonic probe at a predetermined cycle while relatively moving the material to be detected and the ultrasonic probe arranged to face the material to be detected, and the ultrasonic wave is transmitted as a plate wave. A two-dimensional flaw detection image is formed from a plurality of flaw detection signals obtained by receiving each reflected wave as an ultrasonic wave propagating in a direction orthogonal to the moving direction, and flaws in the flaw detection material are flaw-detected based on the two-dimensional flaw detection image. In the apparatus described above, the level of the signal corresponding to each pixel of the two-dimensional flaw detection image is replaced with the maximum value of the signal levels of pixels within a predetermined range including the pixel to create a maximum-value-processed image. Means, a means for compressing the created maximum-value-processed image to obtain a compressed image, a means for binarizing the obtained compressed image based on a preset threshold value, and a defect from the binarized image. And a means for extracting a binary image according to To plate wave ultrasonic flaw detector. 9. The ultrasonic wave is transmitted from the ultrasonic probe at a predetermined cycle while relatively moving the material to be detected and the ultrasonic probe arranged to face the material to be detected, and the ultrasonic wave is transmitted as a plate wave. A two-dimensional flaw detection image is formed from a plurality of flaw detection signals obtained by receiving each reflected wave as an ultrasonic wave propagating in a direction orthogonal to the moving direction, and flaws in the flaw detection material are flaw-detected based on the two-dimensional flaw detection image. In the device, a means for smoothing the two-dimensional flaw detection image to obtain a smoothed image and a level of a signal corresponding to each pixel of the obtained smoothed image within a predetermined range including the pixel. Means for creating a maximum-value-processed image by substituting the maximum value of the pixel signal levels, a means for compressing the created maximum-value-processed image to obtain a compressed image, and the obtained compressed image. Means for binarizing based on a preset threshold and binary Lamb wave ultrasonic testing apparatus, characterized in that it comprises means for extracting a binary image according the image to the defect. 10. The method according to claim 8, further comprising means for extracting a defect image from the compressed image based on the binarized image, and means for determining a degree of defect based on the extracted defect image.
Alternatively, the plate wave ultrasonic flaw detector according to item 9.