JPS62229053A - Defect inspection of pellet end face for nuclear fuel - Google Patents

Abstract

PURPOSE:To enable accurate inspection regardless of variations in inspecting conditions, by sampling image data at a dish section to set a threshold based on the data. CONSTITUTION:A plurality of data at a dish section 3 are sampled from among multilevel image data generated and the multilevel image data are binary coded at points on an end face 2 by a threshold Dt set based on the data sampled to generate a binary coded image 6. In the binary-coded image 6, the position O of the center point of the disk section 3, namely, the center point of a pellet 1 is checked to measure the number of data existing in an area 11 containing the center point O. As the area 11 containing the center point O of the dish section 3 corresponds to the normal part naturally, the number of data in the area 11 is regarded as the number of data at the normal part and hence, the area 12 other than the area 11 can be a defective part. Thus, the number of data in the area 11 is collated with a criterion reference value preset to decide on the quality, the acceptance or reject, of the pellet 1.

Description

[Detailed description of the invention] "Industrial application field"-/7'l D Il[] I-) u:I
tlkt! :L Fn R+7. ,,k/
7'l 6 history Fi7iM /F K? A: It concerns the method of existence.

``Conventional technology'' Nuclear fuel pellets are shown in Figure 3! As shown in , it is formed into a cylindrical shape, and a concave dish portion 3 is formed in the center of the end surface 2 .

During the manufacturing process of this nuclear fuel pellet 1, chips 5 as shown in FIG. 3 are likely to occur from the periphery of the end face 2 to the circumferential surface 4, so it is necessary to inspect the presence or absence of these chips 5 and their size. is important.

Conventionally, inspection for this purpose has generally been carried out visually, but may also be carried out by fj image processing. In that case, as shown in Fig. 4, a photographic light is applied to the pellet end face 2 from the vertical direction and an image is taken with a two-dimensional camera (not shown), and the brightness of the reflected light from each point on the end face 2 is measured. For example, when dividing into 256 stages, masking processing is performed to create multivalued image data for each point on the end face 2. Next, these multivalued image data are binarized by a predetermined threshold (16) and converted into two values as shown in FIG.
A digitized image 6 is created. Then, the size of the defective area 7 or normal area 8 appearing in this binarized image 6, that is, the number of pixels included in each area, is measured, and the measurement result is compared against a preset judgment reference value. I try to judge pass/fail by saying something.

"Problems to be Solved by the Invention" By the way, in the above conventional method, especially when a large number of pellets are to be inspected continuously over a long period of time, the illuminance of the photographing light and the sensitivity of the camera may vary over time. Therefore, the average level of brightness of the captured multilevel image may vary for each pellet, regardless of the existence or size of the chip 5. there were.

However, in the past, even when such inspection conditions changed, binarization was performed uniformly using a fixed threshold value, which made it impossible to perform accurate binarization. Ta.

Therefore, there was a problem in that the accuracy and reliability of the test were not sufficient, and it was not possible to accurately determine whether the pellets I were acceptable or not.

The present invention was made in view of the above circumstances, and an object of the present invention is to provide an inspection method that allows accurate inspection even when inspection conditions vary.

"Means for Solving the Problems" This invention images the end face of a nuclear fuel pellet in which a concave dish portion is formed in the center, creates multivalued image data of each point on the end face, and The multivalued image data is binarized using an appropriate threshold value to create a binarized image, and based on the binarized image, the growth of defects such as chips that occur on the periphery of the end face of the nuclear fuel pellet is determined. In the method of inspecting the absence or size, when the multi-valued image data of the surface is binarized, the multi-valued image data of the dish portion of the end face is sampled, and the threshold value is set based on the sampled data. It is characterized by setting a value.

"Operation" This invention prevents chipping from occurring up to the center of the four-concave dish part formed at the center of the end face of the nuclear fuel pellet, and the image data in this part has the average brightness of the normal part of the surface to be inspected. Focusing on the fact that the image data always includes the level , image data is sampled from this dish portion, a threshold value is set based on the data, and binarization is performed using the threshold value.

"Example" Hereinafter, an example of the present invention will be described with reference to FIGS. 1 and 2.

First, as in the conventional method, an image of the end surface 2 of the pellet I to be inspected is captured using a two-dimensional camera (not shown), and multivalued image data of each point on the end surface 2 is created.

Then, a plurality of pieces of data from the dish unit 3 are sampled from among the multivalued image data created above, and a threshold value is set based on these sampled data.

Setting of this threshold value is performed as follows.

First, as shown in FIG. 115 to I/1 of diameter
Areas AI, At, etc. of approximately 0 are set at approximately 5 to IO locations (5 locations, AI-A in FIG. 1).

Then, calculate the average value of each data in each area A,,A,... and use it as the representative value D of each area.
,,D,... Tozuro. These representative values D,,D
, . . . are the average values of data in the dish portion 3 where no chipping usually occurs, and therefore each represents the average level of brightness of the normal portion of this end surface 2. Then, the maximum one-direction D max and the minimum value D min are determined from among those representatives (riD, D2...), the difference between them (D max - D m1n) is calculated, and the value of the difference is set to a predetermined value. Coefficient K (for example = O, 1~0.3
) and further subtract that value from the minimum value D min, and the value obtained is set as the threshold value Dt. That is, D L = D min -K ・(D max - D
m1n) K=0.1 to 0.3.

Next, the multivalued image data at each point of the end face 2 is binarized using the threshold value Dt set in two steps to create a binarized image 6 as shown in FIG.

Then, in this binarized image 6, confirm the center point of the dish portion 3, that is, the position 0 of the center point of the pellet l,
Only the number of data existing in the region II that includes the center point O is counted. In this measurement, the outline of this area is first confirmed in the X-Y coordinates, and then, as shown in FIG. 2, the number of data in the area is sequentially counted while tracing the outline.

The region 11 including the center point O of the dish portion 3 is naturally a normal region since no defect such as chipping occurs in the dish portion 3 as described above. Therefore, the number of data in this region 11 is Find out the number of data in the normal part. The region 12 other than the region 11 is defined as a defective portion.

Then, by comparing the number of data in the area 11 (the number of data in the normal part) obtained in ``1'' with the preset judgment criterion 1, it is judged whether or not the pellet 1 is acceptable. In general, if the number of data in this area 11 is greater than the criterion bet value, it is determined to be a good product, and if it is smaller, it is determined to be defective.

This completes the inspection of one pellet, and then the next pellet is inspected in the same manner.

According to the above procedure, the value of the threshold DL is set each time based on the data of the dish portion 3 where defects such as chips do not occur, so the average value of the normal portion of the end surface 2 to be inspected is Accurate binarization corresponding to brightness can be performed.

Therefore, if the test conditions change, each pellet
-1 can be tested under almost the same conditions, the accuracy and reliability of the test can be improved, and accurate judgments can be made.

In addition, in the above, the area including the center point O of the pellet 1 is considered to be a normal area and the measurement is omitted, so if the binarized image 6 is divided into many areas, There is no need to judge whether it is a normal part, and the test can be performed extremely quickly. Furthermore, even if there is a pseudo-normal part (a part that appears in the image with the same brightness as the normal part despite being a defective part) in the defective part, the pseudo-normal part can be considered as the original normal part. There is no risk of misidentification and measurement, further improving inspection accuracy.

Although the present invention has been described in detail above, the setting of the threshold value from the data of the dish portion is not limited to the above, and may be changed as appropriate in consideration of the range of expected fluctuations in inspection conditions. For example, the number of areas to be sampled and the values of the coefficients shown above may be changed, and furthermore, the calculation formula for ponds may be used instead of the calculation formula shown above.

In addition, in the above, the area including the center point of the pellet is considered to be a normal area and measured, but in this invention, this is not necessarily the case, and the size of the defective area or normal area can be measured in the same manner as before. So good.

"Effects of the Invention" As explained in detail above, according to the present invention, the image data of the dish portion is sampled and the threshold value is set based on that data, so that the inspection conditions can be adjusted. Even if the threshold value changes, the accuracy and reliability of the test will be improved by determining the optimal threshold value corresponding to the test condition.
This has the effect that pass/fail judgments can be made accurately.

[Brief explanation of drawings]

1 and 2 are diagrams for explaining an inspection method according to an embodiment of the present invention, in which FIG. 1 is a diagram for explaining a method of setting a threshold value, and FIG. 2 is a diagram for explaining a method for setting a threshold value, and FIG. FIG. 3 is a diagram illustrating a state in which an area including FIG. 3 is a perspective view showing the shape of a nuclear fuel pellet, FIG. 4 is a sectional view showing an image of the end face of the nuclear fuel pellet, and FIG. 5 is a diagram showing a binarized image of the end face. 1. Nuclear fuel pellet, 2. End face, 3. Dish portion, 5. Chip (defect portion), 6. Binarized image. 1 applicant Mitsubishi Kinuta Mashishiki-go

Claims (1)

[Claims] The end face of a nuclear fuel pellet with a concave dish formed in the center is imaged, multi-valued image data is created for each point on the end face, and the multi-valued image data is divided into two values using an appropriate threshold. In the method for inspecting the presence or absence of a defect such as a chip on the peripheral edge of the end face of the nuclear fuel pellet and its size based on the binarized image, the multi-valued image is Defect inspection of a nuclear fuel pellet end face, characterized in that when data is binarized, multivalued image data of the dish portion of the end face is sampled, and the threshold value is set based on the sampled data. Method.