JPS62229054A - Defect inspection of pellet end face for nuclear fuel - Google Patents

Abstract

PURPOSE:To enable accurate decision eliminating the effect of any false normal part, by measuring the number of pixels in an area containing an arbitrary point in a dish section to judge on the quality, acceptance or reject of a pellet, based on the results of the measurement. CONSTITUTION:A plurality of data at a disk section 3 are sampled from among multilevel image data generated and multilevel image data at points on an end face 2 are binary coded by a threshold 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 pixels alone existing in an area 12 containing the center point O thereof. As the area 12 containing the center point O of the dish section 3 corresponds to a normal part naturally, the number of pixels in the area 12 is regarded as the number of pixels at the normal part and hence, the area 13 other than the area 12 can be a defective part. Thus, the number of pixels in the area 12 is collated with a criterion reference value preset to decide on the quality, acceptance or reject, of the pellet 1.

Description

DETAILED DESCRIPTION OF THE INVENTION "Field of Industrial Application" The present invention relates to a method for inspecting defects on the end face of a nuclear fuel pellet.

r<, 'fiIIF□tlz The Todoshi electric nuclear fuel pellet is formed into a cylindrical shape as shown by the reference numeral 1 in FIG.
is formed.

During the manufacturing process of this J: eel nuclear fuel pellet 1-1, eel chips 5 are likely to occur from the peripheral edge of the end face 2 to the peripheral surface 4 (not shown in Fig. 3). It is important to check the

Conventionally, inspections for this purpose have generally been performed visually, but there are also cases where they can be performed using image processing. In the case of image processing, 1. l:1'' As shown in Figure 4 (a), a photographing light is applied to the end face 2 of the pellet 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. If it is divided into, for example, 256 levels, massing processing is performed to create multi-valued image data for each point on the end face 2.Next, these multi-valued image data are divided into two values using a predetermined threshold value. to create a binarized image 6 as shown in Figure 4 (b). I+ C+ /J”l
W4'>J+4/i' IM, < N (7, and the normal part is expressed as normal part 8.

Then, from this binarized image 6, the number of pixels representing the defective part 7 or the normal part 8 is measured, and by comparing the measurement result with a determination reference value, it is determined whether the pellet 1 is acceptable or not.

``Problem to be Solved by the Invention J'' However, with the above conventional method, accurate determination may not be possible depending on the shape of the chip 5.

For example, as shown in Figure 5 (a), if there is a part of the chip 5 that is almost parallel to the end face 2, diffuse reflection will not occur in that part and the reflection will be regular like the end face 2. There are times when I get confused. In this case, in the binarized image 6, that part has the same brightness as the normal part 8, and as shown in FIG. , the part that appears as a normal part in the image)
They appear in the form of islands in the defective portion 7.

In such cases, it has not been possible to determine which region is the normal region in the past, and it has not been possible to distinguish between the pseudo-normal region 9 and the original normal region 8 as described in 1. The pseudo-normal part 9 was also measured as the normal part 8, and as a result, the pellet 1, which was a defective product, was sometimes incorrectly determined to be a good product.

For this reason, in the past, in order to avoid such misjudgments, it was necessary to visually check the inspection penetration four times.
There was a problem in that it was difficult to automate testing and improve efficiency.

The present invention has been made in view of the above circumstances, and it is an object of the present invention to provide an inspection method that can exclude the influence of pseudo-normal parts and make accurate determinations.

[Means for Solving the Problems] This invention creates a binarized image by imaging the end face of a nuclear fuel pellet in which a concave dish portion is formed in the center, and then creates a binarized image based on the binarized image. When determining pass/fail by inspecting the growth and size of defective parts such as chips occurring on the peripheral edge of the nuclear fuel pellet end face, among the plurality of areas appearing in the binarized image, the center of the dish part is inspected. It is characterized by measuring the number of pixels in an area including an arbitrary point, and making a determination based on the measurement result.

"Operation" This invention focuses on the fact that chipping does not normally occur at the center of the concave dish formed at the center of the end face of nuclear fuel pellets, and the image data in this part is always a normal part. This is what I did. Then, the number of pixels in a region including an arbitrary point in the dish portion is measured and used as the number of pixels in the normal portion, and based on the measurement result, I'll determination is made as to whether the pellet is acceptable or not.

``Example'' The inspection method of the example of this invention will be described below with reference to FIGS. 1 and 2.
This will be explained with reference to the figures.

First, as in the conventional method, the end surface 2 of the pellet 1 to be inspected is imaged using a two-dimensional camera (not shown), and multivalued image data of each point on the end surface 2 is created.

Then, the denomination 1 of the multivalued image data created above is created. ．． A threshold is set based on the sampled data by performing the data tree kf of the first part 3-i'', 1 contemplation → J-nobling. As shown in FIG. 2, a plurality of (A in FIG.
, ~A5), calculate the average value of each data in those areas, and set based on those average values.

Next, according to the threshold value set above, the end face 2
The multivalued image data of each point is binarized to create a binarized image 1 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 I,
Only the number of pixels existing in the area 12 including the center point O is measured. In this measurement, the outline of this area is first confirmed in the X-Y coordinates, and then, as shown in FIG. 1, the number of data in the area is sequentially counted while tracing the outline.

The region 12 including the center point O of the dish portion 3 is naturally a normal region because no defects such as chips occur in the dish portion 3 as shown in the double series. The number of pixels is taken as the number of pixels in the normal part. And areas other than area I2! 3 is a defective part.

Then, by comparing the number of pixels in the region 12 obtained above (the number of pixels in the normal part) with a preset criterion value, it is determined whether the pellet I is acceptable or not. That is, if the number of pixels in this area 12 is larger than the determination reference value, it is determined to be a good product, and if it is smaller, it is determined to be a defective product.

According to the above procedure, the region 12 including the center point O of the pellet I is measured as a normal region, so even if the binarized image 6 is divided into many regions, it is difficult to determine which region. There is no need to judge whether it is a normal part, measurements can be taken immediately, and inspections can be performed extremely quickly.

Furthermore, even if there is a pseudo-normal area 9 as shown in FIG. Improve your skills and be able to accurately judge pass/fail. Therefore, there is no need for a check by an inspector, which was necessary in the past, and it is possible to automate and speed up the inspection.

Furthermore, in the above embodiment, the threshold value is set to
J1 to the data in the dish section 3 where defects such as J do not occur.
Since the settings are made based on the model, accurate binarization corresponding to the average brightness of the normal portion of the end face 2 to be inspected can be performed. Therefore, even if the inspection conditions change due to a change in the illuminance of the photographing light or the sensitivity of the camera, each pellet I can be inspected under almost the same conditions, improving inspection accuracy and reliability. It is possible to make accurate judgments.

Although the present invention has been described in detail above, the present invention is not limited to the above. For example, in the above embodiment, the area to be measured is the area including the center point 0 of the dish part 3, but if you select an arbitrary point in the dish part 3 and measure the area including that point, the same result will be obtained. Of course there is. In addition, when measuring the number of pixels in the above area 12,
For example, a technique such as a labeling method may be used. moreover,
When determining pass/fail, the above area 12 is selected from the total number of pixels.
Since the value obtained by subtracting the number of fj hydrogens in the middle becomes the number of pixels in the defective part,
The determination may be made based on the value.

In the above embodiment, the threshold value is set from the data of the dish section when performing 21 conversion, but in the present invention, this is not necessarily the case, and binarization is performed using a fixed threshold value. But it's okay.

"Effects of the Invention" As explained in detail above, according to the present invention, the number of pixels in an area including an arbitrary point in the dish part is measured, and a pass/fail judgment is made based on the measurement result. Therefore, if there are many areas in a binarized image, there is no need to judge which area is a normal area, and measurement can be performed immediately. Prevent it from being measured as a normal part. Therefore, although the image processing algorithm is simple, it is possible to quickly and accurately determine whether a pellet is acceptable or not, and highly reliable inspection can be performed efficiently.

[Brief explanation of drawings]

1 and 2 are rounded diagrams explaining one embodiment of the present invention, in which FIG. 1 shows a state in which an area including the center point is being measured, and FIG. 2 shows a threshold value. FIG. 3 is a diagram showing a state in which settings are being made. Figure 3 is a perspective view showing the shape of a nuclear fuel pellet, Figure 4 (a) is a sectional view showing the imaging state of the end face of the nuclear fuel pellet, and Figure 4 (b) is a diagram showing a binarized image. , FIG. 5(a) is a diagram showing the imaging state when a pseudo-normal area appears;
FIG. 5(b) is a diagram showing a binarized image in which pseudo-normal parts appear. 1. Nuclear fuel pellet, 2. End face, 3. Dish portion, 5. Chip (defect portion), 6. Binarized image, 12. Area including any point in the dish portion.

Claims (1)

[Claims] An image of the end face of the nuclear fuel pellet with a concave dish formed in the center is imaged to create a binarized image, and based on the binarized image, defects such as chips occurring on the peripheral edge of the end face of the nuclear fuel pellet are detected. When inspecting the presence or absence and size of the dish part to determine pass/fail, the number of pixels in a region including any point in the dish part is measured among the plurality of regions appearing in the binarized image, and the number of pixels is measured. A method for inspecting defects on end faces of nuclear fuel pellets, characterized in that a determination is made based on measurement results.