JPH11173996A - Nondesturctive detecting method for abnormal egg - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an egg inspection method in which an abnormal part such as a bloody egg or the like detected inside an egg with shell is judged in a short time, nondestructively and precisely. SOLUTION: An egg 3 is placed between a light source 2 and a color CCD camera 1. An obtained transmission image is decomposed into color components (R, G, B) of every pixel by using a video signal and an A/D conversion. Chromaticity coordinates (r, g, b) are found on the basis of the color components. The contour of only an eggshell image is extracted on the basis of the original transmission image which contains a background color and the eggshell image. In addition, an image noise by which a normal part is discriminated errneously as a bloody egg part is a bloody egg or not is discriminated by using the chromaticity coordinates. In addition, the size of the part of the bloody egg is estimated.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an egg inspection method for non-destructively and accurately judging abnormal portions such as blood eggs which appear in shelled chicken eggs in a short time.

[0002]

2. Description of the Related Art As a conventional egg inspection apparatus, an apparatus for detecting a defective liquid egg contained in a chicken egg taken out of a broken egg (for example,
Japanese Patent Application No. 6-43238) and a crack inspection device on the outer surface of a chicken egg (for example, Japanese Patent Application No. 7-220327). It is a common practice for non-destructive detection of foreign substances such as blood contaminating shell eggs without breaking the eggs, based on experience and intuition, using a transmitted light from the eggs. It is.

[0003]

As an absolute condition for the distribution of eggs with shells, it is required that the eggs be quality assured eggs free from abnormal eggs such as blood eggs, meat chunk eggs and broken eggs. However, as described above, a method, a technique, and an apparatus for non-destructively detecting a chicken egg with a shell have not yet been put to practical use, and this egg-checking process has relied exclusively on humans.

[0004] Generally, in an egg inspection process in an egg collecting place, a fluorescent lamp is installed below a conveyor for rotating and moving the eggs, and an egg inspection operator looks directly at the transmitted image of the eggs appearing in the dazzling light. Abnormal eggs are detected by experience and intuition.
Under the proposition that this egg inspection work is not allowed to distribute abnormal eggs, it has been pointed out that ordinary problems have to be carried out in a very nervous state physically and mentally. Was.

[0005] Further, under such severe operating conditions, too much attention is paid to the shipping time, so that the detection standard of abnormal eggs of an egg inspector varies, and as a result, consumers and the like may find that abnormal eggs are mixed. Problem was frequently encountered.

[0006] Fear of consumer complaints,
It is common knowledge of egg collection sites that "if you suspect an abnormal egg, do not distribute it." In this case, despite the fact that the contents are normal, there are actually many eggs that are not shipped, and it has been pointed out that this is a great loss economically and in terms of food resources.

On the other hand, on the consumer side, the sanitation
From a safety perspective, there is an increasing tendency to hate foreign matter contamination. Especially for chicken eggs that are often eaten raw, even if you find a bloody part as small as millet, even if the bloody part does not pose any problems in hygiene and quality,
Complaints were frequently introduced.

[0008] The present invention has been made in view of such problems, and its purpose is to non-destructively deconstruct the contents of a shelled egg by image processing without relying on the experience and intuition of an egg inspector. It is an object of the present invention to provide a method for detecting abnormal chicken eggs in which the measurement is carried out in a short time and as a result, as a result, blood-like foreign substances are mixed therein.

[0009]

In order to achieve the above object, the present invention provides a transmission image obtained by irradiating a transmission light to a chicken egg through a color CCD camera, a video signal processing and an A / D converter. A non-destructive detection method for abnormal chicken eggs, comprising performing a chromaticity comparison by comparing the chromaticity coordinates obtained from the color components with a predetermined value and determining whether or not the egg is a blood egg. There is.

According to the present invention, color components are converted into color components via a color CCD camera, video signal processing and an A / D converter.
The color egg is determined by comparing the chromaticity coordinates obtained from the color component with a predetermined value, and the value egg is discriminated.By applying image processing technology, abnormal egg such as blood egg is non-destructively.
Since the detection can be performed in a short time, blood eggs can be easily and accurately selected in place of a highly skilled specialized egg inspector.

Further, as an effective means of the present invention, a configuration in which only the eggshell image is extracted from the original image including the background color and the eggshell image as described in claim 2 is also an effective means.
By performing image processing only on the eggshell image, it is possible to more accurately detect abnormal chicken eggs by eliminating extra noise.

According to a third aspect of the present invention, when non-destructive detection of an abnormal chicken egg is performed, a pixel comparison is performed by comparing a blood egg pixel constituting the blood egg whose chromaticity has been compared with a predetermined value, and a normal portion is detected. It is also an effective means of the present invention to configure so as to avoid image noise that causes erroneous determination as a blood egg portion.

[0013] In order to prevent abnormal eggs from being placed on the sales route, the threshold value at the time of the chromaticity comparison must be set with a sufficient margin. May be present. However, by performing pixel comparison in which the blood egg pixel is compared with a predetermined value and selecting again, it is possible to avoid image noise in which a normal part is erroneously determined as a blood egg part.

The present invention mainly relates to a light source for transmitted light behind the eggs, a color CCD camera for photographing a transmitted image on the front of the eggs, an image processing board or RG for decomposing the transmitted image one pixel at a time. It is preferable to include an A / D converter for the -B signal, an arithmetic unit for determining a blood egg, or a computer.

[0015]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Preferred embodiments of the present invention will now be described in detail with reference to the drawings. However, the dimensions, materials, shapes, relative arrangements, and the like of the components described in this embodiment are not intended to limit the scope of the present invention unless otherwise specified, and are merely illustrative examples. Only.

As shown in FIG. 1, a color CCD camera 1
The egg 3 is placed horizontally and horizontally between the light source 2 and the light source 2, and a transmitted image of the egg is created using the transmitted light. Further, the transmitted image is separated into color components (R, G, B) of each pixel by an image processing board 4 having a video signal processing and A / D conversion circuit, and is taken into a computer 5. A light-shielding plate 7 having a small hole 6 for illuminating the light source is provided on the light source side when viewed from the egg, and a black support 8 for preventing the illumination from leaking from the rear light source to the outer surface of the egg is provided on the CCD camera side. Is installed. In addition,
A polarizing filter 9 is provided in front of the lens of the CCD camera 1.

At this time, the position where the eggs are placed may not be horizontal and horizontal, but may be in any direction. Further, the image need not be photographed in the horizontal direction as shown in FIG. 1, but may be photographed in the vertical direction or the oblique direction. The light source for the transmitted light may be any type, such as a fluorescent lamp, a laser spot light, or the like.

The transmitted image data is digital image information obtained by decomposing R, G, B per pixel into 256 gradations for each screen by 768 × 480 pixels per screen by the image processing board 4. By displaying the image at this time on the monitor TV, it is possible to proceed with the measurement while confirming that the illumination from the back light source has not leaked to the outer surface of the egg. No installation required.

In order to distinguish a blood egg from a normal egg by image processing, it is necessary to determine a certain threshold value and compare it with the threshold value. "Blood eggs" are roughly divided into eggs in which substances for blood spread throughout the egg white (so-called blood cells) and eggs mixed with bloody clots (blood spots). The whole image of the egg obtained in step (1) appears redder than that of a normal egg, and the latter image may appear red or black depending on the location of blood spots. Therefore, in the latter case, it is important to accurately extract an image region that is likely to be a blood egg portion.

In order to distinguish between a blood egg and a normal egg by image processing, the features of the blood egg part and the normal egg part are extracted,
You need to determine the threshold. In order to extract the “blood egg” portion from the collected image data, the color components (R, G, B) or the chromaticity coordinates (r, g, b) of the transmission image data
We consider an image extraction method based on a threshold using.

Now, 0 ≦ R ≦ 255 gradation, 0 ≦ G ≦ 25
Assuming that 5 gradations and 0 ≦ B ≦ 255 gradations, each chromaticity r, g, b is calculated by the following arithmetic expression using the obtained color components (R, G, B) for each pixel. r = R / (R + G + B) (1) g = G / (R + G + B) (2) b = B / (R + G + B) (3)

In this embodiment, the color component R of the original image data is used as a method for extracting only the eggshell image from the original image including the background color and the eggshell image, and R> 80 is empirically obtained from many preliminary experiments. As a threshold, the contour of the eggshell image could be extracted.

Next, a method for setting a threshold for extracting a blood egg portion from the eggshell image extracted by the above method will be described. Each color component of each egg was determined using fresh white eggs laid on the day. Egg size is 3 of LL, L and M
After the experiment was completed, all the eggs were broken and all eggs were confirmed to be blood eggs or normal eggs.

FIG. 3A is a diagram showing the interrelationship of each color component from the data obtained in the above experiment, in which a normal egg is represented by ○ and a blood egg is represented by +. The left side is a diagram in which the color component G is set on the vertical axis and R is set on the horizontal axis, and the gradations of these color components are plotted. The right side is set with the color component B on the vertical axis and G is set on the horizontal axis. It is the figure which plotted the gradation to take. FIG. 3B is a diagram showing the mutual relationship between the color coordinates calculated from the respective color components. On the left side, chromaticity g is plotted on the vertical axis and r is plotted on the horizontal axis.
It is a diagram in which the values of these chromaticities are plotted, and the right side is a diagram in which chromaticity B is plotted on the vertical axis and G is plotted on the horizontal axis, and the values of these chromaticities are plotted.

In FIG. 3 (a), since there are portions where blood eggs and normal eggs are mixed, it is difficult to distinguish blood eggs when using the color components (R, G, B). You can see that. On the other hand, when the chromaticity coordinates (r, g, b) of FIG. 3B are used, it can be seen that blood eggs and normal eggs can be distinguished.
In the case of FIG. 3 (b), it is difficult to select a blood egg from a normal egg with the chromaticity r and b. It can be seen that setting a threshold value to g makes it possible to detect a blood egg.

Therefore, in order to determine a more accurate threshold value, the chromaticity coordinates g = 0.27 to 0.32 and 0.005
A threshold is set at intervals, and an optimum threshold is determined from the result of the discrimination rate at that time. FIG. 4 shows a change in the discrimination rate at each threshold value. 4 (a) and 4 (b), when the threshold value is small, the discrimination rate of blood eggs is low but the discrimination rate of normal eggs is high. Conversely, when the threshold value is large, the discrimination rate of blood eggs is high but normal. It can be seen that the discrimination rate of eggs is low. Therefore, by integrating these results and using the total discrimination rate (FIG. 4 (c)) obtained by summing the blood egg and the normal egg, further, it is desired to secure a blood egg discrimination rate of 95% or more in the field. The degree coordinate g = 0.29 was set as the optimal threshold.

However, a part where g ≦ 0.29 occurs in the image of a normal egg due to noise or the like at the time of image data collection, and may be erroneously determined as “blood egg part”. In the present embodiment, the erroneous determination can be avoided by the following method.

The total number of pixels of the transmission image of the whole eggshell in this embodiment was 20,000 to 30,000 pixels. The total number of pixels of the blood egg portion is 1,000 to 25.0.
It was found that the range was 00 pixels. On the contrary,
Many measurements have shown that the total number of pixels in the case of misjudgment due to the above-mentioned noise is less than 500 pixels in the whole eggshell of a normal egg.

Then, g ≦ 0.29 for each pixel in the transmission image, that is, the total number γ of “blood egg pixels” when it is determined to be “blood egg portion” is counted, and γ ≧ 0.29
The egg at the time of 500 was determined to be “blood egg”, and when γ <500, it was determined to be “normal egg”.

FIG. 2 is an example of a flowchart of the blood egg determination in the present embodiment. In the figure, a transmission image of a hen's egg is input as a video signal (10), A / D-converted (11), and a color component (R, G, B) is obtained (12). The chromaticity coordinates (r, g, b) are calculated using (1), (2), and (3) (13). Then, only the eggshell image is extracted from the original image containing the background color and the eggshell image (1
4) After that, it is determined whether or not each pixel is a blood egg portion using the threshold value g for determining a blood egg portion (15). And
When the pixel of g ≦ 0.29 is determined to be a blood egg portion, the total value (16) of the pixel is counted, and the whole area of all eggs is terminated (17), and when the blood egg pixel γ is γ ≧ 500. Is determined to be a “normal egg” when γ <500 (18), and whether it is a blood egg (19) or a normal egg (20) is displayed on the display unit of the computer 5.

[0031]

Then, when a blood egg discrimination experiment was performed using the above-described threshold value g = 0.29, as shown in Table 1, an actual blood egg could be correctly discriminated as "blood egg". Only 41
The number was 401 out of the 6 pieces, and the discrimination rate was 96.4%.
In addition, 431 out of 466 eggs were able to correctly determine an actual normal egg as “normal egg”, and the discrimination rate was 92.5%.
%Met. As a result, the total discrimination rate of the two is 9
4.3%.

[0032]

[Table 1]

In order to evaluate the blood egg discriminating ability of the present embodiment, a comparison was made with the discrimination result of a skilled egg inspector at the egg collecting place by manual operation. Table 2 shows the results. This is the result of a skilled egg inspector preliminarily discriminating by the conventional transmitted light method before the experiment in Table 1. The skilled egg inspector had a blood egg discrimination rate of 98.8%, a normal egg discrimination rate of 91.9%, and an overall discrimination rate of 95.0%, which were almost the same as those in the present embodiment.

[0034]

[Table 2]

Further, a blood egg discrimination experiment was carried out using different eggs for different dates with the threshold value fixed at g = 0.29. As a result, a high discrimination rate of approximately 95% or more was maintained throughout.

As described above, it has been found that the blood egg discriminating ability of the present embodiment is sufficiently reproducible and can be put to practical use.

The light source C shown in FIG.
The threshold of blood egg discrimination may fluctuate depending on the characteristics of the CD camera, the image processing board, and the like. Therefore, the threshold is tuned by a preliminary experiment to obtain an optimum blood egg discrimination rate. It is desirable to keep.

[0038]

As described in detail above, the first aspect of the present invention provides a color CCD camera, video signal processing and A /
By converting to a color component via a D conversion device and performing chromaticity comparison by comparing the chromaticity coordinates obtained from the color component with a predetermined value to determine a blood egg, by applying image processing technology, Since abnormal chicken eggs such as blood eggs can be detected nondestructively and in a short time, blood eggs can be easily and accurately selected in place of highly skilled and specialized egg inspectors.

Further, as described in claim 2, since only the eggshell image is contour-extracted from the original image including the background color and the eggshell image, image processing is performed only on the eggshell image. Abnormal chicken eggs can be detected more accurately by eliminating extra noise.

According to a third aspect of the present invention, when non-destructive detection of an abnormal chicken egg, a pixel comparison for comparing a blood egg pixel constituting the blood egg whose chromaticity has been compared with a predetermined value is performed. Therefore, by performing pixel comparison in which the blood egg pixel is compared with a predetermined value and selecting again, it is possible to avoid image noise in which a normal part is erroneously determined as a blood egg part.

Therefore, since the present invention is an egg inspection system mainly based on image processing, humans can be released from the harsh egg inspection work as in the past. In addition, the present invention can be easily used as an egg inspection system not only in an egg collection place but also in a chicken egg processor for eating out foods, delicatessen, confectionery and the like and other mass consumers.

[Brief description of the drawings]

FIG. 1 is a configuration diagram of a nondestructive detection apparatus for abnormal chicken eggs according to an embodiment of the present invention.

FIG. 2 is a flowchart for determining a blood egg.

FIG. 3 is a diagram showing a mutual relationship between color information of an eggshell image.

FIG. 4 is a graph showing a change in a discrimination rate when a chromaticity coordinate g is set as a threshold value.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Color CCD camera 2 Light source 3 Egg 4 Image processing board 5 Computer 6 Small hole 7 Shield plate 8 Black support stand 9 Polarization filter

Claims (3)

[Claims] 1. A transmission image obtained by irradiating a transmission light onto a chicken egg is converted into a color component through a color CCD camera, video signal processing and an A / D converter, and chromaticity coordinates obtained from the color component are converted. A non-destructive detection method for abnormal chicken eggs, which comprises performing chromaticity comparison of comparing the blood with a predetermined value to determine whether or not the eggs are blood eggs. 2. The non-destructive detection method of an abnormal chicken egg according to claim 1, wherein only the eggshell image is contour-extracted from the original image including the background color and the eggshell image. 3. An image in which, at the time of non-destructive detection of an abnormal chicken egg, a pixel comparison is performed in which a blood pixel constituting the blood egg whose chromaticity is compared is compared with a predetermined value, and a normal part is erroneously determined as a blood egg part. 2. The method for non-destructively detecting abnormal chicken eggs according to claim 1, wherein noise is avoided.