JPS62211544A - Inspecting method for egg - Google Patents

Abstract

PURPOSE:To rapidly and certainly identify a crack, dirt, deformation, etc., of an egg and to decide smoothly its quality by irradiating light from one side of the edge to be inspected and picking up the egg to be inspected on image with an image picking device arranged at the other side. CONSTITUTION:First, when the egg E to be inspected is mounted on each rotating roller 2 of a conveyor 1 and each roller 2 is moved from the right to the left, each roller 2 is moved in the horizontal direction and each roller 2 is rotated anticlockwise. Accordingly, the egg E on each roller 2 is move to the left while being rotated clockwise. When an image inspecting instrument 9 detects that the egg E is located right above a stroboscope device 5 based on a signal S2 from a synchronizing switch 8 in this state, the instrument 9 sends a signal S3 to the device 5 and the image of the egg E is taken in a video camera 7. Then, this image is sent to the instrument 9 and converted into the digital image with an A/D conversion part and stored in an digital image storage part. Next, the image processing and decision are performed with an inspection area setting part and each decision part based on the stored digital image.

Description

DETAILED DESCRIPTION OF THE INVENTION "Field of Industrial Application" The present invention relates to an egg inspection method C for determining the quality of eggs by identifying cracks, stains, malformations, and thin parts of eggshells.

``Conventional technology'' Traditionally, sorting and packaging centers, so-called GP centers (Qrading
and pa, cking Centsr) l
``Problems to be Solved by the Invention'' However, in the conventional visual inspection of eggs mentioned above, cracks, dirt, thin parts, malformed eggs, etc. , there are the following problems. That is, (1) the eyes of the inspector who inspects the eggs are severe, and inspection errors are likely to occur;

■ Inspectors (2) Therefore, the quality of sorted eggs is likely to fluctuate due to the criteria for determining substandard eggs.

■ Quality control is difficult because standards for substandard eggs cannot be quantified.

Therefore, in an attempt to solve the problems of the conventional visual inspection, the Queen has used various self-inspection methods, such as applying a narrow beam of light and measuring the intensity of light leaking from the egg. A method of identifying cracks (see Japanese Patent Application Laid-Open No. 112579/1982), or a method of blowing air onto eggs to break defective eggs such as cracked eggs, broken eggs, and eggs with weak eggshell strength (special method). (See Japanese Patent Application Laid-Open No. 58-170416), or a method of applying pressure to the egg to make the crack larger and then detecting and identifying the elastic waves (see Japanese Patent Application Laid-Open No. 58-170416) has been proposed. However, for practical use,
This method has various problems and has not been fully addressed.In some cases, each of the conventional automatic inspection methods described above conducts separate inspections for individual items, such as cracked eggs, etc. In order to completely automate the inspection of inspection items, it is necessary to install a separate transverse strain device for each inspection item, which makes the sorting device complicated and large, making the current sorting and packaging difficult. This is because it cannot be incorporated into the device.

The present invention was made in view of the above circumstances, and its purpose is to be able to quickly and reliably identify cracks, stains, malformations, and thin films in eggs, and to facilitate smooth determination of the quality of eggs. To provide an egg inspection method capable of automatically and continuously ζ:
There are two.

"Means for Solving the Problems" In order to achieve the above object, the present invention irradiates light from one side of the eggs to be inspected and images are captured by an image acquisition device tC2 placed on the other side of the eggs to be inspected. In the egg inspection method of the present invention, the egg inspection method of the present invention involves capturing an image of the inspection area and processing the captured image using the gray scale fJ, Il difference and pattern to determine the quality of the rabbit to be inspected. , irradiate light from one side, take an image of the l1liI image of the egg to be inspected from the other side, and process it from the gradation scale difference and pattern C2 based on this obtained image. Identify cracks, stains, malformations, thin films, etc. in the spinal examination area.

"Actual Example" Hereinafter, based on Figures 1 to 39, g-g of the present invention
A. Explain overgrowth.

1 in the figure Ig1 is a conveyance device for the eggs to be inspected E, and this conveyance device 1 consists of a large number of equally spaced rotating rollers 2
Both ends of the C are rotatably supported by support members (not shown), and the support members are horizontally supported by each rotating roller C and a chain drive @JrC (Fig. 1). While being moved (from right to left), the lower portions 1-1 at both ends of each rotating roller 2 are pressed, and a fixed member (not shown) is placed to rotate each rotating roller 2. and C:, the lower ζ2 light shielding plate 3 of each of the above rotating rollers 2
are installed, and transparent plates 4 such as acrylic plates are fitted into windows formed at four locations on the light shielding plate 3 at the same intervals as the installation intervals of the rotating rollers 2. Below each of the transparent plates 4, there are four strobe devices j15 and lenses 6, and above the transport device l, there are four strobes, and four flashes illuminated by the eI flash device 5C2. Two video cameras (image acquisition devices) 7 are installed to take two images of each of the examination areas E. In addition, the timing of irradiation of the O-bo device 5 and the timing of image capture by the video camera 7e are determined in step (a) above. Horizontal direction C: A synchronization switch 8 for synchronizing each rotating roller 2 that is moving is placed at an appropriate position on the conveying device 1. still image)
The signal S1 and the signal S2 from the synchronous switch 8 are:
When the image inspection device [9C is input 1;, this image inspection device 9 sends a strobe control signal S3 to the strobe device fff5, and a pick-up device that removes the defective eggs E to be inspected from the transport device 1. The pick-up output signal S4 is output for each of the 10 and 10, respectively. In the image inspection apparatus 9C, as shown in FIG.
1 The timing control section 131 of the cloud II synchronization information;
and the vertical synchronization signal and the synchronization switch signal S2
Based on this, the timing control unit 13 outputs the digital image 600 million copies 12, the strobe control unit 14 that outputs the upper epsilon strobe control signal S3, and the above-mentioned output signal 8.4.
The pick-up output control section 15 outputs a command signal, and the pick-up output control section 15 outputs a command signal. Furthermore, the digital image in the digital image storage section 12 is sent to the inspection region setting section 16 (:), where the contour of the subject's spinal region E is set and the digital image is The outline (inspection area) of the egg E to be inspected is determined by the malformation determining section 17, the thin film determining section 18, and the crack determining section 1, respectively
9 and a dirt determining section 20. These determination units 17 to 20 use the digital image, the inspection area, and each parameter table 21 to
Based on the determination reference values set in 24, the quality of the inspected part E is determined for each inspection item (malformation, thin film, crack, stain). The judgment signals of each judgment section 17 to 20 are sent to the judgment result processing section 25, where a comprehensive judgment of the inspected section E is performed, and the result is is stored in the judgment result buffer 26. Then, the output of the judgment result buffer 26 and the timing control section 13 are
Based on the command signal and (;, the pick-up output control unit 15 sends a pick-up output signal S4 to the pick-up device 10 to remove the WIt spine E that has been determined to be defective from the transport device 1. +J71 has been made to send.

Next (=, I: As described above, there are two cases in which the inspection method of the present invention is carried out using the egg inspection apparatus configured as described above.

First, as shown in FIG. Since the rollers 2 move in the horizontal direction while contacting the fixed member ζ2 (not shown), each of the rotating rollers 2 rotates counterclockwise.Therefore, the inspected portion E on each rotating roller 2 rotates clockwise. (
Move C2 to the left while rotating once. In this state,
Based on the signal S2 from the synchronization switch 8, when the image inspection device 9 detects that the part to be inspected E placed directly above C2 of the strobe device 5 and between each rotating roller 2 is located, the image inspection device 9 The inspection device 9 has a strobe device 5C and two signals S:
(), and the video camera 7C captures an image of the inspected area E at the moment when each strobe device 5 irradiates the object, and this image is sent to the image inspection device 9 (= image inspection device!). f
The image is converted into a digital image by the A/D converter 11 of 9 and stored in the digital image storage 12. In addition, this example C
:In case of 1 subject.

There are 2 eggs, 4 eggs) Upper position 1 of a-bo device 5:
Each time it comes, a total of four images are taken, and then, based on the digital image ζ2 stored in the digital image storage section 12, the horizontal vehicle area setting section 16 and each judgment section 17
20, image processing and determination are performed. There are two processing contents C, which will be described in detail based on the flowchart C shown in FIGS. 3 to 8.

First, in FIG. 3C, the digital ii! ii When the information is input (see steps SP1 and SP2), the inspection area setting process shown in step SP3 is performed based on this digital image in the inspection area setting section 16C2 shown in FIG. It is done. This setting process, as shown in FIG.
! Calculate the brightness change rate inward from both ends (
(See step 5P11). Then, based on this result, determine whether there is a peak address (starting point) for the rate of change (see step 5P12), and if there is not,
It is determined that the inspected area E has not been photographed, and an egg-free code is set and stored (see step 5P13). Also, if the starting point of the above rate of change is 9 (= is step 5P
14t=Advance, from the line where the starting point is located, a pair of left and right peak addresses (
Search for the point where the rate of brightness change is maximum), and repeat this until the distance between both peak addresses becomes smaller than the reference value of the distance between the left and right contours (see step 5P15), and then search for both peak addresses. When the distance between the peak addresses becomes smaller than the reference value, proceed to step 5P16 (: Proceed to search for a pair of left and right peak addresses within a certain range every C21 line downward from the starting point, and find the distance between the two peak addresses. Repeat this until the value becomes smaller than the reference value (see step 5P17), determine the outline (inspection area) of the egg E to be examined, and set and store the egg presence code (see step 5P18@). In this case, if the above reference value is set too large, the contour (inspection area) where the upper and lower parts of the IeJ image of the part to be inspected E are cut off in the 9th factor, Figures 17, 27, and @33. is set, but if a suitably small value is selected as the reference value, an inspection area approximately approximated by the contour C2 of the image of the part E to be inspected is set. Once completed! ! Proceeding to step SP4 of the main flowchart shown in 311C2, the presence or absence of the inspected egg E is determined based on the egg absent or egg present code set above, and if the inspected egg E is not imaged ( Second,
Step SP 1ζ2 Return. Further, when the egg to be inspected E has been imaged, based on the malformed egg determination process in step SP5, the malformation determination unit 17 shown in FIG. 2 determines whether or not it is a malformed egg.

This malformed egg determination process: To explain based on the malformation determination flowchart shown in FIG. To calculate t, refer to step 5P21), this length t
is between two reference values set in advance in the parameter table 21, namely the minimum longitudinal direction reference value L1 and the maximum reference value ([L2])
(See step 5 P22). If the lifting length t is not between the two reference values Ll and L2, C: is determined to be a malformed egg, and a malformed egg code is set and stored (step 5).
(see page 23), and two reference values Lt a L2
If it is between C: as shown in step 5P24.
, calculate the diameter r of the front press E to be tested, and determine whether the diameter r is between the two reference values Rx and Rz (step 5P
25), if the diameter r is not between the two reference values R1 and R2, a malformed egg code is set and both reference values R1 and R2 are set.
(Secondly, calculate the ratio t/r between the length t and the diameter r of the front push E to be tested, and this ratio L/r is equal to the two reference values P.
1. Determine whether there is ζ2 between P2 (step 5P
26.5, P27), and in this determination, both reference values P1. P2
If there is no (2), set and store the malformed egg code, and if the above ratio L/r exists between both reference values P1 and P2, set and store the back code (see step 5P23.8P28)
.

When the process for determining the Umi-shogata egg is completed T, the process proceeds to step SP6 shown in FIG. Determine whether or not there are 4 shells. This judgment process takes advantage of the fact that thin film parts are usually formed on the sharp or blunt parts (#I9 (right and left ends) of the egg E to be inspected). , For example, 10(2) to 6 ζ: The grayscale gradation (histogram) of each line ■ to ■ shown c: Based on the magnitude of the difference in brightness between the center and both ends of the front press E to be tested. and,
The quality of the egg is determined by determining whether the bright and dark thin film has a linear boundary between light and dark), and the size of the area of the thin film.

There are two steps in this process, and the explanation will be based on the thin film determination flowchart shown in FIG. 6. First, step 5P31 (:
As shown, the rate of change in brightness is calculated for each line on the acute end side of the front push E to be tested, and its maximum value dsn (n is a positive integer indicating the number of horizontally distorted lines) is found. Then, it is determined whether the maximum value dSn of the rate of change for each line is larger than the reference value DS set in the parameter table 22, and if it is larger, then whether the coordinates of each maximum value asn are equal or not. (See step 5P32.8P33).

If both of these two determination processes are yes (YES),
Step 5P34 + binary advance, area a of the thin film part in acute nm
Fi is calculated, and it is determined whether this area a8 is larger than the reference value A (see step 5P35).

If the surface 11as > reference value A, the sharp end is 4
It is determined that it is the standard outer shell of the shell, and the 4 shell codes are set and stored (see step 5P41). Also, if any of the above determination processes (steps SP 32, SP 33, and SP 35) is negative (NO) (2) is steps 5P36 to 5P40
Perform thin film discrimination processing on the blunt end as shown in . This process is similar to the process for the sharp edge described above, so a description thereof will be omitted. In Fig. 6, aBn is the maximum brightness change rate for each line on the obtuse side, DB is the standard value of the brightness change rate, and all is the obtuse (2)
) Indicates the area of the lateral thin film. And the above winter hair j! II (Step 5P37.5P38.5P4
0), if any of them is negative (NO), it is determined that it is a skin part, the main part code is set and stored, and if each process is all yes (Y1iC8), the sea sets a 4-shell code ( (See step 5P42.5P41).

Next, when the Kamishi thin film determination process is completed, the process proceeds to step SP7 shown in figure fJc3, and the broken egg determination process 3! Based on Ill2, the crack determination unit 19 shown in FIG. 2C determines whether or not the inspected portion E is cracked. This judgment part y!
JiI" - cracks and scratches (not actually cracked,
It is necessary to distinguish between those that are outside the standard and those that are not.

In other words, the crack pattern has many spiderweb-like patterns (or branches, zigzags, and curves), and the pattern has sharp changes in brightness, whereas the scratch pattern has a straight line. This method takes advantage of the fact that most of the patterns are shaped like a circle or an ellipse, and in the case of a straight line, the width is almost the same throughout the entire pattern, and the brightness in the pattern is constant. Based on the pattern and histogram shown in FIG. 32 (2), a crack and a flaw in the fracture+*vertebral region E are distinguished.

Regarding this process, see Figure 7 1: Flowchart C for determining cleavage.
:Explain based on 6 First, step 5P511: As shown in 1:, detect the peak (maximum point) of the bright part of the inspected part E, and calculate the difference p between the brightness of the peak and the brightness of the surrounding area, and the brightness. Calculate the rate of change au when the level of brightness increases and the rate of change 1d when the brightness level decreases, and calculate these values p.

The magnitude of du*da and each reference value P#Du, Dt1 is determined (see step 8528), and
If even one of these three size determinations is negative (NO), the process proceeds to step 5P53, where it is determined whether or not all data to be examined has been determined, and if it has not been completed (; Return to step 5P51, and when completed, store the skin code (step 5P5)
(see 4).

In addition, in step 5P52+2, if all three size determinations are yes (YES), proceed to step 5P55, set the tracking level (reference value of tracking brightness) La, and then set this tracking level LaC :Based on inspected part E
Tracking pixels of the above tracking level La or higher in the image circle,
A pattern for bright areas is created (step 5P56). From this C2, for example, in Fig. 17 (1), two cracked parts are created, and in Fig. 27 (2), Vs, the flawed part is created as a bright pattern. Based on the pattern (2), in steps 5P57 to 5P65, it is determined whether the egg is broken or not. In other words, the above-mentioned cracks and scratches are distinguished, and the pattern (2) where there is a spider web-like branch (see step 5P58) is determined. , pattern C:
There are bends (see Steps 5P59 and 5P62), patterns (: sharp changes in brightness and darkness (Steps 5P61.8)
(See P65) In case C:, it is determined that the egg is cracked, and the broken egg code is set and stored (see step SP668).In other cases C:, the process returns to step SP511m and repeats the data inspection process.

Furthermore, when the above-mentioned cracked egg judgment is completed, in step 5P8 (step 2, dirty part judgment process C2 of FIG. 3), in the dirt judgment part 2 shown in FIG. This process determines whether the egg's smooth part is, for example, the third part.
As shown in Figures 3 to 39 □□□, this method takes advantage of the fact that the area is darker than the surrounding area, and the area of the dark area is measured to make the determination.

That is, as shown in Fig. 8 (:, First, the peak (minimum point) of the dark part of the image of the inspected area E is detected, and the difference between the brightness around the peak and the brightness of the peak is the brightness. Calculate the rate of change fid when the level of brightness is decreasing (when the level of brightness is decreasing) and the change ibu when the level of brightness is increasing, and use these values as bd.

The size of bu and each reference value B # B (L # Bu is determined) (see step 5P71.8P72), and if any of these size determinations is negative (No), the data to be inspected is It is determined whether or not all the processes have been completed, and if not completed, go back to step 5P71+2, and if completed, C: sets and stores the skin part code (see steps 5P73 and 8P74).

In addition, in step 5P72, if all three size determinations are W (YES), C sets the tracking level (standard value of darkness to be tracked) M, and then sets the tracking level M or less with the peak pixel as the center. A dark pattern is created by tracing the pixels of (see steps 5P75.S and P76). Then, based on this pattern (2), check whether the dark part 8t8 is larger than the reference value S (see step 5P77) and the difference MmaX between the darkest pixel in the dark part and the surrounding pixels.
is larger than the reference value N (see step 5P78), and if both are negative (NO), return to step 5P71, and if either is yes (YES), C: is dirty. It is determined that the dirty part coat degree is set and stored (see step 5P79).

In this way (in FIG. 2), each determination unit 17 to 2
After the determination of malformations, thin films, cracks, and stains at step 0 is completed, the results (FF-shaped egg code, thin film code, broken egg code, dirty part code, or skin part code) are sent to the determination result processing section 25. (See step SP9 in FIG. 3), and in this determination result processing section 25, a final determination is made as to whether it is a skin part or a standard contour, and the determination result is stored in a determination result buffer 26. Then, the pick-up output control section 15 determines the standard that has reached the installation position of the pick-up device R1O based on the command signal of the tie-bang system d141 section 13 and the good judgment result stored in the judgment result bag 726. The outer shell is removed from the a feeding device 1 (the picking up device 10 is controlled so that the outer shell is removed from the transport device btl), and only the skin part is used to handle eggs of different sizes in the next stage. As described above, the judgment of the quality of the eggs E to be inspected is carried out smoothly and reliably. Light is applied from below and passes through the egg E to be inspected.The image based on this is taken with a video camera 7, and by processing this image, the entire shape of the egg E to be inspected, the 11v shell, Since cracks and dirt are identified, the configuration of the optical system is simple, and pass/fail judgments can be made quickly.Therefore, the process that was conventionally inspected visually can be automated, and can be done using just one inspection device. and above 4
Since it is possible to inspect items, there is no need to significantly improve the conventional scanning process, and only one inspection device can be introduced and installed in the inspection process. −
Since the strobe device 5 and the video camera 7 are synchronized by the timing control unit 13 based on the signal S2 of the synchronization switch 8, a stable image of the egg E to be inspected is always captured by the image inspection device 9C. .

Therefore, as with visual inspection, there may be oversights, mistakes, etc.! In addition, the selection criteria will not fluctuate and the selection of selected eggs will not vary.Roughly, the image taken into the image inspection device 9 is processed based on the gray scale and the difference C:IS. Since patterns C of cracks and scratches are classified and discriminated, it is possible to easily and reliably distinguish between cracks and scratches, thereby improving the reliability of egg judgment. Furthermore, by changing each reference value set in the parameter tables 21 to 24 of each determination section 17 and 20, egg inspection can be performed that flexibly and immediately responds to changes in environmental conditions such as season and temperature. It will be done.

Note that although the strobe device 5 is used as the light source in this embodiment C2, the present invention is not limited to this, and the video camera 7 can also be used as the light source.
Alternatively, if a shutter device is provided on the transparent plate 4, a continuous lighting type light source may be used.

In addition, in this practical example, one video camera 7 has two
Although images of several eggs E to be inspected were taken at the same time, the number of eggs E to be inspected that are captured by the video camera 7 at one time was taken into consideration, including the processing speed of the image inspection device, manufacturing cost, etc. 1
The number may be 1, or 3 or more. Furthermore, in Example 1, four images should be taken for one egg E to be inspected, but it would be better if seven images covering the entire circumference of the egg could be captured (for example, in a continuous manner). ``Effects of the Invention'' As explained above, the present invention provides an apparatus for irradiating light from one side of an egg to be inspected to capture three images from the other side of the egg to be inspected. The egg to be inspected is imaged by an image acquisition device, and the captured image is processed using gradation differences and patterns to determine the quality of the egg to be inspected. Image acquisition device @I: Intensity/weakness of the image of the captured egg to be inspected (-
By converting and identifying cracks, stains, malformations, thin films, etc. in the fractured spinal region in a single examination, based on this gradation difference and the pattern of bright or dark areas in the image,
It is possible to reliably (and smoothly determine whether eggs are good or not), it is possible to automatically and continuously sort eggs, it is possible to greatly improve the reliability of the sorting, and it is possible to shorten the inspection time. It can shoulder such excellent effects.

[Brief explanation of drawings]

1 to 39 show one embodiment of the present invention, the first factor is a schematic configuration diagram showing the overall configuration, FIG. 2 is a block diagram of the image inspection apparatus, and the third factor to FIG. 8 Figure 3 shows the processing content of each part of the image inspection device. Figure 3 is the main flowchart, Figure 4 is the flowchart for setting the inspection area, Figure 5 is the flowchart for deformity determination, and Figure 6 is the flowchart for thin layer determination. , Fig. 7 is a flowchart for determining a broken egg, Fig. 8 is a flowchart for determining a dirty part, the ninth factor is an explanatory diagram of an egg with 4 shells and 1 sharp part, I! lO diagram or l! Figure 16 is 3
J! A characteristic diagram illustrating the gray scale difference of each inspection line ■ to ■ in Figure 9, Figure 17 is an explanatory diagram showing the state of cracking, and Figures 18 to 26 are each inspection line ■ to ■ in Figure 17. Figure 27 is an explanatory diagram of a damaged egg, and Figures 28 to 32 explain the rtjt gray scale difference of each inspection line ■ to ■ in Figure 827. Characteristic diagram, I! 3
Figure 3 is an explanatory diagram of eggs with dirt, 1M34 to @39
The figure is a characteristic diagram illustrating the gray scale difference of each inspection line (■) to (■) in Figure @33. 5...Strobe device 7...Video camera (image acquisition device) 9...
...1itj <jJ @Inspection device E...Egg to be inspected. Applicant Niigata Iron Works Co., Ltd. Figure 3 Figure 4

Claims (1)

[Claims] Irradiating light from one side of the egg to be inspected, taking an image of the egg to be inspected by an image capturing device placed on the other side of the egg to be inspected, and processing the image using a tone difference and a pattern, An egg inspection method characterized by determining the quality of the egg to be inspected.