JP2009229100A - X-ray foreign matter inspection device for food, and method therefor - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an X-ray foreign matter inspection device for food, and a method therefor, capable of detecting a foreign matter in an inspection object with high accuracy or at high speed. <P>SOLUTION: This X-ray foreign matter inspection device 10 for food is equipped with an X-ray generation part 20 for irradiating an X-ray to the inspection object 12, and an X-ray detection part 30 for detecting the X-ray transmitted through the inspection object 12. The X-ray foreign matter inspection device is also equipped with an image processing part 40 for dividing X-ray transmission image data from the X-ray detection part 30 into a plurality of images with a different division number, setting a threshold in each divided image and binarizing it, and determining existence of a foreign matter 13 by decision by majority from the acquired binarization result of each pixel. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a food X-ray foreign substance inspection apparatus and method, and more particularly, to a food X-ray foreign substance inspection apparatus and method for detecting foreign substances mixed in a product such as food.

  In order to maintain products such as processed foods at a predetermined quality, the products are inspected for foreign matters in the post-production line. Since it is difficult to visually detect fine foreign matters such as glass and resin pieces, a foreign matter inspection apparatus using X-rays has been used.

FIG. 10 is a diagram showing a schematic configuration of a conventional food X-ray foreign matter inspection apparatus. As shown in the figure, the X-ray foreign matter inspection apparatus 1 includes an X-ray generation unit 3 on a product transporting conveyor 2 and an X-ray detection unit 4 below an opposing transporting line 2 and includes a data processing unit 5.
The X-ray generation unit 3 irradiates the inspection object 6 such as food with X-rays. The X-ray detection unit 4 detects X-rays that have passed through the inspection object 6.

  In the food X-ray foreign substance inspection apparatus 1 having such a configuration, the X-rays radiated from the X-ray generation unit 3 pass through the inspection object 6 and are received by the opposing X-ray detection unit 4. For example, the X-ray detection unit 4 scans (scans) the first to 1280th element in one line. When the inspection object 6 moves on the conveyor 2, a two-dimensional image is formed. In the data processing unit 5, the X-ray image data of the X-ray detection unit 4 is input, compared with a threshold value such as reference data when a foreign object is set in advance, and based on the comparison result It is determined whether or not the object 6 has foreign matter mixed therein.

  Patent documents 1 and 2 are mentioned as a conventional food X-ray foreign substance detection device. Among them, Patent Document 1 irradiates a food pack with X-rays and performs threshold selection by trend display when inspecting foreign matter from the obtained data. At this time, the threshold value is manually selected.

In Patent Document 2, when a food pack is irradiated with X-rays and a foreign object is inspected from the obtained data, a threshold is selected using a statistical method, and the threshold is updated each time a new food pack is inspected. Has been.
JP 2004-28685 A JP 2007-333562 A

  However, a conventional food X-ray foreign substance inspection apparatus can easily detect a metal piece having a relatively high detection sensitivity, but it is difficult to detect small pieces such as resin and glass other than the metal piece. It was. This is because resin, glass pieces, etc. are displayed as pale images and it is difficult to determine whether or not they are foreign objects. Moreover, it has been difficult to perform the foreign object detection process in real time on a high-speed belt conveyor.

  In the foreign object detection device disclosed in Patent Document 1, a detection limit value (threshold value) that serves as a reference for detecting the presence or absence of an abnormality is set by key operation, that is, manually. Therefore, the threshold value must be reset every time the inspection object changes. In addition, troublesome parameter setting must be performed every time when foreign matter is determined.

  Furthermore, in the foreign substance detection apparatus shown in Patent Document 2, the threshold value is updated and improved as the number of inspection objects increases, but there is a high possibility of erroneous detection in the initial state before data update.

Accordingly, an object of the present invention is to provide a food X-ray foreign substance inspection apparatus and method that can detect foreign substances in an object to be inspected with high accuracy.
Another object of the present invention is to provide a food X-ray foreign substance inspection apparatus and method that can speed up the process of determining foreign substances in an object to be inspected.

  An X-ray foreign matter inspection apparatus for food according to the present invention includes a X-ray generation unit that irradiates an inspection object with X-rays, and an X-ray detection unit that detects the X-ray transmitted through the inspection object. In the X-ray foreign substance inspection apparatus, a plurality of divided images having different division numbers are generated from the X-ray transmission image data output by the X-ray detection unit, a threshold value is calculated for each block of the divided images, and binarized. The image processing unit is characterized by determining the presence or absence of a foreign object by majority based on the binarization result for each pixel of the divided image.

In this case, the image processing unit stores the X-ray transmission image data output from the X-ray detection unit, and a plurality of divided images having different numbers of divisions of the image area of the X-ray transmission image data. It is preferable that the image forming apparatus includes a dividing unit that divides the image into two and a determination unit that determines a foreign object by majority vote based on the binarization result for each pixel of the obtained divided image.
The image processing unit may include binarization processing means for creating a density histogram for each block of the divided image, calculating a threshold value based on the density histogram, and binarizing the threshold value.

  The X-ray foreign matter inspection method for food according to the present invention irradiates an inspection object on a transport line with X-rays, collects X-ray transmission image data of the inspection object, and inspects the foreign object of the inspection object. In the X-ray foreign substance inspection method for food, the X-ray transmission image data is divided into a plurality of divided images having different division numbers, a threshold value is calculated for each block of the divided images, and binarized. Based on the binarization result, the presence of foreign matter is determined by majority vote.

を利用することにより求めることを特徴としている。 Further, the X-ray foreign matter inspection method for food according to the present invention irradiates an inspection object on a transport line with X-rays, collects X-ray transmission image data of the inspection object, and statistically analyzes the foreign object in the inspection object. In the food X-ray foreign substance inspection method for determination using threshold selection based on the quantity, the threshold selection divides the density histogram of the amount of X-ray transmission into two classes, the foreign object and the inspection object, and the two classes. Is a normal distribution, and in the process of calculating a threshold value that minimizes the average recognition error rate, the variance when the threshold value is k is σ ² _k , each density level is i, and the threshold value is k When the mean is μ _k and the probability that the concentration level appears is p _i

It is characterized by obtaining by using.

  In the X-ray foreign matter inspection method for food according to the present invention, X-rays are applied to the inspection object on the transport line, X-ray transmission image data of the inspection object is collected, and the foreign object of the inspection object is inspected. In the X-ray foreign matter inspection method for food, the X-ray transmission image data is divided into a plurality of divided images having different numbers of divisions, and a density histogram of the X-ray transmission amount is determined for each block of the divided images. Dividing into two classes of objects, assuming that the two classes are normal distributions, in the process of calculating a threshold value that minimizes the average misrecognition rate, in the threshold calculation process, variance calculation is performed with an approximate expression, The calculated threshold value is binarized, and the presence of the foreign matter is determined by majority decision based on the binarization result for each pixel of the divided image.

According to the X-ray foreign substance inspection apparatus for food of the present invention having the above configuration, a plurality of divided images having different division numbers are generated from X-ray transmission image data, a threshold value is calculated for each block of the divided image, and binarized Based on the binarization result for each pixel of the divided image, the presence or absence of foreign matter can be determined by majority vote. Even when the inspection object changes, the threshold value is calculated based on the X-ray transmission image data output by the X-ray detection unit, and therefore it is not necessary to manually set or update the threshold value. Even in the initial stage of inspection, foreign matters are not erroneously detected. Further, the detection accuracy is improved, and even a small piece such as resin or glass can be detected with high accuracy.
In addition, since the presence / absence of foreign matter is determined by majority decision based on the binarization result for each pixel of the divided image, the determination system is improved by narrowing down abnormal foreign material candidates, and the possibility of erroneous detection is reduced. .

Detailed embodiments of a food X-ray foreign substance inspection apparatus and method according to the present invention will be described below with reference to the accompanying drawings.
FIG. 1 is a diagram showing a schematic configuration of a food X-ray foreign substance inspection apparatus according to an embodiment. FIG. 2 is a block diagram of the food X-ray foreign substance inspection apparatus according to the embodiment. As shown in FIG. 1, a food X-ray foreign substance inspection apparatus 10 is attached on a conveyor 14 for a product such as food to be inspected 12. The transport conveyor 14 transports the inspection object 12 to a post-processing step at a subsequent stage at a predetermined transport speed.

The X-ray foreign substance inspection apparatus 10 for food has an X-ray generation unit 20, an X-ray detection unit (line sensor) 30, an image processing unit 40, a selection unit 50, and an image display unit 60 as main constituent requirements. Yes.
The X-ray generation unit 20 uses a cylindrical X-ray tube provided along the transverse direction of the transport conveyor 14. An X-ray tube generates X-rays by irradiating an anode target with an electron beam from a cathode. X-rays from the X-ray tube are emitted toward an X-ray detection unit 30 described later.

  The X-ray detection part (line sensor) 30 is provided in the lower part of the conveyance conveyor 14 facing the X-ray generation part 20 on the conveyance conveyor 14. The X-ray detection unit 30 is provided along the transverse direction of the conveyor 14 like the X-ray generation unit. The X-ray detection unit 30 outputs an electrical signal corresponding to the amount of X-ray transmitted through the inspection object 12. In the present embodiment, a line sensor is used as the X-ray detection unit 30.

The X-ray transmission image data detected by the X-ray detection unit 30 is input to the image processing unit 40 that is electrically connected.
The image processing unit 40 includes a CPU, a memory, and the like. Specifically, the data storage unit 41, the image cutout unit 42, the filter processing unit 43, the image division unit 44, and the binarization processing unit 45 are provided. The determination means 46 is the main constituent requirement.
The data storage means 41 receives and stores X-ray transmission image data of each inspection object detected by the X-ray detection unit 30.

  The image cutout means 42 is set in advance with a rough threshold value that can be determined as the inspection object 12, and among the entire X-ray transmission image data, data having a density value equal to or lower than the threshold value is determined as the inspection object 12 and the foreign object. Cut out as 13 image data. Specifically, as shown in FIG. 3A, the image data portion of the inspection object 12 is left in the X-ray transmission image data, and a blank portion other than the inspection object 12, that is, the background is cut out (FIG. 3). 3 (2)).

  Generally, noise is generated in the X-ray transmission image data detected by the X-ray detection unit 30. Therefore, the filter processing means 43 performs a filtering process in order to remove the generated noise and enhance the foreign matter 13 to enhance the determination accuracy of the presence or absence of the foreign matter 13. Specifically, the filtering process of the present embodiment performs three types of filter processes using a median filter, a smoothing filter, and a Max-Min filter. First, the median filter removes noise by arranging the densities of peripheral pixels of a certain pixel in order and using the median as the pixel density.

The smoothing filter removes noise using the average value of the densities of peripheral pixels of a certain pixel as the density of the pixel.
The Max-Min filter takes out the maximum value and the minimum value of the density values of peripheral pixels of a certain pixel, and adds the difference to the density value of a certain pixel as a density value to emphasize the foreign matter 13 (FIG. 3 ( 3)).

  The image dividing unit 44 generates a plurality of divided images having different numbers of divisions from the X-ray transmission image data after the filter processing by the filter processing unit. As an example of the number of divisions, seven patterns of 3 × 3 to 9 × 9 divisions can be set. The number of divisions can be arbitrarily changed according to the image data of the object to be inspected, such as n × n division or n × m division, and is set to an odd number of at least 3 or more. This odd number of times is set to facilitate determination by the determination means 46 described later.

The binarization processing unit 45 performs binarization processing for acquiring a binarized image from a grayscale image. Specifically, the binarization processing unit 45 creates a density histogram for each block of the divided image divided by the plurality of division numbers generated by the image dividing unit 44. FIG. 4 is an explanatory diagram of image data in the binarization process. In the density histogram shown in FIG. 4A, the horizontal axis represents density values (brightness and darkness), and the vertical axis represents the number of pixels (frequency). As shown in the figure, 25 blocks are generated in a 5 × 5 divided image. A density histogram is created for each block. As shown in the figure, the density histogram is normally generated by two peaks based on the image of the inspection object. The position of the valley portion between the two peaks is set as a threshold value, and the foreign substance side (dark density) is class 1 and the product (inspected object) side (dark density) is class 2. The optimum threshold value is selected by obtaining a threshold value k that minimizes the evaluation function J (k) shown in Equation 2.

Here, when the threshold in Equation 2 is k, the variance of class 1 is σ ₁ (k), the variance of class 2 is σ ₂ (k), the occurrence probability of class 1 is ω ₁ (k), As the occurrence probability ω ₂ (k), the variances σ ₁ (k) and σ ₂ (k) of classes 1 and 2 in Equation 2 are generally calculated using the equation shown in Equation 3.

となり、数式３から数式１が導かれる。しかし、計算の際に数値を有限桁でしか扱えないコンピュータ上で数式１を計算すると数式３に比べ誤差が大きく生じる。通常数式１の計算方法は用いられない。通常は用いられない計算法ではあるが、本実施形態において分散は、最終的な計算対象である閾値の計算過程において利用するのみであり、最終的な結果に対する計算誤差の影響は少なくなる。数式１を用いることで、全体としての計算オーダーはＬに比例することとなり、処理時間が短縮される。 Here, the variance when the threshold value of Equation 3 is k is σ ² _k , each density level is i, the average when the threshold value is k is μ _k , and the probability that the density level appears is p _i .
According to the method using Expression 3, it is necessary to perform calculation for each k. When the maximum density level is L, calculation of L steps is required as a whole, and the overall calculation order is proportional to the square of L. It will be. Therefore, there is a problem that it takes a long time to calculate and the processing time for foreign object detection becomes long. Therefore, in this embodiment, as shown in Equation 4, the dispersion equation (Equation 3) is approximated (Equation 1).
In theory,

Thus, Equation 1 is derived from Equation 3. However, if Equation 1 is calculated on a computer that can handle numerical values with only a finite digit during calculation, an error is larger than Equation 3. Usually, the calculation method of Formula 1 is not used. Although this is a calculation method that is not normally used, in this embodiment, the variance is only used in the process of calculating a threshold that is the final calculation target, and the influence of the calculation error on the final result is reduced. By using Equation 1, the calculation order as a whole is proportional to L, and the processing time is shortened.

  In this example, since the density data of each pixel is expressed by 12 bits, the maximum density level L is 4095. The calculation time of the dispersion can theoretically be shortened to 1/4095 at maximum, and the actual value of the total calculation time including the dispersion is approximately 20 seconds by the method using Equation 3, and approximately 0.2 by using Equation 1. It is 2 seconds and is shortened by about 1/100.

  The threshold value for each block is selected by the above formula 1, and the image data of the block is binarized. The threshold value setting according to Equation 1 is always set for each block. Therefore, when an image of only one of the product (inspection object) and the background is displayed in the block, one peak appears in the density histogram. In this embodiment, even in such a case, a threshold value is set. Then, even if there is no foreign object, it may be erroneously determined that there is a foreign object. Therefore, the following post-processing is performed to avoid this erroneous determination. Post-processing consists of the following two steps.

First, it is determined whether or not the circumference of the block rectangle remains above a certain level. Specifically, as shown in FIG. 4 (2), for a block in which it is determined that there is a foreign object 13, the outer periphery (broken line) of the foreign object 13 surrounded by the block is A and the outer peripheral length B of the block. . If the A / B value P1 (0 <P1 <1) is equal to or greater than the set parameter Pa, it is determined that there is no foreign matter 13 in that portion and that the background is present.
As the setting parameter Pa, a value at which the foreign matter 13 is not detected is set as a parameter based on the X-ray transmission image data of the non-defective inspection object 12 in advance.

  Next, it is determined whether or not a certain threshold value is exceeded. That is, when a threshold value is set for a non-defective image, a threshold value is set for a portion having a high density value in the density histogram due to noise or the like. This is to avoid this. Specifically, as shown in FIG. 4 (2), when the density value C of the foreign matter portion is greater than or equal to the set parameter Pb with respect to the block in which the foreign matter 13 is determined, there is no foreign matter 13 in that portion and the background Judge that there is. The setting parameter Pb is set as a parameter based on the X-ray transmission image data of the non-defective inspection object 12 in advance, as in the case of the circumference of the rectangle described above.

  The determination unit 46 determines the presence or absence of the foreign matter 13 in the inspection object 12 from the binarization result by the binarization processing unit 45. The determination unit 46 checks all the coordinate data (pixels) of the odd-numbered divided images, and determines that there is a foreign object when the majority is determined by the majority of the plurality of divided image data determined to have a foreign object.

  The sorting unit 50 is electrically connected to the image processing unit 40. The conveyance unit 50 removes the equal inspection object that moves on the conveyance conveyor 14 from the conveyance conveyor 14 based on the detection signal of the foreign matter contained in the inspection object of the image processing unit 40. As an example, the sorting unit 50 of the embodiment has a sorting arm 52 disposed on the upper surface of the transport conveyor 14. One end of the sorting arm 52 is rotatably attached to the driving means. With this configuration, the transfer arm 52 swings on the transfer conveyor 14 with one end serving as a fulcrum, and the other end pushes the inspection object on the transfer conveyor 14 out of the transfer path.

  The image display unit 60 is electrically connected to the image processing unit 40 and displays image data processed by the image processing unit 40. Specifically, the image display unit 60 displays image data of the entire inspection object 12 indicating the presence or absence of foreign matter.

Next, an inspection method of the food X-ray foreign substance inspection apparatus having the above configuration will be described below. FIG. 5 is a flowchart showing a foreign matter inspection processing procedure for food X-ray foreign matter inspection.
X-rays are irradiated from the X-ray generator 20 to the object 12 to be transported on the transport conveyor 14. X-rays that have passed through the inspection object 12 are detected by an X-ray detection unit 30 that faces the X-ray generation unit 20. The X-ray detection unit 30 outputs an electrical signal corresponding to the amount of transmitted X-rays to the data storage unit 41 of the image processing unit 40. The data storage means 41 stores the electrical signal as X-ray transmission image data.

As shown in FIG. 5, first, the X-ray transmission image data is transferred from the storage area to the calculation memory (S100), and input to the subsequent image cutout means 42.
Next, the product image is cut out based on the threshold (S110). In the X-ray transmission image data as shown in FIG. 3 (1), the image data portion of the inspection object 12 is left and a blank portion other than the inspection object 12, that is, the background is cut out (see FIG. 3 (2)). Show).

  Then, filtering processing (SUB1) of the cut out X-ray transmission image data is performed (S120). FIG. 6 is a flowchart showing the filtering process. As shown in the figure, in the filtering process (SUB1), first, the filtering process is performed twice by the median filter (S200, 210), thereby removing the abnormal value (the density value that is outstanding compared to the surrounding values).

Next, since it is not possible to remove a sudden change in density value by this alone, it is smoothed and removed by applying a smoothing filter (S220).
The Max-Min filter is applied to emphasize the foreign matter 13 (S230). Finally, the enhanced noise is removed by performing filter processing twice by the median filter (S240, 250). Note that the filter processing according to the embodiment is an example, and the number, combination, order, and the like of the filter processing can be arbitrarily set and changed according to the image data.

Next, binarization / determination processing is performed (S130). FIG. 7 is a flowchart showing binarization / determination processing.
As shown in FIG. 7, first, a memory for calculation corresponding to the number of votes is secured (S300). Next, dynamic threshold setting (SUB3) according to the number of divisions is performed. FIG. 8 is a flowchart showing the binarization process.

  As shown in FIG. 8, a plurality of divided images having different division numbers are generated for the filtered X-ray transmission image data (S400). As an example of the number of divisions, seven patterns of 3 × 3 divisions to 9 × 9 divisions can be set.

  Next, a threshold value is set for each block of the divided image having a different number of divisions (S410). A density histogram is created for each block of the divided image divided by a plurality of division numbers generated by the image dividing means 44. Then, the threshold value is statistically set using Equation 1 in the calculation process.

In the threshold setting, post-processing for avoiding erroneous determination is performed on a block that has been erroneously determined to have foreign matter despite the absence of foreign matter.
First, it is determined whether or not the circumference of the block rectangle remains above a certain level (S420). As shown in FIG. 4 (2), for a block that is determined to have a foreign object, the outer periphery (broken line) of the foreign object surrounded by the block is A and the outer peripheral length B of the block. If the A / B value P1 (0 <P1 <1) is greater than or equal to the set parameter Pa, it is determined that there is no foreign matter and that the background is the background (S440).

  On the other hand, when P1 is a value smaller than the set parameter Pa, it is determined whether or not the next step, that is, a certain threshold value or more (S430). As shown in FIG. 4 (2), if the density value C of the foreign substance portion is greater than or equal to the set parameter Pb with respect to the block determined to have foreign matters, it is determined that there is no foreign matter and that the background is the background (S440). ).

  If the density value C is lower than the setting parameter Pb in S420 and also lower than the setting parameter in S430, it is determined that there is a foreign substance (S450).

  Next, the presence or absence of foreign matter in the inspection object 12 is determined from the binarization result by the binarization processing means 45. It is determined whether the number of foreign material candidates is equal to or greater than a certain value (S320). All the coordinate data (pixels) of the odd-numbered divided images are checked, and it is determined that there is a foreign object when the majority is determined by majority of the plurality of divided image data determined to have a foreign object (S340). FIG. 9 is an explanatory diagram of the divided image data. As an example, in the case of divided image data of 3 × 3 division, 4 × 4 division, and 5 × 5 division as illustrated, it is determined that there is no foreign object 13 in the 4 × 4 divided image. However, in the 3 × 3 divided image and the 5 × 5 divided image, since it is determined that the foreign material 13 is present, it is determined by the majority that the foreign material 13 is present. On the other hand, when there are many divided images without foreign matter 13 due to majority decision, it is determined that there is no foreign matter as a result (S330).

  The image data determined to have foreign matter obtained by the image display unit 60 electrically connected to the image processing means 40 is displayed (S350). In addition, a detection signal of an inspection object determined to have foreign matter is sent to the sorting unit 50, and the corresponding inspection object 12 is removed from the conveyor 14 by the sorting arm 52.

  According to such an X-ray foreign substance inspection apparatus, the threshold setting time can be greatly shortened, and the foreign substance in the inspection object can be detected with high accuracy. Even when the inspection object changes, the threshold value is set based on the X-ray transmission image data output from the X-ray detection unit, so that it is not necessary to manually set or update the threshold value. Even in the initial stage of inspection, foreign matters are not erroneously detected. Further, the detection accuracy is improved, and even a small piece such as resin or glass can be detected with high accuracy.

It is a figure showing the composition outline of the X-ray foreign substance inspection device for foods concerning an embodiment. It is a block diagram of the X-ray foreign material inspection apparatus for foods concerning an embodiment. It is explanatory drawing of image data cut-out and filter processing. It is explanatory drawing of the image data in a binarization process. It is a flowchart which shows the foreign material inspection processing procedure of the X-ray foreign material inspection for foodstuffs. It is a flowchart which shows a filter process. It is a flowchart which shows a binarization / determination process. It is a flowchart which shows a binarization process. It is explanatory drawing of division | segmentation image data. It is a figure which shows the structure outline of the conventional X-ray foreign material inspection apparatus for foodstuffs.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ...... X-ray foreign material inspection apparatus for foodstuffs, 2 ......... Conveyor, 3 ...... X-ray generation part, 4 ...... X-ray detection part, 5 ...... Data processing part, 6 ......... Inspection object, 10... X-ray foreign matter inspection apparatus for food, 12... Inspection object, 13 ... Foreign matter, 14 ... Transfer conveyor, 20 ... X-ray generator, 30 ... X Line detection unit (line sensor) 40... Image processing unit 41... Data storage unit 42... Image cropping unit 43 43 Filter processing unit 44. ... Binarization processing means 46... Determination means 50... Sorting section 52... Sorting arm 60.

Claims (6)

In a food X-ray foreign substance inspection apparatus comprising: an X-ray generation unit that irradiates an X-ray on an inspection object; and an X-ray detection unit that detects the X-ray transmitted through the inspection object.
A plurality of divided images having different numbers of divisions are generated from the X-ray transmission image data output by the X-ray detection unit, a threshold value is calculated for each block of the divided image, and binarized to obtain 2 for each pixel of the divided image. An X-ray foreign substance inspection apparatus for food, comprising an image processing unit for determining the presence or absence of a foreign substance based on a majority result based on a valuation result. The food X-ray foreign matter inspection device according to claim 1,
The image processing unit
Storage means for storing the X-ray transmission image data output from the X-ray detector;
Division means for dividing the image region of the X-ray transmission image data into a plurality of divided images having different division numbers;
Determination means for determining foreign matter by majority based on the binarization result for each pixel of the obtained divided image;
An X-ray foreign matter inspection apparatus for food, comprising: The food X-ray foreign matter inspection device according to claim 1 or 2,
The image processing unit
An X-ray inspection apparatus for food, comprising binarization processing means for creating a density histogram for each block of the divided image, calculating a threshold value based on the density histogram, and binarizing the threshold value. In the X-ray inspection method for foods for irradiating the inspection object on the transport line, collecting X-ray transmission image data of the inspection object, and inspecting the foreign object of the inspection object,
Dividing the X-ray transmission image data into a plurality of divided images having different division numbers;
Threshold value is calculated and binarized for each block of the divided image,
An X-ray foreign matter inspection method for food, wherein the presence of foreign matter is determined by majority based on the binarization result for each pixel of the divided image. X-ray irradiates the inspection object on the transport line, collects X-ray transmission image data of the inspection object, and determines the foreign matter in the inspection object using threshold selection based on statistics In the X-ray foreign matter inspection method,
In selecting the threshold, the density histogram of the amount of X-ray transmission is divided into two classes, the foreign object and the object to be inspected, and the two classes are assumed to be a normal distribution, and a threshold value that minimizes the average misrecognition rate is calculated. In this process, when the threshold is k, the variance is σ ² _k , each density level is i, the average when the threshold is k is μ _k , and the probability that the density level appears is p _i.

An X-ray foreign matter inspection method for foods characterized by being obtained by using In the X-ray inspection method for foods for irradiating the inspection object on the transport line, collecting X-ray transmission image data of the inspection object, and inspecting the foreign object of the inspection object,
Dividing the X-ray transmission image data into a plurality of divided images having different division numbers;
For each block of the divided image, the density histogram of the amount of X-ray transmission is divided into two classes of the foreign matter and the object to be inspected, and the two classes are assumed to be a normal distribution, and the average error recognition rate is minimized. In the process of calculating the threshold value, in the threshold value calculation process, the variance calculation is performed with an approximate expression, the calculated threshold value is binarized, and the presence of the foreign matter based on the binarization result for each pixel of the divided image A method for inspecting X-ray foreign matter for food, characterized in that determination is made by majority vote.

Images