WO2020004068A1 - Inspection device, inspection system, inspection method, inspection program and recording medium - Google Patents

Inspection device, inspection system, inspection method, inspection program and recording medium Download PDF

Info

Publication number
WO2020004068A1
WO2020004068A1 PCT/JP2019/023558 JP2019023558W WO2020004068A1 WO 2020004068 A1 WO2020004068 A1 WO 2020004068A1 JP 2019023558 W JP2019023558 W JP 2019023558W WO 2020004068 A1 WO2020004068 A1 WO 2020004068A1
Authority
WO
WIPO (PCT)
Prior art keywords
unit
inspection
article
processing
learned model
Prior art date
Application number
PCT/JP2019/023558
Other languages
French (fr)
Japanese (ja)
Inventor
廣瀬　修
Original Assignee
株式会社イシダ
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 株式会社イシダ filed Critical 株式会社イシダ
Publication of WO2020004068A1 publication Critical patent/WO2020004068A1/en

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N23/00—Investigating or analysing materials by the use of wave or particle radiation, e.g. X-rays or neutrons, not covered by groups G01N3/00 – G01N17/00, G01N21/00 or G01N22/00
    • G01N23/02—Investigating or analysing materials by the use of wave or particle radiation, e.g. X-rays or neutrons, not covered by groups G01N3/00 – G01N17/00, G01N21/00 or G01N22/00 by transmitting the radiation through the material
    • G01N23/04—Investigating or analysing materials by the use of wave or particle radiation, e.g. X-rays or neutrons, not covered by groups G01N3/00 – G01N17/00, G01N21/00 or G01N22/00 by transmitting the radiation through the material and forming images of the material
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N23/00—Investigating or analysing materials by the use of wave or particle radiation, e.g. X-rays or neutrons, not covered by groups G01N3/00 – G01N17/00, G01N21/00 or G01N22/00
    • G01N23/02—Investigating or analysing materials by the use of wave or particle radiation, e.g. X-rays or neutrons, not covered by groups G01N3/00 – G01N17/00, G01N21/00 or G01N22/00 by transmitting the radiation through the material
    • G01N23/06—Investigating or analysing materials by the use of wave or particle radiation, e.g. X-rays or neutrons, not covered by groups G01N3/00 – G01N17/00, G01N21/00 or G01N22/00 by transmitting the radiation through the material and measuring the absorption
    • G01N23/18—Investigating the presence of flaws defects or foreign matter

Definitions

  • One aspect of the present invention relates to an inspection apparatus, an inspection system, an inspection method, an inspection program, and a recording medium.
  • Patent Document 1 As a conventional inspection apparatus, for example, an apparatus described in Patent Document 1 is known.
  • the inspection device described in Patent Literature 1 includes an X-ray irradiator that irradiates an X-ray to an article, and an X-ray detector that detects X-rays, and an X-ray created based on the X-ray detection result. The image is subjected to image processing to inspect the article.
  • inspection devices have performed inspections of foreign substances and the like using learned models generated by machine learning.
  • image processing based on the learned model, processing is performed on all pixels of the image to be inspected. Therefore, it takes time for the inspection, and the processing capability of the inspection apparatus may be reduced.
  • An object of one aspect of the present invention is to provide an inspection apparatus, an inspection system, an inspection method, an inspection program, and a recording medium capable of improving processing capacity.
  • An inspection device is an irradiation unit that irradiates an article with an electromagnetic wave, a detection unit that detects an electromagnetic wave transmitted through the article, and a pixel number of a transmitted image created from a detection result of the detection unit.
  • a processing unit that performs a reduction process to create a reduced image, and an acquisition unit that obtains a processing result obtained by performing a process using a learned model generated by machine learning on the reduced image created by the processing unit.
  • An inspection unit that inspects the article based on the processing result acquired by the acquisition unit.
  • the reduced image is used for the process using the learned model. Accordingly, in the inspection device, the processing load of the learned model is reduced, and the processing result can be obtained quickly. Therefore, in the inspection device, the processing capacity can be improved.
  • the inspection unit may inspect the article based on the transmission image, and inspect the article based on the determination result of the inspection and the processing result.
  • the inspection of the article is performed based on the determination result based on, for example, the threshold value determination of the luminance value based on the transmission image and the processing result.
  • the inspection device inspects the article based on the two results, so that the inspection accuracy can be improved.
  • an output section that outputs a sorting signal to a sorting apparatus that sorts the article, and the article is shaken after a transmission image is created.
  • a setting unit that sets a reduction ratio of the reduced image based on a first time required to reach the minute device and a second time required for processing by the learned model may be provided.
  • the time required for processing by the trained model changes according to the number of pixels of the reduced image, and becomes longer when the number of pixels is large and becomes shorter when the number of pixels is small. If the number of pixels is large, the processing takes a long time, and the processing may not be completed before the article reaches the distribution device.
  • the accuracy of processing by the learned model may be reduced.
  • the reduction ratio of the reduced image is set based on the first time and the second time.
  • a generator that generates a learned model by machine learning using a teacher image may be provided. With this configuration, it is possible to generate a learned model suitable for inspecting an article.
  • a learning unit that performs a process on a reduced image based on the learned model generated by the generation unit may be provided, and the obtaining unit may obtain a processing result executed by the learning unit.
  • the acquisition unit acquires the processing result obtained by inputting the reduced image to the learned model in the learning unit. Therefore, the inspection device can more reliably perform the inspection based on the processing result.
  • the generation unit may generate a learned model including a neural network.
  • the learned model can be made appropriate, and the inspection accuracy of the article can be improved.
  • the inspection unit may inspect whether or not the article contains a foreign substance.
  • the products may overlap each other. In this case, it is difficult to distinguish a portion where the products overlap each other and a foreign substance based on the luminance value of the transmission image.
  • the inspection device since the inspection is performed based on the processing result obtained by performing the processing based on the learned model, it is possible to distinguish a portion where commodities overlap each other and a foreign substance. Therefore, the inspection device is particularly effective for inspecting whether or not the article contains foreign matter.
  • the irradiation unit may irradiate the article with X-rays, and the detection unit may detect the X-ray transmitted through the article.
  • An inspection system is an inspection system including an inspection device and a server communicably connected to the inspection device, wherein the inspection device includes an irradiation unit that irradiates an article with electromagnetic waves.
  • a detection unit that detects electromagnetic waves transmitted through the article, and performs a reduction process of the number of pixels on the transmission image created from the detection result of the detection unit to create a reduced image, and transmits image information related to the reduced image to the server.
  • An inspection unit for inspecting the article based on the processed result, and the server performs a process using the learned model on the reduced image based on the image information transmitted from the inspection device.
  • a reduced image is used for processing by the learned model. Accordingly, in the inspection system, the processing load of the learned model is reduced, and the processing result can be obtained quickly. Therefore, in the inspection system, the processing capacity can be improved.
  • a generator that generates a learned model by machine learning using a teacher image may be provided. With this configuration, it is possible to generate a learned model suitable for inspecting an article.
  • An inspection method is an inspection method performed by an inspection apparatus including an irradiation unit that irradiates an article with an electromagnetic wave, and a detection unit that detects an electromagnetic wave transmitted through the article.
  • a reduced image is used for processing by the learned model.
  • the processing load of the learned model is reduced, and the processing result can be obtained quickly. Therefore, in the inspection method, the processing capability can be improved.
  • An inspection program is a computer for an inspection apparatus including an irradiation unit that irradiates an article with electromagnetic waves, and a detection unit that detects an electromagnetic wave transmitted through the article, created from a detection result of the detection unit.
  • a processing unit that performs a reduction process of the number of pixels on the transmission image to create a reduced image, and a process that performs a process using a trained model generated by machine learning on the reduced image created by the processing unit
  • An acquisition unit that acquires a result and an inspection unit that inspects an article based on the processing result acquired by the acquisition unit function.
  • a reduced image is used for processing by the learned model.
  • the processing load of the learned model is reduced, and the processing result can be obtained quickly. Therefore, in the inspection program, the processing capability can be improved.
  • a recording medium is a computer that records an inspection program to be executed by a computer of an inspection apparatus that includes an irradiation unit that irradiates an article with an electromagnetic wave and a detection unit that detects an electromagnetic wave transmitted through the article.
  • a readable recording medium wherein the computer performs a reduction process of the number of pixels on a transmission image created from a detection result of the detection unit to create a reduced image, and a processing unit created by the processing unit.
  • An acquisition unit that acquires a processing result obtained by performing a process using a learned model generated by machine learning on the reduced image, and an inspection unit that inspects an article based on the processing result acquired by the acquisition unit. Inspection programs are recorded to function.
  • the recording medium records the inspection program.
  • the reduced image is used for processing by the learned model.
  • the processing load of the learned model is reduced, and the processing result can be obtained quickly. Therefore, with the inspection program recorded on the recording medium, the processing capability can be improved.
  • processing capacity can be improved.
  • FIG. 1 is a diagram illustrating an inspection system according to one embodiment.
  • FIG. 2 is a configuration diagram of the inside of the shield box shown in FIG.
  • FIG. 3 is a diagram illustrating an example of a configuration of a control unit of the inspection device.
  • FIG. 4A is a diagram illustrating a transmitted image
  • FIG. 4B is a diagram illustrating a reduced image.
  • FIG. 5 is a diagram illustrating the distribution device.
  • FIG. 6 is a diagram illustrating an example of a neural network.
  • FIG. 7 is a flowchart showing an example of the operation of the inspection system.
  • FIG. 8 is a diagram showing an inspection program.
  • the X-ray inspection system 1 includes an X-ray inspection device 10, a distribution device 30, and a server 50.
  • the X-ray inspection system 1 inspects the article A using a learned model generated by machine learning.
  • the X-ray inspection apparatus 10 includes an apparatus main body 11, a support leg 12, a shield box 13, a transport unit 14, an X-ray irradiation unit (irradiation unit) 15, an X-ray detection unit (detection unit) 16, and a display.
  • An operation unit 17 and a control unit 18 are provided.
  • the X-ray inspection apparatus 10 generates a transmission image G1 (see FIG. 4A) of the article A while transporting the article A, and inspects the article A based on the transmission image G1 (for example, inspection of the number of stored articles, contamination of foreign matter). Inspections, missing parts inspections, cracks and chipping inspections).
  • the X-ray inspection apparatus 10 performs a foreign substance inspection of the article A.
  • the article A before the inspection is carried into the X-ray inspection apparatus 10 by the carry-in conveyor 19.
  • the article A after the inspection is carried out to the conveyor 31.
  • the article A determined to be defective (having an abnormality) by the X-ray inspection device 10 is sorted out of the production line by the sorting device 30 arranged on the conveyor 31.
  • the article A determined to be good by the X-ray inspection apparatus 10 passes through the distribution apparatus 30 as it is.
  • the apparatus main body 11 houses the control unit 18 and the like.
  • the support legs 12 support the apparatus main body 11.
  • the shield box 13 is provided on the apparatus main body 11.
  • the shield box 13 prevents leakage of X-rays to the outside.
  • An inspection area R in which the inspection of the article A by X-rays is performed is provided inside the shield box 13.
  • the shield box 13 has a carry-in port 13a and a carry-out port 13b.
  • the article A before the inspection is carried into the inspection area R from the carry-in conveyor 19 via the carry-in port 13a.
  • the article A after the inspection is carried out from the inspection area R to the conveyor 31 of the sorting device 30 via the outlet 13b.
  • Each of the carry-in entrance 13a and the carry-out exit 13b is provided with an X-ray shielding curtain (not shown) for preventing leakage of X-rays.
  • the transport unit 14 is arranged in the shield box 13.
  • the transport unit 14 transports the article A in the transport direction D from the entrance 13a to the exit 13b via the inspection area R.
  • the transport unit 14 is, for example, a belt conveyor spanned between the entrance 13a and the exit 13b.
  • the X-ray irradiator 15 is disposed inside the shield box 13.
  • the X-ray irradiator 15 irradiates the article A transported by the transporter 14 with X-rays.
  • the X-ray irradiator 15 has, for example, an X-ray tube that emits X-rays, and a collimator that spreads the X-rays emitted from the X-ray tube in a fan shape in a plane perpendicular to the transport direction D.
  • the X-ray detector 16 is arranged in the shield box 13.
  • the X-ray detection unit 16 is a line sensor including X-ray detection elements arranged one-dimensionally in a horizontal direction perpendicular to the transport direction D.
  • the X-ray detection unit 16 detects the X-ray transmitted through the article A and the transport belt of the transport unit 14.
  • the display operation unit 17 is provided in the apparatus main body 11.
  • the display operation unit 17 displays various information and receives input of various conditions.
  • the display operation unit 17 is, for example, a liquid crystal display, and displays an operation screen as a touch panel. In this case, the operator can input various conditions via the display operation unit 17.
  • the control unit 18 is arranged in the apparatus main body 11.
  • the control unit 18 controls the operation of each unit of the X-ray inspection apparatus 10.
  • the control unit 18 includes a CPU (Central Processing Unit), a ROM (Read Only Memory), a RAM (Random Access Memory), and the like.
  • the X-ray detection result of the X-ray detector 16 is input to the controller 18.
  • the control unit 18 creates a transmission image G1 based on the X-ray detection result.
  • control unit 18 includes a setting unit 20, a processing unit 21, an acquisition unit 22, an inspection unit 23, and an output unit 24.
  • the setting unit 20 sets a reduction ratio of the reduced image G2 (see FIG. 4B) created by the processing unit 21.
  • the setting unit 20 reduces the reduced image G2 based on a first time from when the transmission image G1 is created to when the article A reaches the distribution device 30 and a second time required for processing by the learned model.
  • the first time is set according to the configuration of the X-ray inspection device 10 and the distribution device 30. Specifically, the first time is the dimension (length) of the article A, the distance between the X-ray detection unit 16 and the sorting device 30, the transport speed of the transport unit 14, the transport speed of the conveyor 31, and the like. Is set as appropriate according to.
  • the second time is set according to the processing capacity of the server 50 and the like.
  • the second time is longer as the number of pixels of the image is larger.
  • the setting unit 20 sets the reduction ratio such that the second time is shorter than the first time (first time> second time).
  • the second time is, for example, 200 msec. From the viewpoint of the processing accuracy of the learned model, it is preferable that the second time is not significantly different from the first time. Thus, it is possible to set a reduction ratio at which a predetermined processing accuracy is obtained while reliably performing distribution by the distribution device 30.
  • the setting unit 20 outputs reduction information indicating the set reduction ratio to the processing unit 21.
  • the processing unit 21 performs image processing on the transmission image G1.
  • the processing unit 21 performs a reduction process of the number of pixels on the transmission image G1 to create a reduced image G2.
  • the reduction process is a process for reducing the number of pixels of the transparent image G1, and a known method can be used.
  • the processing unit 21 creates the reduced image G2 by using, for example, the nearest neighbor method, the bilinear method, or the bicubic method.
  • the processing unit 21 creates a reduced image G2 based on the reduction ratio set by the setting unit 20.
  • the processing unit 21 may perform various processes such as contrast adjustment, color change, and format change on the transmission image G1.
  • the processing unit 21 transmits image information relating to the reduced image G2 to the server 50 via a communication unit (not shown).
  • the acquisition unit 22 acquires a processing result obtained by performing a process using a learned model generated by machine learning.
  • the acquiring unit 22 receives the processing information transmitted from the server 50 according to the image information transmitted from the processing unit 21 and acquires a processing result.
  • the acquisition unit 22 outputs the acquired processing result to the inspection unit 23.
  • the inspection unit 23 inspects the article A.
  • the inspection unit 23 determines whether or not the article A includes the foreign matter F based on the processing result obtained by the obtaining unit 22 and the determination result based on the transmission image G1.
  • the inspection unit 23 performs processing (image processing) on the transmission image G1 using an image processing algorithm set for each of a plurality of sensitivity levels (for example, 1 to 7) to generate a processed image.
  • the image processing algorithm is a type indicating a processing procedure of image processing performed on the transmission image G1.
  • the image processing algorithm is configured by one image processing filter or a combination of a plurality of image processing filters.
  • the plurality of image processing algorithms can be obtained from outside via a network such as the Internet. Further, the plurality of image processing algorithms can be obtained from an external storage medium such as a USB memory or a removable hard disk.
  • GA Genetic @ Algorithms
  • the inspection unit 23 performs predetermined image processing on the transmission image G1 based on the image processing algorithm, and determines whether or not the article A contains the foreign matter F based on the determination image obtained by performing the image processing. .
  • the predetermined image processing is, for example, a binarization processing of the transmission image G1.
  • the inspection unit 23 determines that the foreign matter F is included in the article A when the luminance value exceeds the threshold value in the determination image.
  • the threshold value is appropriately set by a test or the like according to the properties of the article A.
  • the inspection unit 23 stores the determination result in a storage unit (not shown).
  • the inspection unit 23 inspects whether or not the article A contains the foreign matter F based on the processing result obtained by the obtaining unit 22 and the above determination result. The inspection unit 23 determines that the foreign material F is not included in the article A when the foreign matter F is not included in the article A in the processing result and the foreign matter F is not included in the determination result. judge. The inspection unit 23 determines that the foreign matter F is included in the article A when the foreign matter F is included in the article A in the processing result and the foreign matter F is included in the determination result. I do. The inspection unit 23 determines the processing result when the foreign substance F is included in the article A in one of the processing result and the determination result and the foreign substance F is not included in the other processing result and the determination result. Is adopted.
  • the inspection unit 23 outputs inspection information indicating an inspection result to the output unit 24.
  • the output unit 24 outputs various signals based on the test information output from the test unit 23.
  • the output unit 24 outputs a display signal for displaying the inspection result on the display operation unit 17 to the display operation unit 17 based on the inspection information.
  • the output unit 24 outputs a distribution signal instructing the distribution device 30 to distribute the article A.
  • the distribution device 30 is provided downstream of the X-ray inspection device 10.
  • the distribution device 30 is provided on the conveyor 31.
  • the sorting device 30 sorts the articles A based on the sorting signal output from the X-ray inspection device 10.
  • the distribution device 30 has a photoelectric sensor 32 and an arm 33.
  • the photoelectric sensor 32 is a sensor that detects the passage of the article A.
  • the photoelectric sensor 32 is installed on the upstream side of the arm 33 and detects the entry of the article A into the distribution device 30.
  • the photoelectric sensor 32 has a light projector 32a and a light receiver 32b. When the light emitted from the light emitter 32a to the light receiver 32b is blocked by the article A, the photoelectric sensor 32 transmits a signal to the X-ray inspection apparatus 10 assuming that the article A has reached the distribution device 30.
  • the tip of the arm 33 swings around the base end by a driving force of, for example, a motor.
  • the arm 33 pushes the article A to one side in the width direction of the conveyor 31 and sorts the article A out of the production line.
  • the server 50 is a device that generates a learned model by machine learning and performs processing using the learned model.
  • the server 50 includes a CPU (Central Processing Unit), a ROM (Read Only Memory), a RAM (Random Access Memory), and the like.
  • the server 50 is managed by a user of the X-ray inspection apparatus 10, for example. As shown in FIG. 1, the X-ray inspection apparatus 10 and the server 50 are communicably connected to each other by a wired or wireless network N such as the Internet or a telephone network, and can transmit and receive information to and from each other. .
  • a wired or wireless network N such as the Internet or a telephone network
  • the server 50 includes a communication unit 51, a learned model generation unit 52, and a learning unit 53.
  • the communication unit 51 communicates with the X-ray inspection apparatus 10.
  • the communication unit 51 receives the image information transmitted from the X-ray inspection apparatus 10 and outputs the image information to the learning unit 53.
  • the communication unit 51 transmits the processing information output from the learning unit 53 to the X-ray inspection device 10.
  • the learned model generation unit 52 acquires learning data used for machine learning, performs machine learning using the acquired learning data, and generates a learned model.
  • the learning data includes a teacher image and other data.
  • the teacher image is, for example, a transmission image acquired by the X-ray inspection apparatus 10, such as a transmission image including the foreign matter F and a transmission image not including the foreign matter F.
  • Other data include, for example, in a transmission image acquired by the X-ray inspection apparatus 10, when a plurality of products are included in the article A, data on an area where the products overlap each other, and data included in the article A. And the like regarding the region where the foreign matter F exists.
  • the learned model generation unit 52 performs machine learning by using each pixel value of the teacher image as an input value to the neural network and processing information corresponding to the teacher image as an output value of the neural network to perform the neural network NW (FIG. 6). See).
  • NW neural network
  • the learned model outputs information indicating the presence or absence of the foreign matter F with respect to the reduced image G2.
  • the trained model includes a neural network NW.
  • the trained model may include a convolutional neural network. Further, the trained model may include a neural network of a plurality of layers (for example, eight or more layers). That is, a learned model may be generated by deep learning.
  • the neural network NW is, for example, a first layer that is an input layer, second, third, and fourth layers that are intermediate layers (hidden layers), and an output layer. And a fifth layer.
  • Each of the second, third, and fourth layers converts the total input into an output by the activation function and passes the output to the next layer.
  • the neural network NW inputs the pixel value of each pixel of the reduced image G2 and outputs information indicating the presence or absence of the foreign matter F.
  • the input layers of the neural network NW are provided with neurons for the number of pixels of the reduced image G2.
  • the output layer of the neural network NW is provided with neurons for outputting information indicating the presence or absence of the foreign matter F.
  • Information indicating the presence or absence of the foreign matter F can be determined based on the output value of the neuron in the output layer.
  • the output value of the neuron is, for example, a value between 0 and 1.
  • the value of the neuron is larger (the value is closer to 1), the foreign matter F is contained in the article A (it is abnormal), and as the value of the neuron is smaller (the value is closer to 0), the article is A indicates that the foreign matter F is not included in A (normal).
  • the learning unit 53 inputs the reduced image G2 based on the image information output from the communication unit 51 to the learned model.
  • the learning unit 53 outputs, to the communication unit 51, processing information indicating a processing result including an output value output from the neural network NW of the learned model.
  • the processing result includes information indicating the presence or absence of the foreign matter F in the article A, and information indicating an area (position) in the article A where the foreign matter F is included when the article A includes the foreign matter F. Have been.
  • the X-ray inspection apparatus 10 creates a transmission image G1 from the detection result of the X-ray detection unit 16 (Step S01). Subsequently, the X-ray inspection apparatus 10 performs a reduction process on the transmission image G1 to create a reduced image G2 (step S02, processing step). The X-ray inspection apparatus 10 transmits the reduced image G2 to the server 50 (Step S03). Further, the X-ray inspection apparatus 10 determines whether or not the foreign material F is included in the article A based on the transmission image G1 (Step S04).
  • the server 50 performs a process using the learned model on the reduced image G2 transmitted from the X-ray inspection apparatus 10 (Step S05).
  • the server 50 transmits the processing result based on the learned model to the X-ray inspection apparatus 10 (Step S06).
  • the X-ray inspection apparatus 10 acquires the processing result transmitted from the server 50 (acquisition step), and performs an inspection as to whether or not the article A contains the foreign matter F based on the processing result (step). S07, inspection step).
  • the X-ray inspection apparatus 10 transmits a distribution signal to the distribution apparatus 30 (step S08). If the X-ray inspection apparatus 10 does not determine that the foreign matter F is included in the article A, the processing ends.
  • the inspection program P can be recorded on a computer-readable recording medium 100.
  • the inspection program P stored in the recording medium 100 includes a setting module P1, a processing module P2, an acquisition module P3, an inspection module P4, and an output module P5.
  • the inspection program P functions by causing the computer to execute the setting module P1, the processing module P2, the acquisition module P3, the inspection module P4, and the output module P5.
  • the functions realized by executing the setting module P1, the processing module P2, the acquisition module P3, the inspection module P4, and the output module P5 are respectively a setting unit 20, a processing unit 21, an acquisition unit 22, an inspection unit 23, and an output unit 24. Is the same as the function of
  • the inspection program P is recorded in a program recording area of the recording medium 100.
  • the recording medium 100 is configured by a recording medium such as a CD-ROM, a DVD, a ROM, and a semiconductor memory.
  • the inspection program P may be provided via a communication network as a computer data signal superimposed on a carrier wave.
  • the reduced image G2 is used for the process using the learned model.
  • the processing load of the learned model is reduced, and the processing result can be obtained quickly. Therefore, in the X-ray inspection system 1, the processing capacity can be improved.
  • the processing capability can be improved by using the reduced image G2 for the process using the learned model. Therefore, in the X-ray inspection system 1, the processing capability can be improved without increasing the performance of the server 50 (such as the specifications of the CPU). Therefore, the X-ray inspection system 1 does not require an expensive device, so that the cost can be reduced.
  • the inspection unit 23 of the X-ray inspection apparatus 10 inspects the article A based on the transmission image G1, and based on the determination result of the inspection and the processing result, The inspection of the article A is performed.
  • the inspection of the article A is performed based on the determination result based on the threshold value determination of the luminance value based on the transmission image G1 and the processing result.
  • the inspection accuracy can be improved.
  • the X-ray inspection apparatus 10 outputs a distribution signal to the distribution apparatus 30 that distributes the article A when the inspection unit 23 detects an abnormality of the article.
  • the output unit 24 that outputs the reduced image G2 based on the first time from when the transmission image G1 is created to when the article reaches the distribution device 30 and the second time that is required for processing by the learned model.
  • the processing takes a long time, and the processing may not be completed before the article A reaches the distribution device 30.
  • the accuracy of processing by the learned model may be reduced.
  • the reduction ratio of the reduced image G2 is set based on the first time and the second time.
  • the server 50 includes the learned model generation unit 52 that generates a learned model by machine learning using a teacher image. With this configuration, it is possible to generate a learned model suitable for inspecting an article.
  • the learned model generation unit 52 generates a learned model including a neural network. With this configuration, the learned model can be made appropriate, and the inspection accuracy of the article can be improved.
  • the inspection unit 23 of the X-ray inspection apparatus 10 inspects whether or not the article A contains the foreign matter F.
  • the article A includes a plurality of products, the products may overlap each other. In this case, it is difficult to discriminate a portion where commodities overlap each other and a foreign substance based on the luminance value of the transmission image G1.
  • the X-ray inspection apparatus 10 since the inspection is performed based on the processing result obtained by performing the processing using the learned model, it is possible to distinguish a portion where the products overlap each other and a foreign substance. Therefore, the X-ray inspection apparatus 10 is particularly effective in inspecting whether or not the article A contains the foreign matter F.
  • the X-ray detection unit 16 of the X-ray inspection apparatus 10 includes one line sensor.
  • the X-ray detector 16 may include two line sensors. In this configuration, the two line sensors are arranged to face each other in the vertical direction.
  • the (upper) line sensor detects X-rays in the low energy band transmitted through the conveyance belt of the article A and the conveyance unit 14.
  • the other (lower) line sensor detects X-rays in the high energy band transmitted through the article A, the conveyor belt of the conveyor 14 and one of the line sensors.
  • the inspection unit 23 of the X-ray inspection apparatus 10 inspects the article A based on the determination result based on the transmission image G1 and the processing result obtained by performing the processing using the learned model generated by the machine learning.
  • the embodiment for performing the above has been described as an example.
  • the inspection unit 23 may inspect the article A based on only the processing result of the processing using the learned model.
  • the X-ray inspection system 1 includes the X-ray inspection apparatus 10 and the server 50 .
  • the server 50 need not be provided.
  • the X-ray inspection apparatus 10 only needs to include the learned model generation unit and the learning unit.
  • the X-ray inspection apparatus 10 may acquire a learned model generated by another apparatus (computer) and store it in the storage unit.
  • the X-ray inspection system 1 includes the X-ray inspection device 10 and the distribution device 30 has been described.
  • the distribution device 30 may be included as a part of the configuration of the X-ray inspection device 10.
  • the server 50 may not include the learned model generation unit 52.
  • the server 50 may acquire the learned model generated by another device, store the acquired model in the storage unit, and perform processing on the reduced image G2 using the learned model stored in the storage unit. .
  • the learned model generation unit 52 acquires learning data used for machine learning and performs machine learning using the acquired learning data to generate a learned model.
  • the generation method of the learned model is not limited to this.
  • the learning data other data may be used in addition to the teacher image and other data.
  • the neural network NW has five layers (four layers when the input layer is excluded) has been described as an example.
  • the number of layers of the neural network constituting the learned model generation unit 52 is not limited at all.
  • the learned model generation unit 52 may use a neural network having an arbitrary number of layers of three or more, which may use a neural network having an intermediate number of one or more layers. means.
  • the configuration of each layer of the neural network (for example, the number of neurons) and the connection between neurons are not limited to the configuration described in the above embodiment.
  • the inspection apparatus is the X-ray inspection apparatus 10
  • the inspection apparatus is not limited to the X-ray inspection apparatus, and may be any inspection apparatus that inspects an article using electromagnetic waves. That is, in the present invention, the electromagnetic waves are X-rays, near infrared rays, light, and other electromagnetic waves.
  • the present invention is not limited to the method for inspecting for the presence or absence of foreign matter contained in an article, but is not limited to a package such as a film packaging material and the like, in which the content such as food is stored and shipped. The inspection may be performed to check for biting of the contents into the section, damage to the contents in the package, and entry of foreign matter into the package.
  • the type of the article is not particularly limited, and various articles can be inspected.
  • the type of foreign matter is not particularly limited, and various foreign matters can be inspected.
  • SYMBOLS 1 ... X-ray inspection system, 10 ... X-ray inspection apparatus, 15 ... X-ray irradiation part (irradiation part), 16 ... X-ray detection part (detection part), 20 ... setting part, 21 ... processing part, 22 ... acquisition part Reference numerals 23, inspection unit, 24, output unit, 50, server, 51, communication unit, 52, learned model generation unit, 100, recording medium, A, article, F, foreign matter, P, inspection program, P2, processing module (Processing unit), P3: acquisition module (acquisition unit), P4: inspection module (inspection unit).

Abstract

Provided is an X-ray inspection device 10 comprising: an X-ray irradiation unit 15 that irradiates an article A with X-rays; an X-ray detection unit 16 that detects X-rays transmitted through the article A; a processing unit 21 that carries out pixel number reduction processing on a transparent image G1 that is created from the detection result of the X-ray detection unit 16 and creates a reduced image G2; an acquisition unit 22 that acquires a processing result of performing processing by a learned model generated by machine learning on the reduced image G2 that is generated by the processing unit 21; and an inspection unit 23 that inspects the article A on the basis of the processing result that is acquired by the acquisition unit 22.

Description

検査装置、検査システム、検査方法、検査プログラム及び記録媒体Inspection device, inspection system, inspection method, inspection program, and recording medium
　本発明の一側面は、検査装置、検査システム、検査方法、検査プログラム及び記録媒体に関する。 One aspect of the present invention relates to an inspection apparatus, an inspection system, an inspection method, an inspection program, and a recording medium.
　従来の検査装置として、例えば、特許文献１に記載された装置が知られている。特許文献１に記載の検査装置は、物品にＸ線を照射するＸ線照射部と、Ｘ線を検出するＸ線検出部と、を備え、Ｘ線の検出結果に基づいて作成されたＸ線画像に対して画像処理を施して物品の検査を行う。 装置 As a conventional inspection apparatus, for example, an apparatus described in Patent Document 1 is known. The inspection device described in Patent Literature 1 includes an X-ray irradiator that irradiates an X-ray to an article, and an X-ray detector that detects X-rays, and an X-ray created based on the X-ray detection result. The image is subjected to image processing to inspect the article.
国際公開第２００６／００１１０７号International Publication No. WO 2006/001107
　近年、検査装置では、機械学習により生成された学習済モデルを用いて、異物の検査等を実施している。学習済モデルに基づく画像処理では、検査対象の画像の全ての画素に対して処理を行う。そのため、検査に時間を要し、検査装置の処理能力が低下し得る。 In recent years, inspection devices have performed inspections of foreign substances and the like using learned models generated by machine learning. In the image processing based on the learned model, processing is performed on all pixels of the image to be inspected. Therefore, it takes time for the inspection, and the processing capability of the inspection apparatus may be reduced.
　本発明の一側面は、処理能力の向上を図ることができる検査装置、検査システム、検査方法、検査プログラム及び記録媒体を提供することを目的とする。 An object of one aspect of the present invention is to provide an inspection apparatus, an inspection system, an inspection method, an inspection program, and a recording medium capable of improving processing capacity.
　本発明の一側面に係る検査装置は、物品に電磁波を照射する照射部と、物品を透過した電磁波を検出する検出部と、検出部の検出結果から作成された透過画像に対して画素数の縮小処理を施して、縮小画像を作成する処理部と、処理部によって作成された縮小画像に対して、機械学習によって生成された学習済モデルによる処理を行った処理結果を取得する取得部と、取得部によって取得された処理結果に基づいて、物品の検査を行う検査部と、を備える。 An inspection device according to one aspect of the present invention is an irradiation unit that irradiates an article with an electromagnetic wave, a detection unit that detects an electromagnetic wave transmitted through the article, and a pixel number of a transmitted image created from a detection result of the detection unit. A processing unit that performs a reduction process to create a reduced image, and an acquisition unit that obtains a processing result obtained by performing a process using a learned model generated by machine learning on the reduced image created by the processing unit. An inspection unit that inspects the article based on the processing result acquired by the acquisition unit.
　本発明の一側面に係る検査装置では、学習済モデルによる処理に縮小画像を用いる。これにより、検査装置では、学習済モデルによる処理負荷が小さくなるため、処理結果を迅速に得ることが可能となる。したがって、検査装置では、処理能力の向上を図ることができる。 検 査 In the inspection device according to one aspect of the present invention, the reduced image is used for the process using the learned model. Accordingly, in the inspection device, the processing load of the learned model is reduced, and the processing result can be obtained quickly. Therefore, in the inspection device, the processing capacity can be improved.
　一実施形態においては、検査部は、透過画像に基づいて物品の検査を行い、当該検査の判定結果と、処理結果とに基づいて、物品の検査を行ってもよい。この構成では、透過画像に基づく、例えば輝度値の閾値判定による判定結果と、処理結果とに基づいて、物品の検査を行う。このように、検査装置では、２つの結果に基づいて物品を検査するため、検査精度の向上が図れる。 In one embodiment, the inspection unit may inspect the article based on the transmission image, and inspect the article based on the determination result of the inspection and the processing result. In this configuration, the inspection of the article is performed based on the determination result based on, for example, the threshold value determination of the luminance value based on the transmission image and the processing result. As described above, the inspection device inspects the article based on the two results, so that the inspection accuracy can be improved.
　一実施形態においては、検査部において物品の異常が検出された場合に、物品の振り分けを行う振分装置に対して振分信号を出力する出力部と、透過画像が作成されてから物品が振分装置に到達するまでの第１時間と、学習済モデルによる処理に要する第２時間とに基づいて、縮小画像の縮小率を設定する設定部と、を備えていてもよい。学習済モデルによる処理に要する時間は、縮小画像の画素数に応じて変化し、画素数が大きい場合には長くなり、画素数が小さい場合には短くなる。画素数が大きい場合、処理に時間がかかるため、物品が振分装置に到達するまでに、処理が完了しないおそれがある。一方で、画素数が小さすぎると、学習済モデルによる処理の精度が低下し得る。検査装置では、第１時間と第２時間とに基づいて縮小画像の縮小率を設定する。これにより、検査装置では、振分装置による振り分けを確実に行わせることができると共に、学習済モデルによる処理精度の低下を抑制できる。 In one embodiment, when an abnormality of an article is detected by the inspection section, an output section that outputs a sorting signal to a sorting apparatus that sorts the article, and the article is shaken after a transmission image is created. A setting unit that sets a reduction ratio of the reduced image based on a first time required to reach the minute device and a second time required for processing by the learned model may be provided. The time required for processing by the trained model changes according to the number of pixels of the reduced image, and becomes longer when the number of pixels is large and becomes shorter when the number of pixels is small. If the number of pixels is large, the processing takes a long time, and the processing may not be completed before the article reaches the distribution device. On the other hand, if the number of pixels is too small, the accuracy of processing by the learned model may be reduced. In the inspection device, the reduction ratio of the reduced image is set based on the first time and the second time. Thereby, in the inspection device, the distribution by the distribution device can be reliably performed, and the decrease in processing accuracy due to the learned model can be suppressed.
　一実施形態においては、教師画像を用いる機械学習によって学習済モデルを生成する生成部を備えていてもよい。この構成では、物品の検査に適切な学習済モデルを生成できる。 In one embodiment, a generator that generates a learned model by machine learning using a teacher image may be provided. With this configuration, it is possible to generate a learned model suitable for inspecting an article.
　一実施形態においては、生成部によって生成された学習済モデルにより、縮小画像に対して処理を行う学習部を備え、取得部は、学習部によって実行された処理結果を取得してもよい。この構成では、学習部において縮小画像を学習済モデルに入力して得られた処理結果を、取得部が取得する。したがって、検査装置では、処理結果に基づく検査をより確実に行うことができる。 In one embodiment, a learning unit that performs a process on a reduced image based on the learned model generated by the generation unit may be provided, and the obtaining unit may obtain a processing result executed by the learning unit. In this configuration, the acquisition unit acquires the processing result obtained by inputting the reduced image to the learned model in the learning unit. Therefore, the inspection device can more reliably perform the inspection based on the processing result.
　一実施形態においては、生成部は、ニューラルネットワークを含む学習済モデルを生成してもよい。この構成では、学習済モデルを適切なものとすることができ、物品の検査精度の向上が図れる。 生成 In one embodiment, the generation unit may generate a learned model including a neural network. With this configuration, the learned model can be made appropriate, and the inspection accuracy of the article can be improved.
　一実施形態においては、検査部は、物品に異物が含まれているか否かを検査してもよい。物品に複数の商品が含まれている場合には、商品同士が重なることがある。この場合、商品同士が重なっている部分と、異物との区別を、透過画像の輝度値に基づいて行うことが困難となる。検査装置では、学習済モデルによる処理を行った処理結果に基づいて検査するため、商品同士が重なっている部分と、異物との区別を行うことが可能である。したがって、検査装置は、物品に異物が含まれているか否かの検査に特に有効である。 In one embodiment, the inspection unit may inspect whether or not the article contains a foreign substance. When a product includes a plurality of products, the products may overlap each other. In this case, it is difficult to distinguish a portion where the products overlap each other and a foreign substance based on the luminance value of the transmission image. In the inspection device, since the inspection is performed based on the processing result obtained by performing the processing based on the learned model, it is possible to distinguish a portion where commodities overlap each other and a foreign substance. Therefore, the inspection device is particularly effective for inspecting whether or not the article contains foreign matter.
　一実施形態においては、照射部は、物品にＸ線を照射し、検出部は、物品を透過したＸ線を検出してもよい。この構成では、物品が包装されている場合であっても、包材や、包材に施された印刷に影響されることなく、物品の検査することができる。 In one embodiment, the irradiation unit may irradiate the article with X-rays, and the detection unit may detect the X-ray transmitted through the article. With this configuration, even if the article is packaged, the article can be inspected without being affected by the packaging material or the printing applied to the packaging material.
　本発明の一側面に係る検査システムは、検査装置と、当該検査装置と通信可能に接続されているサーバと、を備える検査システムであって、検査装置は、物品に電磁波を照射する照射部と、物品を透過した電磁波を検出する検出部と、検出部の検出結果から作成された透過画像に対して画素数の縮小処理を施して縮小画像を作成し、縮小画像に係る画像情報をサーバに送信する処理部と、処理部からサーバに画像情報を送信したことに応じて、機械学習によって生成された学習済モデルによる処理を行った処理結果をサーバから取得する取得部と、取得部によって取得された処理結果に基づいて、物品の検査を行う検査部と、を備え、サーバは、検査装置から送信された画像情報に基づいて、縮小画像に対して学習済モデルによる処理を行う学習部と、処理結果を検査装置に送信する通信部と、を備える。 An inspection system according to an aspect of the present invention is an inspection system including an inspection device and a server communicably connected to the inspection device, wherein the inspection device includes an irradiation unit that irradiates an article with electromagnetic waves. A detection unit that detects electromagnetic waves transmitted through the article, and performs a reduction process of the number of pixels on the transmission image created from the detection result of the detection unit to create a reduced image, and transmits image information related to the reduced image to the server. A processing unit for transmitting, an obtaining unit for obtaining from the server a processing result obtained by performing a process using the learned model generated by machine learning in response to transmitting the image information from the processing unit to the server, and obtaining by the obtaining unit An inspection unit for inspecting the article based on the processed result, and the server performs a process using the learned model on the reduced image based on the image information transmitted from the inspection device. Comprising a learning unit, a communication unit that transmits the processing result to the inspection device.
　本発明の一側面に係る検査システムでは、学習済モデルによる処理に縮小画像を用いる。これにより、検査システムでは、学習済モデルによる処理負荷が小さくなるため、処理結果を迅速に得ることが可能となる。したがって、検査システムでは、処理能力の向上を図ることができる。 検 査 In the inspection system according to one aspect of the present invention, a reduced image is used for processing by the learned model. Accordingly, in the inspection system, the processing load of the learned model is reduced, and the processing result can be obtained quickly. Therefore, in the inspection system, the processing capacity can be improved.
　一実施形態においては、教師画像を用いる機械学習によって学習済モデルを生成する生成部を備えていてもよい。この構成では、物品の検査に適切な学習済モデルを生成できる。 In one embodiment, a generator that generates a learned model by machine learning using a teacher image may be provided. With this configuration, it is possible to generate a learned model suitable for inspecting an article.
　本発明の一側面に係る検査方法は、物品に電磁波を照射する照射部と、物品を透過した電磁波を検出する検出部と、を備える検査装置で実行される検査方法であって、検出部の検出結果から作成された透過画像に対して画素数の縮小処理を施して、縮小画像を作成する処理ステップと、処理ステップにおいて作成された縮小画像に対して、機械学習によって生成された学習済モデルによる処理を行った処理結果を取得する取得ステップと、取得ステップにおいて取得された処理結果に基づいて、物品の検査を行う検査ステップと、を含む。 An inspection method according to one aspect of the present invention is an inspection method performed by an inspection apparatus including an irradiation unit that irradiates an article with an electromagnetic wave, and a detection unit that detects an electromagnetic wave transmitted through the article. A processing step of performing a reduction process of the number of pixels on the transparent image created from the detection result to create a reduced image; and a trained model generated by machine learning on the reduced image created in the processing step. And an inspection step of inspecting an article based on the processing result acquired in the acquiring step.
　本発明の一側面に係る検査方法では、学習済モデルによる処理に縮小画像を用いる。これにより、検査方法では、学習済モデルによる処理負荷が小さくなるため、処理結果を迅速に得ることが可能となる。したがって、検査方法では、処理能力の向上を図ることができる。 検 査 In the inspection method according to one aspect of the present invention, a reduced image is used for processing by the learned model. As a result, in the inspection method, the processing load of the learned model is reduced, and the processing result can be obtained quickly. Therefore, in the inspection method, the processing capability can be improved.
　本発明の一側面に係る検査プログラムは、物品に電磁波を照射する照射部と、物品を透過した電磁波を検出する検出部と、を備える検査装置のコンピュータを、検出部の検出結果から作成された透過画像に対して画素数の縮小処理を施して、縮小画像を作成する処理部と、処理部によって作成された縮小画像に対して、機械学習によって生成された学習済モデルによる処理を行った処理結果を取得する取得部と、取得部によって取得された処理結果に基づいて、物品の検査を行う検査部と、して機能させる。 An inspection program according to one aspect of the present invention is a computer for an inspection apparatus including an irradiation unit that irradiates an article with electromagnetic waves, and a detection unit that detects an electromagnetic wave transmitted through the article, created from a detection result of the detection unit. A processing unit that performs a reduction process of the number of pixels on the transmission image to create a reduced image, and a process that performs a process using a trained model generated by machine learning on the reduced image created by the processing unit An acquisition unit that acquires a result and an inspection unit that inspects an article based on the processing result acquired by the acquisition unit function.
　本発明の一側面に係る検査プログラムでは、学習済モデルによる処理に縮小画像を用いる。これにより、検査プログラムでは、学習済モデルによる処理負荷が小さくなるため、処理結果を迅速に得ることが可能となる。したがって、検査プログラムでは、処理能力の向上を図ることができる。 検 査 In the inspection program according to one aspect of the present invention, a reduced image is used for processing by the learned model. Thus, in the inspection program, the processing load of the learned model is reduced, and the processing result can be obtained quickly. Therefore, in the inspection program, the processing capability can be improved.
　本発明の一側面に係る記録媒体は、物品に電磁波を照射する照射部と、物品を透過した電磁波を検出する検出部と、を備える検査装置のコンピュータに実行させる検査プログラムを記録しているコンピュータ読取可能な記録媒体であって、コンピュータを、検出部の検出結果から作成された透過画像に対して画素数の縮小処理を施して、縮小画像を作成する処理部と、処理部によって作成された縮小画像に対して、機械学習によって生成された学習済モデルによる処理を行った処理結果を取得する取得部と、取得部によって取得された処理結果に基づいて、物品の検査を行う検査部と、して機能させる検査プログラムを記録している。 A recording medium according to one aspect of the present invention is a computer that records an inspection program to be executed by a computer of an inspection apparatus that includes an irradiation unit that irradiates an article with an electromagnetic wave and a detection unit that detects an electromagnetic wave transmitted through the article. A readable recording medium, wherein the computer performs a reduction process of the number of pixels on a transmission image created from a detection result of the detection unit to create a reduced image, and a processing unit created by the processing unit. An acquisition unit that acquires a processing result obtained by performing a process using a learned model generated by machine learning on the reduced image, and an inspection unit that inspects an article based on the processing result acquired by the acquisition unit. Inspection programs are recorded to function.
　本発明の一側面に係る記録媒体では、上記検査プログラムを記録している。検査プログラムでは、学習済モデルによる処理に縮小画像を用いる。これにより、検査プログラムでは、学習済モデルによる処理負荷が小さくなるため、処理結果を迅速に得ることが可能となる。したがって、記録媒体に記録されている検査プログラムでは、処理能力の向上を図ることができる。 記録 The recording medium according to one aspect of the present invention records the inspection program. In the inspection program, the reduced image is used for processing by the learned model. Thus, in the inspection program, the processing load of the learned model is reduced, and the processing result can be obtained quickly. Therefore, with the inspection program recorded on the recording medium, the processing capability can be improved.
　本発明の一側面によれば、処理能力の向上を図ることができる。 According to one aspect of the present invention, processing capacity can be improved.
図１は、一実施形態に係る検査システムを示す図である。FIG. 1 is a diagram illustrating an inspection system according to one embodiment. 図２は、図１に示されるシールドボックスの内部の構成図である。FIG. 2 is a configuration diagram of the inside of the shield box shown in FIG. 図３は、検査装置の制御部の構成の一例を示す図である。FIG. 3 is a diagram illustrating an example of a configuration of a control unit of the inspection device. 図４（ａ）は、透過画像を示す図であり、図４（ｂ）は、縮小画像を示す図である。FIG. 4A is a diagram illustrating a transmitted image, and FIG. 4B is a diagram illustrating a reduced image. 図５は、振分装置を示す図である。FIG. 5 is a diagram illustrating the distribution device. 図６は、ニューラルネットワークの一例を示す図である。FIG. 6 is a diagram illustrating an example of a neural network. 図７は、検査システムの動作の一例を示すフロチャートである。FIG. 7 is a flowchart showing an example of the operation of the inspection system. 図８は、検査プログラムを示す図である。FIG. 8 is a diagram showing an inspection program.
　以下、添付図面を参照して、本発明の好適な実施形態について詳細に説明する。なお、図面の説明において同一又は相当要素には同一符号を付し、重複する説明は省略する。 Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. In the description of the drawings, the same or corresponding elements will be denoted by the same reference symbols, without redundant description.
　図１に示されるように、Ｘ線検査システム１は、Ｘ線検査装置１０と、振分装置３０と、サーバ５０と、を備えている。Ｘ線検査システム１は、機械学習によって生成される学習済モデルを用いて物品Ａの検査を行う。 As shown in FIG. 1, the X-ray inspection system 1 includes an X-ray inspection device 10, a distribution device 30, and a server 50. The X-ray inspection system 1 inspects the article A using a learned model generated by machine learning.
　Ｘ線検査装置１０は、装置本体１１と、支持脚１２と、シールドボックス１３と、搬送部１４と、Ｘ線照射部（照射部）１５と、Ｘ線検出部（検出部）１６と、表示操作部１７と、制御部１８と、を備えている。 The X-ray inspection apparatus 10 includes an apparatus main body 11, a support leg 12, a shield box 13, a transport unit 14, an X-ray irradiation unit (irradiation unit) 15, an X-ray detection unit (detection unit) 16, and a display. An operation unit 17 and a control unit 18 are provided.
　Ｘ線検査装置１０は、物品Ａを搬送しつつ物品Ａの透過画像Ｇ１（図４（ａ）参照）を生成し、透過画像Ｇ１に基づいて物品Ａの検査（例えば、収納数検査、異物混入検査、欠品検査、割れ欠け検査等）を行う。本実施形態では、Ｘ線検査装置１０は、物品Ａの異物混入検査を行う。検査前の物品Ａは、搬入コンベア１９によってＸ線検査装置１０に搬入される。検査後の物品Ａは、コンベア３１に搬出される。Ｘ線検査装置１０によって不良品（異常が有る）と判定された物品Ａは、コンベア３１に配置された振分装置３０よって生産ライン外に振り分けられる。Ｘ線検査装置１０によって良品と判定された物品Ａは、振分装置３０をそのまま通過する。 The X-ray inspection apparatus 10 generates a transmission image G1 (see FIG. 4A) of the article A while transporting the article A, and inspects the article A based on the transmission image G1 (for example, inspection of the number of stored articles, contamination of foreign matter). Inspections, missing parts inspections, cracks and chipping inspections). In the present embodiment, the X-ray inspection apparatus 10 performs a foreign substance inspection of the article A. The article A before the inspection is carried into the X-ray inspection apparatus 10 by the carry-in conveyor 19. The article A after the inspection is carried out to the conveyor 31. The article A determined to be defective (having an abnormality) by the X-ray inspection device 10 is sorted out of the production line by the sorting device 30 arranged on the conveyor 31. The article A determined to be good by the X-ray inspection apparatus 10 passes through the distribution apparatus 30 as it is.
　装置本体１１は、制御部１８等を収容している。支持脚１２は、装置本体１１を支持している。シールドボックス１３は、装置本体１１に設けられている。シールドボックス１３は、外部へのＸ線の漏洩を防止する。シールドボックス１３の内部には、Ｘ線による物品Ａの検査が実施される検査領域Ｒが設けられている。シールドボックス１３には、搬入口１３ａ及び搬出口１３ｂが形成されている。検査前の物品Ａは、搬入コンベア１９から搬入口１３ａを介して検査領域Ｒに搬入される。検査後の物品Ａは、検査領域Ｒから搬出口１３ｂを介して振分装置３０のコンベア３１に搬出される。搬入口１３ａ及び搬出口１３ｂのそれぞれには、Ｘ線の漏洩を防止するＸ線遮蔽カーテン（図示省略）が設けられている。 The apparatus main body 11 houses the control unit 18 and the like. The support legs 12 support the apparatus main body 11. The shield box 13 is provided on the apparatus main body 11. The shield box 13 prevents leakage of X-rays to the outside. An inspection area R in which the inspection of the article A by X-rays is performed is provided inside the shield box 13. The shield box 13 has a carry-in port 13a and a carry-out port 13b. The article A before the inspection is carried into the inspection area R from the carry-in conveyor 19 via the carry-in port 13a. The article A after the inspection is carried out from the inspection area R to the conveyor 31 of the sorting device 30 via the outlet 13b. Each of the carry-in entrance 13a and the carry-out exit 13b is provided with an X-ray shielding curtain (not shown) for preventing leakage of X-rays.
　搬送部１４は、シールドボックス１３内に配置されている。搬送部１４は、搬入口１３ａから検査領域Ｒを介して搬出口１３ｂまで、搬送方向Ｄに沿って物品Ａを搬送する。搬送部１４は、例えば、搬入口１３ａと搬出口１３ｂとの間に掛け渡されたベルトコンベアである。 The transport unit 14 is arranged in the shield box 13. The transport unit 14 transports the article A in the transport direction D from the entrance 13a to the exit 13b via the inspection area R. The transport unit 14 is, for example, a belt conveyor spanned between the entrance 13a and the exit 13b.
　図１及び図２に示されるように、Ｘ線照射部１５は、シールドボックス１３内に配置されている。Ｘ線照射部１５は、搬送部１４によって搬送される物品ＡにＸ線を照射する。Ｘ線照射部１５は、例えば、Ｘ線を出射するＸ線管と、Ｘ線管から出射されたＸ線を搬送方向Ｄに垂直な面内において扇状に広げるコリメータと、を有している。 Ｘ As shown in FIGS. 1 and 2, the X-ray irradiator 15 is disposed inside the shield box 13. The X-ray irradiator 15 irradiates the article A transported by the transporter 14 with X-rays. The X-ray irradiator 15 has, for example, an X-ray tube that emits X-rays, and a collimator that spreads the X-rays emitted from the X-ray tube in a fan shape in a plane perpendicular to the transport direction D.
　Ｘ線検出部１６は、シールドボックス１３内に配置されている。Ｘ線検出部１６は、搬送方向Ｄに垂直な水平方向に沿って一次元に配列されたＸ線検出素子によって構成されているラインセンサである。Ｘ線検出部１６は、物品Ａ及び搬送部１４の搬送ベルトを透過したＸ線を検出する。 (4) The X-ray detector 16 is arranged in the shield box 13. The X-ray detection unit 16 is a line sensor including X-ray detection elements arranged one-dimensionally in a horizontal direction perpendicular to the transport direction D. The X-ray detection unit 16 detects the X-ray transmitted through the article A and the transport belt of the transport unit 14.
　図１に示されるように、表示操作部１７は、装置本体１１に設けられている。表示操作部１７は、各種情報を表示すると共に、各種条件の入力を受け付ける。表示操作部１７は、例えば、液晶ディスプレイであり、タッチパネルとしての操作画面を表示する。この場合、オペレータは、表示操作部１７を介して各種条件を入力することができる。 表示 As shown in FIG. 1, the display operation unit 17 is provided in the apparatus main body 11. The display operation unit 17 displays various information and receives input of various conditions. The display operation unit 17 is, for example, a liquid crystal display, and displays an operation screen as a touch panel. In this case, the operator can input various conditions via the display operation unit 17.
　制御部１８は、装置本体１１内に配置されている。制御部１８は、Ｘ線検査装置１０の各部の動作を制御する。制御部１８は、ＣＰＵ（Central　Processing　Unit）、ＲＯＭ（Read　Only　Memory）、ＲＡＭ（Random　Access　Memory）等で構成されている。制御部１８には、Ｘ線検出部１６のＸ線の検出結果が入力される。制御部１８は、Ｘ線の検出結果に基づいて、透過画像Ｇ１を作成する。 The control unit 18 is arranged in the apparatus main body 11. The control unit 18 controls the operation of each unit of the X-ray inspection apparatus 10. The control unit 18 includes a CPU (Central Processing Unit), a ROM (Read Only Memory), a RAM (Random Access Memory), and the like. The X-ray detection result of the X-ray detector 16 is input to the controller 18. The control unit 18 creates a transmission image G1 based on the X-ray detection result.
　図３に示されるように、制御部１８は、設定部２０と、処理部２１と、取得部２２と、検査部２３と、出力部２４と、を有している。 制 御 As shown in FIG. 3, the control unit 18 includes a setting unit 20, a processing unit 21, an acquisition unit 22, an inspection unit 23, and an output unit 24.
　設定部２０は、処理部２１において作成する縮小画像Ｇ２（図４（ｂ）参照）の縮小率を設定する。設定部２０は、透過画像Ｇ１が作成されてから物品Ａが振分装置３０に到達するまでの第１時間と、学習済モデルによる処理に要する第２時間とに基づいて、縮小画像Ｇ２の縮小率を設定する。第１時間は、Ｘ線検査装置１０及び振分装置３０の構成に応じて設定される。具体的には、第１時間は、物品Ａの寸法（長さ）、Ｘ線検出部１６と振分装置３０との間の距離、搬送部１４の搬送速度、及び、コンベア３１の搬送速度等に応じて適宜設定される。 The setting unit 20 sets a reduction ratio of the reduced image G2 (see FIG. 4B) created by the processing unit 21. The setting unit 20 reduces the reduced image G2 based on a first time from when the transmission image G1 is created to when the article A reaches the distribution device 30 and a second time required for processing by the learned model. Set the rate. The first time is set according to the configuration of the X-ray inspection device 10 and the distribution device 30. Specifically, the first time is the dimension (length) of the article A, the distance between the X-ray detection unit 16 and the sorting device 30, the transport speed of the transport unit 14, the transport speed of the conveyor 31, and the like. Is set as appropriate according to.
　第２時間は、サーバ５０の処理能力等に応じて設定される。第２時間は、画像の画素数が大きいほど長くなる。設定部２０は、第２時間が第１時間よりも短くなるように（第１時間＞第２時間）、縮小率を設定する。第２時間は、例えば、２００ｍｓｅｃである。学習済モデルによる処理精度の観点からは、第２時間は、第１時間と大きく異ならないことが好ましい。これにより、振分装置３０による振り分けを確実に行いつつ、所定の処理精度が得られる縮小率を設定できる。設定部２０は、設定した縮小率を示す縮小情報を処理部２１に出力する。 The second time is set according to the processing capacity of the server 50 and the like. The second time is longer as the number of pixels of the image is larger. The setting unit 20 sets the reduction ratio such that the second time is shorter than the first time (first time> second time). The second time is, for example, 200 msec. From the viewpoint of the processing accuracy of the learned model, it is preferable that the second time is not significantly different from the first time. Thus, it is possible to set a reduction ratio at which a predetermined processing accuracy is obtained while reliably performing distribution by the distribution device 30. The setting unit 20 outputs reduction information indicating the set reduction ratio to the processing unit 21.
　処理部２１は、透過画像Ｇ１に画像処理を施す。処理部２１は、透過画像Ｇ１に対して画素数の縮小処理を施して、縮小画像Ｇ２を作成する。縮小処理は、透過画像Ｇ１の画素数を減少させる処理であり、公知の方法を用いることができる。処理部２１は、例えば、ニアレストネイバー法、バイリニア法、又は、バイキュービック法を用いて、縮小画像Ｇ２を作成する。処理部２１は、設定部２０によって設定された縮小率に基づいて、縮小画像Ｇ２を作成する。処理部２１は、透過画像Ｇ１に対して、コントラストの調整、色の変更、及びフォーマットの変更等の各種の処理を行ってもよい。処理部２１は、縮小画像Ｇ２に係る画像情報を、通信部（図示省略）を介してサーバ５０に送信する。 (4) The processing unit 21 performs image processing on the transmission image G1. The processing unit 21 performs a reduction process of the number of pixels on the transmission image G1 to create a reduced image G2. The reduction process is a process for reducing the number of pixels of the transparent image G1, and a known method can be used. The processing unit 21 creates the reduced image G2 by using, for example, the nearest neighbor method, the bilinear method, or the bicubic method. The processing unit 21 creates a reduced image G2 based on the reduction ratio set by the setting unit 20. The processing unit 21 may perform various processes such as contrast adjustment, color change, and format change on the transmission image G1. The processing unit 21 transmits image information relating to the reduced image G2 to the server 50 via a communication unit (not shown).
　取得部２２は、機械学習によって生成された学習済モデルによる処理を行った処理結果を取得する。取得部２２は、処理部２１から送信された画像情報に応じて、サーバ５０から送信された処理情報を受信して処理結果を取得する。取得部２２は、取得した処理結果を検査部２３に出力する。 The acquisition unit 22 acquires a processing result obtained by performing a process using a learned model generated by machine learning. The acquiring unit 22 receives the processing information transmitted from the server 50 according to the image information transmitted from the processing unit 21 and acquires a processing result. The acquisition unit 22 outputs the acquired processing result to the inspection unit 23.
　検査部２３は、物品Ａの検査を行う。検査部２３は、取得部２２によって取得された処理結果と、透過画像Ｇ１に基づく判定結果と、に基づいて、物品Ａに異物Ｆが含まれているか否かを判定する。 The inspection unit 23 inspects the article A. The inspection unit 23 determines whether or not the article A includes the foreign matter F based on the processing result obtained by the obtaining unit 22 and the determination result based on the transmission image G1.
　検査部２３は、複数の感度レベル（例えば、１～７）毎に設定されている画像処理アルゴリズムを用い、透過画像Ｇ１の処理（画像処理）を行って処理画像を生成する。画像処理アルゴリズムとは、透過画像Ｇ１に施す画像処理の処理手順を示す型である。画像処理アルゴリズムは、１つの画像処理フィルタ、又は、複数の画像処理フィルタの組み合わせによって構成される。複数の画像処理アルゴリズムは、インターネット等のネットワークを介して外部から取得することができる。また、複数の画像処理アルゴリズムは、ＵＳＢメモリ又はリムーバブルハードディスク等の外部記憶媒体から取得することもできる。複数の画像処理アルゴリズムのうちの少なくとも１つ以上は、生物界における遺伝及び進化のメカニズムを応用した手法である遺伝的アルゴリズム（GA＝Genetic　Algorithms）を採用して、Ｘ線検査装置１０の仕様又は検査条件等に基づき複数の画像処理フィルタから自動生成することができる。複数の画像処理アルゴリズムの少なくとも一部は、作業者が表示操作部１７を介して適宜設定することもできる。 The inspection unit 23 performs processing (image processing) on the transmission image G1 using an image processing algorithm set for each of a plurality of sensitivity levels (for example, 1 to 7) to generate a processed image. The image processing algorithm is a type indicating a processing procedure of image processing performed on the transmission image G1. The image processing algorithm is configured by one image processing filter or a combination of a plurality of image processing filters. The plurality of image processing algorithms can be obtained from outside via a network such as the Internet. Further, the plurality of image processing algorithms can be obtained from an external storage medium such as a USB memory or a removable hard disk. At least one of the plurality of image processing algorithms adopts a genetic algorithm (GA = Genetic @ Algorithms), which is a method that applies a mechanism of genetic and evolution in the living world, and specifies the specifications of the X-ray inspection apparatus 10 or It can be automatically generated from a plurality of image processing filters based on inspection conditions and the like. At least a part of the plurality of image processing algorithms can be appropriately set by the operator via the display operation unit 17.
　検査部２３は、画像処理アルゴリズムに基づいて透過画像Ｇ１に所定の画像処理を行い、画像処理して得られた判定画像に基づいて、物品Ａに異物Ｆが含まれているか否かを判定する。所定の画像処理は、例えば、透過画像Ｇ１の２値化処理等である。検査部２３は、判定画像において、輝度値が閾値を超えている場合には、物品Ａに異物Ｆが含まれていると判定する。閾値は、物品Ａの性質に応じて、試験等によって適宜設定される。検査部２３は、判定結果を記憶部（図示省略）に記憶させる。 The inspection unit 23 performs predetermined image processing on the transmission image G1 based on the image processing algorithm, and determines whether or not the article A contains the foreign matter F based on the determination image obtained by performing the image processing. . The predetermined image processing is, for example, a binarization processing of the transmission image G1. The inspection unit 23 determines that the foreign matter F is included in the article A when the luminance value exceeds the threshold value in the determination image. The threshold value is appropriately set by a test or the like according to the properties of the article A. The inspection unit 23 stores the determination result in a storage unit (not shown).
　検査部２３は、取得部２２によって取得された処理結果と、上記判定結果とに基づいて、物品Ａに異物Ｆが含まれているか否かを検査する。検査部２３は、処理結果において物品Ａに異物Ｆが含まれておらず、且つ、判定結果において物品Ａに異物Ｆが含まれていない場合には、物品Ａに異物Ｆが含まれていないと判定する。検査部２３は、処理結果において物品Ａに異物Ｆが含まれており、且つ、判定結果において物品Ａに異物Ｆが含まれている場合には、物品Ａに異物Ｆが含まれていると判定する。検査部２３は、処理結果及び判定結果の一方において物品Ａに異物Ｆが含まれており、且つ、処理結果及び判定結果の他方において物品Ａに異物Ｆが含まれていない場合には、処理結果を採用する。具体的には、例えば、検査部２３は、処理結果において物品Ａに異物Ｆが含まれており、判定結果において物品Ａに異物Ｆが含まれていない場合には、物品Ａに異物Ｆが含まれていると判定する。検査部２３は、検査結果を示す検査情報を出力部２４に出力する。 The inspection unit 23 inspects whether or not the article A contains the foreign matter F based on the processing result obtained by the obtaining unit 22 and the above determination result. The inspection unit 23 determines that the foreign material F is not included in the article A when the foreign matter F is not included in the article A in the processing result and the foreign matter F is not included in the determination result. judge. The inspection unit 23 determines that the foreign matter F is included in the article A when the foreign matter F is included in the article A in the processing result and the foreign matter F is included in the determination result. I do. The inspection unit 23 determines the processing result when the foreign substance F is included in the article A in one of the processing result and the determination result and the foreign substance F is not included in the other processing result and the determination result. Is adopted. Specifically, for example, if the inspection result indicates that the foreign matter F is included in the article A in the processing result and the foreign matter F is not included in the determination result, the foreign matter F is included in the article A. It is determined that it has been done. The inspection unit 23 outputs inspection information indicating an inspection result to the output unit 24.
　出力部２４は、検査部２３から出力された検査情報に基づいて、各種信号を出力する。出力部２４は、検査情報に基づいて、検査結果を表示操作部１７に表示させるための表示信号を、表示操作部１７に出力する。出力部２４は、検査情報において、物品Ａに異物Ｆが含まれている場合、振分装置３０に対して、当該物品Ａの振り分けを指示する振分信号を出力する。 The output unit 24 outputs various signals based on the test information output from the test unit 23. The output unit 24 outputs a display signal for displaying the inspection result on the display operation unit 17 to the display operation unit 17 based on the inspection information. When the inspection information includes the foreign matter F in the article A, the output unit 24 outputs a distribution signal instructing the distribution device 30 to distribute the article A.
　図５に示されるように、振分装置３０は、Ｘ線検査装置１０よりも下流側に設けられている。振分装置３０は、コンベア３１に設けられている。振分装置３０は、Ｘ線検査装置１０から出力された振分信号に基づいて、物品Ａを振り分ける。振分装置３０は、光電センサ３２と、アーム３３と、を有している。 よ う As shown in FIG. 5, the distribution device 30 is provided downstream of the X-ray inspection device 10. The distribution device 30 is provided on the conveyor 31. The sorting device 30 sorts the articles A based on the sorting signal output from the X-ray inspection device 10. The distribution device 30 has a photoelectric sensor 32 and an arm 33.
　光電センサ３２は、物品Ａの通過を検知するセンサである。光電センサ３２は、アーム３３の上流側に設置されており、振分装置３０への物品Ａの搬入を検知する。光電センサ３２は、投光器３２ａ及び受光器３２ｂを有している。光電センサ３２は、投光器３２ａから受光器３２ｂに出射された光が物品Ａで遮光された場合に、物品Ａが振分装置３０に到達したとして、信号をＸ線検査装置１０に送信する。 The photoelectric sensor 32 is a sensor that detects the passage of the article A. The photoelectric sensor 32 is installed on the upstream side of the arm 33 and detects the entry of the article A into the distribution device 30. The photoelectric sensor 32 has a light projector 32a and a light receiver 32b. When the light emitted from the light emitter 32a to the light receiver 32b is blocked by the article A, the photoelectric sensor 32 transmits a signal to the X-ray inspection apparatus 10 assuming that the article A has reached the distribution device 30.
　アーム３３は、例えばモータ等の駆動力により基端側を基軸に先端が揺動する。アーム３３は、コンベア３１の幅方向における一方側へ物品Ａを押し出し、当該物品Ａを生産ライン外に振り分ける。 The tip of the arm 33 swings around the base end by a driving force of, for example, a motor. The arm 33 pushes the article A to one side in the width direction of the conveyor 31 and sorts the article A out of the production line.
　図１に示されるように、サーバ５０は、機械学習によって学習済モデルを生成して、学習済モデルによって処理を行う装置である。サーバ５０は、ＣＰＵ（Central　Processing　Unit）、ＲＯＭ（Read　Only　Memory）、ＲＡＭ（Random　Access　Memory）等で構成されている。 As shown in FIG. 1, the server 50 is a device that generates a learned model by machine learning and performs processing using the learned model. The server 50 includes a CPU (Central Processing Unit), a ROM (Read Only Memory), a RAM (Random Access Memory), and the like.
　サーバ５０は、例えば、Ｘ線検査装置１０のユーザーによって管理されている。図１に示されるように、Ｘ線検査装置１０とサーバ５０とは、インターネット、又は電話網等の有線又は無線のネットワークＮによって通信可能に接続されており、互いに情報の送受信を行うことができる。 The server 50 is managed by a user of the X-ray inspection apparatus 10, for example. As shown in FIG. 1, the X-ray inspection apparatus 10 and the server 50 are communicably connected to each other by a wired or wireless network N such as the Internet or a telephone network, and can transmit and receive information to and from each other. .
　サーバ５０は、通信部５１と、学習済モデル生成部５２と、学習部５３と、を備えている。 The server 50 includes a communication unit 51, a learned model generation unit 52, and a learning unit 53.
　通信部５１は、Ｘ線検査装置１０と通信を行う。通信部５１は、Ｘ線検査装置１０から送信された画像情報を受信して、学習部５３に出力する。通信部５１は、学習部５３から出力された処理情報をＸ線検査装置１０に送信する。 (4) The communication unit 51 communicates with the X-ray inspection apparatus 10. The communication unit 51 receives the image information transmitted from the X-ray inspection apparatus 10 and outputs the image information to the learning unit 53. The communication unit 51 transmits the processing information output from the learning unit 53 to the X-ray inspection device 10.
　学習済モデル生成部５２は、機械学習に用いる学習データを取得して、取得した学習データを用いて機械学習を行って学習済モデルを生成する。学習データは、教師画像、及び、その他のデータを含む。教師画像は、例えば、Ｘ線検査装置１０において取得された透過画像であり、異物Ｆを含む透過画像及び異物Ｆを含まない透過画像等である。その他のデータは、例えば、Ｘ線検査装置１０において取得された透過画像において、物品Ａに複数の商品が含まれている場合に、商品が互いに重なり合っている領域に関するデータ、及び、物品Ａに含まれる異物Ｆが存在する領域に関するデータ等である。 The learned model generation unit 52 acquires learning data used for machine learning, performs machine learning using the acquired learning data, and generates a learned model. The learning data includes a teacher image and other data. The teacher image is, for example, a transmission image acquired by the X-ray inspection apparatus 10, such as a transmission image including the foreign matter F and a transmission image not including the foreign matter F. Other data include, for example, in a transmission image acquired by the X-ray inspection apparatus 10, when a plurality of products are included in the article A, data on an area where the products overlap each other, and data included in the article A. And the like regarding the region where the foreign matter F exists.
　学習済モデル生成部５２は、教師画像の各画素値をニューラルネットワークへの入力値とすると共に、教師画像に対応する処理情報をニューラルネットワークの出力値として機械学習を行ってニューラルネットワークＮＷ（図６参照）を生成する。画素値を入力値とする際には、それぞれの画素（画像上の画素の位置）に対応付いたニューロンの入力値とする。上記の機械学習自体は、従来の機械学習アルゴリズムと同様に行うことができる。 The learned model generation unit 52 performs machine learning by using each pixel value of the teacher image as an input value to the neural network and processing information corresponding to the teacher image as an output value of the neural network to perform the neural network NW (FIG. 6). See). When the pixel value is used as the input value, the input value of the neuron associated with each pixel (the position of the pixel on the image) is used. The above machine learning itself can be performed in the same manner as a conventional machine learning algorithm.
　学習済モデルは、縮小画像Ｇ２に対する異物Ｆの有無を示す情報を出力する。学習済モデルは、ニューラルネットワークＮＷを含む。学習済モデルは、畳み込みニューラルネットワークを含むものであってもよい。更に、学習済モデルは、複数の階層（例えば、８層以上）のニューラルネットワークを含むものであってもよい。すなわち、ディープラーニングによって学習済モデルが生成されてもよい。 The learned model outputs information indicating the presence or absence of the foreign matter F with respect to the reduced image G2. The trained model includes a neural network NW. The trained model may include a convolutional neural network. Further, the trained model may include a neural network of a plurality of layers (for example, eight or more layers). That is, a learned model may be generated by deep learning.
　図６に示されるように、ニューラルネットワークＮＷは、例えば、入力層である第１層と、中間層（隠れ層）である第２層、第３層、及び第４層と、出力層である第５層とで構成される。第１層は、ｐ個のパラメータを要素とする入力値ｘ＝（ｘ０，ｘ１，ｘ２，…ｘｐ）をそのまま第２層に出力する。第２層、第３層、及び第４層のそれぞれは、活性化関数により総入力を出力に変換してその出力を次の層に渡す。第５層も活性化関数により総入力を出力に変換し、この出力は、２個のパラメータを要素とするニューラルネットワークの出力値ｙ＝（ｙ０，ｙ１）である。 As shown in FIG. 6, the neural network NW is, for example, a first layer that is an input layer, second, third, and fourth layers that are intermediate layers (hidden layers), and an output layer. And a fifth layer. The first layer outputs the input value x = (x 0 , x 1 , x 2 ,..., X p ) having p parameters as elements to the second layer as it is. Each of the second, third, and fourth layers converts the total input into an output by the activation function and passes the output to the next layer. The fifth layer also converts the total input into an output by the activation function, and the output is an output value y = (y 0 , y 1 ) of the neural network having two parameters as elements.
　本実施形態では、ニューラルネットワークＮＷは、縮小画像Ｇ２の各画素の画素値を入力して、異物Ｆの有無を示す情報を出力する。ニューラルネットワークＮＷの入力層には、縮小画像Ｇ２の画素数分のニューロンが設けられる。ニューラルネットワークＮＷの出力層には、異物Ｆの有無を示す情報を出力するためのニューロンが設けられる。出力層のニューロンの出力値に基づいて、異物Ｆの有無を示す情報を決めることができる。ニューロンの出力値は、例えば、０～１の値である。この場合、ニューロンの値が大きい程（値が１に近い程）、物品Ａに異物Ｆが含まれており（異常であり）、ニューロンの値が小さい程（値が０に近い程）、物品Ａに異物Ｆが含まれていない（正常である）ことを示している。 In the present embodiment, the neural network NW inputs the pixel value of each pixel of the reduced image G2 and outputs information indicating the presence or absence of the foreign matter F. The input layers of the neural network NW are provided with neurons for the number of pixels of the reduced image G2. The output layer of the neural network NW is provided with neurons for outputting information indicating the presence or absence of the foreign matter F. Information indicating the presence or absence of the foreign matter F can be determined based on the output value of the neuron in the output layer. The output value of the neuron is, for example, a value between 0 and 1. In this case, as the value of the neuron is larger (the value is closer to 1), the foreign matter F is contained in the article A (it is abnormal), and as the value of the neuron is smaller (the value is closer to 0), the article is A indicates that the foreign matter F is not included in A (normal).
　学習部５３は、通信部５１から出力された画像情報に基づく縮小画像Ｇ２を、学習済モデルに入力する。学習部５３は、学習済モデルのニューラルネットワークＮＷから出力された出力値を含む処理結果を示す処理情報を、通信部５１に出力する。処理結果には、物品Ａに異物Ｆの有無を示す情報、及び、物品Ａに異物Ｆが含まれている場合、物品Ａにおける異物Ｆの含まれている領域（位置）を示す情報等が含まれている。 The learning unit 53 inputs the reduced image G2 based on the image information output from the communication unit 51 to the learned model. The learning unit 53 outputs, to the communication unit 51, processing information indicating a processing result including an output value output from the neural network NW of the learned model. The processing result includes information indicating the presence or absence of the foreign matter F in the article A, and information indicating an area (position) in the article A where the foreign matter F is included when the article A includes the foreign matter F. Have been.
　続いて、Ｘ線検査システム１の動作（検査方法）について、図７を参照して説明する。図７に示されるように、Ｘ線検査装置１０は、Ｘ線検出部１６の検出結果から透過画像Ｇ１を作成する（ステップＳ０１）。続いて、Ｘ線検査装置１０は、透過画像Ｇ１に縮小処理を実行して、縮小画像Ｇ２を作成する（ステップＳ０２、処理ステップ）。Ｘ線検査装置１０は、縮小画像Ｇ２をサーバ５０に送信する（ステップＳ０３）。また、Ｘ線検査装置１０は、透過画像Ｇ１に基づいて、物品Ａに異物Ｆが含まれているか否かを判定する（ステップＳ０４）。 Next, the operation (inspection method) of the X-ray inspection system 1 will be described with reference to FIG. As shown in FIG. 7, the X-ray inspection apparatus 10 creates a transmission image G1 from the detection result of the X-ray detection unit 16 (Step S01). Subsequently, the X-ray inspection apparatus 10 performs a reduction process on the transmission image G1 to create a reduced image G2 (step S02, processing step). The X-ray inspection apparatus 10 transmits the reduced image G2 to the server 50 (Step S03). Further, the X-ray inspection apparatus 10 determines whether or not the foreign material F is included in the article A based on the transmission image G1 (Step S04).
　サーバ５０は、Ｘ線検査装置１０から送信された縮小画像Ｇ２に対して、学習済モデルによる処理を行う（ステップＳ０５）。サーバ５０は、学習済モデルによる処理結果を、Ｘ線検査装置１０に送信する（ステップＳ０６）。 The server 50 performs a process using the learned model on the reduced image G2 transmitted from the X-ray inspection apparatus 10 (Step S05). The server 50 transmits the processing result based on the learned model to the X-ray inspection apparatus 10 (Step S06).
　続いて、Ｘ線検査装置１０は、サーバ５０から送信された処理結果を取得し（取得ステップ）、処理結果に基づいて、物品Ａに異物Ｆが含まれているか否かの検査を行う（ステップＳ０７、検査ステップ）。Ｘ線検査装置１０は、物品Ａに異物Ｆが含まれていると判定した場合には、振分装置３０に振分信号を送信する（ステップＳ０８）。Ｘ線検査装置１０は、物品Ａに異物Ｆが含まれていると判定しなかった場合には、処理を終了する。 Subsequently, the X-ray inspection apparatus 10 acquires the processing result transmitted from the server 50 (acquisition step), and performs an inspection as to whether or not the article A contains the foreign matter F based on the processing result (step). S07, inspection step). When determining that the article A contains the foreign matter F, the X-ray inspection apparatus 10 transmits a distribution signal to the distribution apparatus 30 (step S08). If the X-ray inspection apparatus 10 does not determine that the foreign matter F is included in the article A, the processing ends.
　続いて、Ｘ線検査システム１を実現させるための検査プログラムＰについて説明する。図８に示されるように、検査プログラムＰは、コンピュータ読み取り可能な記録媒体１００に記録され得る。記録媒体１００に格納された検査プログラムＰは、設定モジュールＰ１、処理モジュールＰ２、取得モジュールＰ３、検査モジュールＰ４及び出力モジュールＰ５と、を備える。設定モジュールＰ１、処理モジュールＰ２、取得モジュールＰ３、検査モジュールＰ４及び出力モジュールＰ５をコンピュータに実行させることにより、検査プログラムＰが機能する。設定モジュールＰ１、処理モジュールＰ２、取得モジュールＰ３、検査モジュールＰ４及び出力モジュールＰ５を実行することにより実現される機能はそれぞれ、設定部２０、処理部２１、取得部２２、検査部２３及び出力部２４の機能と同様である。 Next, an inspection program P for realizing the X-ray inspection system 1 will be described. As shown in FIG. 8, the inspection program P can be recorded on a computer-readable recording medium 100. The inspection program P stored in the recording medium 100 includes a setting module P1, a processing module P2, an acquisition module P3, an inspection module P4, and an output module P5. The inspection program P functions by causing the computer to execute the setting module P1, the processing module P2, the acquisition module P3, the inspection module P4, and the output module P5. The functions realized by executing the setting module P1, the processing module P2, the acquisition module P3, the inspection module P4, and the output module P5 are respectively a setting unit 20, a processing unit 21, an acquisition unit 22, an inspection unit 23, and an output unit 24. Is the same as the function of
　検査プログラムＰは、記録媒体１００におけるプログラム記録領域に記録されている。記録媒体１００は、例えば、ＣＤ－ＲＯＭ、ＤＶＤ、ＲＯＭ、半導体メモリ等の記録媒体によって構成されている。検査プログラムＰは、搬送波に重畳されたコンピュータデータ信号として通信ネットワークを介して提供されてもよい。 The inspection program P is recorded in a program recording area of the recording medium 100. The recording medium 100 is configured by a recording medium such as a CD-ROM, a DVD, a ROM, and a semiconductor memory. The inspection program P may be provided via a communication network as a computer data signal superimposed on a carrier wave.
　以上説明したように、本実施形態に係るＸ線検査システム１では、学習済モデルによる処理に縮小画像Ｇ２を用いる。これにより、Ｘ線検査システム１では、学習済モデルによる処理負荷が小さくなるため、処理結果を迅速に得ることが可能となる。したがって、Ｘ線検査システム１では、処理能力の向上を図ることができる。 As described above, in the X-ray inspection system 1 according to the present embodiment, the reduced image G2 is used for the process using the learned model. As a result, in the X-ray inspection system 1, the processing load of the learned model is reduced, and the processing result can be obtained quickly. Therefore, in the X-ray inspection system 1, the processing capacity can be improved.
　また、本実施形態に係るＸ線検査システム１では、学習済モデルによる処理に縮小画像Ｇ２を用いることにより、処理能力の向上を図ることができる。したがって、Ｘ線検査システム１では、サーバ５０の性能（ＣＰＵのスペック等）を高めることなく、処理能力の向上が図れる。そのため、Ｘ線検査システム１では、高価な装置を必要としないため、コストの低減が図れる。 In addition, in the X-ray inspection system 1 according to the present embodiment, the processing capability can be improved by using the reduced image G2 for the process using the learned model. Therefore, in the X-ray inspection system 1, the processing capability can be improved without increasing the performance of the server 50 (such as the specifications of the CPU). Therefore, the X-ray inspection system 1 does not require an expensive device, so that the cost can be reduced.
　本実施形態に係るＸ線検査システム１では、Ｘ線検査装置１０の検査部２３は、透過画像Ｇ１に基づいて物品Ａの検査を行い、当該検査の判定結果と、処理結果とに基づいて、物品Ａの検査を行う。この構成では、透過画像Ｇ１に基づく、輝度値の閾値判定による判定結果と、処理結果とに基づいて、物品Ａの検査を行う。このように、Ｘ線検査装置１０では、２つの結果に基づいて物品Ａを検査するため、検査精度の向上が図れる。 In the X-ray inspection system 1 according to the present embodiment, the inspection unit 23 of the X-ray inspection apparatus 10 inspects the article A based on the transmission image G1, and based on the determination result of the inspection and the processing result, The inspection of the article A is performed. In this configuration, the inspection of the article A is performed based on the determination result based on the threshold value determination of the luminance value based on the transmission image G1 and the processing result. As described above, since the X-ray inspection apparatus 10 inspects the article A based on the two results, the inspection accuracy can be improved.
　本実施形態に係るＸ線検査システム１では、Ｘ線検査装置１０は、検査部２３において物品の異常が検出された場合に、物品Ａの振り分けを行う振分装置３０に対して振分信号を出力する出力部２４と、透過画像Ｇ１が作成されてから物品が振分装置３０に到達するまでの第１時間と、学習済モデルによる処理に要する第２時間とに基づいて、縮小画像Ｇ２の縮小率を設定する設定部２０と、を備えている。学習済モデルによる処理に要する時間は、縮小画像Ｇ２の画素数に応じて変化し、画素数が大きい場合には長くなり、画素数が小さい場合には短くなる。画素数が大きい場合、処理に時間がかかるため、物品Ａが振分装置３０に到達するまでに、処理が完了しないおそれがある。一方で、画素数が小さすぎると、学習済モデルによる処理の精度が低下し得る。Ｘ線検査装置１０では、第１時間と第２時間とに基づいて縮小画像Ｇ２の縮小率を設定する。これにより、Ｘ線検査装置１０では、振分装置３０による振り分けを確実に行わせることができると共に、学習済モデルによる処理精度の低下を抑制できる。 In the X-ray inspection system 1 according to the present embodiment, the X-ray inspection apparatus 10 outputs a distribution signal to the distribution apparatus 30 that distributes the article A when the inspection unit 23 detects an abnormality of the article. The output unit 24 that outputs the reduced image G2 based on the first time from when the transmission image G1 is created to when the article reaches the distribution device 30 and the second time that is required for processing by the learned model. A setting unit 20 for setting a reduction ratio. The time required for processing by the learned model changes according to the number of pixels of the reduced image G2, and becomes longer when the number of pixels is large and becomes shorter when the number of pixels is small. When the number of pixels is large, the processing takes a long time, and the processing may not be completed before the article A reaches the distribution device 30. On the other hand, if the number of pixels is too small, the accuracy of processing by the learned model may be reduced. In the X-ray inspection apparatus 10, the reduction ratio of the reduced image G2 is set based on the first time and the second time. Thereby, in the X-ray inspection apparatus 10, it is possible to reliably perform the distribution by the distribution apparatus 30, and it is possible to suppress a decrease in processing accuracy due to the learned model.
　本実施形態に係るＸ線検査システム１では、サーバ５０は、教師画像を用いる機械学習によって学習済モデルを生成する学習済モデル生成部５２を備えている。この構成では、物品の検査に適切な学習済モデルを生成できる。学習済モデル生成部５２は、ニューラルネットワークを含む学習済モデルを生成する。この構成では、学習済モデルを適切なものとすることができ、物品の検査精度の向上が図れる。 In the X-ray inspection system 1 according to the present embodiment, the server 50 includes the learned model generation unit 52 that generates a learned model by machine learning using a teacher image. With this configuration, it is possible to generate a learned model suitable for inspecting an article. The learned model generation unit 52 generates a learned model including a neural network. With this configuration, the learned model can be made appropriate, and the inspection accuracy of the article can be improved.
　本実施形態に係るＸ線検査システム１では、Ｘ線検査装置１０の検査部２３は、物品Ａに異物Ｆが含まれているか否かを検査する。物品Ａに複数の商品が含まれている場合には、商品同士が重なることがある。この場合、商品同士が重なっている部分と、異物との区別を、透過画像Ｇ１の輝度値に基づいて行うことが困難となる。Ｘ線検査装置１０では、学習済モデルによる処理を行った処理結果に基づいて検査するため、商品同士が重なっている部分と、異物との区別を行うことが可能である。したがって、Ｘ線検査装置１０は、物品Ａに異物Ｆが含まれているか否かの検査に特に有効である。 In the X-ray inspection system 1 according to the present embodiment, the inspection unit 23 of the X-ray inspection apparatus 10 inspects whether or not the article A contains the foreign matter F. When the article A includes a plurality of products, the products may overlap each other. In this case, it is difficult to discriminate a portion where commodities overlap each other and a foreign substance based on the luminance value of the transmission image G1. In the X-ray inspection apparatus 10, since the inspection is performed based on the processing result obtained by performing the processing using the learned model, it is possible to distinguish a portion where the products overlap each other and a foreign substance. Therefore, the X-ray inspection apparatus 10 is particularly effective in inspecting whether or not the article A contains the foreign matter F.
　以上、本発明の実施形態について説明してきたが、本発明は必ずしも上述した実施形態に限定されるものではなく、その要旨を逸脱しない範囲で様々な変更が可能である。 Although the embodiments of the present invention have been described above, the present invention is not necessarily limited to the above-described embodiments, and various modifications can be made without departing from the gist of the present invention.
　上記実施形態では、Ｘ線検査装置１０のＸ線検出部１６が、ラインセンサを１個備える形態を一例に説明した。しかし、Ｘ線検出部１６は、２個のラインセンサを備えていてもよい。この構成では、２つのラインセンサは、上下方向において対向して配置される。一方（上方）のラインセンサは、物品Ａ及び搬送部１４の搬送ベルトを透過した低エネルギー帯のＸ線を検出する。他方（下方）のラインセンサは、物品Ａ、搬送部１４の搬送ベルト及び一方のラインセンサを透過した高エネルギー帯のＸ線を検出する。 In the above-described embodiment, an example has been described in which the X-ray detection unit 16 of the X-ray inspection apparatus 10 includes one line sensor. However, the X-ray detector 16 may include two line sensors. In this configuration, the two line sensors are arranged to face each other in the vertical direction. On the other hand, the (upper) line sensor detects X-rays in the low energy band transmitted through the conveyance belt of the article A and the conveyance unit 14. The other (lower) line sensor detects X-rays in the high energy band transmitted through the article A, the conveyor belt of the conveyor 14 and one of the line sensors.
　上記実施形態では、Ｘ線検査装置１０の検査部２３が、透過画像Ｇ１に基づく判定結果と、機械学習によって生成された学習済モデルによる処理を行った処理結果とに基づいて、物品Ａの検査を行う形態を一例に説明した。しかし、検査部２３は、学習済モデルによる処理を行った処理結果のみに基づいて、物品Ａの検査を行ってもよい。 In the above embodiment, the inspection unit 23 of the X-ray inspection apparatus 10 inspects the article A based on the determination result based on the transmission image G1 and the processing result obtained by performing the processing using the learned model generated by the machine learning. The embodiment for performing the above has been described as an example. However, the inspection unit 23 may inspect the article A based on only the processing result of the processing using the learned model.
　上記実施形態では、Ｘ線検査システム１がＸ線検査装置１０とサーバ５０とを備える形態を一例に説明した。しかし、サーバ５０を備えていなくてもよい。この場合、Ｘ線検査装置１０が学習済モデル生成部及び学習部を備えていればよい。あるいは、Ｘ線検査装置１０は、他の装置（コンピュータ）で生成された学習済モデルを取得して、記憶部に記憶していてもよい。 In the above embodiment, an example in which the X-ray inspection system 1 includes the X-ray inspection apparatus 10 and the server 50 has been described. However, the server 50 need not be provided. In this case, the X-ray inspection apparatus 10 only needs to include the learned model generation unit and the learning unit. Alternatively, the X-ray inspection apparatus 10 may acquire a learned model generated by another apparatus (computer) and store it in the storage unit.
　上記実施形態では、Ｘ線検査システム１がＸ線検査装置１０と振分装置３０とを備える形態を一例に説明した。しかし、振分装置３０は、Ｘ線検査装置１０の構成の一部として含まれていてもよい。 In the above embodiment, an example in which the X-ray inspection system 1 includes the X-ray inspection device 10 and the distribution device 30 has been described. However, the distribution device 30 may be included as a part of the configuration of the X-ray inspection device 10.
　上記実施形態では、サーバ５０が学習済モデル生成部５２を備える形態を一例に説明した。しかし、サーバ５０は、学習済モデル生成部５２を備えていなくてもよい。この場合、サーバ５０は、他の装置で生成された学習済モデルを取得して記憶部に記憶させ、記憶部に記憶されている学習済モデルを用いて、縮小画像Ｇ２に処理を施せばよい。 In the above embodiment, an example in which the server 50 includes the learned model generation unit 52 has been described as an example. However, the server 50 may not include the learned model generation unit 52. In this case, the server 50 may acquire the learned model generated by another device, store the acquired model in the storage unit, and perform processing on the reduced image G2 using the learned model stored in the storage unit. .
　上記実施形態では、学習済モデル生成部５２が、機械学習に用いる学習データを取得して、取得した学習データを用いて機械学習を行って学習済モデルを生成する形態を一例に説明した。しかし、学習済モデルの生成方法は、これに限定されない。また、学習データとしては、教師画像及びその他のデータの他に、更に他のデータを用いてもよい。 In the above-described embodiment, an example has been described in which the learned model generation unit 52 acquires learning data used for machine learning and performs machine learning using the acquired learning data to generate a learned model. However, the generation method of the learned model is not limited to this. Further, as the learning data, other data may be used in addition to the teacher image and other data.
　上記実施形態では、ニューラルネットワークＮＷが５層（入力層を除いた場合には４層）である形態を一例に説明した。しかし、学習済モデル生成部５２を構成するニューラルネットワークの層の数は何ら限定されない。例えば、学習済モデル生成部５２は３以上の任意の個数の層を有するニューラルネットワークを用いてもよく、これは、１以上の任意の個数の中間層を有するニューラルネットワークを用いてもよいことを意味する。また、ニューラルネットワークの各層の構成（例えばニューロンの個数）も、ニューロン間の接続も、上記実施形態で示した構成に限定されない。 In the above embodiment, an example in which the neural network NW has five layers (four layers when the input layer is excluded) has been described as an example. However, the number of layers of the neural network constituting the learned model generation unit 52 is not limited at all. For example, the learned model generation unit 52 may use a neural network having an arbitrary number of layers of three or more, which may use a neural network having an intermediate number of one or more layers. means. Further, the configuration of each layer of the neural network (for example, the number of neurons) and the connection between neurons are not limited to the configuration described in the above embodiment.
　上記実施形態では、検査装置がＸ線検査装置１０である形態を一例に説明した。しかし、Ｘ線検査装置に限定されず、電磁波を利用して物品の検査を行う検査装置であればよい。つまり、本発明において、電磁波とは、Ｘ線、近赤外線、光、その他の電磁波である。また、本発明は、物品に含まれる異物の有無を検査するものに限定されず、フィルム包装材等のパッケージ内に食品等の内容物を収容して出荷するような物品において、パッケージの封止部への内容物の噛み込み、パッケージ内での内容物の破損、パッケージ内への異物の混入等を検査するものであってもよい。また、物品の種類は特に限定されず、様々な物品を検査対象とすることができる。同様に、異物の種類は特に限定されず、様々な異物を検査対象とすることができる。 In the above embodiment, an example in which the inspection apparatus is the X-ray inspection apparatus 10 has been described as an example. However, the inspection apparatus is not limited to the X-ray inspection apparatus, and may be any inspection apparatus that inspects an article using electromagnetic waves. That is, in the present invention, the electromagnetic waves are X-rays, near infrared rays, light, and other electromagnetic waves. In addition, the present invention is not limited to the method for inspecting for the presence or absence of foreign matter contained in an article, but is not limited to a package such as a film packaging material and the like, in which the content such as food is stored and shipped. The inspection may be performed to check for biting of the contents into the section, damage to the contents in the package, and entry of foreign matter into the package. The type of the article is not particularly limited, and various articles can be inspected. Similarly, the type of foreign matter is not particularly limited, and various foreign matters can be inspected.
　１…Ｘ線検査システム、１０…Ｘ線検査装置、１５…Ｘ線照射部（照射部）、１６…Ｘ線検出部（検出部）、２０…設定部、２１…処理部、２２…取得部、２３…検査部、２４…出力部、５０…サーバ、５１…通信部、５２…学習済モデル生成部、１００…記録媒体、Ａ…物品、Ｆ…異物、Ｐ…検査プログラム、Ｐ２…処理モジュール（処理部）、Ｐ３…取得モジュール（取得部）、Ｐ４…検査モジュール（検査部）。 DESCRIPTION OF SYMBOLS 1 ... X-ray inspection system, 10 ... X-ray inspection apparatus, 15 ... X-ray irradiation part (irradiation part), 16 ... X-ray detection part (detection part), 20 ... setting part, 21 ... processing part, 22 ... acquisition part Reference numerals 23, inspection unit, 24, output unit, 50, server, 51, communication unit, 52, learned model generation unit, 100, recording medium, A, article, F, foreign matter, P, inspection program, P2, processing module (Processing unit), P3: acquisition module (acquisition unit), P4: inspection module (inspection unit).

Claims (13)

  1. 　物品に電磁波を照射する照射部と、
    　前記物品を透過した電磁波を検出する検出部と、
    　前記検出部の検出結果から作成された透過画像に対して画素数の縮小処理を施して、縮小画像を作成する処理部と、
    　前記処理部によって作成された前記縮小画像に対して、機械学習によって生成された学習済モデルによる処理を行った処理結果を取得する取得部と、
    　前記取得部によって取得された前記処理結果に基づいて、前記物品の検査を行う検査部と、を備える、検査装置。     An irradiation unit that irradiates the article with electromagnetic waves;
    A detection unit that detects the electromagnetic wave transmitted through the article,
    A processing unit that performs a reduction process of the number of pixels on the transmission image created from the detection result of the detection unit, and creates a reduced image.
    An acquisition unit configured to acquire a processing result obtained by performing a process using a learned model generated by machine learning on the reduced image created by the processing unit;
    An inspection unit that inspects the article based on the processing result acquired by the acquisition unit.
  2. 　前記検査部は、前記透過画像に基づいて前記物品の検査を行い、当該検査の判定結果と、前記処理結果とに基づいて、前記物品の検査を行う、請求項１に記載の検査装置。 2. The inspection device according to claim 1, wherein the inspection unit inspects the article based on the transmission image, and inspects the article based on a determination result of the inspection and the processing result.
  3. 　前記検査部において前記物品の異常が検出された場合に、前記物品の振り分けを行う振分装置に対して振分信号を出力する出力部と、
    　前記透過画像が作成されてから前記物品が前記振分装置に到達するまでの第１時間と、前記学習済モデルによる処理に要する第２時間とに基づいて、前記縮小画像の縮小率を設定する設定部と、を備える、請求項１又は２に記載の検査装置。     When an abnormality of the article is detected in the inspection unit, an output unit that outputs a sorting signal to a sorting device that sorts the articles,
    A reduction ratio of the reduced image is set based on a first time from when the transparent image is created to when the article reaches the distribution device and a second time required for processing by the learned model. The inspection device according to claim 1, further comprising: a setting unit.
  4. 　教師画像を用いる機械学習によって前記学習済モデルを生成する生成部を備える、請求項１～３のいずれか一項に記載の検査装置。 The inspection device according to any one of claims 1 to 3, further comprising a generation unit configured to generate the learned model by machine learning using a teacher image.
  5. 　前記生成部によって生成された前記学習済モデルにより、前記縮小画像に対して処理を行う学習部を備え、
    　前記取得部は、前記学習部によって実行された前記処理結果を取得する、請求項４に記載の検査装置。     A learning unit that performs processing on the reduced image by the learned model generated by the generation unit,
    The inspection device according to claim 4, wherein the acquisition unit acquires a result of the processing performed by the learning unit.
  6. 　前記生成部は、ニューラルネットワークを含む前記学習済モデルを生成する、請求項４又は５に記載の検査装置。 The inspection device according to claim 4 or 5, wherein the generation unit generates the learned model including a neural network.
  7. 　前記検査部は、前記物品に異物が含まれているか否かを検査する、請求項１～６のいずれか一項に記載の検査装置。 The inspection device according to any one of claims 1 to 6, wherein the inspection unit inspects whether the article contains a foreign substance.
  8. 　前記照射部は、前記物品にＸ線を照射し、
    　前記検出部は、前記物品を透過した前記Ｘ線を検出する、請求項１～７のいずれか一項に記載の検査装置。     The irradiating unit irradiates the article with X-rays,
    The inspection device according to claim 1, wherein the detection unit detects the X-ray transmitted through the article.
  9. 　検査装置と、当該検査装置と通信可能に接続されているサーバと、を備える検査システムであって、
    　前記検査装置は、
    　　物品に電磁波を照射する照射部と、
    　　前記物品を透過した電磁波を検出する検出部と、
    　　前記検出部の検出結果から作成された透過画像に対して画素数の縮小処理を施して縮小画像を作成し、前記縮小画像に係る画像情報を前記サーバに送信する処理部と、
    　　前記処理部から前記サーバに前記画像情報を送信したことに応じて、機械学習によって生成された学習済モデルによる処理を行った処理結果を前記サーバから取得する取得部と、
    　　前記取得部によって取得された前記処理結果に基づいて、前記物品の検査を行う検査部と、を備え、
    　前記サーバは、
    　　前記検査装置から送信された前記画像情報に基づいて、前記縮小画像に対して前記学習済モデルによる処理を行う学習部と、
    　　前記処理結果を前記検査装置に送信する通信部と、を備える、検査システム。     An inspection system comprising an inspection device and a server communicably connected to the inspection device,
    The inspection device,
    An irradiation unit that irradiates the article with electromagnetic waves;
    A detection unit that detects the electromagnetic wave transmitted through the article,
    A processing unit that performs reduced processing of the number of pixels on the transmitted image created from the detection result of the detection unit to create a reduced image, and transmits image information related to the reduced image to the server;
    An acquiring unit that acquires, from the server, a processing result obtained by performing a process based on a learned model generated by machine learning, in response to transmitting the image information from the processing unit to the server;
    An inspection unit that inspects the article based on the processing result acquired by the acquisition unit,
    The server comprises:
    Based on the image information transmitted from the inspection device, a learning unit that performs processing by the learned model on the reduced image,
    A communication unit configured to transmit the processing result to the inspection device.
  10. 　前記サーバは、教師画像を用いる機械学習によって前記学習済モデルを生成する生成部を備える、請求項９に記載の検査システム。 The inspection system according to claim 9, wherein the server includes a generation unit configured to generate the learned model by machine learning using a teacher image.
  11. 　物品に電磁波を照射する照射部と、前記物品を透過した電磁波を検出する検出部と、を備える検査装置で実行される検査方法であって、
    　前記検出部の検出結果から作成された透過画像に対して画素数の縮小処理を施して、縮小画像を作成する処理ステップと、
    　前記処理ステップにおいて作成された前記縮小画像に対して、機械学習によって生成された学習済モデルによる処理を行った処理結果を取得する取得ステップと、
    　前記取得ステップにおいて取得された前記処理結果に基づいて、前記物品の検査を行う検査ステップと、を含む、検査方法。     An irradiating unit that irradiates the article with electromagnetic waves, and a detecting unit that detects the electromagnetic waves transmitted through the article, an inspection method performed by an inspection apparatus including:
    A processing step of performing a reduction process of the number of pixels on the transmission image created from the detection result of the detection unit to create a reduced image,
    An obtaining step of obtaining a processing result obtained by performing processing by a learned model generated by machine learning on the reduced image created in the processing step;
    An inspection step of inspecting the article based on the processing result acquired in the acquiring step.
  12. 　物品に電磁波を照射する照射部と、前記物品を透過した電磁波を検出する検出部と、を備える検査装置のコンピュータを、
    　前記検出部の検出結果から作成された透過画像に対して画素数の縮小処理を施して、縮小画像を作成する処理部と、
    　前記処理部によって作成された前記縮小画像に対して、機械学習によって生成された学習済モデルによる処理を行った処理結果を取得する取得部と、
    　前記取得部によって取得された前記処理結果に基づいて、前記物品の検査を行う検査部と、して機能させる、検査プログラム。     An irradiation unit that irradiates the article with an electromagnetic wave, and a detection unit that detects the electromagnetic wave transmitted through the article, a computer of an inspection apparatus including:
    A processing unit that performs a reduction process of the number of pixels on the transmission image created from the detection result of the detection unit, and creates a reduced image.
    An acquisition unit configured to acquire a processing result obtained by performing a process using a learned model generated by machine learning on the reduced image created by the processing unit;
    An inspection program that functions as an inspection unit that inspects the article based on the processing result acquired by the acquisition unit.
  13. 　物品に電磁波を照射する照射部と、前記物品を透過した電磁波を検出する検出部と、を備える検査装置のコンピュータに実行させる検査プログラムを記録しているコンピュータ読取可能な記録媒体であって、
    　前記コンピュータを、
    　前記検出部の検出結果から作成された透過画像に対して画素数の縮小処理を施して、縮小画像を作成する処理部と、
    　前記処理部によって作成された前記縮小画像に対して、機械学習によって生成された学習済モデルによる処理を行った処理結果を取得する取得部と、
    　前記取得部によって取得された前記処理結果に基づいて、前記物品の検査を行う検査部と、して機能させる前記検査プログラムを記録しているコンピュータ読取可能な記録媒体。     An irradiation unit that irradiates the article with electromagnetic waves, and a detection unit that detects the electromagnetic waves transmitted through the article, and a computer-readable recording medium that stores an inspection program to be executed by a computer of an inspection device including the inspection apparatus,
    Said computer,
    A processing unit that performs a reduction process of the number of pixels on the transmission image created from the detection result of the detection unit, and creates a reduced image.
    An acquisition unit configured to acquire a processing result obtained by performing a process using a learned model generated by machine learning on the reduced image created by the processing unit;
    A computer-readable recording medium that records the inspection program to function as an inspection unit that inspects the article based on the processing result acquired by the acquisition unit.
PCT/JP2019/023558 2018-06-29 2019-06-13 Inspection device, inspection system, inspection method, inspection program and recording medium WO2020004068A1 (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
JP2018124258A JP7250301B2 (en) 2018-06-29 2018-06-29 Inspection device, inspection system, inspection method, inspection program and recording medium
JP2018-124258 2018-06-29

Publications (1)

Publication Number Publication Date
WO2020004068A1 true WO2020004068A1 (en) 2020-01-02

Family

ID=68986485

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/JP2019/023558 WO2020004068A1 (en) 2018-06-29 2019-06-13 Inspection device, inspection system, inspection method, inspection program and recording medium

Country Status (2)

Country Link
JP (1) JP7250301B2 (en)
WO (1) WO2020004068A1 (en)

Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH08153194A (en) * 1994-11-29 1996-06-11 Hitachi Medical Corp Image processor
JP2002312762A (en) * 2001-04-12 2002-10-25 Seirei Ind Co Ltd Grain sorting apparatus utilizing neural network
JP4317566B2 (en) * 2004-06-24 2009-08-19 株式会社イシダ X-ray inspection apparatus and method for generating image processing procedure of X-ray inspection apparatus
JP2009192519A (en) * 2008-01-17 2009-08-27 Anritsu Sanki System Co Ltd X-ray foreign material detector
JP2010190867A (en) * 2009-02-20 2010-09-02 Chikafumi Teramoto Remote diagnosis method
JP2011089920A (en) * 2009-10-23 2011-05-06 Shimadzu Corp X-ray inspection method and x-ray inspection apparatus using the same
JP2012137387A (en) * 2010-12-27 2012-07-19 Anritsu Sanki System Co Ltd X-ray inspection device
JP2014016239A (en) * 2012-07-09 2014-01-30 Tokyo Electron Ltd X-ray inspection method and x-ray inspection device
JP6537008B1 (en) * 2018-02-14 2019-07-03 株式会社イシダ Inspection device

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP5876116B1 (en) 2014-08-11 2016-03-02 株式会社イシダ X-ray inspection equipment
JP2016156647A (en) 2015-02-23 2016-09-01 株式会社 システムスクエア Inspection device using electromagnetic wave

Patent Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH08153194A (en) * 1994-11-29 1996-06-11 Hitachi Medical Corp Image processor
JP2002312762A (en) * 2001-04-12 2002-10-25 Seirei Ind Co Ltd Grain sorting apparatus utilizing neural network
JP4317566B2 (en) * 2004-06-24 2009-08-19 株式会社イシダ X-ray inspection apparatus and method for generating image processing procedure of X-ray inspection apparatus
JP2009192519A (en) * 2008-01-17 2009-08-27 Anritsu Sanki System Co Ltd X-ray foreign material detector
JP2010190867A (en) * 2009-02-20 2010-09-02 Chikafumi Teramoto Remote diagnosis method
JP2011089920A (en) * 2009-10-23 2011-05-06 Shimadzu Corp X-ray inspection method and x-ray inspection apparatus using the same
JP2012137387A (en) * 2010-12-27 2012-07-19 Anritsu Sanki System Co Ltd X-ray inspection device
JP2014016239A (en) * 2012-07-09 2014-01-30 Tokyo Electron Ltd X-ray inspection method and x-ray inspection device
JP6537008B1 (en) * 2018-02-14 2019-07-03 株式会社イシダ Inspection device

Also Published As

Publication number Publication date
JP2020003387A (en) 2020-01-09
JP7250301B2 (en) 2023-04-03

Similar Documents

Publication Publication Date Title
WO2019159440A1 (en) Inspection device
JP6537008B1 (en) Inspection device
JP6126230B2 (en) Inspection device
WO2017014194A1 (en) Optical inspection system and method for setting image-processing algorithm
JP6068060B2 (en) X-ray inspection equipment
WO2019235022A1 (en) Inspection device
JP2015137858A (en) Inspection device
WO2020004068A1 (en) Inspection device, inspection system, inspection method, inspection program and recording medium
JP2005127962A (en) X-ray inspection system
JP2015083967A (en) Inspection equipment
JP2009080030A (en) X-ray inspection device
JP6420065B2 (en) Inspection device
JP7411984B2 (en) Inspection equipment
US20230258581A1 (en) X-ray inspection apparatus
JP6647873B2 (en) Inspection device
JP2021012107A (en) Inspection apparatus and learning apparatus
JP2016024096A (en) Inspection device
US20230284991A1 (en) X-ray inspection apparatus and adjustment method thereof
JP2019132755A (en) X-ray inspection device
JP6392552B2 (en) Inspection device
JP2018091874A (en) Inspection device
JP2021149653A (en) Image processing device and image processing system
US11138716B2 (en) Inspection apparatus
JP2017167059A (en) Optical inspection device
JP2016176813A (en) Inspection device and inspection method

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 19825947

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 19825947

Country of ref document: EP

Kind code of ref document: A1