EP4376624A1 - Procédés, dispositifs et agencement pour localiser des parties osseuses présentes dans une cuisse de volaille - Google Patents

Procédés, dispositifs et agencement pour localiser des parties osseuses présentes dans une cuisse de volaille

Info

Publication number
EP4376624A1
EP4376624A1 EP21749591.0A EP21749591A EP4376624A1 EP 4376624 A1 EP4376624 A1 EP 4376624A1 EP 21749591 A EP21749591 A EP 21749591A EP 4376624 A1 EP4376624 A1 EP 4376624A1
Authority
EP
European Patent Office
Prior art keywords
poultry
leg
reference points
neural network
image data
Prior art date
Legal status (The legal status 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 status listed.)
Pending
Application number
EP21749591.0A
Other languages
German (de)
English (en)
Inventor
Christoph Fabian Isernhagen
Oliver Schlesiger
Florian Jarck
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Fpi Food Processing Innovation & Co KG GmbH
Original Assignee
Fpi Food Processing Innovation & Co Kg GmbH
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 Fpi Food Processing Innovation & Co Kg GmbH filed Critical Fpi Food Processing Innovation & Co Kg GmbH
Publication of EP4376624A1 publication Critical patent/EP4376624A1/fr
Pending legal-status Critical Current

Links

Classifications

    • A—HUMAN NECESSITIES
    • A22—BUTCHERING; MEAT TREATMENT; PROCESSING POULTRY OR FISH
    • A22C—PROCESSING MEAT, POULTRY, OR FISH
    • A22C21/00—Processing poultry
    • A22C21/0069—Deboning poultry or parts of poultry
    • A—HUMAN NECESSITIES
    • A22—BUTCHERING; MEAT TREATMENT; PROCESSING POULTRY OR FISH
    • A22B—SLAUGHTERING
    • A22B5/00—Accessories for use during or after slaughtering
    • A22B5/0064—Accessories for use during or after slaughtering for classifying or grading carcasses; for measuring back fat
    • A22B5/007—Non-invasive scanning of carcasses, e.g. using image recognition, tomography, X-rays, ultrasound
    • 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
    • A—HUMAN NECESSITIES
    • A22—BUTCHERING; MEAT TREATMENT; PROCESSING POULTRY OR FISH
    • A22B—SLAUGHTERING
    • A22B5/00—Accessories for use during or after slaughtering
    • A22B5/0017—Apparatus for cutting, dividing or deboning carcasses
    • A22B5/0035—Deboning or obtaining boneless pieces of meat from a carcass
    • A—HUMAN NECESSITIES
    • A22—BUTCHERING; MEAT TREATMENT; PROCESSING POULTRY OR FISH
    • A22B—SLAUGHTERING
    • A22B5/00—Accessories for use during or after slaughtering
    • A22B5/0017—Apparatus for cutting, dividing or deboning carcasses
    • A22B5/0041—Electronic, robotic or computer assisted cutting, dividing or deboning carcasses
    • A—HUMAN NECESSITIES
    • A22—BUTCHERING; MEAT TREATMENT; PROCESSING POULTRY OR FISH
    • A22C—PROCESSING MEAT, POULTRY, OR FISH
    • A22C17/00—Other devices for processing meat or bones
    • A22C17/004—Devices for deboning meat
    • A—HUMAN NECESSITIES
    • A22—BUTCHERING; MEAT TREATMENT; PROCESSING POULTRY OR FISH
    • A22C—PROCESSING MEAT, POULTRY, OR FISH
    • A22C17/00—Other devices for processing meat or bones
    • A22C17/0073—Other devices for processing meat or bones using visual recognition, X-rays, ultrasounds, or other contactless means to determine quality or size of portioned meat
    • 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
    • G01N23/083—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 the radiation being X-rays
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N33/00—Investigating or analysing materials by specific methods not covered by groups G01N1/00 - G01N31/00
    • G01N33/02—Food
    • G01N33/12—Meat; Fish
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
    • G06V10/774—Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N2223/00—Investigating materials by wave or particle radiation
    • G01N2223/40—Imaging
    • G01N2223/401—Imaging image processing
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N2223/00—Investigating materials by wave or particle radiation
    • G01N2223/60—Specific applications or type of materials
    • G01N2223/618—Specific applications or type of materials food

Definitions

  • the present invention relates to a method for training at least one neural network for locating bony parts present in a poultry leg and a non-transitory computer-readable storage medium comprising a program comprising instructions for causing the computer to carry out the method. Furthermore, the invention relates to a method for locating bony parts present in a poultry leg and an arrangement for training at least one neural network for locating bony parts present in a poultry leg and a device for locating bony parts present in a poultry leg.
  • Such methods, arrangements and devices are used in the automatic processing of poultry.
  • the positions and/or location of the thigh and lower leg bones as well as that of the patella are of interest.
  • Document EP 2532246 B1 discloses a boning method for meat with bones using an X-ray system.
  • the poultry bone to be processed is transported between an X-ray source and an X-ray detector and the location and position of the bony parts is determined by analyzing the X-ray data recorded.
  • the object is achieved by the method mentioned at the outset for training at least one neural network for the localization of bony parts present in a poultry leg, comprising the steps of: providing a large number of poultry legs; capturing images of the front or back sides of the poultry legs in the optically visible wavelength range using an optical camera to generate optical image data for each of the poultry legs; irradiating the back or front of the poultry legs with X-rays from an X-ray source and taking X-ray images on the side of the poultry legs remote from the X-ray source with an X-ray imager to generate X-ray image data on each of the poultry legs; determining reference points for identifying the positions of the bony parts based on the X-ray image data, superimposing the positions of the bony parts and/or the X-ray image data on the optical image data to generate hybrid image data for each of the poultry legs; establishing reference points for identifying the positions of the bony parts based on the hybrid image data; inputting the image data of the optical camera as input data and the reference
  • the method according to the invention offers the advantage that X-ray technology is used exclusively in the training or learning phase.
  • the position of the bony parts in the poultry legs can be reliably determined solely on the basis of optical images of the poultry legs.
  • the optical image data and the X-ray image data are superimposed on one another.
  • the hybrid image data thus represents an overlay image containing the optical image and the X-ray image of the intact poultry leg. It is therefore possible to use the hybrid image to identify the exact location and position of the bony parts in the poultry leg and to to the external aspect of the intact fowl leg. In this way a correlation is established between the position of the bony parts and the external shape of the poultry leg.
  • reference points include thigh reference points, lower leg reference points and kneecap reference points.
  • the aforementioned reference points offer sufficient accuracy to determine the location and position of the thigh and lower leg bones and the kneecap.
  • the algorithmic complexity for determining these reference points is reduced to a required minimum.
  • the thigh reference points and the lower leg reference points are pairs of points, which each indicate the position of the bone end regions.
  • the position and alignment of said bones in the poultry leg is defined with sufficient accuracy on the basis of the pairs of points.
  • the pairs of points are particularly preferably located on the bone heads, preferably in the center of the respective bone head with respect to the longitudinal direction of the bone.
  • the patella reference points form a point cloud comprising at least one point, with the points of the point cloud referencing edge positions of the patella.
  • a point cloud with one point is usually sufficient to indicate the position of the patella. Alternatively, this one point is chosen so that it is in the middle of the kneecap.
  • a further expedient embodiment of the invention is characterized in that the point cloud comprises at least one upper patella reference point and one lower patella reference point, with the upper and lower patella reference points being localized on a patella edge area.
  • the point cloud particularly preferably includes a third reference point, so that the three reference points form a triangle whose area covers the patella as best as possible.
  • the three reference points mentioned lie then - as already mentioned - preferably on the edge of the kneecap.
  • object-related image areas of the optical image data and the x-ray image data are released from the image background before the hybrid image data are generated.
  • the cropping masks out those image areas that do not show the poultry leg or its bony parts. This prevents structures located in the background from being taken into account when training the neural network. Overall, this increases the reliability and precision of the localization of the bony parts.
  • non-volatile computer-readable storage medium mentioned at the outset comprising a program that includes instructions for causing the computer to execute the method for training at least one neural network for localizing bony parts present in a poultry leg.
  • the object is achieved by the initially mentioned method for locating bony parts present in a poultry leg, comprising the steps: conveying the poultry legs in a conveying direction by means of a conveying device; capturing digital images of the front or back sides of the poultry legs using a first optical imager for each of the poultry legs conveyed past the imager; sequentially providing the digital images as input data to a first neural network configured to locate the bony portions, the first neural network being trained to locate the bony portions according to the aforementioned method of training the at least one neural network; Determining position data of the bony parts by means of the first neural network and providing the position data for display and/or forwarding to a downstream machine for processing the poultry legs on the basis of the position data determined.
  • the method according to the invention offers the advantage that in order to localize the bony parts present in the poultry legs, only images in the optically visible range of the poultry legs have to be recorded.
  • the method according to the invention is therefore free of X-ray technology during operation.
  • the required outlay on equipment is thus considerably reduced compared to the methods known from the prior art.
  • the method according to the invention is therefore considerably less expensive than the methods known from the prior art.
  • reference points include thigh reference points, lower leg reference points and kneecap reference points.
  • the aforementioned reference points offer sufficient accuracy to determine the location and position of the thigh and lower leg bones and the kneecap.
  • the algorithmic complexity for determining these reference points is reduced to a required minimum.
  • the thigh reference points and the lower leg reference points are pairs of points, which each indicate the position of the bone end regions.
  • the position and alignment of said bones in the poultry leg is defined with sufficient accuracy on the basis of the pairs of points.
  • the pairs of points are particularly preferably located on the bone heads.
  • the patella reference points form a point cloud comprising at least one point, with the points of the point cloud referencing edge positions of the patella.
  • a point cloud with one point is usually sufficient to indicate the position of the patella. Alternatively, this one point is chosen so that it is in the middle of the kneecap.
  • a further expedient embodiment of the invention is characterized in that the point cloud comprises at least one upper patella reference point and one lower patella reference point, with the upper and lower patella reference points being localized on a patella edge area.
  • the point cloud particularly preferably includes a third reference point, so that the three reference points form a triangle whose area covers the patella as best as possible. The three reference points mentioned are then - as already mentioned - preferably on the edge of the kneecap.
  • a further expedient embodiment of the invention is characterized by determining a cutting line course of the localized bony parts from the position data provided by means of a control unit of the downstream machine and moving a control-movably equipped knife of the downstream machine by means of the control unit along this cutting line course in order to cut the poultry bones boning
  • the poultry legs can be deboned in an optimal manner.
  • the cutting line can be brought as close as possible to them in order to leave as little residual flesh as possible on the bony parts.
  • the position of the incision which is usually in the area of the kneecap, can also be precisely determined in this way. The present invention thus makes it possible to debone poultry legs fully automatically.
  • a preferred development of the invention is characterized in that the captured digital images of the poultry legs are supplied as input data to the first neural network of a leg side detection device before being provided, which is set up to carry out leg side detection and to determine whether the respective digital image is from a right or a left poultry leg and, if the digital image does not correspond to a predetermined leg side, mirroring the image data of the respective digital image about a virtual axis in order to convert the digital image of a right poultry leg into an apparent digital image of a left poultry leg and vice versa.
  • the virtual axis is a vertical axis.
  • both the complexity of the neural network is reduced and the required training outlay for the neural network is also significantly reduced, since it only has to be trained with one type of poultry leg.
  • the neural network is designed to process left poultry legs, digital images of right poultry legs are first mirrored as described. Due to the reflection, the partial image of the poultry leg then appears as if it were from a taken from a left poultry leg. In this way it is possible to localize the bony parts in poultry legs using the image data from both the right and left poultry legs using the neural network designed for processing left poultry legs. If the neural network is designed for processing right poultry legs, the corresponding procedure is reversed and the partial images of left poultry legs are mirrored as described before processing by the neural network.
  • a further expedient embodiment of the invention is characterized in that the leg side is recognized by means of a second neural network which has been trained with images of poultry legs of the specified leg side.
  • the second neural network is preferably trained exclusively with images of left or right poultry legs. If the second neural network is designed to recognize left poultry legs, for example, it indicates that no left leg has been recognized when an image of a right poultry leg is input. Failure to recognize a left fowl leg implies that it is a right leg.
  • a preferred further development of the invention is characterized in that the digital images of the poultry legs are supplied to a front/back side detection device before being provided as input data to the first neural network and/or before leg side detection, which is set up to detect a front/back side to carry out detection and to determine whether the respective digital image shows the front or the back of the poultry leg and, if the digital image does not correspond to a predetermined front/back side, to leave a hanging receptacle of the conveyor device that holds the respective poultry leg and can be swivel-controlled about its vertical axis, performing a 180° rotation and capturing a digital image of the side of the poultry leg pointing to the first optical imager by means of a second optical imager, which is downstream of the first optical imager with respect to the conveying direction.
  • the front designates the outside or skin side of the poultry leg, while the back designates the inside or flesh side of the poultry leg.
  • the bony parts can therefore be localized fully automatically, regardless of the alignment of the poultry legs.
  • the front of the poultry legs is always chosen as the specified side, because this is due to the larger muscle proportion of flesh and the resulting curved surface in the digital images provide better clues for locating the bony parts.
  • a further expedient embodiment of the invention is characterized in that the front/back side recognition takes place by means of a third neural network which has been trained with images of poultry legs of the specified front/back side.
  • the third neural network is preferably trained exclusively with images of the front or back of poultry legs. If the third neural network is designed, for example, to recognize front sides of poultry legs, this indicates when an image of the back side of a poultry leg is input that no front side has been recognized.
  • an arrangement for training at least one neural network for locating bony parts that are present in a poultry leg comprising a device set up to record and generate images of the front or back sides of the poultry legs in the optically visible wavelength range optical camera formed from optical image data on each of the poultry legs; an X-ray source arranged to irradiate the back or front of the poultry legs with X-ray radiation and an X-ray imager arranged to take X-ray images on the side of the poultry legs remote from the X-ray source and adapted to generate X-ray image data for each of the poultry legs; a display and input device designed to display the X-ray image data and/or to display hybrid image data and to input reference points to be defined, which are used to identify the positions of the bony parts; a superimposition unit designed to superimpose the optical image data with the X-ray image data and/or the reference points while generating the hybrid image data for each of the poultry legs; at least one neural network and a learning cycle control unit,
  • the arrangement according to the invention offers the advantage that X-ray technology is used exclusively in the training or learning phase.
  • the position of the bony parts in the poultry legs can be reliably determined solely on the basis of optical images of the poultry legs.
  • the optical image data and the X-ray image data are superimposed.
  • the hybrid image data thus represents an overlay image containing the optical image of the intact poultry leg and the X-ray image. It is therefore possible to use the hybrid image to identify the exact location and position of the bony parts in the poultry leg and to relate this to the external view of the intact poultry leg. In this way, a correlation between the position of the bony parts and the external shape of the poultry leg is created and learned from the neural network.
  • reference points include thigh reference points, lower leg reference points and patellar reference points.
  • the aforementioned reference points offer sufficient accuracy to determine the location and position of the thigh and lower leg bones and the kneecap. At the same time, the algorithmic complexity for determining these reference points is reduced to a required minimum.
  • the thigh reference points and the lower leg reference points are pairs of points, which each designate the position of the bone end regions.
  • the position and alignment of said bones in the poultry leg is defined with sufficient accuracy on the basis of the pairs of points.
  • the pairs of points are particularly preferably located on the bone heads.
  • a further expedient embodiment of the invention is characterized in that the patella reference points form a point cloud comprising at least one point, with the points of the point cloud referencing edge positions of the patella.
  • a point cloud with one point is usually sufficient to indicate the position of the kneecap. Alternatively, this one point is chosen so that it is in the middle of the kneecap.
  • the point cloud comprises at least one upper patella reference point and one lower patella reference point. reference point, where the superior and inferior patellar reference points are located on a patellar rim region.
  • the point cloud particularly preferably includes a third reference point, so that the three reference points form a triangle whose area covers the patella as best as possible.
  • the three reference points mentioned are then - as already mentioned - preferably on the edge of the kneecap.
  • a further expedient embodiment of the invention is characterized in that the arrangement is designed so that object-related image areas of the optical image data and the x-ray image data are released from the image background before the hybrid image data are generated. By cutting out those image areas are masked out that do not show the poultry leg or its bony parts. This prevents structures located in the background from being taken into account when training the neural network. Overall, this increases the reliability and precision of the localization of the bony parts.
  • a device for locating bony parts present in a poultry comprising a conveying device arranged for conveying the poultry legs in a conveying direction; a first optical imager configured to capture digital images of the front or back sides of the poultry legs; a first neural network designed to locate the bony parts, which has been trained according to a method according to any one of claims 1 to 6, and an input unit set up to sequentially provide the digital images as input data to the first neural network, the first neural network set up is to determine position data of the bony parts and to provide the position data for display and/or forwarding to a downstream machine for processing the poultry legs on the basis of the determined position data.
  • the device according to the invention is therefore free of X-ray technology.
  • the required ap parative effort is thus compared to that known from the prior art Devices significantly reduced.
  • the dangers associated with X-ray technology no longer exist and the use of highly qualified personnel can be dispensed with.
  • no more expensive components of the X-ray technology are required during operation and the required maintenance effort is significantly lower.
  • the device according to the invention is therefore considerably less expensive than the methods known from the prior art. However, it is in no way inferior to X-ray-based devices in terms of the precision of the localization of the bony parts.
  • reference points include thigh reference points, lower leg reference points and kneecap reference points.
  • the aforementioned reference points offer sufficient accuracy to determine the location and position of the thigh and lower leg bones and the kneecap.
  • the algorithmic complexity for determining these reference points is reduced to a required minimum.
  • An expedient embodiment of the invention is characterized in that the thigh reference points and the lower leg reference points are pairs of points which each indicate the position of the bone end regions.
  • the position and orientation of said bones in the poultry leg is defined with sufficient accuracy on the basis of the pairs of points.
  • the pairs of points are particularly preferably located on the bone heads.
  • a point cloud with one point is usually sufficient to indicate the position of the kneecap. Alternatively, this one point is chosen so that it is in the middle of the kneecap.
  • the patella reference points form a point cloud comprising at least one point, with the points of the point cloud referencing edge positions of the patella.
  • a point cloud with one point is usually sufficient to indicate the position of the kneecap. Alternatively, this one point is chosen so that it is in the middle of the kneecap.
  • a further expedient embodiment of the invention is characterized in that the point cloud comprises at least one upper patella reference point and one lower patella reference point, with the upper and lower patella reference points being localized on a patella edge area.
  • the point cloud particularly preferably includes a third reference point, so that the three reference points form a triangle whose area covers the patella as best as possible. The three reference points mentioned are then - as already mentioned - preferably on the edge of the kneecap.
  • a further expedient embodiment of the invention is characterized by a control unit of the downstream machine that is set up to determine a cutting line course of the localized bony parts from the position data provided, with the control unit also being designed to move a control-movably set up knife of the downstream machine along this for processing the poultry legs To move the course of the cutting line in order to debone the poultry leg.
  • the poultry legs can be deboned in an optimal manner.
  • the cutting line can be brought as close as possible to them in order to leave as little residual flesh as possible on the bony parts.
  • the position of the incision which is usually in the area of the kneecap, can also be precisely determined in this way. The present invention thus makes it possible to debone poultry legs fully automatically.
  • the device further comprises a leg side recognition device which is designed to carry out leg side recognition on the basis of the captured digital images of the poultry legs before providing them as input data to the first neural network and to determine whether the respective digital images are from a right or a left Poultry leg originates and, if the digital image does not correspond to a predetermined leg side, to mirror the image data of the respective digital image about a virtual axis in order to convert the digital image of a right poultry leg into an apparent digital image of a left poultry leg and vice versa.
  • the first neural network it suffices to design the first neural network in such a way that it recognizes only one type of leg, namely either a left or a right poultry leg.
  • the neural network is designed to process left poultry legs
  • digital images of right poultry legs are first mirrored as described. Due to the reflection, the partial image of the poultry leg then appears as if it had been taken from a left poultry leg. In this way it is possible, using the image data from both the right and left poultry legs, to localize the bony parts in poultry legs using the neural network designed for processing left poultry legs. If the neural network is designed for processing right poultry legs, the procedure is reversed accordingly and the partial images of left poultry legs are mirrored as described before processing by the neural network.
  • the leg side recognition includes a second neural network that has been trained with images of poultry legs of the specified leg side.
  • the second neural network is preferably trained exclusively with images of left or right poultry legs, respectively. If the second neural network is designed to recognize left poultry legs, for example, it indicates that no left leg has been recognized when an image of a right poultry leg is input. Not recognizing a left poultry leg implies that it is a right leg.
  • a further expedient embodiment of the invention is characterized in that the device further comprises a front/back side recognition device, which is set up before the digital images of the poultry legs are provided as input data to the first neural network and/or before the leg side recognition is carried out, a front -/back side recognition and to determine whether the respective digital image shows the front or the back of the poultry leg and, if the digital image does not correspond to a predetermined front/back side, a hanging receptacle of the conveyor device that holds the respective poultry leg and can be swiveled and pivoted about its vertical axis to leave, perform a 180° turn and capture a digital image of the side of the poultry leg pointing to the first optical imager by means of a second optical imager, which is downstream of the first optical imager with respect to the conveying direction.
  • a front/back side recognition device which is set up before the digital images of the poultry legs are provided as input data to the first neural network and/or before the leg side recognition is carried out, a front -/back side recognition and to
  • the front side of the poultry legs is preferably always chosen as the specified side, because this offers better clues for localizing the bony parts due to the greater proportion of muscle meat and the resulting curved surface in the digital images.
  • the front/back recognition device includes a third neural network that has been trained with images of poultry legs of the specified front/back.
  • the third neural network is preferably trained exclusively with images of the front or back of poultry legs. For example, if the third neural network is designed to recognize front sides of poultry legs, when an image of a back side of poultry legs is input, this indicates that no front side has been recognized. Not recognizing a front side implies that it is the back side.
  • a non-transitory computer-readable storage medium comprising a program comprising instructions for causing the computer to carry out the above-described method for locating bony parts present in a fowl leg.
  • FIG. 1 shows a plan view of the arrangement according to the invention for training at least one neural network for localizing bony parts present in a wing bone 10,
  • FIG. 2 shows a schematic view of a first optical imager and a poultry leg located in front of it
  • 3 shows a schematic view of a hybrid image based on the hybrid image data
  • 4 shows a schematic view of the device according to the invention for locating bony parts
  • Fig. 5 is a block diagram
  • Fig. 7 is a side view of the device shown in Fig. 4 and
  • FIG. 8 shows a schematic representation of the front/back side detection device.
  • FIG. 1 shows a top view of the arrangement according to the invention for training at least one neural network for localizing bony parts present in a poultry leg 10 .
  • these bony parts are in particular the thigh bone 11, the lower leg bone 12 and the kneecap.
  • the following explanations serve both to explain the arrangement mentioned and to describe in more detail the method according to the invention for training the neural network mentioned.
  • the invention comprises an optical camera 14 which is set up to take pictures of the front or back sides of the poultry legs 10 in the optically visible wavelength range.
  • the optical camera 14 is thus designed to generate optical image data for each of the poultry legs 10 .
  • the front of the poultry legs 10 is preferably aligned with the optical camera 14 so that only images of the front of the poultry leg are recorded.
  • the poultry legs 10 with their Rear side to the optical camera 14 are aligned. In this case, only images of the back of the poultry leg are taken.
  • the poultry legs 10 can be transported in the conveying direction 16, for example by means of a conveying device—not shown in the drawing. However, it is also possible for the poultry legs 10 to be positioned in front of the optical camera 14 manually.
  • the invention according to the arrangement further comprises an X-ray source set up for irradiating the back or front of the poultry legs 10 with X-ray radiation 17
  • the X-ray image generator 19 is arranged on the side of the poultry legs 10 facing away from the X-ray source 18 and is designed to generate X-ray image data. In this way, X-ray image data on each of the poultry legs 10 is generated.
  • the optical image data obtained and the X-ray image data form the basis on which the first neural network is trained.
  • the optical image data and the X-ray image data are fed to a superimposition unit - not shown in the drawing - which is set up to superimpose the optical image data of one of the poultry legs 10 with the X-ray image data of the same poultry leg 10 in order to produce hybrid image data on each of the poultry legs 10 to create.
  • the hybrid images 15 for each of the poultry legs 10 therefore represent an overlay image obtained by superposition, in which the position of the bony parts, in particular the femur 11, the lower leg bone 12 and the patella, together with the external shape of the poultry leg 10 are visible.
  • the images are preferably recorded by means of the optical camera 14 and the X-ray image generator 19 in such a way that the respective recorded image sections of the wing bone are already as congruent as possible with one another.
  • the superimposition is preferably additionally set up to produce such congruence of the image sections.
  • the reference points 20 for identifying the positions of the bony parts are first established using the X-ray image data.
  • the determination can be made, for example, by an inspector or semi-automatically. subsequently become the positions of the bony parts determined in this way are overlaid with the optical image data and the hybrid image data for each of the poultry legs 10 is thus generated.
  • the method according to the invention and the device for training the neural network also include the display of the hybrid image data by means of a display and input device—not shown in the drawing. Based on the displayed hybrid data, reference points 20 are now defined, for example by an inspector or semi-automatically, which are used to identify the position of the bony parts. The reference points 20 are entered via the input device.
  • Figure 3 shows an example of a schematic view of such a hybrid image 15 based on the hybrid image data.
  • the present invention further comprises a neural network (not shown in the drawing) and a learning cycle control unit.
  • the learning cycle control unit is designed and set up to input the optical image data and the reference points 20, preferably the reference points 20a, 20b, 20c, 20d, 20e, 20f, 20g, as training data for the neural network.
  • the optical image data thus form the input for the neural network
  • the reference points 20 each correspond to the output data of the neural network expected for the respective optical image data and thus form the target data.
  • the learning cycle control unit is configured to repetitively adjust the weights of the neural network based on the difference between the target data and the output data generated by the neural network.
  • the neural network is preferably a multi-layer neural network with a corresponding number of hidden layers.
  • the weights are preferably adjusted during training using the stochastic gradient descent method. For example, the mean squared error from the difference between the target data and the output data generated by the neural network is used as a loss function.
  • the reference points 20 each include two points for identifying the thigh 11 and lower leg 12 and three points for identifying the patella.
  • the thigh reference points 20a, 20b mark the position of the thigh bone 11, the lower leg reference points 20c, 20d the position of the lower leg bone 12 and the patella reference points 20e, 20f and 20g the position of the patella.
  • the present invention is not limited to the number of reference points 20 mentioned. Rather, it is also possible to specify more reference points 20 .
  • the thigh reference points 20a, 20b and the lower leg reference points 20c, 20d each form pairs of points. These pairs of points preferably characterize the position of the bone end regions 21.
  • the bone end regions 21 each designate those regions of a bone where the joint heads are located.
  • the patella reference points 20e, 20f, 20g preferably form a point cloud 22, the points of which reference edge positions of the patella.
  • the point cloud 22 includes at least one of the patellar reference points 20e, 20f, 20g. However, the point cloud 22 preferably comprises the three patella reference points 20e, 20f, 20g shown in FIG. 3, which reference the edge of the patella.
  • the point cloud 22 comprises at least two of the reference points 20e, 20g, namely a lower patellar reference point and an upper patellar reference point.
  • the superimposition unit is preferably set up to free object-related image areas of the optical image data and the x-ray image data from the image background before the hybrid image data are generated. In other words, image areas that only represent the background are masked in the respective data.
  • the present invention also relates to a non-transitory computer-readable storage medium having a program comprising instructions for causing the computer to carry out the above-described method for training the neural network.
  • All common storage types can be used as a storage medium, with for example, CD-ROMs, DVDs, memory sticks, hard drives or cloud storage services.
  • the present invention also includes an apparatus and method for locating bony parts present in the fowl leg 10 .
  • the device comprises a conveying device 23 which is formed out for conveying the poultry legs 10 in a conveying direction 16 .
  • Digital images 25 of the front or back sides of the poultry legs 10 are recorded by means of a first optical image generator 24, which is shown schematically in FIG. It has turned out to be particularly advantageous to always capture the front sides of the poultry legs 10 . In principle, however, it is possible to always capture the backs of the poultry legs 10 .
  • the device further comprises a first neural network--not shown in the drawing--which is designed to localize the bony parts in the poultry leg 10.
  • the first neural network has previously been trained by means of the method according to the invention for training a neural network for locating bony parts present in a poultry leg.
  • the device according to the invention further comprises an input unit--not shown in the drawing--which is set up to provide the digital images sequentially as input data to the first neural network.
  • the first neural network preferably receives the digital images of the front sides of the poultry legs 10 as input data.
  • the correspondingly trained first neural network is set up to determine position data 31 of the bony parts based on this input data and to provide this determined position data 31 for display and/or forwarding to a downstream machine 26 for processing the poultry legs 10 on the basis of the determined position data 31 .
  • the steps of the method can also be seen from the block diagram according to FIG.
  • the poultry legs 10 conveyed in the conveying direction 16 by means of the conveying device 23 pass through the first optical image generator 24, by means of which digital images 25 of the front or back sides of the poultry legs 10 are recorded at given times.
  • the digital images 25 are subjected to pre-processing 28 and their alignment is then checked.
  • the pre-processing 28 of the digital images 25 as well as the implementation of steps 29, 30 are purely optional. On white- More details of the above steps are discussed below.
  • the position data 31 is determined by means of the first neural network.
  • the poultry leg 10 is illuminated at least at the time of recording the optical image data, preferably by means of a flashlight.
  • the shape of the neural network can be varied. In principle, all multi-layer networks can be considered.
  • a network structure with 29 layers has proven to be particularly advantageous with regard to the recognition accuracy and at the same time a justifiable algorithmic effort.
  • these 29 layers sixteen are preferably two-dimensional convolution layers. More preferably, the convolutional layers are divided into four blocks, each followed by a max-pooling and a dropout layer.
  • the last layer which uses a sigmoid function, all other layers are activated using a rectifier function.
  • the input layer of the first neural network preferably forms an input layer which is set up to process the digital images with a resolution of preferably 300 ⁇ 300 pixels.
  • the output layer of the first neural network preferably comprises fourteen nodes, each representing the x and y coordinates of the seven reference points 20 .
  • the first neural network is also preferably set up to carry out all calculations using floating-point arithmetic. In particular, the calculations are carried out using floating-point numbers, preferably of the “float” type, with a resolution of 16 or 32 bits.
  • the first neural network is also preferably designed using a multiplicity of processors for parallel computation.
  • the reference points 20 each include two points for identifying the thigh 11 and lower leg 12 and three points for identifying the patella.
  • the thigh reference points 20a, 20b mark the position of the thigh bone 11, the lower leg reference points 20c, 20d the position of the lower leg bone 12 and the patella reference points 20e, 20f and 20g the position of the patella.
  • the present invention is a matter of course not limited to the number of reference points 20 mentioned. Rather, it is also possible to specify more reference points 20 .
  • the thigh reference points 20a, 20b and the lower leg reference points 20c, 20d each form pairs of points. These pairs of points preferably characterize the position of the bone end regions 21.
  • the patella reference points 20e, 20f, 20g preferably form a point cloud 22, the points of which reference the edge position of the patella.
  • the point cloud 22 includes at least one of the kneecap reference points 20e, 20f, 20g.
  • the point cloud 22 preferably comprises the three patella reference points 20e, 20f, 20g shown in FIG. 3, which reference the edge of the patella.
  • the point cloud 22 comprises at least two of the reference points 20e, 20g, namely a lower patellar reference point and an upper patellar reference point.
  • the method according to the invention preferably also includes determining a cutting line course from the position data provided by means of a control unit (not shown in the drawing) of the downstream machine 26.
  • a control-movably installed knife is moved along this cutting line course in order to cut the poultry leg 10 to Bone.
  • FIG. 6 shows a flowchart for the pre-processing 28 mentioned at the outset.
  • Step 28 preferably consists in the digital images 25 first being subjected to a lens correction 38 .
  • the background masking 39 is then advantageously carried out in order to release the image area in which the poultry leg 10 is visible.
  • the resolution of the digital image is reduced, for example to a resolution of 300 ⁇ 300 or preferably 100 ⁇ 100 or 128 ⁇ 128 pixels.
  • the present invention is but not limited to the mentioned resolutions. Rather, any other resolution reduction can be used.
  • a color conversion 41 then takes place, for example in a BGR or gray scale image.
  • a conversion to floating point numbers is then preferably carried out by means of a floating point conversion step 42.
  • the recorded digital images 25 of the poultry legs 10 are supplied to a leg side recognition device 32 shown in FIG. 5 as input data to the first neural network before being provided. This is set up to carry out leg side recognition and to determine whether the respective digital image 25 originates from a right or a left poultry leg.
  • a leg side recognition device 32 shown in FIG. 5 This is set up to carry out leg side recognition and to determine whether the respective digital image 25 originates from a right or a left poultry leg.
  • it is preferably only trained on one type of poultry legs 10, namely either left or right poultry legs 10.
  • the method according to the invention and the device are advantageously set up to automatically determine whether the recorded digital image 25 originates from a right poultry leg 10 or a left poultry leg 10 . If the digital image 25 does not match a specified leg side, the leg side detection device 32 is set up to mirror the image data of the respective digital image 25 about a virtual axis in order to convert the digital image 25 of a right poultry leg 10 into an apparent digital image 25 of a left poultry leg 10 and convert in reverse.
  • the digital image 25 is not changed by means of the leg side recognition device 32 if it recognizes that it is a left poultry leg 10.
  • the digital image 25 is then passed on via the signal flow arrow 34, as shown in FIG.
  • the digital image 25 thus arrives along the signal flow arrows 34 or 36 either directly as an input to the first neural network for locating the bony parts in step 27 or after the image mirroring.
  • the side of the leg is preferably recognized by means of a second neural network that has been trained with images of poultry legs 10 of the specified side of the leg, ie with left legs, for example.
  • the leg side detection device 32 thus preferably includes the second neural network. If a digital image 25 is recognized which corresponds to the specified side of the leg, the digital image 25 remains unchanged--as previously described. If the second neural network does not recognize the specified side of the leg, the digital image 25 is mirrored.
  • the digital images 25 of the poultry legs 10 are supplied as input data to the first neural network and/or to a front/back side detection device 37 before the leg side detection.
  • the front/back side recognition device 37 is designed to determine whether the respective digital image 25 shows the front side or the back side of the poultry leg 10 . If the digital image 25 does not match a predetermined front/back side, this is set up to cause the poultry leg 10 to rotate in such a way that the respective other side is directed towards the first optical image generator 24 . This process is shown schematically in FIG.
  • the conveying device 16 comprises a multiplicity of hanging receptacles 43 which are each designed and set up to receive one of the poultry legs 10 .
  • the hanging receptacles 43 are each designed to be pivotable about their vertical axis 44 .
  • Such a hanging receptacle 43 is shown in detail in the side view of the conveyor device 16 in FIG. If the front/back recognition device 37 has determined that the desired side of the respective poultry leg 10 was not directed towards the first optical imaging device 24, this causes the respective hanging receptacle 43 of the conveyor device 16 to rotate 180° to be executed in step 46.
  • the poultry leg side now pointing to the second optical imager 45 is then detected by means of a second optical imager 45, which is arranged downstream of the first optical imager 24 with respect to the conveying direction 16.
  • the front/back side recognition is preferably carried out by means of a third neural network which is fed with images of poultry legs of the predetermined front/back page has been trained. Consequently, the front/back recognition device preferably comprises the third neural network.
  • a non-transitory computer-readable storage medium comprising a program comprising instructions for causing the computer to carry out the method for locating bony parts present in the fowl leg 10.
  • the lenses of the optical camera 14 and the first and second optical imaging devices 24, 45 include polarization filters. These are designed to reduce possible reflections caused by damp or wet surfaces of the poultry legs 10, for example.

Landscapes

  • Engineering & Computer Science (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Food Science & Technology (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • General Health & Medical Sciences (AREA)
  • Zoology (AREA)
  • Wood Science & Technology (AREA)
  • Chemical & Material Sciences (AREA)
  • Evolutionary Computation (AREA)
  • Software Systems (AREA)
  • Biochemistry (AREA)
  • Pathology (AREA)
  • Immunology (AREA)
  • Multimedia (AREA)
  • Analytical Chemistry (AREA)
  • Computing Systems (AREA)
  • Medical Informatics (AREA)
  • Artificial Intelligence (AREA)
  • General Engineering & Computer Science (AREA)
  • Databases & Information Systems (AREA)
  • Biophysics (AREA)
  • Robotics (AREA)
  • Automation & Control Theory (AREA)
  • Medicinal Chemistry (AREA)
  • Human Computer Interaction (AREA)
  • Toxicology (AREA)
  • Data Mining & Analysis (AREA)
  • Mathematical Physics (AREA)
  • Image Analysis (AREA)
  • Image Processing (AREA)
  • Processing Of Meat And Fish (AREA)

Abstract

La présente invention concerne un procédé et un agencement pour entraîner au moins un réseau neuronal à localiser des parties osseuses présentes dans une cuisse de volaille (10). De plus, la présente invention concerne un procédé et un dispositif de localisation de parties osseuses présentes dans une cuisse de volaille (10), le procédé comprenant les étapes de transport des cuisses de volaille (10) dans une direction de transport (16) au moyen d'un transporteur (23), de capture, au moyen d'un premier système d'imagerie optique (24), d'images numériques (25) du côté avant ou du côté arrière de chaque cuisse de volaille (10) transportée de manière à passer dans le système d'imagerie (24), de fourniture séquentielle des images numériques (25) en tant que données d'entrée à un premier réseau neuronal conçu pour localiser les parties osseuses, le premier réseau neuronal ayant été entraîné à localiser les parties osseuses conformément à un procédé selon l'une quelconque des revendications 1 à 6, et de détermination des données de position (31) des parties osseuses au moyen du premier réseau neuronal et de fourniture des données de position (31) pour l'affichage sur une machine aval (26) et/ou la transmission à celle-ci pour traiter les cuisses de volaille (10) sur la base des données de position déterminées.
EP21749591.0A 2021-07-26 2021-07-26 Procédés, dispositifs et agencement pour localiser des parties osseuses présentes dans une cuisse de volaille Pending EP4376624A1 (fr)

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/EP2021/070826 WO2023006173A1 (fr) 2021-07-26 2021-07-26 Procédés, dispositifs et agencement pour localiser des parties osseuses présentes dans une cuisse de volaille

Publications (1)

Publication Number Publication Date
EP4376624A1 true EP4376624A1 (fr) 2024-06-05

Family

ID=77180002

Family Applications (1)

Application Number Title Priority Date Filing Date
EP21749591.0A Pending EP4376624A1 (fr) 2021-07-26 2021-07-26 Procédés, dispositifs et agencement pour localiser des parties osseuses présentes dans une cuisse de volaille

Country Status (7)

Country Link
US (1) US12535436B2 (fr)
EP (1) EP4376624A1 (fr)
KR (1) KR20240037319A (fr)
CN (1) CN117651490B (fr)
AR (1) AR126562A1 (fr)
AU (1) AU2021457761B2 (fr)
WO (1) WO2023006173A1 (fr)

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2025040251A1 (fr) * 2023-08-22 2025-02-27 Fpi Food Processing Innovation Gmbh & Co. Kg Procédé et dispositif d'analyse de carcasses ou de parties de celles-ci, procédé d'entraînement d'au moins un réseau neuronal pour l'analyse de carcasses ou de parties de celles-ci
WO2025104182A1 (fr) * 2023-11-14 2025-05-22 Marel Poultry B.V. Systèmes et procédés d'analyse de séparation de moitié postérieure de volaille en une partie dos, une patte gauche et une patte droite
FR3160616B1 (fr) 2024-03-27 2026-02-27 Clextral Filière pour l’extrusion d’une matière riche en protéines et en eau, ainsi que système de préparation en continu d’un produit alimentaire extrudé, comprenant une telle filière
WO2026037504A1 (fr) * 2024-08-16 2026-02-19 Fpi Food Processing Innovation Gmbh & Co. Kg Appareil et procédé de traitement d'une pluralité de parties de volaille, en particulier de pattes de volaille

Family Cites Families (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
DK2532246T3 (en) * 2010-10-27 2015-10-05 Maekawa Seisakusho Kk Deboning and deboning the meat with bone using X-rays
EP3275313B1 (fr) * 2016-07-29 2024-02-14 Nordischer Maschinenbau Rud. Baader GmbH + Co. KG Dispositif pour saisir et évaluer des informations spécifiques aux produits sur des produits de l'industrie de traitement des aliments et système comprenant un tel dispositif et procédé pour traiter des produits de l'industrie de traitement des aliments
US20210204553A1 (en) 2018-06-01 2021-07-08 Cryovac, Llc Image-data-based classification of meat products
CN109637633B (zh) 2018-11-17 2023-02-07 广州航海学院 一种基于大数据与机器学习的诊断乳腺癌病情的方法
JP2020183876A (ja) * 2019-04-26 2020-11-12 株式会社前川製作所 特徴点認識システムおよびワーク処理システム
CN110503635B (zh) 2019-07-30 2022-04-05 浙江工业大学 一种基于异构数据融合网络的手骨x光片骨龄评估方法
CN111062947B (zh) 2019-08-14 2023-04-25 深圳市智影医疗科技有限公司 一种基于深度学习的x光胸片病灶定位方法及系统
CN112200780B (zh) 2020-09-29 2023-09-29 上海联影医疗科技股份有限公司 骨组织定位方法、装置、计算机设备和存储介质

Also Published As

Publication number Publication date
KR20240037319A (ko) 2024-03-21
WO2023006173A8 (fr) 2023-11-02
CN117651490A (zh) 2024-03-05
CN117651490B (zh) 2025-11-21
AU2021457761B2 (en) 2025-06-05
US20250052699A1 (en) 2025-02-13
WO2023006173A1 (fr) 2023-02-02
US12535436B2 (en) 2026-01-27
AR126562A1 (es) 2023-10-25
WO2023006173A9 (fr) 2023-04-06
AU2021457761A1 (en) 2024-02-22
CA3174181A1 (fr) 2023-01-26

Similar Documents

Publication Publication Date Title
WO2023006173A1 (fr) Procédés, dispositifs et agencement pour localiser des parties osseuses présentes dans une cuisse de volaille
EP3883716B1 (fr) Détection d'erreurs d'usinage d'un système d'usinage laser à l'aide d'un réseau neural convolutionnel profond
DE19527518B4 (de) Röntgenröhren-Strommodulation während der Computertomographie-Abtastung
EP0880109B1 (fr) Procédé pour déterminer la transformation entre un objet et sa représentation tridimensionnelle et dispositif pour la mise en oeuvre du procédé
DE102012204019B4 (de) Verfahren zur Reduzierung von Bewegungsartefakten
EP3210537B1 (fr) Fabrication d'une image tridimensionnelle d'une partie de corps au moyen d'un appareil à rayons x
DE602004009653T2 (de) Vorrichtung, Verfahren, Programm und maschinenlesbares Speichermedium zur Strahlungsbildverarbeitung
DE112007001360T5 (de) Stent-Inspektionssystem
US20210321968A1 (en) X-ray image quality control system
EP1894538A1 (fr) Procédé et appareil pour la détermination de la position des plans pelviens
EP3228250A1 (fr) Dispositif et procédé de fabrication d'une radiographie panoramique
DE102016219887A1 (de) Verfahren und System zur Nutzung von Messdaten
DE102017215059B3 (de) Verfahren zum Betrieb einer bildgebenden Röntgeneinrichtung, insbesondere Computertomograph, Röntgeneinrichtung und Computerprogrammprodukt
DE112020002362T5 (de) Medizinische Bildverarbeitungsvorrichtung, medizinisches Bildverarbeitungsprogramm, medizinisches Gerät und Behandlungssystem
DE3336470A1 (de) Verfahren und anordnung zum vergleich von bildern
WO2024120894A1 (fr) Procédé de soudage correcteur et dispositif de traitement de joint pour soudage correcteur
DE19748082A1 (de) Verfahren und Vorrichtung zur Erfassung von Teilvolumen-Bildartefakten
DE102005022544A1 (de) Verfahren und Einrichtung zum Aufnehmen eines digitalen Röntgenbildes
DE102018211106A1 (de) Verfahren zur Reduktion eines Artefakts in einem Computertomographie-Bilddatensatz
DE102013215043A1 (de) Verfahren zur Bildgebung mittels eines Röntgengeräts und Röntgengerät
WO2011003408A1 (fr) Procédé et dispositif pour la représentation par imagerie tridimensionnelle à résolution spatiale d'informations complètes sur les défauts dans des corps cristallins massifs au moyen de rayons x durs
DE102023210809B4 (de) Bereitstellen angepasster Aufnahmeparameter
DE102021125839A1 (de) Optoelektronischer Sensor und Verfahren zum Absichern einer Maschine
DE112021007285T5 (de) Röntgenbildgebungsvorrichtung und röntgenbildverarbeitungsverfahren
DE102021201600A1 (de) Verbesserung eines Satzes Abbildungsparameter bei der Transmissions-Röntgenbildgebung

Legal Events

Date Code Title Description
STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: UNKNOWN

STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: THE INTERNATIONAL PUBLICATION HAS BEEN MADE

PUAI Public reference made under article 153(3) epc to a published international application that has entered the european phase

Free format text: ORIGINAL CODE: 0009012

STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: REQUEST FOR EXAMINATION WAS MADE

17P Request for examination filed

Effective date: 20240213

AK Designated contracting states

Kind code of ref document: A1

Designated state(s): AL AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC MK MT NL NO PL PT RO RS SE SI SK SM TR

RAP3 Party data changed (applicant data changed or rights of an application transferred)

Owner name: FPI FOOD PROCESSING INNOVATION GMBH & CO. KG

DAV Request for validation of the european patent (deleted)
DAX Request for extension of the european patent (deleted)