Abstract: A monitoring system implements a method for versatile and efficient detection and grading of deviations in packaging containers for liquid food in a manufacturing plant. The method comprises obtaining image data of a packaging container, or a starting material for use in producing the packaging container; analyzing the image data for detection of a current deviation; processing the current deviation in relation to a set of basis functions, which is associated with a deviation type of the current deviation, to obtain a current set of weights that represent the current deviation; and determining a current grading of the current deviation based on the current set of weights. The set of basis function may be pre-computed based on reproductions of packaging containers or starting material comprising different magnitudes of the deviation type.

Abstract: A monitoring system implements a method for versatile and efficient detection and grading of deviations in packaging containers for liquid food in a manufacturing plant. The method comprises obtaining image data of a packaging container, or a starting material for use in producing the packaging container; analyzing the image data for detection of a current deviation; processing the current deviation in relation to a set of basis functions, which is associated with a deviation type of the current deviation, to obtain a current set of weights that represent the current deviation; and determining a current grading of the current deviation based on the current set of weights. The set of basis function may be pre-computed based on reproductions of packaging containers or starting material comprising different magnitudes of the deviation type.

Abstract: A monitoring system implements a method for versatile and efficient training of a machine learning-based model for subsequent detection and grading of deviations in packaging containers for liquid food in a manufacturing plant. The method comprises creating a virtual model of a packaging container or of a starting material for use in producing the packaging container; obtaining probability distributions for features that are characteristic of a deviation type; producing reproductions of the virtual model with deviations included among the reproductions in correspondence with the probability distributions; associating gradings with the reproductions; and inputting the reproductions and the associated gradings for training of the machine learning-based model for subsequent detection and grading of an actual deviation in image data acquired in the manufacturing plant.

Abstract: A monitoring system implements a method for versatile and efficient training of a machine learning-based model for subsequent detection and grading of deviations in packaging containers for liquid food in a manufacturing plant. The method comprises creating a virtual model of a packaging container or of a starting material for use in producing the packaging container; obtaining probability distributions for features that are characteristic of a deviation type; producing reproductions of the virtual model with deviations included among the reproductions in correspondence with the probability distributions; associating gradings with the reproductions; and inputting the reproductions and the associated gradings for training of the machine learning-based model for subsequent detection and grading of an actual deviation in image data acquired in the manufacturing plant.

Abstract: A method for detection of deviations in packaging containers is disclosed, comprising creating a virtual model of a packaging container in a virtual coordinate system (x, y, z), defining a deformation zone on a surface of the virtual model, creating a defined deviation in the deformation zone having a defined geometry and coordinates in the virtual coordinate system (x, y, z) to create a controlled deformation of the virtual model, producing an image rendering of the virtual model with said controlled deformation to generate image features representing a deviation in the packaging container, associating the image features with different categories of deviations, and inputting the image features to a machine learning-based model for subsequent detection of categories of deviations in packaging containers in a packaging machine based on the image features. A system for detection of deviations is also disclosed.

Abstract: A method for detection of deviations in packaging containers is disclosed, comprising creating a virtual model (201) of a packaging container (401) in a virtual coordinate system (x, y, z), defining a deformation zone on a surface of the virtual model (201), creating a defined deviation in the deformation zone having a defined geometry and coordinates in the virtual coordinate system (x, y, z) to create a controlled deformation of the virtual model (201), producing an image rendering of the virtual model (201) with said controlled deformation to generate image features representing a deviation in the packaging container (401), associating the image features with different categories of deviations, and inputting the image features to a machine learning-based model for subsequent detection of categories of deviations in packaging containers (401) in a packaging machine (400) based on the image features. A system for detection of deviations is also disclosed.