Machine vision inspection system and method for transparent containers
AssigneeBall Corporation
InventorAmir R. Novini
Life of this patent
Estimating the term…
Abstract
A system and method for optical inspection of the bottom surfaces of transparent containers comprises an image acquiring means, an illumination source and means for processing, storing and analyzing the image to search for and identify a baffle mark in the acquired image, and when found, to remove or erase the baffle mark from the image memory prior to the defect inspection process. Significant defects present in the area of the image being searched can be identified, even if the defects are located directly on the baffle mark, thus providing high sensitivity capabilities for defect detection while maintaining a low false rejection rate due to baffle marks.
Claims47
A method of inspecting a substantially transparent container being transported by a conveyor means, said container having a baffle mark formed in the bottom surface thereof, said method comprising the steps of:
- illuminating the bottom surface of said container;
- acquiring an image of the bottom surface of the container;
- identifying the baffle mark present in the image;
- analyzing the resulting image while ignoring the baffle mark to identify any remaining marks in the bottom surface of the container;
- determining whether any of said remaining marks are acceptable;
- and
- removing the container from the conveyor means if any of said remaining marks are unacceptable.
A system for inspecting a substantially transparent container having a baffle mark formed in its bottom surface, said system comprising:
- means for presenting the container to an inspection zone and therefrom to a rejection station;
- means for illuminating the bottom surface of the container at the inspection zone;
- means for generating an image corresponding to the bottom surface of the container;
- means for storing the sensed image;
- means for identifying the baffle mark present in the image;
- and
- means for processing the stored image of the bottom surface of the container, while ignoring the baffle mark for detecting unacceptable marks in the bottom surface of the container other than the baffle mark and for effecting the removal of unacceptable containers from the system at the rejection station.
An apparatus for inspecting the bottom surface of a glass container and determining whether the bottom surface is acceptable, said bottom surface having a baffle mark formed therein, said apparatus comprising:
- front end means for acquiring for analysis an image of the bottom surface;
- a vision engine for analyzing and processing the image, said vision engine being operable to identify marks in the image of the bottom surface and distinguish the baffle mark from other marks present in the image and to determine whether the other marks present are unacceptable;
- operator interface means for allowing an operator to communicate with the apparatus;
- and
- machine interface means for allowing the apparatus to communicate with remote machines.
A method employing a machine vision system including an electronic sensor and a computer for inspecting the bottom surface of a glass container having a baffle mark formed therein and for detecting and rejecting unacceptable containers, said glass container being transported by a conveyor means to an inspection zone and from the inspection zone to a removal zone, said electronic sensor including an image sensing means and a diffused light source, said method comprising the steps of:
- sensing when the container is positioned at the inspection zone;
- synchronizing the image sensing means and activating the light source;
- acquiring an image of the bottom surface of the container;
- searching a portion of the image within a predetermined number of annular boundaries at various points along the boundaries to identify edge transition points and determining whether a sufficient number of edge transition points have been identified,
- and if a sufficient number of edge transition points have not been identified, searching another boundary until a sufficient number of edge transition points are identified,
- and if a sufficient number of edge transition points are identified, storing in the computer all transition edge points identified and analyzing, in combinations of three, selected transition edge points to identify centers and radii of annular boundaries and storing the identified centers and radii in the computer;
- determining a standard deviation for the centers and radii, ignoring centers and radii outside of the standard deviation, and determining and storing the average center and radius in the computer, thereby determining the location and size of a baffle mark;
- searching in a second annular inspection area defined by the average center and radius for the image;
- identifying edge transition points on a plurality of lines within the annular boundary and determining if the image characteristics sensed in the second annular boundary corresponds with a baffle mark,
- and if the sensed characteristics correspond with a baffle mark, replacing the baffle transition edge points with an average gray shade value to remove the baffle mark from the image, thereby providing a normalized image ready for inspection;
- searching a plurality of annular boundaries within the resulting image for imperfections in the bottom surface until all annular boundaries have been searched and comparing identified imperfections with a predetermined set of acceptable parameters,
- and if the identified imperfections do not fall within the acceptable parameters, generating a rejection signal to remove the unacceptable container from the conveyor means.
Dependencies are inferred from the claim text.