Image analysis – Applications – Manufacturing or product inspection
Reexamination Certificate
2001-06-18
2004-01-27
Werner, Brian (Department: 2621)
Image analysis
Applications
Manufacturing or product inspection
Reexamination Certificate
active
06683975
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
This invention relates to the field of machine vision, and, more particularly, the use of machine vision to determine the dispersibility grade of a product in particulate form.
2. Discussion of the Art
Infant formula powder manufacturers have used dispersibility to describe the degree of firmness and solubility of particulate materials within a product. Such particulate materials include, for example, gel, mineral, and oil. These particulate materials may cause coagulation of the product or sedimentation in the product, both of which are not desired.
The conventional procedure for determining dispersibility comprises the following steps:
(a) providing a quantity of particulate material, e.g., infant formula powder, for each sample to be tested;
(b) introducing a quantity of tap water at a specified temperature into a container;
(c) applying the quantity of particulate material to the surface of the water and placing a cover on the container;
(d) agitating the container in such a manner, e.g., on a mechanical shaker, to disperse the particulate material to form a dispersion;
(e) pouring the dispersion through an 80 mesh screen;
(f) filling the container approximately ¼ full with tap water to dislodge any particulate material that may remain in the container, and pouring the resultant dispersion through the screen; and
(g) rating the particulate material remaining on the screen by referring to a series of standard photographs of samples of differing grades.
Each sample is tested in duplicate and the average reported as the final result.
This test and similar tests are performed manually. The results are highly dependent on the personal experience of the evaluator and are subject to deviation from evaluator to evaluator, location to location, or even day to day. Manual testing is time-consuming. Hiring experts to perform the tests is expensive. Furthermore, the test results cannot be retrieved and analyzed at a later time.
Accordingly, it is desired to develop a method for determining the dispersibility grade of a particulate material that will reduce discrepancies, reduce consumption of time, eliminate the need for experts, and allow retrieval and analysis of samples at a later time. It would also be desirable to develop a method for determining the dispersibility grade of a particulate material that can be performed at a location remote from the sample. It would also be desirable to develop a method for determining the dispersibility grade of a particulate material, which grade can be transmitted electronically as it is being determined.
Automated vision systems have been described in the art. See, for example, Gregory, “Planning a PC-Based Machine Vision System”, SENSORS, April 1998, pages 12-18 and Claude and Carritte, “Increasing product Yields with Automated Vision Systems”, Medical Device & Diagnostic Industry, May 2000, pages 158-169, both of which are incorporated herein by reference. However, the art has not addressed the problem of classifying the dispersiblity of particulate material according to a plurality of grade levels. The prior art is primarily concerned with a pass/fail decision and does not indicate a means for assigning samples into different categories.
SUMMARY OF THE INVENTION
This invention provides a method and apparatus for determining the dispersibility grade of particulate material by means of an image processing technique that employs machine vision.
In one aspect, this invention provides a method for grading the dispersibility of particulate material, the method comprising the steps of:
(a) obtaining an image, in digitized form, of a sample comprising reconstituted particulate material;
(b) performing a series of image processing steps to determine the presence and magnitude of blobs;
(c) performing a blob analysis to obtain data relating to the distribution of the particulate material in the sample;
(d) comparing the result of the blob analysis with data in a database; and
(e) determining the grade of said particulate material by means of the comparison.
The sample of particulate material, which is typically in the form of a powder, is reconstituted preferably by means of agitating a container filled with the sample. The series of image processing steps preferably includes, but is not limited to the following: identifying the location of the particulate material; assigning grey-scale values to individual pixels that form the image; determining which pixels, by connecting to one another have formed into blobs; and reducing noise. As used herein, the term “pixel” means a picture element in an array of cells that form an image.
The database suitable for use in this invention can be established by recording the results of a large number of tests performed by a powder quality grading authority. As used herein, a “powder quality grading authority” is a person in an organization having the authority to determine grades of quality of a particulate material. The recorded test results are analyzed by one or more statistical methods, whereby they are classified according to grade levels. The grade levels are associated with several ranges that have emerged from the statistical analysis. For example, the ranges can be characterized by measures of central tendency, that is, the moments of a distribution, the median, and mode. In running a test, a digitized image of a sample of particulate material is made as the material is residing on the surface of a screen. A plurality of physical properties of the sample of the particulate material can be recorded and stored in a database. Statistical analyses can be performed to assign a grade level to the particulate material. Such analyses involve comparing the physical properties of the sample of particulate material with the data in the database.
In another aspect, this invention provides an apparatus suitable for grading the dispersibility of particulate material. The apparatus comprises:
(a) a sample holder;
(b) a source of light having a specified frequency band;
(c) a bandpass filter;
(d) a camera, preferably a CCD camera;
(e) a frame grabber; and
(e) a computer.
The apparatus is typically commanded and controlled by software. In an alternative embodiment, the frame grabber can be incorporated into the camera, in which case, the camera and the frame grabber will be an integrated unit.
The method of this invention eliminates the need for human involvement in analyzing images, eliminates environmental influences (e.g., extraneous light), and eliminates subjective physical comparison. Accordingly, this invention eliminates human bias. The digitized images of samples and statistical results are retrievable for later analysis and future applications. The system can be used to provide standardized test results at a single production site or at a plurality of production sites. In addition, the test results can be transmitted to remote locations by means of computer networks, thereby allowing production decisions to be made rapidly.
REFERENCES:
patent: 5436980 (1995-07-01), Weeks et al.
patent: 5880830 (1999-03-01), Schechter
patent: 98/30886 (1998-07-01), None
PCT Search Report, PCT/US02/17566.
Standard Test Method for Bulk Solids Characterization by Carr Indices1, American Society For Testing and Materials, D 6393-99.
Do All Soy Formulas Shake Out? Powder Mixability Varies Greatly Among Soy Formulas, Ross Laboratories, Columbus, Ohio, 1987.
Robert Gregory, “Planning a PC-Based Machine Vision System”, Sensors, Apr. 1998, pp. 12-18.
E.J. Claude and R.M. Carrite, “Increasing Product Yields with Automated Visions Systems”, Medical Device & Diagnostic Industry, May 2000, pp. 158-169.
E.R. Davies,Machine Vision: Theory, Algorithms, Practicalities, 2ndEdition, Academic Press (1997), pp. 19-39, 79-130, 245-269, 348-371, 471-528, 561-601, 603-620, 677-695.
Mondiek David A.
Terflinger Randall Edwin
Wang Jianjun
Zeitler Wilson G.
Abbott Laboratories
Weinstein David I.
Werner Brian
LandOfFree
Apparatus and method for determining the dispersibility of a... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Apparatus and method for determining the dispersibility of a..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Apparatus and method for determining the dispersibility of a... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3195968