WO1998030886A1

WO1998030886A1 - Apparatus and method for quantifying physical characteristics of granular products

Info

Publication number: WO1998030886A1
Application number: PCT/CA1998/000007
Authority: WO
Inventors: Stefan Bussmann; Kathleen A. Harrigan; Bruce Hodgins
Original assignee: Maztech Microvision Ltd
Priority date: 1997-01-07
Filing date: 1998-01-07
Publication date: 1998-07-16
Also published as: CA2194534A1; AU5545498A

Abstract

An apparatus and method for quantifying physical characteristics of granular products includes the placing of product samples into a sample holder, and compression of the product samples against a clear window for identification by the image acquisition sub-system. Using custom software architecture, an electronic image is generated, captured, processed, and specific physical characteristics are quantified based on calibration-trained expert-user criteria.

Description

APPARATUS AND METHOD FOR QUANTIFYING PHYSICAL CHARACTERISTICS OF GRANULAR PRODUCTS

FIELD OF INVENTION

The invention generally relates to the identification and quantification of physical characteristics such as size, shape, colour, and hardness of granular products such as whole kernels, flour, millstreams, semolina, pasta, noodles, spray-dried powders.

BACKGROUND OF THE INVENTION

Physical characteristics of particular material, such as whole kernels, are difficult to quantify using traditional means. Assessment of kernel morphology, for example, is critical to plant breeders in order to select for breeding quality. Such assessment must be performed using a nondestructive method due to a very limited number of kernels in a genotype; the kernels measured must be left intact for planting. Kernel morphology assessment is also necessary for flour millers in order to estimate flour yield based on kernel size and/or shape, and for grain graders to estimate the value and type of kernels based on colour and hardness (i.e. vitreosity) . Physical characteristics that are directly related to quality parameters include minimum and maximum axis length, area, perimeter, aspect ratio, ellipsoidal volume, colour, shape and vitreosity. Typically, an experienced plant breeder would make predictions on the milling quality of kernels based on visual examination of whole kernel samples. This might include manual measurement of the width and length of individual kernels, but is often performed without the aid of measuring devices. Visual assessment is based on the skill and experience of the breeder, miller, or grader, and is highly subjective. Standard procedures do not exist, and generally whole kernel assessment is thought to be more art than science. Whole kernel quality predictions based on chemical methods such as near infrared (NIR) spectroscopy are determinations of moisture, protein, fat, fibre and ash content (i.e. composite analysis) . NIR methods do not measure the physical characteristics of whole kernels. Visual assessment of whole kernels cannot yield quantitative results and subjective analysis cannot be standardized across the industry.

Hardness is a physical characteristic of whole kernels that is highly desirable. Hard kernels are called vitreous, while soft kernels are termed opaque. Hard kernels have better nutritional characteristics, are better for dry milling, and have superior breakage resistance and pathogen resistance qualities than soft or opaque kernels. Vitreosity assessment is important for plant breeders, flour millers and grain graders for the same reasons as indicated previously. Traditional methods for evaluating vitreosity include density determination by ethanol displacement and flotation on salt solutions, both of which are non-destructive approaches. Destructive methods such as milling and grinding techniques are also used. Another method of evaluation of vitreous genotypes has been to view the kernels with candling or backlighting, and to visually score the kernels based on the ratio of vitreous to floury endosperm. Visual classification using this procedure is highly subjective and not reproducible, and relies primarily on the skill, experience and visual acuity of the investigator .

Assessment of physical characteristics is applied to the smallest components of food products as well. Particle size analysis is performed on macro molecules such as starch granules which are the reduction product from milling whole kernels into flour. Size is typically measured as granule diameter, assuming the particles are spherical and therefore circular in cross-section. Particle size is an important physical parameter in food systems and is directly related to flavour and texture interactions, colour and appearance, physicochemical interactions, mixing and blending performance, uniformity of reactions, and process monitoring. Particle size methods vary considerably and are designed for specific product applications. Types of particle size methods used in the food industry include: sieve analysis (wet, dry), microscopic analysis (optical, electron), zone sensing (resistivity, optical) , sedimentation (gravity, centrifugal) , elutriation

(laminar flow, cyclone) , and chromatography (packed column, capillary tube) . The method employed will depend on the type of material to be measured and its particle size range.

Sieve analysis is routinely employed in the grain-based food and ingredient industries. Sieving is a dynamic process based on physical separation. As the material is forced through a mesh screen, it is automatically separated by size. After a specified period of time, material remaining on the screen is assumed to be larger than the mesh opening, and material passing through the screen is assumed to be smaller than the mesh opening. This assumption determines the particle size as a result of a physical process rather than an actual measurement. Exact particle distributions, i.e. the frequency of particles of a particular size, cannot be determined, and therefore no predictive calculations can be made on the individual particles. Only broad particle size ranges can be determined and are based on minimum and maximum mesh openings which are dependent on fixed mesh sizes. Furthermore, the shape of particles cannot be determined since screen openings are designated by diameter, and therefore all particles may be assumed to be spherical. Regardless of sieve type, wire mesh screen sizes are limited to a nominal aperture of 20 μm, which prevents smaller particles from being sized.

Quantifying physical characteristics of food products includes assessment of particle contaminants that are the result of undesirable material being introduced in the grinding, mixing, extruding, spray-drying or other processing operation. For example, the milling of durum wheat produces a granular product comprised of evenly sized starchy endosperm particles called semolina, and is used exclusively in the manufacture of pasta and other paste products. Dark specks in the semolina adversely affect the appearance of the semolina and the finished pasta. Wheat bran is the most common source of brown specks, and black specks are usually caused by discoloured or diseased kernels or weed seeds. Speck counting is a mandatory quality control measurement for ensuring that the semolina meets customer specifications. Despite the importance of speck counting in determining the marketability, price and acceptance of semolina, there is no standard objective procedure for analysis. Specks are generally counted by a manual process where the observer visually identifies and counts the number of specks within a defined area of flattened semolina. Consistent, objective results are difficult to obtain due to observer bias in determining speck size and darkness of specks, observer experience and fatigue levels, inconsistent sample presentation, overall level of speckiness, and the tediousness of visual counting.

Spray-dried dairy powders are routinely analyzed for scorched particles which appear as dark specks in the material. Dark specks arise from the processing machinery used in spray-drying technologies. Standards for grades of dry milks are set by the American Dairy Products Institute (ADPI), and grading must be performed using one of four specific wet chemistry methods. Methods used are dependent on the material to be assessed and how they were processed. The approved standard protocols are: the water disc method for spray-process dry milk products, the Calgon (sodium hexametaphosphate) method for roller-process dry whole milk and nonfat dry milk, the sodium citrate method for roller-process dry whole milk and nonfat dry milk, and the EDTA method for roller-process dry buttermilk and dry buttermilk products. For all methods, the solution is filtered through a disk, and the disk is visually compared to an ADPI photoprint disk. Any filtered disk falling between two ADPI disks is assigned the higher disk's letter for grading purposes. Visual assessment is highly subjective and is dependent on the experience and skill of the technician and the colour quality of the photoprint. With time and exposure to light, the intensity of colour in the photoprint fades, requiring that the print be replaced.

Colour is used to identify, qualify and/or quantify one type of substance or structure from others associated with it in a solid matrix or in a dilute solution. Dyes or stains are usually applied to specific substrates according to precise protocols in order to mark the material to be measured. Once the marker has combined with its target, the detection method becomes all important. Many of the detection methods use spectral responses to qualify and quantify the marked material. Spectroscopy allows the absorbance or transmittance levels of a dilute solution to be quantified using specific wavelengths which correlate to the amount of marked material . Enzyme linked immuno selective assays (ELISA) are typical of this class of measurement and are rapid, automated objective determinations. Specific equipment, reagents procedures and laboratory facilities as well as trained personnel are required to implement the methodology. Another type of spectral detection is analysis with a microscope which is typically used for qualitative determinations where the presence or absence of a marked substance is noted. Bright field, dark field, phase contrast, and fluorescence modes are common microscopic configurations for qualitative analysis. None of these modes can achieve quantitative results. The skill and experience of the microscopist is essential for consistent qualitative analysis, but absolute amounts of a marked substance cannot be determined with this method.

Colour analysis of food and non-food products is often expressed as three primary elements: colour (hue), lightness or brightness (value) , and saturation or vividness (chroma) . The elements are used in a three dimensional system where hue is centred on the outside around the centre axis, value is the vertical axis, and chroma is the horizontal axis. The elements are defined in a 'colour notation system' as L*a*b*, where L* represents value, a* is hue and b* is chroma. This system requires a chromameter or colorimeter in order to generate the L*a*b* values which can be universally reproduced using the same equipment. The measuring area of the instrument is limited to only 50 mm, and multiple measurements must be made of a particular product. With pasta, an important colour to monitor is the greyness or brownness of the material, because consumer acceptance of these colours in pasta is extremely low. L*a*b* values do not measure the brown and grey tinges in pasta directly, and therefore visual assessment is used. Red-green- blue (RGB) values generated by colour CCD arrays have not been used extensively to measure the colour of food and non-food products either alone or in correlation with L*a*b* values. RGB colour can be digitally represented by up to 68 billion colours and is a function of the amount of information used in the system's memory (i.e. 36 bits per RGB channel) .

Quantification of physical characteristics of products are analytical procedures routinely included in the quality assurance and quality control process. Most of these analytical protocols are performed manually through visually subjective means, which are not reproducible, or through chemical and mechanical means which do not produce satisfactory results due to time, resource and information constraints. The use of an objective rapid automated quantification system is preferable to the subjective methods described above.

SUMMARY OF THE INVENTION

The present invention uses a novel combination of integrated hardware and software components to achieve an objective quantification system that is superior to known imaging systems for the identification and quantification of physical characteristics such as size, shape, colour, and hardness of products such as whole kernels, flour, millstreams, semolina, pasta, noodles, spray-dried powders.

It is therefore an object of the present invention to provide a method and apparatus to identify and quantify physical characteristics in granular products.

A further object is to provide an apparatus and method for producing objective, rapid, automated quantification of physical characteristics in granular products.

According to the invention, there is provided an apparatus for determining physical characteristics of a granular product sample, comprising: a sample holder for supporting a product sample; an image acquisition system for generating a digital image of the product sample .

According to the invention, there is further provided an apparatus comprising: a sample holder having a transparent window for supporting a product sample; an image acquisition system for generating a digital image of the product sample through the window.

According to the invention, there is further provided a method for determining physical characteristics of a granular product sample, comprising the steps of: placing the product sample onto a sample holder; generating a digital image of the product sample; identifying physical characteristics of the product sample from the image; and quantifying the characteristics .

Other advantages, objects and features of the present invention will be readily apparent to those skilled in the art from a review of the following detailed descriptions of a preferred embodiment in conjunction with the accompanying drawings and claims.

BRIEF DESCRIPTION OF THE DRAWINGS

Preferred embodiments of the present invention will now be described in greater detail, and will be better understood when read in conjunction with the following drawings, in which:

Figure 1 is a perspective view of a first embodiment of the present invention;

Figure 2 is a side elevation view of the sample holder of the first embodiment;

Figure 3 is a side elevation view of the image acquisition sub-system of the first embodiment;

Figure 4 is a top plan view of the image acquisition sub-system of the first embodiment;

Figure 5 is a side elevation view of the sample holder of a second embodiment of the present invention;

Figure 6 is a side elevation view of the image acquisition sub- system of the second embodiment; and

Figure 7 is a top plan view of the image acquisition sub-system of the second embodiment.

DETAILED DESCRIPTION OF THE INVENTION

Referring to Figures 1 to 4 , an apparatus 17, which is a combination of integrated hardware and software components, includes: a custom designed sample holder 1; a custom designed sample presentation mechanism 2; and an image acquisition sub- system 3.

The sample holder 1 has wing studs 13, acrylic cover 14, acrylic holder 15, and receptacles 16. The sample holder 1 is designed to contain and compress a granular product, such as whole kernels, flour, millstreams, semolina, pasta, noodles, spray-dried powders. A granular product is a product composed of particles of various shapes and sizes with grainy or smooth texture. The product is then presented to the image acquisition sub-system 3 through a clear window 4.

The sample holder 1 is made from three pieces of acrylic; the cover 14 is made of clear acrylic with an opaque vinyl applied to one surface; the holder 15 is made of two pieces of acrylic, one black and the other clear, which are glued together in the form of an open frame in order to hold the clear glass window 4.

The sample presentation mechanism 2 is a drawer assembly that accepts the sample holder 1 in a metal frame 5 with a front plate 6 and handle 7, and presents the clear window 4 to the glass surface 12 of the image acquisition sub-system 3. A drawer slide 11 includes a block 8 and rails 9 attached to an angled plate 10. A customized instrument housing (not shown) covers the sample presentation mechanism 2.

Referring to Figures 5 to 7, apparatus 19 includes: a custom designed sample holder 18; a custom designed sample presentation mechanism 26; and an image acquisition sub-system

20.

The sample holder 18 has acrylic cover 29 and acrylic holder 30. The sample holder 18 is designed to contain a granular product, such as whole kernels, flour, millstreams, semolina, pasta, noodles, spray-dried powders. The product is then presented to the image acquisition sub-system 20 through a clear window 21.

The sample holder 18 is made from three pieces of acrylic; the cover 29 is made of clear acrylic with an opaque vinyl applied to one surface; the holder 30 is made of two pieces of acrylic, one black and the other clear which are glued together in the form of an open frame in order to hold the clear glass window 21.

The sample presentation mechanism 26 is a drawer assembly with a light source 25 that accepts the sample holder 18 in a metal frame 22 with a front plate 23 and handle 24, and presents the clear window 21 to the glass surface 28 of the image acquisition sub-system 20. Drawer slides 27 are attached on both sides of sample presentation mechanism 26. A customized instrument housing (not shown) covers the sample presentation mechanism 26.

The apparatus 17, may also be adapted for use on-line by adding a slide gate for taking a sample from a flow stream of product and conveying the sample to the sample holder 1, 18 by, for example, pneumatic or screw conveyance.

The operation of the invention is rapid and automatic and the measurement results are accurate and reliable. For reflected light applications (see the first embodiment of Figures 1 to 4) , the process begins by filling the acrylic holder 15 with product and securing the acrylic cover 14 with the wing studs 13 into receptacles 16. The product becomes compressed against the clear glass window 4 and is placed in the drawer assembly 2 by fitting sample holder 1 into the metal frame 5. The drawer assembly 2 is closed by sliding the handle 7 attached to the front plate 6 along the angled plate 10 with a block 8 and rails 9. The sample in its sample holder 1 is presented to the image acquisition sub-system 3 glass surface 12 which generates and captures an electronic image of the sample. The physical characteristics of the product are isolated and quantified based on calibration-trained expert -user criteria.

For transmitted light applications (see the second embodiment of Figures 5 to 7) , the process begins by placing the product into the acrylic holder 30 which has an open cover 29 without the compression lid, wing studs or receptacles present in the first embodiment. The product is visible against the clear glass window 21 and is placed in the drawer assembly 26 by fitting holder 18 into the metal frame 22. The drawer assembly 26 is closed by sliding the handle 24 attached to the front plate 23 along the drawer slides 27. The sample in its sample holder 18 is presented to the glass surface 28 of the image acquisition sub- system 20 which generates and captures an electronic image of the sample. The physical characteristics of the product are isolated and quantified based on calibration- trained expert -user criteria.

Preparation of the sample placed in the sample holder can consist of preparing dry material so that all particles are packed together or in a fashion where the particles are physically separate from each other and in a specific orientation for digitization. Sample preparation can also consist of hydrating, mixing, extracting and/or staining individual particles in solution and physically separating them within the sample holder for digitization.

Using electronic imaging techniques, sample images are acquired and analyzed. The architecture of the system software is designed to quantify physical characteristics and include several aspects.

The image acquisition sub-system 3, 20 of the present invention generates and captures an electronic image of the product sample. For reflected light applications (see the first embodiment of Figures 1 to 4) , a line scan camera and fluorescent light tube automatically track below the glass window 4 of the sample holder 1 and scan the entire sample surface. For transmitted light applications (see the second embodiment of Figures 5 to 7) , a fluorescent light tube automatically tracks above the sample holder 18 while the line scan camera tracks below the glass window 21 of the sample holder 18. The entire sample surface is thus scanned.

The line scan camera of the first embodiment includes an array of CCD elements in a single row on a sensor, one or more analog-to-digital (ADC) converters, camera lens and camera housing mounted on a rolling track. While the light bounces off the sample and goes through the camera lens to the CCD elements, the camera scans one line at a time, with one CCD element for each pixel in a line. Each CCD element converts the light to an analog voltage and the ADC converts the analog voltage to a digital value using 8, 10, or 12 bits per colour.

The digital information is written to memory and blob analysis is performed. Blob analysis is an imaging technique whereby measurement attributes are defined as groups of pixels that meet specific grey level and particle size criteria and which must also be attached to each other or contiguous.

Specific algorithms are then applied for isolating and quantifying distinct physical characteristics. Visual assessment of the sample by the expert user is calibrated or matched to the digital information by accessing the blob analysis information of each physical characteristic. Using a point and click user interface approach, blob analysis information such as particle size minimum and maximum, and grey level minimum and maximum are viewed, selected and saved. The saved information is said to be calibration-trained to the expert user and is included in the algorithm used to isolate and quantify specific physical characteristics. An interactive user interface presents the system operator with a functional user interface to control and monitor the quantification process. This includes menus for standardizing the light source (using a calibrated colour tile) , selecting the region of interest, adjusting the illumination source (gain, offset) , defining calibration parameters (particle size and grey level) , selecting measurement parameters (number of sample holders counted, i.e. counted by holder or by area) , and determining results format (counts added or averaged) . An analysis reporting system tabulates the results and organizes the information for statistical analysis. Included are installation and configuration utilities for installing and configuring the software component of the invention on the host PC as well as an encryption code that protects the executable and limits access to it by using a site code that recognizes only one site license.

Counts and measurements determined by the algorithms include black and brown specks in semolina, corn grits, farina, flour, corn starch; dark scorched particles in spray-dried dairy powders; axis length and width of whole kernels; vitreousness or opaqueness of whole kernels; particle size of starch granules and specks; red-green-blue (RGB) colour of granulated products such as semolina, flour, corn starch.

Numerous modifications, variations and adaptations may be made to the particular embodiments of the invention described above without departing from the scope of the invention, which is defined in the claims.

Claims

THE EMBODIMENTS OF THE INVENTION IN WHICH AN EXCLUSIVE PROPERTY OR PRIVILEGE IS CLAIMED ARE DEFINED AS FOLLOWS:

1. An apparatus for determining physical characteristics of a granular product sample, comprising: a sample holder for supporting a product sample; an image acquisition system for generating a digital image of the product sample.

2. An apparatus as defined in claim 1, comprising: a sample holder having a transparent window for supporting a product sample; an image acquisition system for generating a digital image of the product sample through the window.

3. An apparatus as defined in claim 2, wherein the window is substantially horizontally oriented.

4. An apparatus as defined in claim 3, wherein the sample holder further comprises means for compressing the product sample against the window.

5. An apparatus as defined in claim 1, wherein the image acquisition system comprises a scanner.

6. An apparatus as defined in claim 5, wherein the scanner is a line scan camera.

7. An apparatus as defined in claim 6, wherein the image acquisition system comprises a light source for providing the line scan camera with light reflected off the product sample.

8. An apparatus as defined in claim 6, wherein the image acquisition system comprises a light source for providing the line scan camera with light transmitted through the product sample .

9. An apparatus as defined in claim 1, further comprising a presentation mechanism for positioning the sample holder relative to the image acquisition system.

10. An apparatus as defined in claim 9, wherein the presentation mechanism comprises a drawer.

11. An apparatus as defined in claim 10, wherein the drawer comprises a handle, a block, and at least one rail.

12. An apparatus as defined in claim 1, further comprising analyzing means for identifying and quantifying physical characteristics of the product sample from the image.

13. An apparatus as defined in claim 12, wherein the analyzing means comprises blob analysis means.

14. A method for determining physical characteristics of a granular product sample, comprising the steps of: placing the product sample onto a sample holder; generating a digital image of the product sample; identifying physical characteristics of the product sample from the image; and quantifying the characteristics.

15. A method as defined in claim 14, comprising the steps of: placing the product sample onto a sample holder; compressing the product sample; generating a digital image of the compressed product sample; identifying physical characteristics of the product sample from the image; and quantifying the characteristics.

16. A method as defined in claim 14, comprising the steps of: placing the product sample onto a sample holder; positioning the sample holder relative to an image acquisition system; generating a digital image of the product sample; identifying physical characteristics of the product sample from the image; and quantifying the characteristics.

17. A method as defined in claim 14, wherein the characteristics are quantified using blob analysis.