Note: Descriptions are shown in the official language in which they were submitted.
CA 02380392 2002-O1-15
WO 01/06232 PCT/US00/19101
-1-
INTEGRATED OPTICS BLOCK FOR SPECTROSCOPY
RELATED APPLICATIONS
This application is a continuation of and claims
priority to U.S. Application No. 09/354,497 filed July
16, 1999, the entire teachings of which are incorporated
herein by this reference.
BACKGROUND OF THE INVENTION
It has been long recognized that the value of
agricultural products such as cereal grains and the like
are affected by the quality of their inherent
constituent components. In particular, cereal grains
with desirable protein, oil, starch, fiber, and moisture
content and desirable levels of carbohydrates and other
constituents can command a premium price. Favorable
markets for these grains and their processed commodities
have therefore created the need for knowing content and
also other physical characteristics such as hardness.
Numerous analyzer systems have been developed using
near infrared (NIR) spectroscopy techniques to analyze
the percentage concentrations of protein and moisture.
Some of these systems target cereal grains in milled
form as explained, for example, in U.S. Patent No.
5,258,825. The value added by milling in some instances
decreases the economic gain that is obtained by first
sorting, and thus others target the analysis of whole
grains, as in U.S. Patent No. 4,260,262.
NIR spectrophotometric techniques are typically
favored because of their speed, requiring typically only
thirty to sixty seconds to provide results, as compared
with the hours of time which would be needed to separate
CA 02380392 2002-O1-15
WO 01/06232 PCT/US00/19101
-2-
and analyze constituents using wet chemical and other
laboratory methods. NIR spectrophotometric techniques
are also favored because they do not destroy the samples
analyzed. In a typical analysis of wheat grains, for
example, a sample is irradiated serially with selected
wavelengths. Next, either the sample's diffuse
transmissivity or its diffuse reflectance is measured.
Either measurement then lends itself to use in
algorithms that are employed to determine the percentage
concentration of constituents of a substance.
For example, the analyzer described in U.S. Patent
No. 4,260,262 determines the percentage of oil, water,
and protein constituents by using the following
equations:
o i 1 % - Ko + K1 ( + KZ ( oOD K3 (
oOD ) o + DOD
) W ) P
water o - K4 + KS (DOD)+ K6 (oOD) K., (oOD)
W o + D
protein % - Ke + K9 (oOD)+ Klo (DOD) Kll (DOD)
W o + P
where (oOD)W represents the change in optical density
using a pair of wavelengths sensitive to the percentage
moisture content, (oOD)o represents the change in optical
density using a pair of wavelengths sensitive to the
percentage oil content, and (oOD)P represents the change
in optical density using a pair of wavelengths sensitive
to the percentage protein consents. K~-K" are constants
or influence factors.
The change in optical density of any given
constituent may thus be found from the following
equation:
oOD = log (Ii/Ir) 1 - log (Ii/Ir) Z
CA 02380392 2002-O1-15
WO 01/06232 PCT/US00/19101
-3-
where (Ii/Ir)1 is the ratio of the intensity of incident
light to the intensity of reflected light at one
selected wavelength, and (Ii/Ir)2 is the ratio of the
intensity of incident light to the intensity of
reflected light at a second selected wavelength.
Typically, grain analyzers use selected wavelengths
in the range of about 1100 to 2500 manometers. However,
in U.S. Patent No. 5,258,825, particle size effects of
flour were overcome by additionally using a 540
manometer wavelength.
Analyzers of the prior art typically use a filter
wheel or scanning diffraction grating to serially
generate the specific wavelengths that are of interest
in analyzing grain constituents. Because of moving
parts, filter wheels and scanning diffraction gratings
are sensitive to vibration and are not reliable in
analyzing grain during harvesting. They therefore are
not suitable for withstanding the mechanical vibrations
generated by a combine or other agricultural harvesting
equipment, and therefore have not found use in real-time
measurement of grain constituents during harvesting.
Optical systems typically include fiber optic
cables to conveniently direct light from a source to a
destination located at a distance. Unfortunately, fiber
optic cables can not be used in certain applications,
such as a combine or harvester, without further
conditioning of the optical signal because the
mechanical vibrations of the machinery can cause
undesirable modal disturbances within the optical fiber.
These modal disturbances or higher order reflections
create light intensity disturbances that are not related
to the properties of a sample. Therefore, without
incorporating costly conditioning mechanisms, the
CA 02380392 2002-O1-15
WO 01/06232 PCT/US00/19101
-4-
quality of an optical signal can be degraded. This
detracts from the accuracy of the measuring device such
as a grain analyzer, especially when it is expected to
be used in the field.
SUMMARY OF THE INVENTION
This invention is a spectral analysis system and
method for determining percentage concentration of
constituents in a flowing stream of agricultural product
and related substances as they are fed through a combine
harvester, grain processor, or storage equipment. Such
agricultural products may include, but are not limited
to, for example, cereal grains such as wheat, corn, rye,
oats, barley, rice, soybeans, amaranth, triticale, and
other grains, grasses and forage materials.
The invention uses the diffuse reflectance
properties of light to obtain percentage concentrations
of constituents of the flowing stream of an agricultural
substance. In the preferred embodiment, techniques of
the invention involve measuring a spectral response to
short wavelength, near infrared (NIR) radiant energy in
the range from 600 to about 1100 nanometers (nm) as well
as light in the visible spectrum, including wavelengths
as low as about 570 nanometers (nm). Other wavelengths,
up to mid infrared near 5000 nanometers, are optionally
used in the present invention, where the range of
wavelengths analyzed depends on the limited bandwidth of
the detector. The spectral response at shorter
wavelengths helps in the modeling of proteins and other
constituents in conjunction with the response at longer
wavelengths.
The analyzer includes an optical lamp having a
suitably broad bandwidth for simultaneously irradiating
CA 02380392 2002-O1-15
WO 01/06232 PCT/US00/19101
-5-
the flowing agricultural product stream with multiple
wavelengths of light. A detector receives the radiation
diffusely reflected from an agricultural sample so that
the received optical signal can be analyzed by a real-
time computation subsystem to determine constituents in
the sample.
The lamp is angularly positioned in a first chamber
to irradiate the flowing product sample through a window
formed of a suitable protective material such as
sapphire. The light source is focused using a lens or
parabolic mirror to intensify the light irradiating the
agricultural product sample. This enhances reception of
reflected light off the sample into the detector, which
is angularly positioned in a second chamber. The design
of each chamber ensures that stray light from the lamp
is not received by the detector from within the
detection apparatus itself during a sample measurement.
Rather, light received by the detector is essentially
only light reflecting off the product sample back into
the second chamber.
The second chamber includes a diffuser in line with
the light received from the irradiated sample. The
diffused optical signal emanating from the diffuser is
then fed into a wavelength separator, such as a linear
variable filter (LVF) within the second chamber to
spatially separate the wavelengths of interest.
The wavelength separator in turn feeds the optical
signal into a suitable detection device, such as a
charge coupled device (CCD), which is capable of
simultaneously detecting the spatially separated
wavelengths reflected from the irradiated sample.
Electrical signals from the detection device
corresponding to individual wavelengths of light from
CA 02380392 2002-O1-15
WO 01/06232 PCT/US00/19101
-6-
the irradiated sample are converted into digital data
where they are spectrally analyzed by a computation
device to calculate the percentage concentration of
various constituents in the sample.
The invention includes a reflective device in the
first chamber to direct a portion of the optical lamp
light, which serves as an optical reference, into the
detector located in the second chamber. A controllable
shutter mechanism is used to block this reference light
when a sample is spectrally analyzed. Conversely,
another shutter mechanism blocks light reflected from
the sample when the reference light is spectrally
analyzed. Based on a combination of reference and
sample measurements, a precise wavelength analysis is
used to determine constituents in a sample agricultural
product. According to the principles of the present
invention, an entire harvest can be analyzed and
recorded to provide a correlation between a harvested
product and a particular geographic region.
The analyzer advantageously monitors a flowing
sample without requiring an expensive and restrictive
fiber optic cable or mode mixing apparatus. Modal
disturbances caused by mechanical vibrations in the
optical fibers are therefore avoided. Furthermore, the
aperture angle of monitored light from the irradiated
sample can be much larger because there is no need to
incorporate an optical pickup to guide the sample light
into a narrow fiber optic cable. The wide angle optical
return signal results in greater received light
intensity strength which leads to more accurate
constituent measurements.
BRIEF DESCRIPTION OF THE DRAWINGS
CA 02380392 2002-O1-15
WO 01/06232 PCT/C1S00/19101
The foregoing and other objects, features and
advantages of the invention will be apparent from the
following more particular description of preferred
embodiments of the invention, as illustrated in the
accompanying drawings in which like reference characters
refer to the same parts throughout the different views.
The drawings are not necessarily to scale, emphasis
instead being placed upon illustrating the principles of
the invention.
Fig. 1 is a high level schematic illustration of a
short wave near infrared grain quality analysis system
according to the invention.
Fig. 2 depicts a process for measuring aborptivity
of a sample according to the principles of the present
invention.
DETAILED DESCRIPTION OF THE INVENTION
Referring now more particularly to Fig. 1, the
present invention is a system 100 for analyzing the
constituent components of a flowing stream of an
agricultural product as it is being processed or
harvested. The agricultural products which may be
analyzed by the system 100 include, but are not limited
to, cereal grains such as wheat, corn, rye, oats,
barley, rice, soybeans, amaranth, triticale, and other
grains, grasses and forage materials. The constituent
components being analyzed may include, but are not
limited to, protein, oil, starch, fiber, moisture,
carbohydrates and other constituents and physical
characteristics such as hardness. Although the
following discussion describes a particular example
wherein the product being analyzed is a cereal grain, it
CA 02380392 2002-O1-15
WO 01/06232 PCT/US00/19101
_g_
should be understood that other agricultural products
may be analyzed as well.
The system 100 uses a suitable continuous
irradiating device such as an infrared light source 10.
Radiation from the light source 10 shines forward
through a first window 12 to a sample of a flowing
agricultural product 14 being harvested, processed, or
otherwise flowing through a conveyance such as a duct
16.
The light source 10 continuously and simultaneously
produces infrared light of multiple wavelengths in a
region of interest such as between 570 to about 1120
nanometers (nm). Other applications of the present
invention include wavelength analysis in a range between
visible and mid-infrared corresponding to 400 to 5000
manometers and, therefore, require a source light
capable of emitting such wavelengths. The desired range
of analysis depends on the characteristics of the
detector, which typically is bandwidth limited. For
example, a fairly inexpensive silicon photodiode array
is capable of detecting light intensities of wavelengths
between 600 and 1100 manometers. Other detectors
optionally used in the invention are lead sulfide and
lead selenide detectors, which support a response
between 1000 to 3000 manometers and 3000 to 5000
manometers respectively. The detector will be discussed
in more detail later in the application.
The preferred light source 10 is a quartz halogen
or tungsten filament bulb and is widely available. A
typical light source 10 is a tungsten filament bulb
operating at 5 volts (vDC) and drawing one amp of
current. The light source 10 may be further stabilized
by filtering or by using an integral light sensitive
CA 02380392 2002-O1-15
WO 01/06232 PCT/US00/19101
_g_
feedback device in a manner which is known in the art
(not shown).
The light source 10 is positioned to shine upon the
flowing product 14 as it is flowing through a conveyance
such as a duct 16 within an agricultural combine or
other grain processing apparatus. The flow of the
agricultural product 14 through the duct 16 is generally
in the direction of the illustrated arrows, but
optionally is reversed.
The light source 10 and related components
positioned adjacent the duct 16 are positioned within a
suitable housing 11. In such an instance, a first
window 12 is preferably disposed between the light
source 10 and the flowing agricultural product 14. This
prevents the flowing agricultural product from clogging
the system 100. The first window 12 is formed of a
suitable material, such as sapphire, which is
transparent at the wavelengths of interest, and which
does not see a significant absorption shift due to
temperature changes. Sapphire also resists scratching
and, therefore, debris such as stones flowing in concert
with the agricultural product will not damage the
window. The first window 12 may be integrally formed
with the housing 11 or the duct 16 as desired.
The housing 11, including the enclosed light source
10, first window 12, and other related components to be
described, is thus positioned to monitor a continuous
flow of the agricultural product 14 through the duct 16.
This is accomplished by mounting the housing 11 such
that the first window 12 is disposed adjacent an opening
15 in the duct 16 so that the light source 10 shines
through the window 12 and opening 15 onto the flowing
product 14.
CA 02380392 2002-O1-15
WO 01/06232 PCT/US00/19101
-10-
The housing 11 may be a separate physical housing
or it may be integrally formed within the duct 16.
A parabolic mirror or reflector 17 is disposed
within the lamp cavity to direct light from the light
source 10 to the sample 14 being analyzed. In the
preferred embodiment, the light is collimated to enhance
the intensity of the spectrally analyzed light that is
reflected off the sample. However, lens 20 optionally
provides a means of additionally focusing or de-focusing
the light into a more or less intense beam. In other
words, the irradiated light shining on the sample is
optionally angular rather than collimated.
In an alternate, embodiment, more than one light
source 10 can be used, such as an array of infrared
emitters, as long as they are focused on the same point.
It is preferred that the light source 10 be placed
such that it directly illuminates the flowing product 14
through the first window 12 with no fiber optic or other
device other than the first window 12 itself being
disposed between the light source 10 and the flowing
product 14. In the preferred embodiment, the
illumination spot size from the light source 10 onto the
sample of flowing agricultural product 14 is
approximately a half to one inch in diameter. In
particular, incident light 48 from the light source 10
is focused onto a flowing sample 14 practically touching
or near the surface of the first and second window 12,
13. Effectively, the incident light 48 shines through
the first window 12 and flowing sample 14 to produce
sample light 49 directed towards a second window 13 and
analysis chamber where light intensities are analyzed.
A wider illumination spot size, such as up to a few
inches in diameter, is optionally used in applications
CA 02380392 2002-O1-15
WO 01/06232 PCT/US00/19101
-11-
where the sample is located farther away from the
surface of the first and second window 12,13.
A wide illumination spot size and viewing aperture
is preferred because it results in more accurate sample
measurements of the flowing agricultural product. Each
measurement based on the wide spot size typically
involves analyzing multiple pieces of illuminated grain,
resulting in a beneficial averaging effect.
Other spectral analyzers incorporate costly optical
hardware for receiving the light reflected off a sample
49 and directing it to an optical detector located far
away. To view even a small spot with these systems
requires a high intensity light source. This method of
using optical hardware to redirect the reflected sample
light 49 limits the spot size to a narrow diameter
because the reflected light must be focused into a
narrow fiber optic cable.
The present invention, on the other hand,
advantageously positions a detector 52 with a wide
viewing aperture located in a second chamber 65
immediately adjacent the first chamber 68 to receive the
reflected sample light 49. This eliminates the need for
costly fiber optic hardware because received light no
longer needs to be directed to a detector at a remote
location. Rather, reflected sample light 49 naturally
strikes the detector 52 located immediately in the
second chamber. To match the performance of the present
invention, a fiber system would require a very large
fiber bundle for redirecting reflected sample light to a
remote detector.
Eliminating the fiber optic pickup and associated
fiber optic cables has advantages in addition to
enabling the use of a wider illumination spot size.
CA 02380392 2002-O1-15
WO 01/06232 PCT/US00/19101
-12-
Typically, fiber optic cables have a limited bandwidth.
Hence, when they are used to steer reflected light to a
detector located far away, the spectral range of
directed light is limited to the bandwidth of the cable.
This negatively impacts constituent measurements because
a narrow spectral bandwidth can produce inaccurate
results. Moreover, the use of fiber optic cables are
further prohibitive because the fiber optic cables
supporting the a bandwidth of mid infrared are
particularly expensive. In some cases, just a few
meters of this type of cable can be more than a thousand
dollars. The present invention is not bandwidth limited
nor burdened with unnecessary additional cost because it
does not incorporate any fiber optic cables to redirect
light. As a result, any broadband light source and
complementary detector can be used to perform
constituent sampling.
The use of a fiber optic cable to redirect the
reflected sample light 49 is additionally undesirable
because the integrity of the optical signal within a
fiber optic cable is susceptible to heat distortion and
mechanical vibrations. This is especially true when the
fiber optic cable supports the transmission of light in
the infrared region. Both the heat distortion and
mechanical vibrations, particularly prevalent in a
combine or harvester, negatively impact the integrity of
the optical signal used to detect constituents in a
sample. By placing the detector 52 in a second chamber
65 immediately adjacent the light source without
incorporating an optical fiber in the reflected sample
light path 49, the present invention advantageously
avoids the aforementioned problems.
-02 FAn 6I' E3? 7000 FOLE~~ HOAX E:.IOT :.T.P
14-0~-2001 _
International Application No. PCTIUS00/19101
-13-
size can be adjusted to create a variable field of view and allows a large
sample to
be imaged from a distance.
Moreover, the design of each chamber, i.e., the first chamber 68 and the
second
chamber 65, ensures that stray light from the light source 10 is aot received
by the detector
50 from within the housing 11 of the system 100 during a sample measurement.
Rather, the
light received by the detector 50 is essentially only reflected sample light
49. In particular,
a wail 57 separating the fast chamber 68 and the second chamber 65 of the of
housing I 1
extends to the first window 12 and the second window 13 to form an optical
blocking
element that forces the incident light 48 into the sample 14 and intu'bits the
incident light 48
from directly reaching the detector 50.
Light emitted by the light source 10 passes through the firsflwindow 12 and
opening
in the duct 16 onto the flowing agricultural product 14. Incident source light
48 then
15 reflects off the sample 14, where the reflected sample light 49 is
angularly directed back
through opening 15 in the duct and second window I3.
In the preferred embodiment, the angle of the light source and detector unit
52 in ~e
second chamber 65 are optimized so that most of the reflected sample light 49
is directed to
the second chamber 65 for spechal analysis of the flowing agricultural
product. For
example, the light source may be optimally angled at approximately 600
relative to the first
window 12 while the detector unit 52 in the-second chamber 65 may be angled at
approximately 60° relative to the second window as shown in
illustrative Fig. 1.
The-secbnd chamber 65, as mentiened; includes optical devices for detecting
the
reflected sample light,49. Specifically, the reflected sample light 49 passes
through the
second window 13 into the second chamber 65 where it is spectrally analyzed.
Diil'user 59
acts to scatter the reflected sample light 49, spatially distnbuting the
intensity of the light
throughout the second. chamber 65 for more accurate simultaneous spectral
readings. Far
example, reflected sample light 49 of various wavelengths is more evenly
distributed
throughout the second chamber 65. Otherwise, high intensity light regions
caused by
reflected sample light 49 from the flowing product 14 result in less accurate
constituent
measurements due to imaging effects.
Replacement Sheet
CA 02380392 2002-O1-15
CA 02380392 2002-O1-15
WO 01/06232 PCT/US00/19101
-14-
spectral readings. For example, reflected sample light
49 of various wavelengths is more evenly distributed
throughout the second chamber 65. Otherwise, high
intensity light regions caused by reflected sample light
49 from the flowing product 14 result in less accurate
constituent measurements due to imaging effects.
Hermetically sealed chamber 46 is positioned in the
second chamber 65 to receive reflected sample light 49.
An optically transmissive third window 60 allows
diffused light emanating from the diffuser to shine onto
wavelength separator 50 and CCD array detector 52, both
of which are positioned within the hermetically sealed
chamber 46. This airtight chamber protects sensitive
optical components from corrosive and measurement-
inhibiting elements such as humidity and dust. Without
the hermetically sealed chamber 46, a buildup of dust
and other debris on the detection unit 52 and wavelength
separator 50 will negatively effect constituent
measurements. It should be noted that all or part of
the second chamber 65 is optionally designed to be
hermetically sealed.
The wavelength separator 50 within hermetically
sealed chamber 46 in a preferred embodiment provides
spatial separation of the various wavelengths of
diffusely reflected light energy of interest. Suitable
wavelength separators 50 include linearly variable
filters (LVF), gratings, prisms, interferometers or
similar devices. The wavelength separator 50 is
preferably implemented as a linearly variable filter
(LVF) having a resolution (~?~/?s) of approximately one to
four percent.
The now spatially separated wavelengths in turn are
fed to the detector 52. The detector 52 is positioned
CA 02380392 2002-O1-15
WO 01/06232 PCT/US00/19101
-15-
such that it simultaneously measures the response at a
broad range of wavelengths. In the preferred
embodiment, the detector 52 is an array of charge
coupled devices (CODs), which individually measure the
light intensity at each of the respective wavelengths.
In other words, each cell of the CCD array is tuned to
measure the intensity of an individual bandpass of
light.
Other suitable detectors 52, however, are
constructed from fast scan photodiodes, charge injection
devices (CIDs), or any other arrays of detectors
suitable for the task of simultaneously detecting, in
parallel, the wavelengths of interest.
In a preferred embodiment, the detector 52 is a
silicon CCD array product, such as a Fairchild CCD 133A
available from Loral-Fairchild. The device preferably
has a spatial resolution of about 13 micrometers. The
frequency resolution is the selected bandwidth of
interest (as determined by the linear variable filter
50), divided by the number of CCD elements. In the
preferred embodiment the CCD array 52 is a 1,024 element
array processing wavelengths in the range from about 570
to about 1120 nm. Other detectors, as mentioned,
supporting different bandwidths are optionally used.
In addition, the detector 52 such as a CCD array is
typically temperature sensitive so that stabilization is
usually preferred.
In the preferred embodiment, because of the
relatively close positioning of LVF 50 and CCD array 52,
both of these components can be temperature stabilized
together. The temperature stabilization can be by
suitable heat sink surfaces, a thermoelectric cooler
(Peltier cooler) or fan. The preferred embodiment
CA 02380392 2002-O1-15
WO 01/06232 PCT/US00/19101
-16-
of the invention includes a reflector 22 disposed in the
first chamber to reflect reference light rays 23 to the
wavelength separator 50 and detector 52 positioned in
the second chamber 65 depending on the position of light
blocking shutters. The reflector 22 is preferably fixed
such that repeated measurements are based upon the same
reference light intensity.
A first shutter controls the passage of reflected
sample light 49 into the second chamber 65, whereas a
second shutter controls the passage of reference light
rays 23 reflected off reference light reflector 22 into
the second chamber 65. At any given time, the first and
second shutter are controlled to allow the appropriate
light to flow into the second chamber 65. It should be
noted that a single shutter is optionally designed to
provide a way of controlling which light, reference
light rays 23 or reflected sample light 49, eventually
shines on the detector 52.
Control electronics 18 located adjacent to the
second chamber 65 provides a means of controlling the
shutter mechanisms. Shutter position commands are
received via electronic signals transmitted by
controller 35 residing in the electronics block 30.
The shutters are used to perform three
measurements. A first measurement involves blocking
both the reflected sample light 49 and reference light
rays 23. This reference measurement of the "dark"
second chamber 65 serves as a means of calibrating the
detector unit or array 52. A second measurement
involves blocking the reflected sample light 49 and
measuring the reference light rays 23. This measurement
serves to calibrate the system to the light source 10.
Finally, a third measurement involves blocking the
CA 02380392 2002-O1-15
WO 01/06232 PCT/US00/19101
-17-
reference rays 23 and measuring the reflected sample
light rays 49. Details of the measurements and related
computations are further described in Fig. 2.
An electronic signal or signals 27 between the
electronics block 30 and system housing 11 provide a way
for the controller 35 to pass signals controlling the
position of the first and second shutter, and
specifically the flow of reference light rays 23 and
reflected sample light 49 into the second chamber 65
where the detector unit 52 resides. For example, the
first shutter is placed in the open position to allow
light to pass to the sample and to be diffusely
reflected by the flowing product sample 14 during sample
measurement operations, and placed in a closed position
to occlude light from the sample and diffusely reflected
light from the shutter during reference measurements.
The second shutter is used to block the reference light
rays 23 from entering the second chamber 65 during
detector unit 50 calibration and flowing product 14
sampling.
The electronic signals 27 are bundled together in a
wire harness 28 connecting the system housing 11 and
electronics block 30. In a practical deployment of the
system 100 such as in an agricultural harvester, it is
preferred that the cable sheath 28 be sufficiently long
such that housing 11 can be placed adjacent the duct 16
while electronics block 30 may be placed in a less harsh
environment such as the protective cab of the harvester.
This distance is generally three meters, for example,
and is optionally more or less.
The electronics block 30 includes an analog to
digital converter 33, a constituent computation function
34, a controller 35, and a display interface 36. In the
CA 02380392 2002-O1-15
WO 01/06232 PCT/US00/19101
-18-
preferred embodiment, the constituent computation
function 34, controller 35 and display interface 36 are
implemented as software in a computer, microcontroller,
microprocessor and/or digital signal processor.
Electronic signals 27 in wire harness 28 provide
connectivity between the electronics in the system
housing 11 and the electronics block 30.
Typically, the electronics block is mounted in a
shielded environment, such as a cab, while the housing
11 of the optical system 100 is mounted in a position to
detect the flowing agricultural product 14. Therefore,
based on this separation, the electronics are designed
to ensure that signal integrity does not suffer because
of the length of the wire harness 28. For example, the
electronic signals 27 within wire harness 28 are
properly shielded to prevent excess coupling noise,
which may deleteriously effect A/D readings of the CCD
array detector 52. The controller 35 coordinating the
A/D sampling process, as mentioned, controls the shutter
mechanisms positioned in the second chamber 65 for the
various spectral measurements.
The individual electrical signals provided by the
CCD for each wavelength are then fed from the output of
the detector 52 to analog to digital converter 33 where
the electrical signals are converted to digital signals
for processing.
A computation block 34, preferably implemented in a
microcomputer or digital signal processor as described
above, then carries out calculations on the basis of the
received wavelength intensities to obtain percentage
concentrations of constituents of the sample 14. The
percentage of constituents, which are determined using a
chemometric model, are then shown in any desired way
CA 02380392 2002-O1-15
WO 01/06232 PCT/US00/19101
-19-
such as by a meter or presenting them to a display. The
display may be integral to a laptop computer or other
computer placed in the cab of the harvester. The
computation block may be part of the electronics block
30 or may be physically separated from it.
Techniques for calculating percentage
concentrations of grain based upon samples of light and
particular wavelengths are the multi-variate techniques
detailed in the book by Sharaf, M.A., Illman, D.L., and
Kowalski, B.R., entitled "Chemometrics" (New York: J.
Wiley & Sons, 1986).
Preferred wavelengths of interest depend upon the
constituents being measured. For example, when
measuring protein concentration, the algorithms make use
of absorptance attributable to the vibration-rotational
overtone bands of the sub-structure of protein. At
longer wavelengths absorptivity coefficients are large,
the path length is short, and thus one would not sample
the interior of the grain particles. At shorter
wavelengths the absorptivity coefficients are small and
the signal is thus weak.
The system 100 thus provides for irradiation of the
sample followed by spatial separation and detection of
multiple wavelengths in parallel, making for rapid
analysis of this sample. Moreover, because the optical
portions of the unit are stabile to vibrations, it is
substantially insensitive to vibrations such as found in
agricultural combines or other harvesting and processing
equipment. The system 100 may therefore be easily
deployed in environments where real time analysis of
harvested grain or other agricultural produce may be
carried out during harvesting and other processing
operations. The data obtained thereby may be compared
CA 02380392 2002-O1-15
WO 01/06232 PCT/US00/19101
-20-
with reference data to provide percentage concentrations
of constituents for use in mapping field layout
according to the so called global positioning system
(GPS) .
Furthermore, the use of the CCD array as detector
unit 52 provides advantages over prior art techniques
that use discrete or scanned diode arrays. In
particular, the CCD bins are all filled with charge at
the same time in parallel with one another, until one of
them is nearly full. They are then emptied and the
results read out by the controller 35 while the CCD
array begins filling again. Therefore, each pixel has
seen the same grains over the same time intervals. In
contrast, diode arrays must be read sequentially so that
for example, any given element is producing a signal
from a volume of grain if it is distinct from those seen
by previous pixels.
The signal to noise ratio of the system 100 may be
improved by averaging over the course of many
measurements.
The preferred absorptivity measurement includes the
following process (also depicted in Fig. 2):
1. Block both the sample reflection light and
reference light from the wavelength detector
unit (step 301)
2. Perform a reading on the wavelength detector
unit, storing measurement data in D for dark
spectrum (step 302).
CA 02380392 2002-O1-15
WO 01/06232 PCT/US00/19101
-21-
3. Block the sample reflection light and allow
reference light to shine on wavelength
detector unit (step 303) .
4. Perform a reading on the wavelength detector
unit, storing measurement data in R for
reference light spectrum (step 304).
5. Block the reference light and allow sample
reflection light to shine on wavelength
detector unit (step 305).
6. Perform a reading on the wavelength detector
unit, storing measurement data in S for sample
spectrum (step 306).
7. Calculate the absorptance spectrum A, where
the light absorption as derived from these
diffuse reflectance measurements is given by:
A = LOGlo (R-D/S-D) .
In addition, since the absorptivity variations from
the presence of protein are quite small, multiple
realizations, averaging, and second derivative analysis
are typically used to produce the desired absorptivity
number at a particular wavelength. Further data
processing therefore may provide a second derivative of
this function to remove constant and linear offsets so
that only quadratic and higher order features in the
absorptivity spectrum are utilized in the determination
of protein content.
CA 02380392 2002-O1-15
WO 01/06232 PCT/LTS00/19101
-22-
EQUIVALENTS
While this invention has been particularly shown
and described with references to preferred embodiments
thereof, it will be understood by those skilled in the
art that various changes in form and details may be made
therein without departing from the spirit and scope of
the invention as defined by the appended claims. Those
skilled in the art will recognize or be able to
ascertain using no more than routine experimentation,
many equivalents to the specific embodiments of the
invention described specifically herein. Such
equivalents are intended to be encompassed in the scope
of the claims.
