CA2402669C - Apparatus and method for measuring and correlating characteristics of fruit with visible/near infra-red spectrum - Google Patents

Abstract

This disclosure is of 1) the utilization of the spectrum from 250 nm to 1150 nm for measurement or prediction of one or more parameters, e.g., brix, firmness, acidity, density, pH, color and external and internal defects and disorders including, for example, surface and subsurface bruises, scarring, sun scald, punctures, in N-H, C-H and O-H samples including fruit; 2) an apparatus and method of detecting emitted light from samples exposed to the above spectrum in at least one spectrum range and, in the preferred embodiment, in at least two spectrum ranges of 250 to 499 nm and 500 nm to 1150 nm; 3) the use of the chlorophyl band, peaking at 680 nm, in combination with the spectrum from 700 nm and above to predict one or more of the above parameters; 4) the use of the visible pigment region, including xanthophyll, from approximately 250 nm to 499 nm and anthocyanin from approximately 500 to 550 nm, in combination with the chlorophyl band and the spectrum from 700 nm and above to predict the all of the above parameters.

Description

R-lb Oc 16: 08 FRC1M: LIEBLER II~EY COR '~97353SBS Tt7:7~33057T24 PHfaE:06 .
~Cl",~lS 01 / 0 814 6 IPEA/US ~. ~ MAR 2002 J..
~X' 7 ',~. fi ~~' .mm 3 An Apparatus and Moth'"'od And Ter.h.niqytr;s for M~;asurin~ and Correlating 4 Characteristics ofFt'uit With Visible/Near infra-Red Spectrum 6 The present disclosure ielates generally to the use of the combined visible and ,.
7 near irrfi-a red spectrum i>9 aat 8pp8~atus and methc~tl for measuring physical ......, ~ parameters, e.,g., ~t~rrrness, density and internal and external disorders, and chemical . .
~) parameters, e.g., molecules containing O<'C~i:,1'~1~~1H: and C~kl chemical bonds, in fruit I U ~u~d cowelating the resulting m sswrements with fruik quality at~d maturity 11 characteristics, including ~3rix,'' Acidity, density, pH, ftrn~rness, color and interrual and t ? external defects to forecaat ~ConsumeT preferences including taste preferences and -_ Y~x I 3 appearance, as well as hawest; star>qge and shipping variables. With the present 14 apparatus a.nd method, the interior of a san~pla, e.8., fruit including apples, is illuminated and the spectrram 4fa;~swrtaed and scattered light from the sample is i 6 detected and measured. prediction, calibraeian and classification ai~arithms are 17 determined Car the category afsaritpla perrlaittiltf; cc~rrrwlratiott heiween the spectrum of I 8 absorbed and scattered ligfit arid altmgle characteristics, e.g., fruit quality and "°~~ 1~~ maturity characteristics.
?I) sac ~o~,n of t ~ .LnventiioD

2 i The embodiments disclosed herein has a focus an combined visible and near-23 inl~r:~re~J 4.Nllt) spectroscopy and its modes of use, major issues in the application of a(:
23 NlR to the measurement of O~I~,1V'~H and C'-xT containing molecules that arc, 24 indicators of sample quality ineiudin~ fruit quality and iu particular fret fruit quality.
hear-Infrared Spectroscopy Bs~tck~r~xund: Near-infrared spectroscopy has ?c, been used since the 197t)'s for the conyositi.onal cu~alysis of low moisture food 27 produms. However, only in the feel 1U-15 years has N~llt becn successfully applied to c~.~' tl» 31131yS1s Of hl~l'L ntoisritre products su~l1 as i:'rut. Nt~~ ib a form of vibrational 29 spectroscopy that is particularly sensitive to the presence of molecules containing C
., -AMENDED

R-16 D2 16:09 FROM:LIEBLER IVEY CONNDR 597353585 70:7033057724 Pp~E:O?' - ~~ ~IUS of / 0814 6 . .. Ip~UtIS 1$ MAR Z0~2"
I H (carbon-hydrogen), 4-H (oxyg'ert-hydrogen), and N-H (nitrogen-hydrogen) groups.
Therefore, constituents such sugars arid starch (C-i-I), moisture, alcohols and acids 3 (O-H), and protein (N-lI) cariye quantified in Ltqutds, solids and sliuz~tes. in addition, 4 the analysis of gases (e.g., wa vapor, ammonia) is possible. N1R is not a trace analysis teciuiique and it is geneeaily used for tueasurit~ components that ara present _#
at concentrations greater than ~0.1 °/p.
7 Short-Wavelength NXR vs. Long-Wavelength N1R: N'1R has tradition$Ily ..
8 bean curried out in the 1100-2500 nm region of tl» electromagnetic spectrum.
",.
9 However. the wavelw~;th region of 700-1100 rtrn (short wavelength-NIR nr S W-_....- ~.'_..
NIR) has keen gaining increased attention. rl he S W-NtR region offers numet~us I l advantages for on-line and in~sitat balk constituent analysis. This portion of the N!R
I 2 is :~ccessihle to law-cost, high~psc~m,tance silicon. detectors and fiber optics. Ltl 13 addition, high intensity l~acr diodes and low-cost light emitting diodes are becoming 14 increasingly avai table at a variety of NTR wavelength outputs.
K
l5 -fhe rei.atively low extineti4~it (light absorption) coefficients in the SW-NIR
'f t5 region yields linear absorbanee with analyte ccmcentration and permits long, _.;.
17 convenient pathlvngths to be used. The depth o f penetraxion of S W-NIR is also much ..,~;
18 ~,Tealer than that oftho t~nger vsrav~alength N>-R, permitting a more adequate sampling ~x l~~ of the "bulk" material. This is~ofparticular importance wbtn the stlml.,le to be analyzed is heterogeneous such as fruit.
? 1 Diffuse I~eflect~tt~Ce Saotpllng vs. Tranamissiou Samptiog: Traditional 22 NLR analysis has used diffuuse'i~tlectance sampling. ~Chis mode of sampling is 23 convenient for sarrtples that are highly light scattering or samples for which there is ?4 no physical ahility to cmploytransMaission spectrc,scopy, Diffusely reflected light is light that has entered. a sample, undergone multiple scattering events, :md emerged ?C~ tcom the surface in random directions, A portion of light that enters the sample is :L.
?7 Uso absc,rbed. The depth of penetration of the lighr 1S highly dependent on the ?8 sarr~pl~ characteristics and is often affected by the size of particles in the sample and 39 the sample density. Furthermore, diffuse reflectance is biased to the surface of a .n R-16 02 16:09 FROM:LIEBLER IVEY COhINOR ~~~'~~35~15 TO:'fi033057724 PR6E:08 _ ~-,~ p 1 ~ p 814 b tPF.~ ~;~ 16 MAJ~ Z~(~
1 sanil.,le and may not provide mpret~ntatxve data for large heterogeneous samples such 2 as apples.
3 While transmission ssartplin,g is tyl,ically used Cor tha analysis of clear ~ solunons, it also can be used f~bir interrogating solid samples. A
transtvissian measurement is usually perfo~ ad with the detector directly oPposit~ the Iight source G (i.e., at 1 ~O de~~rees) arid vv~th~tha sample in tl~a center, Alternately the detector can 7 he placed closer to the light soiurcc (at angles Less than ! 8U de,grees), which is often 8 necessary to pmvide a more e8aily detected level of light. .Because of the long sample ~S~_ pathlengths and highly light iaag tlature of mast tree tTUrt, transmission I U mvasuren'tents can only be p~armtd in the SW-N'.V:R wavelength raglan, unless ;...~.
1 1 special procedures art employed to improve signal to noisa.
.,t 12 IvIIR Calibratlv>A: ~R tlysis is lar~;aly a.rt emyrical method; the spectra) I3 titres are difficult to assign, anti the sptctrascopy is tiequently carried out an highly 14 light scattering sz~rnples Where sd~ierettce to i~eer's Law is not expecaed.
I 5 Accordingly, statistical calibration t~cbniqutx ara often used to determine if there is a t G relationship between andlyte cortcat~tratiar~ (or sample property) arrd instrument 17 response. T'a uncover this relationship rcduires a rel,resentativt sat of "traitring" or 18 calibration samples. Thcat s~imptts must span the complete range of chemical and ~,: r '' 1 ~ physical properties of all future samples to be soon by ttte instrument.
~~r 20 Calibration begins by scduiring a spectrum of each of tha samples-2 I Constituent values far alf of tEte sutaiytas of interest are then obtained using the beat ,~.,.
?~ reference method availablt with retards to t~curaey and preeisian, It is impanant to ?3 note. that a quantitative sp~t~a1 method develolaed using stntisticat correlation z~
?4 techniques can perfc~rni no bt~ar tl~~n the reference method, 25 After the data has bean acquired, cotxtputer models emlaloying Statistical 26 calibration techrticlues ~tre dev~laped that relates thH l~~Dt sperara, to the maasurad 37 constituent values or properties. 'these calibratic,n mortals cap, be cxpar;dad and must ?8 b~ periodically updated and veriFed using co77vea7tiao~~al testing procedures.
2~) horrors af~ectiog calibt'akion include fr~wit type and variety, seasonal and F~

R-16 X72 16: 10 FROM: LTEBLER TUEY Ct7NM7R 5097353585 T0: '~03305~~24 PACE: ~9 PC'~'/US 01 / 0 814 t t~~ z s ~A~ X06 1 gto~;raphica.l. differe,~ces, and vvhethcr the fru.i.t is fresh or has been in cold or other ~>
2 storage. Calibration varisblesiinelude the particular properties or analyzes to be 3 measured and the concentration or level of the properties, intcrcorreldtions (co-4 linearity) should be minimi2~~c,1'irt calibration saalples so as not to lead to raise :i.~
interpretation of a models predictive ability, Co-linearity occurs when the 6 concentrations of two component~t acre correlated, e.g., an inverse correlation exists ~~:>,r<.
7 when one component is high; the other is always law or vice versa.
J.k I~y .
8 Application of N1R to Tree Fruit and ~.xistiog Un-t.ine NTTt ~4 9 instrumentation: A growing body of research exists for NTR analysis of true fruit.
_.~ ";, NiR has been used far lht measurement of fi~it: juice, flesh, and whole fruit.
In juice, 11 the individual sugars (sucrose, fructose, glucose) and total acridity can be quantified ._,:.
I? with high correlation (:0.95) and acceptable error. Individual sugars can not be J 3 readily measured in whole fruit, prix is tile most succ;essf~ully rncasured N1R
l4 paraweter in whole fruit ~rtd can gener$lly be achieved with an error of+tJ.S-1.0 Brix.
l 5 More tentative recont rese~reh~'r~sults indicate Frmness and acidity measurement in ~:, 16 whop fruit also rnay be possible.
I 7 Only in Japan has tttc~lar~-scale deployment of on-line NIR for fruit sorting ~,~.
):3 occurred. These in,st~rttm~ents~~tmanual placetnentlorientation of the fruit prior 'y' 1 q to measurement and early versions were limited to a measurement rate of three ?0 samples per second. The Japanese NTR. instruments are also limited to a single lam;
2 I of fntit and appear to be ~iff'tcult try adapt to mufti-lane sorting equipment used in the 22 United SCcitES of America. White carlir;r rapanese NTIt instnunents etnployed 23 ret7ectanee sampling, o~ore recent instruments use transmission sampling.
?4 In Koashi et al., U.S. Pat. loo. 4,$$3,J53, d~ere is described a method and ?S apparatus for measuring sugAr e0ncentrtttions in liquids. Measurements are made at 2G two difi~ercnt depths using wdak and strong intra~red radiation. T'he level of sugar at ,x 27 depths hctween these two depths c~i t~>1en be measured. "Che method amd apparatus 38 utilizes wavelength. bands of950-1,154 nm, 1,150-1,300 mn, and 1,3(10-1,45() nm.
29 U.S. Pat. No. 5,0>19,701, to Dull et al,, uses near infrared (NIR) radiation in h:
1~AA~tt?~n cu~~

...-R,-1C ~2 i6: i0 FPOM:LIEBLER IVE'f CO~R 5~'3?3S35B5 T0:'~03305i'Z2~1 PRC~:10 ''~ ~Tius 0 ~ / 0 81 ~ 6~
~PEAJUS ~. ~ MAR X062 i the wavelength range of g00-1;050 stun to demonstrate m~;asuretylent of soluble solids 2 in Honeydew rnelans. An eight-certtimeler or greater distance between the light 3 delivery location to the frwtit~~ the light collection location wax found to be 4 necessary to accurately predi$olublt; solids becKtuse c~f the thick rind.
Jwamoto et al., U,S. Pict. No. 5,324.,945, also use thl~t radiation to predict r r~ su~:~r Content Ot mandarin urges. Jvvam~~ta uailiz,es d tta~asrrtission measurement 7 arraugemettt whereby the ligHt ttavrrses through the entire sample of fruit and is 8 detected at 1.s0 degrees relative to the light input eagle. Moderately thick-skinmed 9 fruit (mandarin oraaycs) tjverC'ustd To demonstrate the method, which reties on a fruit 1 () dit~meter cowcction by normali2iing dividing) the spectra at $44 tun, where, 1 1 according to the disclosed da~, GQrrelation with the sugar content is lowest. MR
l2 wrvelengths in the range of914-919 nm were found to have the highest correlation l 3 with sugar content. Second, tt~t'trd urtd fourth r~°aveltrngths that were added to the 14 nwltiple regression analysis ~quttl~n used to correlate the NiR sp~eeta-a with sugar content were 769-770 nm, 745. teat, end 7$5-78G nm.
I6 fn U.S. Pat. No. 5,708,271, lto et al. demonstrates a sugar content measuring ~.
17 apparatus that utilizes throe different NTR wavslcngths in th4 range from $bCt-9Ci0 nm.
. ~P:, I 8 The angle bctvaeen light deliivaryand collection was varied between 0 and 19 degrees and it was concluded that the low N'Ll~ rttdiatic~n levels that mut3t be detected when a phototdetector is pl~tc~ed apt i $0 degrees reLstive to the radiation source are net 21 desirable hecause of the more complicated procedures attd ecluipmeut that are w.
1? required. A cowclatian oFN'iR ~bsurbaxtce with sugar contcttt of rtwslnttelans and 23 w;ttem,elons was found when an intct~nediate angle, which gave greater NIIt .;, 24 radiation intensity, was detected. No size correction was necessary with this ZS approach, 36 U.S. Pat. No, 4,853,953 to lCoashi et al. uses comparatively long wavelengths 37 of NIR radiation (i.e., :~q50 nm), while in U,S. fat. Nos. 5,089,701 to Dull, and 28 5.708,271 to Jto, wavcletlgths of 1W1R radiation used are gres~.ter then 800 m and $(i0 2t) nrn, respective.Ly. In 11.5. Pat. No. S,a24.'~~S to I'warnoto, the wavelengths of MR
,»
., wk .~
e~uwn c~~r R-16 02 16:11 FR0~1:LIEBLER IVEY CONNOR 50973535A5 70:7833057724 PRCE:11 ' v:
~c~~~s o s i o $1 ~ 6 1~F~'~ ~: ~. 6 MAID ZDO
t radiation with the highest cotTelatiern to sugar content of mandarins were 914 nxn or 2 919 nnt. When the fruit were m~'eRsttred on th.e eduatorial or stem portion, respectively.
3 All of thesE methods use near:tnfrared wavelengt~~s of light to c.ozrelat.e with sugt~r 4 content of whole fruit. I~o or 9ua,litylWrameters are measured by these technicluES. :r~~:
G The faun disclosed patents ~.t~e simil~.r to the apparatus and method described L..
7 here in that tile Present diselosut'e tiso measures sugar content. T'wo of the patents 8 (Pat. No. 5,089,701 and 5,324,945) NfR wavelengths less than 850 nm) Pat.
No.
Y.:. , 9 5,t)sc),7f) 1 discloses the operation of the invention within the range of "from about l () 81:10 nanometers to about 1050~,nar'tormeters." U.S. Pat, Nu, 5,324,945 lists 914 nttt or 1 I 919 nrtt as the primary analytical wavelength cowelated with whole fruit sugar .f-12 content; multiple lint;ar region was used to add successivt~ wavelengths to the 13 model as follows: 769-770 tvii (2t~ wavelengtkt added), 745 nm (3rd wavelength ,L~
14 added), and 7g5-786 nm (4thxw~vel>rttgth added). 7n Pat. No. 5,089,701, addition of 1 S the fourth wavalen~th to the mddesl only reduced the standard error of pr~diction 16 (SAP) by p.1-0.2 Brix, whiah~,i~ 8t#praFtching ar less than the error limits ofthe I 7 retcactometer used. to determine the reference ~"t,rtte") Brix values.
4~
18 Other similarities between the method and apparatus described het;ein with the l a four patents listed above itticlude t~ use ofmultivariate statistical catalysis to ?0 establish correlation of the nee -infrared spectral data with sugar content of whole a 2 I fruit. Most also use data ptocessin~ technidues such as second derivative h,.
22 transformation and some type of spectral nurmali2atian. All of these methods for 23 relating Nl-R spectra to chemical or physical p~ropetties are well known to those 24 practiced in the art of NlJlt spectroscopy.
...
?5 The Foregoing patents~atld punted publications are provided herewith in an ?G b~fortnation Disclosure Statement in accordance with 37 CFR I .97.
27 Se~m~r~t~ary of the .Invention 28 Research groups around the world continue to explore tht~ applic~ttioes of near 29 infrared spectroscopy to tree Emit. The apparatus and process ~Lisclc~sed herein is of .d :10 AMEtf .~y:
R-16 02 16:12 FROM: ~IEBLER IUEY 519?~5~~8~ Tt7; ?03305?'?2~ PR~aE; 12 -PC~',~5 O 1 l 0 814 b ' ' .. IP~AII;~, 1 ~ N11AR' ~O~z I thr nondestntetive determin~ti rt cat' prediction oi' U-'ET, N-H and C-H
cantdinirtg . 4..
2 molecules that are indicators ofsa<nple c]ualities, including fruit such as apples, 3 cherries, oranges, grapes, po ~ta>o~, cereals, and other such sarnples, using near-4 infrated spectroscopy. f~rior at'T has utilized spectrurri ,ti-um 7~Sttrn and above. This ja.
disclosure is of 1) the utilizati!o1t of the: spcctrurn from 2.(~ nrn to 1150 nrn for 6 measurement or predietioa of~ ne rar mare paralzlCto'cs, e.g., Hri.x., lirntness, acidity, 7 density, pH, color artd external ~d itaturnal defects and disorders including, for 8 example, surfaoe and subsutfaee bruises, scar~rirr~;, sun scald, punctures, watercore, Y~i 9 internal browning, in samples ~lu~ling fruit; '?) an apparatus and method of 1l> illuminating the interiorofap~lc and d~aectinb ~nirtod light frmn samples i I exposed to the above spectrum in at least one spectrum range and, in the preferred ,a 12 en,budiment, in at least trove spectrum ranges o(' 25U to 499nn~ and SUl~nm to 115lhun;
I 3 3) the use of the chlorophyl aliao~tiion band, peaking at 6$Unm, in combin~lrion with ~r~
14 the specarum from 700ntn and above to pr°edict one or u'aore of the above parameters;
J 5 4) the use of the visible p~grnent re~ian, in~ludi~~~ xanthophyll, trom approximately 1 b 25()nm to 499nm and anthocyartin Carom approximately S00 tc.~ 5SOnm, in combination I 7 with the chloroptty) band and:fhe spectrum firorrt '~OUnm and above to ~pr~dic~t the all 1$ of the above pararryeters.
St Prior art alas only examined spectrum from fruit far the prediction ofBrix.
?l7 This disclosure is ofthe examination ofa greater spectrum using the combined 21 visible and ryear infrared wavelength regions for the prediction of the above stated 22 characteristics. The appatatus~arld method disclosed eliminates the problem of Z3 saturation of liglJ.t spectrumd~tectors within l~~trtir-ular-spectrun3 1"eglons while 24 gaining data 4vithin other r~agions its the examination, in particular, of fruit. That is, a ?5 spectrometers with CGD (chax'ge coupled device) ~u:ray or PT.~I~
(photorliode. array) 2l, detectors will detect light within the ~SU to 1150rzm region, but when detecting ?7 spec;lrum out o.C Fruit will saturate in regions, e.g., 70C1 to )'?5nm, or the signal to 28 noise (S/lv) ratio will be unsatisfactory at~d not useful fear quantitation in other 29 regions. e.g.. 250 to 6y9nm and .'eater than "~?5nm, thus precludit'ig the gaining of ~U

5~

R-16 D2 16:12 FROM:LIESLER IVEY CONhIOR 5097353585 70:7033057724 PHt~aE:l3 .aE~' . f: ~crtus o m o 814 ~
:u~' IPEA~US l G MAR'~0~2 1 additional inFomiation regarding the parameters above stated, Thus disclosed herein ,:
? is an apparatus and method pe~'rmitting 1 ) the automated measurement of multiple 3 spectra with t~. single pass or singlm t~tteasurcment activity by detecting more than one 4 spectrum range dwrittg a single pass or single measurement ~tcaiviky, 2) cvmhining the more than one spectrunt range deteatEd, 3} compming the combined spectrum witli a ,~,.
ti stored calibration algorithm to'4) preditling the parameters above stated.
".
.r 7 In each instance in the method and apparatus diselrased herein there will be a tS dual or plurn.l spectrum tucquisltion from a sschiple~, from different spectrum regions.
This is accomplished by 1) seiially acquiring data from different spectrum regions _...-Itl using different light source intensities or different detector/spectcometet exposure 1 l times using a single specltromer, 2) acquiring data in parallel with multiple 12 spectrometers using different light intensities, c:,g., by varying the voltage input to a 13 lamp, or different exposttre times to the spec:h wmeters; however, different exposure 14 times leads to s:~mpling errors pdt'tieularly where a sample is moving, e.g,, in a ..,..
., ! 5 processing line, due tv viewing different regions on a sample; and 3) with multiple F, 16 spectrometers using the same'exposure time, constant lamp intensity with c~u~) or a 17 plurality of light detector's including neutral density filtered light detectors (where 1$ f Itered light detectors giving the some effect as using a shorter exposure time). This '' I ~~ approach provides dual or plural spectre with good signal to noise ratio fur a.ll 2C~ wavelengths intensifies using a siTl:gle light source intonsiy and the sxrne exposure 21 time on al! spectrometer det~etors. This approach uses at least one 1:71tered light 2? detector using f ltered input 82 to the spectrometer 170 rather than different exposure 23 times. A filter ca.n be any m>atterial that tlbsorbs light with equal strength over the .A
?4 range of wavelengths used by khe spectrometc;r including but not limited to neutral 25 density filters, Spectralon, Teflon, opal coated glass, screen, Tlte dual intensity 26 approach using two different lamp voltages proves problematic because the high and 27 low intensity speat'rtt arc not easily combined tobether due to slope differences itt tht 2$ spectra. The dual exposure approach yields excellent edmhined spectra, which are 29 necessary For hn~nness and othFx characteristic prediction and also m~.praves Btzx -16 X12 1b: 13 FROM: LIEB<_ER IUEY COIVNOR 50~ ~?.~ 3583. 'TU: ~'ta3305?'~~4 PsaC~:14 -~~ '.
t~CT~S~ 01 / 0 814 b w IPEAI'v~ 16 ~MA~ Z002 I prediction accuracy.
2 Measurements are dieclosød" with the apparatus and process oftttie disclosure, Shy ~..
which are made simul~neousl~ in multi 1e sar~~ 1e t es, e. ., v~here sam les are P p YP g P
4 apples, measureta~ent is indepdndent of a particular a~pplE cultivar, using a single S calibration equation with errors of+ 1-2 1b. and -~. ().5-1,t:D prix. This disclosure n ti pertains to luboratury, poriab~ls antl art-line N'lTt. analyzers for the simultaneous 7 measurement of multiple quality parrmeters of wimples ir~clurling fruit, Depending ,.,:
:,.
8 on the application or particular charactLristic sought to be predicted or treasured, a _r 9 variety of calibration models rnay~be used, .from n.miversa~l to Mighty specific, e.g., the calibration can be speeiftc to a variety, different ~eragraphiea! location, stored v. fresh :.."' ! l fntit and other calibrations.
-_...
,:
12 Disclosed here is the grettt~ar role NIJZ, technology will play as a tool fear 13 grading sample qualities ineluding'Ptuit nu~afity, 'Che unique ability ofN'fit statistical 14 calibration techniques to eaetract non-chemical "properties'' provides a technique for development of. a general N~i.~~"quality index'" for tree fruit. This general "quality ";, 16 index" combines all of the iafottttation that oottl~l be extracted from the N1R spectra 17 and includes infUrmal~ion about prix, acidity, lirmatess, density, pH, color and 18 extenial and internal disorders antd d~efeets, ....
._,.., 19 T'hc near-infrared wavelattgth re~,rion below 74~ nm has net been explored by ~() prior investigations. Generai~y, the prier art clesiga~ and or apparatus utilized was is 21 such that longer wavelength regions pt'owided adequate data. The prior art for 2Z measuring stt~;ar content in liquids and ~rl~ale fruits using near-infrared spectroscopy ,.
2~ utilizes longer wawl~engths oFradiation. No prior art exists for rneastu~ing other 24 important quality parameters such as firmness, acidity, density arid pH. Nu prior art ' ?5 has cowelated consumer taste,prei~'erences with the cornlairled '.~1-(T~
detern~inatiurt of ' ,.
?C~ multiple quality parameters such as sugar, acidity, pH, ~'irtnncss, color, and internal ?7 arrd external defects and disorders.
2$ Ia will be shown in this patont that the wavelength region from 25(?-1150 am '?~ can be used to nondealrucliv~ty measure not onDy sugar content (Brix) in vtttious ~~ ~i~

-16 02 16:13 FROM:LIEBLER IVEY C01. X97353585 TD:?033057729 PRC"aE:lS
;a.~_ ' ~'tfUS _01 / 081 ~
IP~u~uv ~.6 z~
1 whole fivit, hut firntness, de city, acidity, pH, color and internal and external defects ? as well. For example, density oforartges is measured and is correlated to quality, 3 e.g., frcaze damaged fi~uit and ry fruit typically have Lower density than goad quality 4 fruit and lower water content,~i~.e., greater dry matter content). N1~R
density m~dsurca~ent cats be used ko remove poor quality fruit in a sortinglpaclci.ng line or at C. the supermarla;t, lnformatiori'abA~ct color pigments and chlorophyll, ralated to 7 maturity and qualify, tire obtained from ?50 to approximately 699 nm. Ftlont R approximately 700-1 L5U ttun,.the short wavelength NIR region, C-H, N-tC., O-H
.Tc information is obtained. Combini~ the visible and NIR region gives more analytical -- .~~r~
1 C) power to predict chemical, physiC8t1 and consumer properties, particularly for fruit..
11 All of these parameters can be determined simttllanooual~y Franc m:ombined 1? visiblel~ilR spectrum. Multiple parameters can be. combined to arrive at a ~'~uaIity 13 Index" that is a better tncasure of maturity or qualify than a single p:uaimeter.
1.4 Absorption of light bj!~whole fruit in the approximately 250-699 nm region is dominated by pi~t,ents, incl'udin~ ehlc~ruphyll (a green pi~;mrrnt) which absorbs in the ~m 1 ~ ahprox'imately 600-699 nm region. Chlorophyll is composed of a number of t 17 chlorophyll-protein complexes, C:haitges in these chlorophyll-protein complexes and 18 changes in othEr pi~,rtncnts, roost notably 3.nthocyanin (red pigment) and xanthophylls "'', 19 (yellow pigments), are related to the maturation and ripening process.
Chlorophyll 2O and pigments art itroporlant for determ.inirtg tirmn ess.
21 While the N'IR wsve)eitgths of 700-92S nm and lon~ar have been readily 4' 22 accessible to common near-infrared spectrometers, shorter wavelengths have not typically L~een e~cplored for the following reasona: 1 ) lead-salt and outer detector ?4 types, e.g., lnGaAs, were not sat~itive to shorter wavelengths; Z) light diftrnction gratings were blazed al Innge~ wavelengths yielding poor efficiency at short Zfi wavelengths: 3) light sources did not have enough energy output at shorter '_'7 WJYCII'Il~thS I:o overcuma the strong light nbsorpt.ion and scattering of biological i,plant and animal) matarial it1 the visible region ! ~'S0-699 nm).

.,., D. 1 U

-16 ~2 16:19 FROM: LIEBLER IUEY CC.~r09~35~S8~i 1'CJ: °?03305724 PRC~E:16 .' ~Ct'~US 01 / 0 8 ~ v 6 =~~v~~~ 1 s. MAR~ZOOZ
1 Disclosed herein is an apparatus and method for ner.casarement, with the visible/near-infrared ('VXS/N'IR) spectraseopic technique for sugar content (also 3 known as Brix or soluble sale, which is inversely .related to dry matter content), 4 firmness, acidity, density, p'~;~color trod ir~t~:rnal and external defects and disorders.
F.h The appauatus and trethod is i~ucco9sful in. measuring one car more such characteristic ti in apples, gapes, oraxtgcs, potatoes and chccries. Demonstrated in this disclosure is ~~~_ 7 the ability to combine chcttniettl anti physical property data permitting the prediction ...r 8 of consumer properties, suah ss tsste~, appearance and color,; harvest variables, such as x 9 time for harvest; and starage!varitrbles such as prediction of 4ir~n~eness retention and It) time unr.il spoilage.
Yn.6' , 1 l ~dlP~~~tJ!"~l2~ion ,~"~I~y]~ra,~Ying,~
12 The foregoing and pthi~r f'etttures said adv4~.ntag~as c'f the presont disclosure will 3 become more readily appreciated as the same become better understood by re~Ference 14 to the following detailed desc~n,'ptaat~ of the prefrrr-r~ed embodiment and additional S
embodiments of the disclosure whom taken in conjunction with the accompanying l A drawings, wherein;
17 .,,.
_. 18 FtG. 1 is a top plan showing an etnt~odin,ef,t c~~ the disclosure illustratity a sauuplc '' I 19 holder having a securiing or spring biasing article urging a holding article in contact ?1) with a s;~mple having a sampaurfttce, a light detector havitt~ a tight detector 21 securing or spring biasing article std light sources proximal the sample surface with ?2 the light sources positioned in relation to tl~e light sensor ~enet'a.lly orthogonal to the ?3 sample surface. An optional filter may be positioned between eho light source and tha 24 sample or between the sample'and a spectrometer(s). The light sources may he 35 controlled by the CPU. The autpue i~~ott~ tl~e light sensor becomes the input to a tight 26 detector such as a CCD array'withira a sptsctrometer.
2R Fig. 1 A is a side elevation suction of Fig 1.

-16 02 16:14 FROM:LIEBLER I~JEY CONNOR 509735358 Tf7:70330577~4 , PR~aE:1?
y ~'GT~IS ~1 / X814 ~
IPEAt"'u~ 1 G MAR ~~0~
1 Fig, LB is a side elevation suc~on ofFig l with no sample additionally showing a 2 light sourcs securing ~rticJe,.
<.
Fig. 1 C is a .flow diagram delnon~xrating the method of this invention. The flaw S diagram is schematically repreacntu~tive of all emhodirnents ufthis disclosure.
~r 6 _.~~.
7 Fig. I l~ is a liaw diagram demonstrating the method and app~rratus illustrating the 8 light sources) which illwminate a mple, light collection Cha~ulel5 1...n (light ,tr.
_ ~) ~letecto~- l ...n) of the spectra from a sarriple delivr;red as input to a spectra measuring i U device, shown here as speetrameter 1...n. Spectrometer 1...r~ channels output 1...n tare 1 i converted fi-on~ analog to digital $tid becaane, far each channel, input to a CPU. The Ax ~,.
I 2 CPLI is computer program controlled. The CPU output 15 3150 for each charnel 1.,..n.
,~.

,u,, 14 hig. 1 ~ is a flaw dir~grarr~ dem'onslirating the method and apparatus illustrating the w, i 5 light sources) 12Q as a broad band sourco which illumiaatus a sample 3t);
at least one ::
16 discrete wavr:length f'tltcred (bandpa~ss) photodetectors 255 loaviag fitters 13U for light r~
17 collection channels 1...n frorii a sa,~rnple 3O. in this embodiment a light source 7 2U
_. ;
18 with lamp 123 is controlled >sjr a CPU 172. The speetrmn detected front the sample 19 surface 35 may he c:ornmuni aced by Fbe~r optic tubers as light detectors 80 to the 20 photodotectors 255.
is z1 V:
2? Fi6. I F is a flaw diagram deriions~trcating the method and apparatus illustrating the X.
23 light sources) provided by at least one discrete wavelength light emitting diodes 25?
y 34 to illuminate a sample 3U; at least one broadbmd photodetector 255 and at least one ?5 broa,dband photodetector 255fbr each Ll?D 257 tbr light collection channels '1..,n 26 (photodetector l ...n) of the spectra libm a sample.
'' 7 28 Fig. '' is a top plan depicting at least une light suurce, wiCh a single light source a ?9 shown in this illustration, with optional filter and with at least one light detector, with -16 d2 16: 15 FROM: LIEBLER I41EY C0J'~3E3~85 TG: ~'~330~T7E4 PRCaE:18 ',LAS 01 / 08 ~, 4 b w _~. ~pEJ~u~ ~ ~ I~1~ ZOrO
1 a plurality of light detectors iQustrated, proxin7al to the sampl4 surface.
This Z depiction demonstrates an a ~t8rtid'n of light detectors relative to the direction of 3 light cast on the sample surface with one light cietcctor ariemecl ac approximately 45 ,~ r..
dagrees to the direction oi~ thaligltl cast by the: light source aru~i a sv:cuad light dekector oriervtcd at approx imatcly 1$i~r;d~grees fiom the ditvction. of the light cast by the light ",.
6 source. -~
7 ~::.
Ni -,, 8 Fig. 2~l.is a sactian elevation~vicw of Fil; 2 with the sample removed.
i0 Fig. 2B is a top plan depicti ~~~.a single light source, with optional filters) and with 1(aC, 11 multiple light det~;ctors prox"aVT~nal and du~ccted to illumiu.ate the sarnpie surfa~c~; with l2 both light detectors oriented ~ a~proxirnat~;ly 45 de:grces to the direction of the light 13 east by the light source.

5 Fig. 2C is an elevation view.~fFig 2B.
1 (i 17 Fig. 2D is a section from Fig.C depicting a shieldin.,g ~~nethud. or apparatus, e.b., irr 18 the farni of a bellows or other shielding at~ticle shielding the tight detector li~om l9 arnbienr light and directing thia light detector to detect light spectnrm outlput trom the ?tl sample.

*.
22 Fig. 2F is a detail of a shielding device between the light detector of Fig. 2 at~td a 23 sample. Shown in this illustration is a shield in the fr~rt~n of a bellows.
Other 24 shielding apparahrs and rne~ods will provide like shielding structure.
26 Jr~ig. 3 is a cop plan depictiag'an alternakive eml~odimettt of a light source and light 27 dct.cetor configuration where the light source is co.r,.na-runicated by $ber optics.
.>.
8 ~bt ,. ..
~F~~ ~~

-16 02 16:15 FROM:LIEBLER IVEY COI~VOR 5097353595 70:7033057729 PRrE:l9 ~'J ~r oz r o8 ~tS 1 ~.~
' ~PE,~US= ~ 6 N~4R ZO'D
1 Fig. 3A is a section from Fig, 3. The light source and libht detector may be as 2 described for Fig. l.. Alternative (i$ht source rnay he provided by a plurality of light 3 sources, which. niay be sequeiitially fired fight er7~itiing diodes emitting discrete 4 wavelengths; where LEDs ate emptQyed, the light sei7sor or light detector may be a bruadband photodiode detector central to concentrically positioned LEDs. Fig.

x., illustrates light sources or tamps (end alternatively LEL~s) cot~centricell~y positioned 7 around a broadband light det sector (and alternatively a brvadband photodiode detector x.
8 zSS, such light sourr;cs as well as the light sources 120lLEDs 4S 7, can be placed in t) other arrangements. These arid other configurations also apply iu the use of f ltered photodetectors 255 and broadband lamp !?3 design.
1t 12 Fig. 3B is a section from Fig.,~3 slowing art vmbodimenr where light detectors or li6ht 1 z detection Fibers surround a least onrv light source or light source fibers. The light 14 sourer and light detector maybe as described for Fig. 1. In this representation, the .
centrally positioned light source rts~.y be a lamp or ligUt transmitted from a :r, 1 C spectrometer; the light detection may be by fiber optics transmission with discrete I 7 bandwidth filters between the ;fiber optics ftber and the sample; limiting the 18 transmission by any singly or gents of fibers.
?0 Fig. ~I is a top plan depicting an alltornative embodiment of a light source and light 21 detector configuration.

't-!
:'t:
?3 Fig. 5 is a top plan depicting an alternative embadi,tnent of the disclosure in a hand rt'.~
?4 held case showing a light source and light detector configured in a sampling head. in ~i this embodiment ut the sampling head at last one light source, which may be a 2G tungsten halogen lamp, is positioned ulrelation to discrete-wavelength filtered 27 photodetectors, A shield is illustrated as an an~biLnt shield. The operation of this 3$ embodiment is seen in 1G'ig. 1E wherein ~tll components are ent;ased within the case 2y ?SU.
1.4 -16 02 16:16 FROM:LIEBLER IVEY COhI~ 3~8~ T~7:7033057724 PHCE:20 ~ ~s~os~~
.~ ,~"
U~, 1 G MAR Z~~ .
t Fig. 5A is a side elevation df Fig 5 depicting a sample positi~rr,ed do the saunpiing 2 head.
3 ,.
4 Fig. 5B is an illustration ofthcrrrl~odiment o~t'Fi~;. 5 r~~hcre the sampling head 2bU is in the form of a clamp 2C~3. The light detector 8U is de~pieae~l us ~r fiber optic fiber C transmitting spectrum from the ~amplc to an array of tiltarc:d f 3f) photodetectors Z55 7 or a spectrometer 17(1, Tha o'kput ~~ will b~c t:natiagcd as sEaowti in Fig.
1 D or 1F.
y.

'..l 9 Fig. SC is a section front Fig..58 of the tu~ray of ~ Itered 130 photodetecto~rs 255. A
1 e) positioning structure 79 aecuis and positions the .light detccaor 8U
relative to the ! 1 filtered 13U ph.otodetectors 255.
.~~
13 Fig. SD is an illustration of th ott~odiment of Fig. 5 where in at least one clamp jaw 14 2Gb structure at Jcast one arc pfLUtade'toGtor atTay '~0.
1 S ~_ ...
1b Fig. 5E is a section of the photodetector X55 arrant of Fig. ST.~.

18 Fig. 6 is a kop plan depicting an additional ea~bocfiment of the disclosure in a hand 1') held case. The oger~tiou of tltt~'t~s arnbodiment i:~ seeta in Fil;. l F
wharein all ~1) cOmpOnCntS are CilGaBCd Wit,~Ul tlrC CaS~s ~Sb.

'~:i 22 Fig bA is a section elevation ~ifFig b del.~ictu~.g tile samplini; head showing the 23 ambient shield, tight emitting diodes and photudeteatar or light detector fixed by 24 :~ftixing articles within khe sampling head. The ~aUtpllt fr~atl't the light detector is 35 depicted as we~i as is the; C~tfi ~4~
?7 Fig. 6B is an elevation reprosontativc of an adclitiaraal attvc~dirttent ofthe disclosure ofthis invention and of the embodiment of Fig. G.
29 r ~~ 1 S
.x _ _ __~ .__ CA 02402669 2002-09-12 -16 02 16:16 FRDM:LIEBLER IVEY C<~INOR 5097353585 TD:?033057729 t'RCE:21 ' , ,, ~~~s ~ 6 MAR42002 I Jrig. GC is a plan view of the ~nbodiment ofFig. 6B illustrating a plurality of light 2 defectors, illustrated here as ftbee ap'tic light detectors. Shown in this ihustration are 3 two light c9etectors with otte ~xit~nta! the light sattrce anal another distal froth the 4 light source.
f' :~:
J~.
E> fig. (.~D is a sectivo detail view from Fig. 6B illustrating the light source, lamp, Iight 7 source securing article, case, samplinb head., Light detectors positioned proximal and -~t:
g distal from the light source, light source input anal light detector outputs.
._ 9 1 U Fis. GE is ate elevation view of an tmboditnertt of the discfc~sure of Fig. f~ wherein the 1 t sampling head structure Provided t~~ ambient shirrld stru:ctt~r~.
13 Fig. GF is~ a section detail fib~~ig. GF showing light detectors affixed within the ~~c~.
14 S~Illpllfl~' head ambient shield positioned proximal and distal from the lighk source, a I 5 lamp with lamp input, light detector outputs and n case.
lG
17 Fig. 7 is a side elevation showttng another embodiment in a packing/sorting line form 'fir 18 of the disclosure. Tl7e light source and light detector are positioned proximal the I9 sample.

.~~:
? I Fig. 7A is a section elevation of fiig 7 depicting tlZe fight source, and sample 32 ronvey~rnce system, bracket fixture, light source securing artirte, 1W p input xnd 9.
23 spectrometer as a sa':n.ple mov s into illumination from the light source and toward 24 the light detector, ?5 ?G Fig. 7B is a section elevation of Fig 7 depicting the Iight detector, and sample 27 conveyance SYSLC."~1.1, bra.alset $xture, light detector fxrure, Iibht detector output, 2!i spectrometer, and detector as $ satttple moves toward and under the light detector.
n_ 29 -' LG
t ANI~NDED SNFNT

,..w..
-16 02 16:16 FROM: LIEBL~R TIDY 503'353585 T0: X033057724 PR("E:22 ~~'z'/US O1/0$1~6 ~'~AIU~ X16 II~AR ~,~~ fl,~
I I=ii. 7C is an elevation depiattttg ak least one: light c~.etector ~tWnd as shown a .s;.
plurality of light detectors 80 iepresentative o'f measurements of a plurality of Spelal'ut77 r~~;totis. ~f d : .
Fib. 71J is ~ secaian from Fig~,:'7"C showing thd fz~.~np 1'~3 oriented to illutnuiate the 6 sample From the side, As illustrated, the sample as an apple is illuminated from the 7 seem side, 9 Fig. 7r is a section From Fig.~7C showing one of the light detectors $U.
t0 I ! Fig. 8 is a side elevation shoring sit additional embodiment of the apparatus 1? dlSCIOSed in Fig. 7.
I 't I 4 Fig. 8A is a section elevation of Fig $ depicting the light shield ~tnd at least ogle curtain, light source, and ~tmpls conveyance system as a sat~~le moves into contact !ti with and under the light slhi a. Fig. 3B is a section elevation ol'Fig 8 depicting the 17 light shield, at Jeast one curt ~ li,~lrt detector. and sarttple conveyance system, as a ~.
11; satt7ple moves into contact with and under the tight shield.
w _.. l N
?t) Fib') is an elevation denictiti'~°an addition~cl ~:n~.bodiment ot:
tlm invention 21 demonstrating at least one light dekector 8i) having an autpxtt ~2 to a spectrometer 170 ?? l7.:tvin$ a defector 200. A collumina~tin,$ lens 7S is interrneuiatc the at least one ?:, d~.~tector ~(1 and a sample 3Q.:~The detector ~0 positioned tc, detect light from the ?d sample 30. Light source 12A lamps ! 23; a case ?5() ante~TCe.diate the light source 120 ?5 lamp 1 ?3 and a sample 30 conveyed by ssitnplc aoriveyor X95. An aperture 3G allows illumination of the sample 30 by the at light soaccc 12(1 hunp 123.
A least 27 light shutter 3O0 intcrmedistt~thc light source 12U lar7~p 12=~ end aperture 31U. The 3f; Itaht shorter 3llt) raperable by shutter operating n7eans. The: shutter control means 30S
29 receiving control signals from a CpC.J 1'72 having shutter operating control output ~'' t 7 x .;,;~5 -16 02 16:17 FROP1:LIEBLER IVEY Ct~JOR 509?353585 T'C9:703305772~1 PACE:23 ~''~'~~~~1 / X814 ~P,~rus i s ~R~~~
1 307. A reference lighC tr~rtsmittting means 81. including fiber-optics receiving , .1:
2 reference light. output froth the light sauree t2(? lamp 123. A reference light shutter Z 3() I intermediate the light sotu a 1Z0 lamp 123 and the reference light tr~rt~tfng 4 means R I , The refr;rence lighhuller 3611 operable by shutter control means 305.
The reference liht shutter 301 shutter control means 305 receiving control signals C from a CPU 172 having ~t shutter op~rati~tg control output 30'7. The cefert~nc:e light 7 transmitting means 81 pravidyng an input to tL~e spectrometer 17C). The CPU
17'2 8 providing larnp poweroutput~125 to the light source 120 larttp i23. '-fhe spectrotneier I 7(), recziving input from reference iigh.t tratisu,iiting means $1 lyaviug output 82 received its in input to the CPU 172. The spectrotnetar output 82 capable of AID
I 1 conversion to form input to the GPU 172. The spectrometer 17U, receiving input ~8 l2 from defector output $2 receiv~l as In input to the CT~'C3 172. Mounting means .4~
I 3 indic,ared as described in otlzer_figttres to light sources 120 lamps 123, detectors 80, 14 shuuers 30(), shutter eontml means 3f.15, reference light lransmittiag means 81 and I 5 casE'St). 1<ncoderlpulse gen ator 33i1 input to (.~P~1 172 providing sample conveyor 1 h 2t~5 movement data. Cvinputcr prog~tn.~ to operate Cp~U 1.72 in data collection and J
l7 control functions, !g ~) Fy. ) U illustrates using $peClroBCOpic sensors for measuring fruits and vegetables while. in w ?fl maeion on a symple conveyor 295. Shown is a sample 30 with proximity sensing means '' I i~tt7. remonstrated is the sample owcyor 295, a ctlse 250, col)umating lens 78, ''3 Fig 1 An is a seoti«n from >~'ig. 10 illustrating the proximity sensing means 340 in the form ,.
2~ of reflectance rnr;ans.
-;:
?CW=ig I I IIILIStr:9tL~ the toannervftakiri~ zt reference measurenl~n.t t~fthc light source 120 ?7 lamps) 12~ wherz intensity vs, wavelength output can also tie ~,bta~n~d using ret7ccttng 28 means a6(), ftetlecting mcany 360 muy be inserted vta an aperture 31Q, for example in a case 2~7 25(), when a reference mearwremeTtt is to be made as di~takc.d by ret)ecting control tueans t '~ ANIEN~EI1 !;NFFf ,~..
-16 a2 16:1' FROM: LIE=BLER IUEY COh~R 503T3~35E~5 T~: "~~57?24 PF~E:24 ~~ ~'TL1~ 01 / 081 ~ 6 ~s ~ ~ ~R zr~~z 1 308 as an output li-om a CPtJ 17a, 'Ttte CPIJ 172, via means, WIl detect the presence or 2 ah:mce of a sample 30 and, when a .sample 3C~ is sbsent ft~r "n" time in~.~rements or samlple 3 conveyor Z95 niavementS will ode a reflecting control 'means 3i1~8 control signal to "~
4 reflecting position means 30b,' e.g., linear acttttctxir ar rotary solenoid operated by S means, e.g., mechanical dtwcn by electrical, pneumatic, hydraulic or other power C means.
;"'.

8 fib. I? and 13 wllustrate the m~Chantical insertion c~t'referenc;e means 4313 an nr near the 9 loc;~tion where actual xant~lt 3f? is normally meast.ired. Tnsertion is by insartion means I U including but not limited to an"~tuat~r system 40t1, I 2 t~ig. 1~ and 14A illustrate a mCans of reducing the width c~f apparatus structure by mounting 13 lyht source ! 2U lamps l23 disco! from a sample 3t) with spectrum from the sample 3D
directed by retleating mcatns 36~ arid lens 78 or reference light transmission means 320 15 mth spectra received via a31t7.
er°
16 x.
,..
17 Fy. I S arid !5A illustratca specstra detection tiom st~mple 3U other than discrete inet~mertts, 1 F such oa tipples, including, fur example potato chips, where light st~uree 1 z0 lamps 1:23 1 ~ illuminate the samples) 30 with de>aactors RA receiving inpux with Light detector output 82 ~1) conveyed as Input tv spectrx~mc~ers 1'X0 detectors Z00. In this illustration a lens 130 is ,~".
21 depicted between the ss~mple 30 and the detaator 80. lllustraticms I 5 and 1 ~SA depict in .~;,.
22 detail, with ~I'Cel' I 3t) a.nd mounting rrteans, t> single detector SfJ, ? ~ A C'Pl I 17~, controlled by computer program, is aot depicted in tip;, 14, LOA, 11, 12, 13, 14,.
14 14,1, 15 or I SA as ii person of cirdinary skill will ap~sreciatc: suu>y structure from viewing ..;..
25 other drawings presented heroin 26 ... Detailed :Descriratian 37 Tha uppt~ratus imd method disclosed herein is i.llubtrated in Fag. 1 through 8.
?E; FiB, 1 C. 1 D,, 1 E and I 1' a~tre flow diagrams detnc~n,strati.xx8 tfae method of this ?~ invention. Tha Wow diagram Fig. 1C is eepreseutativc of all en7boditneats cafthis all aw;
' 19 .'.
'~,.
>>~AIIEI~nFn cN~cr -16 02 16:18 FROM:LIEBL1=R IVEY C~OR 9097353585 TO:7033057?24 PRtaE:P5 ' . .: '~~lS o l / 0 814 4 _1P~A~US 1 ~ MaR 2~, l disclosure. The flow diz~gr~ttt~~,Fig. 1D illustrates one or more li,l;ht soetrces 12p and '~''~'r~.
2 multiple channels tiom light d~eteetor 50 through final prediction of sample 3 characteystic. Fi~~_ 1 D demon trattts the method anal apparatus of this disclosure 4 illustratinc the light st~urce(s) 120, which rnay be tamps 123 ur other light sources, which illuminate a sample 30,interiar 36, light collection channels 1...n, composed ti for example of fiber optic fibers &0 or photvdetecturs 255, e.g,, light detector l ...n, of 7 the spectra from a sample 30 d livetrrd as input 82 to a spectra measuring device, shown here as spectrometer( )~'l...n. 170. In tlZe preferred embodiment a light source I?0 with lamp 1?3 is ex.ttnnal to die spectz~omater and is controlled by a CPU

1~ which triggers howar l25 to the light source 12f) lamp 123. Spectrometer 1..,n 170 ,c:
I 1 channels output l ...n are converted from analog to ~iigitul by AlD
converters l ...n 13 171 and becocna, For each eh~ iitel; input to a CPU 172. The C'.PU 172 is computet~
,~~"~
I 3 pru~~ram controlled with eaeb;"stcp, following the CPU 172 in this flow diagram is 14 representative of a computer pTOgram controlled activity. A C'P(.J 172 output is ..".
1 S provided for each channel l.,~n wil~re the steps of t) cal.culati.on ofabsorbance i b spectra i 73 occurs for each ChBmnel l ...n, 2) combine absorbance spectra l 74 into a l 7 single spectntm encompassing°;tht entire wavelength ratage detected from the sample 1$ by spectrometers I ...n 170, 3) inanatical preprocessing or preprocess 175, e.g., 19 smoothing or box car smooth or calculate derivatives, precedes 4) the prediction or .,.
2l1 predict 17C~, for each ehatti~el, comparing the preprocessed combined spectra 175 with ,., 21 the storeeJ calibration specttvm orealibration algorithms) I'~7 for each characteristic ::,..
22 I ...x 17S, e.g., l3rix, firmeess; acidity, density, pfl:, color and external and internal 2a detects and disorders, For which the $arrtple is examinr;d, followed by ~) decisions or 24 further combinations a,nd eomptatisotls oJ:'the results of quantification of each 25 characteristic, l ,..x, e.g., determination of internal end or external defects of disorders 'b 179, t S0; determination of color 181; determination of indexes such as eating quality ?7 inde,~: I 82, appearance duality. index 183 vnd cvncludin~ with sorting or other 28 decisions 184. Sorting or other decisions 184 may fur example lee input process 29 controllers to control packing/sorting lines or may determine. the time to harvest, titnc 3 t) _ ,.,.
ni~t;HnFn cuter -16 t32 16 : 18 FROM : t_IEBL.ER I VEY COh~JBR 5~9?353~t35 T~f : ?Q3395??2~+
PA(a~ : 26 PACT, 01 ~ 0814 ~
:; ~ ~~ .tP~EA~JS ~16 MAR 2~~~
I to ren3ove from cold stor~ged time to ship. fhe apparatuses depicted in F'ig. 1 2 through 8 do not all illustrate the entire flow diagz-arn sequence from i.llurr~ination of 3 sample 3U thrvut;lz detarrrrina~t'ian of"the predicte~~i result; as is clel:ricted in Fig. 1 C, 1D, , 4 1. E and 1 F. FOT StgIlal 'pCOC~35$fll,~ illustrations, reference is r~~ude to the indicated Drawings.
6 Absorbance is calculated as fellows: once; Lhe dark spectruat, rcferenc;e '~L
7 spectivtr~~ a.nd san.~.ple spacttu~m are collected, they are procc;ssed tQ
can~.puta the y,~:
~ absorban.ce spectrum, which taker's Law indicates is prolaotrtional to concentration.
y:
The dark spectrum, which m ~ include b;tck.~rou,adlarnl~i~nt Light, is subtracted from both the sample spactrr.~m andalte rofaranc~~ s~pacit~urn, yfhe l~,g Lease 10 of the D,~ ~
1 ! reforenc~ spectmm divided by~the s~arnple spactrun~ is then calculated.
This is the 12 absorbents spectrum. It is nnt.ed t~lslat dark and referelacc tern bG
collected l 3 p~rica~lically, i.e., they da not necaasarily need tc~ be collected alor~
with every saurtple 1-t spectrum. A stared dark mnd reference can be used if light S~uurce and detector- are stable and don't drift. Pte-pte~sin,~ xtses techniques known to those practicacl in 1 G the trri such as bi.nning, smootl ing, wavelenl;kh ratioing, taking dtrivativcs, spectral .<., 17 normalizing, wavelength subdactiTt~, etc. then thc~,prooessed absarbance spectrum s ~,;..
1 Twill be cun~parod With ra stwreil caftbratio.n algorichn~ to produce :rrr output 1 o representative or predictive of~one ar rrLOre ~cl~.arar.teristics, e.g , firmness, Brix, pti, acidity, density, color, a.nd int~rnai ~cl external defects or acidity, of the sample 30.
",;.
21 Fig. lE is a~ llaw diag~m demonstrating the method acrd apparatus iLLustrating ;,: .
22 the Light sources) 120 as a broad band source, such as a tut~~;stan halogen larrtp, ?3 which illmTlinates a sample 3t1; at lamrst ryne, but in e.n ~mk~~aclir~c~ani a plurality, cal 2~ discrete wavelength filtercc! ~b~ndpass) phutodetectors 25~ ha.vin~;
filters 13(~ provide .,..
spectrum detection i~or light collection ~.hnrwnels l...a~ ~phntocletector 1...n) of the ?G spectra kronr a sample 3U, In This embodiment ti li~;h.t source L20 with Latnp 123 is 27 controlled by a Cf'Ll 172 which tri,g,gars power 1:.?5 to the liy;hl source 12() laax~p 123.
38 The specU-um detected fmm the sat'u~pl~: surfac.~ 3~ rna,y be cot~ir~nunicatad by fiber ?9 optic fibers as lil;ht detectors 80 to tote lahutoc3e.tectors 25 ~. ~l~lie ma~.na~;ement of the ewrrcturlCtl ~.HF r~

-16 02 16:19 FROh1: LIEBLER IVEY COhIN~ 50973S3SB5 ~-t7: 703305772+ PA~aE: 27 ice:

:;:
~:
_IP~AIU~ 1 G MAR 2a~~
l detected spectra i.s as de$criby~ for Fig. 1D. An alternative to this embodiment niay ? use an AOTF, (acvusto-optic~t~un~blc filter) to replace the at last one or a plurality of 3 photodetectors 2S5 as the sp~trum detection device.
4 f,'i.g, 1F is a flow diagram demonstrating the method and apparatus illustrating the li6ht sources) provided by at least one, but in an embodiment a plurality of 6 discrete wavelength light errtjtting diodas 257, which may be se;querttially fared or ....
~..
7 lighted by a CPl'J trigger fflr powdr 125 to illutvi.nate a sam.pi.e 3f); at least one $ broadband photodetector 255°rand, in an alternative embodiment a least one 9 broadband photodetector 25S for each I..ED 257, provide spectrum detection for light 1 al collection ehtjnnels l...n (photodete:etor l...n) of~lhe spectra tiorn a sample. The T..
1 1 management of the detected spectc;E<t is as described for Fig. t D.
Alternative tik;ht I 2 sources tvr this etnbodimnnt include but are riot limited to tunable diode lasers, laser LY
F"
I 3 diode arid a filter wheel placed between the light sourctr(s) and sample or between the 14 sample anal photodetectoT(s). ~
z..
15 Fig. 1, LA and 1 B dtpi_et an embodiment of a Nondestructive Fruit Maturity C., an ~l Quality Tester 1 for measiaring and correlating characteristics of fruit with t 7 cornbined Visible and Nes.r Infra-'Led Spectrum showing arf embodiment ofthe 18 disclosure illustrating a s~rmple.h0lder 5 having a securing or spring biasing article S1 ax..
19 urging a holding article 12 against arid in contact with, a sample 30. The holding r 2t) article dNpicted in Fig. 1 is ilhistrated as essentially a hemisphere sized to receive a 1 sample 30. The sample hits a sample surfaco 35. At least one light source 120 will 22 be em~play~d proximal the sarYiple surface 35. The light source 120 is comprised of at -t 23 least one lamp 123, optional filters 130. Here illustrated are two light sources 120 24 each directed essentially orthogonally to the santpl.~ sttrfnce 3S and illuminating the .u ?5 sample 3O approximately 60 TO 9t7 degrees relative to each other. A light detector 26 80 is depicted as directed to detect light from the sample surface 3S at approximately 27 30'10 45 degrEES relative to the direction. o.f the light cast from either tight source 2b 1?i). The light detector $(1 is itlusttated as positioned by a light detector fixture 50 29 having a light detector securing or spray biasing article 60 placing, holding; and or 3c) ?Z
~,..
AMFIUnFn ~FET

-16 02 16: 20 FROM: LIEBL.ER IUEY Ct31'IIS~J?3S35E35 T0: ~~~3305?724 PACE:28 . Ayr.
~~rru~ 01 / 0 8 t 4 b t~~us ~ s ~R z:
t urging a light detector &U into~con~tact with the sample surw.Face 3~, Mouituring of the 2 light source I ?0 is depicted lay~light detectors 8C1 d~pict~:d as directed toward the lamp 3 123 output; the output t12 of these tr~tarenc~ li~;lat decectArs tii:l is detected by a 4 reference spectrometer 170;'~'slt~rn~ttive to trlae ~tse of two spectrometers 170 will be S the secluen.tial measurement of reference light detectors hand the light detector 80 .~~"-ti directed to the sample surface ~5. Hll Light dateotor 80 are f xed by light detector 7 fixtures 5(J by light detector secu i~,g or spring oiasing articl~a 6t! to a plate ? car other ,.
8 cantaining device such as a cs,~a. 'The secus-irtg at-ticlrv 9 urging the holding article 12 afiainst the sample 30 also ur~~s the scat~npl~ against the light detector 81). The 1 () securing at~tiele ~ and holding ~tiah 12 in cambinatyon witlx the light detectar 8l? and 11 libht detector securing article ~0 stature axed prr;vewxt the sample ~U
from movement.
"~,.
I 2 The sample 30 is shown, in Fy. 1, as an apple, "1"he light saurces 120 may be, for I 3 example, tt~ng5ten/halogen tamps. ~~.n ,optional filter 1 ~t7 or filters 13fl functioning as z ,r 14 heat block, bandl.~t~ss and or cutoffi~'lltcrs, s~[aaTately or in combination, may be :,.
I 5 positioned benwecn the lamp ~23 d the sample 3(l or brtwertn the sample 3(1 and the I 6 light detector SU, The light so~w~ees 12i;f may 1»e lamps 12'3, providr~d For eXturtple by 17 external S~Walt, 75 Vlra,tt, or 150 Watt lamp sources e,ontrolled by a CP'l.l 17a.
18 Power l?5 can be providccl by~pow9ar supply .farom a spectrometer 170 ar li4m an ~..
'°~ I ~ alternate power supply. Both the light somce(s ~ and the spectrometers) arr.
2(f controlled by a CPLJ 172 amd~~their operation can be precisely o:antrolled and .
~.,.
?1 optimally synchronised using digital input/output (IIp) trig,g~r. The tight detector 80, 22 shown here as a fiber-optic sonsar, p>~ovidas ~a li,,glat detector output 82 which 23 becomes the input to a spectrometer 17~, or other spectrLt.nv t,~easuring or pma~assing ?4 instrument, which is detected by s detector 2()0, e.g., at least an ~ light detection 25 device or article, such as ~t CC"D at'ray which m4~,y~ bt: a C'.C~.'D
at~cay within a 26 spectrometer i7U, The sampl~~hol~4e~r ~, light detector ~'ixture 5(l and light detector 27 sacurin~; article b0 and light saurecs 1213 with light sour.~~t~ s~:c~.wing article 12.'x. are ?8 aftixed to a plate 7, for experimental pt~z~poses bu.t will be otherwise enclosed and or 29 affixed in a container, teat, cabinet or other. ox rather Future for canunercial purposes, 30 t:
,. 2,3 14M~N~EI1 SHEET

-16 02 16:20 FROM:LIEBLER IIlEY ClNAaOR 50'37353565 T0:703305772~ PR0E:29 41l 081 ~~
:~: ~G~Il~S
1 e.g.. aprlications include $nd~'~are not limited to sample measurements on high speed A,~.~.-2 sorting and packing lines, hsi<Y~eters, trucks, conveyor-halts and experimental and ,.;
3 laboratory. Dther brackets, f futures or articles may be employed tv secure or position 4 either sample holders 5, light iietars S() and or. san~.pies 3U reQturing only that the i~
device or method used retain the saett.ple :l() in position relative to the light source 120 Cf and tight detector SO during the period of measurement; fiK~l7g methods including 7 welds, bolts, screws, glee, sheet tn'e4stl fornting and other methods may be used to R secure such items for either experimental o.r comn-tercial purposes..
9 F'ig. 2, 2A, 2B, 2C, 2D and 2E depicts an alternative embodiment of the Nondestructive Ftvit Ma~turity~and Qtuttlity Tester 1. depicting a single light source I?U, 4vith lamp 123 axtd optional filter l30 and with multiple light detectors 130 in 1? contact with the sample surface 35. '1"his depiction oFthe relative positioning of the I 3 light detectors 80 with the satripla 30 or sample surface 35 is directed to the shielding r.:
I 4 of tho light detector >30 from ambient light and is intended to demonstrate either direct Y' lS contact between the light detector' 80 and the sample surface; ~iS or shielded a shield 1 G 84 composed, for example, by betlows, a foam structure or other pliable or 17 compressible article or apparatus providing a sealing structure or shield method of I ~t Insuring kllat the light detectoi 80 is shielded from ambient libht and Light from the ""'~ I 9 light source 120 and.receives light spectruJn input solely from the sampte 30. The positionitlg of the light source 12U relative to the light datec;tors 80 illustrate a 21 posirioning of one light detectot~ $0 at angle dteta of approximately 4S
degrees to the 22 direction of the light txs directed by fhe light source 120 to illuminate the s:3mple 30.
23 The second light detector i~0, in this illustration, is at angle Comma of approximately 24 I 80 debrres to the direction of the light as directed by the light source 120. The positioo.in~, of the light detector 80 at approximately 180 degrees to the direction of 2G the libht as directed by the light source 120 msy he a position utilized for the 27 detection oFinternal disorders within the sample, e.g., i.nt:emal disorders within 'Casmanir .Tonagold apples. such as water corn, core rot, .intertt.al browning/breakdown, cat'bon dioxide damage, and, in some c~3scs, insect ?4 f A~NDED SHEET

Je~is -16 02 16: 21 FROM: LIEBLER IVEY CCJNNt~ ~'9?'3535I~1'.' TC~: ?033~577c ~#
PRCE: 30 . ' ' ~ P~TIUS 01 / 0 81 ~ ~~~=
~P~UUS ~ ~ I~R ~~aZ
da,i,age/in.festatioa. The light~cletector;~ 80 in this i.ltustration ~~re suggestive of the 2 many light detector 80 positis possible with the positioning dependent on the 3 sample and the chsracteristic~o'~r chr~racteristi~s to be measured or predicted. lay this illustration the light dete,~tors g0 s.re positiota.lvel to detect within the same platt4 as the tyht directed from the light s urue '120. Tr~c oril~ntation ol~ I ~~CI
tlegrecs between light ti source l2() 'slfld Il,~ht Cl~t~'ctOC~8U will be pref'erx-ecl for stzaaller saxaples. Larger 7 sdrnples 30 will attenuate light transmission thus requiring tire location of the tight a; dctector.BU proximal the light sotau°c~e 12Q tra insutae exposcue tar light $pectruxn output 9 82 ch.aracteristie raf the mpl~~0. 'The octant itton oi° the light source 12U artd light 1 t) detectors 80 is senSitiVe to fruit size, Frr'tit skirt and fruit pulp or flesh properties. The 1 1 orientation where the sattiple 3U is ann apple, will likely preclude a 184 degree l2 urienta.tion because oYlimitF~ti~n9 in proximity ~~nd intensity of the light source 120 as 13 being likely to damage of burn the apple skin. Howevet~, orart f;e skins are less 14 sensitive and may with;stsrrd, v~ith~,t~t commerc;i~tl clegaradatior~, a light source 120 of 1 S high intensity and closely poelt~tot~ew~ to the orange surface, ~,enw~rally, the signal 16 output or light detector output 82 is dependent ort the otirrntattion of the tight source ., l7 120 relative tQ the sample 30 and sample surface ~S axt.d the: light detector 131_1.
I g Thv light detector outputs are illustrated sts providing, rr~puts to apectrotneters.
'"' 19 The outputs Jnay be combined.toprovi.de a single input to a single spe;ctrtrm 2t) measuring and detecting instruimanl: or rna.y sl~par~.tely form inputs to separate 21 spectrometers. For the case o~ a single rneas4ring instrument, light shutters may be 22 used and altet-t~ately activated 1~o prcavide light input from each measuring location r 23 separately in Series, thus prodtieiag two spectra from different depths or locations of a 24 sample.
2S F'ig. 2:B arid 2G depict an s~ltcrnati~re orierttation of light detectors 80 wrhere 26 the light detectors SU are oriented ~t angle th~cre elf approximately 45 de,yees to the 2'7 dire~ctic~n of the light as dirlsctad by the light ~;~aarce l;~tt. '.this illustration ?8 den~onstrukes twc~ lighfi detectciirs $U positioned K~~ppraxirvately Of) degrees apart arid A:.:
29 positioned to detect light from'atapraxitnately rhc. Same plane. One oFordinary skill >:
v'14~UI~NDED ~H~~fi -16 02 16:21 FROf~1:LIEBLER IVEY COh~OR 5097353585 ~T0:7033057724 PRC'aE:31 , P~CTIUS 01 / 081 ~ ~
~PEa~us s s MAR z~a 1 in the art will recognize from'.thase illustrations that the positioning of the light 2 source or light sources and light d~tactor or detectors wiU depend on the 3 measurement intended, Fig. 21~ dt'td 2E depict a shielding tn~ahod or apparatus, e.g., 4 in the torm of a bellows or other shield 84 article. shielding the light detector from ambient light and ena.b)itl.g tha~iig~lt detector to snleJ.y detect light spectrum output 6 fi-om the sample. The shield 8~4 structure may be farmed of a. flexible or pliant ::;, 7 rubber, foam or plastic which°will conform to the surface irregularities of the satuple ._~ 8 and will provide a sealing PunCtior~ between the shicldi>zg material dnd sample surface 9 which will eliminate introduction of ~rnbiant light: Into coritrtel with the light datector.
'.. ,p~.
The shield 84 is depicted in the form of a bellows in Fig. 2n and 2E.
t I l;i~,. 1, 2 - 4, G, 7 and $,depiet light Sources which may he provided by ~;.
L2 spectrometers 17U (as in tha case of rig, 3) or axternal lamps controlled by CPU 172 13 (as in case of Figs. 1., 2 , 4 - 8).~ tn ~,l) cases of Fi6. l - 4, C, '7, and 8, tungsten halogen 14 lamps or the equivs.lent atee used wlKich generally produce a spectrum within the range 1 S of ?SO-1 150 em when the fi]~iment temperature is operated at 2500 to 35L~
degrees 16 l:elvin. The light Source, for the invCntian disclosed heroin may be a broadband 17 lamp. which for examplr:, but without limitation, may be a torrgsren halogen lamp or 18 the equivalent, which may produce to spectrum within the range of 250-1150 em;
"' 1 t~ ocher braadband spectrum Irtrrigs may be en~playad dehendi.rig upon the sample 30, characteristics to be ptrd'ed; and embodiment utilized The tight detector 80 output 21 82 in these etnbodimenls will ge~nergtlly be received by a spectrometer 170 having a 22 detector 2Ut) such as a CCD ~,t'f~y.
23 Fig. 3, 3J~ s.nd 3'B depict an alternative embodiment of a hlondestructive Eruit 24 Maturity and (?uftlity Tester-Combined Unir 1 S of a combini;cl unit )2t1 having a comhined source/detector 135. The source of light and method of light detection in ?G this embodiment may he a light source 120, lamp 123 and light detector 80 27 con.liguratian where the light source 123 larn.p 123 is communicated by fiber ohtii;s ?8 from an ihuminutiap source, e,g., a lamp such as the la~nip at a spectrometer 17t); light 39 detective is provided by light detectors 8(1, e.g., fiber optics or ether manner of light _ ~III~NDED SHEEt i~
-16 D2 16:22 FROM: LIEBLER IUEY COh~R 509?~53~6E TC:): ?e:~305??2~F PR(a~:32 ;r:
w~;..
~~NUS .~ ~A 2002 1 transmission, positioned in varying relationships to the lamp 'l 23 as shown in Fig. 3A - r ;~,.
2 and 3B. Fig, 3A is a sectionifram 1~'ig. 3 sh.owin~p the conubit~.ed unit 12b where a .".~.
3 combined source/detector l3~ae an ~ltet~~ative source of iigl~~t and light dctcc;tian;
4 the source of light, depicted ~s~a plurality o.f sut~rces, toay he sequentially fired light en~ittin,g diodes 2S7 emitting di,sete~te wavelengths; the light detection may be a b broa.dhand photodiade detector 2S5 corttral to cancentricalJy positioned L.EDs, The x" .:
7 combined unit 126 and sample holder 5 are mounted to a plate 7 ur other mounting or 8 containing fixture, case, cabi»bt or ether devices srtitable for ccwmntercial or 9 experimental pun?~ses, fior example with a laracket or other rns~unting article, so as to I U be fixed or as to f~ave a spring or other biasing function to tsr~~;e the combined unit 1?6 11 and sample holder 5 against the sample. A light shield g~4, as depicted in Fig. ~D and 12 2E may be used between the combined sourcel'detector 135 and the sample surfs~ce 1 ~ 35. C'ig. 3B is n section from F'iig. 3 showing a~i additiona.l embodiment of a 14 combined unit 126 where a ectitral~y' positioned light source 1'2U lamp 'l 23, for example light via fiber optics from a tungsten halogen lamp, is concentric to at least 16 one and, as depir:ted here a pluralityr, at" discrete wavelength photodetectors. 'fhe 17 output of the at least one dete~tiott fibers ox light detectors ~~ is the input to a 18 spectrometer 17Q or other spcctt~aJ mestsuring instrument sty~:h as a phatodcteator 255.
I 9 Delaicted. is a spectrometer 1717 h>ivi.mg a detector 2t70. Alternatively, light source delivery arid detection hor the embodiment ofFil;.. ~B rraay lse key a biiiixxcatecl ~a '' I reflectance probe; alternatively, it is recopixcci that a reflectance probe may provide '?2 one or more light delivery scrutcces acrd one or more ligi~t detectur~s providing inputs to 23 one or more spectrometer. While Fig. 3A ii l.ustrates L.Et~s ~;~~
concentrically ?4 positioned around a broadhancl photodiode detector 255, it will be rec;a~mazed that the LEDs of"this embodiment, as well as the light sources 1.2.t~) o~'c~ther embodiments, can 26 be placed in other arrangements, e.g,, the phatodiacie detector 255, as well as the 2? det~;cturs b0 of other embodiments, can be 1 ~(7 dat,~rees opposite a circle of 1..J~Ds 257 28 and the sample 3t7 placed between the LEDs 2S7 a,o,d the plaotodiode detector 255, ?9 e.g.. for cherries Ur grebes; alternatively, the: l:..Ffls °?5'7 ca~5 ire placed on an arc, r,.
~IAII~NDED SHEE"~' -16 02 16:22 FROM:LIEBLER IVEY ~097353S85 70:7033057724 Pf~6E:33 ~z . ' . ~'~ PCT'IUS 0i / 081 ~ 6 t~EAlt~s 1 . ._ 6 MAR Zu l equidistant and 1 RO degrees opposite from the photodetector ?55 in relationship to 2 the sample 30. These two err~angetr~ents are suggestive of the C,ositionin6 ~.
3 relationships of LEDs 2S7 (11~ t s~bwrces 120), hh0todiode dcttctors 2S5(light 'H~~
4 detectors 80) and samples 30 ss well as the insta.o.cc: where other types of light source unnd detectors arc; employed ineludi~ng, Cor example, the use of Filtered photodetectors 255 with a broadband lamp 123, at; illustrated in fig. 5. :hi each embodiment the 7 particular sample 30 Type combined with the particular charactc.nstics to be predicted . , 8 wilt dictate the p.~.ttern of light source 1Z0 and light detector 8U in relation to the sample 30. Additionally, it is to b~ rvcogni~ed that light source used herein includes Y~
~~~k bruadbttnd lamps such as tire tun~ten halogen lamp, 'LEDs and other light emitting 11 devicc;s; light detectors used herein includes fiber optic fibers, photodiade detectors j;
12 and other devic~a sensitivte to'and capable ofdetectin.g tight, ,.
13 Fig. 4 is a top plan depicting an alternative crnbodin~ent of a Nondestructive .t..
14 Fruit Maturity a.nd Quality Tester 1 Showing at least one light source 120 and tamp 123 also light detector SO configuration whoa at least one, and as depicted in this t6 illustration two, light source 1'20 and lamps 123 are communicated by fiber optics to 17 or proximal the sample surface 35, From an illumination source, e.g., a tamp 123 or l8 other external Light source. lLight detection is provided by light detectors 80, e.g"
-"r 1 ~) tiber optics or ether method of light transn:tissi.on. in this emhadiment the light ..s.
,~j.3 21) sources 120 and light deteetor$0 are in contact with the sample surface 35, The light 21 detector 80 detects the light spcctrun5 output from the sample 30 and providing light 22 detector input 82 to a spcctturiinrleasuring or processing instrument or method :;
23 including, for example, a spectrometer 170 having a detector 200. For certain 24 samples, the light defector 80Iwi11 be inserted into the samplo 30 thus effecting :t ?5 shielding of the light detector 80 from ambient light, e.g., on harvester-mounted Zty applications or in a processing plant wrhere tlxe product will be ptncessed such as 27 sugar beets or grapes. Otherwise, the light shield 134 depicted in Fig. 2D
and 2E is Z8 applicable to tlae interrelationship of the samplo 30 and sample surface 35 with the 29 light detector 5C1 and light source 120 and lamp 123. Tllustrated in Fig. 4 is the AMENDED SHED' ,cur.
'- 1 6 a2 16 : 23 FROM : LIEBLER I VEY Ct7N~OR 5~97353~8~ TCl : '7p33a57724 PAraE : 34 - . ~~... P~'~'~JS 01 / 0814 6 .
~P~s ~. s ~R ~oo~
1 GOI'tnelal011 Of thl~, light detoCt~r Autputs 82 fiom the at ls;asY o~~~~;
ii~;bt detector 80 ,. r, 2 forming the input to a. specttvin moasurin,g or larocessing itxstrulment. It will be 3 recognized that each compo ~k ofthis embodiment will be affixed by cor'tventional 4 methods to a plate 7 or other rnottnting or c;4ntainin.g ti.~ture, case, cabinet or other device suitalOe for commarciat~or e~cperi.mc:ntal purposes.
~,~w:
r, Fig. 5 is a top plan depicting att alternative embodiment of the Nondestructive 7 Fruit MuturiUy and Quality Tester I in a hand held case 2'iC) showing a light source 8 12t) and.at least one lighk detector >$0, shown here as six light detectors Ht), 9 configuration in the form ofa sampling head 260_ I:n. this embodiment at the sampling head 2bC1 at least one li.8ht source 120 lamp 123 is positioned in relation to 1 1 light detectors SU provided by~"'at leapt one discrete-wavelength photQdetector ?55.
12 Shown in .fig. 5 are a plurality~oFdiscrete-wavelength photcsdetecturs 255, filling the 13 combined functi«n of light detector 80, and specs°rum detecting instrument such as a 14 GCD array detector 2Qp. Thc,~,vpcr~tian of floss embaditoent is sem in Fig.

wherein all components are ericas~li within the case 2507. Electronic and computer :x.
s 6 comnmlnication between the sampling head 250 and t'he can~puter control circuitry is 17 via electronic signal cabtin~ 2155 or wireless including infrared or other such _. 1g transmission method or ~pparatus.1 The samplin8 head 2tit) ambient shield 262 will 19 provide a shielding rnothod 4r~app$ratus, e.g., fulfilling th.e carne or sirnilrar structural ~,, function as the shield 84 in Fi.~. 2D ttnd 21~, in shielding the at: least one photodeteetor ? l 255 and lamp 123 from ambient light. The sxn~plirt8 head 2rit) and ambient shield ~2 2(y2, depicted in FiB. 5 and SA~mey be formed ki°or,n a pliable y~olyfoam within wlhich 23 the at least one lamp l23 and iii least one I~hot~.,detector .255 nxay bye secured by a 24 fixture article. The maleri:~l of structure Forming tha sato.pling head 26fl and ambient shield 262 may be flexible 4r plia~hle foa'~~ in the form o.f a 1~~c~l lows or other shielding 36 article similar to that depietedrin Fig. 2D and 21~. 3.'he tma of a pliable polyfoam to 27 Form the ambient shield Z62 wall serve to :;eal rrut or prec;luete exposure, by a sealing ?8 ac~iun betwc~;n a sample surface 35 and the 4~n 7hiertt shield 2~2, of the at least one 29 photodetector 255 a.nd lamp l23 from ambient light. Other shielding apparatus arid 2~
~~~~ ~H~T' '.-16 02 16:23 FROM:LIEBLER IVEY CO~R: 5'09?3535B5 TQ:'~03305??24 PHCzE:35 ' ~ v ~'t'~of~08i4a IP~S 16 MAR 200 1 methods wit( provide adequa~ hi~ltling structure ineludivg bellows, a case or box 2 enclosing the sun t.pling head 60 ~d sample 30 yr other such article providing 3 shielding structure between ambient light: and the interface between the sampling 4 head ?GC), the at least one ph'v~datscl4r 25S and lamp J 23 a~~tf the sample 30 and .F~~.
sample surface 35. '.!'he oper~han oFthis embodin~.enl i.s seen in Fig. lE
wherein all a.:
.),v 6 components are encases within the case 250.
:, 7 7n this illustration, Figs, tlste sampling head is arranged so that the -~ 8 phutodetectc~rs a.re concentrically arrayed in relati4n to the light source. The light source may be communicated,'by fiber optics Irom an illumination source, e.g., a lamp within the case or by PI>xcement ofa lamp within the sampling head, e,g,, the ! ! broadhand ou ul !urn c, t n tp p, g., _ gsten halogen, is physically located centrally to 12 concentrically arrayed photodetectors. The light source may be present to be in ~;",.
13 contact with the s2tmple surface or ~roxirnal to the sample surface.
Electrical 14 communication is effected betiNCan the light soctree and photvdetectors and ~~yG
>r~~
computer processor.
I G Fig. S and 5A illustrate~the sampling head 260 arranged so that at least one, 17 and as illustrated in dig. 5, a plurality of discrete-wavelength filtered 13C!
.
18 photodetectors 255 are concenkrically errayed in relation to the centrally positioned at y 1 ~) lean one light: source 120. The light source LZO lamp I23 which nay be communicated by f ber optfcS~from tug iilt~mination Source, e.g., ;~ lamp within the 11 case. 250 or mly, Far partic~tla~.samples 30, e.g., nranges, be present to be in contact 22 with or closely proximal the ample surface 35. Electrical communication and light .., 23 communication i.s eFFected betwean the light source 12U and photodetectors 255 and a 24 spectrometer I 70 by txber optics and or wiring, printed circuit paths, cables. The photoci.etectors 255 fulfil) a speetmnteter or spectral measurement flznetiotr, provides 26 the input 82 which 4vill be processed with microprocessor stored calihration ?7 algorithm to produce an output representing one or tnorc parameters of the sample.
28 1=ig. 5,A is a side elevation dfFig 5 depicting a sample positioned an the sampling ?9 lead.

~':;t ~i.
"~;s FENDED SHEET

-16 02 16;24 FROM:LIEBLER IVEY CtlhlM~2 5097353585 T0;703305772~4 PR~aE:36 ~~ !~~f~s 01 /_ 0 ~ 14 6 j~EArUS
s ~R zooz_ 1 Fig. 5B, 5C, 5D and 5E illustrate ernbodinxent of the invention directed ..a.
2 particularly to small samples 30, c,g., grapes and cherri~;s, 4vhero the sampling head 3 26O is in the :form of a clam 3 havin at least two darn aws 266 which receive p g pJ
4 and secure within at leant one jaw ~6~i sttvcturo at least ot~e lamp 123 having a light a:
S source input 125 and in nt least on~c elarap,javv ~.fi6 strur;t~.tre at Ir~ast one light detector r~
6 80 such that the jaws 266, when the cllmp ~f~3 is closed, receive .a sample r.*~
7 positioned to have the at leas one lamp 123 and the at least ono light detector 8U
- . $ proximal the sample surfxuce 35. The light detector 8(1 is depicted as a filaer optic __ 9 fiber transmitting spectrwnt fmm the ss.mple to ~.n ~trr~y c~f.fill~ered !3U photodetectors 255 or a spectrometer 17(l. a output 82 will be matxaged as shown in F'ig. '1D
or 1 1 1 ~. fig- 5i3 depicts a light de~ectot ~ as a F.tber transrrritting spoctrum from a sample ~f I? 3(.1 to be displayed on a ftltar~ad 130 photodr~tesctor array 2.55 where the fiber 80 is a~,~
13 comained and positioned to tr~i'nsn~~it the detected shectrun~
ft°~an ~ the sample 3t) so .~~.>
1 ~t that the f her 84 is central to ~i cunoentrical.ly arrayed ~ltored 130 photodeteciors 255.
1 S A positioning structure '7~, which may be tubes itttexootar~.ected to position tho fiber 16 light detector 80 central tc~ the photodttector array 255, secures anal positions the light 17 detector RO relative to the filtered t30 photodetectors 2,55. A
~.ollimdtin~; lens 78 will I 8 be pusicioned between khe light detector 801'ib~;r ~~.nd the array 255 to insure that light 1 ~~ from the: light detector 80 is norm~il tea the tiltor~d 130 p6otr~detector aa~ray 255. ~ig.
?U 51~' depicts an arc phoGodetactor array 90 received and secured within ~tt least one jaw 2l 266 structure where the pbotodctccwrs 255 within the photodetector anrdy 9Q
are 22 prefer:~bly equidistant from tits lig~ltt source 120 or tamp 1.23.
?3 Fig. SD is an illustration of the emboditneot of Fib;. 5 whore th,e sampling 2~4 head ''G() is in the fotTrt ofaclamp 263 having a.t !cast two clamp jaws 266 which ?5 receive and secure within at least one,j~tw :2C~Ci structure at least one (a,imp 123 and irr 26 at least on.e clamp jaw 266 structure at bast otae arc photodetector .array g!1 such that 27 the jaws 266, when the clamp°~63 is closed, receive a sa.n~ple 3() positioned to have .~.
?8 the at Least ono lump 123 and the at least on.e arc pltQtodetector array 90 proximal the 2~ sample surface 35. Tl7e arc photodetector array i)0 is d~;pi~c~d as an array of filtered ~ c) ;~ 1 ,~' FENDED MEET

'-16 ~2 16: 25 FROt9: L_IEBLEft IVEY C~IOR 597353585 T0: 7033057724 Pf~E: 3'~
~~~ ~'~sv1r08i46 . ~ 1P
~t,rs 16 Jl~~ 2o~t 1 13Q photodetectors 255 which fill j>'ref4ra61y be equidistant Crom the lamp 123 when 2 a sample 30 is received. The o ttput 82 will he managed as shown in Fig. ID
or IE.
3 Fib?. 6 through 6f illustrate an addi.tiona,l embodiment of the Nondestructive 4 Fruit ~M~aturity and Quality Te~~ I . Fig. 6 is ~t top plan depicting an additional t~rnbudinvcnt of the disclosure i1t a. h~u,d held easy 2S(~ form showing tt light source ?t.~ in the form of LEDs 257 aitd tight detector $t?, in the fc,rm of a photudetector 7 ?SS. cunfi~;uration in the forth of a sampling head 260. With the LED 257 anti _ 8 phatodetector 2SS configuration! the photodetector 255 is used without filters, i,.e., wavelength bandpass filters, bird is sensitive from ~25t)-1151:1 too.
Alternative 1 t) dEmccs or methods for providing light source and light detection includes, but is not I 1 limited to diodelasers and vth~ light soarcr~s producing a discrete wavelength l2 spectrum. Ln this embodiment at the sampling head 26U at least one LED
2S'7, zutd :~
!'ta2~
13 illustrated in Fib;. G, a plur$lity ofLZrDs 257, is positioned in relation at least one 14 photodeteator 255. A method~er article is reqr~ircd to shield the LEDs 257 and phutodctector/photodiod~e detector 255 from ambient .tight which is illustrated as an 11f 1 b amhient shield 2tS2 including struoi~.tses of compressible and pliable foam, bellows as 17 indicated by the shield $4 structure of Fig. 2D and 2E and other such m$terials, fa:.
I 8 structures or articles. In this illu9t~'~tion the sampling head 260 is arranged so that the .j~Y
t"
'°'' l~l at least one photodetector/photodl~xdz detector ?55 is central to concentrically arrayed discrete wavelength LEDs 257. !n this embodiment the tight. emitting diodes 21 fulfill the function of light sates aa~td are sequentially fired yr lighted whit the 22 spectrum output detected by the at least one photodetectur/photudiode detector 255. .
23 The phutudetector 255 output 82 is processed as demonstrated in Fig. lF.
24 The photodetector Z55 is responsive !o a broad ran.gc of wavelengths, both visible and near-infr2u~ed (i.e.; --250-1 I50 n.tn). When each LED 2S7 is f°tred, the light ?6 enters the sample 30, interacts with the sample 30, and re-etnerges to be detected by ? 7 the photodetector 255. 'flee photodetector 25 5 produces ~~ c:urrent proportional to the 28 int~:nsity of light detected. Tlte current is cot,verted. to a voltage, which ie than 39 digitized using an a.rta.log-to-digital converter. ~fhe digital signal is then stored by an '.-16 02 16:25 FROM: t_IEBLER IUEY COhIJOR ~97,~53~85 TCt: ~03~05?724 PRt'aE:38 .
x",01 / 0814 ~
..
,~'EAI~IS ~ 6 MAR 200' ! embedded microcontrdtler/mrcroprocess~.~t~, Tl~e micrdccarrtrotlerlmicroproeessor used y 2 in the preferred embodiment i~;an 'Lt~ttel 8151. H:ov~ever, other miicropmcessors and 3 other devices and circuits wilt, perform the needed tasl~s. The signal detected by the 4 photudetactar 255 as each LED 25? is fired is digitized, .r~/D converted and stored.
S Aflcr ez~ch LLD Z57 has beeti'firad ~d the: converted sigl~r~l stored, the c; micmi~rocessc~r atorrrd readings arc combined to create a. spectrum consisting of as 7 many data points as here ace;.~Ds 257. This spectrum is then used by the embedded 8 microprocessor u~ corn hinaticii~t with a ptevi.ously stored calibration algorithm to 9 predict the sample properties"of interest. ,Sipai Ior4cessing then proceeds as shown . .. _ in Fig. 1F. J;ig, GA is a section alcvatian c~fFig G depicting tltc sampling head 26U
11 showing the ~ttbient shiold 2,6?, c~rposeci for example of compressible foam or 12 bellows or other such structurro, e.g., a rublaer plunger, originally d~csigned for a 13 vacuum pickup tool which 1.'ooks t'rtuCh like a toilet pittnger, but. has a more gentle 14 curve and is av~il~tble in a variety of sizas inclrtding lmcn diatueter and larger; in t S certain of these embodiments ~a 20 mm rubber ptustger was used with a pickup fiber 1 (~ optic operating as the "handle" th~lC couples to the plunger. The sample then makes a s~.
17 seal with the plunger prior to ino~trement. f?ther devices yr methods will also ~<
1 » provide the requisite scttling,s~,truclure, as described in this spccitication. Also shown 19 are light emitting diodes 2S1 and lift detectorlphotadiode detector 80 axed by afl"ixing articles within the sampling head 240. "the affixing articles will be Z ! composed of bracket at'tieles and outer tnountiog structure rocog»ized lay one of 22 ordinary skill. The output 82 fmrth the light detector 80 is de,-~picted as well as the case ~:
23 25t) with processing as shown i» trig. lf,..
,. , 24 ,Fib. 6B, 6C and 6D are representative of eat additional embodiment of the disclosure of this invention whore a sarnplit~~; head. 26U is affixed in a case 250, light ?f.. detectors 8() are affixed by afftxing articles within the sampling head 26U. The Z7 sampling head 250 recei.vcs a sample 3() which is positioned to be illuminated by a Z8 light source t2O lamp 123. This embodiment depicks the case Z50 as having a saver 29 which serves as an ambient shiold 262. Additionally, tl'ae structure of the sampling :.
x 'A.L

,~'r -16 02 16:26 FROP1:LIEBLER IUEY Cl7f~lOR 5097353~9~ TD:7033057724 _ PRC~E:39 ~~ ~C1'~ 01 / 0814 :.
:, u~~ru~ ~, s ~R 2ob~
I head 260 nray he of a compressible or lalir~lale t:uam or a bellows which may provide 2 the structure allowing an ambient shield 26Z, Anrbieot Light can also be measured 3 after the sample 30 is ur placet, b_ ut beforE the light source I 20 lamp )Z3 is fumed on.
4 This ambient light signal is then stored and sober acted lccordingly for subsequent S measurements. A light source input power 1.25 is depicted for example from a 6 sp~;ctrometer 170 or may be from a CPU 1 '72 trigger or other external lamp source 7 and/or power supply. Outpu#S 82 E3-om the light detector./photudiode detectors $() are ...
$ depicted and processed as shown in Pig. )r.
9 Fig, 6C is a plan view~oftht~ embodiment of Fig. 68 illustratir;g a plurality of ....
libht detectors, i Ilustrated heicc as f~t~bcr optic light detectors. Shown in this t t illustration are two light deteors with one proximal the light source and another t? distal fi~om the light source with the purpose L»ing to provide two different 13 pathlengths, shallow and deep, by tailing the di.tterence between the far or deep 14 spectrum anti the near or shallow speetrurn, data. oFgrester accuracy can be obtained.
This difference method provJdes a pathlength correctidrr to improve concentration or 16 property or sample charaete istic prGClictlons.
~,., 17 Fig. 6E acrd 61~ are representative of an enrboditnent of the disclosure wherein 18 the lamp 123 is positioned Within the sampling head 2fi0. Alternarively, the (amp 123 - 19 may he positioned by an etFXing article within the ambient shield 262.
:, Another embodiment in a packinglsorting line form of the disclosure is 21 depicted in Fig. 7, 7A and ~B illustrating a light sotttce 12U and Light detector 80 22 affixed surd positioned by bt'~cket at-ticLus 275, libht detector fixture SO and light 23 source securing articles 122 vi~hieh will be recugtiized as mounting structure fmm 24 which at tense one light source 120 a.nd at least pne light detector 80 will be w suspender), rigidly secured acid otherwise positioned including the use of such as rods, w 26 bars anti other such bracket ai*icle 275 ~Gxi'~~res . The at least one light source 12() is 27 positioned to illunr.inste a sampi~C 3C1, depicted in this drawing as ~t~u~r apple. The at ?$ least one ligltt detector 84 is 'positioned by bracket atkicl.es 275 and light detector Z9 fixture 50 to detect the light spectrtun output fii;orn the illuminated sample 3f.1.
3a AMENDED SH~1' '-16 D2 16:26 FROh1: LTEBLER IUEY CO~IhID,R '509753585 Tt~: 7d3305~?24 PRC~E:40 _ ' . ~ ~~os~ a814b ~EArus ~, 6 MAR ~00~
1 Samples 30, in this illustratioare cotiveyed by a sample conveyor 295. Total r~~
2 exposure to the of least one light source ~12i? 4tncl at least :ant; tight detector 80 wi!! be 3 determined by the intensity o~tho light source used and thi; ~~c~ture of the sample 4 heir interro ~atetl. For a I ~' ex asure tin~e;~ of 5-1 t) n~sec ~~r lGss arc commonl g ~ pp ~ P Y
S used to provide multiple n7easuremetlts per: apple at line speeds up to 20 fruidsecond.
".
G The at least one light delector~,80 depicted in Fig. 7 illatstrates a separation of the light ,:.;r 7 detector 80 from the light source 120 of aphrvximatehy 9C) degrees with both light 8 detcclor 8() and light source 120 essentiall~r ortho~;anal to the sample in the same 9 pl:~ne. kl.owever, for eEmh emliodiwent of. this cJisclosure" khe positioning of the light .s".
dekector(a) 80 a.nd of the light sourccs(es) 'I 2t) relative to eae.h other and relative to r;
I l the sample is dependent on the characteristics of tl~e sample and of the qua3iti~es 12 sought to be measured. For exam.~4c, the light source 12ta n,ay he positioned to be 13 directed essentially orthc~gpn~l to tll°re sample surface 3U irt a plane oriented ~)tf degrees J. «~
.t 14 from the plane to which the light detector $U is directed. "Che light source 120 and 1 S light detector 80 are pos~itionecl pr'aaci~mal the sample 3f). The light source 12U lamp t t5 123 may be powered Cram a spectrometer 1.7t3 or otJ~.er extenaal source, as noted in the dLSL'LISSjOil C7f 1~''1~. 1. "fhc light detector $0 may be a single hb~:r optic h.ber with the 1$ light spectr~.im detected fc~rxnitng tl~ output 82 tea a spectrr.trt~
detection instrument "- t~) such as a spectrometer 1 fit) arid detector 2()U. The processing of the light spectrum .<~w 2t) detected is as described and sort owt in Fig. 1C;.
21 Another ernbodirnerlt directed to sortingrpackin~; lines is seen in Fig.
7C, 7D
22 a~~d 7E depictipg at least atie light detector 8t) arid as showy a plurality of light 23 detectors 8U representative afTmeasurervents ot"a pluraliky oC spectrum regiocts. .A
?4 filtered 130 light detector 80 is representative c~f the detectiml caf spectrum of 700 to 25 92Snm, another light detector 80 is representa,tiwe of detection of red pigrrtettts and 26 chlorophyl in the 500 to 699 nm range and water, alcvhols and physical quality (e.g., ?7 firmness, density) infon~nation available in the 926 ko 11 ~U nm rsu~,ge, another light 28 detector $0 is representative of detection of tl~e yellow pigmse~t region its the range of Z9 ?St:) to X99 nm. Two additional light detet;kors 8t) are shwwn positioned opposite a 30 ' w a5 ";
.,.
I~MENDED ~1E~"

.n~.,.
-16 02 26:27 FROM:LIEBLER IVEY CONNDR 5A97353585 70:7033057724 Pfa.~'aE:41 ~"r~
- ' :~.
r~~u~ ~ s ~~ ~a~
1 light source 12() (temp 123 sa h that the sample will pass botween the hemp 123 and ~r ? light detector 80 arid is representative of an input to two reference spectrometers 170, on a monitoring the 2SC)-X99 '~~i w~veleagth region and the other monitoring the 500-4 1 i 50 nm region., Where the sample is an apple it will be expected that the reference channel additionGtlly will not detect spectrum out of the sample and will indicated the 1. presence or absence of a sample. The oulr~tt oi'th~: reference channeL(s) can be used 7 as an object locator to deierniitte which spectra froth. the sa~pole light detectorf s) to t.~.
_ $ retain for use in prediction. Shieldixrg may be u.ti lized between the light source 120 lamp ! 23 and thi; light detectors I~0 and or sample 30, e.g" options include but are not limited to 1) z light shield 284 as a curtain 285 may extend froth a bracket fixture 275 its::
I t hetwecn the light source 120 lamp 123 and light detectors 8f) reducfng the direct l2 exposure of the light detaetons~80 t~ the light soaree 12(7 lamp 123, 2) the light shield I 3 Z85 may extend between the~light sou:ree 12U lamp 123 and ligdt detectors SU and 14 sample 30 ~whercin an aportutc wil) be formed in the light shield 284 between the I S lyht source 120 lamp 1?3 and satrtple 30 limiting surface reflection from the sample I C~ surface 35 to the light detc~cto s 80 and 3) the light slxield 284 may pros ide filter 130 17 function, e.g.. heat biockirtg,~tttoff and. ba.ndpass, between the light source 120 lamp 18 I 23 and st~mpie surface 3S limiting the possibility of heat or burn damage to the ' l9 satrtple 3U.
An additional embodiincnt is seen in Fig. 8, RA and SB wherein at least one 31 fight shield 284 is positioned by ~ bracket article 275 to separate the at least one light 2? source 120 and lamp 123 from the at Toast ono light detector 80 as a sample 30 is ?3 conveyed by s. sample conve~r 29S under and pest a light snurce 12() and lamb t.23 y ?4 toward and under a light detector 80. The light shield 284 may be a curtain 285 and is depicted in Fig. 8 as a cturtain 285 composed of. at least one portions and as shown m FiB. i~A of two portions or a phtraiity of portions, each suspended from a bracket ?7 article 275. Where: thc;re; are a plurality of curtain 285 portions, the respective curtain ?8 285 nnrtions mill overlap and separate as the sample 30 passes.
~ c~

f-.a/, rY,.
.r ~(W~NDED SHE

'-16 02 16:27 FROM:LIEBLER It~Y CL 5t~7353585 T0:703~577?4 Pf~:~?
. , P~~ 01/ 08146 1PI~UUS- ~ 6 MAR 20C
1 tn this embodim~at, aa,'shown in Fig. 8, the sample 30, I~or exarnpie an apple.
2 is conveyed by a packinglsorting corweyauce system 295. A cyClc will be repeated as 3 earh sample 30 moves to~vard,~~'Intu contact witht, under a~aarj past khe light shield 284.
4 rChe ackin /;~ortin. conve tts stem 295 will have saan,ale~; 3U s uentia.ll P g g Y r., y 1 ~i Y
positiunecl on the Gvnveyance,~ystem 295 such that th,e s~rnc;e between sample 30 is c, minyrna.l generally in relation to the site of the samplr~ 30, A5 the sample 3C1 troves 7 toward, but is not in conlatct w h, the liglxt shield 2$4 the samlole 30 will be $ illurr,i~t~.t.~~t by the light source; l20 while tk~e light detector 80 will detect only ,.
9 amhrient light and will be shielded Crom the l.i,ght source t 20 As the sample 30 - .~t~.
1 t) moves into cvntr~ct with and undet the light shield z$~ thrr sample 317 will, while r t I continuing to be illuminated bythG light source 121, be exposed to the Light detector I 2 SU which will detect ~pectrurri$om. idle sample 30, When the sample 3U
moves past 13 the light shield 2$4 the light detector 80 will again be shielded from the light source 14 12U and will detect only sumbient light. The light source 12(? may, for example, be a I S tunbstet>/halogen lamp or light omitted lay r~ptics to illuminate the sample 30. The I 6 light detector 80, far exatrtplc~a optic fiber dotect'or, is positioned such that the sample I 7 surface 35 vi II be proximal to the light detector 80 as khe sampl,~ 30 contacts skid f;
18 passes under the light shield X84, The ligl'~c slair'ld z8~ traay be composed cif a flexible '' 1 ~ or pliable sheet opar~tte to khe;spectra. to which, the light detector 80 is sensitirre and 2U may he com.priscd, for example, of"silicane ntbber, Mylar, thermoplr~,stic~
and other ? 1 niateri~ls. The light detector 80, lyht shield 2$4 and (i.ghk source 120 wilt bt:
22 mechs.nically affixed by bracket articles 275 or ether tnaunting apparatus or methods 23 readily recognized by those Qf ordinary skill. iu the art yr measurement at ?4 packing/sortin.g systems. =~;
25 An alternative configuration o~the embodatncnts of Fig. 7 and 8 will employ a 26 plurality of light sources lZ0 including, for example a light source 120 illuminating __>7 the sample 30 from the top with asecond light snurce l'2 G1 illuaoinating the sample 34 2$ I'rur-rr the side: or two light sources I 2a illumit7ating the sample 3t~
from apposite sides 29 illusrratin); the multiple positions which may he cn~ployed for light sources 1.2(). A

.-16 02 16: E8 FROM:LIEHLER IUEY CO g09735:~58~ T0:7r~33~57724 PRC~E:~3 ~_ p~cT~1 / fl814 ~p~~ 16 BAR ZQC
I plurality of light detectors 80,wtI1 view the same ar different Sample surface 35 2 locations with Each light det~Ctor 80 autput R'" either sensed by a separate c,~,c a ~~, 3 specrrotneter or combined to form ~. single output 82, Where a plurality of outputs R2 4 are received by a plurality of~sf octrometers 17t) at least one spectrometer 1 7U will S have a neutral density tilt$r iii talJed to block soma percentage, e.g. SO%, of the b o~rtput S2 from the light detector $0 with this spectrometer '170 to provide data from a 7 particular spectral rang, e.g.;~approxima.tely 7t)f) to approxitnateiy 925 tvn. .A second W ~;9 8 spectrometer will not us~a a filter and will saturate ti'om approximately 700 to 925 nm 9 but will yield good signal to noise (SIN ~ data fiom approximately 500 to G99 nrx~ and 1 t) approximately 926 to 1130 nm. Ot~tcr oul'puts 82 to filtered utput spectrometers 17C) 1 l will permit the examination of spcoi~c spectra.) ranges. Additionally, this method ~r~' 12 allows the use of the s~mc exposure times on hotly or a plurality oh spectrometers I 3 170 making thetas easier w control ire parallel. Tlvs is essentially the dual exposure 14 approach using filtered tnput.82 to the Spectrometer 170 rather than different I S exposure times. The blncking~of light to une spectrometer t 7U effects the same result 16 :r5 U5i11d a charter exposure time. 'fete dual intensity approach proves problematic t 7 hecause the high and low intensity spectra are not easily pasted or combined together I 8 due to slope differences in theapectrtt, however the dual intensity approach may be ~.
-~~ 1 y preFerred For predicting c~rtainpara~netcrs (~:.b., firmness, density ) with certain 20 sample types (e.g. s#ored fruit or oranges). While the dust exposure approach yields 21 excellent combined speett~a, both approaches provide useable combined spectra, 22 which are nc;cessary For firmness and other paratrtGter prediction and also improved 25 Brix accuracy. .a.
34 Typically, Partial Least Squires (PLS) re~~ression analysis is used during 25 calibration to genErate a regression vector that relates the V1S and N1R
spectra to 2G brtx, firmness, acidity, density, pPl, colar and external and interns) defects and Z7 disorders. This stored regression vector is referred to its a preelictiou or ct~libration 2s algorithm. Spectral pre-processing routines are performed un the data prior to 2~ reg-ression analysis to improve signal-to-noise ( SIN), remove spectral efrects that are 3i) ..

-16 ~2 16: 28 FROM: LIEBI..EF~ IVEY C:~!'d'97:35,:;~8~i TO: 7~330x7724 PR~:4q , '.~ r . _ ~ ~~~s o ~ ~ a s 14 b ~~ws 1 ~ ~aR zooz 1 unrelated to t(~e parameter pf nterest, e,g., baseline ot~'sots and slope changes, and fi~:x 3 "norntalize" the data by attempting to ntat'hematically correct for pathlength and 3 scattering errors, A pre-proeing routine typirally iirclr.rdes °'birining", e.g., 4 averi~.ging 5-10 detector charmls to improve SIJV, hoxcar or gaussian smoothing (to rmprovc SIN) and computation of a deaivative. The Znd derivative is most often s G used, however, the ist det'ivstiv~ Can also be used ~tnd the uy~: of the 4th derivative is 7 also n possibility, p'or firmn~spredictioia. data i~; rtften used after binning, Xx.
_. 8 smuothin~ a.nd a basoline aortvction or nolntafirr~tion; where no derivutivo is used.
9 For Brix and ocher clt,stnical piroparrties, a .'?nd-derivative transfAmtation ofton is best.
I ~ Using a principal Co'rripan8nts Analysis C~"(.°:A) classification algorithm, soR
1 ! fiwit and very firni frail can b~ut'~qualy identified from rnodcrately firm fruit. Also, a..
12 under-ripe and ripe fruit can biv sr~ted Land spoiled, e.g., higher pH, or rotten fruit ,,a 13 0817 be identified fvr segregarion. The N1R spectra of whole apples, and other fruit, in l4 the approximately 2S0-L 150,~t~:giuu also show correlation with pH and total I S acidity, The Z5U-699 em wa~~el~,8th region contains color irtfotmation, ~o.g., 1 G ~anthophylls, yellow pigmcat~, absorb in the 2~tj-499 rim region;
anthocyanin, which 17 is a red pigmeatt, has an absorptidtt band spttturiing the X500-55t') nin rogion, improves 18 claSSiFcation or predictivB pei'forntartce, particularly fiar firmness. An example is the s~
19 prediction. of how red a cherry is by measuring and applying or competing the ?0 unthocyanin absoiptivtt at or,,near 52p nm to the pr~rtinent predictive or classification 21 slg~,iritlun, Undor-ripe. or~.nges, having a rcen color, caxr. lee predicted by 22 measurement of sample spectrum output 82 in the chlorophyll absorption region 23 (g.reen higtrtcnts) at or rear 680 nt~ and applying the measured output 82 spectrum to 2.~ the pertinent predictive algorithm. TDe spcctru:rn ~nutput from the sample, in the 950-25 11 St) mat region has additional information ~rbuut water, alcohols and acids, and 26 protein content. For oxarrtple, sample water content relates tie fineness in most fruit 37 with water loss occur~rin,g during storage. k=tigh phI &uit* often indicative of a~poilago, 38 can also be uniquely identified in the pt~eserrce of other apples using a classification 29 al~oritl~nt.

3 ~) Y.
AMENDED ~E~T

.,..
'-16 02 16:29 FRDM:LIEBLEF~ TLJEY COhINOR 5097353w6~ TD:7033057724 p~,q5 t~'~ 0 l / 0 814 IPIEAf~I~ ~, 6 MRR~2~( 1 The present disclosures a non-destructive method and apparatus for measuring the spectrum oCsca~tter~d and absorbed light, particularly within the NIR , 3 uange of 250-1 15(:f nor, for Thai, urposc of predicting, by use of the dtpplicable 4 predictivr; aigorithyn, particular fntit charaetvristics incltidi~i~, sugar content, firmness, :w..
S density, p1;3, total acidity, eoloi any! inte~rnaf and external defects.
These fruit t, characteristics are lcey parameters for determininb maturity, e; g., when to pick, when 7 to ship, when anal hc~w to store, attd quality, e.g., sweetness/sottmess ratio and 8 firmness or crispness for many fruits at~.d ve,~etables. 'These characteristics are also 9 indicators ofcansumer taste pmfcrences, expected shelf life, economic value and other characteristics. Internal disorders can also be doteeted, e.g., for T:;smania I 1 Jonagold apples, including disorders such as water core, core rot, interns!
. .
l ? browninglb~reakciown, c;nrbon dioxide darnagc, and, in some cases, insect 1 z dam.a~;e/iofestation. The disclosutro simultaneously utilizes 1): the visible absorption ! 4 region (about 250-Ca99 nrn) that ca~ttfiains information about pi~~rneats cad chlorophyh, I 5 Zl the wavelength portion of the shod-wavelength NIIt that has the greatest f",..
1 t~ penetration depth in biological tissue, especially the tissue of fruits and vegetables 17 (7130-9?5 nor), and 3) the region frotrn 926-115(J ma, which contains information I 8 about n,oisr:ire content and other O-H components such as alcohols and organic acids "'J 19 such as malic, citric, ttnd lat~t~aric acid.
~en~htop, handheld, porta6hle and. automated packin~i3ortirsg embodiments 1 ;ire ciiselos~~d. The benchtop embodiment wilt generally brr distinguished from the 22 high speed packinglsortin~ embodiment through the greater easy of examining the ?3 sample 30 with more thorn one intensity lieyt source 12() , l.c., lamps 123 or light 24 sources I 2() controlled with rreore than one voltal;e or power level or more than anE
exposure time. A benchtop emb4diment discussed herein utilises a dual intensity "4.
26 light source 1.?(.?, e.g., by utilizing dual voltages car du~sl exposure times ar other mrahods of varying the intensity of the li.gh( source 12i) used to illuminate the sample ?8 3C1. Alternatively, the light detector ~0 t.~.iay be operated to prcveide at least rrtye 29 exposure at ane lamp 123 intensity and, for example, the light detector 8(~
may i0 4() .t L
hi~~~unFn cr~hxr ,_16 D2 X6:30 FR0~1:LIEBL..ER IVEY C1R ~09T?~a,i~:~N~ TD:7033eJ5'77c'4 PF~C'aE:46 ~c~~s o 1 ~ o s ~ ~
ups ~ s ~R ZooZ
I provide dual or. a plurality ofcitpvsures at 1 lamp intensity. ?i he method ofproviding Z dual or a plurality of ex.poaures at ono It'tmp Intensity is accortrplishcd as fc~llaws, the 3 light defector 8U e~cposure time~'is awilsustsble tluough basic computer software control.
.~ lai the computer progxam, tw'~pectrwrtt of.'differcnt exposure times are collected for eA~..
each sample 30. The benchtopyrnethod may, as preferred by the operator, involve ti dirr:ct physical contact between the ,tamplc surface 3S and the apparatus delivering ~:
7 the light source 120, e.g., $t least one light detecaor 80 rnay renetrate the sample 8 surface 35 into the sample intoriar. ~A high speed packinglsorting embodiment 9 generally will be limited in the~delivery or the exrusure of the tight gaurce i~U, ~Y
1 a relative to or at the sample su~;facs 35, resulting from the limi.tad time, usually a few .a I l milliseconds, the sample 30 will be in range of the light source 120.
Multiple passes 12 or arrangements of multiple light a~aurccs L 70 and multiple light detectors 80, 13 including nhotodetectors 2SS.~and other Light detection devices, will permit, in the I 4 highspEed packinglsorting erri'bodimenr, the exposure of the saample to multiple light source I?0 intensities. The haindl~'I~I embaduoent generally wilt allow sampling of a 1 G limited number c~f items by orch~r'd operators, i.e., in inspection of fruit satmples on i i 7 the plant or tree, and from pmduaa dclive.red for paclcinglxorting, to centralized ._ . 18 grocery distribution centers r~individual ~;~;rocery stores.
..:.
,~,, ~''~ 19 Obtatining date over the wavelength region of 250-115U nm is only possible using a mufti intensity or ruulti expms~ure Tr~easuremeant, i.e., dual intensity or dual 21 exposure as in the preferred einbar~lin~ent. While one spectrometer can be used to n' Y~
22 ccavEr the S00-1 I SO nm region, a second spectrometer is ne~:~.xssary to cower the 2S0-.a ?3 49~) nm region. 'Che number of different tight ,,our°ce intensity or exposures required ., 24 is dependent vn the charnct~aristica of the sample. and of the detector 200, The .~;' 2S spectrum ncc~uired at longer detector X01) exposure titxxes or higher light source 26 intensity saturates the dctectorpixels, for some detectors, e,.g., Sony 1L,?C Sl 1, or 27 Toshiba 1?Ol , from --~7f)fl-'~2Sa,nm, yet yi~:lds e~cc:eltent SIN data from wS00-&99 ntn ?!~ and from ~92a5-1 150 nm. The low intensity or shorter etcposure time spectrum is ' ?'~ optimised to provide good SIN data from 700-925 nm. Accurate firmness predictions ' ~1 a~.
. ' AMENC1ED SME~"~' '..-16 02 16:30 FROf~I:LIEBLER IVEY CC'>ht~fOR '~037~S35g5 Tt7:7033057724 PRC~E:47 - -~'~ ~ 6 MAR 20~~
ru 1 of fresh and stored ftuir. requires the 7U0-925 nm region and the S00-699 non, e.g_.
t -_ ? pigJnent and chlorvlahyil, plus;~the 5126-1150 nm regioaa. Addition of the 250-499 anal 3 region, c.g., yellow pigrnentS kriow~tt Zs x~.nthophylls which absorb light, will ,,.
~i' irnprove prediction of firmness'antl Qther paraniaters such as F3rix, acidity, p~, color s and internal and external defects. There is hi6h correlation between the specta-urrr ti output from the sample 30 inhe 9~G-11 SO ntn region with water content.
Stored 7 Fruit a;ppears to haue higMec r dtive water content thin fiesta fruit and less light ..:
,, 8 se:~tlering_ Tho el>aorophyll aiid pigment of d sample 3U is predicted by correlation ') with the: sample spectrum output $2 in the 25t)-Gl9 nm region, with this correlatioty likely being the most important for prediction. of firmness of fi~esh fruit, while the ~.
1 I longer wavelength water region may be more important fur accurate firmness f S
l ? measurenz~:nt of stored hrult, .;.
I 3 Just as in the lt~riger N1R wavelength regions, the 7U0-925 nrn region also l4 contains absorption bands from r~tfibc~n-hydngcn, oxygen-hydrogen, and nitrogen-1 S hydrogen bands. e.g., (CH, OH, ~. rn the case whore protein is key component of 1!-.i interest, the 926-1150 a~.ro re,~ n is of greatest interest. However, pre-sprout 17 condition in grain, for example, cyan be predicted by examination of the sample y;.
! R output spectrum in the SOD-699 nt~ region.
--' ~ 19 The profi~rred etnbodiin.ent of the apparatus is composed of at Icast ont light sc~~rrce 12n, a samlsle hvldar S~incltrding, For example a sortinglracking sample i conveyer 295 and other devices ~~I nnethods of positioning a sample 30, with at least 22 one: libht detector $0, i_e, optical Ether light sensors in the preferred embodiment, ,:"
?3 detecting the sample spectrum output 82 to be received by a spectrum measuring ?4 instrument such a.S a spectrometer 170 with a dct~ctor 20U, e.g., a CC.U
array, with the signal thus detected to be computer processed, by a CPtJ 172 having memory, z6 and compared with a stored calibration algorithnz, i.e., stored in GPU 172 memory, 27 praclucing a prediction of orie or ITVO~ characteristics of the s~emple.
The at least one y ?8 likht source 12(7 and at least one light detector SO are positioned relative to the sample ?~7 surface 35 to permit detection of scattered and absorbed spc;ctrmn issuing from the .
~2 s..
;AMEt~DEI~ SHED' '-16 02 16: 31 FRAM: LIEBLER IVEY C'OM~R 509736~~Er~ TQ: 7833057724 F'HCE: 4t3 .
f~o~~oss~b ~~'~s ~ s m~R ~o~~
1 s~rnple. l3racfccfi fixtures 2?5~brackets and uth~r r~cr~gni~~d positioning and affxing ~ devices arid methods will be,eutploy~d to positiur7 light' sources 124, light dete~f,ors 3 8U and s;t~oplo hoiders 5. ln,tlio pr'efemed embodiment the positivriing of the light 4 source i 20 acid tight sensor oraligltt detecar~r 80 wilt be such as to shield 84 the Light detector 8t:1 fi'am direct cxpostae 1o the light source l20 and will limit the light G detector 80 to detection or exposwre of light transmitted from the light source I 20 ~a,~:~
7 through !he sample 3U. The light source 1?t) may fee fixed in a cortical or other cup or .~, -$ shielding container which will allow direct exposure of the light source 120 to tk~e -9 sample su.rf~ce while shi~eldi~g the light source 120 from the light detector 80.
__ ! () A(t~rntitiv~ly, the tight dttect~cir 80 may be fixed in a shielding, cantai.ner, e.g., a shield I 1 84 or s,mbieot shield 2h2, thus shuldin,g the light detector RO from the light source 80 1 ? and expcsin~; the light detector' $0 solely t~~ the f iglit spe~trurt~
transmitted through the ! 3 sample 30 From the light source 80 to the 'light detector 8t7_ ~hhs spectrum detected by ~~:;., 14 the; light detectors 80, i.tr., tha stgiFtal output 82, is directed, as input, to at le~ust one spectrometer 1 ?0 or other dvvico 3angitive to and having dte capttbiliry of receiving 16 itnd measuring light spectrum. Ju the preitenred embodiment two or more 17 spectrometers 170 ace employed. One spectrorncter 1 ?0 monitors the stimple I R channel, i.e., the light detector 80 output RZ, and another spectrometer 170 monitom --- 19 the reference, l.c., Light source 12U channel. if the lamp Z23 is turned vn and vt'f 2t7 hetwtJ~n measuraxxents, atnbivnt tight correction carr be done for bath light detector 1 8t) sad light source 120 ahaiririel, c.~., sperwtrurri collected with no light is subtracted 22 frhm spectrum collected whe~a ligl'rts are 0i~ and stabilized.
Aleornatively, the light 23 sot.irce 1 ?f.> can he lev can and ambient light caiu be physically eliminated using a ø:
?4 shield 84 ur r~mbi~nt shield 2~2, shah as ;a lid or cover r~rr aly~a-opriate light-tigttit box.
zS THe discussion. of shielding of the light detector 80 composed of fiber optic fibers t.' r 2b applies as well to photodetoctara 2~S and the utilization of light sowrces other th~ut ?? tungsten halogen lamps including far example; 1i -;ht exnittirig diodes 257.
z8 Another alternative with mwltiple sampling points and thus multiple light ?~) deiectc~rs S0, as with fiber-optic sensors, l, ti.~ converge .all ~nr some sampling points, 3t) ~~,AMC~IDEU SH~EEi' ,-16 D2 16:31 FROM:LIEBLF_f~ IUEY C01~1NOR 597353585 TO:a03305??24 PR6E:49 .
~C'1'~~s OI l 0814 E
~~~us ~ s ~rR Zoc 1 as depicted in Fig. 4, back to ~~siagle sample or light detector 80 channel 2 spectrometer 1.7(.1, e.g., usin~,~ ifWre~ted, trifurcated or other multiple fiber-optic 3 spectrometer 170 input, hvtulttple or a. plurality of sample points, i.e., light detectors 4 Stl, provides better coverage ofa sc~mpl~ 3l, G.g., sampling is mare representative of the san~.ple 3(? as a whole, or (lows multiple paints, e.g., on a conveyor belt full ot' (l praducl, to be measured by ~ single spectrometer 170 thus providing au "average"' :, 7 spectrum that i5 Used to prediiet an averatge property such as Brix for all sample 30 or 8 light detector 80 ehatmeis.
Tn the prefetTed embodiment two or more spectrometers 170, or at least two -- 10 sp~:Gtrameters 170 arc used for rtfertrnce and or measureazent. A
spectrometer 170 ~i 1 I used in gatherin6 data. far this~remtion utilized gratings blazed dt 750 nra to provide 12 coverage from SOp-1150 nm..Add4ttonally, spectmtneters 1 ~0 operating in the 250--x...
13 49!) nm wavelength region earl be included to provide exptu~ded coverage oFthe -.~.
x_ 14 visible region where xanthop~lls, e.g., yellow pigments, absorb light.
Information in the ou.tPut 82 spectrum detected from 10OQ-1 l4!) nm also contains repeated.
~,.
f. intormr~tion, if a cutoff or long-pass falter is not used, tom 5U0-SSU nm, e.b., 17 regarditlg Anthocyanin, which is a red pigment, has an absorption band spanning the 18 SOU-550 nm region, which it»ptoves elassiticatiori or predictive perlonnance, ~_~
19 particularly far firmness, ?0 The spectrotnerors 174 t~ts~d in the prefi;rred embodiment have charge-coupled ? I dwice (C'C1J) away detecaors 200 with 2U~18 pixels or chx~tnels, but odter array ?3 detect4rs 2(l0, otht;r light detectors 80, including other detectaa~ 2t)Cl si~.es vis-a-vis ?3 array size or other method of detector size characterization, may be used as would be 24 recognized try one of ordinary skill in the atl. One of the two spectrometers 17(f ?5 monitors the light source 120~'t~nte~nsity and wavelength output directly, providing a 26 li~~ht source reference signal 81 that corrects for ambient light hnd lamp, detector, and _'7 cl~ucrortics drift which are largely ea.useil by temperature changes Find lamp aging.
?R 1'he other spectrometers) 170 receives the light detector 8t3 signal output 82 ft-oii~
?9 one or more light detectors 80 which are sensing light output from one ar more ~a ., ~° AAAFNn~n cucct '-16 02 16: ~E FROM: LIEBLER IUEY COI~R ~9'~~~3c;P~; 'rt7 ~ T03~05'~724 PR~aE:

?-~r~~s o ~ i o 8 ~. 4 b~
IPEAru~ ~ s ~R ~ooZ
I samples 3() and/or one or more locations on a sample 3tJ, a.g., nt multiple points over A.
2 et single sample 30, such as an'apple, or at multiple points over a sample conveyor Z 295 belt of apples, grapes or oherri~, or a. different sample 3t1, e.g., a different lane on a p~rc:kir~g/sotting line, c Vibe measured with e~:h additional spectrometer 1.7t)-Each tight sensor, e.g., light ~etesetor 8U(photodetector 25S or other light sensing o appnralus ur rttcthod), in the prr~fetrcd ewbodimenc represents a separate sample 3U or 7 dil'I~;ront location on the samme'sam'plo ;lU c~~- group of samples 30.
Spectra from all 8 spectrometers 17U are acquired, in the preferred embodiment, simult~uneously.
~k.
~) Depending on the type of speatroxnsatcr, A,f~D car~version can acctrr in parallel ar series for each spectrometer (parttllet prt~ferred). The; computer then processes the spectra 11 and produces an output. Curt daringly Cl'L' csampaters process spectra in series. A
12 dual C'PCJ comp~~ter, two computers, or digital si~,mal processuy (DSP) hardware can 13 perform spectral pr~cessirig and provide outpost in parallel, 14 I~n atn nltcrnative cmbimam spectra from the wavelength region from about 250-115U mn, the near-infrared spectra, is exacrtiued &orn s~u~iptes 3U, e.g., trait I 6 including apples. In this particular experiment, a reflectance tibcr-optic probe was -. 17 used tts the lighr detector Bt). k~'Vt~ttile the spectrophotometer 1701 used to collect the a <, x~
1$ data; i.e., sense the spectrum Autput $2 frcant the light rlctector ~0, wtts a DSquarcd I 9 D4valopntent, La~Gra~nde. (he;, Made) DPA 20, anc of ordinary 5k:ill in the art will ~, 2i) rcro~,miLa drat other spectroriieters end spech~pharomet~rs 1'7U may be used. The 21 spcctrophotameter 170 nefer~icsd ernployecl a Uve watt tungsten halogen Light source 22 1?~, a fiber-optics light sensor to detect the spcclrum. or output 82 Pram the sa~npie 30 23 and provide the light sensor signal input L~? tea the spectrortteter 17t), other lamps 123 2~ ur libht sources !?t1 ratty be substitut~;d a.s well as other light sorisars uc light detectors ? 5 ~0. 'fhe light detector sig;naiinput g2 to the spectrometer 17(x, in this embodiment, is ~2c~ detected Gy a charge cawplad device arra>> detector 2(.i~(f. 't'he output from the charge 27 coupled device arr~~.y detector is processed as described above. Firmness and Brix 28 were measured using the standard destructive procedures c~f Tvlagrtess-Taylor finzuness ?9 ("hunch test") and refractometry, respectively. In this emhosliment the N(R
spectra is 3t) ~S

~:°- ~'r'I~ 01 / 0814 ~~~ ~, s BAR-Zap YF.
I detected by m array detector- 20U which pem,its recordin,~ or. detection of 1.1724 data 2 prints. The 1024 data points are smoothed using d ni~nc~polzat gaussian smooth, 3 fol lowed by a 2nd-derivative transfon»ation using a "'' sixa of rti.nG
paints. Partial 4 least squares lPLS) regression was used to retake tile: Zrid-derivative NIR
spectra to Brix and firmness. To ensure that false correiatiora vas not occurt~ing, the method u1 G Iehve-one-out crass-validation was used to ~era~,~rate ctand~td errors of prediction. In "-:,"
_ 7 cross-validation, the prediction model is constructed using all but cane sample: the 8 Brix and firmness of the sample left aut 1.s theu predicted and the process repeated ..;:.:
__ 9 until all samples have been predicted. 'hhe validated mortal sari then be used to ::
1 U nondestructively p~redici. Brix and Firmness ire unknown whole Cntit scruples. This 1 1 information guides harvest decisions indicating tirxte to harvest, which Fruit is szritahle 12 for cold storage, where the fruit is classified from acceptable tQ
unwcoptable 13 Characteristics of quality or consumer taste, which fruit to be removed tr'on.~ the 1.4 sorting/packirig operation as not meeting reiluired characteristics, e.g., Crrrtrtess, Hrix, M. .1 color and other characteristics.
16 'hhis disclosure oFrmbodiments of an appals and metlyod is dixetteas to the ~~.
17 simultaneous mcasurecrsent and use of more than ot~ spectral xegion from a sample.
18 In this.embodiment the use of the chlorophyll trbsorption~region urtd tlae '~T1R region, 19 including the highly absorbing 95U-1 t St'1 O-.H region, is: accomplished by exposing the sample, e.g. apple, to more than one intensity sauce of light or by e.cpositig the 21 light detector 8p at snore Chart one exposure tin~.e, e.g., a dual intensity source: of light 22 ur at lea..st kwu intensitira of lil;ht, or by detecting light From ~.
sample with n~rore than 23 one light detector SU such that each light detecke~r $O is sensitive tc~ a ~iiCter~nt n z4 spectrum.. e.g., by filleting one or more light detectors 8t7 with ~Iterity eiths:r between the sample 3t) anal the light dr,tcctor $0 ur betw4en the light detector SU or tput 82 and ..
2t1 the spectrometer. 170 input. Fig. 1 illustrates filterod light sources 1zU
allowing 27 exposure of the sample 3U to diffErent lyht intensities.gPig. 2 illustrated th.e use of ;..
2$ mare than on.e light detector $0 where tilterin.g betv~seit'tlae sample 3t?
and light :.
2y di;tcctar 8U allows detection of different spectral r.~gions~. Shown its f 1g. ~A, where ~G
..~
h....
AMENDED ~IE~' '~:
~r _ _ _ _ _ . , . , . ~..~ r. . r rsr W r,m ~r~~ t -,.n r-J~7'tC7~ 7~ °~
. t.ls7~! .J 7c~ ~ p T 7f.'1 Q T _ ;Jf v. ".~.~5 Q 1 / 0 814 b ~-° ~~ru~ ~~, s ~aR zooz :. ..
1 the light source is a plurality of discrete wave;letxgth GEhs 257, i,s an embodiment 2 wherein tiae sample i5 ex~rosod to a plurality of ligh4 i~titi~c. ff'he intensity of the 3 light source 12C) will be selected to provide li~;hl output to~~th~ light detector SO which 4 will give optimal SIN data in the desired spectral r~giota.~ in a first pass a light source.
e.g,, a lower intensity light source, is used to illtttrtittsttc tlae satt~apl~~, e.g. apply, to ,.~..
6 ohtain data, with an acceptable S/'vl ratio, in tlr~ 700~'9s5tun rtsgion.
A.t higher (='925 7 t>Irt) atad lower {=7U(f nm) wavelengths, the spectnu» is dorni rated by noi.so due to the ~,~
8 low light levels and is not useful. Tn a second pass a i~iger intensity light source is ,_ 9 selected to illuminate the sample, saturating the ~letc~;tor~~'arx'ay at the 700-925nm 1 U regions while obtaining data with an acceptable SIN rativ~"in thr; red pigment rel;iun 11 of S()t)-6U0 nm, the chlorophyll region of 600-b~~»~ raratd iw the C)-~
region of 9~C
"a 12 10U0nm. The data from each of the. two passes con~ris~ sap~u~ate data inputs 13 delivered to an analog to digital converter for co~pu~rr processing. Same 14 spectrometer and AID for benchtop unit, where the t~vo ~puctra 4tre acquired I 5 sequentially. For on-link two spectrot~~elors are used, each with its own A/D, ;hs one r,, .
16 embodiment AVD cards external to tl,e computer t~t~a utilixcd which arr.
serial ~~d are _ . 17 provided by Ocean Optics. This process rs provides ~or,~tttltiplo chs~t~nele into a data ~.".
t$ analyzer for ;analysis by software. In this cntboditn.~rtt C~~Ceart Optics drivers, herta$er 19 refcn-cd to as drivers, accept MS "C°' or ~° is~tal Ba~i~c~
toy) dctetmive the spectrum 20 detected from the sample or 2) subject the data to floe pc~~,ictive algori.th.m and 21 produce the output. Displ~~y control rompu~er;~rrogram~or soltwara periodically 22 requests dt-iv~rs to deliver the spectntnas to be combined: Ttte digital combination 23 t,ben produces, with standard display software, the outptifdisplay representing the 24 entire. spectrum ranges clctccted from the each satxtplc. There hay be, for each ., 25 sample, n~ultiplo spectrt>am data. F'ur exanxple the s~ctru~t:n ssxmpling protocol rnay 26 seek 5U sp~;ctrum satt>lples during each of the multiple ~rassw~s, e.g., 5t) spectrum 27 sarnples during the pass subjecting the 1»ruit sample to ttie lower intwsity li~,ht source 28 and separately 50 spectrum samples dunnl; tlae~ paes subjeatittg the fruit sample to the 29 higher intensity light s~~urce. The total duration ox aach~ pass will be dc,ter>Lrrined by "l F..

4~
x~
. _.._...~.- . . ,., . , .. . ....., .».~.., . it<w...wr, . -, n ~r ~. i-v~r.,-,~.~ . m,~, n , rr . nT 7fa OT _ JI-JI 1 ~T~l~ 01 / 0 814 !
T~p~~ 1 s BAR zoo2 1 the speed of the sorting/paclcing line and may be limaited to approximately Stns pc.x~
2 sample. However, it will be recognized, for all embodiments aad sample types, >hat 3 other sampling times and strategies will be within the realm ofusc for the inv~ntiorz 4 disclosed herein as different samples and different tstahbodimeets are employed.
Where the samples being processed, on a surting/packmg~line, are appl~;s, there is C expected to be little space hetween each successive apple. Spechum obtained fi-om ,: .
w ~ 7 the space between apples a,nd at the leading and trailing s aes of the sample or apple ~, $ will be i~iscarded. As the sampler, i.e.. apple or other fruit, moves under the light . _ ~ 9 detector 80, the spectrum. data detected will be that exiting the sample 30 1U representative ofthe poriiun of the sample 30 cpnstifinting the path between the point 11 of exposure of. the sample 30 with the light source 120 arid the point of spectrum exit ,p.
12 for detection by the light detector 8f1. By mathematical iiispaction. afc~h spectrum, l3 e.g., automated tnspectivn via a computer, this method ~c:~n determine whEther light 14 detected by the light detector 8U is from au apply or the empty space between apples in a satting/paekiug line sample conveyor 295. 2h~is method can also detect the 1(i leading and trailing edgos of. an apple as it passes by the~7li~ht detECtox 80 having an 17 output 82 to a spectrometer 170. .prom this data, disoriminstion can c~c:our to select ."~
18 specific spectra samples which, for exarnplc, are cxpectcil to be from the midsection _". .;.
s 19 of the sample or apple. t,Jsing mathr,matical inspection cafeach speetrutxt (on-line) to determine i.f it is a good apple spectrum or a speetrtttn o~'the litre material. The cycle 2 l detected. by the light detector 80 thus, for each sample 30 itx the on the sample 22 conveyor 295 of a sot~ting/pacl~ing line. is composed afan initial segtnent where the 23 iil;ht detector 8U or pickup fiber is exposed to only ambient light with a ligt~tt shield 24 2g4 between the light detector $(1 and the 1ig12t source 120. As the sarmpie 3U, c.,g., apple, n~taws into contact with artd under the light shield 284, which m~~y for example 2Ci be s curtain 285, the loading edge or side oFthe apple will continence to be revealed -., 27 permitting the light detector 8(,') to detect spectrum output 82 Crow the apple.
~w 28 Continued movement of the sample 30 under the light sb~ield 284 ~:xposes the light ~.,, 2d detector 80 t:o Sptch~um output 82 from thr: san7plc 30 until the sample;
3Q moms to 3t;1 ~s .
,.
a..~.
_ ,. ~.
_ . .. ..-..., . .-...,... . arm vr,... ~ -r., r ~r,-~~~ T-~ . "np ..W 'r~ .
r~T ~~~ OT _ -!1-!P'~1'I~JS 01 / 0 814 ~
_~~:~1~ . ~. s BAR-ZOQz the paint where the trailing rdge oa- side ot'the sample 3~O~is rema~uLn~;
exposed to the 2 light sotarcc 120. The Sample 3(1 then moves pit tJ~ Mitt st~ieltl 284 :ictd all 'light 3 from the (i.ght sotuce l.2tl is blocked i~etweetl the li~;h~t detector 8fJ
at~d the light source 4 120. Thus the i:rtiiial spectra detected by tlyG light detact~r 80 will be at the le~aditig edge or side of the sa,mpl~; 3G1 as it appr~.~aahes the cwt~taita~ 285. 'Che ivtemtediate 6 spectrum measuretoents, between tl7e initial tune at which the leading edge of the 7 s;~n~ple 30 is exposed to the light sours a 12t) ~c,d the timb whrSn the trailing erige or ~~.
8 side of the sample 30 is exposed to the ligltl source l~tJ,'iwilJ include those where ttte .- 9 light detector RQ or light pickup is aptimal.ly pusitior~d to detect spectra most representative of the charatcteristics of the li,gkmt spectra oixtput 82 from the sample 30 I I tts the i.ight source 12t) iltunuittates the sutrtple 30, e.~g., apple, other fruit or otktet O-H.
12 C-H or N-.H materials. In cht: preferred embodirttertt, fnt:''ease of data processing, the I3 libht detector 80 analog output 82 is converted to digital da~ka by an AID
card.
14 Computer program or software tests the. data for au:oejatexice or discarding. T'he criteria for acceptance ofeach spectrurrt ss~x~ple 30 Ix a .e~tcrrnined spectral feature 16 determined by the expected spectral output 82 oaf the ~ai~riple ~U, a.g., where the 17 sample 3Q is curt apple, l.c., the c»teria will be to deta~ci 2i~s~Ct~rum fmm 250 to 18 l 150nrrt. falling within the spectra expected for art atppte.'~'I"he detection of the space 19 between apples, irt the sortinglpacki.ng lin~;~ will be t~cog~tiratd as not apples. This spcctrutn ;acyuit:ed for each sarr~ple ?~(~ is the intl5ut to t~t~ predictive algorithms as 21 indicated by the flow diagram of Fig. 1.C. Multiple spectntm, for cx~tnpie fifty 22 spr;ctrutn, are detected by the liyltt detector ~U for ~~uch ~i~tnple.
'~'he computer 23 program compares each detected discrete specta~unn with an expected spectrum from 24 the partictalar sarnplc, the sp~ecttvm not tr~ee;tin,g the ct~it~txa are discarded, the retained spectrum, e.g., 40 - SO samples, are combined to provide the spectrum which 2(i becomes the input for the predictive al~;orithna. lvlttltipe spectra fratn the saztze apple 27 are averaged to provide a single average spectrum representing nnultiple points ou the 28 apple, the apple may be spinninb as it trave.l.s by the seri~or, e.g., clockwise car counter 2y clockwise in relation to the direction of sorting Line travel with better measurement .
,.
~,:
-. . _____. .~... ....-,.-..-....P.,...mr. .,rn~n,.Mrr, r~n.r vl-e-W
7T'1~yn~.1.1 L.S~~t~T 71a OT_SJHI~

~'ftiS 01 / p 8 ~, 4 E
1PEA~U~ 1 ~ MdQ-~nn~
). indicated with caurtterclockwise motion of the saxxtpllas, thus giving even eater 2 coverage of its surface. Uuce the average aUsonbata~c~ spectrtattt Far a san~.ple is 3 calculated, the spectreml is multiplied by the regression vector (via a vector ...
4 multiplication dot product)_ The re6rcssion vector is obtained From previous calibration efforts and is stored on the camp~tter. Tttete iav separate regression b vector for each p~3ratneter b~;ing predicted - e.g., lim~rressBrix. The results of the - 7 processing the spectrum output 82 by the predictive alsoritbums will d~tennit~~ xh~
8 predicted characteristics o.f the sample zU. T.he charaeteiistirx detmnined.
for each :~.
9 discrete sample 30, e.g., apple or other fruit, will be us~dr decision.
making in handling or dispositioh of the sample 3U including, FOr e7~$mpl~, 1} in the t 1 packing/sorting line different characl~:ristics will be :fox sorting and pack.inl;
"' ~
1Z decisions, e.g., by color, sire, firmness, tasty;. as predicted by acidity anct Brix anc~ 2;) ,.
I3 characteristics indicating spoilage m.ay trigger t~.eti~ds of e)i~r~~ination of the 14 particular sarmpl~ 30 tTOm the packing/sorting line.
Packing and sorting of. apples wilt likely involveriittltipla packing/sortin,g .,.
l6 illumination. or light source 120 and light. detector $0s fo~~eacta line.
W'lyere the 17 sample 30 is comprised of smaller Emit, e.i;., cherrios or,grapes, there;
may be :,.
18 multiple light sensors with single or multiple light lo interrogate or examine and __ 1 y gather data ~rom a tray of such smaller fi-uit rather tJ~ on the basis of exatninat~ion of 2U each discrete cherry or b~rape. For each sample 30, dots is~acquircd, tested tc~
.,:
21 determine of the data conesponds to preset criteria with data sclectecl which meets 22 preset criteria and discarded if it fails to meet preset criteria, Dale received by light fR-, 23 sensors is then combined tv compose the total speett'um~ampled. T'he total spectrum 24 is then cotnpared with the pretiictim algorithm and decisions are n~~rdo regardin g the sample 3U including, for example, sartinglpackmg decins. The results of the ..
26 comparison of the total spectrum with the predictive algorithm provides a number or 27 other output For end use inefuding information 1'or cotnnuter diAecte;u sncting M
28 ecpipme~,t. i'~,,:
..
29 r 5 0 ';.,~
a ., r;.
~!3;.
_ . . _~.-... . . .-. .
,. , .--,.~ ,-. s rr-m~ yn~.W n~°o l ~,/,,T -,.17"S~'~7T't ~
f.ll'~1,~.1 J 8~.r~ ~ qT ?f..'1 qT .-.'.-tHl~

4, ' ~'1'fUS 01 / 0 814 6 ~~~u~ .~ s ~aR za~~z ,r 1 Operation of. the light source 120 is eoa.bles the rapid ~cyuisition of 2 reproducible data with good S/N, even ire the l~aigh(y kight~scat~arinh and absorbing 3 25(.)rv(~yy nm and the strongly a.bsorhing >950 rrrn region. , The lamp 123 in tlae 4 preferred embodiment is a 12~Volt, 75~~Watt tungsten halragcn °~atn;p. However, other light sources whii;h naay 1~e used iraelude (~rrt are not iiart~it~d to light enutting diode, 6 laser diode, tunable diode laser, flash lamp aald other such sot~rcers which will provide 7 equivalent light source and will be t'a~oiliar to those ~rsuctided in. the art. T°he lamp is y~ 4.>
8 held at a resting voltage u(' 2-Volts. Whim :~ meaa~tre~raeiat is tah.en, the lamp is ... 9 r~amped up to the desired voltage, a brief delay allows ch~lamp output 82 to stabilize, 1 U then spectra are acquired. Aver data acrduisitioaa, the Isnrip is ra.mped down to kkte l t resting voltage. This procedure extends lamp Ii:E'e mnd pr~v~mts burrt(ng the sample.
°~a.
12 In high speed operations the lamp may always be Iigk~te~d~°"e.g., uaa m~ high-speed 13 packinglsorting line or used on harvest ~:quipn~er~t, rrr~ a ligttt''ahopper'~ or shutter or ,.~~
14 other e4.uivalent articlr; or method could he utilized to deliver light to the passim;
I 5 sample for v determined period of tix,~e. T°tae operatiatt iaftl~ae light source is l6 important in extending lamp fife, redu~:~ng upc.~°aaittg ox 4psc at.°td xeducitag disruption 17 of opsraticms. The lamp 1 Z3 voltage ib rarnped ttp and d~bwn to preserve lamp 1'13 18 life and to lessen the likelihood ofburning irc~it, ,~, atll~y voltage to keeps the lamp 19 123 filament warm. An ao,(a(ent/rc~orn light ba.clrgrc~~tnd"~ncaaawt~remertt is trade to 2U correct tilt the dark spectrum" which rn.3y ioctudu atnbiein°t light, It is stored and Z1 subtracted lrotn the sample and ref~reaace iilvapp(icxlxle) so that there is no 22 contribution of ambient light to the sample: spee;trwrn, which would ttffeat 3ccura,Cy..
23 Dual intensity illu.n~in~ttustl is employed to: 1) improve data accuracy above X25 nm Z4 and below 700 ram and 2) to narmali~.e path length olaan~es due to scattering° ~~yal 25 exposure time increases the likelihood of itacr-eascd dat~quality with large and small 2ti fruit. Utilization of more than one ti~;lat detc~ct~,~r i~0, wit~t each positioned at different r, 4,.
27 distrnces from the sample. will likewise iwlcredse the ability to obtaizp incxeased data 28 quality throughout each portion of the spectr.~.un. irocxt approxymately 250ram to 11 1.50 Z9 rang. . ' 30 ~( . . _ _ _ . _ - . . ,r.._. "~.,. ° .-,.."... M ~.,rvc ru nr,.., , _.,~.r ~. rte--nrw~r~p . y-n. n i r~r . n t 701 .'~ 7 _ J4-11 I

. ~'1'fUS 01/08,4!
-.~-- r~:Anu ~. s MaR-2o~
..
Othez steps in determining predictive al~;otythmts included reference 2 del~rminativa~ of pH using electrode measurement ~t~d raferenev determination of 3 tot31 acidity using end-point tit.ratiun of extracted juice. Correlation between the 1~1IR
4 spectra and the reference data (pl:l and total acidity) was conducted.
Methods l:no~-n S to those practicEd in the art such as partial lEast squat'es (P'LS) are used to determine 6 the correlation of the NCR spectrum with a chosen parameter such as p.H,.
lance 7 correlation is established, PLS is used to generate a. xcgrossion vector, froth the.
ry 8 calibration samples, This regression vector is khen tt&ed to~ predict sample pzopetties - 9 by taking the dot product of the sample spectmm and t'egi'ession vector.
N1R analysis can be ca,rtied our directly on the juice yieldinb very high corrclations with Brix, p:H, I l and total acidity. A eumtxtercially available "'dip probe"t is usse! that is a comrnun 12 item available from optical Ctber fabricators or from. catripanies involved in process 13 analysis. in addition to tlae use of PI.S for quantifyi:~anx, ftrn~ttess, phh and acidity, 14 Principal Components Analysis (PCA) was performed o l thia NTR sp~etral data. PCA
differs from fLS in that no reference data is required. ACA allows classiUcation of l6 firm vs, svft apples and low pH vs. high pH samples. This classification algorith.n, is 17 suf~tcient to achieve the goat of product segregation. Using PGA, poor quality fruit .
I $ can be removed frvm a batch and the hil;hest quality frui( t;at1 be segregated into a 19 pzemium class. Poor qualify fruit was obs~.~rved tw tts''have a higttet pht level than *.
2U good quality fruit. 't 21 Fig. 4 illustrates art alternative embodimetti ofthe disclosure and includes at 22 le:~st one light source 12U transmitted by a transmitting a.rticl.e, !'ur example a fiber Y' -, 23 optic fiber ox other equivalent article fur tnansmittin,g light; a sample 30 having an Z4 sample surface 35; input mechanism of positioning ligh't;from the at least oue Light source 120 proximal the sample surface; at least ones illumination detector;
output 26 mechanism of hositiomng the at least one illunt.ination detector proximal the sample 27 surface; the at least one light source 12U and the at leasfone illumination detector 28 rnay he positioned in relation to the surface or against tbe.surtace by a.
positiunittg 29 article provided, for example, by a positioning atfiicle spring biased against the .
:~,.
5?
_ ...~.;_ .. ,_.-, ,_-. , ._,...,.. w .r,r a .,-...~ . --,.n , . r ,-,r,-, r °-, . , ,r,~. . , ,,...,, ~ ,'~ T 7ri o r _ '..1~

t~r~s 01 ~ 0 ~ 14 6 IFE~W~ ~, 6 ~~ 282 t, surCacc of the sample; thu pressure against ~ sa.a.-tplc st,~e~'ace~ by an at least one ,light Z source 1 ZO or an at least ont. illumi~aatio" det~;ctc~r, vr~i;~l'bb limited by st.trface 3 characteristics of the sample anchor the character of the me~asurctnetat process, i.e., ~., 4 pressure may b4 reduced where a sample is suiaj~;ct to surface damage or where the measurcrncnt process is m at high speed fi~,zitin~ tfie titnt~ permitted for each separate P..
6 sample contact. The illumination is transrsutted to the surface, for example by 1~iber 7 optics or ot),aer equivalent nxa~aa~er; and at f~.lst one davic~ ox method of rneasuriiag the 8 illumination detected frvn, the sample. Tl he l.s~~htt s~wrcell'or the disctasure herein _ J rnay be a broadband lamp, which for exan -whle;, but witt~tt~ut limitation" may be a 1 U tungsten halogen lamp or the cquivalEnt, which xayaty prowitacc a spectmtaa within the w~-l l range 25U-115U ntn and have a filament tetn~peratur~ of,of25UtJ to 35170 degrees 12 kElvin; other broadband spectrum lamp: i~~~siy be eta~tphay~d depending upon tla.e 13 sample 3U, characteristics to,b~; predicted, a~~td embadin~nt Wtali~ed; the a.t least one 14 devicE or method of measuring the illumirsation rrta~rbo'8~~p~clronmter having at least one input; the at least unc spectrotoeter may include, fot' exannple, a 1e)24 linear array lei detector with those of~ordln3ry skill in the ~trrt rec.ug~ixit~g that other such detectors 17 will provide equivalent ah;tcction; the at lu.ast one iUumi~tiation detector tnay be a light 18 pickup fiber or other Equivalent det~;tar incCt~ding Cor eatample a fiber optics fight 19 pickup; the rat least one illumination detector collects a~peetrum which is received by ZU the W least one spectrometer input; the sample in ti9is ernbo~liment is fratx~ the 21 chemical group of C.'~!, :~1H:, 4H or t:lae phySiLal cl~aractet~istics of Frn~ness, ~lertsity,' ?2 color vitld internal and e~te,rnal detects A.a~litiona,lly, the light source; 12(7 may 23 comprises a plurality of illumination fibers. :ltro this catbodiment a plurality of .#
?4 illumitaat~on fibers may be. an-ayed such tl7~.u ~;ach of tits plurality afillunaination fibers is equidistant ftutxa. adjacent illumination ftbe~; the at least ono illumination v', 2b detector may, in this embodiment, be positioned centrally itt floe affray of il.lun,ination 27 fibers. In an emhadime~it of this disclosure, tlae plurality at illumination rbers may, ,sS~M
28 for example, be comprised of 32 illumin~.xtiun ~ibet~ ~uctcl ~e light source 12U may be 29 provided, fur example, by a 5w tun,~sten halogen lamp~~~~t' otb.c~.' equivalent light 3U 5~

k.~, -.

.,, ~Q~'~~

,,.

_ . _ _ _ _ . _ . .... _.... ,.~., , , ,,~,~ ,~ ".,..,. r -,~ r ,m-a-v~-~T--i ..,.,.,.., ., m., . nr . nr -ira r~r _vW

w ~T'fUS 01 / 0 814 ~
~PEAfu~ ~ s BAR-zoo;
1 source or by a plurality of illumination sourcc;s prov'i~Ied fOr example by at least two 2 light sources such. as, for example, at least two 50 Watt light sources.
Illumination ., 3 sources may bt composed, far ~;xamplc, of sources h:rvi ~ a focusing ellipsoidal ',.;
4 reflector with confine fan. 1n this embodiment the at least one illurtzittation detector .niay comprise a plurality of light detectors 50, which mtty~far ~xarn~plc, be: arrayed 6 such that each illumination detector is equidistant f~rom~adjoining fight detecaors $0:
- . 7 where at least two light sources are positioned axe employed, they may for example 8 be positioned 45 degrees relative to the illumination detectors. in the array of 9 illumination .fibers. to an additional emhodimettt ofitfiis'''dlsclosure, a plurality oi~
light defectors 80 may be comprised of twenty-two i~l~lunriitriatiar~
detectors. ~n 11 embodin~~nt of the disclosure may be cotttpJ~,setl csFat least one Light souxce 1?0 .P
12 composed of a 5 w tungsten halogen lan7p; the at least o>ae Illunllriatx~ull detECtor is a 13 single detection fiber.; the light source 12U is positiCuaed 1 ~ainst the sample 30 degrees l 4 distal to the detection fiber, lf" the n,easumment of the s mple stt~rrac~
is tnad~ i.n. tt p..k l5 non-contacting manner, an alternative embodiuienl ray include a, polaniralion Cite,r ..
l6 between the light source: 120 and the sample, provided, for example by a linear 17 polarization fl Iter or au c~quivalenl as understood by oi~ of ordinary skill ixt the art; a 18 matching poluizadon f.Vter is positioned between tkte at''least one illumination _.., o.rt.
19 detector and the sample, which niay he provided, foa example by a.
li.aea.,r pohwitation ,;.:
filter rotated y0 d~g~es in relation to the polarization alter between the light sourc.~:
21 12U and the sample.
'.f.
2Z The method dtscribcd above, which uses wavelengths of'both visible 23 radiation (25t)-(i99 nm) speci0cally chosen to include the absorption band fdr yellow 24 color pigments (25U-499nm), red color pigments (500-500 ttrrt) a.nd green pi~nents or ebloraphyh (bt71-li!~9 am), as well as hll~, (700-1150~n) radiation to correlate 26 with Brix, firmness, pl-.l, acidity. density, i:ular and internal and external defects can 27 be carried out using a. variety of apparaW ses.
28 ADDf.'TIfONAL f)FTA1LLD DESCRLt'T'TON :~'.
.:
.Y.
2g Qverview Qf calibration o( visible~NIR seusor~~

>,..~

,~L
''~
. _ _ _ . _ . .,._. "r", ~ ..r-~.-. .,.nr,~ e~ nor, l °~,~,,-r .a-~-~nwr...~ , iar~ J, l ,-, . qT aGf aT _~',~I

ML'N
._ ~'oi~osl~6 ..~ ' 1P~,ArU~ ~ 6 II~R 2~C~
:x 1 Required cdlibra.tian wds addressed i,au the. Pa~nnt ~,.pplicatian t?9/524,3a9, in 2 paragraphs, idenkitled by pagclline by pn/ln, as fouows: 1/18; 3/17, 22, 2S;
4~'Z; 8/8;
3 9/4; y/14; 12/lfi; 16/8; 22/5; 31/21; 33119; 3!)11.t); 43I~:47/~; 52/13 etc.
Calibration ~:.
4 of spectroscopic maturity and quality sensors i_nvolvr~ btiitding algoxithxns that relate the visible arid near infrarrd spectrum of an individual fruit or vegetable to one or mare of the following: Brix (including, but not lirxutcd to ~u~r uontenk, or sweetness, ~ 7 or soluble; solids content); acidity (mcludin~; but not limitad to total acidity, or 8 sourness, or malie acid content or eitxic acid content or t~it~t~aric t~cid coatent); pH;
*~
.__ 9 Frmness (including but not limited to crispness or tt~.cdn'~s); intct~al di5orda~rs or ,r defects including but nut limited to watcrco.re, brovrrting;'cores rot, insect infestation.
! 1 Furthcn~~ore, the individual prapet~ty data collected abov~~can be combined as 12 follows; using the ratio of the sugar contempt to ~~tc;id cnnt~t to better t.'redict gating l3 qu~rlity, taste, swoet/sour ratio; using the co~.nbittcd ~tatal~fmm two or more of the 14 following; sugar content, acid content, pH, lirnuncss, cot,~t', c~ctcrnal and internal.
".
disorders to better predict eating quality.
16 nti v 1 rs '~ no c ce 17 no ' ni ~ ac i . ..o y w' ,. , si i o 1 !~ Sensing sample data including the prtrscnce yr abiacnce of a sample wss 2() adtlr~ased in the parent in paragraphs, identi~.cd by pagc~litte by px~/lxt, as follows:
~xk,".
21 2UI20; 36/8 etc. Using spectroscopic: sensors for wring Fruits amd vegetables 22 while in motion on a sample conveyor ~~S Sybtem in, sotrting and pacJ'ing warehouses 3Z is illustrated in i'ig. 10 ~.nd I~ig, l0A anc! is done r~ follows: TMc presenca err ahsenc;e 24 of a sample 30 and the pusiti.oatllocation of the sample 3U telative to the point of spectrum measuren,ertt is determined usinS, cane or mor~~ofthc folfo~ving means: 1) 2~ simple all position detern~inatioyx means a~7ti or sample iconveyor 2~5 position 2'7 determination means, provided for example by an ~nco'dct' or pulse generator 330. as 2~ seen in Fig. ~), integral to the sample conveyor 295 deed detecting stttnpl.e conveyor w.
9 29S movement, lyromdes one or n~nre, electronic or di,git.~tl signals to a ,;~..

!~':

_ _ ..,.,."-. , ,-,., .. . ,r,c a w,.v , -.,n.t yi_,-,mr..f ., ,nlW 1 f C1T 7fd OT _'.lNl.l ~r~us o s / 0 814 I
.~p~ ~, s ~R ~o~~
1 which initiates, by computer program control, control sig~ls to initiate and stop 2 acquisition of spectra, Z) the spectrum itself is autornxtieally it~speeted using 3 computer programs or programmed hardware, e,g., digital sigt'tal processors, is ' ~:r 4 determine if the sample 3(1 beinf; measured is at the optional location(s') for spectrum measurement, 3) a proximity sertsirtg means 340, including p~7ciiattity sensors oi; hut 6 vat limited to, magnetic, inductance, optical, mechaelica) sehsors; and al.s4 known as 7 object presence sensors, such as thru-beam or rcflcctancC sensors 341., i,s used to 8 provide informatioc~. shout the position, i.r;., urientativh o~ (ocetlon of l.he product o0 9 the packing or sorting lfnc relatme to the NllZ sensor, e,8., light detector 80, and/or size of the sample 3U, such proxtmity sensing means 340~tnd tl'tt~ir use being of 1 I. common knowledge to those practiced m the art ot~it~dustc~iaf ~rucessrng object 12 presence s~'naing. 'The proximity sensing means 34Q caaa a placed 1, 2, 3 or ...n t3 units of length, e.g., cups or pockets or conveyor halt leligkh, before the N1R sensor, 14 e.g.. detector 80, to indicate if 1, 2, or 3 or.., n mace emp~ s[~c:es, ~.~,., cups or ,.
t5 pockets or a defined and known length of conveyor belt;rarc present in sequence, thus .:ts, 16 al Ivwin g a ~,~reater amount of time far perfort~tting d~tk sp~tt~t ttt'td/or reference 17 spectra and/or standardlc;~libration sgmples. Usir~ eke of nrtw~ of the above 1.8 methods, the presence yr absence of samples) 3U is determined over ;~
defined length 19 of the particular sample conveyor 295 system. 1f surnple(s) 30 is present, multiple 2U visible and near-infrared spectra are uccluired as the sample 30 passes by the light ;,, 21 source 120 lamps) 123 providing light detector output 82 and spectremetcr(s,) 170 22 detector 200 input; such light collection n,;ty be achieved using a collimating lens 78 23 and or other light transmission means includi~ for example ether-optics to transfer -_ 24 the light that has interacted with the sample 3U to the specirotneter(s) 1.7t) detectors 25 2UU. It no sample 3t) is present, other reference measttfements ace ~madc to improve h'!.';
fi'M
2li stability and 3cot~r;my such as previously n~~:ntioned dark, spectra, referent;e spectra ,.;..
27 (lamp intensity anJ color ~~ulpul), null sttindardlcalibration st~.mpies, whici~ may be 2~t optical filters or polytrers or organrc material wikh known and re,~peatahie spectcwl ."fix,' 29 characteristics. Measurements that are made when no sample is pr~sunt include, bt~t S G ', A~IISHEET ~~.
_, :.
. . . _,..~..... . _ .,-, , .,.- .-., ,-"-. ~ .-.-,:-. v,r,~. a. nn-, min t ~.
r'wsawt' ~ ~ 4 ~w~.i . t 0c- ~ 4 T afa C T ._ ;~lH

~'~s 01/~0$14~
r~~~;~ ~. s ~aR zooz are, not limited to I) txteasurinQ a ref$rencE sp~cti~u~ ~i~ty vs.
wave.length) ofthe light source(s), 2) measuring the dark curr~;nt (rao Ji,gl~t ~ttditiw5ns) of one or more spectrometeys) 170 detectors) 200, including but not lbnitlad to the sample 4 spectrometers) 170 and the reference spe~troin~ter~~) 170, ~.h~d 3) standard or' calibration san~tples or filters 130 or material.
G
8 fer a or are wit a to tlae ,:.
9 product's ab nrbance Rpe ~ um.
Reforence to retorence, bas4littc and ds~tk ~fwaa auddressed io the parent - ~, 11 u~ paragraphs, identified by pagelline by pnl.lx~, xs Follows )2118; 39/10;
52/14 etc.
12 The rcfereaacc measurements to accout~.t for c~han,8e~r ion Iij~ht source antsnsity or color ~, t t 3 outpu(. can be obtameJ using a referonc;e light tx~nstmissioint rne;~ns~
320, e.g., a fiber-14 optic bundle which may be furcated, a ligln. pipe orcathnr~mean:;
ofkrwsmitti.tyg li~lit, t 5 with a common end 322 providing input to a refer~tpe ~trt~rta~otor 170, and, where 1 G furcated, one or more branched cads $1., ~:~oh oh which is"taa~ntcd by moans to allow n:
17 only light from the light source t20 lan7p(s) 123 to antler the reference light 18 transmission moans 320- A light shutter 300 is pha~d batw~ each light source 120 '.
19 lamp 123 and each reference light transmission rns 320. The at least one ligk~.t ~~:;.
shutter 300 can be opened :md closed separately by butter control means 305 ~..
21 including, for example, driven by a Iir~~ar ackuakor or rotary slrtlenoid ur other 22 ii~echanical or pneumatic device, or all at or7ce. ~ ~~ ' 23 Each light source 120 lamp 123 in floe systlatts ca~i ba measured separately to 24 determine if it is faulty or if it will so~.~n n~~d ropl~eme~t based on a stor,rd intensity vs. wRvCleh$t11 spectrum profile. ?he combined intcnsiihes from the refc;rence light 26 btansmission means 320 is used as the reli:r~t;t~a spoctxii~n for put'posr~s of calculatin8 27 an absorbance (or log 1./R) spectrum, which is linear wtconcentration (~.g., percent ~~ ~,k, 28 8rix or acidity or pounds of firmness, etc.'1, w ~.
~ .~ ~,~.fw ~~k x~
s9x;
. , . . ~...._.-.... . ~.. ".-~(-rr-.-. r r,ar~ W rw.~.lrr~ ~ 7.1T
~1T1G7aT°1 ~.1.11-I?J.J f-1C' :4T ?fl 4T-?~,,J

~T~ 01/ 08141 ~P~ ~. s ~R zoo rK
1 Closing all. of the light Shutters 33(~ of the r~.f~rettce light tra~nsrnission means 2 320 allow a dark current (no light condition) measilrenoteiat of khe spectrometer 17U
3 detector(sj 200. T"he dark current is Isrl;tly affected by temperature and must be 4 periodically measured and its intensity valor at each wavelength (or detector] pixel .,r~:
suhtract~:cl Croat the reference Spectrum obtained with the shutters 330 open.
~..~
6 'The sample spectrometer's I 7U detectoz 20Q dark'curtent must also be 7 periodically measured by closing light shutters 33U thstt are pls~:ed between the light 8 source and the saalple 30, or between the satllple f0 attd the san~pl~
spectrometer a:~, 9 light collection tiber, seen here as detector SO and detcctax output 82, or between the light collection .fi.her and the spectrometer 17U. ~isniloarly,to the reference ~r 1 l measurement, the dhrlc current of the sa~npl~; spectrc~Mtet~r 17U must Ue subtracted l2 .from the sample spectrum obtained with tho shutters 33U open. It will be appreciated 13 that reference measurement must be made wiih respmct to,tltE spectrometer 17U used 14 for light source 120 ]amp I 23 measurement as well as fiir~the spectrozt~tetets 170 used ,~.. , tv acquire detector 8U spectrum output 82 as processed iri the conxputer program I G controlled L Pl:l 172 in assaciat~on with algorithms for the ch~rtu:l~t~izatio~n of samples 17 30. ,. .
y o.
IS The reference measurement, utilizing 1 shutt~rme~tts, is demonstrated in Fig.
.E=..
19 9. Fig. 9 is art elevation depicting an additional ecxtbpdimettt orth~
invention den~onSirating at least one light dete~aor 8() having at least one output 82 ko at least 21 one spectrometer 17U having at least one detector 200. A,1 least onE
eollurrtinatiz~g 22 lens 7$ initrmcdiate the at lert5t ono light detector 813 attd a st~n'tple 30. The, at least .ll.
23 une light detector 80 positioned to ~Jele:ct libht front the sample 30. At least one light 24 source 1.2U lamp 123; a shielding marts inteuvediate the at least one light sou.re~ 120 lamp 123 and a sample 34 ounveyed by sarrapl~ convey t 295. At bast one aperture -rE .
?6 3 I U in the shielding tneans to show l I l~.m~inatiot~ of the satnplc 34 by the of Iet,st uw 27 light source 12i) lamp 123. T1 will be appreciated by those ef ordinary skill i~~ the 28 instntmEnt containment arts that an instrument ease or coutain~:r will b~ a oaeaos of 29 mounting the elements of thv disclosed invLntivn ia7 all its etnbc~din~et~tts. It wi.l.l be 5$
_. _ .,.-,._.,....-.-..-. ~er,mrr~~, .~,w,~ ~.i°'r-toy°~rwt~tirl\.~J :CC'~OT af.1 qT_~

s ~T~IS 01 / 0 814 b ~~~;~ ~ s ~aR 2ooz ~.::
1 appreciated that a case 25t1 n~.ay provide shi4lding a~~ud rn'ciundttg means for tb~~
2 invention. At least one !'t~lit interruption means lntetrloedi~tt~ the s~C
bast one lift 3 source 121) lamp 123 and the at least one ap~;rtura 310. ~:~ght interruption mecums re 4 provided, fo.r example, by light shutter 3ClCI m~;u~. T.hte~~st leuu;t one light shutter 3UU
,o operable by at Least one shutter control means 3015, ~.,~., linoar actuatoa~
or' rotary ,~ .
6 solenoid operates! by means, e,g., meehanical driven byaleCtriical, pneumatic, - 7 hydraulic or other power means or other shutter t~ne~a iitcludin8, for example liquid ayY
8 crystal screen operated by means, The at least Anne shutter control n~e~ns ,.
9 receiving control signets from at least one CPU !"I2 l~uv't~"t1g set !cast one shutter __.
operating contrr~l output 3U7, At Toast one reference ligkit transmitting rrreans 81 1 I. including, far exrarnple, Tiber-optics mcluctictg bifurcp~od~fib~er optics, receiving ::' 1.2 reference light output from the at toast one light s~ 1201s~aap l23. A.t feast am 13 rehorence light interruption means, con~pris~:d for exa~tpie of shuttEr 30'1, .c :"~' 14 iJttern »diake the rat least one light source 1 "2U lacr~,p X23 did the at toast one zeference light transmitting merans 81. The at least one rel"bte light shutter 301 operable by ~...
16 at least one shutter control means 3175. ~.g., linear fct~t~r or rtrtaty salenoic! up~rrc;ted _..
17 by means, e.g., tnechanical driven by electrical, pneut~tiG, hydraulic or ath~r power :, ~.
18 means or other shutter mcam including for exempla latqu~d crystal screen operated by ~,a i 9 means. 'The at least one reference light shutter 301 s~hut~"er control mexans 305 ~~,,,.
receiving Control signals from at least one C.~U 172 lutavutg al leu$t one shutter h~
21 operating control output 3tJ7. The at least one rsfi~nee,~li~bt trantsmitting moans 81 22 providing an input to the at least one spectrometer 170 detector 2U0, T(~c at feast one 23 CF'U 172 providing at least one lamp power output 125 to the ;~t least one light source 24 120 lamp 123. The at least o~~e spectrometer 1,70, raucciwing input firom at least one i..
.~.5 relercnce light transmittir~.g means 81 ha4 iry at toast ors output 82 received as in 2(i input to the at least one C:PU 172. 'the spectroxrtetcr a hut 82 capable of Al.D
2? conversion to form input to khe at Least o~,~ C:'PLJ 172. '~"he at !cast one sp~:etrometer >i:F%~
28 171), receiving input from at least onr: detector output 82~recoivcd as in input to the at Zq least one (:PU 172. The spectrometer output ~2 capubhe;of ~iL7 ~onv~,.rsion to farm x'' ,.., ,"j.a.
",re ,.
.,r~T
_ _._. . . . , , ._,".,~,_.~~, c,~,r ~~~nmnr~~9 mnT ~f:a'~awr'1 n.ln;~ a ~c :4T ~f1 ~7i-~IHL~

~"~S01I0814 1P~:AIUS 16 MAR 20 1 input to the at least one CPU 1.72. Mounting means t4 light sources 1.2t) lamps t 2B, 2 detectors 80, shutters 3t70, shutter control means 345, ref enae light transmitting 3 rrteans 81 and case 251. FncodeNpulse generator 330 input to CPU 172 providing 'f 4 sample conveyor 295 movement data. Computer pain to operate C~'U l7in t, drtta collection and control functions.
G A rclerence measurement of the light source l2UTlamp(sj 123 intensiay vs.
t..
7 wavelength output can also be obtained using reflecting moat 3Gt), as seen in Fib;.
t:..
8 1.1, including but n.ot limited to, for example. mirrors yr otltet reflecting or difiusin.g 9 materiai, including roughened aluminum, gold, Spectralon~l, Teflon, ground glass, steel_ Reflecting means 360 will be positioned to tIeet light source 120 lamp 11 tight to a detector $0 having an output 82 received try a tmmete:r 170 detector 12 200. A colluminating lens 78 may be positioned iritet'tnediato the detector 80 and the 13 light reflected by the reflecting moans 3C0. R~ilooti.t~g trieatts 3$0 may be positioned, 14 e.g., inserted via do aperture 310, for examplE where a case 250 is utilized, when a reference measurement is to be made as dictated by reflecting control means 3U8 as l6 an output from a CPU 172. The CPLi 172, via mews, ~1 detect the presence or 17 absence of. a sample 30 and, when a sample 3,0 is abssrit~For "n" time incr~n~.'nts or 1$ sample conveyor 2>5 muvernents wit I provide a reflecti~ control means 3()8 content 19 signal to reflecti.ug position means 306, e.g., linear actuator or rotary sc~lenuid operated by means, c.g., mechanical driven by eluctrical,~pneumatie, hydraulic or 21 other powEr means. The reflecting means 36U capable ofbeing withdrawn as dictated ~.
22 by ref>t~cting cvntrpl means 308 as an output front the CPU 172 when roference ,~.~.
23 measurement is to be ceased and spc~tra me,asureianent cifa sample 3U
resumed.
24 A light re .fleeting cc diffusing body .Foe obtaining the reference spectr~.~rn may also be obtained by mechanical insertion of reference means 430, as seen in Fig. i2 .y,.
2G and Fig, 13. in or near the location where actual samp1e~30 is t~.ormally measured, 27 which is between the light source 120 lamps) 123 artd ~ fr~er3co llgtat transmission ~x ..
.t y'i'_ 28 means 3Z0 leading to the sample specirumater 17Q detector 20U(s). insertion 3s by ;.,.
i 29 insertion means including but not linuited to am actuatot,systea~ 4()0 ea,pable, upon ,.
lit) a ~;
,.
. ., . .-,. ,-..-..-. ~ rnr. v.m~, m en~~ a w,~ r ~.,J~-~a7r~ ~ I.Ir~~t J rat.
° p T 7rt ~ T __ ~.~~

P'~" 01 / 081 ~ b.
~t~~ .~ 6 MAC 20fl~
1 receiving control signals or means as rerogniz~d by those af~rdinary skill innluding Y..
2 control signals or tn.eans provided from a ~'PlJ 17a, of op~eratian of an actuator 410 3 causing a piston 420 to extend 421 and retract 422 ~ sin ua Fit;. 12 and 13.
Power, 4 i~~cluding far cxarnple clrctrical, pneun~;rtic. hy~lraul~tc mid other me;;uys, is p~ravided S to operate the actuator by power transmission mans 44Q as will be appreciated by ..
6 those of ordinary skill.
S' 7 A CPU 172, control led by computer prograttt, is trot depicted in Fig. 1 ~1, I UA~, 8 11, 12 or 13 r~.s a parsnn of ordinary skill will appr~iax ~t~ch structure from vieuving 9 other drawings pt~esented herein. ''~
Achi v' hol r dut4 i rr a aslareltneet).
12 To improve the rrrea.surerrrent of the ~;ntire pttmlduct, twc~ of more light sources l 3 120 lamps 123 and/or detection 80 points arc used. xha ~roduot can be measured .A d, 14 rolling or not rolling with a rolling rneasurer~.~Lr~t felnerally irnprovirrg wh4le product 1 S measurement, while a non-rolling measurcn~,atat prpvid s bcttet accuracy and ',~-I6 introduces less spectral noise due to r,~ove:r~n~.~r:t.
17 i0.s a single fruit or vegetable sample 3U passe~B b~,,t~te point of spectruatZ
._ 18 acquisition, multiple spectra are acquired, ~:ach spect~rtt ropr~sentirr,g d diFterent I9 measurement location or arcs on the product.
r.
x aal-to- o' a r:ac ~'ze 21 p~odaat.
22 One or more n iestns may brr used to d~turmit~a the size ox<- weight of the 23 individual fruit or vegetable sample 3t). M~:ans xox dekerrxrinin,g product size includes, :,."
24 but is not limited to l ) a separately rlctcrrrtirred weight ox mass using sensors common .t~~
ZS t.o the industry, 2) utilizing the color sorter or. defect sorter data (e.g., from camera or 2G 4CD irnages), 3) utilising otbar size :~cnsnrs based otta~agn~ti~:, inductive, light:
'~!:
27 reflectance or multiple light bearzr, crartains, c~mnort to attrer ind>JStries. The relative ,r_, 28 sire of the sample ?~0 can then be, used to adjust the hacdw~l'~ shectr~tm acquisition f' ~
29 parameters or the amount of light (by varying the aperture 310 size) to provide an ax ..ci . ._ . . ..~... __ . ... , ~.. ..~,.. ~ .-rar, .,u~,~~~,w~~ ~ ~n~ y~waa~w~ ~
I,In:~ a f.~i.. ~dT ~(a C1T-->lHlrl ~'T~JS 01 / 0 81 ~
tPE~I'U~ ~ 6 MAR-20 ~r :r 1 in~provecl signal-to-noise ratio spectnnn for large samplds 30 and/or to prevent 2 dekector ~fl saturation by light for small product sample 30, e.g., detector SC? exposure 3 or integration time can be set for i0ng~~r time periods f4r large product samples 30 4 and for shorter time periods for small produe:l.
6 f m a in 1 ' ruduct attd emwin ~ oor ua it or you Ier" ect a. ",~ hen 7 calcufatine the absvrbs~nc~ guectrum frv~nlhe~~a coJlECted for dark.
-- >3 reference and sgmiale.
Each individual spectrum from the series of spectra a~cyuired for each individual product sample 3t1 are then inspected Lay x cOtnputer program or l l programmed hardware. Poor duality spectra are doleted'from this batch at'spectra 1.2 and the remaining spectra are used for constituent or pmpariy prediction.
The I 3 retained spectra of the product are combined with the appropriate reference and dark l4 curnrrtt tneasuren~ents to produce an absorbance spectrum as follows:
~;-A,bsorbanee Spectrum = -1og10 ~(sampte irttcnaity spectrum- sample darn 16 current speclrtun) l (refere~~ce intensity spectrum - reference dark cutn:eaat spectrum)]
r 7 i.e. the absorbanee spectrurt is equal to the n~;gative logarithm (base 1 U) of the ratio ...
18 of the dark current corrected sample sp~ctruat to the dptrkeurteat corrected refer~;ncu 19 spectrum.
x 2U All of the absorbalice spectra For each product sample 3U can then be , :y.
21 combined to produce a mean or average absorbmee spectrum of the product sample.
'e~
22 This average absorbance spech:a can then bmsed to compute the comp4yyent or 23 property of interest based on a previously stared CBlib'ration~ algorithm.
Alternatively, xr;.., Z~l each absorbance spectrum can be used individually wxtli'a previously stored calibration algorithm to cornpute multiple results of the~~ mponeatt or property of 2G interest for an individual product, followed by detEnnittation o~Cth~
average or m~;an 4.
27 component or property value computed by summing all of the vatuES and dividing the 28 resultant sum by the number of absorbance spectra used"-~fi 29 .-3.
3U G' rr~~_ ~E~~ ~~
,, fi .-, .~..,. . . . -.. n ,.,-,.-,.-.r, v .m :-~n,~~n~r r rrar S.~r~l.nln~v y/~T
~.~~"Ir~~T"1 ~ 4 Irk~:~ J Tr. ~0T 7fi ciT _ J

:- -~'~us01i08i4b .4~ ~ ~ u~~nus ~ ~. 6 MAR ~D~
i.
z 3 a r tnrr r hni' 4 Calibration is pertirmed as follows: 1 ) ~hr~ctra utpraduct sample 3Q are measured and absorbance spectra (corrr;cted for x~ferr~ttc~ atld dark cu,~rrextt) arv b stored, 2) Standard laboratory measurements (which arc often destructive) are madE
7 on the product sample 3U. Note: it is intportaat to rhea st~~ccss of the NIR
rr~ethad 8 that the portion of the sample 3t) that is interr08ated betr~een the light sourcc,(s) 120 9 lamps I 23 and light collectivn(s) detectors, e,g" ii,gh~t de~t~tors $U, leading to the :,.
spectrometer(~s) 170 detectors 200 i.s the same as lhat partiiatt measured by the 1 i standard laboratory technique. ~: , F~.
12 .por many sample conveyors 295 that are use ~"or~iirhr~c f~tit and vegetable ".~~;~
13 sorting and packing operations, the product cai~r be trrunspoiCted past the 14 measurement location rolling or nai rolling. It abae spectra arc collected ~rom I S the prp~lucl as it is roll ink, the exact location of ttny cue ttiicasurcu~ent (one spectrum) I 6 is not usually known, and therefore the entire. product (~ opposed to one localised 17 spot) must be analyzed kor the con~poncnt or property ofauuttorest, l.~calibratzori 18 algorithms are constructed in this way (using measiita car rolling product), all of 19 the rctalned sp~ltr'a fur that individual product are averaged to produce pan average absorbance spectrum and the total product compot~~t~t ox property is assigned to this x;.
21 vne absorbance spectrum. 'a;r 22 Because most fruits and vegetable are het~ato8en~ot~s and vary in component 23 lcvei with IIQCatlall, it is preferable to develop a oalil~rafi~ion model on product sample 24 30 that is not rolli.~~ so that each acquired spectrum is from a brawn physical location on the product sample 3t;1- Then, laboratory merasurcments a,re made on the.
2b same portion of product sample 3C) tl7at spectra were taken froraa. When this 27 procedure is used, a whole fruit ar vcget~tble~ satxtple 34wnn~y be:
separated, e,.g., cut or Z8 sliced, into smaller sub-portions prior tee laboratory analysis., Thesse smaller. sub-29 portions each correspond to Nflt data collected over the~rsamc locations within the ,r?a:
.i0 :y I~~1$~' . »,.. _. _..,.." . ,~,...,. , "..,> ~, ,"_,.,.". I .-,," r w.-,-,n~ rw . i.
tflvt I T L ~ OT 7f1 OT _'1~N1.1 ~Tf.~S01/081~
v- ~ t~~us ~ s t~,~R 2 1 product sample :311; the time period of NI:R data acquis~tiori watt be adjusted to sborter 2 or longer tunes, corresponding to the measurement of smaller or ltxrger product 3 samples 3U, respectively, fn this cast;, each sub-poctiott oFthc product sample 3D will 4 have unc ur more spectra FISboClated with that partict~lac location. The laboratory detcrtnined compot,tnt or property is then assigned to etch sps:4trum or spectra :from 6 that particular location.
7 Mathematical processing, i~ p~riormed~ oe absarbance s~ectra~rrior to - 8 conductioE statistical correlation analysis pad c~.~ratioa. Hno~d~el builidlng:
9 Absorbance spectra are pre..processed using a bia-aad smtooth function.
i~ ~'.
If? Parti:~l least squar~;s analysis (or variants rh~reufsu~c~h asptecewisc dii'evt 11 standardic,ation) are then used to relate the processed absorba~tce spectrum to the 12 assigned component and property values such as k3.rix, acidity, pH, firmness, color, 13 iutternal or external disorder severity and type, and asatiti~ ,quality.

1.5 calibration rnvdel.
16 To n,inimizc the number of calibration sarn~~lss that ace necessary, the 17 following method can be used: 1 ) spectra arc collected one ail test samples 30, 2j prior 18 to destructive laboratory measurements, principal components ~aa.aJysis (PCAj is 1 y perfornied on the absorbance spectra, 3) Resultant Soore~plots f~ron't PCA
(e.l;., Score 2U 1 vs. Score 2, Score 3 vs. Score 4, ate. j are then ga~erated, 4) A subset oJ: the original 21 samples (e.g., 4Q% of the original number of samples] ate selected from the Score 22 plots in either a random fashion or by selecting samples that, as a gxoup, yield a .x, 23 similar range, mean and standard deviation. of score values compared to the entire 24 group ol' original samples 30.
.;,s.
25 Calibration updates are periodically requixed to maintaiv .rueasur~ment 26 accuracy, particularly with agricultural product ss~,mples 3U that carp vary in 27 cofiiposition with growing conditions and vW ety. SeVe~al tnethc~cls can be used to ,.~:
2t3 minimize the et~orts of calibration upd:~tes. As Cruit oc' vegetable samples 30 are 29 analysed itt a packinb and svrtin6 warehouse, their visible/ttear infrared spectra can 3 (.) Ci4 NpEp StiLET
,~, .--,,:
_ ___--. _. _. ........-.nrm-.r. w,tn,m,~ mlhr ~I-_~-W5T"1~~Il'W.1J 71-.~OT
~fil t~T_?.I!

~y~ p~~y 01 / 0 81 ~4 6 ~ ~ .~P~:pru~ ~ s ~aR ~o o a 1 be examined by sotitwac-e to determine if thr~ sampla qualifies as a potential a~ 9.
2 calibration. update sample 30. Good calibra.tiou up~ata saxx'yles fit) will cover low to 3 high compon~;nt values mui will have ,'icore values l~at covrc'th~c same range as the 4 original sdn~ple's 30 Score values.
.,..
While a preferred en,bodin,ent of"the pr~c~nt disclosure has been shown and ~ described, il witl be apparent to tl°rose sk~lle~l in the art y~thi~tt rr~~ny elian~;es and 7 m~aditications m.ay be made without departing from ~osclosurc ia~ its broader 8 aspects. The appefi~itd claims are therefc~r~ i~~tendcd~ ~ tier all such .changes and ~~ inudifications a~ fgll within the true spiut ;a;nd scrp~tlie diiaulosure, 1 () II

12 , , ,.;~
~~
13 ;_ l. a ,~~"
1. 5 >.
16 ;;:

1$

".
,:
2t) ."N
2 l ,' 2~

?h z~
28 l.
-.,., (~~
d;;, _. , , . . ...".."., , .,.~ . , , ,. . , r-~ n rrar ~,ar~e..~9.un~ t'.~.~~' ~,1~'n07'twl ~ LIf'WI.I 7i.. ~. AT 7fl AT-:.IHLI

Claims (66)

CLAIMS:

1. A method of simultaneously determining multiple characteristics of produce samples comprising:
A. building algorithms of the relationship between sample characteristics and absorbed and scattered light from a produce sample having an interior;
B. illuminating the interior of the produce sample with a frequency spectrum;
C. detecting a spectrum of absorbed and scattered light from the sample;
and D. calculating the characteristics of the produce sample.

2. The method of claim 1 further comprising:
A. building the algorithms to generate a regression vector that relates the VIS (visible) and NIR. (near infra-red) spectra to sample characteristics selected from Brix, firmness, acidity, density, pH, color and external and internal defects and disorders;
B. storing the regression vector, in a CPU having a memory, as a prediction or classification calibration algorithm;
C. illuminating the produce sample interior with a spectrum of 250 to 1150nm;
D. detecting a spectrum of absorbed and scattered light from the produce sample interior, and inputting the detected spectrum to a spectrometer;
E. converting the detected spectrum from analog to digital and inputting the converted spectrum to the CPU; combining the spectrum detected;
F. comparing the combined spectrum with a stored calibration algorithm;
and G. predicting the characteristics of the produce sample.

3 The method of claim 1 wherein the characteristics are chemical characteristics selected from acidity, pH, sugar content, and molecules containing O-H, N-H and C-H chemical bonds.

4. The method of claim 1 wherein the characteristics are physical characteristics selected from firmness, density, color, appearance and internal and external defects and disorders.

5. The method of claim 1 wherein the characteristics are consumer characteristics.

6. The method of claim 1 further comprising:
A. sampling is of produce samples having molecules containing O-H, N-H and C-H chemical bonds;
B. illuminating of the interior of the produce sample is with a frequency spectrum including visible and near infrared light;
C. building algorithms for the correlation analysis separately of Brix, firmness, pH and acidity in relation to the light spectrum output from the illuminated produce sample; and D. detecting the spectrum of absorbed and scattered light from the produce sample with a light detector.

7. The method of claim 6 further comprising:
A. illuminating of the interior of the produce sample with a frequency spectrum of 250 to 1150 nm;
B. shielding the light detector from the illuminating spectrum;
C. measuring the spectrum for chlorophyl at around 680 nm; and D. correlating the characteristics of Brix, firmness, pH and acidity with the measured spectrum.

8. An apparatus performing the method of claim 1 comprising:
A. at least one illumination source for illuminating the produce sample, wherein the produce sample has a sample surface; an input mechanism of positioning the at least one illumination source proximal the sample surface;
B. at least one illumination detector; output mechanism of positioning the at least one illumination detector proximal the sample surface;
C. at least one mechanism of measuring the illumination detected from the sample; and D. a CPU receiving an output from the at least one mechanism of measuring the illumination detected and comparing the output with a stored calibration algorithm.

9. The apparatus of claim 8 wherein:
the at least one illumination source produces a spectrum within the wavelength range of 250 to 1150 nm;
the at least one mechanism of measuring the illumination is a spectrometer; the spectrometer has at least one spectrometer input;
the at least one illumination detector is a light pickup fiber; the at least one illumination detector collects a spectrum which is received by the at least one spectrometer input; the spectrometer has at least one spectrometer output channel;
the CPU having at least one CPU input; the at least one CPU input receiving the at least one spectrometer output; at least one computer program; the CPU is controlled by the at least one computer program; the CPU having at least one CPU output; the at least one computer program causing the at least one CPU
output to perform the steps of:
1) calculating an absorbance spectra for each at least one spectrometer output channel 1...n;
2) combining the absorbance spectra into a single spectrum encompassing the wavelength range;
3) mathematical preprocessing of the single spectrum;

4) predicting for each at least one spectrometer output channel, a sample characteristic for which the sample is examined by comparing the single spectrum with at least one stored calibration spectrum or at least one calibration algorithm;
5) determining sorting decisions based upon the sample characteristic;
6) inputting the sorting decisions to process controllers to control packing/sorting lines or to determine the time to harvest, time to remove from cold storage, and time to ship; and D. the sample includes molecules containing O-H, N-H and C-H chemical bonds.

10. The apparatus of claim 9 wherein the at least one spectrometer output is converted from analog to digital by at least one A/D converter which becomes, for each at least one spectrometer output channel, input to at least one CPU input; the at least one CPU output provided for each at least one spectrometer output.

11. The apparatus of claim 8 wherein:
the least one illumination source is a tungsten halogen lamp; the illumination is transmitted to the sample surface by fiber optics;
the at least one illumination detector is a fiber optics light pickup; and the at least one mechanism of measuring the illumination is a spectrometer.

12. The apparatus of claim 8 wherein the at least one illumination source is an illumination fiber.

13. The apparatus of claim 8 wherein the at least one illumination source comprises a plurality of illumination fibers; and the plurality of illumination fibers are arrayed such that each illumination fiber is equidistant from adjoining illumination fibers;

the at least one illumination detector is positioned centrally in the plurality of illumination fibers.

14. The apparatus of claim 13 wherein the plurality of illumination fibers are comprised of 32 illumination fibers.

15. The apparatus of claim 11 wherein the at least one illumination source is a watt tungsten halogen lamp.

16. The apparatus of claim 11 wherein the at least one illumination source comprises two 50 watt light sources; and the at least one illumination detector is comprised of a plurality of illumination detectors.

17. The apparatus of claim 16 wherein the plurality of illumination detectors are arrayed such that each illumination detector is equidistant from adjoining illumination detectors.

18. The apparatus of claim 16 wherein the plurality of illumination detectors comprise twenty-two illumination detectors.

19. The apparatus of claim 12 wherein:
the illumination source is comprised of an ellipsoidal reflector to direct illumination into the illumination fibers for transmission to the sample surface, and the illumination fiber and the at least one illumination detector are spring biased against the sample surface.

20. The apparatus of claim 11 wherein the at least one illumination source is a watt tungsten halogen lamp; the at least one illumination detector is an optical fiber fiber; the illumination source is positioned against the sample surface about 180 degrees relative to the at least one illumination detector.

21. The apparatus of claim 12 further comprising:
A. a polarization filter is positioned between the at least one illumination source and the sample; and B. a matching polarization filter is positioned between the at least one illumination detector and the sample.

22. The apparatus of claim 21 wherein the polarization filter is a linear polarization filter; the matching polarization filter is a linear polarization filter rotated 90 degrees in relation to the polarization filter.

23. An apparatus performing the method of claim 1 comprising:
A. at least one light source for illuminating the produce sample, wherein the produce sample has a sample surface; an input mechanism of positioning the at least one light source proximal the sample surface; at least one shutter intermediate the at least one light source and the sample; the at least one light source having a lamp output;
B. at least one light detector; output mechanism of positioning the at least one light detector proximal the sample surface; at least one collimating lens intermediate the at least one light detector and the sample surface; at least one mechanism of measuring the illumination detected from the sample surface; and C. at least one reference light detector directed to the lamp output; at least one shutter intermediate the at least one reference light detector and the at least one lamp output; at least one mechanism of measuring the illumination detected from the lamp output.

24. The method of claim 2 further comprising:

using the predicted characteristics of the sample in combination as follows:
using a ratio of the sugar content to acid content to better predict eating quality, taste, sweet/sour ratio; and using combined data from two or more of the following: sugar content, acid content, pH, firmness, color, external and internal disorders to better predict eating quality.

25. The method of claim 2 further comprising:
sensing sample data including sensing by sample presence sensing means the presence or absence of the sample conveyed on a sample conveyor while in motion; sensing by sample position sensing means the position/location of the sample relative to a point of spectrum measurement; the presence sensing means and the position sensing means having outputs to a computer program controlled CPU; the computer program controlled CPU determining if the sample being measured is at an optimal location for spectrum measurement and determining if the sample is present.

26. The method of claim 25 wherein the presence sensing means is a proximity sensing means.

27. The method of claim 26 wherein the position sensing means is an encoder or pulse generator that detects sample conveyor movement and provides one or more electronic or digital signals to the CPU which initiates, by computer program control, control signals to initiate and stop acquisition of spectra.

28. The method of claim 27 further comprising:
determining by the computer program controlled CPU timing for performing a reference testing of a light source lamp and spectrometer.

29. The method of claim 28 wherein the reference includes measurement of dark spectra or reference spectra or standard/calibration samples.

30. The method of claim 29 wherein the spectrum of absorbed and scattered light is light achieved using a collimating lens or a fiber optic light transmission means.

31. The apparatus of claim 8 further comprising:
a sample presence sensing means for sensing of the presence or absence of a the sample conveyed on a sample conveyor while in motion and a sample position sensing means of the position/location of the sample relative to the point of spectrum measurement; the presence sensing means and the position sensing means having outputs to a computer program controlled CPU; the computer program controlled CPU determining if the sample being measured is at a desired location for spectrum measurement and determining if the sample is present.

32. The apparatus of claim 31 wherein the presence sensing means is a proximity sensor.

33. The apparatus of claim 32 wherein the position sensing means is an encoder or pulse generator that detects sample conveyor movement and provides one or more electronic or digital signals to a CPU which initiates, by computer program control, control signals to initiate and stop acquisition of spectra.

34. The apparatus of claim 33 wherein the computer program controlled CPU
determines timing for- performing a reference testing of a light source lamp and spectrometer.

35. The apparatus of claim 34 wherein the reference testing includes measurement of dark spectra or reference spectra or standard/calibration samples.

36. The apparatus of claim 35 wherein the spectrum of absorbed and scattered light is achieved using a collimating lens or a fiber optic light transmission means.

37. The method of claim 2 further comprising:
measuring by reference measurement changes in at least one light source lamp intensity or color output, a reference spectrometer output and output of spectrometer receiving sample spectra input from detectors; transmitting light from one or more light source lamps to a reference spectrometer detector using one or more reference light transmission means.

38. The method of claim 37 wherein the reference light transmission means comprises fiber-optics.

39. The method of claim 37 wherein the reference light transmission means comprises a light pipe.

40. The method of claim 37 further comprising:
positioning the reference light transmission means at the at least one light source lamp to allow only light from the at least one light source lamp to enter the reference light transmission means.

41. The method of claim 40 further comprising:
placing at least one light shutter intermediate each light source lamp and each reference light transmission means; and opening and closing the at least one light shutter by shutter control means.

42. The method of claim 37 further comprising:
measuring, by the reference spectrometer, each at least one light source lamp separately; inputting the reference spectrometer output to the CPU, wherein the CPU is computer controlled; storing in the CPU a measurement of intensity vs. wavelength spectrum profile for each at least one light source lamp;

comparing the stored intensity vs. wavelength spectrum with the reference spectrometer output; and determining from the comparison the condition of the at least one light source lamp.

43. The method of claim 2 further comprising:
using the detected spectrum as a reference spectrum, for purposes of calculating an absorbance or log 1/R spectrum, which is linear with concentration.

44. The method of claim 41 further comprising:
closing all of the light shutters of the reference light transmission means;
allowing a dark current measurement of the spectrometer detector; measuring the dark current and its intensity value at each wavelength; subtracting the measured dark current from a reference spectrum obtained with the shutters open.

45. The method of claim 37 further comprising:
measuring the reference spectrometer output and a sample spectrometer output dark current; inputting the reference spectrometer output and the sample spectrometer output to the CPU; subtracting the output measured from the reference spectrometer; and subtracting the output measured from the sample spectrometer.

46. The apparatus of claim 8 further comprising:
at least one collimating lens intermediate the at least one illumination detector and the sample; at least one illumination source lamp; an illumination shielding means intermediate the at least one illumination source lamp and the sample; at least one aperture in the illumination shielding means to allow illumination of the sample by the at least one illumination source lamp; at least one illumination interruption means intermediate the at least one illumination source lamp and the at least one aperture; the at least one illumination interruption means operable by at least one illumination interruption control means; the at least one illumination interruption control means receiving control signals from at least one CPU having at least one illumination interruption operating control output; at least one reference illumination transmitting means receiving reference illumination output from the at least one illumination source lamp; at least one reference illumination interruption means intermediate the at least one illumination source lamp and the at least one reference illumination transmitting means; the at least one reference illumination interruption means operable by at least one reference-illumination interruption means control means; the at least one reference illumination interruption means control means receiving control signals from the at least one CPU having at least one reference illumination interruption operating control output; the at least one reference illumination transmitting means providing an input to the at least one spectrometer detector; the at least one CPU providing at least one lamp power output to the at least one illumination source lamp; the at least one spectrometer, receiving input from the at least one reference illumination transmitting means having at least one output received as in input to the at least one CPU; the spectrometer output capable of A/D
conversion to form input to the at least one CPU; the at least one spectrometer, receiving input from at least one detector output received as in input to the at least one CPU; the spectrometer output capable of A/D conversion to form input to the at least one CPU; mounting means to mount the at least one illumination sources lamp, the at least one illumination detector, the at least one illumination interruption means, the reference light transmitting means; encoder/pulse generator input to the CPU providing sample conveyor movement data; computer program to operate the CPU in data collection and control functions.

47. The method of 37 further comprising:
measuring, as a reference measurement, output values of intensity vs.
wavelength from one or more light source lamps using reflecting means;
positioning the reflecting means to reflect light from the one or more light source lamps to a light detector having a light detector output which is received by a spectrometer detector.

48. The method of 47 further comprising:
positioning the reflecting means to a position to reflect light from the one or more light source lamps to the light detector as dictated by reflecting control means, as an output from the CPU; detecting the presence or absence of the sample and, when the reference measurement is to be made, inserting the reflecting means as dictated by the reflecting control means; and withdrawing the reflecting means by the reflecting control means.

49. The apparatus of claim 8 further comprising:
reflecting means, which may be positioned to reflect light from illumination source lamps to the illumination detector by reflecting control means, as an output from the CPU.

50. The apparatus of claim 8 further comprising:
a light reflecting or diffusing body for obtaining a reference spectrum which is obtained by mechanical insertion of the light reflecting or diffusing body in or near the location where the sample is normally measured.

51. The method of claim 2 further comprising:
illuminating, with at least one light source lamp, the sample interior while the sample is rolling or revolving.

52. The method of claim 2 further comprising:
illuminating, with at least one light source lamp, the sample interior while the sample is not rolling or revolving, where a non-rolling measurement provides better accuracy and introduces less spectral noise due to movement.

53. The method of claim 2 further comprising:

obtaining, as the sample passes by a point of spectrum acquisition, multiple spectra, where each spectrum represents a different measurement location or area on the sample.

54. The method of claim 2 further comprising:
optimizing signal-to-noise and accuracy by 1) determining the size or weight of the sample by weight or mass sensors; 2) utilizing a color sorter or defect sorter to provide data; and 3) utilizing other size sensors based on magnetic, inductive, light reflectance or multiple light beam curtains

55. The method of claim 54 further comprising:
adjusting, in accordance with the size of the sample, a spectrum acquisition parameter, to provide an improved signal-to-noise ratio spectrum for larger samples and to prevent detector saturation by light for smaller product sample.

56. The method of claim 2 further comprising:
improving accuracy by inspection of multiple individual spectra collected from the sample; removing "outlier" spectra; and calculating an absorbance spectrum from raw data collected for a dark, reference and the sample spectra.

57. The method of claim 56 further comprising:
combining all of the absorbance spectra for each product sample to produce a mean or average absorbance spectrum of the product sample; and using this average absorbance spectra to compute the sample characteristic of interest based on the previously stored calibration algorithm.

58. The method of claim 56 further comprising:
using each absorbance spectrum individually with the previously stored calibration algorithm to compute multiple results of the sample characteristic of interest for an individual product sample; and determining the average or mean component, characteristic or property of interest.

59. The method of claim 2 further comprising:
transporting the sample, by a sample conveyor, to a NIR measurement location including to a the light detector, wherein the sample is rolled to permit the entire sample to be analyzed for the characteristic of interest; averaging all of the spectra for the sample to produce an average absorbance spectrum and a total sample characteristic assigned to the absorbance spectrum.

60. The method of claim 2 further comprising:
transporting the sample, by a sample conveyor, to a NIR measurement location including to a light detector wherein the sample is not rolled by the conveyor; identifying a portion of the sample where the spectrum was detected;
performing laboratory measurements on the portion of the sample to determine sample characteristics of interest to each spectrum from the measurement location.

61. The method of claim 2 further comprising:
performing mathematical processing on the spectrum of absorbed and scattered light prior to conducting statistical correlation analysis and calibration model building; pre-processing the spectrum using a bin and smooth function;
relating by statistical analysis the processed spectrum to the sample characteristics.

62. The method of claim 2 further comprising:
minimizing the number of samples needed to develop a calibration model;
collecting absorbance spectra on a plurality of samples; performing, prior to destructive laboratory measurements, principal components analysis (PCA) on the absorbance spectra; generating Resultant Score plots from the PCA; selecting a subset of the original samples that yield a range, mean and standard deviation of score values comparable to the entire group of original samples.

63. The method of claim 62 further comprising:
periodically performing calibration updates to maintain measurement accuracy.

64. The apparatus of claim of claim 9, wherein the mathematical preprocessing is selected from smoothing the spectrum, box car smoothing the spectrum, and calculating derivatives of the spectrum.

65. The apparatus of claim of claim 9, wherein the sample characteristic is selected from Brix, firmness, acidity, density, pH, color, external defects, internal defects, and disorders.

66. The method of claim 61 wherein the sample characteristics are selected from Brix, acidity, pH, firmness, color, internal defects, external defects, external disorder severity and type, and eating quality.

80.