Relationship between fermentation behaviour, measured with a 3D vision Structured Light technique, and the internal structure of bread.

Transcription

1 Reltionship between fermenttion behviour, mesured with 3D vision Structured Light technique, nd the internl structure of bred. Smuel Verdú 1*, Eugenio Ivorr 2, Antonio J. Sánchez 2, Jose M. Brt 1, Rúl Gru 1 1 Deprtmento de Tecnologí de Alimentos. Universidd Politècnic de Vlènci, Spin. 2 Deprtmento de Ingenierí de Sistems y Automátic, Universidd Politècnic de Vlènci, Spin *Author for correspondence: Smuel Verdú Address: Edificio 8G - Acceso F - Plnt0 Ciudd Politécnic de l Innovción Universidd Politécnic de Vlenci Cmino de Ver, s/n VALENCIA SPAIN E-mil: sverm@upvnet.upv.es Phone :

2 Abstrct The bred-mking process is set of opertions where could be relevnt to use monitoring methods. Specificlly, the fermenttion phse is crucil step in which the qulity of the product cn be ffected. Severl methods hve been developed to monitor this stge bsed on different technologies. The im of this study ws to nlyze nd obtin informtion bout the internl structure of bred dough during the fermenttion process using 3D vision system bsed on Structured Light (SL). The differences bout fermenttion behvior of two whet flours clssified s high strength flour by compny providing were studied. The prmeters of the internl structure of the bked product (finl bubble size nd their popultion density) were nlyzed with 2D imge segmenttion. An importnt correltion (r=0.865) ws observed between the 3D nd 2D informtion, specificlly between the trnsversl re nd height (3D), nd finl bubble size nd number of bubbles (2D). Although t the end of the dough fermenttion process (Tf) the re (A) nd mximum height (H) were different, the reltionship between both prmeters were similr, reching similr bubble size s consequence of colescence phenomen, independent of the bubble growth rte. This could be bse for the development of prediction models nd devices to monitor the fermenttion phse of the bredmking process. Keywords: Structured light, bubble size, internl structure, monitoring, fermenttion, bred dough, behvior. 2

3 1. Introduction The preservtion nd improvement of product qulity nd properties is of utmost importnce in the food industry. This must be s constnt s possible to mintin competitiveness s well s to stisfy the expecttions of consumers (Mirlbes, 2004). Therefore, obtining detiled informtion bout the fctors which influence the different phses in their processes is one of the min issues in the food industry. The present study is focused on this context, specificlly the industril bred-mking process. Severl fctors ffect productivity in the bred industry due to the modifictions of whet flour properties nd hence their behvior during processing (Bdj & Serš., 2011; Wng et l. 2011). In prticulr, chemicl composition nd rheologicl properties my seriously ffect both the dynmics of the process nd the finl homogeneity of products (Brk et l. 2013; Cocchi et l. 2005). Process vribles (time, temperture, humidity, proportions of ingredients, etc) nd relevnt qulity ttributes (texture, pltbility, rom profile) could be ffected by these modifictions ( Le-Bil et l. 2009; Novotni et l. 2011). Specificlly, one of the most influentil phses, dough fermenttion, is n esily lterble phse due to smll chnges in the chrcteristics of the rw mterils, which thereby brings out significnt modifictions in the usul development of production. During fermenttion, the gs produced by yest ctivity expnds the ir bubbles previously incorported into the dough system in the mixing phse (Ktenioudki et l. 2009), therefore nything tht modifies this phse could lter the finl ttributes of the product. These ltertions occur becuse the vribility in the gs phse distribution in the dough system plys crucil role in crumb structure formtion, since the stbility nd the growth of gs bubbles generted will determine the finl volume of the lof s 3

4 well s the texture of the bked product (He nd Hoseney, 1991). Thus, vribles such s the volume nd density of the dough, to control the fermenttion process, s well s their reltionship with the properties of the gs phse hve been widely studied. To improve the knowledge bout the fermenttion phse, mny studies hve been crried out using different points of view nd techniques. Among them, there re sttic controls of vribles like volume, density nd bubble size (Perez-Nieto et l. 2010; Updhyy et l. 2012). There re lso studies ddressing the development of pplictions to dynmiclly monitor the sme vribles (Flcone et l. 2005; Zúñig et l. 2009; Lucs et l. 2010) bsed on ultrsound, MRI, 2D nd 3D imging nlysis. Although 2D imge nlysis, in which segmenttion imges hve been pplied, is usully employed s sttic control for checking the bubble size or bred crumb grin (Lssoued et l. 2007, Gonzles-Brron & Butler, 2008, Scnlon & Zghl, 2001, Pérez- Nieto et l. 2010), in order to be relted to different recipes or processing, minly bking. There re vrious techniques used to obtin 3D imging, one of them bsed on structured light (Verdú et l. 2013). It is bsed on the projection of pttern of light on smple nd the clcultion of 3D dimensions from the deformtion of the pttern using cmer (Verdú et l. 2013). This technique permits the monitoring of continuous processes nd could be pplied on-line. In previous study (Ivorr et l. 2013), ten whet flours, without physicochemicl nd rheologicl differences, were monitored nd nlyzed during their fermenttion evolution, employing the Structured Light method. Results showed differences in their fermenttion behvior (peks nd vlleys tht tke plce during fermenttion, when the vrition of the totl trnsversl re is relted to the mximum height) which were relted with the fermenttion cpcity. 4

5 Thus, the objective of this work is to focus on tht study, relting the fermenttion behvior, mesured with 3D vision Structured Light technique, to the evolution of the internl structure of bred, mesured with 2D imge nlysis. 2. Mteril & methods 2.1. Physicochemicl chrcteriztion of flours. A bttery of physicochemicl nlyses ws crried out to obtin informtion bout the generl chrcteristics of the smples. Ech nlysis ws relised ccording to the stndrd methods of the Interntionl Assocition for Cerel Science nd Technology (ICC). The nlyses performed were: moisture (ICC stndrd No.110/1), percentge of gluten (ICC stndrd No.106/2), flling number (ICC stndrd No.107/1, FN 1500, Perten, Sweden) nd rheologicl prmeters (ICC stndrd No.121, Alveogrph, Chopin Technologies). All nlyses were crried out in triplicte. Tble 1 lists the verge nd stndrd devition of the evluted prmeters. 5

6 Tble 1.Vlues nd stndrd devition of lveogrph prmeters (P=mximum pressure (mm), L=extensibility (mm); W=strength (J -4 ), moisture, dry-gluten, nd flling number of the two different whet flours employed. Different letters in rows men significnt differences t p Prmeter F1 F2 P 98 ± 1 97 ± 1 L 106 ± ± 1 W 378 ± ± 4 P/L 0.92 ± ± 0.01 %Moisture 15 ± ± 0.1 Dry Gluten (g/100 g) 12.9 ± ± 0.4 Flling number 410 ± ± Dough preprtion nd fermenttion process The whet flours employed were obtined from two different btches produced by Molí del Picó-Hrins Segur S.L (Vlenci-Spin). Both btches, without physicochemicl nd rheologicl differences (Tble 1), were selected from the previous study (Ivorr et l. 2013).One hd the lowest fermenttion cpcity (F1) nd the other the mximum (F2). In ddition third btch (Fm), prepred mixing F1 nd F2 (50%) ws lso used. The ingredients nd their percentges for the doughs were: 56% whet flour, 35% wter, 2% refined sunflower oil (mximum cidity 0.2º. Koipesol Semills S.L - Spin), 2% commercil pressed yest (Scchromyces cerevisie. Lesfre Ibéric S.A - Spin), 4% white sugr ( 99.8 % scchrose. Azucrer Ebro, S.L Spin) nd 1.5% NCl (refined mrine slt 97 % NCl. Sliner Espñol, S.A Spin). The three doughs were mde using the sme procedure. 6

7 The doughs were mde by combining ll the ingredients in food mixer (Thermomix TM31, Vorwerk, Germny) ccording to the following procedure. At the first step, the liquid components (wter nd oil), sugr nd NCl were mixed for 4 minutes t 37 C. Then, the pressed yest ws dded nd mixed t the sme temperture for 30 seconds. Finlly, the flour ws dded nd mixed with the rest of the ingredients using specific defult progrm for dough mixing. At this step, the device mixes the ingredients with rndom turns in both directions of the mixer helix (550 revolutions/minute), in order to obtin n homogeneous dough. Then, 450 g of the dough ws plced in metl mold (8x8x30cm) for its fermenttion. This process ws crried out in chmber with controlled humidity nd temperture (KBF720, Binder, Tuttlingen, Germny), where 3D imging Structured Light (SL) device ws developed nd clibrted. The conditions of the fermenttion process were 37 C nd 90 % Reltive Humidity (RH). The smples were fermented until the dough lost its stbility nd size (Tf), specificlly when growth depletion occurred. 4 replictes were crried out for ech dough. 2.3.Fermenttion monitoring by Structured Light method (SL) The objective of the 3D vision system is to obtin the 3D smple profile during fermenttion. In order to ccomplish this objective 3D vision system ws developed specificlly to monitor fermenttion. This vision system ws formed of red linel lser (Lsiris SNF 410, Coherent Inc. Snt Clr, Cliforni (USA)) nd network grycmer (In-Sight 5100, Cognex, Boston, Msschusetts (USA)). Both of them were instlled inside the fermenttion chmber (Figure. 1). This ws possible becuse the cmer hs n index protection of 67 (IP67) nd the lser is robust enough to work sfely in these conditions. 7

8 The 3D visul system developed hs resolution of m nd m for the X nd Z xes respectively. This resolution in derived from lser ngle β of 0.65 rdins (Fig 1) in combintion with the resolution of the cmer (640x480) nd its distnce from the smple. The working rnge chieved with this resolution is 0.1 m in the X xis nd 0.08m in the Z xis. Figure 1. Clibrtion of 3D vision system instlled in the fermenttion chmber. Clibrtion pieces cpture with illumintion (top right) nd lser response under processing conditions (bottom right). Although the cmer cn work t up to 60 fps, the cquisition rte ws 1 fps due to the long period of time tht fermenttion requires (round 2 hours). Clibrtion of the equipment ws firstly performed by tking 10 regulrly distributed points in the lser projection plne with known coordintes (Trobin et l. 1995) nd then using these 3D 8

9 points nd their correspondent points in the imge to clculte n homogrphy trnsformtion (Zhng et l ). 2.4.SL method imge processing In order to obtin the 3D profile of the smple, the first step is the segmenttion of the lser points cptured by the cmer. This segmenttion ws performed s follows: using Otsu s globl threshold (Otsu et l. 1979) the lser pixels were selected. Then, these pixels were filtered removing non-connected pixels with n re lower thn 100px. Finlly, exct row coordintes were clculted by weight men for ech column using the intensity vlue. Following this method, subpixel precision ws chieved. The second step is the trnsformtion from imge coordintes to 3D locl coordinte system. This ws done using the homogrphy trnsformtion clculted in the clibrtion step. The lst step is to chnge the locl coordinte system using rottion mtrix which mkes the z xis norml to the surfce s cn be seen in the world coordinte system of Fig. 1. The 3D smple profile is 3D curve composed of the 3D points which re between the known 3D points from the edges of the mold. The following informtion ws extrcted from ech imge in order to nlyze the growth of the smples during fermenttion: Mximum height (H): The mximum Z vlue for the smple nd its position. Trnsversl re (A): The integrtion of the Z vlues long the X direction of the smple. 9

10 Acquisition nd dt processing were crried out using own code developed in the Mtlb computtionl environment (The Mthworks, Ntick, Msschussets, USA). 2.5.Smpling procedure to 2D imge cquisition. The internl structure of the doughs ws studied with 2D imge segmenttion. The im ws to obtin informtion bout the finl bubble size (Bz) nd its popultion density (Dp) t different times during the fermenttion process fter they were bked. Smple times (T) were selected bsed on the finl fermenttion time of different doughs (Tf). The first point (T1) ws smpled t 50 minutes, s this ws round ½ of the shortest Tf for the doughs (F1), in order to obtin informtion in the process erly. The rest of the points were smpled t lmost Tf, just before the depletion of ech dough. Therefore, points T2, T3 nd T4, were smpled t 100, 150 nd 180 minutes for F1, Fm nd F2 respectively (Figure 4). Finlly, different numbers of smples were obtined in function of the dough Tf. Hence, the number of smpling times for ech dough ws 2, 3 nd 4 for F1, Fm nd F2 respectively. Smpling ws relized by stopping the fermenttion nd bking the doughs t ech time t 180ºC/50 minutes. Ech test ws crried out in triplicte nd 6 slices of 1 cm thickness were cut from the centrl zone of the breds D imge cquisition nd segmenttion. Both sides of ech slice of bred were cptured with scnner (Aficio MP C300- Ricoh, Tokyo, Jpn) to be nlyzed through imge segmenttion, so for ech flour nd time smple point informtion ws extrcted from 36 imges. The imges were cquired with resolution of 300 dpi (Figure 2 A). The use of the scnner directionl nd homogenous light is very suitble for bubble segmenttion. In ddition, blck 10

11 bckground ws used in order to enhnce the mesurement of the pore cell wll structure nd porosity of the bred. (A) (B) Figure 2. (A) Gry imge (drk res represent bubbles nd light res represent structure); (B) segmented imge (blck pixels represent bubbles nd white pixels re structure). After trying severl lgorithms, bnd thresholding method followed by growing process ws finlly selected to segment the bubbles nd structure pixels (Figure 2 B). The bnd thresholding method ws used to introduce flexibility to the globl threshold selection decision. Gry pixels, lower thn first threshold (Th1), re clssified s bubbles nd pixels, higher thn second threshold (Th2), re clssified s structure (Figure 3). These two thresholds were selected in order to obtin high confidence for 11

12 the first clssifiction process. After this initil clssifiction, pixels between these two thresholds were clssed s undetermined t this first step Th 1 Th 2 Pixels Bubble Structure Gry level Figure 3.Thresholds for clssifying undetermined pixels. In order to clssify the undetermined pixels second growing step ws performed. This second step consists of clssifying undetermined pixels tking into ccount the previous clssified neighbours in growing process. This technique ws previously used for detecting weed ptches in cerel crops (Benlloch et l. 1995, Benlloch et l. 1996, Benlloch et l. 1996b). Imge processing ws crried out using own code developed in the Mtlb computtionl environment (The Mthworks, Ntick, Msschussets, USA) 3. Results & discussion 3.1 SL method results 12

13 The evolution of dough volume during fermenttion ws chrcterized from dt obtined employing the structured light method. The dtset ws expressed s the trnsversl re (A), the mximum height (H) nd the rtio between both prmeters Q t ech time. Figure 4 shows the evolution of the trnsversl re (A) ginst the time of the fermenttion process for F1, Fm nd F2 until Tf, s this is the most representtive prmeter for the evolution of the process. Figure 4. Evolution of trnsversl re (A) of F1 ( ), F2 ( ) nd Fm (---) during the fermenttion process nd smpling point times (T) to nlyse imge segmenttion. The results showed different vlues of Tf, A nd H for the doughs. F2 ttined the highest vlues (180 minutes, 17.2 m nd 7.7 m 10-2 respectively). F1 presented the lowest vlues (100 minutes, 11.7 m nd 5.1 m 10-2 respectively) nd Fm reched 13

14 intermedite vlues (Tble 2). It is interesting to note tht behvior ws similr until round 100 minutes, when sttisticl differences between the A nd H of the doughs were not found, then F1 depleted, nd Fm strted to chnge its growth rte. Fm incresed in size till round 150 minutes, then reduced its growth rte nd ttined intermedite vlues of A nd H between F1 nd F2 t its Tf. The reduction in Fm s growth rte could be due to fetures contributed by F1 nd F2. Fm s behvior ws quite logicl s expected. This helped to confirm tht informtion bout the behvior of F1 nd F2 obtined by SL ws relible. Tble 2. SL nd 2D segmenttion imge nlysis results. Different letters within columns men significnt differences t p Time Smple Are (m ) (A) Height (m.10-2 ) (H) Q(m) Medium Bubble Size (m ) (Bz) Density popultion (bubbles / m ) (Dp) F1 6.5 ± ± ± 0.1 b 2.7 ± 0.1 b 15.0 ± 0.8 b T1 (50 min) Fm 5.8 ± ± ± ± ± 1.2 F2 5.8 ± ± ± ± ± 0.4 F ± ± ± ± 0.3 c 12.3 ± 0.7 b T2 (100 min) Fm 11.3 ± ± 0.4 b 2.23 ± ± 0 b 14.0 ± 1.0 F ± ± 0.2 b 2.29 ± ± ± 0.4 T3 (150 min) T4 (180 min) Fm 14.1 ± 0.8 b 6.3 ± ± ± 0.2 b 9.3 ± 0 F ± ± 0.1 b 2.24 ± ± ± 1.1 F ± ± ± ± ± 0.2 b 14

15 3.2 2D imge segmenttion results Informtion bout verge finl bubble size (Bz) nd the popultion density (Dp) of smples ws obtined through imge segmenttion. Tble 2 shows the results of different smpling times (T1, T2, T3 nd T4) for ech flour. Bz incresed during fermenttion, while Dp presented inverse behvior (Figure.5). These behviors greed with studies relized previously by the uthors, which showed tht the rte of bubble size gin hd gret influence on the structure of the dough during the fermenttion process (Updhyy et l. 2012; Prud homme nd Khn, 1996; Wilde, 2003; Autio nd Lurikinen, 1997). The results showed n increse of Bz, but with different rtes between flours. At T1 nd T2, F1 presented greter Bz thn Fm nd F2. Furthermore t T2, F1 presented similr differences. These results explined why F1 presented the lowest Tf, s bubble size t T2 ws too high to resist dough structure so it reched colescence nd consequently overll depletion. Figure 5b shows the bubble sizes of ll smples t T2, evidencing differences in the internl structure of the doughs bsed on significnt differences between Bz nd Dp, while differences between A were not found. Moreover, the sme behvior of Fm nd F2 in comprison with F1 t T2 were observed t T3 nd T4. Figure 5c shows tht just before the Tf of ech dough, Bz nd Dp presented the sme tendency to rech similr vlues t the end. So, lthough A nd H were different t Tf for the doughs, the internl structure bsed on bubble number nd size ws not different for ll smples before colescence. 15

16 Figure 5. Correltion between Popultion Density of bubbles (Dp) vs. Averge bubble size (Bz) for (A) t ll smpling times, (B) T2, nd (C) Tf of ech dough. (Series of B nd C sections re: F1 ( );Fm ( ) nd F2 ( )). 16

17 3.2 Joined nlysis of SL nd 2D imge segmenttion results The dt obtined by both imge nlysis techniques were compred nd nlyzed jointly. The objective ws to see whether the SL results greed with the imge segmenttion nlysis. The evolution of the dough during fermenttion ws registered by SL, which provided informtion bout dough growth kinetics, volume nd shpe t ech second. SL prmeters (A nd H) were used in order to further nlyze of the dough shpe, giving rtio between both Q (A/H). This prmeter linked the increment of height nd trnsversl re. It ws used s it is possible to find smples with nondifferent totl volumes but different heights in function of flour properties s well s rheologicl fetures. The first study ws to relte verge finl bubble size (Bz) to the trnsversl re of the doughs (A) t ech smpling point (T). Figure 6 shows n exmple of the reltionship presented between A nd Bz t ech T. 17

18 4,0 3,5 T2 T3 Tf T4 Bz (m ) 3,0 2,5 2,0 1, A (m ) Figure 6. Evolution of Averge bubble size (Bz) ginst the increse of the trnsversl re of doughs (A) t ech smpling point (P). (Series : F1 ( ); Fm ( ) nd F2 ( )). A tendency cn be seen bout how Bz increses for ech dough during the process. The kinetics of increse of Bz were similr becuse t the end ll the doughs hd nondifferent vlue, however the doughs grew until reching different vlues of Tf nd A. Therefore, this vribility could be interpreted s differences in the rte of bubble growth due to differences in rheologicl properties nd chemicl composition. Anlysing ech smpling point individully, it cn be observed tht A did not present differences between doughs t T1, nd nor did Bz. At T2, the behvior ws similr to T1. The A vlues were similr, but differences between Bz were greter thn t T1. T2 18

19 corresponded to the Tf of F1, which implied mximum Bz nd A, which ment tht it ws not ble to endure the gs pressure resulting in colescence nd depletion. The vlues t T3 presented differences in both A nd Bz. The process prmeters of Fm should be between its precursor smples s it hd logiclly intermedite behvior between F1 nd F2 during fermenttion. Bsed on tht theory, the results were successful. Fm presented greter Bz thn both, while it hd the lowest A. Moreover, Bz t T3 presented non-different vlue with F1 t T2, s Fm reched colescence t T3 so it lso presented its mximum Bz s well s A. The differences between F1 nd Fm were due to fetures proportioned by F2, which incremented its gs pressure resistnce nd thus colesced lter s well s hving greter A.T4 correspond to the Tf of F2. It is the highest vlue of A, however Bz hd similr vlue to F1 t T2 nd Fm t T3. Therefore F2 ws ble to resist greter mounts of gs for longer time thn the other two doughs. Vribility between the Tf of doughs ment different levels of gs resistnce in the dough mtrix. An erly Tf indicted tht the gluten-strch mtrix ws not ble to endure gs divided into smll bubbles. So, smller bubbles completely disppered s consequence of the mss trnsport of gs from smll to lrge bubbles, cusing n overll depletion of the structure with time (Mills et l. 2003). The next step ws to nlyze the influence of Bz on the shpe of the dough. The Q prmeter t ech T ws studied jointly with Bz. Figure 7 shows the behvior of the Bz vs. Q correltion t ech T (r=0.865). The tendency of Q during the fermenttion process ws to decrese though the centrl zone of the dough hd higher increse rte thn the lterl zones, while Bz presented the opposite behvior. Differences in Q indicted tht doughs with non-different A presented different vlues of H. 19

20 This behvior could be consequence of different growth kinetics for the bubbles nd their distribution in the dough, s well s friction forces between the continer nd the dough, resulting in the typicl curvture of the lof (Pour-Dmnb et l. 2011; Ktenioudki et l. 2009). It is notble to observe the different loction of doughs for ech smpling time in Figure 7. Differences in Q vlues were observed for ech dough t ech P. 4,0 Bz (m ) 3,5 P4 P2 P3 3,0 2,5 P2 P1 P1 2,0 P3 P2 P1 1,5 2,2 2,22 2,24 2,26 2,28 2,3 2,32 2,34 2,36 2,38 Q (m) Figure 7. Reltionship between bubble size (Bz) vs. shpe prmeter of SL (Q) of ech dough t ech smpling point (P). (Series: F1 ( ); Fm ( ) nd F2 ( )) The gretest difference ws presented t T1, where F2 hd n importnt difference in comprison with F1 nd Fm. This ment tht lthough ll the doughs presented nondifferent A t T1, they hd significnt differences between both Bz nd Q. Therefore, F2 hd lower increse of H thn F1 nd hence Bz too. The rest of P followed the sme 20

21 pttern. These behviors suggested tht the distribution of the lrger F1 bubbles could ccentute the centrl zone of the dough, deforming it nd incresing the H vlues but not A. Modifiction of the gluten-strch mtrix during fermenttion could lso hve n influence on shpe chnges. As doughs reched Tf, the gluten-strch mtrix ws wekened nd gs pressure incresed (Pyler et l. 1988b; Wieser et l. 2006; Dreese et l. 1988), consequently, s bubbles strted colescing, decrese of dough surfce resistnce ws produced nd hence n increse H in the centrl zone of the dough. According to T2, T3 nd T4, doughs followed the sme tendency equlling their prmeters t Tf. These results suggested tht these experimentl doughs could not chieve Bz higher thn 3-3,3m obtining between for Q. Differences in Q t Tf could be explined becuse smpling ws crried out just before the Tf of ech smple, so smll differences in this vlue could ffect the results, incresing their dispersion. Furthermore, Q showed strong sensibility to vritions in Bz nd Dp. However, it is difficult to ttin better ccurcy for this chrcteristic, so it is necessry to improve the system to obtin less dispersion of the dt. The results could be explined by vribility in the mss trnsport of gs from smll to lrge bubbles, cusing differences in the stbility of the dough mtrix, nd in turn influencing dough shpe. Therefore, retention of gs ws ffected more or less depending on the dough, thus retrding the overll wekening of the structure nd resulting in higher yields of A.(Shh et l. 1998;Gn et l. 1995). 4. Conclusions 21

22 Reltionships between the informtion bsed on SL nd the internl dough structure were found. It ws possible to link the behvior of SL prmeters to bubble growth kinetics using 2D imge nlysis s tool. Although in the end, the dough fermenttion cpcity (Tf), the re (A), nd mximum height (H) were different, the reltionship between the prmeters were similr, reching similr finl bubble size s consequence of the colescence phenomen, independent of the bubble growth rte. The reltionships obtined could be useful to determine the stte of doughs during the fermenttion process using the SL method, in order to improve the chrcteristion of different flour btches s well s improve monitoring of the processes. It could represent bse to develop prediction models s well s devices to monitor the fermenttion phse of the bred-mking process. 5. Acknowledgements We wish to thnk the Polytechnic University of Vlenci nd Generlitt Vlencin for the finncil support they provided through the PAID nd GVPRE/2008/170 Projects, respectively. 6. References Autio, K., & Lurikinen, T. (1997). Reltionships between flour/dough microstructure nd dough hndling nd bking properties. Trends in Food Science & Technology. Volume 8, Issue 6, June 1997, Pges

23 Bjd, F. & Serš, I. (2011).Continuous monitoring of dough fermenttion nd bred bking by mgnetic resonnce microscopy. Mgnetic Resonnce Imging, 29(3), Brk, S., Mudgil, D. &Khtkr, B.S. (2013).Reltionship of glidin nd glutenin proteins with dough rheology, flour psting nd bred mking performnce of whet vrieties.food Science nd Technology, 51, Benlloch, J.V., Sánchez, A., Agusti, M., Alberto, P. (1996b). Weed Detection in Cerel Fields Using Imge Processing Techniques,Precision Agriculture. Precision gricu 3, Americn Society of Agronomy, Crop Science Society of Americ, Soil Science Society of Americ. Benlloch, J.V., Agusti, M., Sánchez, A., Rods, A. (1995). Colour segmenttion techniques for detecting weed ptches in cerel crops. Proc. of Fourth Workshop on Robotics in Agriculture nd the Food-Industry, Benlloch, J.V., Sánchez-Slmerón, A., Christensen,S., Wlter, M. (1996). Weed mpping in cerel crops using imge nlysis techniques. Conference on Agriculturl Engineering (AgEng '96), Universidd Politécnic de Mdrid Cocchi, M., Corbellinib M, Foc,G., Lucisnoc, M., Pgnic, M. A., Lorenzo T. &Alessndro, U. (2005). Clssifiction of bred whet flours in different qulity ctegories by wvelet-bsed feture selection/clssifiction lgorithm on NIR spectr.anlyticchimicact, 544, Dreese, P.C., Fubion, J.M., Hoseney, R.C., The effect of different heting nd wshing procedures on the dynmic rheologicl properties of whet gluten. Cerel Food World 33,

24 Flcone, P.M., Bino, A., Znini, F., Mncini, L, Tromb, G., Dreoss D. (2005). Three-dimensionl quntittive nlysis of bred crumb by X-ry microtomogrphy. Journl of Food Science, 70(4):E Gn, Z., Ellis, P.R. & Schofield, J.D. (1995).Gs Cell Stbilistion nd Gs Retention in Whet Bred Dough.Journl of Cerel Science, 21(3), Gonzlez-Brrón, U., Butler. F. (2008). Frctl texture nlysis of bred crumb digitl imges. Eur Food Res Technol (2008) 226: He, H., Hosene, R.C. (1991)Differences in gs retention, protein solubility, nd rheologicl properties between flours of different bking qulity. Cerel chemistry.68(5) Ivorr, E., Verdú, S, Sánchez A., Brt, J.M., Gru., R. Continuous monitoring of bred dough fermenttion using 3D vision Structured Light technique. Journl of Food Engineering 130 (2014) Ktenioudki, A., Butler, F., Gonzles-Brron, U., Mc Crthy, U., Gllgher, E. (2009).Monitoring the dynmic density of whet dough during fermenttion.journl of Food Engineering, 95, Le-Bil, A. Boumli, K., Jury, V., Ben-Aiss, F., Zunig R. (2009). Impct of the bking kinetics on stling rte nd mechnicl properties of bred crumb nd degssed bred crumb.journl of Cerel Science, 50, Lssoued, N., Bbin, P., Dell Vlle, G., Devux, M.F., Reguerre, A.L., Grnulometry of bred crumb grin: contributions of 2D nd 3D imge nlysis t different scle. Food Res. Int. 40, Mills, E. N. C., Wilde, P. J., L. Slt, J. &Skeggs, P. (2003). Bubble formtion nd stbiliztion in bred dough. Food Bioproducts Process, 81,

25 Mirlbes, C. (2004).Qulity control in the milling industry using ner infrred trnsmittnce spectroscopy. Food Chemistry, Otsu, N., (1979). A Threshold Selection Method from Gry-Level Histogrms.2eeeTtrnsctions on Systrems, mn, nd cybernetics, smc-9(1). Novotni, D., Ćurić, D., Glić, K., Škevin, D., Neđerl, S., Krljić, K., Gbrić, D., Ježek, D. (2011).Influence of frozen storge nd pckging on oxidtive stbility nd texture of bred produced by different processes. LWT - Food Science nd Technology, 44, Pérez-Nieto, A., Chnon-Perez, J.J, Frrer-Rebollo, R.R, Gutierrez-Lopez, G.F, Almill-Beltrn, L., Clderon-Dominguez, G. (2010). Imge nlysis of structurl chnges in dough during bking. LWT Food Science nd Technology. 43, Porosity in bred dough during proving. Food nd bioproducts processing 87, Pour- Dmnb A.R.S, Jfry, A. & Rfiee, Sh. (2011). Monitoring the dynmic density of dough during fermenttion using digitl imging method. Journl of Food Engineering, 107(1), Prud homme, R.K., Khn, S.A., Experimentl results on fom rheology. In: Prud homme, R.K., Khn, S.A. (Eds.), Foms: Theory, Mesurements, nd Applictions. Mrcel Dekker, New York, pp Pyler, E. J. (1988b). Physicl chemistry nd colloidl systems (3rd edition.)bking science nd technology, pp ). Knsns City, Missouri: Soslnd Publishing Compny. Scnlon, M.G., Zghl, M.C. (2001). Bred properties nd crumb structure. Food Res. Int. Volume 34, Issue 10, 2001, Pges

26 Wng, S., Austin, P., Bell, S. (2011). It s mze: The pore structure of bred crumbs. Journl of Cerel Science, Shh, P., Cmpbell, G. M.,. Mckee S. L, &. Rielly, C. D. (1998). Proving of bred dough: modelling the growth of individul bubbles. Trns IChemE, Vol 76, Prt C, June. Lucs, T., Grenier, D., Bornert, M., Chllois, S., Quellec, S., (2010). Bubble growth nd collpse in pre-fermented doughs during freezing, thwing nd finl proving. Food Reserch Interntionl. 43, Trobin, M. (1995).Error model of coded-light rnge sensor. Technique Report, Communiction Technology Lbortory. Updhyy, R., Debjni, G. &Mehr, A. (2012). Chrcteriztion of bred dough: Rheologicl properties nd microstructure Originl Reserch Article. Journl of Food Engineering, 109(1), Verdú, S., Ivorr, E., J. Sánchez, A., Girón, J., Brt, J.M. & Gru, R. (2013). Comprison of TOF nd SL techniques for in-line mesurement of food item volume using niml nd vegetble tissues. Food Control, 33(1), Wieser, H., Bushuk, W., &McRitchie, F. (2006).The polymeric glutenins.inc. Wilde, P., Fom formtion in dough nd bred qulity. In: Cuvin, S.P. (Ed.) Zhng, Z. (2000). A flexible new technique for cmer clibrtion. Pttern Anlysis nd Mchine Intelligence, IEEE Trnsctions on, 22(11), doi: / Zuñig, R., Le-Bil, A. (2009). Assessment of therml conductivity s function of porosity in bred dough during proving. Volume 87, Issue 1, Mrch 2009, Pges