Papers by Ahmad Ghazanfari

Canadian Agricultural Engineering, 1998
Table I: USDA standards for size grading of pistachio nuts (California !'istachio Commission 1995... more Table I: USDA standards for size grading of pistachio nuts (California !'istachio Commission 1995) sorting pistachio nuts .lre not precise and because of their direci contact with the nuts, cause l11eclwnical damaoe to the kernels. Electro-optical sorting devices can only c1as~fy the nuts into two classes, namely "rejecls" and "accepts". while in mosl cases multi-category classification is sought. Machine vision c1assificalion of pistachio milS is an alternative to the conventional eleclro-optical and mechanical sorting devices. During Ihe past two decades Ihere has been an increasing interest toward the use of machine vision for sorting (Ind grading agriculwral products. Examples arc: maturity deteclion in peanuls by Ghate el al. (1993). sepamtion of mixed lOIS of tall fescue and ryegrass seeds by Churchill el al. (1993). anel cr"ck deteclion in eggs by Goodrulll and EISler (1992). Ghazi1n fari and Irudayaraj (1996) uscd a machine vision system in conjunction with a string matching technique for separating four varieties of pistachio nuts. They rcportcd that the classification accuracy was sensitive to the parameters of the cost functions Llsed in that method. Ghnzanfari, A., Wulfsohn. D. and lrudayaraj. J. 1998. J\Jachillc "isi~n grading of pislachio nuls using gray-level histognul1. Can. Agnc. Eng. 40:061-066. A machine vision SYSICll1 was lIsed to classify "in the shelr" pistachio nuts based on USDA l.!n:ldes. The gray-level histogram data oblained from Ihe gray scale i~l:lge of the nuts were analyzed to select a set of suitable recQ[witioll fcmures Based on Ihe analyses, 'he meall of Inc !\l1ly-Ieve\ hi~logram ovcr 50 10 60 gray-level range and (he area of each nut (the imcgral of ils 6"tJ)-I\.~' ct hlslOgrom) were sclcclCd as the recognition fealUres. The selectcd fcaturcs wcre used as input to threc c1assificmion schcmcs: a Gnussian. a decision tree. ilnd a I11ulti-lnycr neural network (t\.'ILNN). The Ihree classifiers hnd similnr recognition rates. However, the MLNN classifier resulted in slightly higher performance with morc uniform classification accuracy than lhe Dlher two classifiers. Keywords: machine vision. neuralnelworks. pallent recognition, pistachio nuts. classification. Un systellle de vision artificiellc a etc utilise pour classifier des pistaches ell ecales. scion les categories dll Dcpartclllent Alllcricain de l'Agriculture (USDA). Un histogramlllc lllolltranties di ffercnts tOllS de gris contenus dans I'image a etc analyse pOllr iderllifier des parametres de reconnaissance satisfaisallls. A panir de ces analyses. Ia moyenne de I'histogrnmme des tons de gris sur une palelle de 50~i 60 tOilS de gris. et I'aire de 13 pistachc (integrale de son histogranulle de tons de gris) ont etc retenues comllle parametres de reconnaissance. Les parametres choisis ont ele utilises comme intrants dans trois methodes de classification: systeme de Gauss. arhre de decision et reseau neuronal multi-couches (M LNN). La vitcsse de reconnaiss;:lnce ctait semblable pour les trois mcthodes. Cependant. Ie rcse~1U ncuronal multi-collches a etc Icgcrcll1ent plus perfomlant que les deux ,llItres methodes ct a pemlis unc classification plus uniformc. Mots-des: vision nnificicllc. rcscaux ncuronenux, reconnaissance de fonncs. pisl<lches. classification.
Asian Journal of Plant Sciences, Aug 15, 2006
Numerical Heat Transfer Part A-applications, Aug 1, 1996
ABSTRACT
Fuel, Jul 1, 2013
ABSTRACT
ASABE/CSBE North Central Intersectional Meeting, 2006
Transactions of the ASAE, 1996
A modified cyclic string matching algorithm was developed and applied to classify four classes of... more A modified cyclic string matching algorithm was developed and applied to classify four classes of pistachio nuts, based on their two-dimensional shapes. In this method, a pistachio nut was represented by a string consisting of N angularly equispaced radii extending from the centroid to the boundary. An algorithm was also developed for determining a prototype for each class. The recognition algorithm, based on a string matching technique, calculated the cumulative distance between an unknown and a class prototype. The class of the unknown was determined by the minimum-distance classification rule. The developed algorithm gave an overall accuracy of 90%. This procedure could be extended to classify other agricultural materials. Keywords.

Journal of Zhejiang University-science B, Jul 1, 2010
Date pits for feed preparation or oil extraction are soaked in water to soften before milling or ... more Date pits for feed preparation or oil extraction are soaked in water to soften before milling or extrusion. Knowledge of water absorption by the date pits helps in better managing the soaking duration. In this research, the process of water absorption by date pits was modeled and analyzed using Fick's second law of diffusion, finite element approach, and Peleg model. The moisture content of the pits reached to its saturation level of 41.5% (wet basis) after 10 d. The estimated coefficient of diffusion was 9.89×10 −12 m 2 /s. The finite element model with a proposed ellipsoid geometry for a single date pit and the analytical model fitted better to the experimental data with R 2 of 0.98. The former model slightly overestimated the moisture content of the pits during the initial stages of the soaking and the latter model generally underestimated this variable through the entire stages of soaking process.
Timothy hay was compressed using a laboratory hydraulic press. The applied pressure was increased... more Timothy hay was compressed using a laboratory hydraulic press. The applied pressure was increased from 0.0 to 21.0 MPa at intervals 3.5 MPa. The moisture content of hay varied from 7 to 20% wet basis (wb). During each compressing interval, the thickness of compressed hay was measured and the density of the resulting bales was calculated. In general, the density of the bales increased and their thicknesses decreased with the applied pressure, however the trends were not linear. The maximum bale density was about 1163 kg/m 3 which occurred at 21 MPa pressure for both 16% and 20% moisture levels. The 16% moisture content (MC) hays resulted in more dense final bales.

Iranian Food Science and Technology Research Journal, 2016
روغن دانه درخت مورینگا در صنایع غذایی، دارویی، آرایشی و بهداشتی مورد استفاده قرار می گیرد. در این... more روغن دانه درخت مورینگا در صنایع غذایی، دارویی، آرایشی و بهداشتی مورد استفاده قرار می گیرد. در این پژوهش روغن خام مورینگا با استفاده از آب و اسید فسفریک صمغگیری و برخی خصوصیات فیزیکوشیمیایی و پایداری اکسایشی آن اندازه گیری و مورد مقایسه قرار گرفتند. آنالیز گاز کروماتوگرافی نشان داد که حدود 71 درصد اسیدهای چرب روغن مورینگا را اسید اولئیک و 24 درصد آن را اسیدهای چرب اشباع پالمیتیک، استئاریک و بهینیک تشکیل می دهند. صمغگیری باعث کاهش عدد پراکسید از 4 به 35/2 میلیاکیوالان/ کیلوگرم روغن؛ کاهش اسیدهای چرب آزاد از 05/2 به 07/0 درصد؛ کاهش عدد صابونی از 11/190 به 55/180 میلی گرم KOH/گرم روغن و کاهش جزئی عدد یدی گردید. همچنین صمغ گیری سبب افزایش ویسکوزیته و چگالی روغن حاصله گردید، اما تاثیری بر ضریب شکست نداشت. بهعلاوه در اثر صمغگیری نقطه اشتعال روغن خام مورینگا از 115 به 205 درجه سانتی گراد افزایش یافت، ولی در مجموع عدد پراکسید روغن خام بالاتر از روغن صمغگیری شده بود. بر اساس نتایج حاصله با افزایش دما از 3 به 120 درجه سانتیگراد عدد پراکسید روغن خام و صمغگیری شده افزایش یافت. بطوریکه د...

Journal of Microwave Power and Electromagnetic Energy, 2005
The feasibility of microwave dehydrating flax fiber was evaluated using a commercial domestic mic... more The feasibility of microwave dehydrating flax fiber was evaluated using a commercial domestic microwave oven at four power settings representing 200, 300, 400 and 500 Watt (W) power level. Due to the possibility of local heating and consequent fiber degradation, the changes in color of the flax fiber at different levels of temperature were also investigated. The dehydration processes at various power levels were simulated by Page model. Based on visual inspection, color analysis and scanning electron microscopy (SEM) of the fiber, it was revealed that discoloration of the fiber occurred at about 170 degrees C. At 200 and 300 W power level, after 10 minutes of dehydrating, the moisture content of the fiber reached from initial 7.9% close to 2.0 and 1.0%, respectively. For 400 W power level, the moisture content of the fiber dropped to 0. 10% in about 9.5 minutes. Major discoloration of the fiber was noticed when dehydration was proceed beyond 4.5 minutes for 500 W treatment. The Page model very well fitted the experimental data. The coefficients of determination calculated from the model and the experimental data increased with increase in applied microwave power
International Journal of Neural Systems, 1997
A multi-structure neural network (MSNN) classifier consisting of four discriminators followed by ... more A multi-structure neural network (MSNN) classifier consisting of four discriminators followed by a maximum selector was designed and applied to classification of four grades of pistachio nuts. Each discriminator was a multi-layer feed-forward neural network with two hidden layers and a single-neuron output layer. Fourier descriptor of the nuts' boundaries and their area were used as the recognition features. The individual discriminators were trained using a biased technique and a back-propagation algorithm. The MSNN classifier gave an average classification performance of 95.0%. This was an increase of 14.8% over the performance of a multi-layer neural network (MLNN) with similar complexity for classifying the same set of patterns.
ASABE/CSBE North Central Intersectional Meeting, 2006
Transactions of the ASAE, 1996
A multi-structure neural network (MSNN) classifier was proposed and applied to classify four vari... more A multi-structure neural network (MSNN) classifier was proposed and applied to classify four varieties (classes) of pistachio nuts. The MSNN classifier consisted of four parallel discriminators (one per class), followed by a maximum selector Each discriminator was a feed-forward neural network with two hidden layers and a single-neuron output layer The discriminators were individually trained using physical attributes of the nuts extracted from their images as input. The performance of MSNN classifier was compared with the performance of a multi-layer feed-forward neural network (MLNN) classifier The average classification accuracy of MSNN classifier was 95.9%, an increase of over 8.
Uploads
Papers by Ahmad Ghazanfari