3:14 AM | апр. 11, 2018
In the developed system, first, images of milled rice grader samples are captured. The image processing operations are then executed to eliminate undesirable noises from images. After kernel segmentation, a primary feature vector is created based on some shape, size and color features. To have a high classification accuracy, it is necessary to prepare a proper input vector for the classifiers. For this purpose, the primary extracted features are subjected to a correlation-based feature selection procedure to reject the inferior features. Finally, the best classifier is selected for milled rice grading by examining four commonly used metaheuristic approaches. The entire applied methodology is described in the following sections.
Rice is a member of the grass family (Gramineae). There are more that 10,000 species of grasses distributed among 600 genera. Grasses occur worldwide in a variety of habitats. They are dominant species in such ecosystems as prairies and steppes, and they are an important source of forage for herbivorous animals. Many grass species are also primary agricultural crops for humans. As well as rice, they include maize, wheat, sorghum, barley, oats, and sugar cane.
Quality testing parameters including the milling parameters head rice recovery (HR%), broken fraction (%), husk (%), bran (%); physical characteristics kernel length (mm), breadth (mm), thickness (mm), shape ( length breadth ratio LBR), color, appearance, quality index; cooking characteristics cooked kernel length (mm), elongation ration (ratio of post cooked kernel length to pre-cooked kernel length), cooking time, bursting %, stickiness(%), water uptake were determined for each lot basmati, coarse and hybrid category separately. Generally, a good quality rice should have a high percentage of whole unbroken grains, little or no chalk, translucent appearance, uniform coloration and good for the purpose for which it has been produced (white for raw-milled rice and with a yellowish tinge for parboiled rice), shape (length and length-width ratio) should be right for the variety type, excellent cooking properties- should satisfy the consumers' preference for cooked rice for the particular kind of food preparation (Anonymous, 2014). Keeping in view these points, moisture was determined using Kett® moisture testing meter PB ID2 after calibration through moisture testing standard plate PB ID2 Tester 15±0.1 accompanied with the instrument. Length (l), breadth (b), thickness (t), length breadth ratio-lbr(shape), and type-quality index (lbr/t) average for randomly selected 1000 or more grains of replicate was measured on the office table glass of known length (cm) by placing kernels in end to end or side to side arrangement for length and width respectively. Steel scale 30cm long was used to measure width of the kernels placed in queue on the office table glass.
From the analysis of the separator of indent cylinder type, the separating factors were clearly defined and their effects were discussed. In addition, the separating mechanisms were modelled, the fractional recovery curve was calculated and compared with the experimental values. The obtained results were as follows: 1) Besides the factor of grain length, the separating factors included the shape and coefficient of friction of grain. The distribution form of the values of physical properties of these factors also had strong effects on the separating performance. 2) The factors which had effects on the separating performance are: size, shape and arrangement density of indent parts; size, number of revolution and material of cylinder; separating time, hopper angle and supplied quantity. It is possible to improve the separating performance if these factors are suitably selected. 3) The boundary value of width of broken and whole grain, and half of this value can be used to determine the width and depth of indent parts respectively. 4) The separating performance was good when the coefficient of friction of grain-cylinder was high, and the improvement of this performance was assured if the grain of which thickness is small was removed by a thickness grader before separating. 5) The separating accuracy was good when the separating time was long, and the hopper angle could be properly selected by regulating this angle in such a way to get the value PB which should be high and small in terms of broken and whole grain respectively. PB was the probability of grain entering the hopper being carried by indent parts.
To assess the effectiveness of using Teflon material in search of locally available quality/effective roller material for a prototype rice destoner machine developed in Nigeria making use of four locally cultivated rice varieties. This required determination of the effectiveness of Teflon material as roller for the developed prototype dehusker/destoner. Microsoft Excel 2007 and Minitab 16 were used to analyze the ANOVA of two main factors (paddy varieties and moisture content). The dehusker cleaning efficiency was found to be 94.73 %, coefficient of dehusking was 0.63, coefficient of wholeness was 0.85, dehusking efficiency was 50.54 %, dehusking capacity was 10.56 kg/h. Also the effect of moisture content and test paddy on coefficient of wholeness and dehusking efficiency were significant at p ≤ 0.05, while only moisture content effect was significant on cleaning efficiency. Further search for better roller material to replace the Teflon material to improve the prototype dehusking performance for different rice varieties was recommended.