Milled Rice Grain Grading using Raspberry Pi with Image Processing and Support Vector Machines with Adaptive Boosting

As it is obliged that cultivars, location and year of production should be displayed on bag of milled rice grains, technology to identify them by the objective methods is indispensable. Milled rice grains produced at the several agricultural experiment stations were subjected to the RAPD method. Each DNA was extracted by CTAB method and amplified by PCR method using 10 mer primers. It became possible to identify the rice cultivars using milled rice grains as samples by the electrophore-sis of the PCR products among the dominant 10 rice cultivars in Japan.

Pathum Thani 1 (PTT1) is a photoperiod-insensitive, aromatic Thai rice variety that is grown year-round. The rice from some locations is often priced lower than others due to sub-standard grain quality. This study sought to determine the limiting grain quality characteristic(s) in dry season PTT1 and their distribution across Thailand's irrigated rice regions. To do so, we evaluated the milling and physical quality of milled rice grain of dry season PTT1 from 24 provinces in 5 regions in Thailand. Sixty-seven paddy rice samples were collected and evaluated for head rice yield, chalkiness, whiteness, and translucency. Head rice yield varied by region, with the highest (48.1%) found in samples from the Central region, compared to 34.5-39.7% elsewhere. Head rice chalkiness was the physical quality that varied more widely among the provinces within each region than among regions, with chalkiness at a level that would adversely affect price in more than one-half of the samples. Contrary to the general perception that chalky grain is less resistant in milling, head rice yield actually increased with total chalkiness, expressed as % chalky grain by weight. The total chalkiness correlated negatively with translucency, while the head rice chalkiness correlated positively with whiteness. Grain chalkiness was identified as the grain quality attribute of PTT1 rice that varied with location; this directly affected the price of milled rice grain, as well as indirectly through its relationship with head rice yield and visual appearance of the milled rice grain.

Rice is a very important item of food and commerce in Sierra Leone and so information o n the quality of available rice grains can serve as a useful indicator for the technical status and level of competitiveness in the local rice industry . During this study a n objective evaluation was conducted to measure and characterize the quality of rice grains available in the local markets of Sierra Leone. A total of 315 randomly selected rice samples from 45 markets selected from the four major cities of Sierra Leone (Makeni, Bo, Kenema and Freetown, representing urban communities from the northern, southern, eastern and western parts of the country , respectively ) were evaluated. Quality evaluation involved measurement of moisture content, number of paddy (unmilled rice kernels) in 1 kg of milled rice and other quality factors normally used for the grading of milled rice. Rice samples were then graded based on criteria adapted from the Philippines Rice Grading Standards for milled rice grains. The results showed that the quality of all grain samples evaluated was generally poor, with 63.2% of the samples failing to meet the criteria set for Grade III rice quality ( meaning that the quality level was worse than grade III ) . Quality measures obtained for imported samples appeared to be superior to that obtained for the local samples in terms of higher proportions of superior grades (grade II or better ) . Comparison of measures of grading factors revealed that among the four cities considered in this study, grains from Kenema were of the lowest quality . Further examination of grade limiting factors revealed that the most critical factors responsible for poor quality outcomes in grain sample were (i) the number of paddy in 1kg of milled rice and (ii) the moisture content of grains . The study provided quantitative measures of the quality status of rice grains available in Sierra Leone , as well as a means of identifying the major binding constraints to rice grain quality, in terms of the grade limiting factors. It is speculated that t he low quality of rice grains observed in this study could be the result of avoidable quality defects that may be linked to an undeveloped national system for rice milling and handling .

A possible mechanism of mineral responses to elevated atmospheric CO2 in rice grains

ABSTRACT: Chemometric formulas for predicting amylose content (AC) and protein content (PC) were developed based on the surface and overall physical properties of single‐cooked milled rice grains as predictive variables. The formula with 8 predictors for AC determination exhibited the highest accuracy (coefficient of determination: R2= 0.952, standard error of calibration: SEC = 1.77) for calibration and the lowest standard error of prediction (SEP = 2.07) for unknown samples (AC: 0 to 30%). The formula with 7 predictors was effective in enhancing the prediction accuracy (SEP = 1.32) among non‐waxy samples in a narrow range of AC (15 to 20%). The formula with 8 predictors for PC determination showed a higher accuracy (R2= 0.449 and 0.470) for both calibration and prediction than that of a conventional 3‐grain method (R2= 0.168 and 0.124). It was possible to accurately determine AC using physical measurement of single grains instead of chemical analysis.

The consumption of resistant starch (RS), which can increase satiety and reduce calorie intake, may help ameliorate lifestyle‐related diseases such as diabetes, hypertension, and obesity. The identification of genes or quantitative trait loci (QTLs) may facilitate improvement in the amounts and properties of RS in cereals such as rice. A genome‐wide association mapping of the RS content in milled rice grain was carried out to identify QTLs in a rice germplasm collection. The RS contents among 105 rice accessions ranged from 0.3 to 2.42%, and the apparent amylose content (AAC) was positively correlated with the RS content (R2 = 0.75). Four QTLs were detected, and each explained a total phenotypic variation of 10–13%. Two of these, on chromosome 6, were identified at the Waxy gene locus encoding granule‐bound starch synthase I and starch synthase IIa. The other two QTLs, on chromosomes 8 and 9, were close to candidate genes encoding the starch debranching enzyme isoamylase I and ADP‐glucose pyrophosphorylase small subunit 1, respectively. The results of this study should improve our understanding of the genetic basis of RS in milled rice and may assist plant breeders attempting to develop new rice varieties and food processors working to develop functional foods.

Using Check-All-That-Apply (CATA) method for determining product temperature-dependent sensory-attribute variations: A case study of cooked rice

Brown rice over-milling causes high economic and nutrient loss. The rice degree of milling (DOM) detection and prediction remain a challenge for moderate processing. In this study, a self-established grain image acquisition platform was built. Degree of bran layer remaining (DOR) datasets is established with image capturing and processing (grain color, texture, and shape features extraction). The mapping relationship between DOR and the DOM is in-depth analyzed. Rice grain DOR typical machine learning and deep learning prediction models are established. The results indicate that the optimized Catboost model can be established with cross-validation and grid search method, with the best accuracy improving from 84.28% to 91.24%, achieving precision 91.31%, recall 90.89%, and F1-score 91.07%. Shapley additive explanations analysis indicates that color, texture, and shape feature affect Catboost prediction accuracy, the feature importance: color>texture>shape. The YCbCr-Cb_ske and GLCM-Contrast features make the most significant contribution to rice milling quality prediction. The feature importance provides theoretical and practical guidancefor grain DOM prediction model. PRACTICAL APPLICATION: Rice milling degree prediction and detection are valuable for rice milling process in practical application. In this paper, image processing and machine learning methods provide an automated, nondestructive, and cost-effective way to predict the quality of rice. The study may serve as a valuable reference for improving rice milling methods, retaining rice nutrition, and reducing broken rice yield.

The effects of pH on botulinal toxin production in boiled rice and the heat resistance of the spores were investigated. Toxin was detected in boiled rice (pH above 5.0) packed in a plastic film pouch or semi-rigid plastic tray with deoxidant and inoculated with mixed spores (ca. 106 per container) of Clostridium botulinum 62A, 90A, 213B and B1G4 strains following incubation for one to six months at 30°C. Lowered pH (5.0) and cooking, for 19 min at 99°C, failed to prevent toxigenesis. The heat resistance of C. botulinum spores in rice gruel (pH 5.3) was assessed. D (decimal reduction time at 100°C) and z (a slope index of the thermal death time curve) were 9 min and 10°C for 62A, and 7 min and 9°C for 213B, respectively. Raw rice was also examined for contamination with clostridial spores. No clostridial spores could be found in milled rice grains. The number of spores per 100 g was 3 to 73 CFU for brown rice and more than 100 CFU for bran.

Effects of Storage Period, Milling Degree of Stored Brown Rice Grains, and Soaking of Milled Rice Grains on the Properties of Cooked Rice Grains

This paper combines multispectral imaging and simple image processing techniques for the non-destructive identification of Thai rice breeds. Especially, we exploit only two fluorescent wavelengths in a 500-580 nm wavelength band and utilize simple image thresholding, blob filtering, and blob analysis techniques in order to identify 8 different Thai rice breeds. Other key features include no waste produced and fast identification time. In our experimental study, UVC light is used as our exciting light, a liquid crystal tunable optical filter is used for our wavelength selection, and a camera with 644×488 active pixels is used to capture desired wavelength images. Milled rice grains from 8 different Thai rice breeds having similar size and shape are tested. There is also one glutinous rice breed in our experiment. Our experimental result shows that by suitably applying image thresholding, blob filtering, and blob analysis to fluorescent images, all Thai rice breeds can be effectively identified.

Background and objectivesProtein composition analysis of 81 rice samples was performed on three cultivars harvested different regions in Japan, and the effects of the protein composition on the nitrogen composition in sake.FindingsCrude protein content showed a significant positive correlation with glutelin content. Yamadanishiki cultivars showed lower crude protein and glutelin contents compared with Gohyakumangoku cultivars harvested from the same field. Total nitrogen content in the enzyme digest of steamed rice grains showed a significant positive correlation with the crude protein and glutelin contents. Ten polished (70%) white rice samples were investigated in a small‐scale sake brewing test. Total nitrogen compounds, bitter‐tasting peptides, and free amino acids in the brewed sake positively correlated with the glutelin content of the rice samples. There were no clear correlations between α‐amylase and acid proteinase activities with the nitrogen compounds in the sake.ConclusionsThe results indicated that the glutelin content in the cultivar rice affects sake quality through taste‐active and other nitrogen‐containing compounds produced from rice glutelin in the sake brewing process.Significance and noveltyWe revealed that the glutelin content in the milled rice grains is important to brew high quality of sake clearly. The present results would be useful for cultivation or breeding of sake rice.

The relationship between rice protein composition and nitrogen compounds in sake

The amylose content, amylographic characteristics and storage proteins of milled rice grains on the primary and secondary rachis branches were investigated. The amylose content of grains on the primary rachis branches with superior palatability was higher than that on the secondary rachis branches. Grains on the primary rachis branches showed higher maximum viscosity and breakdown values, and lower gelatinization temperature than those on the secondary rachis branches. These results suggested that the palatability of cooked rice of grains with high amylose content and amylogram values was superior to that with low amylose content and amylogram values within a cultivar under the same cultural conditions. In terms of protein fraction in the starchy endosperm, there was little difference in prolamin content and in albumin + globulin content between grains on primary and those on secondary rachis branches, whereas total protein content and glutelin content of grains on the primary rachis branches were noticeably lower than those on the secondary rachis branches for all the cultivars. Total protein, glutelin and prolamin content of a good palatability cultivar, Koshihikari were lower than those of moderate palatability cultivars Nipponbare and Reihou. These results indicate that, among storage proteins, glutelin and prolamin play important roles in rice palatability and that it is possible to use glutelin as an indicator of palatability within a single cultivar and prolamin as an indicator of palatability among cultivars.

Abstract. The demand for food authentication is increasing worldwide in recent years. Rice (Oryza sativa L.) is a staple food and is traded globally in great amount. This study proposed an image-based method for discriminating rice grains of 45 cultivars. Images of rice grains were taken by using a digital camera. Morphological, color, and sterile lemma traits of the grains were quantified using image processing algorithms. The traits were analyzed by principle component analysis to investigate the discrepancies among the cultivars. Support vector machines were then developed to identify the cultivars of the grains by using the traits as the inputs. The overall accuracy of the classifiers achieved 85.02%.