[4-1015-D] Food Quality (1)

[4-1015-D-04] Chalkiness Index of Sake Rice “Yamada Nishiki” Using Ultraviolet-Near-Infrared Transmission

*Khokan Kumar Saha^1,2, Firmanda Al Riza Dimas¹, Yuichi Ogawa¹, Tetsuhito Suzuki¹, Naoshi Kondo¹, Takuma Sugimoto³ (1. Lab of Bio-sensing Engineering,Graduate School of Agriculture, Kyoto University, Kitashirakawa-Oiwakecho,Sakyo-ku ,606-8502.(Japan), 2. Department of Agricultural Engineering, Bangabandhu Sheikh Mujubur Rahman Agricultural University,Salna,Gazipur-1706.(Bangladesh), 3. Senior Researcher, Hyogo Prefectural Agriculture, Forestry and Fisheries Technology Research Center, Addo City, Hyogo Prefecture ,671-3441.(Japan))

Keywords:White core, Transmittance, Chalky thickness , Sake rice

White core refers to the opaque area in the kernel which is an important consideration for grading the raw material (sake rice) in the brewing industry. There is a relation between white core level and polishing ratio for sake rice to make premium quality sake. Currently, the evaluation of white core level in grains of sake rice is performed by the internal chalky thickness which is a destructive method. The aim of this research is the grading of sake rice chalkiness in non-destructively. Non-white core and white core [light, medium] and chalky rice, cultivar “Yamada Nishiki”, were used for this experiment. In this research, the transmission optical property in the ultraviolet-near-infrared (UV-NIR) region of rice was measured in order to characterize the suitable wavelength region in sake rice. The transmission optical property was measured by Jasco V-670 spectrophotometer equipped with an integrating sphere (ISH-723). The thickness of the chalky part was measured destructively. The evaluation of transmission optical property was done by image processing techniques. The images acquired by transmission mode of 4 types of intact sake rice kernel by using ring UV LEDs, Blue LEDs, Green LEDs, Red LEDs, and NIR LEDs in which the peak wavelength of the LEDs was 365nm, 465nm,525nm,630nm, and 830 nm respectively. The rice samples were more penetrated by light in the NIR region. The result indicates that there is more contrast between white core and non-white core rice in NIR image and followed by the red, green, blue and UV region images. Therefore, the wavelength region in NIR showed better discrimination between white core and non-white core sake rice. Finally, four images features; the ratio of the area (opaque area to full area), transmittance gray level intensity, the ratio of axis length in opaque (minor axis length to major axis length of opaque) and distance of centroid (opaque and full) were extracted from NIR transmittance images. The proposed chalkiness index (ratio of chalky thickness to whole kernel thickness) was inversely correlated with NIR transmittance image gray level intensity. The classification accuracy of the white core and chalky rice was 94.34% by support vector machine (SVM). Therefore, these results can be used to developed machine vision-based white core detection of sake rice.