Non-destructive identification of breeder rice Seed using transparent image analysis

摘要

In this paper, The transparent digital image analysis is applied for non-destructive identification of breeder seed class. The Rice Research Center and Development, Thailand difficult task is a classification breeder rice seed because of the similar shape and surface color. This paper presents the appropriate frequency light shines through to the grain that is analyzed for the histogram RGB and the separation by color property categories automatically. The rule of RGB histogram classification in ‘Jasmine’ rice seed and sticky rice seed are almost 97% and in “Red Sukhothai 1” red colored rice seed and dark-red colored rice seed almost 88% of rice seeds are identified correctly. In the light shines frequency appropriate to the grain that is analyzed the histogram RGB can be isolated red, white or yellow rice seeds are identified correctly 91%.