Myanmar Rice Grain Classification Using Image Processing Techniques

Modern technologies are being used in agriculture such as quality control and classification of grains that are very important for more productive and sustainable production. Classification of the similar small rice grains can be also made with the help of image processing techniques. This paper studies different characteristics of Myanmar rice grains and their varieties. The classification of various varieties of rice grains is made by using image processing techniques and algorithms. Five types of rice grains in Myanmar such as Paw San Hmwe, Lone Thwe Hmwe, Ayeyarmin, Kauk-Nyinn-Thwe and Kauk-Nyinn-Pu are considered for present study in classifying the rice seeds and quality. Firstly, each grain image is preprocessed to enhance the grain image and then segmented by using the edge detection methods such as threshold method. Five morphological features are extracted from each grain image. This system emphasizes on the development a computer vision-based system that is combined with proper heuristic algorithms for automatic classification of Myanmar’s rice grain samples. This research is very significant in Myanmar because Myanmar is great producer of different qualities of rice grains and therefore the study and basic implementation would greatly help the researchers, agriculturist and other stakeholders of agricultural growth.