Development of Rice Grain Phenotype Quality Verification System using Machine Learning

Main Article Content

Kulwadee Tanwong Poonpong Suksawang Yunyong Punsawad


The objectives of this research were (1) to examine the rice grain phenotype quality using machine learning by convolution neural networks (CNNs) techniques and ( 2) examine the effectiveness rice grain phenotype program quality to the classified rice grain according to jasmine rice classification criteria under the Agricultural Product Export Act By creating rice grainy digital image. The sample obtained from Pathumthani Rice Research Center amount 2,150 images covering 4 jasmine rice groups, teaching machinery according to Auzuble meaningful learning theory. The proportion of practicing model and testing was 60:40. Calculated practicing examine the rice grain quality model by 1,850 images repeat in 25 testing model layers. The training efficiency model result was approaching to 100 percent. After that, repeat training model by 1300 images with developed applications found that, accuracy ratio to discover correct image from over all image 0.907 and inaccuracy (F-measure) 0.028


Article Details