Agricultural Engineer Develops Low-Cost Modern Analysis for Milled Rice
A multi-awarded agricultural engineer at the Bureau of Postharvest Research and Extension (BPRE) has developed a state-of-the art but low-cost computer vision system (CVS) for analyzing the quality of milled rice.
He is Dr. Manolito C. Bulaong. In his CVS, an ordinary scanner serves as the “eye” of the system. It replaces the expensive digital cameras used in the conventional CVS for image acquisition. The image processing software, on one hand, extracts the shape and color patterns from each grain image.
The artificial neutral network (ANN) meanwhile acts as the “brain” of the system. It recognizes the shape and color patterns from each grain and learns the quality category it belongs.
This CVS can compute the percentage by weight of good quality grains and defective grains—such as broken grains, brewer’s grains, damaged, chalky, discolored, immature, and red kernels—present in a sample. It can also count the number of palay and measure the grain length and output of the grade of milled rice according to the specification of the National Grains Standard. And it can do a complete analysis of a 100-gram sample in less than 30 minutes.
The conventional milled rice quality analysis however, is a tedious, slow and expensive process. The analysis per sample takes more than an hour, and complete analysis costs P550.
The result is also subjective as it is affected by the skill and physical condition of the classifier, lighting, and other working conditions.
At the National Food Authority (NFA), trained classifiers visually inspect each milled rice grain based on size and color; each grain is classified according to quality category it belongs.
“The development of the low-cost CVS for milled rice quality analysis will ensure objective, accurate and fast results, and will modernize the method used by the grain industry,” Bulaong said.
“As our country gears up for globalization, grains standardization is one of the strategies for modernizing the agriculture sector, particularly improving the efficiency and global competitiveness of the grains industry,” he added.
With him in this project “Quality Analysis of Milled Rice Using Computer Vision” are Engr. Ruben E. Manalabe and Jayson T. Carbonel of the Postharvest Engineering Department of BPRE and Dr. Oliver C. Agustin of Vera Equinox Technologies. It was funded by the Philippine Council for Agriculture, Forestry and Natural Resources Research and Development in collaboration with NFA.
By Erwin S. Embuscado