Quantitative Measurement of the Texture and Cracks of Hulled Nonglutinous Rice Grains Using Image Data Processing System

Abstract

A quantitative method of mechanical grading of the quality of nonglutinous hulled rice has been developed, using the image data processing system.
The linear relationship was obtained between the attenuances of the images displayed on the color monitor and the readings of the scale of the multi-color data system. The regression was
y=-0.08285x+2.31360
where x is the reading of the scale, and y is the attenuance.
A mean value of the attenuances of the image of a sample was
E_m=∑a_j·s_j¹⁰_j=0/∑a_j¹⁰j=0
where a_j is the distribution of attenuance expressed as the projected area of equal light intensity, and s_j is the slice level of attenuance.
Texture of rice samples was well determined by the mean attenuance, when they were illuminated by diffused, transmitted ray; 1st grade≤0.485<2nd grade ≤0.530<3rd grade ≤0.611
Cracked grains were well detected under dark field illumination by oblique transmitted ray through the sample holder which was developed.
The number of cracked grains was estimated by the following formula.
y=-7.03757-0.60781a₇+0.954693c₆
The content of whole grains or normal grains, a major factor for quality evaluation could be estimated from the distribution of attenuances.