Monitoring geometric characteristics of rice during processing by image analysis system and micrometer measurement

Variations of the geometric characteristics of three Iranian rice varieties, namely Tarom Mahalli, Fajr and Neda, at different processing levels were determined using micrometer and image processing methods. The results obtained by both procedures showed that the geometric characteristics of all three varieties ie length, width, height and projected area, are decreased and the sphericity increased by removing the outer and the brownish layers. The relationships between the micrometer data and the image processing ones were obtained linearly as represented by regression equations. It was found that the micrometer data are underestimated for all the geometric factors and that the true size and sphericity ie ones obtained using image analysis, can be estimated with root mean square error (RMSE) of less than 6% from the dimensional features provided by micrometer procedure.