Detection and Estimation of the Weight of Filled Grains From a Paddy Rice Sample, Using Digital Image Processing with ImageJ®

Abstract

The identification and determination of the weight of full and empty grains of rice extracted from a panicle is a variable related to genetic characteristics of the plants, to the environmental conditions during grain filling and influence yield. Digital Image Processing (DIP) is an objective, effective tool applicable to different identification and classification processes. It is useful for the morphological characterization of grains and seeds. The objective of this work is to present a technique based on PDI for the identification of full grains and the estimation of their weight. It uses an image of the sample obtained with a desktop scanner with a transparency adapter, a scale and the ImageJ® program. The detection of full grains was carried out using the grain histogram. The grain recognition percentage was on average 85%. For weight estimation, the pixels corresponding to the projected area of the grains were correlated with their weight. The values obtained have an error of 1.2% on average. The proposed method is operationally simple, efficient and does not require prior training or special equipment.