MACGRAN-IJ: AN EFFICIENT METHOD TO ANALYZE DIMENSIONS AND COLOR OF WHEAT GRAINS

Abstract

The dimensions, shape, and color of wheat grain are important in plant breeding, although their manual measurement is difficult and time-consuming. While it is true that digital image analysis (DIA) with specific software has made it easier, some are semiautomatic, exclude color, and require the digitization of individual grains. The goal of this study was to compare the precision, accuracy, and efficiency of MacGran-IJ versus SmartGrain, GrainScan, and manual caliper measurement. The morphological characteristics of durum wheat (Triticum durum Desf.) and bread wheat (Triticum aestivum L.) grains were measured using DIA with MacGran-IJ, including color (in RGB and CIELab color spaces), size (mm2), lateral length (LLG, mm), ventral length (VLG, mm), lateral (LWG, mm) and ventral (VWG, mm) width, perimeter (mm), and circularity. Length, LWG, and VWG were measured with a digital caliper. The time (s) taken in each method was recorded. Size, LLG, VLG, and VWG were the same between methods for both species, but the LWG measurement obtained with vernier showed differences between the three programs (p ≤ 0.01), as well as low accuracy and precision. VWG obtained manually had R2 values of 0.92, 0.91, and 0.89 (p ≤ 0.01) for MacGran-IJ, SmartGrain, and GrainScan, respectively. The ventral and lateral color red of wheat grain differed between MacGran-IJ and GrainScan; the lateral part of the bread wheat grain showed differences for red, green, and blue (p ≤ 0.05). DIA in CIELab color required 11.57 s, and in RGB it needed 7.25 s (p ≤ 0.01). Manual measurement of 50 wheat grains required 1528.07 s, whereas SmartGrain took 101.84 s, MacGran-IJ 45.4 s, and GrainScan 54.11 s (H = 36.59, p ≤ 0.01). MacGran-IJ is an automated, simple, and efficient method for accurately measuring the size, length, and width of wheat grains using digitized images