Resumen

SAUCEDA ACOSTA, Carlos Patricio et al. Size and number of wheat grains analyzed by digital image processing. Rev. Mex. Cienc. Agríc [online]. 2017, vol.8, n.3, pp.517-529. ISSN 2007-0934.  https://doi.org/10.29312/remexca.v8i3.28.

The visual record of grain number and size uses limited samples, causing fatigue and increase of error; disadvantages that can be solved with the processing and analysis of digital images (PAID) as proposed in this paper. In this regard, seeds of five durum wheat cultivars (Triticum durum Desf.) and five of common wheat (Triticum aestivum L.) were used. Scanned images were obtained. The PAID was performed with ImageJ 1.49b, length, width, size and number of grains were measured in 317 samples. The time for grain registration was compared between visual counting, PAID and electromechanical counter. Correlation analyzes, X2, ANDEVA, and comparison of means were performed. The grain size of common wheat was from 15.41 to 18.37 mm2, in durum it went from 17.82 to 19.6 mm2 (DMS = 0.385, p≤ 0.01), these were longer (>7.1 mm; p≤ 0.01) but thinner (DMS = 0.043, p≤ 0.05). The number of grains observed and calculated showed concordance (X2= 2.73, p= 1) and close correlation (rs=1, p≤ 0.01). The counter and visual record of grains number showed linear increase of time as the sample size increased, while the PAID required approximately 55 s. Image processing facilitates the individual dimensions of pooled grains and is an accurate and fast tool for analyzing the size and number of wheat grains.

Palabras llave : Triticum aestivum; Triticum durum; grain counting; ImageJ; scanner.