Optimization of the "Germinator" as a complement for the analysis of seed germination quality in rice (Oryza sativa L.)

Introduction. The use of digital image analysis with the "Germinator" allows the automatic evaluation of seed germination. In addition to maximum germination, other parameters associated with seed germination vigor can be quantified simultaneously. Objective. To optimize the "Germinator" as a complement for the automatic analysis of seed germination quality in rice. Materials and methods. The experiments were conducted at the Centro para Investigaciones en Granos y Semillas (CIGRAS) from 2015 to 2018. The "Germinator" software package was optimized on the Palmar 18 variety, and then tested with a panel of 126 rice seed samples comprising fourteen varieties. The germination curves were quantified, the automatically acquired data were compared with manual counting based on radical protrusion and a standardized method. Besides, accelerated aging and controlled aging experiments were conducted to show the sensibility of the automatic method. Results. The relationship between germination quantified automatically and quantified manually was high (R2= 0.99). Maximum germination, quantified by image analysis, ranged from 69 % to 100 % in the panel of 126 samples. The correlation between germination assessed by digital images and the standard method was rho (spearman)= 0.34. The "Germinator" allowed the simultaneous quantification of other variables associated with seed vigor, such as the t50 parameter, which is the time the seed lot takes to reach 50 % of the germination. In addition, the automatic method revealed the differential effect of two seed aging protocols. Conclusions. The use of digital image analysis made it possible to evaluate automatically seed germination based on radicle protrusion and also made it possible to quantify other complementary variables associated with seed vigor (t50).