Akademik Veri Yönetim Sistemi
Maydica, cilt.67, sa.3, 2025 (SCI-Expanded)
The protein content and quality in maize significantly influence grain quality, driving global efforts to develop high-protein-quality genotypes. Opacity serves as a key phenotypic selection criterion in these efforts due to its relationship with essential amino acid content. This study investigates the differentiation of opaque maize kernels using computer-aided software and explores the relationship between opacity levels and color spaces (RGB, HSV, Lab). Seed samples from 10 maize genotypes (1000 seeds) with varying opacity levels were imaged on a light table in embryo-up and embryo-down orientations. Particle analysis and thresholding performed in R determined opacity levels and provided numerical data for RGB, HSV, and Lab color spaces. Protein, lysine, and tryptophan contents were analyzed through reference methods. Correlation and regression analyses assessed relationships between opacity levels (visual and image-processed) and biochemical components, and color space channels. Protein content ranged from 6.66% to 11.62%, lysine from 0.266% to 0.450%, and tryptophan from 0.034% to 0.092% among opacity groups. Relationships between visual and image-processed opacity levels showed R² = 0.57 (embryo-up) and R² = 0.65 (embryo-down). Notably, channels of the HSV color space correlated with lysine and tryptophan contents. This study demonstrates that image processing effectively evaluates opacity levels and protein quality in maize using color space data, offering a promising tool for phenotypic selection.