Predicting Munsell color for turfgrass leaves

Abstract

AbstractLinking turfgrass color to hue, value, chroma (H V/C) in the Munsell Plant Tissue Color Book is a visual comparison process for specifying and communicating plant color. If subjectivity of visual comparison can be mitigated, then the accuracy of color matching may be improved. Research was conducted to develop an algorithm predicting H V/C from CIE‐xyY color (xyY) in digital images of leaves of four turfgrasses. First, value‐chroma (V/C) arrays for Munsell hue groups 5Y, 2.5GY, 5GY, 7.5GY, 10GY, and 2.5G were converted to xyY. Next, chromaticity (xy) plots from each array were fitted with 95% prediction bands. Then, xy points for leaves (N = 60) for each turfgrass were plotted versus prediction bands. The prediction interval containing the most leaf xy points specified leaf H. Pairing each V from each hue group (N = 98) with its corresponding Y regressing log10V on log10Y predicted leaf V from leaf Y (log10V = 0.0715 + 0.5303logY – 0.0332logY2). Pairing each C (N = 98) with its corresponding xy predicted leaf C from leaf xy (C = −11.6021 − 11.2908x + 48.0880y). Predicted H V/C for “Champion” and “TifEagle” hybrid bermudagrass [Cynodon dactylon (L.) Pers. var. dactylon × Cynodon transvaalensis Burtt‐Davy] was 10GY 8/4 and 7.5GY 8/4 mirroring H V/C predicted from xyY using open‐source algorithms in R and Python. The experimental algorithm predicted H V/C accurately, and objectively advancing the way turfgrass color is specified and communicated. Specifying turfgrass color in situ could be possible by interfacing the algorithm with smartphone technology.