Image-based tracking of ripening in wheat cultivar mixtures: a quantifying approach parallel to the conventional phenology

Abstract

The lack of quantitative methods independent of the conventional qualitative phenology, may be a vital limiting factor to evaluate the temporal trends in the crop growth cycle, particularly in the heterogeneous canopies of cultivar mixtures. A digital camera used to take ground-based nadir images during two years of a field experiment conducted at the College of Agriculture, Shiraz University, Iran; in 2014-15 and 2015-16. The experimental treatments consisted of 4 early- to middle-ripening wheat cultivars and their 10 mixtures, under post-anthesis well- and deficit-irrigation conditions, arranged in a randomized complete block design with 3 replicates. Then the images were processed and three image-derived indices including CC (canopy cover), GR [(G-R/G); RGB color system], and CCGR (CCxGR) were used as the quantifying criteria. The declining trends of these indices during ripening showed strong fits to binomial equations, based on which simple prediction models were suggested and validated. Furthermore, the split linear trends and their slopes were estimated to assess the short-term variations. Some agronomic aspects were also evidenced using the mixtures-monoculture diversions, and the relationship between CC and GR. The frameworks evaluated appears to provide the reliable and simple solutions for quantifying the crop temporal trends parallel to the conventional phenology.