OBJECTIVE PHENOTYPING THE TIME OF BUD BURST BY ANALYZING GRAPEVINE FIELD IMAGES

Abstract

In grapevine breeding programs and other studies, the acquisition of phenology is usually performed directly in the vineyard, mainly by visual estimation. In order to acquire the time of bud burst, the BBCH scale is applied so that the person in charge estimates the BBCH stage for each genotype over a time period of at least two weeks. Phenotyping is thus strongly limited by time, cost, and the subjectivity of records. In the present study an efficient image based phenotyping method was developed and validated. Based on the sum of degree-day, image acquisition was carried out at only one single time point directly in the vineyard, and image analysis was conducted by using Matlab®. For each grapevine image a weighted length Lw (ratio of counted buds, number of shoots and average shoot length) was calculated. Lw was compared with reference evaluation data and significant correlations of r=-0.87 and r=0.76 were obtained. The range of Lw was used for classifying the trait ?time of bud burst? into an objective five-class system from very early to very late genotypes. The classification method was applied in order to rate the trait within a mapping population in comparison to traditional cultivars used as references. It was observed that the time of bud burst distributed into all five classes.