Gaps in breeders’ plots (spatial variability) are caused by multiple biotic and abiotic stresses (i.e. poor seed gemination, frost, heat, drought, and/or disease). The gaps are either ignored or visually scored by human inspection. However, this method is time-consuming, labour-intensive, and prone to human error. Low seed availability at the early stages of crossing programs also mean that many treatments/genotypes are un-replicated; hence gaps lead to selection errors. The aim of this study is accurately to estimate the size and distributions of gaps in individual plots using an unmanned aerial vehicle (UAV) and image analysis. A wheat breeding trial was conducted in York, WA, which had 1608 plots with 804 cultivars and 2 phosphorus treatments. Severe gaps in the plots were caused by root disease. Images were collected with by an unmanned aerial vehicle flown over the wheat trial around flowering time on 12th October 2017. The ortho-mosaic is reconstructed using commercial software, then the post-processing was conducted using the integrated platform PhenoCopter including plot segmentation and phenotypic extraction. The ortho-mosaic in each plot was firstly segmented into background and vegetation using classification and regression tree. The gaps were counted by several methods including ground coverage, plot slices into multiple sub-plots, and patch sizes with certain threshold. The new estimated phenotypic value can be used to indicate the gap size in a breeding trial, then in the statistics model of breeding selection as co-variance.