Plant Phenomics / 2021 / Article / Fig 6

Research Article

Qualification of Soybean Responses to Flooding Stress Using UAV-Based Imagery and Deep Learning

Figure 6

The comparisons of NDVI measured at three flight heights. (a) The boxplot of NDVI of all soybean plots collected at three flight heights. The end of the boxes defines the 1st and 3rd quartile, with a plot at the median and error bars defining the 10th and 90th percentiles. The lowercase letters above bars indicate the significant difference among these means at a 0.05 significance level. (b) The false-color images of an example plot (Figure 5(d)) in NDVI. (c) FIS map of the experiment field. Each grid represents the FIS of a single soybean line. (d–f) NDVI maps at 20, 50, and 80 m with each grid representing an averaged NDVI value of each plot. Soybean plots at 50 m were segmented from four images. The polygons in (e) mark the four images that the soybean plots were segmented from and the images were taken from the bottom left, top left, bottom right, and top right corner. (g) Four pairs of boxplots with each pair comparing the difference of ten points screened at two different corners. The rectangles in (e) display the regions that the points were randomly selected from.