Training datasets of Japanese barberry and multiflora rose

<p>This dataset of two invasive non-native shrub species, Japanese barberry (<em>Berberis thunbergii</em>) and multiflora rose (<em>Rosa multiflora</em>)&nbsp;was used to test&nbsp;a method of detection with&nbsp;aerial imagery and convolutional neural networks. All aerial imagery was captured at the low altitude of 40 m above ground level using an unpiloted aerial vehicle over deciduous forest stands from late March through early May in northeast Connecticut, USA.&nbsp;The training dataset consists&nbsp;of 1,855 images of size 224 x 224 x 3 pixels in the red&ndash;green&ndash;blue (RGB) spectrum.</p>