Copula Based Bird Cherry Plant Identification

The European Bird Cherry plant (scientifically known as Prunus Padus L.) is a non-native deciduous tree that is quickly displacing native trees in many regions of northern US and the State of Alaska. The need for its eradication is urgent and critical. In this paper, we introduce a copula-based method for Bird Cherry plant identification. This method uses leaf rich and contextual features and copulas ensemble to proficiently identify the plant. To evaluate its performance, we experimented on a large dataset of 2,000 leaf images of bird cherry leaves, sour cherry leaves and red oak leaves. We compared our method performance results with published deep learning methods and found that our results are comparable or better. Finally, this method is fast, and has low computational complexity which makes it an ideal solution for binary plant identification especially for the sought-after medicinal plants. This method can be converted to an application for consumer technology devices such as smartphones and hand-held devices.