ABSTRACT
An automated plant species identification system may aid the efficient identification of plants by laymen and botanists. For the extraction of features, deep learning is powerful because it is powerful in presenting richer image knowledge. Introducing the benefits of modern technology to improve agricultural field productivity is necessary with significant concerns over the increasing world population and limited food supplies. There are thousands of species of plants in the natural ecosystem, so identifying between them can be very challenging. Nonetheless, by using the attributes of the leaves, botanists and those who research plants can identify the type of tree at a glance. Machine learning is used to identify forms of leaves automatically. In this paper, an architecture based on convolutional neural network (CNN) has been proposed to distinguish the plant type from the sequences of photographs. The Swedish dataset which has 15 different leaf species has been used to prove the correct identification. The comparison has been done with correct identification rate of other methods, which proves that the proposed architecture outperformed the state-of-the-art methods.
