A Novel Method for Leaf Identification using Machine Learning

Plants play a very important role for sustenance of life on earth. The leaf of a plant is one of the significant factors to identify various plant characteristics like their species, health and even the climate. The growing interest in biodiversity and the increasing availability of digital images help us to review and analyze plant leaf. We use features like leaf outlines, shape, vein structures and textures and 
use wide range of analytical methods to describe the mentioned characteristics. We use image processing techniques for a set of training images to generate the feature vectors. Once the machine is trained, we apply test images for feature detection and classification of the plant species. Application of this method can be used  in  prototypes  of  hand-held  digital  field  guides  and  various  robotic  systems in agriculture. We conclude with a discussion of ongoing work and challenging problems in this area