Combination of Local Descriptors and Global Features for Leaf Recognition
Maliheh Shabanzade, Morteza Zahedi, Seyyed Amin Aghvami · Signal & Image Processing An International Journal · 2011
Automatic leaf recognition system is a case coming to improve time-consuming and troublesome tasks which have mainly been carried out by botanists manually.This application as judged by common characteristics is popular in institutes for discovering new plant species, modernizing the management of botanical gardens and horticulture fields.In order to conduct a leaf recognition system, the features must be sufficiently distinctive to identify specific objects among many alternatives, where contain both local and global properties.So far, many researchers have represented some techniques which use local or global features only where face problems, such as many images are captured in different intensity, they are maybe sick or calamity, leaves have been damaged or cropped and so on.In this paper, a new method for leaf recognition system is proposed where both local descriptors and global features are employed, combined and finally the most discriminant features are selected by employing a linear discriminant analysis method.The experimental results show that using the feature vector containing the local features and global characteristics leads us to obtain 94.3% recognition rate.