Image Processing for Leaf Recognition Using Neural Network Algorithm

Abstract

Plants are one of the most important and most usable of resources for people in the different fields, so it is important to distinguish between plant species and it called as detection system. So far, this task was carried out by expert botanists which this task is onerous and time-consuming which in addition to, this task has the memory loss and the human error, thus the researchers tried to fix this flaws using artificial intelligence algorithms. For this purpose, we present a method for identifying the leaf and plant species, which includes pre-processing, feature extraction, training, and classification. In this paper, we combine the useful features of all three feature categories (leaf shape, leaf texture, leaf color) and provide a system for identifying a large number of plant species. At the end, the feature vectors will be formed and then we classify them using the feed-forward back-propagation multi-layer perceptron artificial neural network algorithm and compare the results with similar methods in this field. The results are evidence of the proper performance of this system for a large number of species in different conditions, such as pests, seasons changes and lighting.