Plant Leaf Image Recognition Based on Convolutional Neural Network

Abstract

The objective of this research is to develop a computer system which can recognize Thai leaf images by using convolutional neural network. The developed system is called “plant leaf image recognition system (PLIRS)”. The system consists of four modules, namely: 1) image acquisition, 2) training dataset, 3) image recognition and 4) result presentation module. The system dataset consists of 10,800 leaf-images of 54 Thai plant-leaf species. The average time to train the dataset is 1.4357x104 seconds. The precision rate of the system is 96.99 percent with the average access time of 1.3649 second/image.