Rice plant nitrogen level assessment through image processing using artificial neural network

Abstract

This paper presents a program which identifies the 4-panel LCC equivalent of rice plants using image processing techniques and pattern recognition of the Backpropagation neural network. Images of the fully expanded healthy leaves were captured by digital camera and processed through RGB acquisition, color transformation, image enhancement, image segmentation and feature extraction procedures. The extracted features were computed using basic statistical methods, then served as the input to the neural network for LCC panel identification. Thirty (30) samples of IRR 82372H - Mestiso 26 variety were tested; divided into three sets with 10 leaf samples per field. The system was observed to provide an accuracy of 93.33%.