Rice pests and diseases identification using sift feature

Abstract

The growth and development of crops in agriculture depends on many factors including seed, nutritional status, weather conditions, and etc., in which pests and diseases directly affect the yield of crops and can spread widely. Rice, a major crop of Vietnam, plays an important role of food security and export. Although the pest and disease control procedures for rice are strictly applied, it is still unable to fully control the germ of pests. With a large-scale cultivation mode, the human eye is found very difficult to detect signs of pests in the early stages of development. In this paper, authors propose a model using SIFT (Scale-Invariant Feature Transform) features and SVM (Support Vector Machine) for processing images of rice’s leaves. This model can detect and identify 4 pests in rice: zebra blast, rice blast, leaf rollers, and brown backed hoppers. Experimental results on the model can achieve the accuracy from 80 to 85%.