Development of automatic rice plant pest detection system based on convolutional neural network

Abstract

Rice is the most important food crop in Indonesia, one of the plants that is quite significant and as a daily staple food for the people of Indonesia. The Indonesia government through the ministry of Agriculture tries to maintain the rice production by maintaining of the rice field. One the challenging situation that can involve the rice production is pest. Pests became one of the reasons which can reduce rice production. The decrease in rice production due to pest attack is an important problem in rice plant care. In this work, a pest detection system in rice plants developed using an intelligent system technique. The system involved image processing and intelligent technique. The system recognizes the kind of pests of the rice plant based on the feature of the image of the pets. Rice plant pest detection systems based on the pest image proceed by image procession technique and Convolution Neural Network (CNN). The system is working properly, since it resulting the training and testing accuracy of 99% and 90%, respectively.