IMAGE CLASSIFICATION FOR PLANT STRESS MONITORING USING UAV BASED ON RBF NEURAL NETWORK

Abstract

Agricultural products that are very important in the lives of Indonesian people include rice. However, one of the challenges in agricultural development is an increasing downward trend. One effort to help him is information that can be used to collect data about the description of the physical condition of plants (high level or density) as assistance data to obtain agricultural products. The purpose of this study is to monitor the stress level of rice plants and calculate rice productivity based on varieties using image data processing based on classification techniques. The stages of this research start from image data collection and field measurements using drones, then image data will be processed using Matlab Software. Measurements made on several samples that have different variations using the RBF Neural Network classification technique. This technique is used to assess whether the rice plant is in good condition or not. Then do the data validation between the two measurements so that it can be obtained. In addition, measurements were also made using test data that functioned as an analysis of the estimated value of the measurement of the physical condition of plants. In fact, from the paddy’s healthy condition, we can predict the harvest of paddy field quickly. In this research have tested some areas which have been made ground truth and compare the harvest result from manually and automatically method, then the accuracy of this prediction is 90%.