Prediction of soil nitrogen content using E-nose and radial basis function

• Main Authors: Jigme Norbu, Theerapat Pobkrut, Treenet Thepudom, Thinley Namgyel, Teerayut Chaiyasit, Yu Thazin, Teerakiat Kerdcharoen
• 其他作者: Mahidol University
• 格式: Conference or Workshop Item
• 出版: 2020
• 主題: Computer Science; Engineering; Mathematics
• 在線閱讀: https://repository.li.mahidol.ac.th/handle/123456789/50656

© 2018 IEEE Existing soil nutrient determining methods are still a concern as most of them entail arduous field sampling followed by rigorous testing procedures, both of which are time consuming and expensive. Conversely, farmers require cheap and instant information about the soil nutrient management for quick decision making or they will have to risk their crop. Electronic nose (E-nose) is an emerging technology that has potential application in monitoring of soil nutrient abundance. In this work, e-nose coupled with Radial Basis Function (RBF) is employed to determine the amount of nitrogen (N) which is one of the main nutrients in the soil. The results demonstrate that not only does the e-nose clearly discriminate the odors of soil with different N concentration, but also can evidently predict the total N content with accuracy of 96.2% using RBF. Hence, e-nose and RBF network could be a promising alternative to conventional soil testing methods.