Citation

Muliady, Muliady and Lim, Tien Sze and Koo, Voon Chet Machine Learning Based Algorithm for Rice Leaf Nitrogen Status Estimation. In: 2nd FET PG Engineering Colloquium Proceedings 2023, 1-31 December 2023, Multimedia University, Malaysia. (Submitted)

Other (Poster) 9. PPM Poster - Muliady - Lim Chee Siong.pdf - Submitted Version Restricted to Repository staff only Download (337kB)

Abstract

In order to implement site-specific nutrient management for rice plants, farmers must assess the Nitrogen levels in the plants. The objective of this research is to create a machine-learning algorithm that can predict the Nitrogen status of rice leaves using multispectral images from a single sensor. The dataset includes information from 49 rice plants in pots and 33 rice plants in the field. Two cameras, a Red Edge and an Orange Cyan Near Infrared (OCN) MAPIR Survey3, were utilized along with a custom light intensity device built from a GY1145 light sensor and ESP32 microcontroller. A total of 306 data pairs were collected, incorporating features such as NDVI value, RE value, visible light intensity, infrared intensity, and time, with SPAD value as the output. The Support Vector Regression (SVR) algorithm demonstrated optimal performance with a Radial Basis Functions kernel, gamma value of 6, and epsilon value of 0.1. The training data achieved a performance score of 0.94, while the validation data performance score was 0.85. Testing the model with a separate dataset resulted in an average error of 3.72%, indicating that the model can accurately estimate the Nitrogen status

Item Type:	Conference or Workshop Item (Poster)
Uncontrolled Keywords:	Machine leaning
Subjects:	Q Science > Q Science (General) > Q300-390 Cybernetics
Divisions:	Faculty of Engineering and Technology (FET)
Depositing User:	Ms Rosnani Abd Wahab
Date Deposited:	03 Apr 2024 01:17
Last Modified:	03 Apr 2024 01:17
URII:	http://shdl.mmu.edu.my/id/eprint/12342

Downloads

Edit (login required)