NITROGEN ESTIMATION SYSTEM IN LETTUCE USING MULTISPECTRAL CAMERA ON EDGE DEVICE

Authors

  • Quang Hung Ha VNU University of Engineering and Technology
  • Truong Trinh Phung VNU University of Engineering and Technology
  • Hoang Vu Le VNU University of Engineering and Technology
  • Minh-Trung Vu VNU University of Engineering and Technology
  • Trong-Minh Hoang Posts and Telecommunications Institute of Technology
  • Minh Trien Pham VNU University of Engineering and Technology

Keywords:

Leaf Nitrogen Concentration, Machine Learning, Multispectral Image, Edge Device, All-optical dot product, image processing, multimode interference coupler, optical signal processing, Lettuce

Abstract

Farmers can improve their decision-making by accurately diagnosing nutrient deficiencies, resulting in more efficient fertilizer use and reducing the environmental impact of over-fertilization. This study presents an automated system detecting nitrogen stress in lettuce through multispectral imaging, running on an edge device. The system estimates nitrogen levels in leaves using a machine learning algorithm, which is calibrated against reference measurements taken in the field, achieving a mean squared error (MSE) of 0.4 between the estimated and actual values. Based on threshold values determined through ground-truth experiments, plant health data is transmitted to a cloud database, which can be accessed via a web or desktop application. The proposed method guarantees efficient monitoring and regulation of nitrogen levels in crops.

Downloads

Published

2025-09-15

Issue

Vol. 1 No. 4 (2024): Journal of Science and Technology on Information and Communications

Section

Electronics and Telecommunications