Estimating Plant Nitrogen Concentration of Rice through Fusing Vegetation Indices and Color Moments Derived from UAV-RGB Images

Haixiao Ge,Zhengzheng Tan,Changwen Du,Fei Ma,Zhenwang Li,Haitao Xiang,Zhengchao Qiu

doi:10.3390/rs13091620

Abstract

Estimating plant nitrogen concentration (PNC) has been conducted using vegetation indices (VIs) from UAV-based imagery, but color features have been rarely considered as additional variables. In this study, the VIs and color moments (color feature) were calculated from UAV-based RGB images, then partial least square regression (PLSR) and random forest regression (RF) models were established to estimate PNC through fusing VIs and color moments. The results demonstrated that the fusion of VIs and color moments as inputs yielded higher accuracies of PNC estimation compared to VIs or color moments as input; the RF models based on the combination of VIs and color moments (R2 ranging from 0.69 to 0.91 and NRMSE ranging from 0.07 to 0.13) showed similar performances to the PLSR models (R2 ranging from 0.68 to 0.87 and NRMSE ranging from 0.10 to 0.29); Among the top five important variables in the RF models, there was at least one variable which belonged to the color moments in different datasets, indicating the significant contribution of color moments in improving PNC estimation accuracy. This revealed the great potential of combination of RGB-VIs and color moments for the estimation of rice PNC.

Highlights

Rice (Oryza sativa L.) is one of the most important crops in the world, feeding more than half of the world’s population [1]
Maimaitijiang et al [48] compared the performance of partial least square regression (PLSR), random forest (RF), extreme learning regression (ELR), and support vector regression (SVR) in estimating crop leaf N concentration (LNC), using satellite-based vegetation indices (VIs) and Unmanned aerial vehicle (UAV)-based canopy structure information as inputs
Liang et al [50] concluded that the RF model was preferred to predict LNC compared to the least square support vector model (LS-SVR)

Summary

Introduction

Rice (Oryza sativa L.) is one of the most important crops in the world, feeding more than half of the world’s population [1]. And accurate assessment of PNC to detect N excess or deficiency is essential for farmers to improve rice production and N use efficiency [3,4]. Laboratory analysis is one of the important ways to obtain crop N nutrition status. It is time-consuming and laborious to carry out field investigations and collect representative samples, and usually the obtained results are delayed [5]. Compared with traditional laboratory analysis (e.g., micro Kjeldahl method), the non-destructive and timely methods or tools to detect crop N status have been significantly increased over recent decades [6,7,8]. Remote sensing with the advantages of fast and non-destructive characterizations has been proved useful for acquiring related information of crop nutrition status [9]

Methods

Discussion

Conclusion