Monitoring of sugar beet growth indicators using wide-dynamic-range vegetation index (WDRVI) derived from UAV multispectral images
Introduction
Sugar beet is an important cash crop in northern China and the second-largest sugar crop in the country. Beet is grown principally in Inner Mongolia, Xinjiang, and Heilongjiang. The beet biomass can reach more than 60 t/ha, thus much greater than those of corn, wheat, and other crops. However, few reports on the use of multispectral platforms to monitor sugar beet growth have appeared (Li et al., 2017a, Li et al., 2017b). With the expansion of beet planting in Inner Mongolia, the establishment of a precise sugar beet management system has become necessary. Efficient non-destructive monitoring of crop growth is necessary for accurate crop management. The Leaf Area Index (LAI), Fresh Weight of Leaves (FWL), and Fresh Weight of Roots (FWR) are important parameters used to monitor beet canopy structure and growth.
In the time since the initial development of remote sensing technology in the 1950s, satellite and manned aircraft platforms, and ground spectral equipment have been used to monitor crop growth. However, certain issues remain in play. Satellite platforms are constrained by both altitude and orbit, and they do not afford the spatial, temporal, or spectral resolution required for growth monitoring (White et al., 2009, Ortiz et al., 2011). Satellite views are constrained by clouds and suspended particles. The operating costs of manned platforms are high; personnel require advanced flight training. Ground spectral equipment is cumbersome; the operating speed is only about 0.56 m/s, and the crop canopy is inevitably damaged. Newer unmanned aerial vehicle (UAV) technology affords cost-effective crop growth monitoring at high spatial, temporal, and spectral resolution. Enciso J A found that coefficient of determination of 0.72 was observed between canopy cover estimated with the UAV and leaf area index measured with the ceptometer (Enciso et al., 2019). Drone sensors are either multispectral or hyperspectral cameras facilitating LAI monitoring, biomass estimation, and yield prediction. Earlier fixed-wing UAVs were fast and covered large areas. ZHANG found that UAV images with high spatiotemporal resolution can be combined with Aqua Crop model to estimate wheat growth (Zhang et al., 2019). The UAV-based technique was able to detect and count citrus trees with high precision (99.9%) in an orchard of 4931 trees and estimate tree canopy size with a high correlation (R = 0.84) with the manual collected data (Arnpatzidis et al., 2019). However, it was difficult to obtain high-quality spectral information, and take-off usually required manual assistance or the use of a stand. Additionally, landings were generally hard, compromising safety. Few crop-monitoring studies employed fixed-wing UAV platforms. Unlike fixed-wing aircraft, multi-rotor UAVs are stable and controllable, fly at low altitudes, are simple to operate, and pose no problems in terms of take-off or landing. Drones now find many applications in agricultural crop growth monitoring. RGB images captured by UAV, and used to estimate the canopy projected area of individual trees, proved to be the best of the options. This was shown by the high correlation (R = 0.85) between this area and the fruit load (Uribeetxebarria et al., 2019). Devia monitored the agronomic parameters of rice using a seven-point index (Devia et al., 2019). Duan collected a great deal of information on rice growth using a UAV vegetation index and spectral characteristics, and successfully predicted the extent of rice heading (Duan et al., 2019). Multispectral sensors on UAV platforms have been successfully used to monitor the yields and above-ground biomasses of grass, wheat, and sunflowers (Elmore et al., 2000, Honda et al., 2014). Liu found that the use of RVI obtained by drone can monitor the aboveground biomass of winter oilseed rape (Liu et al., 2019). Over the past 30 years, vegetation indices have been widely used to evaluate crop canopy structures and growth. The most widely used index is the normalized vegetation index (NDVI); this assesses the biophysical characteristics of canopies. UAV based hyperspectral images linked to a radiative transfer model can provide a promising approach for high throughput monitoring of plant nitrogen (N) status. Results suggest that Multi-LUTs of leaf area index, leaf N density and two spectral indices (MSR and MCARI/MTVI2) in winter wheat demonstrate good performance of canopy N density (CND) estimation (Li et al., 2019). Zhou found that there is a strong relationship between green normalized difference vegetation index (GNDVI), NDVI and yield data for multiple potato varieties (r = 0.77–0.90) (Zhou et al., 2016). The NDVI-derived LAI and leaf clump biomass (Rouse et al., 1974) are widely used to monitor growth, but medium and high biomasses saturate the NDVI (Sellers, 1985, Gitelson et al., 2002). Thus, a wide-range dynamic vegetation index (WDRVI) was developed; the WDRVI performs weighted calculations using NVDI near-infrared band reflectance, reducing the weights of such bands under medium-to-high biomass conditions. This enhances linearity and reduces saturation, facilitating accurate growth monitoring (Huete et al., 2002, Asrar et al., 1984). Here, we used an improved WDRVI to monitor the early-stage growth of sugar beet. We adjusted the weight coefficient of the near-infrared band and identified an optimal vegetation index.
Section snippets
Experimental methods
Field trials were conducted in two growing seasons with nitrogen applied at different rates to several beet varieties (Fig. 1). The trials were conducted in Wulanchabu City (112.442°E, N40.446°N) in the Inner Mongolia Autonomous Region. The details follow. Test 1 was performed in the 2018 growing season; the test beet variety was KWS7156. The sowing date was May 20, 2018; the five nitrogen application rates were 75, 150, 225, 300, and 375 kg/ha. Each plot was 6 m in length and 5 m wide. Line
Changes of sugar beet growth indexes
Comparisons of the growth indices and canopy NDVI growth periods at different nitrogen levels (Fig. 3) revealed that sugar beet produces a great deal of biomass; the fresh weight of leaves and roots can attain more than 100 t/ha. As growth advances, the LAI and FWL both increase and then decrease. The NDVI of the beet canopy is basically the same as those of the two growth indicators. The LAI increased from a seedling age of 40 days (‘rapid growth of leaf clumps 1′) to a maximum at 100 days
Discussion
We found that the NIR level changed significantly at high and medium beet biomasses, but the RED level did not (Fig. 4a). This can be used to enhance the dynamic monitoring range of the NDVI. The WDRVI method enhances the dynamic range of the green atmospheric resistance index (Gitelson et al., 1996). The WDRVI has been used to monitor growth indicators of wheat and soybeans, and corn canopy coverage, LAI, and FWL; linear relationships between the WDRVI and the growth indicators were evident (
Conclusions
In this study, four WDRVI indexes were calculated by added α weight coefficients to NDVI to estimate the FWL, FWR, and LAI of the study area. Sensitivity analysis of 5 indexes (NDVI, WDRVI1, WDRVI2, WDRVI3, WDRVI4) and 3 growth indicators (FWL, FWT and LAI) (Fig. 4d), when NDVI > 0.8, WDRVI index is more sensitive than NDVI. Fig. 5b shows that at 40–60 days of seedling age, WDRVI changes more than NDVI. The 5 indexes and 3 growth indicators of sugar beet in the early growth stage showed a
CRediT authorship contribution statement
Shao ying Zhang: Funding acquisition, Supervision, Conceptualization, Investigation, Methodology, Software, Writing - original draft, Writing - review & editing. Yang Cao: Conceptualization, Investigation, Methodology, Software, Writing - original draft, Writing - review & editing. Guo long Li: Data curation, Formal analysis, Project administration, Resources. Yuan Kai Luo: Data curation, Formal analysis, Project administration, Resources. Qi Pan: Data curation, Formal analysis, Project
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgement
This work was supported by the China Agriculture Research System (Grant No. 170201).
References (36)
- et al.
Validation of agronomic UAV and field measurements for tomato varieties
Comput. Electron. Agric.
(2019) - et al.
Use of green channel in remote sensing of global vegetation from EOS-MODIS
Remote Sens. Environ.
(1996) - et al.
Novel algorithms for remote estimation of vegetation fraction
Remote Sens. Environ.
(2002) - et al.
Overview of the radiometric and biophysical performance of the MODIS vegetation indices
Remote Sens. Environ.
(2002) - et al.
Multi-LUTs method for canopy nitrogen density estimation in winter wheat by field and UAV hyperspectral
Comput. Electron. Agric.
(2019) - et al.
Determination of differences in crop injury from aerial application of glyphosate using vegetation indices
Comput. Electron. Agric.
(2011) - et al.
Remote estimation of nitrogen and chlorophyll contents in maize at leaf and canopy levels
Int. J. Appl. Earth Obs.
(2013) - et al.
Predicting grain yield in rice using multi-temporal vegetation indices from UAV-based multispectral and digital imagery
ISPRS J. Photogramm. Remote Sens.
(2017) - et al.
Aerial multispectral imaging for crop hail damage assessment in potato
Comput. Electron. Agric.
(2016) - et al.
Citrus rootstock evaluation utilizing UAV-based remote sensing and artificial intelligence
Comput. Electron. Agric.
(2019)
Estimating absorbed photosynthetic radiation and leaf area index from spectral reflectance in wheat
Agron J.
QuickBird satellite versus ground-based multi-spectral data for estimating nitrogen status of irrigated maize
Precis. Agric.
Monitoring of cotton canopy coverage based on wide range dynamic vegetation index
Cotton J.
High-throughput biomass estimation in rice crops using UAV multispectral imagery
J. Intell. Rob. Syst.
Remote estimation of rice LAI based on Fourier spectrum texture from UAV image
Plant Methods
Quantifying vegetation change in semiarid environments: precision and accuracy of spectral mixture analysis and the normalized difference vegetation index
Remote Sens. Environ.
Remote estimation of leaf area index and green leaf biomass in maize canopies
Geophys. Res. Lett.
Cited by (52)
Downscaling estimation of NEP in the ecologically-oriented county based on multi-source remote sensing data
2024, Ecological IndicatorsImproving estimation of maize leaf area index by combining of UAV-based multispectral and thermal infrared data: The potential of new texture index
2023, Computers and Electronics in AgricultureUAS-based imaging for prediction of chickpea crop biophysical parameters and yield
2023, Computers and Electronics in AgricultureCitation Excerpt :The problematic aspects of saturation have been documented in studies that used VIs to evaluate biomass and LAI over a wide range of growth stages (e.g., tillering to booting stage in rice) for crops with different morphologies and/or physiological performances (Fu et al. 2014; Gaso et al. 2019). The main insight provided by those studies was that reliability could be ensured only by limiting the use of VIs to specific growth stages (Cao et al. 2020; Xie and Yang 2020). Nonetheless, even though our experiment was limited to short periods (42 and 28 days in 2019 and 2020, respectively; Supplementary Table S-1), the canopy coverage hindered the ability of the VIs to predict changes in crop biophysical parameters.
Unmanned aerial vehicles for agricultural automation
2023, Unmanned Aerial Systems in Agriculture: Eyes Above FieldsDrones in agriculture: A review and bibliometric analysis
2022, Computers and Electronics in AgricultureCitation Excerpt :Accordingly, more research is needed to understand how thermal and multispectral imaging can be integrated with AI techniques (e.g., deep learning) to detect plant stress (Ampatzidis et al., 2020; Ampatzidis & Partel, 2019; Jung et al., 2021; Santesteban et al., 2017; Syeda et al., 2021). Such insights will help ensure more efficient and accurate detection as well as monitoring of plant growth, stress, and phenology (Buters et al., 2019; Cao et al., 2020; Neupane & Baysal-Gurel, 2021; L. Zhou et al., 2020). Cluster 4.