Comparison of Machine Learning Methods for Leaf Nitrogen Estimation in Corn Using Multispectral UAV Images

2021 ◽  
Vol 64 (6) ◽  
pp. 2089-2101
Author(s):  
Razieh Barzin ◽  
Hamid Kamangir ◽  
Ganesh C. Bora
Keyword(s):  
Machine Learning ◽  
Random Forest ◽  
Vegetation Indices ◽  
Leaf Nitrogen ◽  
Gradient Boosting ◽  
N Content ◽  
Spectral Bands ◽  
Red Edge ◽  
Leaf N Content ◽  
Leaf N

HighlightsLeaf nitrogen percentage in corn was estimated using various vegetation indices derived from UAVs.Eight machine learning methods were compared to find the most accurate model for nitrogen estimation.The most influential vegetation indices were determined for estimation of leaf nitrogen.Abstract. Nitrogen (N) is the most critical component of healthy plants. It has a significant impact on photosynthesis and plant reproduction. Physicochemical characteristics of plants such as leaf N content can be estimated spatially and temporally because of the latest developments in multispectral sensing technology and machine learning (ML) methods. The objective of this study was to use spectral data for leaf N estimation in corn to compare different ML models and find the best-fitted model. Moreover, the performance of vegetation indices (VIs) and spectral wavelengths were compared individually and collectively to determine if combinations of VIs substantially improved the results as compared to the original spectral data. This study was conducted at a Mississippi State University corn field that was divided into 16 plots with four different N treatments (0, 90, 180, and 270 kg ha-1). The bare soil pixels were removed from the multispectral images, and 26 VIs were calculated based on five spectral bands: blue, green, red, red-edge, and near-infrared (NIR). The 26 VIs and five spectral bands obtained from a red-edge multispectral sensor mounted on an unmanned aerial vehicle (UAV) were analyzed to develop ML models for leaf %N estimation of corn. The input variables used in these models had the most impact on chlorophyll and N content and high correlation with leaf N content. Eight ML algorithms (random forest, gradient boosting, support vector machine, multi-layer perceptron, ridge regression, lasso regression, and elastic net) were applied to three different categories of variables. The results show that gradient boosting and random forest were the best-fitted models to estimate leaf %N, with about an 80% coefficient of determination for the different categories of variables. Moreover, adding VIs to the spectral bands improved the results. The combination of SCCCI, NDRE, and red-edge had the largest coefficient of determination (R2) in comparison to the other categories of variables used to predict leaf %N content in corn. Keywords: Corn, Gradient boosting, Machine learning, Multispectral imagery, Nitrogen estimation, Random forest, UAV, Vegetation index.

scholarly journals Dissection of hyperspectral reflectance to estimate nitrogen and chlorophyll contents in tea leaves based on machine learning algorithms

2020 ◽  
Vol 10 (1) ◽  
Author(s):  
Hiroto Yamashita ◽  
Rei Sonobe ◽  
Yuhei Hirono ◽  
Akio Morita ◽  
Takashi Ikka
Keyword(s):  
Machine Learning ◽  
Learning Algorithms ◽  
Visible Region ◽  
Machine Learning Algorithms ◽  
Hyperspectral Reflectance ◽  
N Content ◽  
Green Leaves ◽  
Leaf N Content ◽  
N Status ◽  
Leaf N

Abstract Nondestructive techniques for estimating nitrogen (N) status are essential tools for optimizing N fertilization input and reducing the environmental impact of agricultural N management, especially in green tea cultivation, which is notably problematic. Previously, hyperspectral indices for chlorophyll (Chl) estimation, namely a green peak and red edge in the visible region, have been identified and used for N estimation because leaf N content closely related to Chl content in green leaves. Herein, datasets of N and Chl contents, and visible and near-infrared hyperspectral reflectance, derived from green leaves under various N nutrient conditions and albino yellow leaves were obtained. A regression model was then constructed using several machine learning algorithms and preprocessing techniques. Machine learning algorithms achieved high-performance models for N and Chl content, ensuring an accuracy threshold of 1.4 or 2.0 based on the ratio of performance to deviation values. Data-based sensitivity analysis through integration of the green and yellow leaves datasets identified clear differences in reflectance to estimate N and Chl contents, especially at 1325–1575 nm, suggesting an N content-specific region. These findings will enable the nondestructive estimation of leaf N content in tea plants and contribute advanced indices for nondestructive tracking of N status in crops.

scholarly journals Continuous Monitoring of Cotton Stem Water Potential using Sentinel-2 Imagery

2020 ◽  
Vol 12 (7) ◽  
pp. 1176 ◽  
Author(s):  
Yukun Lin ◽  
Zhe Zhu ◽  
Wenxuan Guo ◽  
Yazhou Sun ◽  
Xiaoyuan Yang ◽  
...  
Keyword(s):  
Machine Learning ◽  
Linear Regression ◽  
Water Potential ◽  
Large Scale ◽  
Temporal Frequency ◽  
Vegetation Indices ◽  
Stem Water Potential ◽  
Spectral Bands ◽  
Red Edge ◽  
Sentinel 2

Monitoring cotton status during the growing season is critical in increasing production efficiency. The water status in cotton is a key factor for yield and cotton quality. Stem water potential (SWP) is a precise indicator for assessing cotton water status. Satellite remote sensing is an effective approach for monitoring cotton growth at a large scale. The aim of this study is to estimate cotton water stress at a high temporal frequency and at a large scale. In this study, we measured midday SWP samples according to the acquisition dates of Sentinel-2 images and used them to build linear-regression-based and machine-learning-based models to estimate cotton water stress during the growing season (June to August, 2018). For the linear-regression-based method, we estimated SWP based on different Sentinel-2 spectral bands and vegetation indices, where the normalized difference index 45 (NDI45) achieved the best performance (R2 = 0.6269; RMSE = 3.6802 (-1*swp (bars))). For the machine-learning-based method, we used random forest regression to estimate SWP and received even better results (R2 = 0.6709; RMSE = 3.3742 (-1*swp (bars))). To find the best selection of input variables for the machine-learning-based approach, we tried three different data input datasets, including (1) 9 original spectral bands (e.g., blue, green, red, red edge, near infrared (NIR), and shortwave infrared (SWIR)), (2) 21 vegetation indices, and (3) a combination of original Sentinel-2 spectral bands and vegetation indices. The highest accuracy was achieved when only the original spectral bands were used. We also found the SWIR and red edge band were the most important spectral bands, and the vegetation indices based on red edge and NIR bands were particularly helpful. Finally, we applied the best approach for the linear-regression-based and the machine-learning-based methods to generate cotton water potential maps at a large scale and high temporal frequency. Results suggests that the methods developed here has the potential for continuous monitoring of SWP at large scales and the machine-learning-based method is preferred.

scholarly journals Relationship between vegetation indices and agronomic performance of maize varieties under different nitrogen rates

2020 ◽  
Vol 36 (5) ◽  
Author(s):  
Marcela da Silva Flores ◽  
Willian Meniti Paschoalete ◽  
Fabio Henrique Rojo Baio ◽  
Cid Naudi Silva Campos ◽  
Ariane de Andréa Pantaleão ◽  
...  
Keyword(s):  
Plant Height ◽  
Precision Agriculture ◽  
Vegetation Index ◽  
Block Design ◽  
Vegetation Indices ◽  
Agronomic Performance ◽  
N Content ◽  
Maize Varieties ◽  
Leaf N Content ◽  
Leaf N

Precision agriculture is a set of techniques that assist the monitoring of the agronomic performance of the maize crop by using vegetation indices. This study aimed to verify the relationship between vegetation indices, plant height, leaf N content, and grain yield of three maize varieties, grown under high and low N as topdressing. The experiment was carried out at the Fundação de Apoio à Pesquisa Agropecuária de Chapadão (Fundação Chapadão), located in the municipality of Chapadão do Sul, during the 2017/2018 season. The experiment consisted of a randomized block design with four replications, arranged in a 3x2 split-plot scheme. The first factor (plots) corresponded to three open-pollinated maize varieties (BRS 4103, BRS Gorotuba, and SCS 154), and the second factor (subplots) consisted of two N rates applied as topdressing (80 and 160 kg- 1). All the evaluated variables showed varieties x N interaction. Vegetation indices in maize varieties were influenced by the N rate applied as topdressing. Normalized Difference Vegetation Index (NDVI) and Soil-adjusted Vegetation Index (SAVI) showed a higher correlation with plant height. At the same time, Normalized Difference Red Edge (NDRE) had a stronger association with leaf N content.

scholarly journals Whole-plant optimality predicts changes in leaf nitrogen under variable CO2 and nutrient availability

2019 ◽  
Author(s):  
Silvia Caldararu ◽  
Tea Thum ◽  
Lin Yu ◽  
Sönke Zaehle
Keyword(s):  
Plant Growth ◽  
Nutrient Limitation ◽  
Leaf Nitrogen ◽  
List Type ◽  
Carbon Export ◽  
N Content ◽  
Foliar N ◽  
Whole Plant ◽  
Leaf N Content ◽  
Leaf N

SummaryVegetation nutrient limitation is essential for understanding ecosystem responses to global change. In particular, leaf nitrogen (N) is known to be plastic under changed nutrient limitation. However, models can often not capture these observed changes, leading to erroneous predictions of whole-ecosystem stocks and fluxes.We hypothesise that an optimality approach can improve representation of leaf N content compared to existing empirical approaches. Unlike previous optimality-based approaches, which adjust foliar N concentrations based on canopy carbon export, we use a maximisation criteria based on whole-plant growth and allow for a lagged response of foliar N to this maximisation criterion to account for the limited plasticity of this plant trait. We test these model variants at a range of Free-Air CO2 Enrichment (FACE) and N fertilisation experimental sites.We show a model solely based on canopy carbon export fails to reproduce observed patterns and predicts decreasing leaf N content with increased N availability. However, an optimal model which maximises total plant growth can correctly reproduce the observed patterns.The optimality model we present here is a whole-plant approach which reproduces biologically realistic changes in leaf N and can thereby improve ecosystem-level predictions under transient conditions.

Monitoring crop N status by using red edge-based indices

2017 ◽  
Vol 8 (2) ◽  
pp. 338-342 ◽  
Author(s):  
J. González-Piqueras ◽  
H. Lopez-Corcoles ◽  
S. Sánchez ◽  
J. Villodre ◽  
V. Bodas ◽  
...  
Keyword(s):  
Vegetation Indices ◽  
Growth Cycle ◽  
N Content ◽  
Spectral Bands ◽  
Red Edge ◽  
Use Efficiency ◽  
Edge Based ◽  
Linear Correlations ◽  
N Status ◽  
Sentinel 2

Intensive agriculture has the objective to increase nutrients use efficiency. Nitrogen (N) is a key nutrient for crops and the estimations of crop N status allow adjusting the fertilization levels to crop requirements, while reducing the environmental costs and optimizing the benefits for farmers. In this work the N status of wheat in a commercial plot has been monitored, varying the N supply taking into account the variability of the soil. The N content in the cover has been monitored simultaneously by sampling at field level and by using vegetation indices based on reflectance in the red-edge band. The results of the field campaign along a crop growth cycle show that the REP, MTCI, AIVI and CCCI calculated from narrow spectral bands show good linear correlations (R2>0.93) with respect to N content (g·m−2). These indices are stable when passing to broad bands as the case of Sentinel 2 with R2>0.9.

Heterogeneity of photosynthesis within leaves is associated with alteration of leaf structural features and leaf N content per leaf area in rice

2015 ◽  
Vol 42 (7) ◽  
pp. 687 ◽  
Author(s):  
Dongliang Xiong ◽  
Tingting Yu ◽  
Xi Liu ◽  
Yong Li ◽  
Shaobing Peng ◽  
...  
Keyword(s):  
Crop Yields ◽  
Biomass Accumulation ◽  
Structural Features ◽  
Leaf Nitrogen ◽  
Gas Exchange Parameters ◽  
N Content ◽  
Chlorophyll Contents ◽  
Functional Changes ◽  
Leaf N Content ◽  
Leaf N

Increasing leaf photosynthesis rate (A) is considered an important strategy to increase C3 crop yields. Leaf A is usually represented by point measurements, but A varies within each leaf, especially within large leaves. However, little is known about the effect of heterogeneity of A within leaves on rice performance. Here we investigated the changes in gas-exchange parameters and leaf structural and chemical features along leaf blades in two rice cultivars. Stomatal and mesophyll conductance as well as leaf nitrogen (N), Rubisco and chlorophyll contents increased from base to apex; consequently, A increased along leaves in both cultivars. The variation in A, leaf N content and Rubisco content within leaves was similar to the variations among cultivars, and the extent of A heterogeneity within leaves varied between cultivars, leading to different efficiencies of biomass accumulation. Furthermore, variation of A within leaves was closely associated with leaf structural and chemical features. Our findings emphasise that functional changes along leaf blades are associated with structural and chemical trait variation and that variation of A within leaves should be considered to achieve progress in future breeding programs.

scholarly journals Effect of Cultivar on Chlorophyll Meter and Canopy Reflectance Measurements in Cucumber

Sensors ◽  
2020 ◽  
Vol 20 (2) ◽  
pp. 509 ◽  
Author(s):  
Romina de Souza ◽  
Rafael Grasso ◽  
M. Teresa Peña-Fleitas ◽  
Marisa Gallardo ◽  
Rodney B. Thompson ◽  
...  
Keyword(s):  
Optical Sensors ◽  
Optical Sensor ◽  
Chlorophyll Concentration ◽  
Vegetation Indices ◽  
Canopy Reflectance ◽  
Cucumis Sativus L ◽  
N Content ◽  
Chlorophyll Meter ◽  
Leaf N Content ◽  
Leaf N

Optical sensors can be used to assess crop N status to assist with N fertilizer management. Differences between cultivars may affect optical sensor measurement. Cultivar effects on measurements made with the SPAD-502 (Soil Plant Analysis Development) meter and the MC-100 (Chlorophyll Concentration Meter), and of several vegetation indices measured with the Crop Circle ACS470 canopy reflectance sensor, were assessed. A cucumber (Cucumis sativus L.) crop was grown in a greenhouse, with three cultivars. Each cultivar received three N treatments, of increasing N concentration, being deficient (N1), sufficient (N2) and excessive (N3). There were significant differences between cultivars in the measurements made with both chlorophyll meters, particularly when N supply was sufficient and excessive (N2 and N3 treatments, respectively). There were no consistent differences between cultivars in vegetation indices. Optical sensor measurements were strongly linearly related to leaf N content in each of the three cultivars. The lack of a consistent effect of cultivar on the relationship with leaf N content suggests that a unique equation to estimate leaf N content from vegetation indices can be applied to all three cultivars. Results of chlorophyll meter measurements suggest that care should be taken when using sufficiency values, determined for a particular cultivar

PREDIKSI KANDUNGAN NITROGEN DAUN PADI DENGAN ANALISIS PERGESERAN TEPI KANAL MERAH (RED EDGE SHIFT) DATA HIPERSPEKTRAL

2013 ◽  
Vol 11 (3) ◽  
Author(s):  
Nadirah Nadirah ◽  
Bangun Muljosukojo ◽  
Teguh Hariyanto ◽  
M Sadly ◽  
M Evri ◽  
...  
Keyword(s):  
Nitrogen Content ◽  
Biochemical Parameter ◽  
Leaf Nitrogen ◽  
Growth Stages ◽  
Similar Response ◽  
Leaf Nitrogen Content ◽  
N Content ◽  
N Level ◽  
Red Edge ◽  
Leaf N

Canopy hyperspectral with various growth stages measured by using field spectroradiometer (350 - 1000 nm) corresponded to leaf Nitrogen content of three rice cultivars (Ciherang, Cilamaya and IR64) during growth season in Java Island,Indonesia. Coinciding with hyperspectral measurement, biochemical parameter such as leaf Nitrogen content (g/100 gr) was analyzed from destructive biomass sample through laboratory analysis. The potential narrow band in the red edgeregion was investigated to predict leaf nitrogen content (N content) with applying modified polynomial interpolation (MPI) and modified four points linear interpolation (MFLI) methods. First derivative reflectance derived from reflectance data andsubsequently used in analysis of Red Edge Position (REP). The correlation REPMFLI was generally stronger than REP-MPI attributed to leaf N content for several level of N application that indicated by value of R2. The response of REP-MFLItoward N level 69 kg/ha exhibited the most significant correlation (R2 = 0.754) than other correlations. Meanwhile, the response of REP-MPI toward N level 161 kg/ha denoted the most significant correlation (R2 = 0.8) than other correlations. The highest correlation using REP-MPI (R2 = 0.8) to predict leaf N contentdemonstrated slightly higher than that of REP- MFLI (R2 = 0.754). In general both REP-MFLI and REP-MPI represented somewhat similar response toward N levels, such as 103.5 kg/h, 115 kg/ha. The exploration of characteristics of red edge shiftis a fundamental point in developing rapid and precise prediction for biochemical parameter. In addition, its prediction capability was promising to support crop farming management.

scholarly journals Estimation of Nitrogen in Rice Crops from UAV-Captured Images

2020 ◽  
Vol 12 (20) ◽  
pp. 3396
Author(s):  
Julian D. Colorado ◽  
Natalia Cera-Bornacelli ◽  
Juan S. Caldas ◽  
Eliel Petro ◽  
Maria C. Rebolledo ◽  
...  
Keyword(s):  
Machine Learning ◽  
Large Scale ◽  
Vegetation Indices ◽  
Ground Truth ◽  
Leaf Nitrogen ◽  
Growth And Yield ◽  
Support Vector ◽  
Image Processing Algorithm ◽  
Stabilization Control ◽  
Leaf N Content

Leaf nitrogen (N) directly correlates to chlorophyll production, affecting crop growth and yield. Farmers use soil plant analysis development (SPAD) devices to calculate the amount of chlorophyll present in plants. However, monitoring large-scale crops using SPAD is prohibitively time-consuming and demanding. This paper presents an unmanned aerial vehicle (UAV) solution for estimating leaf N content in rice crops, from multispectral imagery. Our contribution is twofold: (i) a novel trajectory control strategy to reduce the angular wind-induced perturbations that affect image sampling accuracy during UAV flight, and (ii) machine learning models to estimate the canopy N via vegetation indices (VIs) obtained from the aerial imagery. This approach integrates an image processing algorithm using the GrabCut segmentation method with a guided filtering refinement process, to calculate the VIs according to the plots of interest. Three machine learning methods based on multivariable linear regressions (MLR), support vector machines (SVM), and neural networks (NN), were applied and compared through the entire phonological cycle of the crop: vegetative (V), reproductive (R), and ripening (Ri). Correlations were obtained by comparing our methods against an assembled ground-truth of SPAD measurements. The higher N correlations were achieved with NN: 0.98 (V), 0.94 (R), and 0.89 (Ri). We claim that the proposed UAV stabilization control algorithm significantly improves on the N-to-SPAD correlations by minimizing wind perturbations in real-time and reducing the need for offline image corrections.

scholarly journals The Relationship between Leaf Nitrogen Content and Photosynthesis in Apple Leaves

HortScience ◽  
1996 ◽  
Vol 31 (4) ◽  
pp. 578c-578
Author(s):  
Lailiang Cheng ◽  
Sunghee Guak ◽  
Leslie H. Fuchigami
Keyword(s):  
Nitrogen Content ◽  
Net Photosynthesis ◽  
Leaf Nitrogen ◽  
Leaf Nitrogen Content ◽  
Nitrogen And Phosphorus ◽  
N Content ◽  
Wide Range ◽  
Leaf N Content ◽  
The Relationship ◽  
Leaf N

Fertigation of young Fuji/M26 apple trees (Malus domestica Borkh.) with different nitrogen concentrations by using a modified Hoagland solution for 6 weeks resulted in a wide range of leaf nitrogen content in recently expanded leaves (from 0.9 to 4.4 g·m–2). Net photosynthesis at ambient CO2, carboxylation efficiency, and CO2-saturated photosynthesis of recently expanded leaves were closely related to leaf N content expressed on both leaf area and dry weight basis. They all increased almost linearly with increase in leaf N content when leaf N < 2.4 g·m–2, leveled off when leaf N increased further. The relationship between stomatal conductance and leaf N content was similar to that of net photosynthesis with leaf N content, but leaf intercellular CO2 concentration tended to decrease with increase in leaf N content, indicating non-stomatal limitation in leaves with low N content. Photosynthetic nitrogen use efficiency was high when leaf N < 2.4 g·m–2, but decreased with further increase in leaf N content. Due to the correlation between leaf nitrogen and phosphorus content, photosynthesis was also associated with leaf P content, but to a lesser extent.

Sign in / Sign up

Export Citation Format

Share Document