# Estimation of Relative Chlorophyll Content in Spring Wheat Based on Multi-Temporal UAV Remote Sensing

^{1}

^{2}

^{*}

## Abstract

**:**

^{2}), root mean square error (RMSE), and mean absolute error (MAPE). The results showed that the optimal SPAD value estimation models for different periods of independent reproductive growth stages of wheat were different, with PLS as the optimal estimation model at 7 and 14 days after heading, RF as the optimal estimation model at 21 days after heading, and Ada as the optimal estimation model at 28 d after heading. The highest accuracy was achieved using the PLS model for estimating SPAD values at 14 d after heading (training set R

^{2}= 0.767, RMSE = 3.205, MAPE = 0.060, and R

^{2}= 0.878, RMSE = 2.405, MAPE = 0.045 for the test set). The combined analysis concluded that selecting multiple vegetation indices as input variables of the model at 14 d after heading stage and using the PLS model can significantly improve the accuracy of SPAD value estimation, provides a new technical support for rapid and accurate monitoring of SPAD values in spring wheat.

## 1. Introduction

## 2. Materials and Methods

#### 2.1. Study Site and Experimental Design

^{3}/ha using flood irrigation. The N application subplot had six levels: CK (no fertilizer), N0 (0 kg/ha), N1 (75 kg/ha), N2 (150 kg/ha), N3 (225 kg/ha), N4 (300 kg/ha). The experiment had a total of 12 treatments with three replications, resulting in 36 experimental plots of 42 m

^{2}each. Phosphorus fertilizer was applied as a base fertilizer at sowing, and the sowing rate was set at 375 kg/ha. Rainfall and temperature data are shown in Figure 2.

#### 2.2. UAV Multispectral Data Acquisition and Processing

#### 2.3. Construction and Selection of Spectral Indices

#### 2.4. Ground Data Acquisition and Processing

#### 2.5. Construction of Regression Model

#### 2.6. Segmentation of Dataset and Accuracy Evaluation

^{2}), the root means square error (RMSE), and the mean absolute prediction error (MAPE). R

^{2}is used to indicate the degree of fit between the estimated and measured values, with a value closer to 1 indicating a higher accuracy of the model fit. RMSE reflects the deviation of the estimated value from the measured value, with a smaller value indicating a higher accuracy of the model fit. MAPE is the average of absolute errors, which more accurately reflects the actual errors in the prediction value.

## 3. Results

#### 3.1. Basic Statistical Information of Measured SPAD Values

#### 3.2. Correlation Analysis of SPAD Values and Vegetation Indices

#### 3.3. Model Development and Evaluation

#### 3.3.1. Estimation of SPAD Values after Heading 7 d

^{2}of the training set and the test set of the four models are above 0.70, with the accuracy of the training set higher accuracy than the test set. Using the accuracy of the test set as the evaluation criterion of the models, the accuracy of the models was found to be in the following order: PLS > RF > Ada > DNN. The R

^{2}, RMSE, and MAPE values of the test set of the PLS model were 0.762, 3.048, and 0.052, respectively, which were 7.32%, 2.28%, and 7.17% higher than the R

^{2}of the RF, Ada, and DNN models, respectively. RMSE decreased by 9.25%, 3.33%, and 9.20%, and MAPE decreased by 21.25%, 0%, and 7.69%. Based on these results, it can be concluded that the PLS model had the highest accuracy and stability in estimating the SPAD values of wheat 7 d after heading stage.

#### 3.3.2. Estimation of SPAD Values after Heading 14 d

^{2}values for the training and test sets of the four models are above 0.75, and the accuracy of the training set of the DNN and PLS models was lower than that of the test set, while the accuracy of the training set of the RF and Ada models was higher than that of the test set. When using the accuracy of the test set as the evaluation criterion for the models, the accuracies of different models were found to be PLS > Ada > RF > DNN in descending order. The R

^{2}, RMSE, and MAPE values for the test set of PLS model were 0.878, 2.405, and 0.045, which were 12.13%, 10.72%, and 7.33% higher than the R

^{2}values of the RF, Ada, and DNN models, respectively, and the RMSE is reduced by 25.05%, 23.41%, and 18.28%, and the MAPE decreased by 44.44%, 53.33%, and 35.56%. The comprehensive analysis concluded that the PLS model had the highest accuracy and stability in estimating the SPAD values of wheat 14 d after heading stage.

#### 3.3.3. Estimation of SPAD Values after Heading 21 d

^{2}of training set and test set of all four models were above 0.65, with the training set accuracy of the DNN and PLS models higher than the test set, while the training set accuracy of the RF and Ada models were lower than the test set. The accuracy of different models, from largest to smallest, was DNN > RF > Ada > PLS. The R

^{2}value of the DNN test set model was 0.737, the RMSE was 4.806, and the MAPE was 0.086, which were 12.18%, 1.80%, and 5.74% higher than the R

^{2}values of the PLS, RF, and Ada models, respectively. The RMSE was 12.41%, 2.42%, −11.02%, and the MAPE decreased by 28.33%, 18.10%, and −4.88%. Based on this analysis, it can be concluded that the DNN model was the most accurate and stable in estimating the SPAD values of wheat 21 d after heading stage.

#### 3.3.4. Estimation of SPAD Values after Heading 28 d

^{2}values for the training and test set of all four models are above 0.65, and the accuracy of the training set is higher than the test set for all models except the DNN model. When using the test set accuracy as the evaluation criterion, the Ada model was found to be the most accurate, followed by RF, PLS, and DNN in descending order. The R

^{2}value for the Ada model was 0.815, the RMSE was 5.904, and the MAPE was 0.237, which were 20.38%, 14.63%, and 14.47% higher, respectively, than those of the DNN, PLS, and RF models. The RMSE was reduced by 62.77%, 60.67%, and 60.59%, and the MAPE was reduced by 22.80%, 20.74%, and 31.30% for the Ada model compared to the other models. Overall, the Ada model was found to have the highest accuracy and stability in estimating the SPAD values of wheat 28 d after heading stage.

#### 3.4. Comparison of Accuracy of Four Estimation Models at Different Growth Stages

^{2}value at 7 and 14 days after heading was PLS, the model with the highest R

^{2}value at 21 d after heading was RF, and the model with the highest R

^{2}value at 28 d after heading was Ada. The model with the lowest RMSE value at 7 and 14 days after heading was PLS, and the model with the lowest RMSE value at 21 and 28 days after heading was Ada. The model with the lowest MAPE value at 7 days after heading was PLS and RF, the lowest at 14 days after heading was PLS, and the lowest at 21 and 28 days after heading was Ada.

## 4. Discussion

^{2}of 0.641. Multiple vegetation indices were used in this study to jointly participate in the estimation, and the model with the lowest precision among the four periods, R

^{2}= 0.657, was also above this precision. Yuan et al. [47] found in their study on estimating chlorophyll content of plant seedlings that most vegetation indices, although significantly correlated with SPAD values, had low R

^{2}for regression models with a single vegetation index as the explanatory variable, and the accuracy of the regression models was significantly improved when multiple vegetation indices were used as explanatory variables. In this study, all four regression models based on 26 vegetation indices demonstrated good performance, with test set R

^{2}ranging from 0.657 to 0.878, RMSE ranging from 2.934 to 7.801, and MAPE ranging from 0.045 to 0.345. This suggests that the inclusion of multiple vegetation indices in the regression models allowed for the incorporation of more valid spectral information, resulting in improved estimation accuracy.

^{2}of the training set of the Ada and RF models at 7, 14, and 28 days was significantly higher than that of the test set, and showed a significant overfitting phenomenon, probably due to the high autocorrelation among the 26 vegetation indices of the input, which was influenced by the multicollinearity of the input feature variables. The regression modeling method of PLSR converts a set of highly correlated independent variables into a set of mutually independent, non-linearly related principal component variables by extracting principal components in the process of establishing regressions, which can effectively capture most of the information of the original data and eliminate the covariance among vegetation indices, resulting in the best estimation accuracy of all four models being obtained at 7 d and 14 d after sampling. In the case of deep learning models, a large amount of diverse data is typically required for model training in order to understand the relationship between data and estimates, and the number of samples in this study was not sufficient to support a deep learning network with multiple hidden layers, leading to a low level of estimation accuracy for the DNN model. However, the estimation accuracy (R

^{2}) for all four periods was found to be greater than 0.677, indicating the strong potential for the DNN model to be used for estimation. This finding is in line with the results of Liu et al. [50].

## 5. Conclusions

^{2}= 0.767, RMSE = 3.205, MAPE = 0.060, and test set R

^{2}= 0.878, RMSE = 2.405, MAPE = 0.045).

## Author Contributions

## Funding

## Data Availability Statement

## Acknowledgments

## Conflicts of Interest

## References

- Jiang, H.L.; Yang, H.; Chen, X.P.; Wang, S.D.; Li, X.K.; Liu, K.; Cen, Y. Research on Accuracy and Stability of Inversing Vegetation Chlorophyll Content by Spectral Index Method. Spectrosc. Spectr. Anal.
**2015**, 35, 975–981. [Google Scholar] - Yuan, Y.C.; Zhou, Y.; Song, Y.F.; Xu, Z.; Wang, K.J. Estimation Method of Wheat Canopy Chlorophyll Based on Information Entropy Feature Selection. Trans. Chin. Soc. Agric. Mach.
**2022**, 53, 186–195. [Google Scholar] - Liu, T.; Zhang, H.; Wang, Z.Y.; He, C.; Zhang, Q.G.; Jiao, Y.Z. Estimation of the leaf area index and chlorophyll content of wheat using UAV multi-spectrum images. Trans. Chin. Soc. Agric. Eng.
**2021**, 37, 65–72. [Google Scholar] - Deng, S.Q.; Zhao, Y.; Bai, X.Y.; Li, X.; Sun, Z.D.; Liang, J.; Sun, Z.H.; Cheng, S. Inversion of chlorophyll and leaf area index for winter wheat based on UAV image segmentation. Trans. Chin. Soc. Agric. Eng.
**2022**, 38, 136–145. [Google Scholar] - Yi, H.; Li, F.; Yang, H.B.; Li, Y. Estimation of canopy chlorophyll in potato based on UAV hyperspectral images. J. Plant Nutr. Fertil.
**2021**, 27, 2184–2195. [Google Scholar] - Ma, L.F.; Hu, N.Y.; Li, W.; Qin, W.L.; Huang, S.B.; Wang, Z.M.; Li, F.; Yu, K. Using multispectral drone data to monitor maize’s response to various irrigation modes. J. Plant Nutr. Fertil.
**2022**, 28, 743–753. [Google Scholar] - Li, Z.F.; Su, J.X.; Fei, C.; Li, Y.Y.; Liu, N.N.; Fan, H.; Chen, B. Estimation of chlorophyll content in sugar beet under drip irrigation based on hyperspectral data. J. Agric. Resour. Environ.
**2020**, 37, 761–769. [Google Scholar] - Zhou, M.G.; Shao, G.M.; Zhang, L.Y.; Liu, Z.K.; Han, W.T. Multi-spectral Inversion of SPAD Value of Winter Wheat Based on Unmanned Aerial Vehicle Remote Sensing. Water Sav. Irrig.
**2019**, 9, 40–45. [Google Scholar] - Wei, Q.; Zhang, B.Z.; Wei, Z.; Han, X.; Duan, C.F. Estimation of Canopy Chlorophyll Content in Winter Wheat by UAV Multispectral Remote Sensing. J. Triticeae Crops
**2020**, 40, 365–372. [Google Scholar] - Zhou, M.G.; Shao, G.M.; Zhang, L.Y.; Yao, X.M.; Han, W.T. Inversion of SPAD value of winter wheat by multispectral remote sensing of unmanned aerial vehicles. Trans. Chin. Soc. Agric. Eng.
**2020**, 36, 125–133. [Google Scholar] - Niu, Q.L.; Feng, H.K.; Zhou, X.G.; Zhu, J.Q.; Yong, B.B.; Li, H.Z. Combining UAV Visible Light and Multispectral Vegetation Indices for Estimating SPAD Value of Winter Wheat. Trans. Chin. Soc. Agric. Mach.
**2021**, 52, 183–194. [Google Scholar] - Mao, Z.H.; Deng, L.; Sun, J.; Zhang, A.W.; Chen, X.Y.; Zhao, Y. Research on the Application of UAV Multispectral Remote Sensing in the Maize Chlorophyll Prediction. Spectrosc. Spectr. Anal.
**2018**, 38, 2923–2931. [Google Scholar] - Han, W.T.; Peng, X.S.; Zhang, L.Y.; Niu, Y.X. Summer Maize Yield Estimation Based on Vegetation Index Derived from Multi-temporal UAV Remote Sensing. Trans. Chin. Soc. Agric. Mach.
**2020**, 51, 148–155. [Google Scholar] - Chen, Q.; Xu, H.G.; Cao, Y.B.; Duan, F.Y.; Chen, Z. Grain Yield Prediction of Winter Wheat Using Multi-temporal UAV Based on Multispectral Vegetation Index. Trans. Chin. Soc. Agric. Mach.
**2021**, 52, 160–167. [Google Scholar] - Zhou, X.; Zheng, H.B.; Xu, X.Q.; He, J.Y.; Ge, X.K.; Yao, X.; Cheng, T.; Zhu, Y.; Cao, W.X.; Tian, Y.C. Predicting grain yield in rice using multi-temporal vegetation indices from UAV-based multispectral and digital imagery. ISPRS J. Photogramm. Remote Sens.
**2017**, 130, 246–255. [Google Scholar] [CrossRef] - Huete, A.; Justice, C.; Liu, H. Development of vegetation and soil indices for MODIS-EOS. Remote Sens. Environ.
**1994**, 49, 224–234. [Google Scholar] [CrossRef] - Gitelson, A.A.; Merzlyak, M.N. Remote sensing of chlorophyll concentration in higher plant leaves. Adv. Space Res.
**1998**, 22, 689–692. [Google Scholar] [CrossRef] - Gamon, J.A.; Surfus, J.S. Assessing leaf pigment content and activity with a reflectometer. New Phytol.
**1999**, 143, 105–117. [Google Scholar] [CrossRef] - Birth, G.S.; McVey, G.R. Measuring the color of growing turf with a reflectance spectrophotometer 1. Agron. J.
**1968**, 60, 640–643. [Google Scholar] [CrossRef] - Jasper, J.; Reusch, S.; Link, A. Active sensing of the N status of wheat using optimized wavelength combination: Impact of seed rate, variety and growth stage. Precis. Agric.
**2009**, 9, 23–30. [Google Scholar] - Filella, I.; Penuelas, J. The red edge position and shape as indicators of plant chlorophyll content, biomass and hydric status. Int. J. Remote Sens.
**1994**, 15, 1459–1470. [Google Scholar] [CrossRef] - Qi, J.; Chehbouni, A.; Huete, A.R.; Kerr, Y.H.; Sorooshian, S. A modified soil adjusted vegetation index. Remote Sens. Environ.
**1994**, 48, 119–126. [Google Scholar] [CrossRef] - Rondeaux, G.; Steven, M.; Baret, F. Optimization of soil-adjusted vegetation indices. Remote Sens. Environ.
**1996**, 55, 95–107. [Google Scholar] [CrossRef] - Xing, N.C.; Huang, W.J.; Xie, Q.Y.; Shi, Y.; Ye, H.C.; Dong, Y.Y.; Wu, M.Q.; Sun, G.; Jiao, Q.J. A Transformed Triangular Vegetation Index for Estimating Winter Wheat Leaf Area Index. Remote Sens.
**2019**, 12, 16. [Google Scholar] [CrossRef][Green Version] - Sims, D.A.; Gamon, J.A. Relationships between leaf pigment content and spectral reflectance across a wide range of species, leaf structures and developmental stages. Remote Sens. Environ.
**2002**, 81, 337–354. [Google Scholar] [CrossRef] - Chen, J.; Jönsson, P.; Tamura, M.; Gu, Z.H.; Matsushita, B.; Eklundh, L. A simple method for reconstructing a high-quality NDVI time-series data set based on the Savitzky–Golay filter. Remote Sens. Environ.
**2004**, 91, 332–344. [Google Scholar] [CrossRef] - Chen, J.M. Evaluation of vegetation indices and a modified simple ratio for boreal applications. Can. J. Remote Sens.
**1996**, 22, 229–242. [Google Scholar] [CrossRef] - Huete, A.R. A soil-adjusted vegetation index (SAVI). Remote Sens. Environ.
**1988**, 25, 295–309. [Google Scholar] [CrossRef] - Fitzgerald, G.; Rodriguez, D.; O’Leary, G. Measuring and predicting canopy nitrogen nutrition in wheat using a spectral index—The canopy chlorophyll content index (CCCI). Field Crops Res.
**2010**, 116, 318–324. [Google Scholar] [CrossRef] - Kimura, R.; Okada, S.; Miura, H.; Kamichika, M. Relationships among the leaf area index, moisture availability, and spectral reflectance in an upland rice field. Agric. Water Manag.
**2004**, 69, 83–100. [Google Scholar] [CrossRef] - Wu, C.Y.; Niu, Z.; Tang, Q.; Huang, W.J. Estimating chlorophyll content from hyperspectral vegetation indices: Modeling and validation. Agric. For. Meteorol.
**2008**, 148, 1230–1241. [Google Scholar] [CrossRef] - Devadas, R.; Lamb, D.W.; Simpfendorfer, S.; Backhouse, D. Evaluating ten spectral vegetation indices for identifying rust infection in individual wheat leaves. Precis. Agric.
**2009**, 10, 459–470. [Google Scholar] [CrossRef] - Zha, Y.; Gao, J.; Ni, S. Use of normalized difference built-up index in automatically mapping urban areas from TM imagery. Int. J. Remote Sens.
**2003**, 24, 583–594. [Google Scholar] [CrossRef] - Ju, C.H.; Tian, Y.C.; Yao, X.; Cao, W.X.; Zhu, Y.; Hannaway, D. Estimating leaf chlorophyll content using red edge parameters. Pedosphere
**2010**, 20, 633–644. [Google Scholar] [CrossRef] - Clevers, J.G.P.W.; Gitelson, A.A. Remote estimation of crop and grass chlorophyll and nitrogen content using red-edge bands on Sentinel-2 and-3. Int. J. Appl. Earth Obs. Geoinf.
**2013**, 23, 344–351. [Google Scholar] [CrossRef] - Verrelst, J.; Schaepman, M.E.; Koetz, B.; Kneubühler, M. Angular sensitivity analysis of vegetation indices derived from CHRIS/PROBA data. Remote Sens. Environ.
**2008**, 112, 2341–2353. [Google Scholar] [CrossRef] - Wang, W.; Cheng, Y.K.; Ren, Y.; Zhang, Z.H.; Geng, H.G. Prediction of Chlorophyll Content in Multi-Temporal Winter Wheat Based on Multispectral and Machine Learning. Front. Plant Sci.
**2022**, 13, 896408. [Google Scholar] [CrossRef] - Fu, Z.P.; Yu, S.S.; Zhang, J.Y.; Xi, H.; Gao, Y.; Lu, R.H.; Zheng, H.B.; Zhu, Y.; Cao, W.X.; Liu, X.J. Combining UAV multispectral imagery and ecological factors to estimate leaf nitrogen and grain protein content of wheat. Eur. J. Agron.
**2022**, 132, 126405. [Google Scholar] [CrossRef] - Revill, A.; Florence, A.; MacArthur, A.; Hoad, S.P.; Rees, R.M.; Williams, M. The value of Sentinel-2 spectral bands for the assessment of winter wheat growth and development. Remote Sens.
**2019**, 11, 2050. [Google Scholar] [CrossRef][Green Version] - Basso, B.; Cammarano, D.; De, V.P. Remotely sensed vegetation indices: Theory and applications for crop management. Riv. Ital. Agrometeorol.
**2004**, 1, 36–53. [Google Scholar] - El-Hendawy, S.E.; Hassan, W.M.; Al-Suhaibani, N.A.; Schmidhalter, U. Spectral assessment of drought tolerance indices and grain yield in advanced spring wheat lines grown under full and limited water irrigation. Agric. Water Manag.
**2017**, 182, 1–12. [Google Scholar] [CrossRef] - Bowman, B.C.; Chen, J.; Zhang, J.; Wheeler, J.; Wang, Y.; Zhao, W.; Nayak, S.; Heslot, N.; Bockelman, H.; Bonman, J.M. Evaluating grain yield in spring wheat with canopy spectral reflectance. Crop Sci.
**2015**, 55, 1881–1890. [Google Scholar] [CrossRef][Green Version] - Asner, G.P. Biophysical and biochemical sources of variability in canopy reflectance. Remote Sens. Environ.
**1998**, 64, 234–253. [Google Scholar] [CrossRef] - Ustin, S.L. Remote sensing of canopy chemistry. Proc. Natl. Acad. Sci. USA
**2013**, 110, 804–805. [Google Scholar] [CrossRef] [PubMed][Green Version] - Zheng, H.B.; Li, W.; Jiang, J.L.; Liu, Y.; Cheng, T.; Tian, Y.C.; Zhu, Y.; Cao, W.X.; Zhang, Y.; Yao, X. A comparative assessment of different modeling algorithms for estimating leaf nitrogen content in winter wheat using multispectral images from an unmanned aerial vehicle. Remote Sens.
**2018**, 10, 2026. [Google Scholar] [CrossRef][Green Version] - Kooistra, L.; Clevers, J.G.P.W. Estimating potato leaf chlorophyll content using ratio vegetation indices. Remote Sens. Lett.
**2016**, 7, 611–620. [Google Scholar] [CrossRef] - Yuan, Y.; Wang, X.F.; Shi, M.M.; Wang, P. Performance comparison of RGB and multispectral vegetation indices based on machine learning for estimating Hopea hainanensis SPAD values under different shade conditions. Front. Plant Sci.
**2022**, 13, 2615. [Google Scholar] [CrossRef] - Peng, Y.; Fan, M.; Song, J.Y.; Cui, T.T.; Li, H. Assessment of plant species diversity based on hyperspectral indices at a fine scale. Sci. Rep.
**2018**, 8, 4776. [Google Scholar] [CrossRef][Green Version] - Peng, M.M.; Han, W.T.; Li, C.Q.; Huang, S.J. Improving the Spatial and Temporal Estimation of Maize Daytime Net Ecosystem Carbon Exchange Variation Based on Unmanned Aerial Vehicle Multispectral Remote Sensing. IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens.
**2021**, 14, 10560–10570. [Google Scholar] [CrossRef] - Liu, S.B.; Jin, X.L.; Nie, C.W.; Wang, S.Y.; Yu, X.; Cheng, M.H.; Shao, M.C.; Wang, Z.X.; Tuohuti, N.; Bai, Y.; et al. Estimating leaf area index using unmanned aerial vehicle data: Shallow vs. deep machine learning algorithms. Plant Physiol.
**2021**, 187, 1551–1576. [Google Scholar] [CrossRef]

**Figure 1.**Geographical location of the experimental site. Red frame is 2W treatment; Blue frame is 4W treatment.

**Figure 2.**Rainfall and temperature of experimental site. Mar: March, Apr: April, Jun: June, Jul: July.

**Figure 9.**Violin plots of the spectral indices. (

**a**–

**d**) are violin plots of vegetation indices at 7, 14, 21, and 28 days after heading, respectively.

Index Name | Calculation Formula | References |
---|---|---|

Leaf chlorophyll index | $\mathrm{LCI}\text{}=({\mathrm{R}}_{\mathrm{nir}}-{\mathrm{R}}_{\mathrm{rededge}})/\left({\mathrm{R}}_{\mathrm{nir}\text{}}+{\mathrm{R}}_{\mathrm{red}\text{}}\right)$ | [16] |

Difference vegetation index | $\mathrm{DVI}={\text{}\mathrm{R}}_{\mathrm{nir}}-{\mathrm{R}}_{\mathrm{red}}$ | [16] |

Enhanced Vegetation Index | $\mathrm{EVI}=2.5\times \left({\mathrm{R}}_{\mathrm{nir}}-{\mathrm{R}}_{\mathrm{red}}\right)/\left({\mathrm{R}}_{\mathrm{nir}}+6\times {\mathrm{R}}_{\mathrm{red}}-7.5\times {\mathrm{R}}_{\mathrm{blue}}+1\right)$ | [16] |

Green Normalized Difference Vegetation | $\mathrm{GNDVI}\text{}=\left({\mathrm{R}}_{\mathrm{nir}\text{}}-{\mathrm{R}}_{\mathrm{green}\text{}}\right)/\left({\mathrm{R}}_{\mathrm{nir}\text{}}+{\mathrm{R}}_{\mathrm{green}\text{}}\right)$ | [17] |

Ratio Between NIR and Green Bands | ${\mathrm{VI}}_{(\mathrm{nir}/\mathrm{green})\text{}}={\mathrm{R}}_{\mathrm{nir}\text{}}/{\mathrm{R}}_{\mathrm{green}\text{}}$ | [18] |

Ratio Between NIR and Red Bands | ${\mathrm{VI}}_{\left(\mathrm{nir}\text{}/\text{}\mathrm{red}\right)}={\mathrm{R}}_{\mathrm{nir}\text{}}/{\mathrm{R}}_{\mathrm{red}}$ | [19] |

Ratio Between NIR and Red Edge Bands | ${\mathrm{VI}}_{\left(\mathrm{nir}/\mathrm{rededge}\right)}={\mathrm{R}}_{\mathrm{nir}\text{}}/{\mathrm{R}}_{\mathrm{rededge}}$ | [20] |

Napierian Logarithm of The Red Edge | ${\mathrm{ln}}_{\mathrm{RE}}=100\times \left({\mathrm{ln}}_{\mathrm{nir}}-{\mathrm{ln}}_{\mathrm{red}}\right)$ | [21] |

Modified Soil-Adjusted Vegetation Index 1 | $\mathrm{MSAVI}1=\left(1+\mathrm{L}\right)\left(\frac{{\mathrm{R}}_{\mathrm{nir}}-{\mathrm{R}}_{\mathrm{red}}}{{\mathrm{R}}_{\mathrm{nir}}+{\mathrm{R}}_{\mathrm{red}}+\mathrm{L}}\right)\left(\mathrm{L}=0.1\right)$ | [22] |

Modified Soil-Adjusted Vegetation Index 2 | $\mathrm{MSAVI}2={\mathrm{R}}_{\mathrm{nir}}+0.5-\sqrt{{\left(2\times {\mathrm{R}}_{\mathrm{nir}}+1\right)}^{2}-8\times \left({\mathrm{R}}_{\mathrm{nir}}-{\mathrm{R}}_{\mathrm{red}}\right)}/2$ | [22] |

Optimized Soil-Adjusted Vegetation Index | $\mathrm{OSAVI}=\left(1+0.16\right)\times \frac{\left({\mathrm{R}}_{\mathrm{nir}}-{\mathrm{R}}_{\mathrm{red}}\right)}{\left({\mathrm{R}}_{\mathrm{nir}}+{\mathrm{R}}_{\mathrm{red}}+0.16\right)}$ | [23] |

Modified Triangular Vegetation Index 2 | $\mathrm{MTVI}2=\frac{1.5\times \left[1.2\times \left({\mathrm{R}}_{\mathrm{nir}\text{}}-{\mathrm{R}}_{\mathrm{green}\text{}}\right)-2.5\times \left({\mathrm{R}}_{\mathrm{red}\text{}}-{\mathrm{R}}_{\mathrm{green}\text{}}\right)\right]}{\sqrt{{\left(2\times {\mathrm{R}}_{\mathrm{nir}}+1\right)}^{2}-\left(6\times {\mathrm{R}}_{\mathrm{nir}}-5\times \sqrt{{\mathrm{R}}_{\mathrm{red}\text{}}}\right)-0.5}}$ | [24] |

Normalized Difference Red Edge Index | $\mathrm{NDRE}=\frac{\left({\mathrm{R}}_{\mathrm{nir}\text{}}-{\mathrm{R}}_{\mathrm{rededge}\text{}}\right)}{\left({\mathrm{R}}_{\mathrm{nir}\text{}}+{\mathrm{R}}_{\mathrm{rededge}\text{}}\right)}$ | [25] |

Normalized Difference Vegetation Index | $\mathrm{NDVI}=\frac{\left({\mathrm{R}}_{\mathrm{nir}\text{}}-{\mathrm{R}}_{\mathrm{red}\text{}}\right)}{\left({\mathrm{R}}_{\mathrm{nir}\text{}}+{\mathrm{R}}_{\mathrm{red}\text{}}\right)}$ | [26] |

Modified Simple Radio | $\mathrm{MSR}=\left({\mathrm{R}}_{\mathrm{nir}}-{\mathrm{R}}_{\mathrm{red}}-1\right)/\left(\sqrt{{\mathrm{R}}_{\mathrm{nir}\text{}}+{\mathrm{R}}_{\mathrm{red}\text{}}}+1\right)$ | [27] |

Soil-Adjusted Vegetation Index | $\mathrm{SAVI}=\frac{\left({\mathrm{R}}_{\mathrm{nir}}-{\mathrm{R}}_{\mathrm{red}}\right)}{\left({\mathrm{R}}_{\mathrm{nir}}+{\mathrm{R}}_{\mathrm{red}}+0.5\right)}\times \left(1+0.5\right)$ | [28] |

Simplified Canopy Chlorophyll Content Index | $\mathrm{SCCCI}=\frac{\mathrm{NDRE}}{\mathrm{NDVI}}$ | [29] |

Modified Chlorophyll Absorption Reflectance Index | $\mathrm{MCARI}\text{}=\left({\mathrm{R}}_{\mathrm{rededge}\text{}}-{\mathrm{R}}_{\mathrm{red}\text{}}-0.2\times \left({\mathrm{R}}_{\mathrm{rededge}\text{}}-{\mathrm{R}}_{\mathrm{green}\text{}}\right)\right)\times \left(\frac{{\mathrm{R}}_{\mathrm{rededge}\text{}}}{{\mathrm{R}}_{\mathrm{red}\text{}}}\right)$ | [30] |

Modified Chlorophyll Absorption Reflectance Index 2 | $\mathrm{MCARI}2=1.5\times \frac{\left(2.5\times \left({\mathrm{R}}_{\mathrm{nir}\text{}}-{\mathrm{R}}_{\mathrm{rededge}\text{}}\right)-1.3\times \left({\mathrm{R}}_{\mathrm{nir}\text{}}-{\mathrm{R}}_{\mathrm{g}}\right)\right)}{\left(2\times {\left({\mathrm{R}}_{\mathrm{nir}}+1\right)}^{2}-\left(6\times {\mathrm{R}}_{\mathrm{nir}}-5\times {\left({\mathrm{R}}_{\mathrm{red}}\right)}^{2}\right)-0.5\right)}$ | [31] |

Transformed Chlorophyll Absorption Reflectance Index | $\mathrm{TCARI}\text{}=3\times \left(\left({\mathrm{R}}_{\mathrm{rededge}\text{}}-{\mathrm{R}}_{\mathrm{red}\text{}}\right)-0.2\times \left({\mathrm{R}}_{\mathrm{rededge}\text{}}-{\mathrm{R}}_{\mathrm{green}\text{}}\right)\times \left(\frac{{\mathrm{R}}_{\mathrm{rededge}\text{}}}{{\mathrm{R}}_{\mathrm{red}\text{}}}\right)\right)$ | [32] |

Normalized Difference Index | $\mathrm{NDI}=\frac{\left({\mathrm{R}}_{\mathrm{nir}\text{}}-{\mathrm{R}}_{\mathrm{rededge}\text{}}\right)}{\left({\mathrm{R}}_{\mathrm{nir}\text{}}+{\mathrm{R}}_{\mathrm{red}\text{}}\right)}$ | [33] |

Red-Edge Chlorophyll Index 1 | $\mathrm{Cl}1=\frac{{\mathrm{R}}_{\mathrm{nir}\text{}}}{{\mathrm{R}}_{\mathrm{rededge}\text{}}}-1$ | [34] |

Red-Edge Chlorophyll Index 2 | $\mathrm{Cl}2=\frac{{\mathrm{R}}_{\mathrm{rededge}\text{}}}{{\mathrm{R}}_{\mathrm{green}\text{}}}-1$ | [35] |

Structure-Insensitive Pigment Index | $\mathrm{SIPI}=\frac{\left({\mathrm{R}}_{\mathrm{nir}\text{}}-{\mathrm{R}}_{\mathrm{blue}\text{}}\right)}{\left({\mathrm{R}}_{\mathrm{nir}}+{\mathrm{R}}_{\mathrm{red}}\right)}$ | [36] |

TCARI/OSAVI | $\frac{\mathrm{TCARI}}{\mathrm{OSAVI}}$ | [31] |

MCARI/OSAVI | $\frac{\mathrm{MCARI}}{\mathrm{OSAVI}}$ | [31] |

Irrigation | N Treatment | 7 d | 14 d | 21 d | 28 d |
---|---|---|---|---|---|

2W | N0 | 41.99 bcd | 43.42 bcd | 43.11 d | 36.33 bc |

N5 | 46.90 abc | 47.75 abc | 47.05 bcd | 42.99 ab | |

N10 | 48.42 abc | 50.23 ab | 51.89 ab | 43.82 ab | |

N15 | 50.10 ab | 51.05 a | 55.50 a | 45.29 a | |

N20 | 48.58 abc | 46.05 abcd | 51.53 abc | 46.51 a | |

CK | 39.95 cd | 36.26 de | 33.46 e | 14.73 e | |

4W | N0 | 41.70 bcd | 35.75 ef | 28.47 e | 10.47 e |

N5 | 46.61 abc | 42.75 bcde | 45.09 cd | 24.35 d | |

N10 | 47.97 abc | 41.80 cde | 45.81 bcd | 29.48 cd | |

N15 | 51.15 a | 46.77 abcd | 50.62 abc | 34.55 c | |

N20 | 41.65 bcd | 45.75 abcd | 46.13 bcd | 26.15 d | |

CK | 36.29 cd | 29.63 f | 31.37 e | 12.87 e |

Indices | 7 d | 14 d | 21 d | 28 d | ||||
---|---|---|---|---|---|---|---|---|

2W | 4W | 2W | 4W | 2W | 4W | 2W | 4W | |

DVI | 0.912 | 0.867 | 0.650 | 0.867 | 0.701 | 0.830 | 0.708 | 0.802 |

EVI | 0.908 | 0.866 | 0.695 | 0.871 | 0.708 | 0.856 | 0.727 | 0.798 |

NDVI | 0.842 | 0.815 | 0.790 | 0.871 | 0.703 | 0.867 | 0.740 | 0.790 |

GNDVI | 0.882 | 0.825 | 0.776 | 0.870 | 0.741 | 0.903 | 0.741 | 0.779 |

NDRE | 0.912 | 0.846 | 0.739 | 0.863 | 0.764 | 0.925 | 0.749 | 0.774 |

LCI | 0.912 | 0.846 | 0.739 | 0.863 | 0.764 | 0.925 | 0.749 | 0.774 |

OSAVI | 0.879 | 0.847 | 0.745 | 0.873 | 0.714 | 0.876 | 0.734 | 0.799 |

VI(NIR/G) | 0.893 | 0.867 | 0.671 | 0.830 | 0.765 | 0.919 | 0.742 | 0.756 |

VI(NIR/R) | 0.842 | 0.863 | 0.637 | 0.819 | 0.761 | 0.918 | 0.726 | 0.754 |

VI(NIR/RE) | 0.917 | 0.863 | 0.698 | 0.849 | 0.765 | 0.926 | 0.748 | 0.766 |

lnRE | 0.856 | 0.844 | 0.735 | 0.853 | 0.747 | 0.916 | 0.742 | 0.782 |

MSAVI1 | 0.870 | 0.840 | 0.759 | 0.873 | 0.713 | 0.875 | 0.736 | 0.797 |

MSAVI2 | 0.729 | 0.706 | 0.815 | 0.855 | 0.589 | 0.796 | 0.734 | 0.719 |

MTVI2 | 0.895 | 0.868 | 0.677 | 0.863 | 0.711 | 0.859 | 0.722 | 0.801 |

MSR | 0.915 | 0.859 | 0.640 | 0.866 | 0.692 | 0.801 | 0.679 | 0.793 |

SAVI | 0.898 | 0.860 | 0.706 | 0.872 | 0.712 | 0.867 | 0.727 | 0.802 |

SCCCI | 0.924 | 0.840 | 0.723 | 0.864 | 0.772 | 0.925 | 0.231 | −0.652 |

MCARI | −0.824 | −0.782 | −0.785 | −0.854 | −0.739 | −0.881 | −0.783 | 0.463 |

MCARI2 | 0.929 | 0.870 | 0.663 | 0.853 | 0.764 | 0.910 | 0.746 | 0.750 |

TCARI | −0.817 | −0.850 | −0.638 | −0.830 | −0.766 | −0.921 | −0.562 | 0.682 |

NDI | 0.902 | 0.840 | 0.755 | 0.866 | 0.758 | 0.920 | 0.749 | 0.779 |

CL1 | 0.917 | 0.863 | 0.698 | 0.849 | 0.765 | 0.926 | 0.748 | 0.766 |

CL2 | 0.869 | 0.850 | 0.710 | 0.840 | 0.759 | 0.910 | 0.743 | 0.763 |

SIPI | 0.834 | 0.809 | 0.779 | 0.865 | 0.713 | 0.867 | 0.735 | 0.788 |

TCARI/OSAVI | −0.826 | −0.841 | −0.689 | −0.846 | −0.753 | −0.919 | −0.766 | −0.618 |

MCARI/OSAVI | −0.826 | −0.786 | −0.797 | −0.870 | −0.695 | −0.847 | −0.757 | −0.742 |

Model | Training Set | Test Set | ||||
---|---|---|---|---|---|---|

R^{2} | RMSE | MAPE | R^{2} | RMSE | MAPE | |

DNN | 0.754 | 2.784 | 0.048 | 0.710 | 3.359 | 0.063 |

PLS | 0.786 | 2.595 | 0.043 | 0.762 | 3.048 | 0.052 |

RF | 0.957 | 1.169 | 0.021 | 0.745 | 3.153 | 0.052 |

Ada | 0.968 | 0.829 | 0.014 | 0.711 | 3.357 | 0.056 |

Model | Training Set | Test Set | ||||
---|---|---|---|---|---|---|

R^{2} | RMSE | MAPE | R^{2} | RMSE | MAPE | |

DNN | 0.716 | 3.538 | 0.067 | 0.783 | 3.209 | 0.065 |

PLS | 0.767 | 3.205 | 0.060 | 0.878 | 2.405 | 0.045 |

RF | 0.924 | 1.835 | 0.036 | 0.793 | 3.140 | 0.069 |

Ada | 0.934 | 1.711 | 0.029 | 0.818 | 2.943 | 0.061 |

Model | Training Set | Test Set | ||||
---|---|---|---|---|---|---|

R^{2} | RMSE | MAPE | R^{2} | RMSE | MAPE | |

DNN | 0.881 | 2.922 | 0.044 | 0.737 | 4.806 | 0.086 |

PLS | 0.777 | 4.009 | 0.080 | 0.657 | 5.487 | 0.120 |

RF | 0.678 | 4.815 | 0.097 | 0.724 | 4.925 | 0.105 |

Ada | 0.684 | 5.028 | 0.092 | 0.697 | 4.329 | 0.082 |

Model | Training Set | Test Set | ||||
---|---|---|---|---|---|---|

R^{2} | RMSE | MAPE | R^{2} | RMSE | MAPE | |

DNN | 0.691 | 7.054 | 0.262 | 0.677 | 7.801 | 0.307 |

PLS | 0.713 | 6.803 | 0.299 | 0.711 | 7.383 | 0.299 |

RF | 0.926 | 3.442 | 0.115 | 0.712 | 7.368 | 0.345 |

Ada | 0.971 | 2.168 | 0.067 | 0.815 | 5.904 | 0.237 |

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |

© 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).

## Share and Cite

**MDPI and ACS Style**

Wu, Q.; Zhang, Y.; Zhao, Z.; Xie, M.; Hou, D.
Estimation of Relative Chlorophyll Content in Spring Wheat Based on Multi-Temporal UAV Remote Sensing. *Agronomy* **2023**, *13*, 211.
https://doi.org/10.3390/agronomy13010211

**AMA Style**

Wu Q, Zhang Y, Zhao Z, Xie M, Hou D.
Estimation of Relative Chlorophyll Content in Spring Wheat Based on Multi-Temporal UAV Remote Sensing. *Agronomy*. 2023; 13(1):211.
https://doi.org/10.3390/agronomy13010211

**Chicago/Turabian Style**

Wu, Qiang, Yongping Zhang, Zhiwei Zhao, Min Xie, and Dingyi Hou.
2023. "Estimation of Relative Chlorophyll Content in Spring Wheat Based on Multi-Temporal UAV Remote Sensing" *Agronomy* 13, no. 1: 211.
https://doi.org/10.3390/agronomy13010211