Monitoring Aboveground Biomass And Nitrogen Nutrition In Winter Wheat With Multi-angular Imagery From Unmanned Aerial Vehicles

The rapid,real-time,non-destructive and accurate assessment of crop aboveground biomass(AGB)and nitrogen nutrition status is of great significance for predicting crop yield,managing nitrogen application and reducing environmental pollution.Remote sensing technique,as a kind of non-destructive quantitative monitoring method,provides a new way and reliable technical support for crop growth monitoring.The unmanned aerial vehicle(UAV)is a new crop non-destructive remote sensing monitoring platform and has been widely used in crop growth monitoring due to its advantages of low cost,easy operation,data acquisition with high temporal and spatial resolution.Given by the research of UAV remote sensing in crop AGB and nitrogen nutrition estimation,most of them used nadir observation to obtain crop spectral data.However,the nadir observation can only obtain the single direction information of crop canopy surface and ignore the influence of observation angle on crop parameter inversion.Thus,the vegetation index derived from the single observation angle information is easy to saturate and the estimation accuracy is low when the biomass is high.The multi-angular remote sensing provides the radiation direction information and can generate structure information of crops,which is conducive to improve the accuracy of crop parameter estimation in theory.Therefore,it is of practical significance for crop precise management to explore the method of UAV-based multi-angular remote sensing to improve crop parameter estimation accuracy.In this study,two wheat field experiments were conducted,involving different growing seasons,cultivars,planting densities and nitrogen rates.Three sensors(RGB,color infrared(CIR)and multispectral(MS)cameras)mounted on UAV were used to acquire canopy imagery at critical growth stages of wheat.We evaluated the performance of vegetation indices(VIs),canopy height metrics and their combination derived from UAV-based RGB imagery for AGB estimation,explored an inexpensive approach consisting of the random forest(RF)algorithm and the combination of RGB imagery and point cloud data derived from a low-cost UAV system at the consumer-grade level for improving the accuracy of AGB estimation.For the estimation of nitrogen nutrition status,we analyzed the sensitivity of wheat nitrogen nutrition parameters to observation angle derived from the UAV-based multi-angular multispectral images and proposed the multi-angular observation method for the improvement of the estimation of nitrogen nutrition parameters.Moreover,multi-view images,as a special kind of mutli-angular data,can be acquired by a UAV-based nadir observation.Thus,we explored the multi-view information generated by high overlapping images of the UAV-based near-infrared camera,compared the performance of the single image,mosaic image and multi-view images for the estimation of nitrogen nutrition parameters.The multi-view information extraction method of UAV-based high overlapping images was established.The results would be helpful to provide theoretical foundation and technical support for improving the accuracy of crop AGB and nitrogen nutrition estimation with UAV multi-angular remote sensing technique.Firstly,this study established a low-cost and efficient monitoring method with the integration of UAV-based RGB images and its derived point cloud data for wheat AGB.We used a low-cost UAV system to acquire imagery at 30 m fight altitude at critical growth stages of wheat and evaluated the performance of VIs,canopy height metrics and their combination for AGB estimation in wheat with the stepwise multiple linear regression(SMLR)and three types of machine learning algorithms(support vector regression,SVR;extreme learning machine,ELM;random forest,RF).Our results demonstrated that the combination of VIs and canopy height metrics improved the estimation accuracy for AGB of wheat over the use of VIs or canopy height metrics alone.Specifcally,RF performed the best among the SMLR and three machine learning algorithms regardless of using all the original variables or selected variables by the SMLR.The best accuracy(R2=0.78,RMSE=1.34 t/ha,rRMSE=28.98%)was obtained when applying RF to the combination of VIs and canopy height metrics.Our fndings implied that an inexpensive approach consisting of the RF algorithm and the combination of RGB imagery and point cloud data derived from a low-cost UAV system at the consumer-grade level can be used to improve the accuracy of AGB estimation and have potential in the practical applications in the rapid estimation of other growth parameters.Secondly,we proposed a multi-angle observation method based on UAV multi-angular multispectral images to improve the estimation of nitrogen nutrition parameters.This study employed a UAV-based five-band camera to acquire multispectral images at seven view zenith angles(VZAs)(0°±20°,±40°and±60°)for critical growth stages of winter wheat.Four representative vegetation indices encompassing the Visible Atmospherically Resistant Index(VARI),Red edge Chlorophyll Index(CIred-edge),Green band Chlorophyll Index(CIgreen),Modified Normalized Difference Vegetation Index with a blue band(mNDblue)were derived from the multi-angular images.They were used to estimate the N nutrition status in leaf nitrogen concentration(LNC),plant nitrogen concentration(PNC),leaf nitrogen accumulation(LNA),and plant nitrogen accumulation(PNA)of wheat canopies for a combination of treatments in N rate,variety and planting density.The results demonstrated that the highest accuracy for single-angle images was obtained with CIgreen for LNC from a VZA of-60°(R2=0.71,RMSE=0.34%)and PNC from a VZA of-40°(R2=0.36,RMSE=0.29%).When combining an off-nadir image(-40°)and the 0°image,the accuracy of PNC estimation was substantially improved(CIred-edge:R2=0.52,RMSE=0.28%).However,the use of dual-angle images did not significantly increase the estimation accuracy for LNA and PNA compared to the use of single-angle images.Our findings suggest that it is important and practical to use oblique images from a UAV-based multispectral camera for better estimation of nitrogen concentration in wheat leaves or plants.The oblique images acquired from additional flights could be combined with the nadir-view images for improved crop N status monitoring.Finally,a multi-view information extraction method from UAV-based high overlapping images was established.This study employed a UAV-based CIR for the collection of high overlapping images at critical growth stages of wheat.Three kinds of machine learning algorithms(SVR,ELM and RF)were used to compare and analyze the performance of image information extraction methods,including single-view and multi-view images for the estimation of nitrogen nutrition.The results showed that the performance of the single image was the worst for the estimation of nitrogen nutrition,while the mosaic image obtained the moderate estimation accuracy.The highest estimation accuracy of LNC and PNC(LNC:R2=0.61,RMSE=0.37%;PNC:R2=0.52,RMSE=0.24%)was achieved with high overlapping multi-view images.However,among the three machine learning algorithms,ELM and RF did not substantially improve the estimation accuracy of LNA and PNA,but SVR significantly improved the estimation accuracy of nitrogen nutrition(LNC,PNC,LNA,PNA)based on high overlapping multi-view images,indicating that SVR performed the best in small samples.A multi-view information extraction method based on UAV high overlapping images we established provides a new idea for mining multi-view information of UAV-based high overlapping images and improving the estimation accuracy of crop AGB and nitrogen nutrition.