Unmanned aerial system (UAS)-based remote sensing is one promising technique for precision crop management, but few studies have reported the applications of such systems on nitrogen (N) estimation with multiple sensors in rice (Oryza sativa L.). This study aims to evaluate three sensors (RGB, color-infrared (CIR) and multispectral (MS) cameras) onboard UAS for the estimation of N status at individual stages and their combination with the field data collected from a two-year rice experiment. The experiments were conducted in 2015 and 2016, involving different N rates, planting densities and rice cultivars, with three replicates. An Oktokopter UAS was used to acquire aerial photography at early growth stages (from tillering to booting) and field samplings were taken at a near date. Two color indices (normalized excess green index (NExG), and normalized green red difference index (NGRDI)), two near infrared vegetation indices (green normalized difference vegetation index (GNDVI), and enhanced NDVI (ENDVI)) and two red edge vegetation indices (red edge chlorophyll index (CI_{red edge}), and DATT) were used to evaluate the capability of these three sensors in estimating leaf nitrogen accumulation (LNA) and plant nitrogen accumulation (PNA) in rice. The results demonstrated that the red edge vegetation indices derived from MS images produced the highest estimation accuracy for LNA (R²: 0.79–0.81, root mean squared error (RMSE): 1.43–1.45 g m⁻²) and PNA (R²: 0.81–0.84, RMSE: 2.27–2.38 g m⁻²). The GNDVI from CIR images yielded a moderate estimation accuracy with an all-stage model. Color indices from RGB images exhibited satisfactory performance for the pooled dataset of the tillering and jointing stages. Compared with the counterpart indices from the RGB and CIR images, the indices from the MS images performed better in most cases. These results may set strong foundations for the development of UAS-based rice growth monitoring systems, providing useful information for the real-time decision making on crop N management.

Deep-learning-based object detection algorithms have significantly improved the performance of wheat spike detection. However, UAV images crowned with small-sized, highly dense, and overlapping spikes cause the accuracy to decrease for detection. This paper proposes an improved YOLOv5 (You Look Only Once)-based method to detect wheat spikes accurately in UAV images and solve spike error detection and miss detection caused by occlusion conditions. The proposed method introduces data cleaning and data augmentation to improve the generalization ability of the detection network. The network is rebuilt by adding a microscale detection layer, setting prior anchor boxes, and adapting the confidence loss function of the detection layer based on the IoU (Intersection over Union). These refinements improve the feature extraction for small-sized wheat spikes and lead to better detection accuracy. With the confidence weights, the detection boxes in multiresolution images are fused to increase the accuracy under occlusion conditions. The result shows that the proposed method is better than the existing object detection algorithms, such as Faster RCNN, Single Shot MultiBox Detector (SSD), RetinaNet, and standard YOLOv5. The average accuracy (AP) of wheat spike detection in UAV images is 94.1%, which is 10.8% higher than the standard YOLOv5. Thus, the proposed method is a practical way to handle the spike detection in complex field scenarios and provide technical references for field-level wheat phenotype monitoring.

Leaf area index (LAI) is a significant biophysical variable in the models of hydrology, climatology and crop growth. Rapid monitoring of LAI is critical in modern precision agriculture. Remote sensing (RS) on satellite, aerial and unmanned aerial vehicles (UAVs) has become a popular technique in monitoring crop LAI. Among them, UAVs are highly attractive to researchers and agriculturists. However, some of the UAVs vegetation index (VI)—derived LAI models have relatively low accuracy because of the limited number of multispectral bands, especially as they tend to saturate at the middle to high LAI levels, which are the LAI levels of high-yielding wheat crops in China. This study aims to effectively estimate wheat LAI with UAVs narrowband multispectral image (400–800 nm spectral regions, 10 cm resolution) under varying growth conditions during five critical growth stages, and to provide the potential technical support for optimizing the nitrogen fertilization. Results demonstrated that the newly developed LAI model with modified triangular vegetation index (MTVI₂) has better accuracy with higher coefficient of determination (R_c² = 0.79, R_v² = 0.80) and lower relative root mean squared error (RRMSE = 24%), and higher sensitivity under various LAI values (from 2 to 7), which will broaden the applied range of the new LAI model. Furthermore, this LAI model displayed stable performance under different sub-categories of growth stages, varieties, and eco-sites. In conclusion, this study could provide effective technical support to precisely monitor the crop growth with UAVs in various crop yield levels, which should prove helpful in family farm for the modern agriculture.

The rapid and non-destructive monitoring of the canopy leaf nitrogen concentration (LNC) in crops is important for precise nitrogen (N) management. Nowadays, there is an urgent need to identify next-generation bio-physical variable retrieval algorithms that can be incorporated into an operational processing chain for hyperspectral satellite missions. We assessed six retrieval algorithms for estimating LNC from canopy reflectance of winter wheat in eight field experiments. These experiments represented variations in the N application rates, planting densities, ecological sites and cultivars and yielded a total of 821 samples from various places in Jiangsu, China over nine consecutive years. Based on the reflectance spectra and their first derivatives, six methods using different numbers of wavelengths were applied to construct predictive models for estimating wheat LNC, including continuum removal (CR), vegetation indices (VIs), stepwise multiple linear regression (SMLR), partial least squares regression (PLSR), artificial neural networks (ANNs), and support vector machines (SVMs). To assess the performance of these six methods, we provided a systematic evaluation of the estimation accuracies using the six metrics that were the coefficients of determination for the calibration (R²_C) and validation (R²_V) sets, the root mean square errors of prediction (RMSE_P) for the calibration and validation sets, the ratio of prediction to deviation (RPD), the computational efficiency (CE) and the complexity level (CL). The following results were obtained: (1) For the VIs method, SAVI(R₁₂₀₀, R₇₀₅) produced a more accurate estimation of the LNC than other indices, with R²_C, R²_V, RMSE_P, RPD and CE values of 0.844, 0.795, 0.384, 2.005 and 0.10 min, respectively; (2) For the SMLR, PLSR, ANNs and SVMs methods, the SVMs using the first derivative canopy spectra (SVM-FDS) offered the best accuracy in terms of R²_C, R²_V, RMSE_P, RPD, and CE, at 0.96, 0.78, 0.37, 2.02, and 21.17, respectively; (3) The PLSR-FDS, ANN-OS and SVM-FDS methods yield similar accuracies if the CE and CL are not considered, however, ANNs and SVMs performed better on calibration set than the validation set which indicate that we should take more caution with the two methods for over-fitting. Except PLS method, the performance for most methods did not enhance when the spectrum were operated by the first derivative. Moreover, the evaluation of the robustness demonstrates that SVM method may be better suited than the other methods to cope with potential confounding factors for most varieties, ecological site and growth stage; (4) The prediction accuracy was found to be higher when more wavelengths were used, though at the cost of a lower CE. The findings are of interest to the remote sensing community for the development of improved inversion schemes for hyperspectral applications concerning other types of vegetation. The examples provided in this paper may also serve to illustrate the advantages and shortcomings of empirical hyperspectral models for mapping important vegetation biophysical properties of other crops.

Unmanned aerial vehicle (UAV)-based remote sensing (RS) possesses the significant advantage of being able to efficiently collect images for precision agricultural applications. Although numerous methods have been proposed to monitor crop nitrogen (N) status in recent decades, just how to utilize an appropriate modeling algorithm to estimate crop leaf N content (LNC) remains poorly understood, especially based on UAV multispectral imagery. A comparative assessment of different modeling algorithms (i.e., simple and non-parametric modeling algorithms alongside the physical model retrieval method) for winter wheat LNC estimation is presented in this study. Experiments were conducted over two consecutive years and involved different winter wheat varieties, N rates, and planting densities. A five-band multispectral camera (i.e., 490 nm, 550 nm, 671 nm, 700 nm, and 800 nm) was mounted on a UAV to acquire canopy images across five critical growth stages. The results of this study showed that the best-performing vegetation index (VI) was the modified renormalized difference VI (RDVI), which had a determination coefficient (R²) of 0.73 and a root mean square error (RMSE) of 0.38. This method was also characterized by a high processing speed (0.03 s) for model calibration and validation. Among the 13 non-parametric modeling algorithms evaluated here, the random forest (RF) approach performed best, characterized by R² and RMSE values of 0.79 and 0.33, respectively. This method also had the advantage of full optical spectrum utilization and enabled flexible, non-linear fitting with a fast processing speed (2.3 s). Compared to the other two methods assessed here, the use of a look up table (LUT)-based radiative transfer model (RTM) remained challenging with regard to LNC estimation because of low prediction accuracy (i.e., an R² value of 0.62 and an RMSE value of 0.46) and slow processing speed. The RF approach is a fast and accurate technique for N estimation based on UAV multispectral imagery.

This paper evaluates the potential of integrating textural and spectral information from unmanned aerial vehicle (UAV)-based multispectral imagery for improving the quantification of nitrogen (N) status in rice crops. Vegetation indices (VIs), normalized difference texture indices (NDTIs), and their combination were used to estimate four N nutrition parameters leaf nitrogen concentration (LNC), leaf nitrogen accumulation (LNA), plant nitrogen concentration (PNC), and plant nitrogen accumulation (PNA). Results demonstrated that the normalized difference red-edge index (NDRE) performed best in estimating the N nutrition parameters among all the VI candidates. The optimal texture indices had comparable performance in N nutrition parameters estimation as compared to NDRE. Significant improvement for all N nutrition parameters could be obtained by integrating VIs with NDTIs using multiple linear regression. While tested across years and growth stages, the multivariate models also exhibited satisfactory estimation accuracy. For texture analysis, texture metrics calculated in the direction D3 (perpendicular to the row orientation) are recommended for monitoring row-planted crops. These findings indicate that the addition of textural information derived from UAV multispectral imagery could reduce the effects of background materials and saturation and enhance the N signals of rice canopies for the entire season.