Unmanned aerial vehicle (UAV)-based remote sensing provides a flexible, low-cost, and efficient approach to monitor crop growth status at fine spatial and temporal resolutions, and has a high potential to accelerate breeding process and improve precision field management. In this study, we discussed the use of lightweight UAV with dual image-frame snapshot cameras to estimate aboveground biomass (AGB) and panicle biomass (PB) of rice at different growth stages with different nitrogen (N) treatments. The spatial–temporal variations in the typical vegetation indices (VIs) and AGB were first investigated, and the accuracy of crop surface model (CSM) extracted from the Red Green Blue (RGB) images at two different stages were also evaluated. Random forest (RF) model for AGB estimation as well as the PB was then developed. Furthermore, variable importance and sensitivity analysis of UAV variables were performed to study the potential of improving model robustness and prediction accuracies. It was found that the canopy height extracted from the CSM (Hcsm) exhibited a high correlation with the ground-measured canopy height, while it was unsuitable to be independently used for biomass assessment of rice during the entire growth stages. We also observed that several VIs were highly correlated with AGB, and the modified normalized difference spectral index extracted from the multispectral image achieved the highest correlation. RF model with fusing RGB and multispectral image data substantially improved the prediction results of AGB and PB with the prediction of root mean square error (RMSEP) reduced by 8.33–16.00%. The best prediction results for AGB and PB were achieved with the coefficient of determination (r2), the RMSEP and relative RMSE (RRMSE) of 0.90, 0.21 kg/m2 and 14.05%, and 0.68, 0.10 kg/m2 and 12.11%, respectively. In addition, the result confirmed that the sensitivity analysis could simplify the prediction model without reducing the prediction accuracy. These findings demonstrate the feasibility of applying lightweight UAV with dual image-frame snapshot cameras for rice biomass estimation, and its potential for high throughput analysis of plant growth-related traits in precision agriculture as well as the advanced breeding program.

Remote estimation of flower number in oilseed rape under different nitrogen (N) treatments is imperative in precision agriculture and field remote sensing, which can help to predict the yield of oilseed rape. In this study, an unmanned aerial vehicle (UAV) equipped with Red Green Blue (RGB) and multispectral cameras was used to acquire a series of field images at the flowering stage, and the flower number was manually counted as a reference. Images of the rape field were first classified using K-means method based on Commission Internationale de l’Éclairage (CIE) L*a*b* space, and the result showed that classified flower coverage area (FCA) possessed a high correlation with the flower number (r² = 0.89). The relationships between ten commonly used vegetation indices (VIs) extracted from UAV-based RGB and multispectral images and the flower number were investigated, and the VIs of Normalized Green Red Difference Index (NGRDI), Red Green Ratio Index (RGRI) and Modified Green Red Vegetation Index (MGRVI) exhibited the highest correlation to the flower number with the absolute correlation coefficient (r) of 0.91. Random forest (RF) model was developed to predict the flower number, and a good performance was achieved with all UAV variables (r² = 0.93 and RMSEP = 16.18), while the optimal subset regression (OSR) model was further proposed to simplify the RF model, and a better result with r² = 0.95 and RMSEP = 14.13 was obtained with the variable combination of RGRI, normalized difference spectral index (NDSI _{(944, 758)}) and FCA. Our findings suggest that combining VIs and image classification from UAV-based RGB and multispectral images possesses the potential of estimating flower number in oilseed rape.

The advances of hyperspectral technology provide a new analytic means to decrease the gap of phenomics and genomics caused by the fast development of plant genomics with the next generation sequencing technology. Through hyperspectral technology, it is possible to phenotype the biochemical attributes of rice seeds and use the data for GWAS. The results of correlation analysis indicated that Normalized Difference Spectral Index (NDSI) had high correlation with protein content (PC) with R NDSI 2 = 0.68. Based on GWAS analysis using all the traits, NDSI was able to identify the same SNP loci as rice protein content that was measured by traditional methods. In total, hyperspectral trait NDSI identified all the 43 genes that were identified by biochemical trait PC. NDSI identified 1 extra SNP marker on chromosome 1, which annotated extra 22 genes that were not identified by PC. Kegg annotation results showed that traits NDSI annotated 3 pathways that are exactly the same as PC. The cysteine and methionine metabolic pathway identified by both NDSI and PC was reported important for biosynthesis and metabolism of some of amino acids/protein in rice seeds. This study combined hyperspectral technology and GWAS analysis to dissect PC of rice seeds, which was high throughput and proven to be able to apply to GWAS as a new phenotyping tool. It provided a new means to phenotype one of the important biochemical traits for the determination of rice quality that could be used for genetic studies.

The symptoms of the nutrient stress in plant canopies differ among different growth stages. It is a challenge to develop individual diagnosis models to evaluate the nutrient status for a specific growth stage in time. Therefore, this study encoded spatiotemporal information of plants in a single time-series model to evaluate the nutrient status of oilseed rape more efficiently. Specifically, we combined the convolutional neural network and long-short term memory (CNN-LSTM) to classify the oilseed rape crops according to their nutrient status. The model was validated on a large number of sequential images acquired from oilseed rape canopies at different growth stages during a two-year experiment. Different pre-trained CNNs were used to extract distinctive features from every time step of sequential images and then these features were considered as the input of LSTM to classify the oilseed rape into nine classes of nutrient statuses. We demonstrated that the LSTM outperformed the traditional machine learning method and the deep features showed better performance compared to hand-crafted features. The Inceptionv3-LSTM obtained the highest overall classification accuracy of 95% when tested on the dataset of 2017/2018 and it also provided a good generalization when using a cross-dataset validation, with the highest overall accuracy of 92%. Our proposed approach presents a pathway toward automatic nutrient status diagnosis during the whole life cycle of the plants and the LSTM technique would play an essential role in the near future for time-series analysis for precision agriculture.